part - b speech language pathology - aiish.ac.inaiish.ac.in/images/pdf/08-09slp.pdf · aiswarya...
TRANSCRIPT
STUDENT RESEARCH AT A.I.I.S.H. MYSORE
(ARTICLES BASED ON DISSERTATION DONE AT AIISH)
VOLUME VII: 2008-2009
PART - B
SPEECH LANGUAGE PATHOLOGY
Compiled by
Dr. Vijayalakshmi Basavaraj
Director
Dr. G. Malar
Reader in Special Education
All India Institute of Speech and Hearing
Manasagangothri, Mysore – 570 006
© 2010
A Publication of the All India Institute of Speech and Hearing
Under the title: “Student Research at A.I.I.S.H Mysore”
Articles based on dissertations done at AIISH: Vol. VII 2008-09
Telephone : 0821-2514449, 2515410, 2515805, 2515218, 2514313
Guest House : 2515786, Security (after office hours): 2514449
Fax : 0821- 2510515
E-mail : [email protected], [email protected]
Website : www.aiishmysore.org
Work Hours : 9.00 am to 5.30 pm – Monday through Friday
Holidays : Central Government Holidays
Price: Rs. 150/-
Published by Dr. Vijayalakshmi Basavaraj
Director, AIISH, Mysore
Foreword
AIISH presents with great satisfaction, the seventh volume of full length articles based on the
dissertation work done by our post graduate students in part fulfillment of their PG degrees in
Audiology, Speech Language Pathologist and Special Education (HI) for your reading.
This volume includes articles based on dissertations done by the post graduate students during
the year 2008-09. There are 44 articles in total . Part A comprises of 20 papers related to Audiology.
Several students have shown interest in the areas of developing diagnostic/screening tests, issues
related to understanding, evaluating and comparing the benefits of hearing aids/FM system. The
interest in the hot topics of Cochlear dead region, Auditory dys-synchrony, Vestibular Evoked
Myogenic Potential (VEMP), Post Auricular Muscle Response (PAMR) has continued.
Multifrequency, Multicomponent Tympanometry is a „fresh‟ topic in the list. Evoked potential testing
continues to attract our students. Part B comprises of 20 papers in the area of Speech-Language
Pathology. Our faculty have generated interest in the areas of swallowing disorders, sign language,
dementia and relevance of Yoga in Speech Language Pathology which is encouraging. Topics in the
area of voice assessment, analysis, issues related to language development and bilingualism, dyslexia
and aphasia continue to interest our students. Part C contains 4 papers of students of M.S Ed (HI).
These papers cover the area of teaching methods, curricular adaptation and attitudes of teachers. The
M.S.Ed (HI) program is not attracting more students. However, those who have enrolled are trained
and groomed well as master trainers. Even though there are only four papers in this section, it is
published as a separate section as the readership for each section may be different. To the best of our
knowledge, this is the first effort to publish articles based on M.S. Ed (HI) dissertations. Since AIISH
library is digitized, these articles will be available for review to all the researchers, in India and
abroad, in the area of Special education.
The titles of the articles are the titles of the dissertations. The first authors are the II M.Sc
(Aud), M.Sc (SLP) & M.S.Ed (HI) students of 2008-09 and the second authors are their respective
guides who have supervised and guided the research work. The AIISH faculty members who have
guided the dissertations have modified and edited the papers to bring it to the present shape to the best
of their abilities in spite of their busy academic schedules. Dr. G.Malar, Reader, Special Education
has put in great efforts to procure and compile the edited articles and has herself corrected english in
many of the articles. This is the highly appreciated. The neat formatting by Ms. N. Parimala in a very
short time is acknowledged. The unattended mistakes in print and references, if any, in spite of best
efforts put in is regretted.
You may please e-mail your valuable feedback about this volume to [email protected]
with the subject “Student research, Volume VII A/B/C, 2008-09”.
Dr. Vijayalakshmi Basavaraj
Director
Table of Contents
Sl.
No. Title
Page
No.
1. Effect of Pranayama in the Management of Stuttering – Aiswarya
Anand & Y. V. Geetha
1
2. Reliabilty of Perceptual Evaluation of Voice Using Cape-V Rating
Scale in Indian Context – Gupta Akansha & M. Pushpavathi
17
3. A preliminary investigation in to the cognitive abilities of persons
with stuttering using Simon and Stroop tasks – Amit K. & Y. V.
Geetha
34
4. Norms for Forced & Slow Vital Capacity in adult Dravidian
Population – Annapurna S. B. & N. Sreedevi
47
5. Manual for Adult Fluent Aphasia Therapy – in Kannada (MAFAT-K)
– Chaitra S. & S. P. Goswami
61
6. Normative Nasalance Value in Malayalam Language – Devi T.R. &
Pushpavathi M.
67
7. Symbolic Play and Language: Its Relationship in Late Talkers –
Devika M. R. & N. Swapna
83
8. Subtyping of Children with Developmental Dyslexia: Implications
through Dual Route Cascaded (DRC) Model in the Indian Context –
Gnanavel K. & Jayashree C. Shanbal
99
9. Dyslexia Assessment Profile for Indian Children (DAPIC) – Kuppuraj
S. & Jayashree C. Shanbal
115
10. Clinical Protocol for Assessment of Swallowing in Adults - (CP-
ASA) – Meera Priya C. S. & R. Manjula
130
11. Comprehensive Language Assessment Tool for Children (3-6 years) –
Navitha U. & K. C. Shyamala
140
12. The Order of S, O, V Structures in Sign Language Users With
Hearing Impairment: Influence of Verbal Native Language? – Pallavi
Malik & R. Manjula
156
13. Articulatory Acquisition in Kannada Speaking Urban Children: 3-4
Years – Prathima S. & N. Sreedevi
172
14. Normative Nasalance Value in Hindi Language – Pravesh Arya & M.
Pushpavathi
188
15. Metalinguistic Abilities in Children with Developmental Dyslexia:
Implications for Reading and Writing – Priya M. B. & R. Manjula
200
16. Language Proficiency Questionnaire: An Adaptation of LEAP-Q in
Indian Context – Ramya Maitreyee & S. P. Goswami
214
17. Paraphasias in Bilingual Aphasia - Ridhima Batra & K. C. Shyamala 233
18. Meta-Phonological Abilities in Monolingual and Bilingual Children:
A Comparative Study – Samasthitha S. & S. P. Goswami
249
19. Dementia Assessment Battery – Kannada – Sunil Kumar Ravi & K. C.
Shyamala
262
20. Vocal registers in Classical Carnatic Singers: An inquiry using
Electroglottography – Sweety Joy & K. Yeshoda
276
Effect of Pranayama on Stuttering
1
Effect of Pranayama in the Management of Stuttering
Aiswarya Anand. & Y.V. Geetha*
Abstract
The present study was mainly aimed at investigating whether traditional prolongation
technique when used in combination with Pranayama would be more effective in the management of
stuttering than the prolongation technique when used alone and if it has a long term effect on the
maintenance of fluency. PWS considered in the study were randomly distributed to the treatment
programs of prolongation technique and prolongation technique with Pranayama. The experimental
group involving Pranayama with prolongation therapy consisted of five subjects and the control group
provided only Pranayama consisted of four subjects. The study focused on finding the outcomes in the
various evaluations using Stuttering Severity instrument (SSI), Treatment efficacy scale for fluency
disorders, Situational assessment checklist for PWS, aerodynamic measures for both the treatment
groups of PWS, during the pre therapy, post therapy and one month post therapy follow up conditions.
The results reveal that within both the groups the parameters such as avoidance, anxiety, attitudinal
changes, listener’s reaction, satisfaction, and total scores of treatment efficacy scale showed
significance within both experimental and control groups. The situational assessment checklist showed
significant difference within the control group alone. In the aerodynamic measurement MVV was
better for the control group and SVC did not show much significance between the groups.
Introduction
Theories that have attempted to explain the causes of stuttering have often echoed the
prevailing beliefs of the time. There have been frequent shifts in the view points about the
etiological factors for the onset and development of stuttering from physiogenic to
psychogenic and back and forth, ultimately to the combination of both. For example, in the
mid 20th century there was a trend to believe that many diseases were psychosomatic in
origin, that is, they were caused by psychological factors such as anxiety. Similar shifts in the
focus of management issues in stuttering have been noticed over the decades. The
contemporary treatment for stuttering can take widely differing forms, both in the emphases
and the procedures used to bring about the alleviation of the problem. Most of the fluency
management therapies involve a cluster of different components and is difficult to determine
which are crucial to effective treatment.
Prolongation is one of the oldest and most commonly adopted procedures in the
management of stuttering with fairly good treatment outcome. Gifford (1940) made extensive
use of prolongation in her therapy programs. There have been contemporary practitioners
since then. Both in therapy and in research in related fields there was resurgence of interest in
relaxation (Slorach, 1971). Dalton & Hardcastle (1977) expressed the opinion that relaxation,
per se, was not effective. However, they recommended its inclusion as part of an overall
remediation program.
________________________________ * Professor of Speech Language Sciences, All India Institute of Speech and Hearing, Mysore, Indi
email: [email protected]
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
2
Perkins (1979) focuses on speech act in his definition of stuttering as a “disco-
ordination of phonation with articulation and respiration” and in his treatment program he
worked directly on achieving fluency through management of breath stream in order to initiate
voice with a gentle onset and maintain airflow throughout the phase.
Although the use of relaxation with the techniques for stuttering treatment is widely
implemented, the use of Yoga, one of the ancient Indian methods concerning regulation of
breathing and management of anxiety, has not been explored. Yoga aims to improve people's
inner tranquility and free them from fears and anxieties. Since it is known that stuttering
includes an element of anxiety and fear, Yoga can help reduce this. Yoga, a science not less
than 5000 years old, has addressed a normal man to move towards higher states of harmony
and peace both as an individual and also as a social being. Extensive research on yoga therapy
over the last few decades have brought out the usefulness of yoga in dealing with these
ailments as an effective adjunct to medical management and also for long term rehabilitation.
Pranayama, one of the Yoga techniques, focuses on regulating breathing. Pranayama is
derived from two Sanskrit words - Prana (life force) and Ayama (control). Therefore, in its
broadest description, Prananyama would mean the control of the flow of life force. During
breathing for Pranayama, inhalation (Puraka) stimulates the system and fills the lungs with
fresh air; retention (Kumbhaka) raises the internal temperature and plays an important part in
increasing the absorption of oxygen; exhalation (Rechak) causes the diaphragm to return to
the original position and air full of toxins and impurities is forced out by the contraction of
inter-costal muscles. These are the main components leading to Pranayama which massage the
abdominal muscles and tone up the working of various organs of the body. There are no
studies in either the Indian or the Western literature showing the effectiveness of Pranayama
in stuttering therapy. A lot of people have benefited from the practice of Pranayama for
various psychogenic and or physiogenic disorders and this preliminary attempt might throw
some light on its efficacy in the management of ever alluding disorder of fluency – stuttering.
Need for the study
The classical means for determining the value of a component of therapy is to include
it in experimental treatment and compare the outcome with that produced by treatment
without the component. Thus, a treatment encompassing the traditional practice of Pranayama
is to be determined to find its effectiveness as a treatment component in the management of
stuttering. Also, considering the various benefits of Pranayama as mentioned below, it is
important that its effect be studied.
1. Release of acute and chronic muscular tensions around the heart and digestive organs
2. Role in helping sufferers of respiratory illnesses such as asthma and emphysema to
overcome the fear of shortness of breath
3. Role in increasing lung capacity, proper nervous stimulus to the cardio-vascular
system, dramatic reduction in emotional and nervous anxiety
4. Improvement in detoxification with increased exchange of carbon dioxide and oxygen
5. Amplification of the auto immune system by increased distribution of energy to the
endocrine system
Effect of Pranayama on Stuttering
3
6. Calming the mind and integration of the mental/physical balance, its contribution to both
vitality and relaxation through this single practice
Aim of the study
The present study aims at studying the role of Pranayama in the treatment of stuttering
and also its long term effect in comparison with a group which does not practice Pranayama.
Yoga has been practiced in India for thousands of years for better control of mind and body.
Pranayama being one of the yogic practices which is very effective in breath regulation and
also control of anxiety and tension could be effective in the management of stuttering. The
current experiment aims to investigate how it applies to stuttering.
Objectives of the study
1. To investigate whether Pranayama is more effective in the management of stuttering than
the traditional prolongation technique when used in combination with it.
2. To study if the effect of Pranayama practices with prolongation when compared to
prolongation alone as a treatment technique is maintained during the period after the
termination of therapy.
3. To see if the severity of stuttering has any effect in the treatment out come with
Pranayama.
Method
The present study aimed at evaluating the changes seen in stuttering in a group of
PWS while completing a comprehensive but brief and simple Pranayama program along with
prolongation (experimental group) as against a group which follow the traditional
prolongation technique (control group).
A. Subject selection criteria
The subjects were divided into two groups and the summary is given in Table 1
Table 1: Subjects distributed in the two groups
Particulars Group A Group B
Number 5 4
Age (in years) 18 - 30 18 - 30
Gender Male Male
Severity Moderate - Severe Moderate - Severe
Group A (Experimental Group) - Pranayama as treatment program along with the
prolongation treatment program.
Group B (Control Group) - Prolongation treatment
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
4
Inclusion criteria
The subjects should not have undergone speech therapy in the past or should have had a
relapse after treatment, exhibiting more than moderate degree stuttering at the time of
enrollment
They should not be practitioners of Yoga
All the participants of the study have to be diagnosed by qualified speech language
pathologist as having moderate to severe stuttering, based on SSI
B. Instrumentation/Materials
The following materials were used for the study:
1. Stuttering Severity Instrument (SSI) developed by Glydon D. Riley (1994).
2. MP3 player (INVION-QBP38-IUS)
3. Treatment efficacy scale for fluency disorders (self rating) by Geetha, Sangeetha &
Anjana (2007)– see appendix A
4. Situational assessment checklist for PWS (self rating) – see appendix B
5. RMS Helios 501(Recorders and Medicare Systems)
6. 300 word passage for reading in Kannada (Jayaram & Savithri, 1985). It incorporates all
the phonemes in Kannada, with their respective frequency of occurrence
7. Set of questions for conversation
8. Specific topics given for obtaining narrative samples.
C. Procedure
All the subjects were evaluated individually using the above tests during pre therapy,
immediately after therapy and during one month post therapy follow-up. Subject’s written
consent was taken before starting the program, after briefing about the technique adopted.
1. Perceptual assessment
Subject’s reading, spontaneous speech and conversation were recorded using an MP3
player. The recordings were subjected to offline perceptual analyses to rate the severity of
dysfluency. The reading sample of 300 word passage, monologue and conversation sample
was recorded. Treatment efficacy scale for fluency disorders was used to find the efficacy of
treatment. This is an unpublished project titled “Treatment efficacy and variables for
stuttering management” (Geetha, Sangeetha & Anjana, 2007). This is a self report assessment
procedure which included separate measures of frequency of stuttering, duration, secondary
behaviors, confidence, avoidance behaviors, anxiety features, attitudinal changes, naturalness
of speaking, listeners reaction, satisfaction with treatment, self monitoring skills, feeling about
generalization and maintenance of fluency. A 5- point descriptive scale is used to allow the
subject to show how often or how much each of the aspects applies to them. Situational
assessment checklist for PWS was also employed which again the subjects had to rate on a
Effect of Pranayama on Stuttering
5
five point scale (given in appendix B). Subjects were assessed at the beginning (day 1), at the
end (day 12) of the intervention and one month follow up after the 12 day treatment period.
The recorded speech samples of the subjects were subjected to perceptual rating by 5
qualified SLPs on a 3- point scale (3- good, 2- average and 1- poor). This rating was done for
the parameters rate of speech, continuity, effort, stress and articulation. The post therapy
speech samples of experimental and control group were randomized and given to five judges
for perceptual evaluation and rating. This was to check the inter judge reliability in terms of
therapy outcomes using the two techniques.
2. Aerodynamic measurements
Respiratory measurements included the measurement using RMS Helios 501
(Recorders and Medicare Systems) - a window based program. The parameters recorded
included, slow vital capacity (SVC) and maximum voluntary ventilation (MVV). SVC is the
maximum volume of air that can be exhaled slowly after slow maximum inhalation, measured
in liters. MVV is a measure of the maximum amount of air that can be inhaled and exhaled in
one minute and it is measured in liters/minute.
D. Therapy Program
a. Experimental Group
The treatment program consisted of an integrated package comprising teaching
sessions and practice sessions using prolongation technique with Pranayama. It was
administered in the form of a 12-day outpatient course, 45 minutes each day, spread over a
period of four weeks, being interrupted by a 2-day weekend break. The course was given to a
group of 5 patients individually who were randomly assigned to the experimental group.
The Pranayamas taught included Vibhagiya Pranayama and Nadisuddi (Nagendra &
Nagaratna, 2007).
1. Vibhagiya svasana (sectional breathing): This is a preparatory breathing practice for
Pranayama. It corrects the wrong breathing pattern and increases the vital capacity of the
lungs. It has four sections and 4 steps; each practiced 5 times in each session.
Abdominal (diaphragmatic) breathing
Thoracic breathing
Clavicular breathing
Full yogic breathing.
2. Nadisuddi Pranayama: Practice involves six steps of Nadisuddi Pranayama which is
repeated 9 times: (i). Closing the right nostril with the right thumb and exhaling
completely through the (left) nostril (ii). Inhaling deeply through the same left nostril. (iii).
Closing the left nostril with ring and little finger and releasing the right nostril (iv).
Exhaling slowly and completely through the right nostril (v). Inhaling deeply through the
same (right) nostril (vi). Closing the right nostril and exhaling through the left nostril.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
6
The session involved practice for each of reading, conversation and monologue for the
control group using the prolongation technique and for the experimental group, the session
was divided to encompass Pranayamas before each of the activity. The session concluded with
nadanusandana, which can be attributed to a vocal exercise. Nadanusandana encompassed the
phonation of /a/, /u/ and /m/.
b. Control group
The treatment program comprised of teaching sessions and practice sessions using
prolongation technique. It was administered in the form of a 12-day outpatient course, 45
minutes each day, spread over a period of four weeks, being interrupted by a 2-day weekend
break. The course was given to a group of 4 patients individually who were randomly
assigned to the control group.
Results and Discussion
The present study was aimed at finding difference if any was present in the therapy
programs using prolonged speech for the control group when compared to prolonged speech
in combination with Pranayama for the experimental group in the treatment of stuttering. The
study also aimed at answering the research question related to the long term effect of the
treatment program. The two groups of PWS who underwent treatment were assessed based on
Stuttering Severity Instrument (SSI), Treatment efficacy scale for fluency disorders,
Situational assessment checklist for stuttering, perceptual evaluation by judges and
aerodynamic evaluations before therapy (PT), post therapy (PoT) & FU (1 month post therapy
follow up). Results of these assessments are discussed under different headings below.
Friedman’s test was done to find if there existed any significant difference between the two
groups and further Wilcoxon's test was done to find if there existed any differences within the
two groups
I. Perceptual assessment
a. Stuttering severity instrument (SSI)
Table 2: The severity of disfluencies
Groups Pre therapy Post therapy FU
SSI Score Severity SSI Score Severity SSI Score Severity
E - 1 24.00 Moderate 16.00 Very mild 13.00 Very mild
E - 2 22.00 Moderate 14.00 Very mild 18.00 Mild
E - 3 22.00 Moderate 10.00 Very mild 12.00 Very Mild
E - 4 37.00 Very Severe 21.00 Mild 21.00 Mild
E - 5 27.00 Moderate 19.00 Mild 13.00 Very Mild
C - 1 23.00 Moderate 6.00 Very mild 8.00 Very Mild
C - 2 26.00 Moderate 13.00 Very mild 17.00 Mild
C - 3 33.00 Severe 26.00 Moderate 23.00 Moderate
C - 4 30.00 Moderate 22.00 Moderate 15.00 Mild
Subject’s reading, spontaneous speech and conversation were recorded using an MP3
player. The recordings were subjected to offline perceptual analyses to rate the severity of
Effect of Pranayama on Stuttering
7
dysfluency on SSI. Table 2 gives the summary of the severity of disfluencies in the three
conditions of pre therapy, post therapy and one month post therapy.
1) Frequency scores of SSI: There is no much difference in the SSI Frequency scores in the
three conditions of pre therapy, post therapy and follow up.
2) Duration scores of SSI: There is no much difference in the scores between the two
groups for the duration scores in SSI.
Figure 1: Mean ratings on various parameters of the treatment efficacy scale for the experimental
group
Parameters
Mai
ntai
n.
Selfm
onL.
Rea
c
Nat
u.
Atti.
Anxi
.
Avoi
d.
Con
f.
S.be
h
Dur
.
Freq
.
Mea
n R
atin
gs -
Con
trols
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
.5
0.0
Conditions
Pre
Post
After a month
Figure 2: Mean ratings on various parameters of the treatment efficacy scale for the control group
[Freq = Frequency of stuttering; Dur = Duration of stuttering; S.beh = Secondary Behaviors; Conf = Confidence
in Speaking; Avoid = Avoidance Features; Anxi = Anxiety Features; Atti = Attitudinal Changes; Natu =
Naturalness of Speaking; L.Reac = Listeners Reaction to speech ; Selfmon = Self monitoring Skills ; Maintain =
Feeling about generalization and maintenance about fluency; Pre = Pre therapy; Post = Post therapy; After a
month = A month after post therapy]
3) Physical concomitant scores in SSI: There is no much difference in the scores between
the two groups for the physical concomitant in SSI.
4) Total scores of SSI: It is clear that there is no significant difference in total scores on SSI.
b. Treatment efficacy scale for fluency disorders was used to find the efficacy of
treatment.
Treatment efficacy scale for fluency disorders was used to find the efficacy of treatment.
The scores are from 1 to 5, where 1 is for more severe condition and five for the less
severe condition.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
8
1) Frequency of stuttering: From the mean and standard deviation obtained it is clear that
there is no significant difference in the frequency of stuttering in the treatment efficacy
scale between the experimental and control groups.
2) Duration of stuttering: There exists a significant difference in the conditions within the
control group. The pairs of conditions are pre therapy – post therapy and pre therapy – 1
month post therapy follow up conditions (Figures 8 and 9) but not within the control
group.
3) Secondary Behavior: Figures 8 and 9 shows that there exists no significant difference
between the two groups in the three conditions.
4) Confidence in speaking: From the mean and standard deviation obtained (Figures 8 and
9), it is clear that there is no significant difference in confidence in speaking in the
treatment efficacy scale.
5) Avoidance behaviors: The results show that pre therapy - 1 month post therapy follow up
is the pair for which significant difference is there in the experimental group. In the
control group, the pairs of conditions for which there exists difference is there are pre
therapy - post therapy and pre therapy - 1 month post therapy follow up conditions
respectively.
6) Anxiety features: From Figures 8 and 9, it can be seen that there exists no significant
difference in the scores.
7) Attitudinal changes: From figure 8 and 9 it can be noted that there exists no significant
difference between the two groups in the three conditions. Pre therapy - post therapy
conditions and the pre therapy - 1 month post therapy follow up conditions are the pairs of
conditions for which significant difference is present in both the experimental and control
groups.
8) Naturalness of speaking: There exists no significant difference between the two groups in
the three conditions (Figure 8 and 9). The study by Ingham and Onslow (1985) describes
two studies that illustrate the utility of listener ratings of speech naturalness for measuring
and modifying speech naturalness during a stuttering therapy program.
The program
involved 5 adolescent PWS who were receiving an intensive treatment incorporating a
prolonged speech procedure. The
results showed that each subject's speech naturalness
ratings could be modified toward a target level of speech naturalness. In the present study,
the naturalness scores are better for the experimental group than the control group. The
higher scores of the experimental group could be attributed to the therapy program which
is used which takes into account the breathing coordination and the relaxation aspect.
9) Listeners reaction: Figures 8 and 9 reveal that in the pre therapy condition, the scores are
lower for the experimental group, in post therapy condition, the scores are lower for
experimental group and in 1 month post therapy follow up condition, scores are higher for
experimental group. This suggests that in the long term the listeners’ reaction has
improved better for PWS who was in the experimental group. However, the difference is
very slight.
Effect of Pranayama on Stuttering
9
10) Satisfaction with treatment: The results revealed that pre therapy - post therapy and pre
therapy - 1 month post therapy follow up are the pairs of conditions for which significant
difference is there in the experimental group. In the control group, the pairs of conditions
for which there exists difference are pre therapy - 1 month post therapy follow up
condition.
11) Self Monitoring skills: The results revealed that no significant difference existed in the
experimental and the control groups.
12) Feeling about generalization and maintenance of fluency: From figures 8 and 9 it can
be seen that there exists no significant difference within in the experimental group. A
significant difference in the conditions within the control group was obtained. In the
control group, the pairs of conditions for which there exists difference were pre therapy -
post therapy and pre therapy - 1 month post therapy follow up conditions respectively.
13) Total scores of the treatment efficacy scale for stuttering: Pre therapy - post therapy ;
pre therapy - 1 month post therapy follow up were the pairs of conditions for which
significant difference existed in the experimental group and pre therapy - post therapy and
pre therapy - 1 month post therapy conditions in the control group.
c. Inter judge reliability for the perceptual evaluation by judges
The recorded speech samples of the subjects were subjected to perceptual rating by 5
qualified SLPs on a 3- point scale (3- good, 2- average and 1- poor). The post therapy speech
samples of experimental and control group were randomized and given to five judges for
perceptual evaluation and rating. There was good reliability obtained for all the parameters.
d. Situational assessment
Situational assessment checklist for PWS was also employed which again the subjects
had to rate on a five point scale. In the experimental group pre therapy - post therapy and pre
therapy - 1 month post therapy conditions were the conditions having significant difference.
In the control group, the pairs of conditions for which there existed a significant difference
were pre therapy - post therapy and pre therapy - one month post therapy conditions.
2. Aerodynamic measurements
Aerodynamic measurements were done during pre therapy; post therapy and 1 month
post therapy follow up condition to see if there were any differences in the experimental and
control groups. Respiratory measurements included the measurement using RMS Helios 501 -
a window based program. The parameters recorded included, slow vital capacity (SVC) and
maximum voluntary ventilation (MVV).
Slow Vital Capacity (SVC): There is no significant difference in the mean SVC scores for
either the experimental or the control group for the three conditions of pre therapy, post
therapy and one month post therapy follow up.
Expiratory Reserve Volume (ERV): There exists no significant difference between the
two groups in the three conditions. The mean scores of the experimental group are slightly
higher for all the three conditions than the control group.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
10
Inspiratory Reserve Volume (IRV): The results revealed no significant difference within
the experimental condition and a significant difference in the conditions within the control
group. The IRV scores are better for the experimental group than the control group for the
three conditions of pre therapy, post therapy and 1 month post therapy follow up
conditions.
Tidal Volume (VT): There is no much difference in the mean scores of the tidal volume
for the two groups for the three conditions. The results of Friedman test shows no
significant difference between the groups in the experimental condition and in the control
condition there was a significant difference. The pairs of conditions found to have
significant difference were between pre therapy - post therapy and pre therapy - 1 month
post therapy follow up.
The study by Story & Alphonso (1996) compared the pre and post treatment
measurements of kinematics of fluent speech of PWS. The study reports changes in acoustic,
respiratory, laryngeal
and articulatory kinematics of 3 males who stutter, following
participation in a version of the Hollins Precision Fluency Shaping Program. These results
show that behavioral treatment can produce significant changes in the fluent speech of persons
who stutter with respect to respiration, laryngeal valving and articulation.
5. Maximum Voluntary Ventilation (MVV)
The results show that pre therapy - post therapy and pre therapy - one month post
therapy were the pairs of conditions for which significant difference was there. There was
difference in the scores on the parameters of MVV between the two groups where MVV is
better for the control group in the post therapy condition. MRF in the pre therapy condition is
higher for the experimental group. There is no difference in scores for the other parameters
between the two groups. There is no significant difference between the two groups in the three
conditions. The pairs of conditions for which significant difference existed were post therapy -
1 month post therapy follow up and pre therapy - 1 month post therapy follow up.
Tables 3 and 4 shows the experimental and control groups respectively and the
conditions within those groups for which significant difference existed.
Table 3: The results of the different parameters significant (from Friedman’s test) and the
pairs of conditions significant (from Wilcoxon’s test) in the experimental group.
Parameters Experimental
group
Conditions for which significant
difference is there
Avoidance 6.61* PT & 1 month PoT
Anxiety 7.53* PT & PoT ; PT & 1 month PoT
Attitude 6.53* PT & PoT ; PT & 1 month PoT
Naturalness 7.60* PT & PoT ; PT & 1 month PoT
Listeners reaction 7.60* PT & PoT ; PT & 1 month PoT
Satisfaction 8.31* PT & PoT ; PT & 1 month PoT #
Total 8.40* PT & PoT ; PT & 1 month PoT #
Effect of Pranayama on Stuttering
11
SSI Frequency 8.40* PT & PoT ; PT & 1 month PoT #
SSI Duration 9.33* PT & PoT ; PoT & 1 month PoT #
SSI Physical concomitant 9.33* PT & PoT ; PT & 1 month PoT #
SSI Total 7.89* PT & PoT ; PT &1 month PoT #
Situational assessment 8.40* PT & PoT ; PT &1 month PoT #
[SSI = Stuttering Severity Instrument Scores; * = significance at p< 0.05 (Friedman’s test); PT = Pretherapy ;
PoT = Post Therapy ; 1 month PoT = 1 month post therapy; # = significance at p< 0.05 (Wilcoxon’s test)]
Table 4: The results for the different parameters for the three conditions that are significant
(from Wilcoxon’s test) in the control group.
Parameters Control
group
Conditions for which significant
difference is there
Duration 6.53* PT & PoT ; PT & 1 month PoT
Avoidance 7.53* PT & PoT ; PT & 1 month PoT
Anxiety 7.53* PT & PoT ; PT & 1 month PoT
Attitude 6.50* PT & PoT ; PT & 1 month PoT
Listeners reaction 6.50* PT & PoT
Satisfaction 7.00* PT & 1month PoT
Maintenance 8.00* PT & PoT ; PT & 1 month PoT
Total 6.50* PT & PoT ; PT & 1 month PoT
SSI Physical concomitant 6.61* PT & 1 month PoT
SVT 6.50* PT & PoT ; PT & 1 month PoT
MVV 7.60* PT & PoT ; PT & 1 month PoT
MVT 6.50* PoT&1 month PoT; PT & 1 month PoT
Situational assessment 6.50* PT & PoT ; PT & 1month PoT
[; SSI = Stuttering Severity Instrument Scores; SVC = slow vital capacity; ERV = expiratory reserve volume;
IRV = Inspiratory Reserve Volume; SVT = Tidal Volume; MVV = Maximum Voluntary Ventilation; MVT =
MVV maneuver tidal volume; * = significance at p< 0.05 (Friedman’s test); PT = Pretherapy ; PoT = Post
Therapy ; 1month PoT = 1 month post therapy; # = significance at p< 0.05 (Wilcoxon’s test)]
It can be seen that there was no significant difference between the two groups for the
perceptual evaluations done. However, within the experimental and control groups there was
significant difference in the parameters of SSI. The results of the Treatment efficacy scale for
Fluency Disorders showed parameters to have significant difference within the experimental
and control groups. The inter judge reliability was however good. The situational assessment
checklist also showed significant difference within the various conditions. The aerodynamic
evaluations showed that for the parameters; MVT = MVV maneuver tidal volume, significant
difference was present. MRF scores in the pre therapy condition were higher for the
experimental group. MVV scores were lower in the experimental group in 1 month post
therapy follow up and higher in the pre therapy and post therapy conditions.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
12
Summary and Conclusions
The current study was aimed at investigating whether prolongation technique when
used in combination with Pranayama (experimental condition) would prove to be more
effective in the management of stuttering than the prolongation technique (control condition)
when used alone. It was also aimed to study if the effect of Pranayama practiced with
prolongation when compared to prolongation alone as a treatment technique has a long term
effect on the maintenance of fluency. The study focused on finding the outcomes in the
various evaluations using Stuttering Severity instrument (SSI) developed by Glydon D. Riley
(1980), Treatment efficacy scale for fluency disorders (self rating) by Geetha, Sangeetha and
Anjana (2007), Situational assessment checklist for PWS (self rating), aerodynamic measures
using RMS Helios 501 for both the treatment groups of PWS, during the pre therapy, post
therapy and one month post therapy follow up condition. Appropriate statistical analysis was
done and the results of the study can be concluded as follows:
It can be concluded from the study that there was no significant difference between the
two groups for the perceptual measures. However, within the experimental group significant
difference was obtained for the parameters of frequency, duration and total scores of SSI for
the experimental group and for physical concomitant scores significance was obtained within
both the groups. The parameters avoidance, anxiety, attitudinal changes, listener’s reaction,
satisfaction, and total scores of treatment efficacy scale showed significance within both
experimental and control groups. The inter judge reliability was good for all the fluency
parameters rated. The situational assessment checklist showed significant difference within
the control group and not in the experimental group. In the aerodynamic measurement, MVV
was better for the control group and SVC did not show much significance between the groups.
The objective of finding the role of treatment program in maintenance of fluency reveals that
in all parameters for which significant difference was obtained, the significance was also
found in the one month post therapy when compared with pre therapy in the experimental and
control groups.
The attempt to find if the severity of stuttering has any effect in the treatment outcome
with Pranayama revels that in post therapy condition, 60% of PWS in experimental group had
very mild stuttering and 40% had mild stuttering which was maintained in the one month
follow up evaluation. In the control group, 50% of the PWS had very mild stuttering and 50%
of PWS had moderate stuttering. In the one month post therapy condition in the control group,
25% had very mild stuttering, 25% had moderate stuttering and 50% had mild stuttering. This
reveals that the severity of stuttering had an effect which could be attributable to the treatment
delivered.
Implications
The present study throws light on the fact that Pranayama, a widely practiced exercise
could be encompassed in the speech therapy program for individuals with stuttering.
Effect of Pranayama on Stuttering
13
Limitations of the study
Limitation of subject availability was the major drawback of the study because of time
constraints. There were drop outs in the subjects selected for the experimental and control
groups which limited the total number of subjects. Due to ethical issues, Pranayama alone was
not tried in the experimental group subjects. Follow up maintenance was assessed after only
one month post therapy due to time constraints.
References
Dalton & Hardcastle (1977). Relaxation and desensitization. In R. Ham (Eds.), Techniques of
Stuttering Therapy. Prentice-Hall, Inc., Englewood Cliffs, New Jersey 07632.
Geetha, Y. V., Sangeetha, M. & Anjana, R. (2007). Treatment efficacy in stuttering.
Unpublished ARF Project.
Gifford, M. (1940). Correcting Nervous Speech Disorders. Englewood Cliffs, N.J.: Prentice-
Hall.
Ingham, R.J. & Onslow, M. (1985). Measurement and Modification of Speech Naturalness
during Stuttering Therapy. Journal of Speech and Hearing Disorders, 50, 261-281.
Nagendra, H.R. & Nagarathna, R. (2007). Yoga (Practices - Breathing Practices &
Pranayama, Sithilikarana & Asanaa, Kriyas, Mudras, Bandhas, Om & Cyclic
meditations & Yoga counseling. India: Swamivivekananda Yoga Prakashana.
Perkins, W.H. (1979). Respiration and Phonation. In R. Ham (Eds.), Techniques of Stuttering
Therapy. Prentice-Hall, Inc., Englewood Cliffs, New Jersey 07632.
Riley, G. (1994). Stuttering Serverity Instrument for Children and Adults (3rd ed.). Austin,
TX: PRO-ED.
Slorach, N. (1971). Twenty years of stuttering therapy. Journal of Australian College of
Speech Therapists, 21: 19-23.
Story, R.S. & Alfonso, P.J. (1996). Pre- and Post treatment Comparison of Kinematics of
Fluent Speech of Persons Who Stutter, Journal of Speech and Hearing Research, 39,
991-1005.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
14
Appendix A
Treatment Efficacy Scale for Fluency Disorders
Name: No: Age/Sex: Date:
Ph. No: e-mail id: Address:
Clinician: Supervisor:
General information
1. How many therapy sessions did you take at AIISH? 1: <5; 2: 5-10; 3:10-20; 4: 20-30; 5: >30
2. What was the therapy technique taught to you?
3. Which technique did you find most effective?
4. Has therapy helped you control your stuttering? 1: No; 2: Yes
5. After therapy how is your stuttering? 1: increased; 2: Same; 3: Decreased
6. How much do you stutter now? (1: 0%; 2: 5-10%; 3: 10-25%; 4: 25-50%; 5: 50-75%; 6: > 75%)
7. How do you rate your stuttering severity when you first visited us?
1: Severe; 2: Moderately Severe; 3: Moderate; 4: Mild; 5: Very Mild
8. Have you taken any other treatment for your stuttering? 1: Yes; 2: No. If Yes, specify
9. Do you think you can improve some more by taking our help again? 1: No; 2. Yes
10. Has your attitude toward stuttering changed since you attended therapy at AIISH? 1: No; 2: Yes
I. Frequency of stuttering
1. I have problems in speaking very often, more than three times a day
2. I face the problem at least two or three times a day
3. I face the problem more than five times per week on an average
4. I face the problem two to three times a week
5. I face the problem occasionally once or twice in a week or two
II. Duration of stuttering
1. Often I get stuck during my speech for a long time
2. I get stuck on particular sounds or words for more than five to ten seconds
3. I get stuck for long duration occasionally during speech
4. I get stuck occasionally for very short durations
5. I do not ever get stuck in my speech for any perceptible duration now
III. Secondary behaviours
1. My secondary behaviours are severe and painful to look at
2. My secondary behaviours are very distracting
3. My secondary behaviours are distracting
4. My secondary behaviours are not noticeable unless looking for
5. I do not have any secondary behaviours now
IV. Confidence in speaking
1. I’m not at all confident while speaking
2. Most of the time I’m not confidant while speaking
3. There are some situations where I am not confidant while speaking
4. I am confidant to speak in most situations
5. I am confidant to face any situation
V. Avoidance behaviours
1. I avoid speaking situation every time I possibly can
2. I try to avoid a speaking situation most of the time
3. I try not to avoid a situation, but sometimes avoid it
4. I don’t try to avoid any speaking situation, but sometimes I feel like doing so
5. I never try to avoid any speaking situation
Effect of Pranayama on Stuttering
15
VI. Anxiety features
1. I feel extremely tensed/ scared and anxious whenever I have to speak
2. I feel tensed/scared and anxious in most of the situations when I have to speak
3. I get tensed/scared in some situations and in others I am able to speak easily
4. I get anxious only on few words, but can keep myself calm in most of the situations
5. I remain calm and relaxed in all speaking situations
VII. Attitudinal changes
1. I feel very inferior to others because of my speech
2. I have come to terms with my problem but still feel low most of the time
3. I tell myself that I am not inferior but sometimes cannot help feeling so
4. I have begun to realize I’m not inferior to anybody
5. I have understood my problem well and I know I’m not inferior to anybody
VIII. Naturalness of speaking
1. My speech is very unnatural/monotonous most of the time
2. My speech is unnatural in most situations
3. My speech is unnatural in some situations
4. My speech sounds natural in most situations
5. My speech sounds natural in almost all situations
IX. Listeners’ reaction to your speech
1. I feel all listeners are very impatient toward my speech
2. Most listeners laugh at me and tease/comment on my speech
3. I feel most listeners are sympathetic and complete the sentences for me
4. I perceive subtle reactions from the listeners to my speech sometimes
5. I do not notice any negative reactions to my speech from the listeners now
X. Satisfaction with treatment
1. I am dissatisfied; therapy offered did not satisfy me at all
2. Treatment was not satisfactory and not of much help
3. Treatment has helped reduced my problem to a small extent
4. Treatment was good, has helped me reduce my problem to a large extent
5. Treatment was excellent, has helped me completely overcome my problem
XI. Self monitoring skills
1. I am not able to monitor my speech fluency at all
2. I am not able to control my stuttering in most situations
3. I can monitor my speech fluency in most situations
4. I am able to monitor my speech except in occasional situations
5. I am able to monitor and control my speech always
XII. Feeling about maintenance and generalization of fluency
1. I do not feel adequate to maintain my fluency and I am worried about relapse
2. I feel I can maintain my fluency with regular guidance from the therapist
3. I feel I can maintain my fluency with periodic guidance from therapist
4. I feel I can maintain my fluency with occasional guidance from the therapist
5. I feel confidant that I can always maintain my fluency without further guidance.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
16
Appendix B
Situational Assessment Checklist for PWS
Please answer the following questions by putting a mark among the options
[0 – Nil; 1 – Very Less; 3 – More; 4 – Too much] 0 1 2 3 4
1. While speaking to friends
2. While speaking to parents
3. While speaking to siblings
4. While speaking to young children
5. While speaking to class mates/ colleagues
6. While speaking to teachers/ officers/boss
7. While speaking to strangers
8. While speaking to people of opposite sex
9. While speaking alone
10. While speaking simultaneously in a group
11. While speaking or discussing in a group
12. While addressing a small group
13. While addressing a large group
14. While singing or reciting songs or poems
15. While speaking to people in the market/railway station/bus
16. While reading loudly alone
17. While reading loudly in front of others
18. While answering questions in the classroom or work spot
19. While speaking your mother tongue
20. While speaking a language which you are not competent
21. While initiating a sentence
22. In the middle of a sentence
23. While enacting a role in a play or drama
24. While facing an interview
25. While speaking over a telephone
26. Teaching/demonstrating in a group
27. Arguing a point in a group
28. Asking/giving directions
29. When excited or happy
30. When anxious, scared, nervous or tensed
Reliability of Voice Evaluation using Cape-V Rating Scale
17
Reliability of Perceptual Evaluation of Voice using CAPE-V
Rating Scale in Indian Context
Akanksha Gupta & M. Pushpavathi*
Abstract
Human ears have the ability to identify and recognize the speaker’s voice. A trained voice
clinician is often able to determine the causative pathologies on the basis of psychoacoustic
impression of voice. Perceptual voice evaluation is an integrated process of listening to and
describing a particular voice. The Consensus Auditory-Perceptual Evaluation of Voice (ASHA, 2002)
is one of the most widely used scales in Western countries which has both categorical rating system
and visual analog scale (VAS). However no published reports are available in Indian context on
consistency of the scale as a reliable tool for perceptual analysis of voice. The aim of the study was to
evaluate the reliability across judges on different speech tasks on categorical ratings and on 100 mm
VAS and to evaluate the correlation between categorical descriptors and numerical value (on VAS)
assigned across three tasks and also to determine the most suitable voice sample (phonation/
sentences/ spontaneous speech) for perceptual evaluation of voice using CAPE-V. Twenty one adult
males who were diagnosed to have hoarse voice were taken in the study and were asked to phonate
vowels, read the sentences and spontaneously speak about their voice problem. The sample was audio
recorded and randomized and then was given to six judges for evaluation on CAPE- V scale. The
obtained value was tabulated on SPSS software for statistical analysis. Cronbach’s alpha reliability
coefficients were computed for reliability across judges on VA and ordinal scales. In general, high
reliability was observed across judges on different speech tasks. High reliability is seen in all the
parameters except for loudness, which demonstrates moderate reliability in most of the tasks. Good
correlation was observed between the two scales. That is, the judges selected VA scale value and
ordinal descriptor (degree) in such a way that it shows a relationship with each of the two. Given the
strong reliability of both, it may be noted that the higher resolution of the VA system positively
impacts the reliability of the ordinal scale. Also, it was observed that connected speech serves as the
best sample for perceptual analysis of voice. The findings of the study are helpful in evaluation of
voice disorders using perceptual scale and can be used in combination with objective analysis for
diagnosis as well as for evaluation of efficacy of voice therapy. Perceptual evaluation of voice is a
quick and reliable method of voice evaluation. Hence, its use is warranted for all the persons having
dysphonia.
Introduction
“Every human Society, no matter how primitive has developed the ability to
communicate through speech and our ability to communicate through spoken and written
language has been cited as one single most important characteristic that sets the human apart
from other animals” (Curtis, 1978). The underlying basis of speech is voice. According to
Green 2 (1964) “voice plays the musical accompaniment to speech rendering it tuneful,
pleasing, audible and coherent and is an essential feature of efficient communication by
spoken word”.
______________________________________ Reader in Speech Pathology, All India Institute of Speech and Hearing, Mysore, India
email: [email protected]
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
18
A good voice is a clear, resonant, stable, well supported by adequate breath control. It
is at a pitch level that is appropriate to the speaker and the message. Rate of speech is such
that the messages are clearly understood. An effective speaking voice should have the
following characteristics (Anderson, 1961).
Adequate loudness
Ease and flexibility
A vibrant, sympathetic quality
Pleasing and effective pitch level
Ease of diction
Clearness and purity of tone
Voice disorders arise when an individual‟s quality, pitch or loudness differs from
voice characteristics typical of speakers of similar age, gender, cultural background and
geographical location. The range of etiologies of voice disorders is large and these
differences may result from a variety of factors. Structural, medical or neurologic alterations
of the respiratory, laryngeal, and vocal tract mechanisms may create a voice disorder. Voice
quality is a term that subsumes a wide range of possible meanings covering both laryngeal
and supra-laryngeal aspects. It is perception of the physical complexity of laryngeal tone
modified by cavity resonation.
Voice can be evaluated objectively as well as subjectively in many ways. Objective
measurements include acoustic, aerodynamic and physiologic parameters using instruments.
Similarly, subjective evaluation includes perceptual ratings of voice on various parameters
like roughness, breathiness, resonance, loudness etc. But when objective acoustic measures
alone are used to analyze vocal quality there appear to represent only a friction of the set of
all the measures used by the human listener.
Perceptual Evaluation
Perceptual voice evaluation is an integrated process of listening to and describing a
particular voice. The clinician needs intensive training in voice dimensions that identify
pathology most effectively. Rating voice quality perceptually is universally acknowledged as
difficult task and one requires considerable experience in perceptual judgments. The
perceptual importance of different aspects of voice depends on context, attention, a listener‟s
background and the listening task (Kreiman, Garratt, Kempster, Erman & Berke, 1993).
Advantages of Perceptual Evaluation
The importance of perceptual measures is also demonstrated by their frequent use as a
standard against which acoustic measures are validated or compared (Kreiman et al, 1993).
Perceptual voice evaluation using any standardized scale is an inexpensive, readily available
and practical tool for evaluation purposes.
It has been found to be reliable in its findings in both inter-judge and intra-judge reliability.
Reliability of Voice Evaluation using Cape-V Rating Scale
19
Limitations of Perceptual Evaluation of Voice
The subjective evaluation of voice quality are not highly regarded as either clinical or
research tools because of the following reasons, like they are considered to lack objectivity
and do not require great technical sophistication (Weismer & Liss, 1981). Also, there is no
accepted set of perceptual scales used by the clinicians (Yumoto, Gould & Baer, 1982).
The factors like reliability and uncertainty regarding the use and meaning of various
rating scales have led some to abandon perceptual measures in favor of instrumental
approaches to voice assessment and because of inherent problems with inter-judge and intra-
judge reliability (Ludlow, 1981).
Because of these views the subjective assessment of voice has received back seat in
evaluation of voice pathology with objective evaluation as a primary means of assessment. In
literature, there are many types of perceptual scales available for the judgment of the voice
disorders. They may be a Categorical rating scale, Equal Appearing Interval (EAI) scales,
Visual Analog (VA), Direct Magnitude Estimation (DME) or Paired Comparison. Various
scales have been developed by several authors like The Voice Profile (Wilson, 1987), The
GRABS Scale (Hirano, 1981), Buffalo III Voice Profile (Wilson, 1987), The Consensus
Auditory-Perceptual Evaluation of Voice (ASHA, 2002).
The Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V) was developed as
a tool for clinical auditory-perceptual assessment of voice from a consensus meeting
sponsored by ASHA held in Pittsburg (2002). Its primary purpose is to describe the severity
of auditory perceptual attributes of a voice problem, in a way that can be communicated
among clinicians. Its secondary purpose is to contribute to hypotheses regarding the anatomic
and physiological bases of voice problems and to evaluate the need for additional testing.
CAPE-V is not intended for use as the only means of determining the nature of the voice
disorder. It is not to be used to the exclusion of other tests of vocal function. Finally, it is not
expected to demonstrate a 1:1 relation to results from other tests of vocal function.
Reliability and Validity in Perceptual Evaluation of Voice
Reliability refers to the degree to which test scores are free from errors of
measurement (American Psychological Association, 1985) and the construct of reliability
then has to be defined as relatively free of random errors of measurement (Crocker and
Algina, 1986). Random errors of measurement affect the score of a person because of purely
chance happenings. They are not consistent and will smooth down over time if a test is
repeated several times. Sources of such random errors may include “guessing, distraction in
the testing situation, administration errors, content sampling, scoring errors, and fluctuations
in the individual examinee‟s state” (Crocker and Algina, 1986).
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
20
By contrast, systematic measurement errors are those that consistently affect the score
of a person because of particular characteristic of the person or the test that has nothing to do
with the construct being measured (e.g., a rater who always uses the scale in the same
manner). Such tendencies are supposed to persist across repeated ratings with the same
instrument and affect the score of the rater in a consistent manner. Even if both error types
are of concern in score interpretation, systematic measurement errors do not result in
inconsistent measurement. Still, they may lead to low validity in the ratings and thereby
reduce the utility. Random error, however, may reduce both the reliability and the validity
and thereby the utility of the ratings (Crocker et al, 1986).
Although reliability is an important attribute, the most critical property of any test is
its validity. Validity refers to what the measurement actually measures and how useful the
measurement is to researchers. Reliability is a condition for validity and it places an upper
limit on the validity of a test. Unreliable measures will allow tests to show little, if any,
validity. Reliability is a necessary, but not a sufficient, prerequisite for the test to have
validity (Crocker et al, 1986). These factors change slowly over time and thus hypothetically
affect inter-judge reliability more than intra-judge reliability. Additional factors related to
listeners include fatigue, attention lapses and mistakes. These error factors should affect both
interrater and intra-rater reliability (Kreiman et al 1993). Systematic interactions among
listener and task factors may also occur. Listener sensitivity may interact with scale
resolution, which adds noise to the data or results in information loss.
Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V)
It was developed by Consensus meeting sponsored by ASHA held at Pittsburgh
(2002). It has additional feature of visual analog scale (100 mm scale) where the judge is
required to mark the voice quality on the scale with left end of scale indicating no
abnormality and the right end indicating severe voice problem.
Description of the Tool
The CAPE-V (Appendix I) indicates six salient perceptual vocal attributes namely
roughness, pitch, loudness, breathiness, overall severity and strain. The preferred attributes
were regarded as most identifiable and important attributes for the analysis of voice disorders.
Also they reported them to be most commonly used and easily comprehendible ones.
Each attribute is displayed and represented by a 100- millimeter line forming a visual
analog scale (VAS). The clinician indicates the degree of perceived deviance from normal
voice for each parameter on this scale, using a tick mark. Judgments may be assisted by
referring to general regions indicated below each scale on the CAPE-V: “MI” refers to
"mildly deviant," “MO” refers to “moderately deviant,” and “SE” refers to "severely
deviant." A key issue is that the regions indicate gradations in severity, rather than
Reliability of Voice Evaluation using Cape-V Rating Scale
21
discrete points. The clinician may place tick marks at any location along the line.”
Ratings are based on the clinician‟s direct observations of the patient‟s performance
during the evaluation, rather than patient report or other sources.
Scale
A 100 mm line scale with unlabelled anchors, commonly known as a visual analog
scale, is used to assess each of the six quality features. The left most portion of the scale
reflects normal voice (in the case of judging Severity, Pitch, or Loudness) or none of the
quality being judged (in the case of Roughness, Breathiness, and Strain). The right end of the
scale is to reflect the listener‟s judgment of the most extreme example of deviance.
Measurement from the left end of the scale to each tick mark, in millimeters, is denoted on
the blank to the far right of the scale (___/100).
Additional CAPE-V Elements
A nominal rating judgment allows the clinician to classify the consistency or
intermittent presence of the voice quality feature within and across evaluation tasks. Sections
devoted to resonance or other features supplement the CAPE-V protocol by allowing other
salient descriptors to document a patient‟s voice quality. This flexibility is needed to capture
the spectrum of voice disorders and associated conditions or features. The list of terms
provided on the form is not inclusive, meant only as examples of specific features that may
help describe auditory-perceptual attributes.
Concurrent Validity and the CAPE-V
Karnell, Melton, Childes, Coleman, Dailey, & Hoffman, (2006) published a
preliminary report comparing the reliability of clinician based auditory-perceptual judgments
using the CAPE-V to those made with the GRABS voice rating scheme (Hirano, 1981).
These protocols were compared after use in voice assessment of forty males and sixty one
females by certified speech language pathologists. They found comparable estimates of inter-
judge reliability for the two scales, both at high levels but suggested that the CAPE-V may
offer “more sensitivity to small differences within and among patients than the GRABS
scale”.
Study done by Wuyts, De Bodt, Van de Heyning (1999) aimed at finding and
comparing the reliability of visual analog scale and an ordinal scale by perceptual voice
evaluation of fourteen pathological voices by twenty nine listeners using GRABS scale.
Agreement was found to be higher with original 4 point scale than with the visual analog
scale version for the scale items G, R, A, B and S. A tendency was noted to rate the voice on
the middle of the visual analog scale and with increased freedom of judgment the inter judge
agreement decreases considerably but it also seemed that finer judgment of voice quality is
possible with VA scale. But, the authors reported that, it is logical to assume that the listener
perceives the vocal characteristics along a continuum, rather than quantifying perception by
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
22
intervals. The VAS appears to be advantageous for comparison with absolute acoustic
measurements as it offers more detailed information.
Effect of Speaking Task/Sample on Reliability of Perceptual Assessment
De Krom (1994) conducted a perceptual experiment in which six listeners used the
GRABS scale to rate voice fragments from seventy eight dysphonic speakers. Four different
types of stimuli were presented to each listener: one based on connected speech fragments
and the other three on segments of a sustained vowel. Analyses focused on the consistency
and reliability of ratings and results indicated that stimulus type had virtually no effect on
either intra-rater or inter-rater reliability. When determined as a function of the overall degree
of severity of a voice, the reliability of ratings for the breathiness and roughness parameters
was slightly higher for vowel stimuli than for connected speech.
Munoz, Mendoza, Fresneda, Carballo and Ramirez, (2002) examined the agreement
and reliability of ratings made by thirty four expert listeners using the Buffalo Voice Profile
System. A sustained vowel and a short utterance of connected speech were presented to each
listener. Results revealed that for the evaluation of the sustained vowel, interrater agreement
was moderate for judgments of breathiness, hyponasal resonance, and overall severity of
dysphonia and for connected speech agreement was moderate for most voice qualities.
Need for the study
Perception of a patient‟s voice is at the heart of evaluating and treating patients with
voice disorders. Patients and their families decide whether treatment has been successful
based largely on whether the patient sounds better. Similarly clinicians make many decisions
about managing the speech and voice disorder based upon perceptual judgment. Thus, there
is always a need of a reliable perceptual voice rating scale, which has good inter-judge and
intra-judge reliability. Since CAPE–V is relatively new in its coming, not many studies have
been reported in literature on the reliability of the scale. Thus, a need was felt to assess the
reliability of the perceptual evaluation of voice using the CAPE–V scale.
Aims of the study
The present study is aimed at investigating the reliability of perceptual evaluations of
voice disorders using CAPE-V scale (Consensus Auditory Perceptual Evaluation of Voice,
2002) for different tasks like phonation, sentences and spontaneous speech, in Indian context.
The aims of the study were to evaluate the
Reliability of Voice Evaluation using Cape-V Rating Scale
23
Reliability across judges on different speech tasks (phonation, sentences and spontaneous
speech) on categorical ratings (mild, moderate, severe) and on 100 mm visual analog
scale.
Correlation between categorical naming (mild, moderate, and severe) and numerical value
(on VAS) assigned across three tasks.
Determination of the most suitable voice sample (phonation/ sentences/ spontaneous
speech) for perceptual evaluation of voice using CAPE-V.
Method
Twenty one male adult participants were included in the study (in the age range of 25-
45 yrs) who were diagnosed as having hoarse voice by a qualified speech language
pathologist using acoustic analysis and perceptual assessment (without any use of
standardized scale). Total fifteen native Kannada speakers and six English speakers were
considered as subjects. The audio recordings of the speech samples were done in a sound
treated room, free from all distractions and minimum ambient noise. The room was well
ventilated and well lighted. Recordings were done using Cool Edit (Version II) software in a
Compaq Lap top with an „HP Microphone‟. The microphone was placed at a distance of 6 cm
and slightly to the side of the subject‟s mouth to minimize breathing noise. Gain was adjusted
to avoid saturation and ensure optimal use of recording dynamics. Subjects were instructed to
for three different tasks at a comfortable pitch and volumes as naturally as possible. The
complete sample was audio recorded using the Cool Edit software.
Procedure
The subjects were enrolled for the study after the completion of assessment. Informed
consent was obtained from each subject.
Data Collection/ Sample Recording - The data (voice sample) was collected by asking the
subjects to carry the following three tasks-
Task I - Maximum Phonation Duration: The subjects were asked to phonate vowels /a/, /i/
and /u/ after a deep inhalation, as long as they could in their comfortable pitch.
Task II - Reading Sentences (Repetition task if the subject was illiterate).
Three oral and three nasal, Kannada or English sentences were given to the participants who
were native speakers of respective languages. Kannada sentences were taken from stimuli
developed for nasometer assessment by Jaya Kumar, 2004 (Appendix III). English oral
sentences were taken directly from the original CAPE-V rating scale given by ASHA, 2002.
English nasal sentences were taken from the stimuli developed by Kummer, 2008 (Appendix
IV).
Task III - Spontaneous Speech – The subjects were asked to describe about their voice
problem at a comfortable loudness level. The participants were instructed to speak in
response to questions like, “Describe your voice problem”? or “What problems are you
facing because of the voice problem”?
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
24
The three tasks were recorded with an interval of twenty seconds between each task.
The samples were recorded on Cool Edit (version II). After the speech recordings the samples
from different subjects were randomized. As, there were twenty one participants, twenty one
sets were made which had different speech tasks recording of different participants. This was
done to avoid biasing by judges while rating the voice. These randomized sets were then
copied to compact discs, which were given to the six judges of the study. These samples
along with the rating sheets (Appendix II) were given to the judges for rating the voice
quality based on CAPE- V (2002).
Six speech and language pathologists, who have an experience of 3-4 years in
assessment and management of voice disorders, were selected as judges for the study. The
randomized samples were played through the compact disc (CD) and they were asked to rate
the samples based on CAPE-V rating scale. The analysis sheets (Appendix II) were given to
the judges for perceptual judgment.
The complete perceptual analysis of all the twenty one samples by each judge was
done in two sessions. Six samples were played in the first session, then followed by a week
other fifteen samples were given to the judges for perceptual ratings on CAPE-V. The rating
sheets from judges (Appendix V) were then rearranged based on randomization which was
previously done. Each subject‟s speech samples were then grouped together. The obtained
data was tabulated on to SPSS software (version -16). Appropriate statistical measures were
applied to the get the reliability co-efficient between the parameters and judges.
Results and Discussion
I. Reliability of perceptual evaluation of voice on different tasks -
To evaluate the reliability of judges in rating the voice samples Cronbach‟s alpha (α)
coefficient was computed. Inter judge reliability was evaluated for each parameter of each
task. Table 1 depicts degree of reliability on phonation task, sentences (reading sample) and
spontaneous speech. In general high reliability was found across judges and across all the
tasks (except for loudness parameter on ordinal scale in spontaneous speech). The reliability
range was within acceptable limits of 0.05. This indicates that the judges were reliable in
giving ratings on VA scale and on ordinal scale. The results of the present study indicate that
high reliability for perceptual evaluation of voice across tasks and judges was obtained by
using CAPE- V.
Reliability of Voice Evaluation using Cape-V Rating Scale
25
Table 1 depicts the reliability coefficients obtained on different tasks.
Table 1: Overall inter rater reliability across tasks on both the scales of CAPE-V
* indicates reliability between 3rd & 4th judge
The results obtained from the present study support the findings of Karnell et al
(2006), Zraick et al. (2007) and Kelchner et al. (2009) who have reported significant
reliability for the perceptual evaluation using CAPE-V. Kreiman et al (1992) have suggested
that all the listeners have similar, relatively stable internal standards for “normal” voice
quality because of the everyday experience they have with normal voices. Hence, consistency
is observed when they rate normal or near normal voices. The internal standard for
pathological voice may vary from a listener to listener depending upon their experience or
exposure to it. The study does not support the findings of Wuyts et al (1999) who opined that
though VA offers finer judgment of voice quality but with increased degree of freedom the
inter rater agreement decreases considerably. They also reported that on VA scale a general
trend was exhibited i.e. the raters tend to score the voice more to the middle of the 100 mm
line. They suggested that VA scale has more variability in rating the voices than the ordinal
scale.
II. Correlation between Ordinal and VA Scale
The dimensions on which both the scales of CAPE-V evaluate a voice sample are
different. The ordinal scale has the classification as mild, mild- moderate, moderate,
moderate- severe, severe where as VA scale has values assigned from 1-100. The perceptual
scores provided by the judges were compared to find agreement between the two scales. In
order to estimate the correlation between VA and ordinal scale, Spearman's rank correlation
coefficient was computed. Table 2 illustrates the correlation coefficients between the two
scales across the three tasks, for which Spearman’s correlation coefficient was computed
and compared.
Table 2: Correlation between ordinal and VA scale on phonation, reading sentences and
spontaneous speech task. Here, “*” indicates P < 0.05; “**” indicates P < 0.01; “***”
indicates P < 0.001.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
26
The present is aimed to evaluate the correlation between the two scales (VA and
ordinal) and also to examine the pattern of ratings given by the judges. Since the two scales
are fundamentally different the statistical analysis is also different. After converting the VA
to mVA scale cross tabulations were acquired for both the scales. It was observed that, very
few judges rate the voice quality as mild- moderate and moderate- severe i.e. ratings of 2 and
4 are being given much lesser than that of 1, 3 and 5 (degrees) on ordinal scale. Similar trend
was observed on all the three types of voice samples being rated. Hence during cross
tabulation the “2” and “4” (mild- mod and mod- severe degrees) have not been taken.
Table 3 shows the cross tabulation obtained for the task of phonation, sentences and
spontaneous speech. Rating patterns for mild, moderate and severe degree are been depicted
in the table.
Table 3: Cross tabulation for across all samples.
Reliability of Voice Evaluation using Cape-V Rating Scale
27
It can be clearly noticed similar rating patterns of judges was observed in phonation,
sentences samples and spontaneous speech task. Hence the degree assigned to particular
voice sample and the VA value given by judges, correlated well with each other. This
signifies good correlation between the VA and ordinal scale. Thus the above cross tabulations
for phonation, sentences and spontaneous speech indicates that there is a predictable pattern
in rating the voice quality on both the scales. On an average if a judge assigns VA value from
10 to 39 then he/she refers voice sample to have “mild” degree. If he/she assigns VA value
from 40 to 69 he refers the voice sample to be of “moderate” degree. Finally, if he/she gives a
VA value > 70 then it points to “severe” degree.
The findings from the present study support the results of Yu et al (2002) who
reported that the correlation between perceptual and objective voice judgment is better using
modified visual analog scale (r = 0.88) than a conventional ordinal scale (r = 0.64). This can
be attributed to the fact that distinguishing between normal and severely dysphonic voice is
easy and independent of listener’s experience or level of skill (Yu et al, 2002) but the real
challenge in perceptual analysis is recognition of intermediated grade of dysphonia (G1 and
G2 and variations in degree of dysphonia). Hence when rating on ordinal scale, judgments
even by skilled judges are subject to great variability probably because of lack of precise
internal standard to distinguish intermediate grades of dysphonia (Yu et al, 2002). This also
leads to a disadvantage that ordinal scale becomes too insensitive for small variations in voice
quality (Wuyts et al, 1999) and an advantage that it limits inter and intra rater variability by
providing a broad band for each level of severity (Yu et al, 2002). While, on the other hand it
was observed that VA scale offers increased of freedom of judgment and thus finer judgment.
But this may contribute to decreased inter rater agreement (Wuyts et al, 1999). Hence from
the present study the correlation between VA and ordinal scale can be drawn as follows,
Figure 1: Visual representation of relationship between VA and ordinal scales
III. Comparison of perceptual task across voice samples
The voice parameters were combined within phonation, sentences and spontaneous
speech. Since the study was aimed to determine the suitable task for better perceptual
evaluation of voice. In view of this, the perceptual attribute on all the dimensions were
compared across the three tasks, spearman‟s correlation coefficient was computed in order to
determine the task which shows highest correlation between the two scales of CAPE-V.
Table 4 shows the correlation coefficients obtained for each of the tasks of the study.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
28
Table 4: Overall correlation between the two scales on all the three tasks.
Here „***‟ indicates P < 0.001
High degree of correlation was obtained for all the tasks. Since spontaneous speech
task demonstrates highest correlation coefficient (r = 0.82) compared to sentences (r = 0.76)
and phonation (r = 0.73), consequently this parameter can be considered to be most
appropriate for perceptual evaluation of voice. The following graphs represent the patterns of
ratings by judges on the two scales of CAPE- V. X axis represents the ordinal scale where 1
refers to mild, 2 refers to mild moderate, 3 refers to moderate, 4 refers to moderate to severe
and 5 refers to severe. Y axis refers to VA scale values from 0 to 100 on 100 mm VA scale.
Figure 2: Graph representing the rating patterns on the phonation task
The above scatter plot was obtained for the phonation sample. From the figure it can
be derived that as ratings or severity on ordinal scale enhances i.e. from mild to severe along
the X axis the scatter plots show a clear rise in the VA scale on Y axis as well. A rise in the
height of plots towards severe degree represents good correlation between the two scales in
the phonation task. The following figure was obtained for sentence task.
Figure 3: Graph representing the rating patterns on the sentence reading task.
Similarly the following figure was obtained for sentence task.
Reliability of Voice Evaluation using Cape-V Rating Scale
29
Figure 4: Graph representing the rating patterns on the spontaneous speech task.
Analysis of sustained vowels has always been given major importance in the
perceptual (Aronson, 1980) as well as the acoustic evaluation (Greene, 1992) of voice
disorders. However, many investigators have concluded that sustained vowels do not
adequately represent continuous speech. According to Askenfelt and Hammarberg (1986), a
sustained vowel tends to be representative of voice function status only in those cases where
the dysphonia is due to severe laryngeal pathology, like laryngeal cancer or unilateral
paralysis. Thus, for the majority of the subjects it is necessary to analyze running speech in
order to obtain an adequate estimation of the voice status. Furthermore, according to
Takahashi and Koiki (1975), the initial and the terminal parts of the voice may carry
abundant information not contained in the steady-state vowel. Similarly, Hammarberg,
Fritzell, Gauffin, Sundberg and Wedin (1980) stated "Changes in running speech such as
vocal onset and termination, voice breaks, etc., are crucial to voice quality, and are not likely
to appear in a single vowel sound."
Summary and Conclusions
Perceptual voice rating scales can be of different types, like - Categorical ratings,
Equal Appearing Interval scales, Visual Analog (VA), Direct Magnitude Estimation (DME)
and Paired Comparison. In literature, many types of perceptual scales are available for the
judgment of the voice disorders. But the reliability of the perceptual data varies from study to
study. CAPE-V is relatively a new scale for voice evaluation; hence only limited studies are
available on its reliability and validity as a perceptual scale. The scale was developed as a
tool for perceptual analysis of voice from a consensus meeting by speech language
pathologists and invited experts in human perception held in Pittsburg (2002). The present
study was aimed to study the reliability of perceptual evaluation of voice using CAPE- V
rating scale.
The study was an initial attempt to explore the reliability of CAPE-V on various
parameters in Indian context. The study aimed at investigating the reliability of perceptual
evaluations of voice disorders using CAPE-V scale for different tasks like phonation,
sentences and spontaneous speech, in Indian context.
Reliability across judges on different speech tasks (phonation, sentences and
spontaneous speech) on categorical ratings (mild, moderate, severe) and on 100 mm Visual
analog scale.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
30
Correlation between categorical naming (mild, moderate, and severe) and numerical
value (on VAS) assigned across three tasks.
Determination of the most suitable task or voice sample (phonation/ sentences/
spontaneous speech) for perceptual evaluation of voice using CAPE-V.
The participants considered for the present study were twenty one males (in age range
of 25-45 years) diagnosed as having hoarse voice quality by a qualified speech language
pathologist. The participants were native speaker of Kannada or English. The voice samples
was collected by asking the participants to carry out three tasks i.e. Phonating /a/, /i/ and /u/;
repetition/ reading of sentences (of respective language), and spontaneously speaking about
their voice problem at their comfortable loudness level and in their native language. The tasks
were recorded with an interval of 20 seconds between each of tasks. The samples were
recorded on Cool Edit (Version II) software in a Compaq laptop using “hp” microphone. To
avoid biasing the samples were randomized. Six speech and language pathologists, who have
an experience of 3-4 years in diagnosis and management of voice disorders, were selected as
judges. The analysis sheets along with the randomized voice samples were given to the
judges for perceptual evaluation of voice. The complete evaluation of voice samples of
twenty one participants was done in two sessions by each of the judge.
Obtained data was reorganized and was tabulated on to SPSS software for statistical
analysis. Cronbach‟s alpha coefficient was computed to assess the reliability across judges.
Spearman‟s correlation coefficient was computed to determine the correlation between the
two scales i.e., VA and ordinal scale. Also, in order to determine the most suitable task for
perceptual evaluation of voice spearman‟s rank correlation coefficient was computed.
The first aim of the study was to evaluate the reliability across judges on different
speech tasks. Cronbach‟s alpha reliability coefficients were computed for both the scales of
CAPE-V i.e., VA and ordinal scales across the tasks. High reliability is seen in all the
parameters except for loudness, which demonstrates moderate reliability in most of the tasks.
The reliability range was within acceptable limits of 0.05 (among a particular voice parameter
in the two scales). This shows that the judges were reliable in giving ratings on VA and on
ordinal scale. The results of the present study indicate that high reliability for perceptual
evaluation of voice across tasks and judges was obtained using CAPE- V.
The second aim of the study was to investigate the correlation between ordinal and
VA scale across task. Spearman‟s rank correlation coefficient was computed for each of the
parameter between both the scales. The results of the computation revealed that there is high
correlation between the two scales amongst all the voice attributes of CAPE-V. However,
some of the parameters like loudness consistently demonstrate lower correlation in
comparison to parameters like roughness, breathiness, overall quality etc.
The third aim of the study was to determine the most appropriate task for perceptual
analysis of voice using CAPE-V. Spearman‟s rank correlation coefficient was computed in
order to determine overall correlation in all the three types of voice samples. It was observed
Reliability of Voice Evaluation using Cape-V Rating Scale
31
that overall correlation for spontaneous speech (r = 0.81) is better than that for oral- nasal
sentences (r = 0.73) and phonation (r = 0.70) tasks. Hence, following conclusions can be
drawn from the present study.
CAPE-V scale which incorporates a Visual analog and an ordinal scale has good
reliability. The use of the two scales together (VA and ordinal) facilitates reliability and
correlation across parameters and judges. It can be concluded that, spontaneous speech
sample (or connected speech sample) elicits more reliable perceptual evaluation of voice than
sustained phonation of vowel and reading sample (sentence).
Implications of the study
The findings of the study are very helpful in evaluation of voice disorders using
perceptual scale and can be used in combination with objective analysis for diagnosis as well
as for evaluation of success rates after voice therapy.
Perceptual evaluation of voice is a quick and reliable method of voice evaluation.
Hence, its use is warranted for all the patients who have dysphonia.
Limitations of the study
Sample size taken for the study was small.
Number of judges taken for study was less.
Patients only with hoarseness were included in the study.
Only inter judge reliability has been gauged in the study. No evaluation for intra judge
reliability has been done.
Directions for future research
The present study considered only trained judges who have an experience of 3-4 years
in diagnosis and management of voice disorders. The same method can be employed to
determine the effect of experience or less experience on perceptual evaluation of voice.
It may be interesting to examine the reliability of ratings of other groups of
professionals also, such as general medical practitioners or otolaryngologists, who are often
the first person called to make judgment about dysphonia and its significance in the
presentation of signs and symptoms in the patients.
Intra rater reliability can be assessed by giving the judges the same voice sample 2-3
times with an interval of some days/ week. This would be helpful in determining the internal
standard of judges as stable or unstable.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
32
References
ASHA Special Interest Division 3, Voice and Voice Disorders, (n.d.). CAPE-V form and
procedures from http://www.asha.org/about/membership- certification/
divs/div3member/default, Retrieved on April 2, 2009.
Askenfelt, A. G., & Hammarberg, B. (1986). Speech waveform perturbations analysis: A
perceptual acoustical comparison of seven parameters. Journal of Speech and
Hearing Research, 29 (1), 50-64.
Curtis, J. F. (1978). Process and disorders of communication. New York, Harper Row
Publication.
De Bodt, M. S., Wuyts, F. L., Van de Heyning, P. H., & Lambrechts, L. (1996). The
perceptual evaluation of voice disorders. Acta Oto Rhino Laryngologica, 50, 283-
291.
De Krom, G. (1994). Consistency and reliability of voice quality ratings for different types of
speech fragments. Journal of Speech Hearing Research, 37, 985–1000.
Green, M. C. L. (1964). The Voice and its disorders. Mitman Medical, London.
Hammarberg, B., Fritzell, B., Gauffin, J., Sundberg, J., & Wedin, L. (1980). Perceptual and
acoustic correlates of abnormal voice qualities. Acta Otolaryngoly, 90, 441-451.
Hirano, M. (1975). Phonosurgery: Basic and clinical investigations. Otologia (Fukuoka), 21,
239-240. Cited in Baken, R. J., (1987), Clinical measurement of Speech and Voice.
Taylor & Francis Ltd.
Hirano, M. (1981). Clinical examination of voice. Disorders of Human Communication, 5,
Springer, Wein.
Jaya Kumar (2004). Normative Scores in Nasometry in Kannada. Published master‟s
dissertation submitted to Mysore University, Mysore.
Karnell, M. P., Melton, S. D., Childes, J. M., Coleman, T. C., Dailey, S. A., & Hoffman, H.
T. (2007). Reliability of Clinician-Based (GRBAS and CAPE-V) and Patient-Based
(V-RQOL and IPVI) Documentation of Voice Disorders, Journal of Voice, Vol. 21
(5), 576–590.
Kelcner, L. N., Brehm, S. B., Weinrich, B., Middendorf, J., Dealarcon, A., Levin, L., Elluru,
R. (2009) Perceptual evaluation of severe pediatric voice disorders: Rater reliability
using the Consensus Auditory Perceptual Evaluation of Voice. Journal of Voice,
Article in Press.
Kreiman, J., Gerratt, B. R., Kempster, G. B., Erman, A., Berke, G. S. (1993). Perceptual
evaluation of voice quality: Review, tutorial and a framework for future research.
Journal of Speech and Hearing Research, Vol.36, 21-40.
Ludlow, C. (1981). Research needs for the assessment of phonatory function. ASHA Reports,
11, 3 – 8. Rockville, MD: American Speech Language Hearing Association.
Reliability of Voice Evaluation using Cape-V Rating Scale
33
Munoz, J., Mendoza, E., Fresneda, M. D., Carballo, G., Ramirez, I. (2002). Perceptual
analysis in different voice samples: agreement and reliability. Percept Motor Skills,
94, 1187-1195.
Takahashi, H., & Koike, Y. (1975). Some perceptual dimensions and acoustic correlates of
pathological voices. Acta Otolaryngologica (Suppl. 338), 2-24.
Weismer, G., & Liss, J., (1991). Reductionism is a dead end in speech research: Perspective
on a New Direction. Cited in C. Moore, K. Yorkston & D. Beukelman (Eds.),
Dysarthia and Apraxia of Speech: Perspective on Management (15-27) Baltimore:
Brookes.
Wilson, F. B. (1987). Voice Disorders. Texas, Austin.
Yu, P., Revis, J., Wuyts, F. L., Zanaret, M., Giovanni, A., (2002). Correlation of Instrumental
Voice evaluation with perceptual voice analysis using a Modified Visual analog scale,
Folia Phoniatrica et Logopaedia, 54, 271-281.
Yumoto, E., Gould, W. J., & Baer, T. (1982). Harmonics to Noise Ratio as an Index of the
Degree of Hoarseness. Journal of Acoustical Society of America, Vol. 71, pp.1544-
1550.
Zraick, R. I., Wendel, K., & Olinde, L. S. (2005). The effect of speaking task on Perceptual
Judgement of the Severity of Dysphonic Voice. Journal of Voice, Vol. 19 (4), 574-
581. 21.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
34
A preliminary Investigation into the Cognitive Abilities of Persons with
Stuttering using Simon and Stroop Tasks
Amit K & Y.V. Geetha*
Abstract
The study was planned to investigate the cognitive abilities of persons with stuttering using
Simon and Stroop tasks. The main objectives of the study were to compare the performance of PWS
and PWNS 1) to see if there is spatial interference using Simon effect, 2) to see if there is interference
of automaticity on speed of processing in PWS compared to PWNS, 3) To see if there is semantic
interference using Stroop effect in PWS and 4) To see if there is any effect on cognitive loading
(combined Simon and Stroop effect) in PWS compared to PWNS. 15 normal individuals (Control
Group) and 15 individuals with stuttering (Experimental group) in the age range of 18-30 years who
met the specified criteria were taken for study. Both groups of subjects were administered the Simon
and Stroop tasks using the standard protocol. The data was analyzed using Mixed ANOVA which was
computed for within subject factors tasks (3), color conditions (2) and other conditions (3) and
between subject factors. The results showed that there was interaction effects for task * condition,
condition * group, color*condition and task*color. No interaction effects were found for task * group
and color * group. These results indicate that overall RT for the three tasks is different between
groups. Also RT for three conditions, that is control, incongruent and congruent, differed significantly
between groups. Present study showed longer RT in incongruent conditions than congruent
conditions in Stroop task. This indicates some processing delay in incongruent condition. Simon effect
was greater for PWS group only in sb2 condition and all the other color conditions produced either
negative effect or PWS group had lesser Simon effect compared to PWNS group. But Stroop effect
showed no statistically significant difference between groups. The results of the present experiment do
not completely prove or agree with the notion that PWS group might have some deficit in their
cognitive processing with respect to information processing. At the same time the results of this
experiment should be considered with some caution due to its simple method and materials involved.
Introduction
Speech production is a very complex process which requires great amount of
cognitive processes. In the early stage of human development through cognitive ability a
child acquires his/her basic skills for survival. Cognition also helps a child to acquire his
speech and language ability without any delay or deviancy. A child with a cognitive deficit
due to congenital or acquired problems in early stages shows delay or deviant language
development. Speech language processing has been investigated in many communication
disorders. Speech language processing is nothing but a information processing which requires
cognitive components like attention, memory, decision making, etc,.
Stuttering has been viewed as a puzzling disorder by the speech language pathologists
and by many other professionals investigating it due to its unknown cause or complex nature.
_____________________________ * Professor of Speech Language Sciences, All India Institute of Speech and Hearing, Mysore, India
email: [email protected]
Cognitive Abilities of Persons with Stuttering
35
Over the past six to seven decades stuttering has been investigated for its cause, nature,
therapeutic management by professionals from various disciplines which deal with cognition,
speech-language, behavior, information processing, etc. However, so far none of them could
explain the underlying nature of stuttering.
The findings and theory on cognition and stuttering give us an indication that the
cause does not necessarily lie at the level of peripheral neural level but might lie at a more
central level. Cognition is central for any processing and planning (motor or speech). In order
to understand the language processing, it will be useful to understand some basic processes of
cognition that is information processing. Simon Effect and Stroop Effect are a couple of tasks
evolved to study the cognitive processes involved in language processing.
The Simon Effect refers to the finding that reaction times are usually faster and more
accurate when the stimulus occurs in the same relative location as the response, even if the
stimulus location is irrelevant to the task (Simon 1960). Simon's original explanation for the
effect was that there is an innate tendency to respond towards the source of stimulation.
“Stroop Effect” is named after J. Ridley Stroop who discovered this strange
phenomenon in the 1930‟s. The Stroop Task is a psychological test of mental (attention)
vitality and flexibility. The task takes advantage of our ability to read words more quickly
and automatically than we can name colors. If a word is printed or displayed in a color
different from the color it actually names; for example, if the word "green" is written in blue
ink, we will say the word "green" more readily than we can name the color in which it is
displayed, which in this case is "blue”. The cognitive component involved in this task is
attention that inhibits or stops one‟s response in order to say or do something else.
Only few studies used Stroop task to investigate cognitive ability of PWS
(Subramanian & Yairi 2006; Caruso, Chodzko-Zajko, Bidinger & Sommers, 1994). But
none of the studies used Simon task to investigate the cognitive ability of PWS. The present
study therefore aimed at investigating information processing in PWS using both Simon and
Stroop tasks.
Aims of the study
The aim of the present study is to investigate the information processing ability in
PWS compared to Persons with No Stuttering (PWNS) using Simon and Stroop tasks by
comparing the reaction time under the following conditions: 1) Simon task in both 2 color
and 4 color conditions, 2) Stroop task, 3) RT for combined tasks (Stroop & Simon) and 4) To
compare the performance of both groups (PWS & PWNS).
Method
Subjects
Two groups of subjects were taken in the age range of 18-30 years. The control group
consisted of 15 normal individuals. Participants in the experimental group were taken from
those who registered at AIISH clinic for fluency evaluation and those who were recently
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
36
enrolled for therapy at the clinic. All the participants were native Kannada speakers. The
experimental group consisted of 15 PWS. All the participants were males. A trained speech
language pathologist administered Stuttering Severity Index (SSI) to assess the severity of the
problem.
Selection criteria for experimental group: The subjects fulfilled the following criteria to
be included in the study:
Normal hearing sensitivity
No cognitive deficit
No neurological deficits
No orofacial anomalies
Normal vision or corrected vision
No participation in any of the fluency shaping/modification therapy at least for
one year prior to the experiment
Should be literate (high school) and be able to read English
Control group
15 age matched male PWNS were considered for control group. Selection criteria
were same as for the experimental group except that the control group participants had no
fluency disorder.
Instrumentation
The experiment was carried on a computer with 15” color monitor. The sequence of
events and collection of data was randomized and controlled by using software DMDX (3.0).
Materials
5 cm squared color blocks in red, black, brown, green, blue, yellow and colored words
in Aerial font, font size 14, bold and capital words in the same color as of the blocks were
also used.
Procedure
The standard procedure recommended for testing Simon and Stroop task was
employed as follows (Simon, 1960 & Stroop 1935):
The present experiment consisted of three tasks. Task I was a Simple task which used
colored blocks. Task II was also same as task I but used color words instead of blocks. Task
III involved both Simon task and Stroop task. Participants were made to sit in front of the
monitor on a comfortable arm chair. Care was taken to provide adequate ventilation and light.
Task I: The Task I consisted of four conditions, A) Control 2 condition, B) Experimental 2
condition, C) Control 4 Condition and D) Experimental 4 Condition.
Task II: The Task II was similar to Task I with four conditions but in task II colored written
words were used instead of blocks. This would give us some information on whether word
form makes any difference in RT when compared to blocks
Cognitive Abilities of Persons with Stuttering
37
Task III: In this task in all the conditions the colors of the words were different from what the
actual written word mean (Stroop task); for example, the word “red” appeared in black color
and the word „black‟ appeared in „red‟ color. The participants were instructed to respond to
the color of the word than the meaning of the actual written word, i.e. if the word „red‟
appears in „black‟ color participants were instructed to press „z‟ key and when they saw a
written „black‟ word which appeared in „red‟ color they were supposed to press „m‟ key. This
experiment is color naming rather than reading the actual word.
Results and Discussion
Present experiment had total of 23040 (64*12*30) trials. 1518 Trials were deleted
from the analysis because they were error responses or no responses. Remaining data was
considered for further analysis. Mixed ANOVA was computed for within subject factors
tasks (3), color conditions (2) and other conditions (3) and between subject factors. The
following Table 1 shows the mean and SD for both the groups across the three tasks for all
the conditions.
Table 1: Mean and SD for all the conditions of both groups
Tasks Conditions
Group
PWNS PWS
Mean SD Mean SD
Task I
sbcon 2 399.18 25.8 517.35 82.4
sbcong 2 433.21 41.9 516.27 50.6
sbincong 2 439.26 47.1 545.87 61.0
sbcon 4 531.60 74.2 632.52 79.9
sbcong 4 480.98 39.6 640.69 104.6
sbincong 4 490.36 39.5 651.33 107.0
Task II
swcon 2 356.78 27.9 476.30 59.6
swcong 2 390.51 24.2 508.83 61.3
swincong 2 413.91 19.3 534.62 62.1
swcon 4 452.81 34.6 551.04 85.8
swcong 4 456.39 28.3 639.47 73.4
swincong 4 474.85 28.2 617.71 90.0
Task III
smstcon 2 403.96 25.3 506.33 89.1
smstcong 2 400.63 34.3 526.36 74.1
smstincog 2 426.46 27.2 549.22 68.2
smstcon 4 569.71 48.8 662.64 140.9
smstcong 4 570.20 65.6 720.60 111.7
smstincog 4 547.94 51.3 703.52 122.9 (sbcon2- Simon task 2 color condition with bocks; sb2cong- Simon task 2 color condition with bocks congruent
condition; sbincong2 - Simon task 2 color condition with bocks incongruent condition; sbcon4 -Simon task 4
color condition with blocks; sbcong4 -Simon task 4 color condition with bocks congruent condition; sbincong4-
Simon task 4 color condition with bocks incongruent condition; swcon2- Simon task 2 color condition with
written color words; swcong2- Simon task 2 color condition with written color words congruent condition;
swincong2- Simon task 2 color condition with written color words incongruent condition; swcon4- Simon task 4
color condition with written color words; swcong4- Simon task 4 color condition with written color words
congruent condition; swincong4- Simon task 4 color condition with written color words incongruent condition;
smstcon2-Simon and Stroop task 2 color condition; smstcong2-Simon and Stroop task 2 color congruent
condition ; smstincon2-Simon and Stroop task 2 color incongruent condition; smstcon4-Simon and Stroop task 4
color condition; smstcong4-Simon and Stroop task 4 color congruent condition; smstincong4-Simon and Stroop
task 4 color incongruent condition)
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
38
Mixed model ANOVA was computed for tasks (3), color conditions (2), and different
conditions (3) as within subject factors and group as between subject factor. The Table 2
shows the df, F, sig values for main effects and interaction effects.
Table 2: df, F, and Sig values of tasks, color condition and different conditions (control,
congruent, incongruent) between both groups.
Source df F Sig.
Task 2 34.715 .000*
Task * Group 2 .196 .823
Color 1 189.845 .000*
Color * Group 1 2.338 .137
Cond 2 42.452 .000*
Cond * Group 2 16.290 .000*
Task * Color 2 20.951 .000*
Task * Color * Group 2 .409 .666
Task * Cond 4 9.942 .000*
Task * Cond * Group 4 1.780 .138
Color * Cond 2 6.805 .002*
Color * Cond * Group 2 9.774 .000*
(* p < .00)
Results of mixed model of ANOVA showed main effect for tasks, color and for
conditions (control, congruent and incongruent). Interaction effects for task *condition,
condition*group, color*condition and task*color were found. No interaction effect were
found for task *group and color*group. Also between subject factor showed significant
difference. Boneferroni test was computed to see the pair-wise difference across tasks and
conditions. Results showed a significant difference across all the tasks and conditions (p
<.05). These results indicate that overall RT for the three tasks were different between
groups. Also RT for three conditions, that is control, incongruent and congruent, differed
significantly between groups.
Between group analyses (Task I)
The Figure 1 shows mean and SD for both the groups for task I across control,
congruent, and incongruent conditions of both color conditions between PWS and PWNS.
normal
stg
group
sbcon2 sbcon4 sbcong2 sbcong4 sbincong2 sbincong4
Condition
200.00
300.00
400.00
500.00
600.00
700.00
RT
Figure 1: Mean and SD for different conditions in task I.
Cognitive Abilities of Persons with Stuttering
39
Independent t test was computed to see the significance of difference between both
the groups across color conditions and different conditions.
The results clearly indicate that PWS group significantly differs from PWNS group in
color naming in task I. That is, the interference effect and RT for different conditions indicate
longer processing time for PWS group.
Considerably more research has been done on manual reaction times, generally by
having the subject press a button with a finger. The results have been conflicting. Borden,
(1983) reported that PWS were slower in the execution of finger counting tasks but not in
their initiation. Cross and Luper, (1983) found that voice and finger reaction times of PWS
were highly correlated but Starkweather, Franklin and Smigo (1984) did not. Wilkins,
Webster and Morgan (1984) and Hurford and Webster (1985) reported that manual reaction
times for PWS were faster after speech therapy than before. Such results shed some light on
the inability of PWS in the execution of movements on demand in complex environments.
Before execution of any movement it has to be programmed and sequenced. Using tasks
which probe planning of movements would reveal the reason for delayed responses.
Working memory cost for task I
The relative effects of increasing the number of possible stimuli from two to four -
referred to here as working memory costs - are assessed by subtracting RTs for two-color
from four-color conditions for the groups (Bialystok, Craik, Klein & Viswanathan, 2004).
That is, RT of control, congruent and incongruent 4-color condition is subtracted from
control, congruent and incongruent 2- color conditions within each task. PWS group showed
more memory cost in congruent and incongruent conditions but failed to show any significant
difference (p > .05). In control condition PWS group showed lesser memory cost than PWNS
group.
Between group analysis ( Task II)
normal
stg
group
swcon2 swcon4 swcong2 swcong4 swincong2 swincong4
Condition
0.00
200.00
400.00
600.00
RT
Figure 2: Mean and SD of different conditions in task II of PWS and PWNS
The Figure 2 shows mean and SD for both the groups for task II across different
conditions. From Figure 2 it is clear that PWS showed longer RT in all the conditions
compared to PWNS. 4-color conditions showed longer RT compared to 2-color condition.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
40
This indicates that as the complexity increased RT also increased for both groups.
Independent t test was computed to see the significance of the difference between both the
groups across color conditions and different conditions (control, congruent and incongruent).
Table 3: t, df, sig value of between group comparison
Conditions t df Sig
swcon2 -7.025 28 .000
swcong2 -6.943 28 .000
swincong2 -7.176 28 .000
swcon4 -4.108 28 .000
swcong4 -9.004 28 .000
swincong4 -5.863 28 .000
From the Table 3, showing the t values, df and p values for different conditions of
task II and Figure 2, it is clear that PWS group showed significantly longer RT (p <.05) when
compared to PWNS group across all the conditions in task II. Though the task II used
linguistic stimuli instead of blocks, the results showed a similar trend as in task I. Providing a
linguistic stimuli would involve linguistic processing as well with the spatial interference.
Both the groups performed similarly where the incongruent condition showed longer RT
compared to congruent and control condition. It was expected that involvement of linguistic
component would slower the processing ability of PWS compared to PWNS but such an
effect was not found in task II. It was hypothesized by many researchers that PWS group is
slower in their language processing than their normal peers (Kent, 1984; Postma & Kolk
1993; Conture 2001).
Memory cost for task II
normal
stg
group
smstcon2 smstcon4 smstcong2 smstcong4 smstincog2 smstincog4
Conditions
200.00
300.00
400.00
500.00
600.00
700.00
800.00
RT
Figure 3: Mean & SD of different conditions in task III
PWS group showed more memory cost for congruent and incongruent conditions than
control condition, wherein PWS group showed lesser memory cost than PWNS group. Only
congruent condition showed statistically significant difference between group (p <.05).
Cognitive Abilities of Persons with Stuttering
41
Between group analysis (Task III)
The Figure 3 indicates that PWS group showed longer RT in all the conditions
compared to PWNS, 4- color conditions showed longer RT compared to 2- color condition.
This indicates that as the complexity increased RT also increased for both groups. The results
showed some interesting findings. 2-Color condition showed shorter RT compared to 4 color
condition in both the groups as expected. But as far as the control, congruent and incongruent
conditions are concerned it failed to show a significant difference between groups. Also in 4-
color condition, incongruent condition showed shorter RT compared to congruent condition
in both the groups. It could be due to learning effect. Results showed a significant difference
across all the conditions between groups. As the complexity increased RT also increased in
both the groups. PWS and PWNS group showed longer RT in task III compared to task I and
II.
Memory cost for task III
PWS group showed more memory cost in congruent and incongruent conditions than
PWNS group. This indicates memory requirement is more in PWS group than PWNS group.
But the difference between both groups failed to show significance in all the conditions.
Simon Effect
Simon effect shows the difference between incongruent and incongruent conditions.
The difference between RTs to congruent and incongruent stimuli (the Simon effect) reflects
the efficiency of inhibitory processes. That is, the participants‟ task is to press the key
associated with the stimulus color regardless of spatial position; therefore, smaller Simon
effects reflect less inhibition cost and more efficient inhibitory processes. More Simon effect
indicates that the participant showed longer RT in incongruent than congruent condition. The
Table 4 shows mean and SD for Simon effects across three tasks.
Table 4: Mean and SD of Simon effect between both groups.
Simon effect PWNS PWS
Mean SD Mean SD
sb2eff 6.04 21.05 29.60 30.48
sb4eff 9.37 20.37 10.64 33.74
sw2eff 23.40 24.18 25.79 33.03
sw4eff 18.45 10.25 -21.76 36.91
com2eff 25.83 31.828 22.86 32.78
comb4eff -22.26 36.82 -17.08 37.91
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
42
normal
stg
group
sb2eff sb4eff sw2eff sw4eff com2eff comb4eff
Effects
-50.00
-25.00
0.00
25.00
50.00
Sim
on
Eff
ect
(ms
)
Figure 4: Mean and SD for Simon effects
(Sb2eff- S Simon block 2 color condition effect; sb4eff Simon block 4 color condition effect; sw2eff- Simon
word 2 color condition effect; sw4eff- Simon word 4 color condition effect; com2eff- combined 2 color
condition effect (SMST); comb4eff- combined 4 color condition effect)
Figure 4 shows mean and SD of Simon effect for all the three tasks across color
conditions for both the groups. A negative Simon effect was found for combined Simon and
Stroop 4 color condition (comb 4 eff) in both the groups. Independent „t‟ test was computed
to see the significant difference between groups for Simon effects. It revealed equivocal
results. Simon effect was greater for PWS group only in sb2 condition and all the other color
conditions produced either negative effect or PWS group had lesser Simon effect compared
to PWNS group.
The Table 5 shows the t values, df, and p values for Simon effect between groups.
Results in the table shows that there is significant difference in the Simon effect in block 2-
color condition that is task I and Simon effect in word 4-color condition that is task II and
other conditions failed to show statistically significant difference. This indicates that PWS
group had longer RT (M=29.60) than PWNS (M= 6.0493). But in sw4 condition PWNS
showed negative effect (M = -21.76) than PWNS group (M= 18.45). Also, in task III
combined effect was negative for both the groups.
Table 5: t, df, sig values for Simon effects across task and color condition between groups.
Effect t df Sig
Sb2eff -2.462 28 .020
Sb4eff -.124 28 .902
Sw2eff -.226 28 .823
Sw4eff 4.066 28 .000
comb2eff .252 28 .803
Comb44eff -.379 28 .707
In the experiment task I was presented first followed by task II and finally task III was
presented. This could have led to the learning process in both the groups. Results clearly
indicate that in task I Simon effect for both color conditions were significantly different
between groups. And as the tasks progressed the effects started to diminish.
Cognitive Abilities of Persons with Stuttering
43
Increased RT shows some deficit in information processing at the level of response
selection stage. Since no other studies have been carried out with respect to Simon task and
stuttering, the results of this investigation could not be compared. Though the PWS group
showed increased RT across all the tasks, the Simon effect failed to show such a significant
difference across tasks and color condition. Results revealed significant difference in Simon
effect only in sb2 t(28) = -2.462, P = .020) and sw4 t(28) = 4.066, P = .000) condition. Sb2
Simon effect clearly shows a greater Simon effect in PWS (M = 29 ms) compared to PWNS
(M = 6 ms). This indicates that PWS group showed reduced ability to inhibit the spatial
interference effect in selecting the target response. In sb4 and sw2 Simon effect PWS (m = 9,
23 respectively) group showed slightly higher Simon effect but not statistically significant
difference compared to PWNS (M = 10, 25). Interestingly, in task III which probed combined
effect of Simon and Stroop task showed negative effects in both the groups in 4 color
condition. It was expected that combining both Simon and Stroop tasks would maximize the
cognitive loading thus producing greater effect and increased RT. RT of combined task
seemed to be higher when compared to task I and task II in PWS and PWNS, indicating
increased processing time. But Simon effect failed to produce any significant difference
between groups, also showed a negative effect. This can be attributed to learning effect.
In within group comparison, PWNS group showed increased Simon effect across
tasks. That is task I showed lesser Simon effect compared to task II, and task III showed
increased Simon effect than task I and task II. Such a trend was not found in PWS. This could
be an indication that PWS group somehow differ from PWNS group in processing of
information. Another important factor could be that individual variability. Inter subject
variability could have masked the results.
Stroop Effect
The Figure 5 shows mean and SD of Stroop effect for different color conditions of
both the groups. The Figure 5 shows that there is increased RTs for PWS in both Stroop 2-
color condition and as well as Stroop 4-color condition than PWNS.
normal
stg
group
stroo2 stroo4
Category
0.00
100.00
200.00
Stro
op E
ffect
(ms)
Figure 5: Mean and SD for Stroop effect for both color condition.
Independent „t‟ test was computed to see the significant difference between groups for
Stroop effects. Results indicate that there is no statistically significant difference for both
Stroop 2 and Stroop 4 between groups. But PWNS group showed more Stroop effect than
PWS group in both color conditions.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
44
As far as the Stroop effect is concerned there was no statistically significant
difference were found between both the groups. Results also indicated that PWS showed
lesser Stroop effect compared to PWNS group. However, Stroop effect was more in 4 color
condition compared to 2 color condition in PWNS ( M = 47, 116 ms respectively) and PWS
(M = 30, 111 ms respectively). Increased Stroop effect shows increased processing time for
both groups. This result should be noted with some caution due to simplicity of method and
color combinations. Also, variability between subjects would have masked the results. SD in
Stroop-2 color condition in PWNS was 50ms and in PWS group it was 61 ms and in Stroop
4-color condition, PWNS showed SD of 43 ms and PWS group showed 78 ms.
Subramaniyan and Yairi (2006) reported better performance by PWS group compared
to normal and high risk group. Results of their study (Subramanian & Yairi, 2006) indicated
that reaction time was greater for conflict condition, that is, when the „blue‟ word is seen in
„green‟ color. Unexpectedly, the PWS group had shorter RTs when compared to its control
group and the high risk group. Specifically, in the conflict condition, the PWS group
exhibited shortest RTs. When the ratios of the conflict-to-congruent reaction time were
calculated, the PWS group yielded a ratio of 1.16; it was 1.14 for the high risk group. The
ratios for the control groups were 1.22 and 1.26 respectively. Thus, the conflict condition did
not affect the reaction time for the experimental groups to the extent it did for the control
group. Accuracy values were 100% for all groups in the congruent condition. In fact accuracy
was high for all groups in all other conditions although still lowest for the conflict condition.
Based on these results investigators speculated that when placed under external interferences,
people who stutter may use different speech motor control (including stages of processing,
planning and production) strategies that result in lower reaction time for this group.
Present study showed longer RT in incongruent conditions than congruent conditions
in Stroop task. This indicates some processing delay in incongruent condition. But Stroop
effect showed no statistically significant difference between groups. But the procedure,
method, stimuli were entirely different between these two studies. It is not clear whether to
accept or reject the notion that cognitive ability may be deficient in PWS group. Further
studies are warranted in the same line which utilizes more complex Stroop tasks like semantic
variation, priming on a larger population for this.
Summary and Conclusions
In summary, findings from task which used blocks to probe Simon effect showed
increased RT for PWS compared to PWNS in 2 and 4 color conditions across control,
congruent and incongruent conditions. However Simon effect showed a significant difference
in 2-color condition where PWS group showed more Simon effect. This indicates some
deficits in inhibitory processes in PWS. But 4-coor condition failed to show such a difference
between groups, also the effect was reduced in PWS. Memory cost indicated that PWS
showed more memory cost compared to PWNS in congruent and incongruent conditions.
This indicates more processing time required by PWS group compared to PWNS. In task II
which used colored words instead of color blocks, PWS showed longer RT compared to
PWNS in 2 and 4-coor conditions across control, congruent and incongruent conditions.
Simon effect showed a scattered result, that is, in 2-color condition PWS showed slightly
Cognitive Abilities of Persons with Stuttering
45
more Simon effect but a negative effect in 4-color condition. Memory cost in task II also
showed a similar result as task I. PWS showed more memory cost in congruent and
incongruent conditions compared to PWNS. The results of the present study converge to
indicate an inconclusive evidence to prove any inadequacy of cognitive processes in PWS
group or failed to show any difference in terms of their inhibitory processes.
Stroop effect
Task III probed Stroop effect using colored words. Though there was a difference
between the two groups in Stroop effect, that is, PWS group showed more Stroop effect
compared to PWNS in 2 and 4-color conditions, it was non-significant. This indicates some
deficit in inhibiting interference caused by automaticity that is word reading in PWS
compared to PWNS.
Task III which probed the combined effect of Simon and Stroop showed longer RT in
PWS compared to PWNS in 2 and 4-color condition across control, congruent and
incongruent conditions. In 2-color condition PWS showed marginally less effect than PWNS
and in 4-four color condition both groups showed a negative effect. This negative effect may
be due to learning effect. In task III memory cost effect failed to show any significant
difference between both groups in all conditions. But PWS showed more memory cost effect
in congruent and incongruent conditions than PWNS.
To conclude, the results of the present experiment do not completely prove or agree
with the notion that PWS group might have some deficit in their cognitive processing with
respect to information processing. At the same time the results of this experiment should be
considered with some caution due to its simple method and materials involved. With the
current knowledge and experimental methodology it is not possible to say that PWS group
does not exhibit any cognitive deficits.
References
Bialystok, E., Craik, F.I., Klein, R. & Viswanathan, M. (2004). Bilingualism, aging,
cognitive control: Evidence from the Simon task. Psychol. Aging. 2004 June; 19(2):
290-303.
Borden, G. J. (1983). Initiation versus execution time during manual and oral counting by
stutterers. Journal of Speech and Hearing Research, 26, 289-296.
Caruso, A. J,, Chodzko-Zajko, W.J., Bidinger, D.A. & Sommers, R.K. (1994). Adults who
stutter: responses to cognitive stress. Journal of Speech and Hearing Research. 37(4),
746-54.
Cross, D. E. & Luper, H. L. (1983). Relation between finger reaction time and voice reaction
time in stuttering and nonstuttering children and adults. Journal of Speech and
Hearing Research, 26, 356-361.
Hurford, D. P. & Webster, H. I. (1985). Decrease in simple reaction time as a function of
stutterers‟ participation in a behavioral therapy. Journal of Fluency Disorders, 10,
301-310.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
46
Postma, A. & Kolk, H. H. J. (1993). The covert repair Hypothesis: Pre-articulatory repair
processes in normal and stuttered disfluencies. Journal of Speech and Hearing
Research, 36, 472-87.
Subramanian, A. & Yairi, E. (2006). Identification of traits associated with stuttering. Journal
of Communication Disorders. 39 (3), 200-16.
Starkweather, C., Franklin, S. & Smigo, T. (1984). Vocal and finger reaction time stutterers
and nonstutterers: Differences and correlations. Journal of Speech and Hearing
Research, 27, 193-196.
Simon, J. R. (1969). Reaction toward the source of stimulation. Journal of Experimental
Psychology, 81, 174-176.
Norms for Vital Capacity in Adults
47
Norms for Forced & Slow Vital Capacity in Adult Dravidian Population
Annapurna S B & N Sreedevi*
Abstract
The present study established norms for some of the air volume measurements using a dry
Spirometer in the Dravidian population. 120 healthy adults in the age range of 20-40years who were
natives of south India (Dravidians) participated in the study. Subjects were sub divided into two
groups’ i.e 20-30 years and 30-40 years. Each group comprised of 30 males and females. Spirometer
Helios 501 (RMS) was used for the study. The results indicated that there is no significant difference
between the two age groups for all the parameters considered in both males and females. When
compared across postures, the results indicated that respiratory measures are significantly higher in
standing posture than in sitting posture. It was also found that FVC and FEV1 are more sensitive
parameters in indicating the significant differences across gender and posture. The obtained norms
can be used as reference standard for estimation of lung volume among subjects with voice disorders,
chronic obstructive pulmonary disease like asthma etc.
Key words: Air volume, Dravidian, FVC, FEV1, Respiratory measures.
Introduction
The act of inhaling and exhaling air in order to exchange oxygen for carbon-dioxide is
called respiration. Apart from the primary aim of sustain life, it is the source for the speech
production. Pulmonary measures like lung volumes and lung capacities are useful to
understand the working of respiratory system. Both lung volumes and lung capacities are
responsible for breathing for life and breathing for speech. Hence respiration serves both a
life sustaining function and a speech production function. Any abnormality in respiration
involves in coordination of breathing patterns for speech production.
The respiratory features important for speech production are pressure, volume, flow
and chest wall shape. Pressure refers to the forces generated by the respiratory process.
Volume refers to the amount of air in the lungs and airways. Flow refers to the measure of
volume of air moving in a certain direction over a period of time. Chest wall shape refers to
the positioning of the chest wall (ribcage, diaphragm and abdominal muscles) for speech
breathing act.
The steady outflow of air causes vibration in the vocal folds to produce voice by the
modification of the vocal tract. It has been found that during speech the abdomen is smaller
and the rib cage is larger than in their respective relaxation positions. This positioning is
efficient for speech, because when the abdominal wall moves inward it pushes the diaphragm
upward and expands the lower rib cage. This allows the diaphragm to make quick, strong
_______________________________ * Lecturer in Speech Sciences, All India Institute of Speech and Hearing, Mysore, India
email: [email protected]
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
48
contractions, which facilitates the quick inspirations and the constantly changing pressures
needed for speech (Hixon et al., 1973). Basically, there are four changes that occur when we
switch from breathing for life to breathing for speech: the location of air intake, the ratio of
time for inhalation versus exhalation, the volume of air inhaled per cycle and the muscle
activity for exhalation.
CHANGE LIFE SPEECH
Location of air intake
Ratio of time for
inhalation versus
exhalation
Volume of air
Muscle activity for
exhalation.
Nose
Inhale: 40%
Exhale: 60%
500 cc
10%VC
Passive: Muscles of
thorax and diaphragm
relax
Mouth
Inhale: 10%
Exhale: 90%
Variable, depending on length and
loudness of utterance, 20 to 25% VC
Active: thoracic and abdominal
muscles contract to control recoil of
ribcage and diaphragm.
The Speech Language Pathologist needs to focus on respiration as it relates to speech
production and also there is a need to evaluate the respiratory system as a part of voice
evaluation. Studies in the physiopathology of the lungs are aimed primarily at the
investigation of so called functional weakness of the respiratory organ. In the assessment of
pulmonary functions spirometry is commonly used. Spirometers are useful in determining
volumes such as vital capacity and tidal volume. Vital capacity is used as an index of lung
capacity for speech and voice functions. A person does not use the entire vital capacity in
functional, quite respiration or for speech. Thus a small percentage of total vital capacity is
used in quite breathing. In normal healthy persons, the volume of air in the lungs primarily
depend on body size and build however the body position also influences the pulmonary
measures. It has been found that most of the volumes and capacities decrease when a person
is lying down rather than standing.
Kent (1994) compiled norms for many aspects of respiration including standards of
measurement, appropriate instrumentation, capacities and respiration rates by age and gender,
physiological requirements for speech production, body size factors related to respiration,
flow volume relationships and effects of smoking on the respiratory tract. He reported that
between 25% and 40% of vital capacity is used in speech by typical adults.
The pulmonary measures are influenced by a number of factors particularly height,
age, usual habitat, geographical condition and ethnic and racial origin. In different parts of
the world several investigators have investigated the different lung function measurements
and established the normal standards (Cotes and Ward, 1966; Da Costa, 1971, Sider and
Peters 1973). In a comparative study of lung function among the American, European,
Jordanian, Negro and the Pakistani subjects, it was found that the former three groups were
superior to the remaining groups. The vital capacity varies with age, sex, height, weight, body
Norms for Vital Capacity in Adults
49
surface area, body build and other factors (Zemlin, 1981). Hutchinson (1979) had explained
the relationship between lung capacity and body size and weight.
In the Indian context, such studies were carried out for the north-west region (Jain &
Ramaiah, 1969; Mathew, 1984; Verma, 1983) and southern geographical region (Kamat et
al., 1982; Reddy and Shastry, 1944). These studies observed that the Spirometric functions
varied between these two populations. Chatterjee (1988) reported that the Spirometric
functions of Eastern region Indians are comparable to north-west Indians and superior to
southern Indians. Vital capacity can be predicted on height and weight and there is no
significant difference between the vital capacities and mean flow rates for both males and
females (Krishna Murthy, 1986). Sudhir Banu (1987) found significant differences in mean
airflow rates in dysphonics. It has been found that mean vital capacity values in Indians were
significantly lower than the western subjects (Battacharya, 1963).
It is a well established fact that respiratory measures vary across geographical
locations. There are no established norms for the respiratory measures in the Dravidian
population. Hence the present study is taken up to establish norms for some of the air volume
measurements using a dry Spirometer in the Dravidian population.
Method
120 healthy adults in the age range of 20-40years who were natives of south India
(Dravidians) were selected for the study. Subjects were sub divided into two groups i.e 20-
30years and 30-40years. Each group comprised of 30 males and females.
The subjects were selected based on the following criteria
1. Should be of Dravidian origin
2. Should be free from any history of respiratory, circulatory and neuromuscular
diseases
3. Should be non-smokers and non- alcoholics (heavy)
4. There should be no history of any serious illness and syndromatic conditions
5. Should be free from any obesity related problems
The instrument used in the study was Spirometer Helios 501 (RMS). Helios 501 is a
portable handheld monitoring spirometer.
The following measures were deduced from the study:
FVC (Forced Vital Capacity): This is the total amount of air that can forcibly be blown
out after full inspiration, measured in liters.
FEV1 (Forced Expiratory Volume in 1 Second): This is the amount of air that can be
forcibly blow out in one second, measured in liters.
FEV1 / FVC (FEV %): This is the ratio of FEV 1 to FVC. In healthy adults this should
be approximately 75 - 80%. FEV1/FVC is the FEV1 expressed as a percentage of the
FVC.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
50
Slow Vital Capacity (SVC): Is the maximum volume of air which can be exhaled or
inspired in a slow/steady maneuver in liters.
TV (Tidal Volume): During each respiratory cycle, a specific volume of air is drawn into
and then expired out of the lungs. This volume is tidal volume.
Subject’s height and weight were recorded before the testing. Body Mass Index needs
to be calculated for ruling out obesity. For calculating body mass index the individual's body
weight is divided by the square of their height using the following formula (WHO, ’95,
2004).
BMI= Weight (Kg)/Height ² (m²)
The subjects were considered for the study only when the BMI score was within the
normal range. Each subject is tested individually at a time and was instructed about the test
procedures. The test begins with a model given by the researcher. Before starting the test,
mouth tubes were cleaned and sterilized properly.
Procedure for obtaining FVC (Forced Vital Capacity)
The subject is instructed to begin with a relaxed breathing, then to take a deep breath
in and to immediately blow air out as hard and fast as possible into the mouth piece until no
more air can be exhaled and then the subject should take another deep breath back in, with
the mouthpiece still in the mouth, until the lungs are full. Ask the subject to do this activity
when on the start button is clicked. The same task is repeated thrice each in both sitting and
standing postures for every subject. The best maneuver out of the 3 trials is selected by the
instrument as well as through visual examination and was considered for further analysis.
Procedure for obtaining SVC (Slow Vital Capacity)
The subject is instructed to begin with relaxed breathing through the mouth piece for
two to three cycles and then to take a deep breath followed by a deep exhalation. Both
inhalation and exhalation should be performed to the maximum extent but slowly and
following this the subject should take a few gentle and normal breaths. The subject is asked
to do this activity when the start button is clicked. From the results displayed on the screen,
values for SVC and tidal volume were noted for further analysis.
Results and Discussion
The purpose of the present study was to develop normative values for the respiratory
measures using a Spirometer. The parameters considered were forced vital capacity, slow
vital capacity and some of their related parameters in sitting and standing postures. A total
number of 120 healthy adult subjects in the age range of 20-40 years were studied. They were
divided into 2 age groups i.e. 20-30 years and 30-40 years with 30 males and females in each
group. Data obtained were analyzed using descriptive and inferential statistical procedures.
SPSS (Ver 16 & Ver 10) were used for the statistical analysis. Using Descriptive statistics
group means and standard deviations for each measure was obtained and are depicted in
Tables 1 and 2.
Norms for Vital Capacity in Adults
51
Table 1: Shows the Mean & SD of the respiratory measures for males. M
easu
res
Males
20-30 years 30-40 years
Sitting Standing Sitting Standing
Mean(SD) Range Mean
(SD)
Range Mean(SD) Range Mean
(SD)
Range
FVC 3.71(.27) 2.85-
4.07
3.78
(.33)
2.50-
4.18
3.81(.27) 1.12-
4.17
3.89(.28) 2.85-4.32
FEV1 3.39 (.38) 2.67-
3.98
3.48
(.39)
2.69-
4.05
3.78(.25) 2.85-
4.04
3.85(.21) 3.20-4.12
FEV1/
FVC
(%)
98.69
(1.98)
94.20-
100
99.32
(1.27)
94.65-
100
98.69
(2.10)
92.45-
100
99.22
(1.35)
95.60-100
SVC 3.5(0.39) 2.68-
4.37
3.61(.41) 2.77-
4.55
3.75(.14) 3.46-
3.98
3.82
(0.15)
3.50-4.11
TV 1.43(.61) 0.38-
2.38
0.57(.56) 0.40-
2.40
2.56(.63) 0.16-
2.57
1.71(0.56) 0.65-2.70
On observation of Table 1, it is evident that all the respiratory measures considered in
this study were higher in the older age group of 30-40 years compared to the younger group
of 20-30 years in men in both sitting and standing postures. The standard deviation was not
much different in the two groups. The distribution of tidal volume was scattered compared to
the other parameters in both the postures.
Table 2: Shows the Mean & SD of the respiratory measures for females.
Mea
sure
s
Females
20-30 years 30-40 years
Sitting Standing Sitting Standing
Mean(SD) Range Mean(SD) Range Mean(SD) Range Mean(SD) Range
FVC 2.38(.37) 1.76-
3.37
2.47(.35) 1.98-
3.20
2.29(.16) 2.05-
2.61
2.43(.21) 2.14-
2.90
FEV1 2.26(.37) 1.76-
3.22
2.41(.34) 1.98-
3.31
2.67(.35) 1.86-
3.22
2.76(.35) 1.98-
3.31
FEV1/
FVC
(%)
99.63
(1.02)
94.57-
100
99.93
(.66)
96.35-
100
99.69
(1.71)
94.57-
100
99.87
(.66)
96.35-
100
SVC 2.51(.41) 1.76-
3.80
2.50(.59) 0.65-
3.33
2.45(.24) 2.10-
2.90
2.52(.23) 2.18-
2.99
TV 0.99(.33) 0.70-
1.85
1.12(.38) 0.12-
2.00
0.64(.34) 0.70-
1.27
0.77(.41) 0.12-
2.00
In female subjects, considering the mean values, all the parameters had higher values
in the younger group of subjects of 20-30years compared to the older group of 30-40 years in
the sitting posture. However, in the standing posture, FEV1 and SVC were higher in the older
group of women compared to the younger group. In general, the standard deviation of the
respiratory measures was higher in females compared to their male counterparts. The
normative values obtained in the two age groups of males and females are presented in
Appendix 1.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
52
Initially two way MANOVA was carried out to ensure the main effect of age, gender
and interaction between age and gender within each measure in sitting and standing postures.
Gender effect was significant in all the measures at 0.05 level of significance. Age was also
significant in FEV1 in both postures at 0.05 level of significance. There was significant
interaction between age and gender in SVC & TV in both postures. Since there is interaction
between age and gender further analysis was carried out separately for age groups and
gender.
Comparison across age
For comparison across age, Independent t test was used. Comparing males across the
two age groups, it was found that FEV1 and SVC are significantly higher in the 30-40 years
group compared to the 20-30 years group in both sitting and standing postures at 0.01 level of
significance (Graphs 1 & 3). Comparing females across the two age groups, it was found that
FEV1 and TV are significantly higher in the 20-30 years group compared to the 30-40 years
group in both the postures at 0.01 level of significance (Graphs 2 & 4).
As evident from the statistical analysis, there is no major difference between the two
age groups for all the parameters considered in both males and females as there may not be
significant changes in the chest volume over a span of just 10 years as considered in this
study. Another possible reason is that FVC and FEV1 are relatively less sensitive than
maximal mid expiratory flow (MMF) which is found to be a more sensitive test of subtle,
asymptomatic changes in pulmonary function (Williams, Miller and Taylor , 1978). This
finding is also in accordance to Ayub, Zaidi and Burki (1987), who reported that in Pakistani
men and women, effects of age on FVC and FEV1 were found to be minimal and they state
that only after the 4th
decade significant changes are expected. In the present study all the
subjects were with in 40 years of age.
Comparison across gender
Using independent t test, it was found that all the measures were significantly higher
in males compared to females in both the age groups of 20-30 years and 30-40 years at 0.01
level of significance. This finding is supported by Ali Baig and Qureshi (2007) who reported
that mean FVC and FEV1 were found to be higher in males than in the females at all ages.
Comparison across postures (standing vs sitting)
Using paired t test, in males it was found that all the measures were significantly
different across sitting and standing postures in both the age groups at 0.01 level of
significance (Table 3). Only tidal volume was found to be significantly higher in the sitting
posture all the other measures were higher in the standing posture.
Norms for Vital Capacity in Adults
53
Table 3: Shows the comparison of measures in sitting Vs standing posture in males.
Measures
Males
20-30 years (N=30) 30-40 years (N=30)
Sitting Vs Standing Sitting Vs Standing
“t” value “t” value
FVC 3.56 * 9.14 *
FEV1 4.31 * 6.04 *
(%) 2.89 * 3.18 *
SVC 4.16 * 9.08 *
TV 3.49 * 5.79 *
* indicates significant difference p<0.01
On the same lines, using paired t test, it was found that all the measures except
FEV1/FVC were significantly higher in standing than in sitting posture among females in the
30-40 years group at 0.01 level of significance. Similarly it was found that all the measures
except FEV1/FVC and SVC were significantly higher in the standing posture than in sitting
posture for females in the 20-30 years group at 0.01 level of significance (Graphs 2 & 4).
The t values obtained are shown in Table 4.
Table 4: Shows the comparison of measures in sitting Vs standing posture in females.
* indicates significantly higher in standing posture (p<0.01).
Measures Females
20-30 years
(N=30)
30-40 years
(N=30)
Sitting Vs Standing Sitting Vs Standing
“t” value “t” value
FVC 2.64 * 5.67 *
FEV1 4.70 * 9.42 *
FEV1/FVC(%) - -
SVC - 8.00 *
TV 2.83 * 2.59 *
Hence, in both males and females most of the respiratory measures were significantly
higher in standing posture than in sitting posture. This can be explained on the basis that the
subjects take slightly larger inspirations in this posture than in the sitting position (Townsend,
1984). This finding is supported by Hixon, Goldman and Mead (1973) who investigated the
effects of various body positions on respiratory behavior during oral reading. They found
lung capacities are maximum in the upright position compared to other postures. In the
upright position, gravity acts in an expiratory direction on the ribcage and in an inspiratory
direction on the abdomen. The effect is mainly on the abdomen, being greater at low than at
high lung volumes because the height of the abdomen is greater and its walls less stiff in the
standing posture (Agostoni and Mead, 1964).
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
54
Respiratory Parameters
TVSVCFEVFVC
Mea
n V
alu
es (
litr
es)
- S
itti
ng
- M
ale
s
4.0
3.5
3.0
2.5
2.0
1.5
1.0
.5
0.0
Age Groups
20-30 yrs.
30-40 yrs.
Figure 1: Shows mean values for males across age group in sitting position.
Respiratory Parameters
TVSVCFEVFVC
Mea
n V
alu
es (
litr
es)
- S
itti
ng
- F
emal
es
4.0
3.5
3.0
2.5
2.0
1.5
1.0
.5
0.0
Age Groups
20-30 yrs.
30-40 yrs.
Figure 2: Shows mean values for females across age group in sitting position.
Respiratory Parameters
TVSVCFEVFVC
Mea
n V
alue
s (lit
res)
- St
andi
ng -
Mal
es
4.0
3.5
3.0
2.5
2.0
1.5
1.0
.5
0.0
Age Groups
20-30 yrs.
30-40 yrs.
Figure 3: Shows mean values for males across age group in standing position.
Norms for Vital Capacity in Adults
55
Respiratory Parameters
TVSVCFEVFVC
Mea
n V
alu
es (
litr
es)
- S
tan
din
g -
Fem
ale
s
4.0
3.5
3.0
2.5
2.0
1.5
1.0
.5
0.0
Age Groups
20-30 yrs.
30-40 yrs.
Figure 4: Shows mean values for females across age group in standing position.
As observed in the present study, Lalloo, Becklake and Goldsmith (1991) found that
on average all the Spirometric indices examined, except the peak expiratory flow rate
(PEFR), were higher in the standing position compared to the sitting position although the
change was only significant at the 5% level for FEV1 in women. Also, Townsend (1984)
reported that the forced expiratory volumes in one and in six seconds and forced vital
capacity were significantly larger in the standing than in the sitting posture.
Some of the literature reports on respiratory measures are predicted on height and
weight. Hence, looking at the data obtained in the present study, it was felt that normative
values could be determined for the respiratory measures considered in this study based on
height, weight and body mass index measures. And as the results indicated that most of the
respiratory measures obtained in the standing posture are significantly higher than in the
sitting posture, for comparisons based on height, weight and BMI, sitting posture
measurements are not considered and only standing posture measures were considered.
Height: Based on height, 120 subjects in the study were grouped into 3 groups, i.e. 150-
160cms (Group 1), 160-170 cms (Group 2) and 170-180 cms (Group 3). Groups 1, 2 and 3
consisted of 24, 32 and 64 subjects respectively including both males and females. Among
the 3 groups, the range of height varied between 153 to 182 cms. The mean and standard
deviation for the parameters with respect to height are presented in Table 5. As the 3 groups
based on height did not have equal representation or rather unequal distribution of males and
females, post hoc test was used to determine the significant difference in the respiratory
measures across groups as a whole. And therefore gender difference was not considered
based on height.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
56
Table 5: Shows the Mean and SD of the respiratory measures based on height.
Measures 150-160cms
(N=24)
160-170cms
(N=32)
170-180cms
(N=64)
Mean (SD) Mean (SD) Mean (SD)
FVC 2.43 (0.27) 2.49 (0.31) 3.74* (0.48)
FEV1 2.47 (0.39) 2.67 (0.38) 3.60* (0.44)
FEV1/FVC(%) 99.98* (0.09) 99.95 (0.27) 99.26 (1.32)
SVC 2.44 (0.51) 2.53 (0.40) 3.66* (0.39)
TV 1.04 (0.44) 0.88 (0.41) 1.60* (0.59)
* Indicates that measures are significantly higher compared to the other groups (p<0.01)
Post Hoc analysis was used to determine the significant difference for the respiratory
measures across the 3 groups. It was found that all the measures except FEV1/FVC were
significantly higher in the 170-180 cms group (Group 3) compared to the other two groups.
And FEV1/FVC was found to be significantly higher in 150-160 cms group (Group 1)
compared to the other two groups. Groups 1 and 2 were combined as a single group (150-170
cms) as there was no significant difference for most of the parameters across these groups.
So, two groups were made based on height. In general, the tallest group had significantly
higher respiratory measures. Similarly Virani, Shah & Celly (2001) have reported that in both
men and women height showed significant correlations in all the measures except FEV1.
Weight: Based on weight, the subjects were grouped into 5 groups i.e. <50kgs (Group 1), 50-
60kgs (Group 2), 60-70kgs (Group 3), 70-80kgs (Group 4), 80-90kgs (Group 5). The Mean
and Standard deviation are shown in Table 6. Groups 1, 2, 3, 4 and 5 consisted of 8, 37, 21,
49 and 5 subjects respectively. Among the 5 Groups, the weight range varied between 45kgs-
83kgs.
Table 6: Shows the Mean and SD of the respiratory measures based on weight.
Using Mann-Whitney test, comparisons were made across the groups based on weight
for obtaining the significant difference (Table 7). It was found that almost all the measures
were significantly higher in Groups 4 and 5 compared to the other three groups. Therefore the
first three groups were combined as a single group i.e. <50 – 70 Kg and Groups 3 and 4 are
combined as a single group i.e. 70-90 Kg. Hence there are two groups based on weight for
determining the normative values.
Measures < 50kgs
(N=8)
50-60kgs
(N=37)
60-70kgs
(N=21)
70-80kgs
(N=49)
80-90kgs
(N=5)
Mean
(SD)
Mean
(SD)
Mean
(SD)
Mean
(SD)
Mean
(SD)
FVC 2.37 (0.25) 2.52 (0.43) 2.79 (0.61) 3.81 (0.37) 3.91 (0.15)
FEV1 2.37 (0.35) 2.71 (0.46) 2.70 (0.45) 3.67 (0.38) 3.80 (0.23)
FEV1/FVC(%) 1.00 (.00) 99.94 (0.26) 99.97 (0.10) 99.13 (1.44) 99.30 (1.06)
SVC 2.49 (0.79) 2.59 (0.48) 2.82 (0.66) 3.67 (0.38) 3.84 (0.21)
TV 1.11 (0.55) 0.92 (0.46) 1.14 (0.48) 1.67 (0.59) 1.33 (0.50)
Norms for Vital Capacity in Adults
57
Table 7: Shows the measures which are significantly different between the groups
Weight < 50kgs 50-60kgs 60-70kgs 70-
80kgs
80-
90kgs
< 50kgs -- -- -- -- --
50-60kgs NS -- -- -- --
60-70kgs NS NS -- -- --
70-80kgs FVC,
FEV1
FVC,FEV1,
FEV1/FVC,SVC,TV
FVC,FEV1,
FEV1/FVC,SVC,TV
-- --
80-90kgs FVC,
FEV1,
SVC
FVC,FEV1,
FEV1/FVC,SVC
FVC, FEV1, SVC NS --
(p<0.01) NS: Not significant.
As seen in height, subjects with more weight showed higher respiratory capacities.
However, this observation needs to be viewed cautiously as all the subjects in the present
study were with in normal limits of BMI. So it is erroneous to state that people with greater
height and weight will always have higher respiratory capacities.
Body Mass Index: Based on body mass index, the subjects were grouped into 3 groups i.e.
18-20 (Group 1), 21-23 (Group 2) and 24-25 (Group 3). Groups 1, 2 and 3 consisted of 29, 61
and 30 subjects respectively. The normal range of body mass index varied between 18-25
(WHO,’95, 2004) and therefore only those subjects within this range were considered for the
study. Mean and standard deviation of BMI obtained are presented for each group in Table 8.
Table 8: Shows the Mean and SD of the respiratory measures based on BMI.
Measures 18-20
(N=29)
21-23
(N=61)
24-25
(N=30)
Mean Mean Mean
FVC 2.52 (0.43) 3.23 (0.76) 3.57 (0.60)
FEV1 2.69* (0.47) 3.20 (0.66) 3.38 (0.62)
FEV1/FVC (%) 99.93 (0.29) 99.46 (1.16) 99.50 (1.16)
SVC 2.63* (0.58) 3.19 (0.70) 3.43 (0.64)
TV 0.9* (0.5) 1.32 ( 0.61) 1.54 (0.60)
* Indicates that measures are significantly lower compared to the other two groups (p<0.01).
Using Post Hoc analysis, it was found that all the measures except FVC and
FEV1/FVC were significantly lower in 18-20 BMI (Group 1) compared to the other two
groups (Table 9). Therefore Groups 2 and 3 are combined as a single group (BMI = 21-25)
for determining the normative values.
Hence in the present study, an attempt was made to establish normative values for
some of the respiratory measures based on height, weight and body mass index. The
normative values obtained are listed in Tables 9 and 10.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
58
Table 9: Shows the Normative values based on height and weight for standing posture
Measures Males Females
150-170cms 170-180cms 150-170cms 170-180cms
<50
-
70k
gs
70-
90kgs
<50-
70kgs
(N=7)
70-
90kgs
(N=53)
<50-
70kgs
(N=56)
70-
90kgs
<50-
70kgs
(N=3)
70-
90kgs
(1)
FVC
*
3.77 3.84 2.46
*
**
** FEV1 3.36 3.71 2.57
FEV1/FVC (%) 99.94 99.23 99.97
SVC 3.78 3.71 2.49
TV 1.67 1.65 0.93
* Indicates no subjects were present in the group.
** Indicates less number of subjects in the group.
Table 10: Shows the normative values based on body mass index in standing posture.
Measures
Males Females
18-20
(N=2)
21-25
(N=58)
18-20
(N=27)
21-25
(N=33)
Mean Mean Mean Mean
FVC
**
3.84 2.43 2.47
FEV1 3.67 2.60 2.55
FEV1/FVC (%) 99.29 99.94 99.88
SVC 3.71 2.54 2.49
TV 1.66 0.91 0.95
** Indicates less number of subjects in the group
When the subjects were combined, the overall results have indicated that subjects
with greater height and weight and BMI in the upper limit of the normative range (21-25) had
higher respiratory capacities.
The norms obtained based on categorizing subjects according to height, weight and
BMI are also provided in the result section. However, these values cannot be recommended
for clinical utility at present as there are less number of subjects with in those groups. Hence
this only serves as a preliminary attempt to classify the subjects for their respiratory
capacities based on these variables.
To conclude, out of the respiratory measures FVC, FEV1, FEV1/FVC (%), SVC and
TV considered in the present study, it was found that FVC and FEV1 are more sensitive in
indicating the significant differences across gender and posture. The normative values
Norms for Vital Capacity in Adults
59
obtained in the present study can be used clinically for obtaining of respiratory measures in
voice disorders and other pulmonary conditions.
References
Agostoni, E, Mead, J. (1964). Statics of the respiratory system. In: Fenn WO, Rahn H,
(Editors) Handbook of Physiology. Respiration. Vol. I. Washington, DC: American
Physiological Society; p.387-409, 82 (1), 89-95.
Ali Baig, M.I, Qureshi, R.H. (2007). Pulmonary function tests: normal values in non-smoking
students and staff at the Aga Khan University, Karachi. Journal of College of
Physicians and Surgeons Pakistan, 17 (5), 265-8.
Ayub, M, Zaidi, S.H, Burki, N.K. (1987). Spirometry and flow-volume curves in healthy,
normal Pakistanis. British Journal of Diseases of the Chest, 81 (1), 35-44.
Bhattacharya, D.K. (1963). Vital capacity of the Jat males of Punjab: modification of the
formulae existing for its computation. Indian Journal of Medical Research, 51, 361-
365.
Chatterjee, S, Nag, S. K, Dev, S.K. (1988). Spirometric standards for non-smokers and
smokers of India (eastern region). Japanese Journal of Physiology, 38 (3), 283-98.
Chatterjee, S., Saha, D., and Chatterjee, B. P. (1988). Pulmonary function studies in healthy
non - smoking men of Culcutta. Annals of Human Biology, in press.
Cotes, J. E., & Ward, M. P. (1966). Ventilatory capacity in normal Bhutanese. Journal of
Physiology, (Lond.), 163, 36-37.
Dacosta, J. L. (1971). Pulmonary function studies in healthy Chinese adults in Singapore.
American Review of Respiratory Disease, 104, 128-131.
Hixon, T.J., Golman, M., & Mead, J. (1973). Kinematics of the chest wall during speech
production: Volume displacements of the ribcage, abdomen, and lung. Journal of
Speech and Hearing Research, 16, 78-115.
Hutchinson, B.B., & Hanson, M.L. (1979). Diagnostic Handbook of Speech Pathology.
Baltimore: Williams & Wilkins.
Jain, S. K. and Ramiah, T. J. (1969). Normal standards of pulmonary function tests for
healthy Indian men 25-40 years old. Comparison of different regression equations
(prediction formula). Indian Journal of Medicine Research, 57, 1453-1463.
Kamat, S. R., Tyagi, N.K., & Rashid, S.S.A. (1982). Lung function in Indian adult subjects.
Lung India, 1, 11-21.
Kent .D. (1994). Reference manual for communication sciences and disorders. Austin, TX:
PRO–ED.
Krishna Murthy B.N (1986). The measurement of mean airflow rate in normals. Unpublished
Dissertation submitted to the University of Mysore.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
60
Lalloo U.G, Becklake M.R, Goldsmith CM (1991) Effect of standing versus sitting position
on spirometric indices in healthy subjects. Respiration. 58 (3-4):122-5.
Mathew, L., Sengupta, J., Lakhura, S. C., & Raman, C. V. (1984) Age related changes in
lung functions in Indian servicemen. Indian Journal of Medical Research. 79: 529-
537.
Reddy, D.V.S. & Sastry, B. (1944). Studies in vital capacity. Indian Journal Medical
Research, 32, 237- 246.
Sider, R. & Peters, J. M. (1973). Difference in ventilatory capacity of Irish and Italian fire
fighters. American Review of Respiratory Disease, 108, 669-671.
Sudhir Banu (1987). Mean airflow rates in Dysphonics. Unpublished Dissertation submitted
to the University of Mysore.
Townsend M. C. (1984). Spirometric forced expiratory volumes measured in the standing
versus the sitting posture. American Review of Respiratory Disease, 130(1), 123-124.
Verma, S .S., Kishore, N ., Kaman, C .V., Lakhera, S.C., & Dass, S. K. (1983). Prediction
of some ventilatory norms in healthy Indian males 21-69 years age. Indian Journal of
Physiology Pharmacology, 27, 45-49.
Virani, N, Shah, B, Celly, A. (2001). Pulmonary function studies in healthy non-smoking
adults in Sri Aurobindo Ashram, Pondicherry. Indian Journal of Medical Research,
114, 177-84.
Zemlin, W.R. (1981). Speech and hearing science: Anatomy and Physiology (2nd
Ed),
Englewood Cliffs; New Jersey, Prentice Hall, INC.
Zemlin, W. R. (1998). Speech and Hearing Science: Anatomy and Physiology (4th Ed.).
Needham Heights, MA: Allyn & Bacon.
http://www.ncbi.nlm.nih.gov/pubmed/
Appendix - 1
Measures Males Females
Mean (SD) Range Mean (SD) Range
FVC 3.83 (0.31) 2.50 - 4.32 2.45 (0.28) 1.98 - 2.90
FEV1 3.66 (0.36) 2.69 - 4.12 2.57 (0.39) 1.98 - 3.31
FEV1/FVC(%) 99.31 (1.28) 94.65 - 100 99.91 (0.50) 96.35 - 100
SVC 3.71 (0.32) 2.77 - 4.11 2.51 (0.45) 0.65 - 2.99
TV 1.65 (0.57) 0.40 - 2.70 0.93 (0.45) 0.12 - 2.00
Norms for respiratory measures in Dravidian adults (20 – 40 years) for standing posture
Note: As the respiratory measures were not significantly different between the two age groups
(20-30 Vs 30-40 years) the above norms are for Dravidian adults in the age range of 20-40 years.
MAFAT - K
61
Manual for Adult Fluent Aphasia Therapy in Kannada (MAFAT-K)
Chaitra. S & S. P. Goswami*
Abstract
The language deficits seen in persons with aphasia are vast and heterogeneous in nature.
Thus it is essential that treatment should focus on all the communication deficits faced by an
individual with aphasia. The purpose of the present study was to develop a treatment manual for adult
fluent aphasia therapy in Kannada. The present manual was designed based on the documented
principles and guidelines prescribed in the literature for persons with fluent aphasia. The treatment
parameters are formulated under six main domains i.e., functional communication, auditory
comprehension, repetition, naming, expression, reading and writing. Each of these domains was
further divided into several subsections. Line drawing (coloured) pictures have been used under
various categories as stimulus. Stimulus has been judged by SLPs speaking Kannada for its relevancy
and familiarity. Stimulus has been arranged in hierarchical order of simple to complex.
Introduction
The management of persons with aphasia is a complicated task that involves the
coordinated efforts of a rehabilitation team representing several disciplines. The speech
language pathologist (SLP) being one of the key member of the team, do face number of
challenges while treating persons with aphasia. The reason is mainly due to the varied nature
of the disorder manifesting impairment in all aspects of language. Thus, the speech language
pathologists must use language treatment programs that have been described in detail and
proved to be effective (Shewan & Bandur, 1986).
Aphasia is classified into various types, one of the classifications being the basic
taxonomy of the aphasia syndrome (Kertez, 1979). According to this classification, aphasia is
broadly divided into non-fluent and fluent aphasia types.
Based on the site of lesion, the symptoms will vary accordingly. However, no two
aphasic individuals with similar subtype of aphasia will exhibit an identical symptom profile.
Thus, it is of utmost importance that the rehabilitation process is modified to suit the needs of
a particular patient.
Hence, it is imperative to have materials which are language specific and are readily
available with specific activities, based on the therapeutic approaches documented in the
literature.
Authors have proposed a number of therapy techniques and manuals for the
treatment of aphasia but, majority of them refers to the western population (English language)
for example, the Manual for Aphasia Therapy developed by Longerich (1968), An Auditory
and Verbal Task Hierarchy by Ross and Spencer (1980). In the Indian context, Manual for
Adult Non-Fluent Aphasia Therapy-in Hindi (MANAT-H, Deshpande, 2004), Manual for
Adult Non-Fluent Aphasia Therapy-in Kannada (MANAT-K,Venugopal, 2005), have been
_________________________
* Reader in Speech Pathology, All India Institute of Speech and Hearing, Mysore, India
email: [email protected]
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
62
developed which mainly focuses on the non-fluent type of aphasia and Word Retrieval
Manual:Hindi Aphasics (WORM- H,Praful, 2006) focuses on anomic type of aphasia. Till
date fewer attempts have been made to develop a manual for the fluent type of aphasia (for
adults) in the Indian context.
Hence, due to the vast ethno cultural and language variances, it is imperative to have
language-based manuals in Indian clinical situation. Thus, the present manual in Kannada is
proposed for the treatment of adults with fluent type of aphasia.
The aim of this study is to develop a manual for the treatment of fluent aphasia in
Kannada for adults.
Method
The present manual is designed based on the documented principles and guidelines
prescribed in the literature for persons with fluent aphasia. The illustrations of various
activities are based on the principles of aphasia management. Literature regarding fluent
aphasia and intervention strategies was reviewed from books, journals and internet sources.
The collected information from these resources are compiled and organized.
The treatment parameters are formulated under the following headings:
Functional communication (FC)
Auditory comprehension (AC)
Repetition (R)
Expression (E)
Naming (N)
Reading and writing (R&W)
Vocabulary and sentences used in everyday situation were chosen as training material.
Pictures for the manual were drawn by professional artists. The training material was tested
for familiarity by five speech language pathologists (SLPs), who were native speaker of
Kannada. The pictures were also being tested for ambiguity. The SLPs were asked to
comment on the appropriateness and hierarchy of the items used in the training material.
Development of the manual
The present manual comprises of six domains, which are taken based on the deficits
seen in various skills in persons with fluent aphasia. The various domains are as listed below:
Functional communication (FC)
Auditory comprehension (AC)
Repetition (R)
Expression (E)
Naming (N)
Reading and writing(R&W)
Each of these domains is further sub-divided into several sub-sections:
MAFAT - K
63
Functional Communication (FC)
In this domain, aspects related to daily living like nouns, common verbs which are basic and
applicable in daily life is considered. The various aspects covered under functional skills are:
A. Responding to own name
B. Recognition of family members
C. Recognition of familiar objects
D. Comprehension of simple verbal commands
E. Comprehension of action verbs
F. Functional verbal language
G. Activities of daily living
H. Activities of independence
Auditory Comprehension (AC)
The focus of this section is to improve the auditory comprehension aspect at various
linguistic levels. These levels are
I. Semantic level
II. Syntax level
III. Discourse level
I. Semantic level
A. Gross phonemic level:
Level 1: Widely variant phonemes in a word.
Level 2: Lesser variant phonemes in a word
B. Finer phonemic level
Level 1: Finer phonemic difference in bisyllabic words.
Level 2: Finer phonemic difference in trisyllabic/polysyllabic words.
C. Word level
i. Vocabulary
ii. Antonyms
iii. Synonyms
iv. Syntagmatic and paradigmatic relations
II. Syntax level
A. Person Number Gender markers (PNG)
B. Tenses
C. Plurals
D. Answering yes-no (polar) questions
E. Following body part command
F. Following commands with visual stimuli
G. Identification of objects described by function
H. Identification of objects named serially
I. Sentence types
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
64
III. Discourse level
A. Listening comprehension
B. Reading comprehension
Repetition (R)
This domain is again sub-divided into the following sub-sections:
A. Automatic speech
B. Vocabulary
C. Phrases and Sentences
The person with aphasia is expected to repeat using auditory and graphic cues.
Naming (N)
This domain is again sub-divided into the following sub-sections:
A. Confrontation naming
B. Responsive naming
C. Lexical generative naming: Phoneme fluency, word fluency, category specific
Cueing hierarchies are provided for each sub-section.
Expression (E)
Main aim of this domain is to reduce paraphasias, perseverations, neologisms, and
jargon utterances.
The material and strategies designed to improve the expression skills have been
chosen keeping the therapy techniques as base.
Reading and Writing (R&W)
This domain aims at improving the reading and writing skills of persons with aphasia
at both functional and advanced levels. A sub- section on arithmetic skills has also
been included through which the concept of money and its usage can be worked upon.
This domain is sub-divided into four sub-sections as listed below:
A. Functional reading and writing
B. Advanced reading
C. Advanced writing
D. Arithmetic skills
The activities of each sub-section have been arranged in hierarchical order along with
its stimulus and response mode hierarchy. Scoring pattern and progress criteria have been
provided in the beginning of each sub-section. Overall progress criterion is also provided for
each domain and its sub-sections.
Treatment recording sheet
The treatment recording sheets (Adopted from Treatment protocols for language
disorders in children, Hegde, M.N., 2003) have been included for each domain and sub-
MAFAT - K
65
section. It is expected that the clinician follows these sheets to record the target activity,
mode of presentation of the stimuli, mode of responses, target skills, number of trials and
percentage of the correct responses.
Scoring
0 = No response/ incorrect response/ unintelligible response
1/2= Partially correct and intelligible response
1 = Fully correct and intelligible response
Progression criterion list
Begin with Functional Communication (FC) and Auditory Comprehension (AC)
domains simultaneously. Only when the total score in each of these domains reaches
50%, move to the next level.
Next, begin with the activities of functional reading and writing domain. When a score
of 25% is achieved in functional reading and writing and 75% in FC, AC each, proceed
to the next level.
Now, introduce the activities of repetition. When the person with aphasia scores 25% on
repetition, 50% in functional reading and writing and 100% on FC, AC, advance to the
next level.
Now, introduce the activities of expression. When 25% score is reached in expression
domain and 50% in repetition and 75% in functional reading and writing move to the
next level.
The activities in expression should be continued till a score of 75% is achieved.
Next, proceed to the naming domain and work on it till at least 75% score is achieved.
The advanced reading, writing and arithmetic skills should be introduced now and
worked upon till 75% of the score is achieved.
At the end of the manual it is expected that in each domain the performance of person
with aphasia is 75% or above and in overall 90% and above.
Implications of the study
The manual is quite flexible and can be used by speech language pathologists,
student clinician and caregivers of persons with fluent aphasia. Further, the training
material in this manual can be easily modified by the clinicians to suit the individual
needs of persons with aphasia.
References
Ball, M.J., & Damacio, J.S. (2007). Clinical Aphasiology: Future Directions. New York:
Psychology Press, Taylor and Francis Group.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
66
Bandur, D.L., & Shewan, C.M. (1986). Treatment of Aphasia: A Language-Oriented
Approach. London: Taylor & Francis Ltd., Benson, D.F., & Ardilla, A. (1996).
Aphasia. New York: Oxford University Press.
Code, C., & Muller, J. (1989). Aphasia Therapy. London: Whurr Publishers.
Despande, R. (2004). Manual for Adult Non-Fluent Aphasia Therapy - In Hindi. Unpublished
master’s dissertation submitted to University of Mysore, Mysore, India.
Duchan, J.F., & Byng, S. (2004). Challenging Aphasia Therapies: Broadening the Discourse
and Extending the Boundaries. New York: Psychology Press, Taylor and Francis
Group.
Goodglass, H., & Kaplan, E. (1983). The Assessment of Aphasia and Related Disorders. (2nd
Ed). Philadelphia: Lea and Febiger.
Goswami, S.P., & George, A. (2006). ISHA Monograph, Adult Aphasia: Language
Intervention. A publication of the Indian Speech and Hearing Association.
Hegde, M.N. (2003). Treatment protocols for language disorders in children. Vol. 2. Social
Communication. San Diego: Plural Publishing Inc.
Helm-Estabrooks, N., Fitzpatrick, P.M., & Baressi, B.N. (1982). Visual Action Therapy for
Global Aphasia. Journal of Speech and Hearing Disorders, 47, 385-389.
Helm-Estabrooks, N., & Holland, A.L. (1998). Approaches to the Treatment of Aphasia. San
Diego: Singular Publishing Group Inc.
Kertesz, A. (1979). Aphasia and Associated Disorders. New York: Grune & Stratton, Inc.
LaPointe, L.L. (1990). Aphasia and Related Neurogenic Language Disorders. New York:
Thieme Publishers, Inc.
Martin, B.R. (1962). Communicative Aids for the Adult Aphasic. Illinois: Charles.C.Thomas
Publishers.
Prema, K.S. (1997). Reading Acquisition Profile in Kannada. Unpublished Doctoral Thesis,
submitted to University of Mysore, Mysore, India
Ross, D., & Spencer, S. (1980). Aphasia Rehabilitation- An Auditory and Verbal Task
Hierarchy. Illinois: Charles.C.Thomas.
Shinde, P. (2006). Word Retrieval Manual: Hindi Aphasics. Unpublished master’s
dissertation submitted to University of Mysore, Mysore, India.
Venugopal. (2005). Manual for Adult Non-Fluent Aphasia Therapy - In Kannada.
Unpublished master’s dissertation submitted to University of Mysore, Mysore, India.
MAFAT - K
67
Normative Nasalance Value in Malayalam
67
Normative Nasalance Value in Malayalam Language
Devi T.R. & M. Pushpavathi*
Abstract
Assessment of nasality disorders in speech is traditionally proved to be a difficult perceptual
task for speech pathologists. Clients with velopharyngeal inadequacy are suggested for surgery or
speech therapy as a treatment option. Hence an accurate assessment of the nasality is critical, as this
provides valuable information for the suitable treatment. The use of instrumentation has become an
important part of the assessment and treatment of individuals with velopharyngeal dysfunction. Awan
(1996) developed a computer hardware/ software system called Nasal View (the Nasal Acquisition
System). NasalView is a cost-effective, software/hardware system designed for the data acquisition,
clinical analysis and treatment of nasal resonance disorders. Several studies have shown that
nasalance of normal speech is sensitivity to the phonetic composition of the speech stimuli, native
language, regional dialect, age, and gender. As there are very few established normalized nasalance
score, there is a strong need for establishment of regional norms. Hence ,the present study aimed at
developing the normative nasalance scores for Malayalam language across gender and stimuli in
adults. The study consisted of two groups with 50 normal adult males and females in the age range of
18 years to 35 years in each group who were native speakers of Malayalam. Stimulus set- I consisted
of five oral and five nasal sentences and stimulus set- II consisted of one nasal and one oral
paragraph. The subjects were instructed to read the sentences and paragraphs and NasalView system
was used to acquire the nasalance value for all the stimuli. The results indicated higher nasalance
values for females than males across nasal and oral stimuli. In males, the nasalance value for oral
sentence was 51.19%, nasal paragraph value was 51.43%, oral sentences value was 21.64% and oral
paragraph value was 21.36%. In females, nasalance value for nasal sentences was 57.55%, for nasal
paragraph was 56.93%, for oral sentences was 24.78% and for oral paragraph was 23.16%. The
presence of gender differences could possibly be attributed to the underlying structural and functional
differences across gender.
Key words: Nasal View, Normative nasalance value, Malayalam speaker.
Introduction
To assess and study nasalization and disorders of nasalization, speech language
pathologists and otorhinolaryngologists relay on a combination of direct and indirect
assessment procedures (Shprintzen & Bardach, 1995). Direct methods of visualization of the
velopharyngeal valve include multi view Video fluoroscopy and Nasopharyngoscopy, where
as indirect or non-visualizing procedures are illustrated by the mirror test and aerodynamic
and acoustic investigations (Van Lierde et al., 2001).
Speech language pathologist prefers the indirect methods, since it is noninvasive and
does not require additionally the medical professional support. Therefore, there is a need, in
the field of speech pathology for convenient and reliable systems to monitor velar action
___________________________
* Reader in Speech Pathology, All India Institute of Speech and Hearing, Mysore, India
email: [email protected]
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
68
during speech, both to give the clinician a measure of such action and to provide a means of
feedback for the person trying to improve velar control.
The speech stimuli typically used to measure nasality have considered the nasal
passage, the Zoo passage and the Rainbow passage (Fletcher et al., 1989). These passages
differ in that the Zoo passage contains no nasal phonemes, the Rainbow passage contains
about 11% nasal phonemes and the Nasal sentences are 35% nasal phonemes. Apart from this
the oral syllable, nasal syllable and sentences loaded with high pressure consonants, oral
consonants and nasal consonants are also used for the assessment.
Development of Normative Data across languages
Since the NasalView was introduced in 1996, many articles have appeared in the
literature on developing the normative data across stimuli in various languages. But there are
only few articles which focused on development of normative data on nasalance value and
comparing nasalance value across stimuli and across gender. The following review
summarizes the articles which aimed at developing normative data on nasalance, comparing
nasalance across gender and stimuli.
Awan, S. (I998) gave preliminary normative data collected from groups of 161 male
and female children between the age range of 5 and 14 years and 20 male and female adults
between the age range of 18 and 30 years. To assess varying degrees of normal nasal
resonance, NasalView system was used and subjects were asked to read the Zoo passage, the
Rainbow passage and the Nasal sentences. Results revealed a mean nasalance value of
48.44% for nasal passage, 34.19% for Rainbow passage and 24.67% for Zoo passage.
Tim Bressmann et al. (2000) did a study to evaluate two new measures derived from
mean nasalance data: the nasalance distance (range between maximum and minimum
nasalance) and the nasalance ratio (minimum nasalance divided by maximum nasalance) and
proposed that implementing these two measure would help to overcome the problems of
differences in magnitudes of mean nasalance in speakers with perceptually normal nasal
resonance and the problems in disagreement between categorizations of speech based on
nasalance scores and perceptual judgments. Subjects were 133 cleft lip and palate patients in
the age range of 10 to 66 years with a mean age of 17 years with normal nasal resonance or
varying degrees of hypernasality. Results for the sentence stimuli ranged from 64.4% to
89.6% sensitivity and from 91.2% to 94.1% specificity. The cutoff scores for the nasalance
distance, nasal sentences–oral sentences, were 23.16 and cutoff scores for the nasalance ratio
oral/nasal sentences were 0.539. When the analysis was limited to only one nonnasal and one
nasal sentence, results ranged from 79.7% to 87.5% sensitivity and from 88.2% to 97.1%
specificity. The cutoff scores for the nasalance distance nasal sentence– oral sentence were
26.1 and cutoff scores for the nasalance ratio oral/nasal sentence were 0.502. The authors
concluded that nasalance distance and nasalance ratio measures derived from sentence stimuli
are two useful measures that are easy to apply.
Awan, S (2001) conducted a study to investigate the age and gender effects on
measures of RMS nasalance. The NasalView system was used in the collection of normative
data from groups of male and female children between the ages of 5 to14 years with an
Normative Nasalance Value in Malayalam
69
addition of 20 adult males and 20 adult females in the age range of 18- 30 years. Subjects
were asked to read the Zoo passage, the Rainbow passage and the Nasal sentences while
wearing the NasalView headgear. Subjects were divided into 6 groups based on the age
groups. The nasalance values were 46.68% for Nasal Passage, 32.95% for Rainbow passage
and 23.39% for Zoo passage. Results indicated no significant differences between males and
females in terms of RMS nasalance in any age group. Though no significant differences in
RMS nasalance were observed across gender, the female subjects were observed to produce
adult levels of RMS nasalance at a younger age than the male subjects (11to12 years vs.
13to14 years). The author interpreted that this observation may be due to earlier facial
skeletal and nasopharyngeal maturation in females as compared to males.
Awan. S et al, (2001) provided normative data in English language from 255 normal
subjects who included both 203 children in the age range of 5 to 12 years and 52 adults who
were 13 years and above and all the subjects were native speakers of English language. The
author used NasalView system to collect normative nasalance data for Zoo passage, Rainbow
passage and Nasal sentences. The results revealed approximately equal nasalance score for
both males and females across age and gender. Mean nasalance values for adult females were
26.71% for Zoo passage, 37.11% for Rainbow passage and 51.18% for nasal passage. Mean
nasalance score for adult males were 26.02% for Zoo passage, 35.58% for Rainbow passage
and 51.18% for nasal passage. There was a slight increase in nasality in females compared to
males in all the three passages. The results also revealed a significant difference across the
three stimuli in both males and females.
Keuttner, C et al., (2003) obtained normal nasalance values in German language using
NasalView system. A total of 50 individuals with normal speech development were examined
with NasalView. The median age was 14 years (range 11-20 years). The test material used
comprised the vowels /a:/, /e:/, /i:/, /o:/, /u:/, the sentences S(1): "Die Schokolade ist sehr
lecker."("This chocolate is very tasty.") and S(2): "Nenne meine Mamma Mimmi."("Call my
mummy Mimmi."), and three mixed oral-nasal text passages, long text LT (1): "Nordwind
und Sonne"("Northwind and sun"), long text LT(2): "Kindergeburtstag"("A child's birthday
party") and long text LT(3): "Der grosse Gesang"("A famous song"). Results revealed that
the mean nasalance for the vowels were 35.9%, for S (1) (containing no nasal consonants)
24.9%, and for S (2) 69.6% (with many nasal sounds). The results for the mixed oral-nasal
text passages were 42.1% for LT (1), 36.9% for LT (2) and 38.2% for LT (3). The above
results revealed significant differences across the nasalance values for all the three stimuli.
The nasalance value for the nasal sentence was significantly greater than the oral sentence.
The nasalance values also varied significantly across the three oral-nasal text passages.
Tim Bressmann (2005) did a comparison of nasalance values obtained with the
Nasometer, the NasalView, and the Oronasal system. Objective of the study was to compare
nasalance values obtained with the Nasometer, the NasalView, and the OroNasal system;
evaluate test-retest reliability of the three systems; and explore whether three common text
passages used for nasalance analysis could be shortened to a sentence each. Subjects were
seventy-six adults with normal speech and hearing (mean age 26.5 years). Subjects read the
complete Zoo passage, Rainbow passage, and Nasal sentences. The subjects also read the first
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
70
sentence from the three passages for the purpose of comparing the nasalance values for the
long passages with their corresponding sentences. The combined mean nasalance for adult
males and females obtained from NasalView system were 21.1% for Zoo passage, 35.2% for
Rainbow passage and 55.74% for Nasal sentences. The mean nasalance for the first sentence
from Zoo passage was 21.2%, for the first sentence from Rainbow passage was 34.72% and
for the first sentence from Nasal sentences was 56.9%. Results revealed that the Nasometer
had the lowest nasalance values for the Zoo passage (13.45%). The NasalView had the
highest nasalance values for the Rainbow passage (35.2%). The OroNasal system had the
lowest nasalance values for the Nasal sentences (49.72%).
André Eckardt et al., (2007) conducted a study to compare the nasalance using
NasalView system in patients with maxillary defects who had undergone reconstruction
versus obturation. Further, these groups of patients were compared with normal controls. This
study included 28 subjects. Ten patients had obturators (group l) and in 18 patients the
maxilla was biologically reconstructed with different techniques (group 2). Sound pressure of
nasal and oral airways was assessed separately using NasalView system and standardized
German texts and the results were compared with an uncompromised sample of patients.
Results revealed no significant differences across both the groups. Furthermore, the achieved
values of nasalance were similar to healthy individuals.
Tim Bressmann, Paula Klaiman and Simone Fischbach (2006) compared the data
from normal subjects and cleft palate speakers between three systems for nasalance analysis.
The purpose of this study was to collect and compare the data with all three systems from a
group of normal speakers and a group of cleft palate patients with hypernasality. In particular,
the authors were interested in the diagnostic efficacy of the three systems, as indicated by
sensitivity and specificity. They considered fifty subjects with the mean age of 26.0 years.
Thirty-one participants were female and nineteen were male. Nineteen hypernasal patients
with cleft palate were also included in the experimental group with the age range of 11-19
years. Eight participants were female and eleven were male. The Zoo passage and the Nasal
sentences were used to assess the nasalance value. The results showed that the nasometer had
the lowest nasalance values for the non-nasal Zoo passage (13.45%) for the normal
participants but the highest values for the moderately hypernasal cleft palate patients
(34.06%). The lowest values for the Nasal sentences were recorded with the oronasal system
for the normal participants (49.72%) as well as the patients with cleft palate with moderate
hypernasality (45.93%). The lowest nasalance values for the mildly hypernasal group was
collected with the nasometer. The authors attributed these differences between the nasalance
magnitudes of the same speakers when measured with the three different systems to the
differences of the signal processing routines of the three instruments.
Arya Pravesh (2009) conducted a study to establish the normative nasalance values in
Hindi language, using NasalView system, and to investigate differences in nasalance values
across gender and across stimuli. Subjects were fifty adult males and females in the age range
of 18- 35 years. The stimulus materials included Nasal sentences, oral sentences, nasal
paragraph and oral paragraph in Hindi language. Mean nasalance values in males were
47.04% for nasal sentence, 22.48% for oral sentence, 46.85% for nasal paragraph and 21.59%
Normative Nasalance Value in Malayalam
71
for oral paragraph. Mean nasalance values in females were 50.42% for nasal sentence,
22.59% for oral sentence, 49.74% for nasal paragraph and 22.23% for oral paragraph. The
results revealed significant differences across gender for nasal stimuli but not for the oral
stimuli. The significant differences were not seen across oral stimuli and across nasal stimuli
within gender groups. There was significant difference across nasal and oral stimuli at a
0.001 level of significance.
The above review of literature indicates the need of establishing normative data in
different languages, which is useful for assessing resonance disorders. One of the main aims
of the current study was to establish normative nasalance values for adults in Malayalam
language across gender and stimuli.
Aims of the study
To develop normative data on nasalance value across gender and nasalance deviation
in Malayalam language for oral and nasal sentences and paragraphs for adults.
To study the effect of stimuli on nasalance value across gender.
To study the effect of gender on nasalance value across stimuli
Method
Subjects
One hundred subjects were included in the study. The subjects were selected based on the
inclusionary and exclusionary criteria by experienced speech language pathologist.
Inclusion criteria
The subjects should be native speakers of the Malayalam language in the age range of
18-35 years.
The minimum educational qualification should be higher secondary education.
Should possess sufficient auditory and visual acuity.
Exclusion criteria
No neurologic, sensory, motor, cognitive or communicative impairments.
Subjects were divided into two groups. First group consisted of fifty females in the
age range of 18-35 years and second group consisted of fifty males in the age range of 18-35
years. Table 1 depicts the subject details.
Table 1: The subject details.
Gender Age range Sample size
Male 18-35 years 50
Female 18-35 years 50
Total 100
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
72
Stimuli
Two sets of stimulus material were prepared by an experienced speech language
pathologist whose mother tongue is Malayalam.
Set-1 consisted of 5 nasal and 5 oral sentences of equal length (5-6 syllables)
Set-2 consisted of nasal and oral paragraph containing 6 sentences each.
Procedure
Initially to find the content validity of the stimulus materials, 10 nasal and 10 oral
sentences were prepared using common nasal and oral words selected from Malayalam
phonetic reader (B S Kumari, 1972). The words selected satisfied the criteria of above 90%
nasal sounds for nasal words and below 10% nasal sounds for oral words. These ten
sentences, each for oral and nasal sentences, were audio recorded by a speech language
pathologist who was a native Malayalam speaker and were given for perceptual judgment of
degree of nasality to 10 speech language pathologists who had at least one year experience in
the field. A five point perceptual rating scale was used. Rating of „0‟ indicated fully oral or
no nasality; „4‟ indicated highly nasal stimuli. After perceptual evaluation 5 nasal sentences
with a perceptual nasality judgment of greater than or equal to 3, and 5 oral sentences with
perceptual nasality judgment of less than or equal to 1 were selected as the final stimulus
material set 1. Similarly the nasal and oral paragraph was prepared for the stimulus material
set 2.
Instrumentation
The Nasal View system was designed to run on a 486 or Pentium PC- compatible
system (also requires VGA graphics capability and 4 Mbytes RAM) the system includes the
following components: Headgear, Preamplification and calibration hardware, Computer
hardware, and Computer software
Data collection
The subjects were assessed and recorded individually. After selecting the subjects,
they were seated comfortably and the NasalView head gear was placed on the subject‟s head.
The position of the NasalView head gear was adjusted and secured firmly in accordance with
the manufacturer‟s instructions. The subjects were instructed to read the speech stimuli. The
nasalance trace was continuously monitored throughout each recording to ensure that the data
were being captured. After the completion of each speech sample, the nasalance trace was
stored in a computer file for further analysis.
Data analysis
Data was obtained for the two groups of stimuli. The files for all speech samples were
subjected to calculation of the mean, maximum and minimum nasalance for each stimulus
category. All computations were performed using the SPSS statistical package (SPSS 16.0 for
Windows). The data was analyzed for adult males and adult females separately using various
programs like,
Independent t test to compare nasalance values across gender.
Normative Nasalance Value in Malayalam
73
Paired t test to find the effect of type and the effect of nasality on nasalance values
within gender, and
Mixed ANOVA to perform multiple pair wise comparisons between the mean
nasalance values for; the two within subject variables i.e. type and nasality with gender
as between subject factor, and, the between subject variable, i.e. gender.
Results
I. Normative value for nasalance
Table 2: indicates the normative nasalance value for adults across gender. The mean
nasalance values for sentence and paragraph are depicted along with standard deviation for
oral and nasal stimuli. The upper and lower limit for oral sentences, nasal sentences, oral
paragraph and nasal paragraph is also showed in the table.
Table 2: Normative nasalance value (in %) for adult males and adult females.
In mixed ANOVA, significant interaction was observed between the variables under
study. Hence the variables were separately analyzed as explained below. Independent t test
was done for comparison of nasalance value across gender within each type and nasality. The
following table: 3 shows the results.
Table 3: Comparison of stimulus type and nasality across gender.
Type/ nasality t (1, 98)
Oral sentence 3.168**
Oral paragraph 1.504*
Nasal sentence 6.087***
Nasal paragraph 5.608***
* Significance at 0.05 level.
** Significance at 0.01 level.
*** Significance at 0.001 level.
Gender N Category Mean (S.D) Max (SD) Min (SD)
Adult
Male
50 Oral
sentences
21.64 (4.22) 57.56 (9.16) 7.95 (1.90)
Nasal
sentences
51.19 (5.32) 85.89 (4.59) 18.61 (5.39)
Oral
Paragraph
21.36 (4.73) 71.26 (10.31) 5.92 (1.33)
Nasal
paragraph
51.43 (5.29) 87.38 (4.19) 7.93 (2.11)
Adult
Female
50 Oral
sentences
24.78 (5.61) 63.93 (7.74) 8.36 (2.92)
Nasal
sentences
57.55 (5.10) 90.22 (2.71) 22.98 (6.48)
Oral
paragraph
23.058 (6.46) 73.25 (8.74) 6.40 (2.63)
Nasal
paragraph
56.93 (4.47) 91.10 (2.13) 9.19 (3.89)
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
74
Results revealed significant differences for nasal sentences (p< 0.001), nasal
paragraphs (p< 0.001), oral sentence (p< 0.01) and oral paragraphs (p< 0.05) across gender.
Within gender differences between nasal stimuli and oral stimuli and within gender
differences between sentences and paragraphs were investigated using paired t test. Table 4:
shows the results.
Table 4: within gender comparison across stimuli.
* Significance at 0.01 level.
** Significance at 0.001 level.
The results revealed that there was significant difference across oral and nasal stimuli
in males and females (p< 0.001). There were no significant differences found across nasal
sentence and nasal paragraph neither in males (p> 0.05) nor in females (p> 0.05). There was
no significant difference across oral sentence and oral paragraph in males (p>0.05) but
significant difference was found across oral sentence and oral paragraph in females (p> 0.01).
II. Effect of Gender on Nasalance Values
Females exhibited higher nasalance value for nasal sentences which was 57.54% and
nasal paragraph which was 56.92% compared to males whose values were 51.19% for nasal
sentence and 51.43% for nasal paragraph. Independent t test revealed significant difference
across gender for nasal stimuli. Female subjects also exhibited higher nasalance value for
oral sentences which was 24.78% and for oral paragraphs which was 23.05% compared to
males for whom nasalance value for oral sentence was (21.64%) and for oral paragraph
which was 21.35%. Independent t test revealed significant difference across gender for oral
stimuli.
Interaction of stimulus type, nasality and gender
Mixed ANOVA was carried out to find out the significant differences across the two
within subject variables, i.e. type and nasality with gender as between subject factor, and the
between subject variable, i.e. gender. Table 5: depicts the F value as well as the level of
significance for these variables and their interaction effect.
Gender Source compared T (1,49)
Male Nasal sentence vs. oral sentence 38.70**
Nasal paragraph vs. oral paragraph 32.62**
Nasal sentence vs. nasal paragraph 0.54
Oral sentence vs. oral paragraph 0.66
Female
Nasal sentence vs. oral sentence 44.01**
Nasal paragraph vs. oral paragraph 45.29**
Nasal sentence vs. nasal paragraph 1.23
Oral sentence vs. oral paragraph 3.69*
Normative Nasalance Value in Malayalam
75
Table 5: interaction between stimulus type nasality and gender.
* Significance at 0.05 level.
** Significance at 0.01 level.
*** Significance at 0.001 level.
Table 5: reveals the interaction effect of the within and between subject variables. The
table also shows the level of significance for each interaction. The significant difference was
seen for within type (p< 0.05), within nasality (p< 0.001), and within gender (p < 0.001),
interaction between type and gender (p < 0.05), interaction between nasality and gender (p<
0.01) and interaction between type and nasality (p< 0.05). There was no main effect seen in
the interaction between type, nasality and gender (p> 0.05).
Discussion
Effect of gender on nasalance values
The presence of differences in nasalance value across gender could possibly be
attributed to the underlying structural and functional differences across gender. There are
several studies reported in the literature which supports the findings of the present study.
Jayakumar (2005) in his study on developing normative data of nasalance values for
Kannada language in 50 children and 50 adults reported significant differences across gender
in adults. The author attributed this difference to the basic structural and functional
differences across gender.
Mahesh and Pushpavathi (2008) compared the nasalance values for Rainbow passage
across gender and reported significant differences across gender. Females (31.39%) had
higher nasalance values compared to males (27.93%). Also the standard deviations were
higher in females compared to males indicating higher variability among females. The
authors attributed the gender differences in nasalance to basic structural and functional
differences. They also opined that the resonance of voice is influenced by the size, shape and
surface of infraglottal and supraglottal resonating structures and cavities.
Arya P (2009) conducted a study to establish the normative nasalance values in Hindi
language, using NasalView system, and to investigate differences in nasalance values across
gender and across stimuli. Subjects were fifty adult males and females in the age range of 18-
35 years. The stimulus included nasal sentences, oral sentences, nasal paragraph and oral
paragraph in Hindi language. The results revealed significant gender differences for nasal
stimuli but not for the oral stimuli.
Source F (1, 98)
Main effect of Type 5.023*
Main effect of Nasality 3505.312***
Main effect of Gender 25.258***
Interaction between Type* Gender 4.587**
Interaction between Nasality* Gender 10.777*
Interaction between Type* Nasality 4.829*
Interaction between Type*nasality*Gender 0.625
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
76
In an attempt to determine the physiology behind higher nasalance values observed in
females, several authors have investigated the anatomical and/or physiological
velopharyngeal parameters. (Kuehn & Moon, 1998; McKerns & Bzoch, 1970; Zajac &
Mayo, 1996).
McKerns and Bzoch (1970) used cinefluoroscopy to observe velopharyngeal valving
and found that two basic configurations of VP closure existed in relation to gender during
connected speech production. Females were observed to have a shorter velum, use less velar
elevation and a greater amount of velar contact against the pharyngeal wall to achieve
closure. The authors attributed these differences to the sites of muscle insertion involved in
velopharyngeal closure (i.e., levator palatini, palatoglossus, and palatopharyngeus muscles),
and oral and pharyngeal dimensions (McKerns & Bzoch, 1970).
Zajac and Mayo (1996) studied the aerodynamic and temporal aspects of
velopharyngeal function in 42 normal young Caucasian adult (21 male and 21 female; mean
age 24, 23) speakers during production of the nasal-plosive sequence /mp/ in the strong-weak
stressed word, “hamper.” Significant differences across gender was demonstrated where
females exhibited significantly lower levels of peak intraoral air pressure and longer
durations in the rise of pressure during the production of the /p/ segment, indicating less
occlusion of the velopharyngeal port. It was speculated that differences in preferred intensity
levels and respiratory and velar physiology between genders might have accounted for some
of the findings (Zajac & Mayo, 1996).
Kuehn and Moon (1998) investigated velopharyngeal closure force in varying
phonetic contexts using 14 normal college-aged participants (seven male, seven female).
Statistically significant differences across gender was not found in velopharyngeal closure
force using different vowel and consonant productions, voicing, place, manner, or sequencing
conditions. It has been shown that in regard to gender, there are several possible factors
causing variation within nasalance values. There is no accepted explanation of these
differences.
Many studies have reported that a significant difference was not evident in nasalance
values across gender (Trindade et al., 1997; Whitehil (2001); Christina E et al, 2000; Daniel,
Z. H, 2001; Awan,S, 1999; Awan,S, 2001; Keuttner C et al, 2003). The results of the present
study does not support the findings.
Trindade et al., (1997) compared the nasalance values across gender. Based on the
results, they opined that a statistically significant differences was not found in the nasalance
values across gender.
Awan,S (1999, 2001) discussed the comparison of nasalance across age groups and
gender. The study included age groups from 5-14 years with an addition of 20 adult males
and 20 adult females in the age range of 18- 30 years. Results indicated no significant
differences between males and females in terms of RMS nasalance at any age group.
Keuttner, C et al.,(2003) obtained normal nasalance values in German language using
NasalView system. A total of 50 individuals with normal speech development were examined
Normative Nasalance Value in Malayalam
77
with NasalView. The median age was 14 years (range 11-20 years). This study also reported
of no differences in nasalance s across gender.
I. Effect of Stimuli on Nasalance Values
The nasalance values were compared across stimuli. Results revealed a significant
difference across nasal and oral stimuli at 0.001 level of significance. The reason for this
difference could be attributed to the characteristic phonetic structure of the nasal and oral
stimuli. Production of nasal stimuli induces transfer of acoustic energy into the nasal cavity
through the open velopharyngeal port which is picked up by the nasal microphone of
NasalView. During the production of oral sounds, the velopharyngeal port is closed which
accounts for the reduction in transfer of acoustic energy into the nose and an increase in oral
acoustic energy which is picked up by the oral microphone of the NasalView headgear. The
observed variation in nasalance across oral and nasal stimuli could also be attributed to the
influence of phonetic nasal content of individual stimuli on the nasalance values, an effect
demonstrated by Fletcher, Adams and Mc Cutcheon (1989). On the whole, transpalatal
transfer of energy accounts for nasalance of speech stimuli (Emily et al, 2006). Mean
nasalance scores may be influenced by the phonemic characteristics of a language.
Consequently, the number of nasal sounds in that language as well as frequency of
occurrence of nasal sounds may be an important factor. Among the Indian languages
Malayalam has got more nasal resonance than any other languages. Malayalam has six nasal
consonants, all of which are prevalently used (bilabial, alveolar, palatal, retroflex and velar).
In addition to these nasal sounds, nasalization of vowels is also highly prevalent, which may
account for increased nasal resonance (Ramakrishna B.S, 1962).
When sentences and paragraph were compared within gender using pair t-test, there
were no significant differences across the stimuli (p>0.05) except within oral stimulus group
in females which showed significant difference at 0.001 level. The results from the present
study is supported by past studies done by Tim Bressmann, 2005; Watterson et al, 1996;
Watterson et al. 1999; Jan Van Doom, 2006.
Watterson et al, (1996) compared the nasalance values obtained from Turtle passage
and Mouse passage with the nasalance values obtained from Zoo passage and Rainbow
passage respectively. The results of the study revealed that there were no significant
differences between the passages compared. The authors concluded that nasalance measures
although sensitive to the degree of nasal phonemes in a stimulus, are not sensitive to the
remaining phonetic content.
Watterson et al. (1999) compared the nasalance measures for speech stimuli of four
different lengths. The standard for comparison was a 44- syllable passage. The 44- syllable
passage was compared to a 17- syllable passage, a 6- syllable passage and a 2- syllable word.
All stimuli were devoid of nasal consonants and were composed of only low pressure
consonants and vowels. The results showed that comparable measures of nasalance can be
obtained using stimuli as short as 6 syllable sentence, which, along with the 17- syllable
stimuli achieved high criterion validity, indicating that stimuli of that length could be
substituted for the longer 44- syllable passage.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
78
Bressmann et al. (2000) compared the nasalance for individual sentences from the
Heidelberg Rhinophonia Assessment to the complete nonnasal and nasal passages and
concluded that individual sentences could substitute the complete passages as the mean
values and mean differences for the two stimuli were similar.
Jan Van Doorn et al. (2008) conducted a study with the aim of establishing whether
nasalance values from shorter sections of three Swedish speech stimuli were equivalent to
those from their corresponding whole stimulus. Nasalance recordings for three Swedish
speech stimuli (oral, nasal and oronasal) were obtained from 29 typically developing Swedish
children (7–11 years). Results showed that shorter sequences of sentences could be
considered equivalent to the whole passage for the oral and nasal stimuli, provided that the
sentences were sequenced in order of increasing difference from the whole stimulus and that
those sentences with values significantly different from the whole passage were not included.
Tim Bressmann (2005) in his study on comparison between short and long stimuli
across instruments gave results for the Nasometer and the Oronasal system that were
contradictory to the present study. The author did not find equivalence of nasalance for a
single sentence compared with full length sentence for a group of normal adult speakers for
the standard English passages, the Zoo passage and the Rainbow passage when evaluated
using the Nasometer and Oronasal systems. The discrepancies in the findings from these
studies could be due to the differences in defining “equivalence” of scores, different types of
speech stimulus or different speaker groups (Jan Van Doorn et al. 2008).
Conclusions
The present study established normative nasalance score for adults and also
investigated the effect of gender and stimuli on nasalance value. Significant difference was
evident across gender. In males, the nasalance value for oral sentence was 51.19% (5.3),
nasal paragraph value was 51.43% (5.3), oral sentences value was 21.64% (4.2) and oral
paragraph value was 21.36% (4.7). In females, nasalance value for nasal sentences was
57.55% (5.7), for nasal paragraph was 56.93% (4.5), for oral sentences was 24.78% (5.6) and
for oral paragraph was 23.16% (6.4).
The presence of gender differences could possibly be attributed to the underlying
structural and functional differences across gender. As McKerns and Bzoch (1970) suggested
that velar length and elevation is greater for men compared to women. Differences in
preferred intensity levels and respiratory and velar physiology between genders may be
accounted for the differences in nasality (Zajac & Mayo, 1996). Additionally research has
demonstrated gender differences in vocal fold vibration pattern (Oates and Dacakis, 1997).
The present study also investigated the effect of stimuli on nasalance values. The
results revealed that the differences were evidence across gender for nasal stimuli and oral
stimuli. Significant difference was not observed across sentences and paragraphs within each
oral and nasal stimulus groups except within oral stimulus group in females. Over all the
present study adds to the body of evidence that there are gender and stimulus dependent
differences in nasalance scores. Clinically the normative data reported in the present study
may help identify and treating individuals with resonance disorders.
Normative Nasalance Value in Malayalam
79
References
Andre, E., Thomas, T., Andrea S., Marijana, H., and Christian, K. (2007). Nasalance in
patients with maxillary defects- Reconstruction versus obturation. European
Association for Cranio-Maxillofacial Surgery.
Arya, P. (2009). Normative Nasalance Value in Hindi Language. Unpublished mater‟s
dissertation submitted to the University of Mysore, Mysore (Personal
Communication).
Awan, S.N. (1996). Development of a low-cost nasalance acquisition system. In: Powell T,
(ed.) Pathologies of Speech and Language: Contributions of Clinical Phonetics and
Linguistics. New Orleans, LA: International Clinical Linguistics and Phonetics
Association, 211-217.
Awan, S.N. (1998). Analysis of nasalance: NasalView. In: Ziegler W, Deger K, (ed.),
Clinical Phonetics and Linguistic. London: Whurr-Publishers, 518-525.
Awan, S. N. (2001). Age and gender effects on measures of RMS nasalance. Clinical
linguistics & phonetics, 15, 117-122.
Awan, S.N., Jordan, R, Daniel Z.H., (2001). Preliminary Normative Data. In: Daniel, R,
Green., and Awan, N Shaheen, (Ed), Dr. Speech: User’s Guide. Tiger DRS, INC, 71-
74.
Bressmann, T., Sader, R., Whitehill, T.L., Awan, S.N., Zeilhofer, H.F., & Horch, H.H.
(2000). Nasalance distance and nasalance ratio: two new measures. Cleft Palate
Craniofacial Journal, 37, 248-256.
Bressman, T. (2005). Comparison of nasalance scores obtained with the Nasometer, the Nasal
View, and the OroNasal System. Cleft Palate Craniofacial Journal, 42, 423-433.
Bressman, T., Klaiman, P., & Fischbach, S. (2006). Same noses, different nasalance scores:
data from normal subjects and cleft palate speakers for three systems for nasalance
analysis. Clinical Linguistics and Phonetics, 20, 163-170.
Emily, L., & David J. Zajac. (2006). Estimation of Transpalatal Nasalance During Production
of Voiced Stop Consonants by Noncleft Speakers Using an Oral-Nasal Mask. Cleft
Palate Craniofacial Journal, 43, 6, 691-701.
Fletcher, S.G., Adams, L.E., & McCutcheon, M.J. (1989). Cleft palate speech assessment
through oral-nasal acoustic measures. In: Bzoch, K.R, (ed.). Communicative
Disorders Related to Cleft Lip and Palate. Boston: Little Brown, 246-257.
Jan, Van, Doorn., Ida, Bergh., & Karin, Brunnega. (2008). Optimizing stimulus length for
clinical nasalance measures in Swedish. Clinical Linguistics & Phonetics, 22, 355-
361.
Kumari, B. S. (1972). Malayalam Phonetic Reader. Mysore: Central Institute of Indian
Languages publication.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
80
Kuehn, D.P., & Moon, J.B. (1998). Velopharyngeal closure force and levator veli palatini
activation levels in varying phonetic contexts. Journal of Speech and Hearing
Research, 41, 51-62.
Küttner, C., Schönweiler, R., Seeberger, B., Dempf, R., Lisson, J., & Ptok, M. (2003).
Normal nasalance for the German language. Nasometric values for clinical use in
patients with cleft lip and palate. HNO, 51, 151-156.
Mahesh, S. M., & Pushpavathi, M. (2008). Nasalance value for Rainbow passage:
Normative value for non- native speakers. Journal of All India Institute of Speech and
Hearing, 27, 22-28.
McKerns, D., & K, R, Bzoch. (1970). Variations in velopharyngeal valving: The factor of
sex. Cleft Palate Journal, 7, 652-62.
Oates, J., & Dacakis, G. (1997). Voice changes in transsexuals. Venereology, 10, 178-187.
Ramakrishna, B. S., Nair, K. K., Chiplunkar, V. N., Atal, B. S., Ramachandra, V., &
Subramanian, R. (1962). Some Statistics of Indian Languages. In: Some Aspects of
the Relative Efficiencies of Indian Languages, Catholic Press, Ranchi, India, 12-17.
Shprintzen, R., & Bardach, J. (1995). Cleft palate speech management: A multi-disciplinary
Approach. Missouri: Mosby.
Jayakumar, T. (2005). Normative score for Nasometer in Kannada. An article published
from dissertation submitted to the University of Mysore, Mysore.
Trindade, I.E.K., Genero, K.F., & Dalton, R. M. (1997). Nasalance score for normal Brazilian
Portuguese Speakers. Brazil Journal of Dysmorphology Speech & Hearing Disorders,
1, 23-34.
Van Lierde, K.M., Wuyts, F.L., De Bodt, M., & Van Cauwenberge, P. (2001). Normative
value for normal nasal resonance in the speech of young Flemish adults. Cleft Palate
Craniofacial Journal, 38, 112-118.
Watterson, T., Hinton, J., & McFarlane, S. (1996). Novel stimuli for obtaining nasalance
measures from young children. Cleft Palate Craniofacial Journal, 33, 67-73.
Watterson, T., Lewis, K.E., & Foley-Homan, N. (1999). Effect of stimulus length on
nasalance scores. Cleft Palate Craniofacial Journal, 36, 243-247.
Zajac, D.J., Lutz, R., & Mayo, R. (1996). Microphone sensitivity as a source of variation in
nasalance scores. Journal of Speech and Hearing Research, 39, 1228-1231.
Normative Nasalance Value in Malayalam
81
Appendix PPENDIX- I
MALAYALAM STIMULI
Nasal Sentences
1. id¡ CË¡I i¡µy.
2. iz© iYêy« pzY¡.
3. id¡ i»y« pzY¡.
4. i» ixµ pzY¡.
5. ixdp© CË¡I pË¡.
Oral Sentences
1. Bq PeðxÀy ekÀy.
2. Pyeðy sxky DU¡À¡.
3. Ap¬ e¼y¤j ¥eUy¶¡.
4. e¼y Fmø§ KUy¶¡.
5. e¡Zyj pzU§ pyl÷¡.
Nasal Paragraph
pyixdI iYêy« p˧ dyË¡. DYêy¥ixd§ pyixdI pxµx© ¤ZxËy. DYêy¥ix¤©÷ pxdyd¡ i¡Ëy« iydy
px© pË¡ dyË¡. ey¤Ë DYêy pyixdI KxYx© dyËymø. DYêy i¡Ëy« dyË¡ dzµy.
Oral Paragraph
qqy PyöZ¥jxU§ Pxj ¤Pxby¶¡. Ap¬ Pxjjy« Alyjx¤Z De𧠥PªÀ¡. qqy PyöZ¤j qxsy¶¡.
PyöZ AZ§ ¤K¼¡ Pyky¶¡. A¥eðx¬ qqy K¢U¡Z« ¥Kxey¶¡. C¦ Ka ¥K¼§ Apª Pyky¶¡.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
82
APPENDIX- II
MALAYALAM IN IPA FORMAT
Symbolic Play and Language in Late Talkers
83
Symbolic Play and Language: Its Relationship in Late Talkers
Devika M. R. & N. Swapna*
Abstract
Symbolic play involves the representational use of objects and pretending or acting out a
concept. It has been reported that symbolic play corresponds to language development in typically
developing children and in children with developmental language disorders although there have been
some mixed views. Such studies investigating the relationship between symbolic play and language
are limited in the Indian context especially in the subgroup of children who are termed ‘late talkers’.
Hence the present study was undertaken to investigate the symbolic play behaviours and its
relationship with language in late talkers particularly those with an expressive language delay. In
addition, an attempt was made to determine the effect of manipulation of physical and social context
on symbolic play. The sample included 10 typically developing Kannada speaking children and 10
late talkers in two different age groups 24 to 30 months and 30-36 months. The symbolic play
behaviors and their play and language age were assessed. The results indicated that the late talkers
exhibited lesser presymbolic and symbolic play behaviours compared to the typically developing
group. A developmental trend in the play behaviours which was observed in the typically developing
group was not seen in the late talking group. Further the manipulation of the physical and social
context increased the overall frequency of the targeted response in both the groups. While the
typically developing group exhibited more number of play behaviours with instructions alone, the late
talking group required modeling along with instructions to elicit the majority of play behaviours. It
was also found that in both the groups play age correlated with receptive and expressive language.
The developmental differences between various play behaviors and implications for clinical
assessment and management are discussed.
Introduction
Play is defined as any voluntary activity engaged for the enjoyment it gives without
consideration of the end result (Piaget, 1962). Symbolic play is a type of play which involves
simultaneous representation of objects in two ways (real and pretend) and acting out a
concept as perceived by the performer (Venkatesan, 2004). This type of play develops
between 12-36 months in an orderly and predictable manner (Piaget, 1962; Mc.Cune-
Nicolich, 1977; Watson & Fischer, 1977; Bretherton, 1984).
Symbolic play reflects both symbolizing ability and conceptual knowledge and
therefore is considered to have closer links to language (Lewis, Boucher, Lupton & Watson,
2000). Human language whether written, spoken or signed, is unique in being a symbolic
communication system. The first few years of life are the crucial time in which a child
acquires their native language. One of the powerful tool or a prime medium for developing,
learning language and practicing their new acquisition is play. The infants‟ early knowledge
about the world of objects is reflected in their symbolic play behavior which contributes to
__________________________________
* Lecturer in Speech Pathology, All India Institute of Speech and Hearing, Mysore, India
email: [email protected]
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
84
later language development. Symbolic play skills are highly representational and abstract and
an assessment of the same would provide us with an insight about the child‟s communicative
abilities.
Several investigators have examined the relationship between play and language.
Although, many studies carried out in the West reveal a parallel relationship between play
and language (Piaget, 1962; Bates, 1976; Westby 1980; McCune-Nicolich, 1981; Baron-
Cohen, 1987; Casby & Della Corte, 1987; Beeghly, Weiss-Perry, & Cicchetti, 1990; Ogura,
1991; Doswell, Lewis, Boucher, & Sylva, 1994; McCune, 1995; Lyytenin & Laakso, 1997;
Lifter & Bloom, 1998; Tomasello, Striano & Rochat, 1999; Kitty, 2000), some studies have
evidenced paradoxical results. Some studies have also revealed that the interdependencies
between language and symbolic skills change over time. The symbolic play abilities of
children with language impairment (developmental/specific language impairment/expressive
language delay) also have been studied by many investigators (Lovell, Hoyle, & Siddall,
1968; Udwin & Yule, 1982; Terrell, Schwartz, Prelock, & Messick, 1984; Roth & Clark,
1987; Terrell & Schwartz, 1988; Rescorla & Goosens, 1992) who found that these children
do not appear to develop complex imaginative symbolic play or, if they do, it happens very
slowly. Moreover some of these studies have reported that symbolic play is delayed relative
to play of age mates but may be more advanced than the play of younger children matched on
expressive language at least in some respect. Thus the exact nature of play and language
relationship remains unclear. There is a dearth of such studies in the Indian context carried
out on late talking children especially those with only expressive language delay in the age
group of 2-3 years.
It is also a known fact that the quality and quantity of symbolic play can be enhanced
or dampened by manipulations of physical materials or social context. For example, McCune-
Nicolich and Fenson (1984) demonstrated that the presentation of organized toy sets appears
to facilitate the production of symbolic play sequences. Using Feuerstein‟s (1980)
terminology, modeling might be considered a form of social mediation by which the child
attains a higher level of performance than he/she can attain independently. Therefore an
attempt was also made to investigate the effect of the social mediation strategies namely,
instruction and modeling on the symbolic play abilities.
Aims of the study
1) To investigate the symbolic play behaviors in typically developing and late talking
children.
2) To study the effect of social mediation strategies in the form of instruction and
modeling on symbolic play behaviours.
3) To examine the relationship between the symbolic play development and language
development in both the receptive and expressive domains.
In addition, the age related changes in the symbolic behavior and the toy and play
preference between genders in both the groups of children were examined.
Symbolic Play and Language in Late Talkers
85
Method
Subjects
A total of twenty children between the ages of 24 to 36 months served as subjects for
the study. The clinical group consisted of ten children in two age groups (24-30 months and
30-36 months) diagnosed as Expressive Language Delay (late talkers). The control group
consisted of ten children matched for gender, age range, socio economic status and child care
history. The criterion for inclusion of children in both the groups was on the basis of their
receptive and expressive language age revealed through the results of Three-Dimensional
Language Acquisition Test (3D-LAT, Geetha Harlekhar, 1986). The receptive and expressive
language age of the children in the control group was within 3 months of their chronological
age. On the other hand, the children in the clinical group had a receptive age which was well
within 3 months of their chronological age while their expressive age was 6 months or more
below their chronological age i.e. the gap between the chronological age and the expressive
age was greater than 6 months. (expressive language delay). The children included in both the
groups had no history of medical problems, emotional, behavioral, cognitive or sensory
disturbances. In addition the WHO Ten-question disability screening checklist (Singhi,
Kumar, Malhi, & Kumar, 2007) was used to rule out any disability for the children in the
control group. Ethical procedures were used to select the participants. The parents were
explained the purpose and the procedures of the study and an informed verbal and /or written
consent were taken.
Procedure
The procedure consisted of two phases; Phase I included the investigation of symbolic
play behaviours and the phase II included the investigation of the relationship between play
and language.
Phase I
Investigation of symbolic play behaviours
To study the symbolic play behavior, two sessions of play were organized in which all
the children participated in two types of play situations namely structured play and free play.
In the free play format a specific set of toys and/or objects was provided to play with in a
natural setting i.e., within their homes. In the first session of the structured play format
thematically related toy sets were given to each child to elicit symbolic play behaviours and
in the second session the same was elicited through social mediation strategies namely
instruction and/or modeling.
First session
a) Structured play: Each child was presented with four sets of thematically related toys, one
set at a time, and they were allowed to interact with them for approximately 5 minutes each.
The sets included several standard toys which would facilitate symbolic play and either a
stick or a block as an item to be transformed.
Set 1: Doll, baby bottle, blanket, stick.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
86
Set 2: Stuffed bear, comb, blanket, stick.
Set 3: Two small human figures, horse, soap, block.
Set 4: Truck, human figure, toy screwdriver, two blocks, stick.
b) Free play: Each child was presented toys such as kitchen set, furniture set, doll, blanket,
truck, two small human figures, a comb, small plastic animals etc. which were spread in the
vicinity of the child. The child was invited to play with the toys. The mother was seated in the
room but was asked not to intervene in the child‟s play. This session lasted for approximately
10 minutes. The experimenter did not engage in the ongoing activities, but only redirected the
child‟s focus to the toys if his/her attention wandered before the end of 10 minute period.
Second session: The second session was taken a week after the first session.
a) Free play: The child was presented with the same toys as mentioned above. The same
procedure as in first session was carried out. The play behaviours were assessed for a second
time to examine test to retest reliability of play behaviors.
b) Structured play: Instruction and modeling conditions:
The child was presented with the same four sets of toys as listed in first session. The
codable pretend behaviors at different levels were elicited by the experimenter by giving the
children specific instructions. If these behaviors did not occur on instructions alone, then the
toys were presented once again and the remaining desired actions were demonstrated
accompanied by verbal instructions. For e.g., on presenting the child with a doll and a bottle,
the experimenter said, “Can you give the dolly a bottle?” If the child did not perform the
action requested, a second instruction was given in the form of a command, “Give her some
juice”. In case the child did not perform the behavior with instruction, the experimenter
elicited them by demonstration. For e.g., the experimenter picked up the necessary toys,
pretended to feed the doll with the bottle and handed over both items to the child with the
instruction that, “Can you give the dolly a bottle like that? Now do it”.
All these sessions mentioned above were videotaped. These toys have been selected
on the basis of literature support (Rescorla & Goosens, 1992) with suitable modifications for
Indian context.
Phase II: Investigation of the relationship between play and language
In addition to the free and structured play session, the Assessment Checklist for Play
Skills (Swapna, Jayaram, Prema, & Geetha, 2006) was administered to get their age-
equivalent play scores. This is a checklist standardized on Kannada speaking Indian children
to assess the overall development of play between birth to three years of age. The play
behaviors observed during the free and structured play were used to rate the child‟s play skill
on the assessment checklist. This was done in addition to the information obtained about their
play behavior though parental interviews.
Symbolic Play and Language in Late Talkers
87
Analysis
The symbolic play behaviors such as two basic categories of presymbolic play
(functional play including functional conventional, functional to self and functional to other
and sequential play sub classified into four types A, B, C & D), two types of symbolic play
(symbolism sub classified into three types - A, B & C and verbal symbolism) as well as a
variety of other nonsymbolic behaviors such as grouping, manipulation, wandering and social
interaction were studied.
Data coding for free play: Various types of play behaviors exhibited during free play during
both the sessions were coded from the videotape for frequency of the specified play
behaviors. For e.g., if the child puts a cup on the saucer, it was coded under „functional
conventional‟. Thus each time a play behavior occurred (details provided in the appendix), it
was marked in the response sheet and documented descriptively if required. The qualitative
differences in frequency of symbolic play in the two different age groups were also analyzed.
Data coding for structured play: The following scoring pattern was used. The spontaneous
occurrence of a desired behavior (details provided in the appendix) in the first session was
scored as „3‟, the occurrence of a desired behavior in response to instruction in the second
session was scored as „2‟ and the occurrence of a desired behavior following modeling in the
second session was scored as „1‟. The non occurrence of a desired behavior in the first or
second session was scored as „0‟. The maximum score that can be obtained by a child is „5‟
for each target behaviour with toy sets (3 for spontaneous performance in the first session
plus „2‟ for performance with instruction in the second session).
The raw data from phase I and Phase II was tabulated and further subjected to
appropriate statistical analysis.
Results and Discussion
Two qualified speech-language pathologists scored the play behaviours seen during
free and structured play. Their ratings were compared to check for the inter-rater agreement.
It was found that there was a high reliability between the two raters i.e., Cronbach‟s alpha
value being (> 0.7 to 1) for both free and structured play. The results obtained on the play
behaviours in typically developing children and late talkers from different statistical analyses
have been presented and discussed under phase I and II:
Phase I: Investigation of symbolic play behaviours
a) Free play analysis:
i) Comparison between control and clinical group: The total frequency of each type of
play was aggregated across the two sessions and the mean, standard deviation and t-values
were calculated which has been depicted in table 1.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
88
Table 1: Mean and standard deviation along with t-values for free play behaviours across
both baskets for typically developing children and late talkers
Play behaviour Typically
developing group
Late talking
group t-values
( 18 ) Mean SD Mean SD
Non
symbolic
play
Grouping 5.00 3.16 4.50 4.06 0..31
Manipulation 8.60 4.59 15.60 8.34 *2.32
Social 2.30 1.56 1.10 1.66 1.66
Unoccupied 2.00 1.76 0.40 0.69 *2.66
Pre-
symbolic
play
Functional
conventional 10.30 6.53 5.40 3.37 *2.10
Functional to
self 3.90 4.35 2.90 2.92 0.60
Functional to
other 8.50 6.11 5.80 5.18 1.06
Sequence A 0.90 1.37 0.20 0.42 1.54
Sequence B 2.70 3.71 0.30 0.67 2.01
Sequence C 1.10 1.91 0 0 1.81
Sequence D 0.40 1.26 0 0 1.00
Symbolic
play
Symbolism A 3.10 2.46 1.40 1.64 1.81
Symbolism B 0.70 .823 .30 .483 1.32
Symbolism C 0 0 0 0 -
Verbal
symbolism 1.10 1.85 0 0 0.30
„-‟ indicates both are equal. [*p<0.05]
It can be seen from the table that the mean frequency of different types of play
behaviour observed in late talkers was relatively lesser than that observed in the typically
developing group. However the late talkers exhibited higher frequency of manipulation
behaviour i.e. (mouthing, squeezing etc.) when compared to the typically developing
children. It is a known fact that proper development of symbolic play requires the perceptual
and memory process and the ability to sustain and regulate attention (Tamis-Le Monda and
Bornstein, 1996). In object based symbolic play a child has to divide his or her attention
between toys, transformation, scheme sequencing and the signifier and signified relationship
(Casby, 1997). Thus it would be possible that the late talkers had mild deficits in the higher
level functions of the brain which could have resulted in their poor performance in symbolic
play.
The results of the independent t-test indicated that there was a significant difference
for manipulation (t(18)=2.32), unoccupied (t(18)=2.66) and functional conventional play
behaviours (t(18)=2.10) between the groups at 0.05 level. However, the other play behaviours
failed to show any significant difference. This can be attributed to the scatter in the age of the
subjects considered.
A closer examination of the non symbolic play across all the subjects revealed that the
mean scores were less for unoccupied behaviors. Further, there was a difference in the
unoccupied behaviours seen in both the groups. The unoccupied behaviours in the control
Symbolic Play and Language in Late Talkers
89
group consisted of children moving around the room, checking on with their mothers and
later coming back to the toys for playing. The late talkers, on the other hand, tended to have
fewer occurrences of off-play behaviour, but these appeared to be interludes during which the
child would sit but not be engaged with the toys. The results obtained in the current study for
the manipulation and unoccupied behaviours are in consonance with the results reported by
Rescorla and Goosens (1992). However, the results obtained for grouping and social
behaviours are not in agreement with their study.
The examination of the presymbolic play behaviours revealed maximum number of
functional conventional play behaviors in both the groups of children. This was followed by
functional to other type of play The least score was obtained for functional to self type of
play behavior. But these results appear to diverge from the findings of Watson & Fischer,
(1977); Bretherton, (1984); McCune-Nicolich (1977) and Patterson & Westby (1998) who
reported that the functional to self play behavior occurs before the functional to other kind of
play behavior in the developmental hierarchy of play. This could be because of the fact that
when the child sees many objects, he/she tries to relate these objects to the objects seen in
real life and depict its use in some manner or the other rather than doing any action on self.
Thus the wide variety of toys they are exposed to at a time could have restricted them from
performing the action on self. The functional sequential play behaviours occurred the least.
These results are consistent with Rescorla and Goosens (1992).
With respect to the symbolic play, children in both groups displayed greater number
of symbolism in their play (play behaviors in sequential order and using blocks and sticks as
a substitute for some other object) when compared to the verbal symbolism which involved
verbally creating any action, absent person or verbally substituting one object for another.
These results are also in agreement with Rescorla and Goosens (1992).
In totality, it was found that both the groups of children displayed greater number of
presymbolic play behaviours followed by nonsymbolic and relatively lesser number of
symbolic play behaviours which revealed that the play behaviours are slowly progressing into
the more advanced form. In addition, it was observed that the quality and type of play of the
late talkers was different compared to that displayed by the control group. The late talkers
involved less number of toys in free play, i.e., they continued to play with one or two toys
throughout the session and repeated the same behaviours over time but were able to
appropriately use the toys whereas, the typically developing group exhibited a wide variety of
play behaviours incorporated more number of toys. Thus the late talkers appeared to be less
responsive to the social, thematic, or representational qualities of the toys than the typically
developing children. This has also been reported by Rescorla and Goosens (1992) in their
study.
ii) Comparison of performance between groups within age group: The mean of the free
play behaviours obtained for both the groups of children was analyzed age group wise using
Mann Whitney U test. It was found that there was no statistically significant difference
among the groups within both the younger and the older age group. This could have occurred
because of the smaller sample size considered for the study.
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
90
iii) Comparison of performance in both groups of children across age groups: The mean
of the free play behaviours in two different age groups considered namely 2- 2.6 years and
2.6-3 years, was subjected to independent t-test to examine if any statistically significant
differences existed between the two age groups. The results indicated that there was no
statistically significant difference across both the age groups in both the control and the
clinical group. However, a developmental trend was observed in the typically developing
group in which the younger group showed more number of nonsymbolic play behaviours
(grouping and manipulation) compared to the older group who exhibited more of presymbolic
(except for functional to self) and symbolic behaviours. These findings can be supported with
the studies of Mc Cune-Nicolich, (1977); Patterson and Westby, (1998); Lyytinen, Laakso,
Poikkeus, and Rita (1999): Katz, (2001); and Casby, (2003) who reported that the play
typically follows a developmental progression in a sequential pattern: simple to complex, self
to others, concrete to abstract. When new types of play develop, „older‟ types of play do not
disappear, although they decrease in frequency.
However, such a developmental trend was not observed in the late talking group
across their age. Some of the play behaviours are scored higher by the younger group
whereas some are scored higher by the older group. This lack of a developmental trend which
was seen in the late talking group indicated that there was a variable performance among this
group and they had delayed/deviant play patterns.
b) Structured play analysis
i) Comparison between both the groups w.r.t different toy sets: The group difference in
play behaviour with respect to different toy sets were analyzed using paired t-test. The
weighted play score, mean and SD values obtained along with the F values obtained for both
the groups are depicted in table 2.
Table 2: Maximum weighted scores, Mean, SD and t-values during structured play with toy
sets, for typically developing and late talking group.
Toy
sets
Maximum
weighted
score
Typically
developing group
Late talking group t-value
(18)
Mean SD Mean SD
Set 1 60 25.00 7.55 17.20 4.91 *2.73
Set 2 60 28.00 6.74 22.10 7.75 1.81
Set 3 55 25.70 10.18 16.80 6.46 *2.33
Set 4 40 22.00 4.54 17.30 4.13 *2.41
[*p<0.05]
It is seen from the table 2 that the typically developing group performed better when
compared to the late talking group with respect to all the four sets and this difference was
statistically significant for three sets i.e., set 1, (t(18)=2.73, p<0.05), set 3 (t(18)= 2.33,
p<0.05) and set 4 (t(18)=2.41, p<0.05). However with respect to set 2 the difference was not
significant. This could be attributed to the familiarity factor of the items included in set 2
Symbolic Play and Language in Late Talkers
91
(comb, brush, quilt, bear). These objects are commonly found in every household and are
used on a daily basis. This resulted in late talking group performing better on this set when
compared to the other sets. This finding is not in consonance with the study done by Rescorla
and Goosens, (1992). They found a significant difference in performance between the two
groups only on sets 2, 3, & 4.
ii) Performance of the groups under social mediation strategies such as modeling and
instruction (Comparison across session I and II of structured play): The performance of
the groups between the structured play session I- spontaneous occurrence and II- elicitation
using social mediation strategies were compared using independent t-test. The maximum play
behaviour in each toy set, mean, SD and the results of the test set wise has been represented
in the table 3 below.
Table 3: Mean, SD and t-values number of targeted responses in structured play (session I)
vs. structured play (session II) by groups.
Toy
sets
Structured
play session
Max. play
behaviour
Typically developing group Late talking group
Mean
scores
SD t-value
(9)
Mean
scores
SD t-value
(9)
Set 1
Session I 12 3.30 1.40
*9.78
2.30 1.15
*10.47 Session II 12 9.10 2.20 8.10 1.59
Set 2
Session I 12 3.30 1.63
*8.61
3.80 2.20
*5.82 Session II 12 9.60 2.31 8.90 1.59
Set 3
Session I 11 3.40 1.57
*9.26
2.40 1.50
*11.00 Session II 11 9.10 1.91 7.90 2.28
Set 4
Session I 8 2.40 1.73
*14.45
2.70 1.05
*8.66 Session II 8 7.70 0.67 7.50 1.08
[*p<0.01]
The data depicted in the table 3 revealed there was a significant difference between
structured play (session I and II) in both the groups. This difference was significant for all the
four sets of toys used (p<0.01). Therefore it can be inferred that the frequency of the targeted
responses increased with social mediation (structured play session II) provided by the
experimenter. Although the late talkers did know and understand the conventional use of the
objects, they did not initiate to come up with related symbolic activities but when they were
prompted to do so by verbal or gestural instructions and modeling they were readily able to
carry out activities meaningfully.
Support can be drawn from the studies done by Watson and Fischer, (1977) and Largo
and Howard, (1979) who reported that the social mediation brings about a positive change in
the play behaviours and it helps to trigger matured forms of play. These results are also in
consonance with the study by Rescorla and Goosens (1992).
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
92
iii) Comparison between both groups using social mediation strategies: Independent t-test
was carried out to find if any significant difference existed between the late talkers and
typically developing group in each of the social mediation strategy namely instruction and
modeling. The mean, SD and the t-values are depicted in the table 4 given below.
Table 4: Mean, SD and t-values across both the groups under conditions of instruction and
modeling during structured play (session II)
Toy
sets
Max. Instruction t-
value
(18)
Modeling t-
value
(18) Typically
developing
group
Late talking
group
Typically
developing
group
Late talking
group
Mean SD Mean SD Mean SD Mean SD
Set 1 12 4.50 2.22 2.30 1.25 *2.72 2.50 1.43 4.90 1.28 *3.94
Set2 12 7.00 2.35 2.60 1.57 *4.90 1.00 0.94 5.70 1.33 *9.08
Set3 12 6.10 2.55 3.00 2.16 *2.92 1.40 0.69 4.10 2.13 *3.80
Set4 12 5.50 1.58 1.90 1.19 *5.74 0.90 1.10 4.60 1.64 *5.9
[*p <0.01]
The results revealed that both the groups performed better with instructions and
modeling. It was also seen that the typically developing group exhibited more number of play
behaviours in each set of toys with instructions alone and required modeling to elicit only a
few play behaviours, the late talking group exhibited lesser number of play behaviours with
instructions alone and required modeling to elicit the majority of play behaviours. There was
a significant difference between the performance of both the groups with respect to
instruction and modeling on all the four sets at p<0.01 level.
During the data coding a general difference in toy and play preference across gender
was also observed. It was seen that certain sets of toys elicit different symbolic play
behaviours in male and female children. Toys such as doll, bottle, quilt and a stick, bear,
brush, and comb elicited more symbolic play behaviours in girls whereas toys such as truck,
tools, little men, stick and blocks elicited more symbolic play behaviours in boys.
Phase II: Investigation of the relationship between play and language
The Spearman‟s rank correlation coefficient was applied to examine the correlation
between play age and the receptive and expressive language between the age group of 2-2.6
yrs and 2.6-3 years in both the groups of children the results of which have been depictd in
table 5.
Symbolic Play and Language in Late Talkers
93
Table 5: Correlation values between play age, receptive and expressive age across the age
groups in typically developing group and late talkers.
3D-LAT
Typically developing group Late talking group
2-2.6 yrs 2.6 – 3yrs 2-2.6 yrs 2.6 – 3yrs
Play assessment
checklist
RLA ELA RLA ELA RLA ELA RLA ELA
PLA a a 0.60 0.54 0.83 1.00** 0.73 1.00**
[**p< 0.01], RLA- Receptive Language Age, ELA - Expressive Language Age,
PLA - Play Age. a- represents constant value and therefore could not be computed.
The data in table 5 indicates that the play age remained constant in the age group of 2
– 2.6 years in the typically developing children, therefore the correlation among the play age
and other domains could not be computed. However, in the higher age group, that is, 2.6-3
years, the data revealed that there was a correlation between play age and receptive and
expressive language age, but this was not statistically significant. A closer look at the data
revealed that the correlation was high between receptive language age and play age (r=.60)
when compared to the correlation between expressive language age and play age (r=0.54).
In the late talking group, play age correlated with receptive and expressive language
age in both the age groups. The correlation between play age and expressive language age
was significant (p<0.01) in both the age groups considered, while the correlation between
play and receptive language age in the lower and higher age group was not significant. It was
also seen that the correlation between play and the receptive language age was higher in the
younger age group (r=0.83) than the older age group (r= 0.73). It can be inferred that across
the early age of 2-3 years the play age correlated with expressive rather than receptive
language in late talkers. However in late talkers it was observed that play age correlated more
with receptive language in both the age groups (r=0.83, 0.73) than in the typically developing
group (r=0.60). This could be attributed to the fact that that there was not much variation in
the play age and the receptive language age across subjects.
Thus it can be inferred that play is associated with language and that play parallels
expressive language in late talkers which in turn reflects slower maturation of a complex
developmental system of symbol use. Hence it could be argued that play and language are
mediated by a general developmental factor. The results of the present study are in agreement
with several studies carried out on typically developing children (McCune-Nicolich, 1981;
Baron-Cohen, 1987; Casby & Della Corte, 1987; Beeghly et al., 1990; Ogura, 1991; Doswell
et al., 1994; McCune, 1995; Lyytenin & Laakso, 1997; Lifter & Bloom, 1998; Tomasello et
al., 1999; Kitty, 2000; Lewis et al., 2000).
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
94
Conclusions
To conclude the late talkers in general have poorer symbolic play abilities compared
to the typically developing children. They exhibited lesser presymbolic and symbolic play
behaviours compared to the typically developing group. While a developmental trend was
observed in the typically developing group in which the younger group showed more number
of nonsymbolic play behaviours (grouping and manipulation) compared to the older group
who exhibited more of presymbolic (except for functional to self) and symbolic behaviours,
such a trend was not seen in the late talkers. Moreover, the social mediation (instructions and
modeling) had a definite positive effect on the symbolic play performance in both the groups.
The typically developing group exhibited more number of play behaviours with instructions
alone, however, the late talking group required modeling along with instructions to elicit the
majority of play behaviours. With regard to the play-language relationship, it was found that
in both the groups play age correlated with receptive and expressive language. In the
typically developing children, play age correlated more with receptive language, however in
the late talking group play age correlated more with expressive language age. This indicated
that the lower expressive language ability in late talkers was associated with poorly
developed symbolic play skills.
Cautions must be taken while drawing inferences from this study given the small
numbers of participants and reliance on correlation analysis, which does not clarify causal
relations. Nevertheless, the study has important implication for early childhood assessment
and intervention. This study suggests that it is crucial to examine several domains of
functioning within the same child, only then can the relationships between various skills be
revealed. Furthermore, focusing on skills common to symbolic play, cognition and language
during intervention should also be an important aspect of intervention as they have been
found to impinge on language capacities as the child matures. The findings of such research
might contribute to theories of language development as well as assist clinicians in designing
accurate screening procedures.
References
Bates, E. (1976). Language and context: The acquisition of pragmatics. New York: Academic
Press.
Baron-Cohen, S. (1987). Autism and symbolic play. British Journal of Developmental
Psychology, 5, 139-148.
Beeghly, M, Weiss-Perry, B. & Cicchetti, D. (1990). Beyond sensorimotor functioning: early
communicative and play development of children with Down syndrome. In D.
Cicchetti and M. Beeghly (Eds.). Children with Down syndrome: A Developmental
Perspective.
Bretherton, I. (1984). Representing the social world in symbolic play: Reality and fantasy. In
I. Bretherton (Ed), Symbolic play- The development of social understanding. Orlando,
FL: Academic Press.
Casby, M.W. (1997). Symbolic play of children with language impairment: A critical review.
Journal of Speech, Language, and Hearing Research, 40, 468-479.
Symbolic Play and Language in Late Talkers
95
Casby, M. W. (2003). Developmental Assessment of Play. A Model for Early Intervention.
Communication Disorders Quarterly, Vol. 24, (4), 175-183.
Casby, M.W., & Della Corte, M. (1987). Symbolic play performance and early language
development. Journal of Psycholinguistic Research, 16, 31-42.
Doswell, G., Lewis, V., Boucher, J. & Sylva, K., (1994). Validational data on the Warwick
symbolic play test. European Journal of Disorders of Communication, 29, 291-300.
Feuerstein, R. (1980). Instrumental enrichment: an intervention program for cognitive
modifiability. Baltimore: University Park Press.
Geetha Harlekhar. (1986). Three Dimensional Language Acquisition Test (3D-LAT).
Unpublished dissertation (Speech and Hearing). University of Mysore.
Katz, J.R. (2001). Playing at home the talk of pretend. In D.K. Dickson and P.O. Tabors
(Eds.), Beginning literacy with language. Baltimore: Paul H. Brookes, 53-73.
Kitty C.H. (2000). „Correlation between symbolic play and language in normal developing
Cantonese speaking children.’ Dissertation submitted as a part of partial fulfillment
for the Bachelor of Science, speech and hearing sciences. University of Hong Kong.
Largo, R. H., & Howard, J. A. (1979). Developmental progression in play behavior of
children between 9 and 30 months. Developmental Medicine and Child Neurology,
21, 299-310.
Lewis, V., Boucher, J., Lupton, L., & Watson, S. (2000). Relationships between symbolic
play, functional play, verbal and non-verbal ability in young children. International
Journal of Language & Communication Disorders, 35(1), 117-127.
Lifter, K., & Bloom, L. (1998). Intention and the role of play in the transition to language. In
A. Wetherby, S. Warren, & J. Reichle (Eds.). Transitions in prelinguistic
communication (Vol. 7, pp. 161–196). Baltimore: Brookes.
Lovell, K., Hoyle, H. W., & Siddal, N. Q. (1968). A study of some aspects of the play and
language of young children with delayed speech. Journal of Child Psychology and
Psychiatry, 9, 41-50.
Lyytinen, P., & Laakso, M. (1997). Language and symbolic play in toddlers. Journal of
Behavioural Development, 21(2), 289-302.
Lyytinen, P., Laakso, M. L., Poikkeus, A. M., & Rita, N. (1999). The development and
predictive relations of play and language across the second year. Scandinavian
Journal of Psychology, 40, 177-186.
McCune, L. (1995). A normative study of representational play at the transition to language.
Developmental Psychology, 31, 198-206.
McCune-Nicolich, L. (1977). Beyond sensorimotor intelligence: Assessment of symbolic
maturity through analysis of pretends play. Merrill Palmer Quarterly, 243(2), 89-101.
McCune-Nicolich, L. & Fenson, L. (1984). Methodological issues in studying early pretend
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
96
play. Cited in T.D. Yawkey & A.D. Pellegrini, (Eds.), Child‟s play: Developmental
and applied (pp. 81-103). Hillsdale, N.J; Lawrence Erlbaum.
Ogura, T. (1991). A longitudinal study of the relationship between early language
development and play development. Journal of Child Language, 18, 273-294.
Patterson, J., & Westby, C. (1998). The development of play. In B. Shulman and W. O.
Haynes (Eds.), Communication Development (2nd ed.; pp. 135-164). Englewood
Cliffs, NJ: PrenticeHall.
Piaget, J. (1962). Play, dreams, and imitation in childhood. New York: Norton and Company.
Rescorla, L., & Goossens, M. (1992). Symbolic play development in toddlers with expressive
specific language impairment. Journal of Speech and Hearing Research, 35, 1290-
1302.
Roth, F.P., & Clark, D.M. (1987). Symbolic play and social participation abilities of language
impaired and normally developing children. Journal of Speech and Hearing
Disorders, 52, 17-29.
Singhi, P., Kumar, M., Malhi, P., & Kumar, R. (2007). Utility of the WHO Ten Questions
Screen for Disability Detection in Rural Community-the North Indian Experience.
Journal of Tropical Pediatrics, 53, 6, 383-387.
Swapna, N., Jayaram, M., Prema, K.S., & Geetha, Y.V. (2006). An ARF project
‘Development of intervention module in preschool children with communication
disorder’. Unpublished Project Report, AIISH, University of Mysore, Mysore.
Tamis-LeMonda, C. S., & Bornstein, M. H. (1996). Variation in children‟s exploratory,
nonsymbolic, and symbolic play: An exploratory multidimensional framework. In R.
Rovee-Collier & L. Lipsitt (Eds.). Advances in Infancy Research, 10, 37-78.
Terrell, B., & Schwartz, R. (1988). Object transformations in the play of language-impaired
children. Journal of Speech and Hearing Disorders, 53, 459-466.
Terrell, B.Y., Schwartz, R.G., Prelock, P.A., & Messick, C.K. (1984). Symbolic play in
normal and language impaired children. Journal of Speech and Hearing Research, 27,
424-329.
Tomasello, M., Striano, T., & Rochat, P. (1999). Do young children use objects as symbols?
British Journal of Developmental Psychology, 17 (4), 563-584.
Udwin, O., & Yule, W. (1982). Validational data on Lowe and Costello‟s Symbolic Play
Test. Child Care, Health and Development, 8, 361-366.
Venkatesan, S. (2004). Children with developmental disabilities: A training guide for
parents, teachers and caregivers. India: Sage Publication.
Watson, M.W., & Fischer, K.W. (1977). A developmental sequence of agent use in late
infancy. Child Development, 48, 828-836.
Westby, C. E. (1980). Assessment of cognition and language abilities through play.
Language, Speech, and Hearing Services in Schools, 11, 154-168.
Symbolic Play and Language in Late Talkers
97
Appendix
I List of play behaviors observed during free play
Nonsymbolic play
Wandering/unoccupied – Behaviours not involving any active interaction with objects or
individuals in the room.
Manipulation/handling- Involves child‟s visual and kinesthetic exploration and
manipulation of toys.
Grouping: Placing two or more like objects together in a group, line or stack.
Presymbolic play
Functional conventional (Presymbolic scheme): Behaviours indicating that child knows the
functionally appropriate use of an object
Functional to self (Auto symbolic scheme): Functionally appropriate use of an object on
oneself.
Functional to other (De-centered symbolic games): Involves performance of a pretend
action upon a recipient other than self.
Sequence Type A: Two or more consecutive but different actions, one or more which is
functional conventional.
Sequence Type B: Same recipients (two or more consecutive and) but different actions.
Sequence type C: Two or more recipients/ (two or more consecutive and) same actions.
Sequence Type D: Two or more recipients/ (two or more consecutive but) different actions.
Symbolic play
Symbolism Type A: Substitution of one object for another. Using objects in a manner
different from its intended functional use.
Symbolism type B: Pretending to use an absent object, creating an absent person, or
referring to an absent substance.
Symbolism Type C: Animating the doll or animal as an independent and active agent.
Verbal symbolism:
Verbal Transformation: Verbal substitution of one object for another
Verbal creation of object: Verbally creating an absent person or object by referring to it.
Verbal animation: Verbally creating action, animating an object or toy with no
accompanying action.
Social interaction
Child initiated social games: Play behaviour involving an adult in the room without
functional or symbolic use of an object.
Child-initiated social interaction: Active or verbal non- play behaviour initiated by the child
and directed by an adult
II Targeted behavior for structured play:
I Set 1(doll, baby bottle, blanket, stick):
Functional conventional – Spreading the Blanket
Functional to self 1- Drinking from bottle
Functional to self 2- covering self with blanket
Functional to other 1- Feeding the doll with baby bottle
Functional to other 2- Covering doll with blanket
Sequence A- Spread the blanket and put the doll in the blanket
Sequence B- Feeding doll and then putting it to sleep
Dissertation Vol.VII, 2008-09, Part – B, SLP, AIISH, Mysore
98
Sequence C-Feeding doll, make it drink water and then put to sleep
Sequence D- Feeding doll, make it drink water, cover self with blanket and then both go to sleep
Symbolism A- Feeding doll with stick as bottle
Symbolism B- Feeding doll with pretend bottle
Symbolism C- Having doll to spread the blanket
II Set 2 (stuffed bear, comb, blanket, stick)
Functional conventional – Spreading the blanket
Functional to self 1- Combing own hair
Functional to self 2- Covering self with blanket
Functional to other 1- Combing bear with comb
Functional to other 2- Covering bear with blanket
Sequence A- Spread the blanket and put animal to sleep
Sequence B- Combing bear‟s hair and putting it to sleep
Sequence C- Combing bear‟s hair and own hair
Sequence D- Combing bear‟s hair, covering self with blanket, and both go to sleep
Symbolism A - Combing bear with stick as brush
Symbolism B- Combing bear with pretend brush
Symbolism C- Making bear to spread blanket
III Set 3 (two small human figures, horse, soap, block)
Functional conventional- Moving the truck
Functional to self 1- Soaping own body
Functional to self 2- Self riding on the horse
Functional to other - Washing horse with soap
Functional to other 2- Giving the little man a ride on horse
Sequence B- Washing horse with soap and giving the little man a ride
Sequence C- Soaping horse and then own body
Sequence D- Making the little man pat horse and giving the little lady a ride
Symbolism A- Washing horse with block as soap
Symbolism B- Washing horse with pretend soap
Symbolism C- Making the little man wash the horse
IV Set 4 (truck, human figure, toy screwdriver, two blocks, stick)
Functional conventional – Loading truck with block
Functional to other 1- Fixing truck with toy screwdriver
Functional to other 2- Giving the little man a ride on truck
Sequence A- Loading truck with blocks and making the little man ride on truck
Sequence B- Fixing truck with screwdriver and giving the little man a ride in truck
Symbolism A- Fixing truck with stick as a tool
Symbolism B- Fixing truck with pretend tool
Symbolism C- Making the little man drive the truck
Subtyping of Children with Developmental Dyslexia
99
Subtyping of Children with Developmental Dyslexia: Implications through
Dual Route Cascaded (DRC) Model in the Indian Context
Gnanavel. K & Jayashree. C. Shanbal*
Abstract
Reading is a complex cognitive process. It involves the co-ordination of a series of functions
which include visual functions such as configurational (feature) and orthographic (word form)
analyses and verbal or language functions such as phonological, semantic and syntactic coding and
decoding and other cognitive functions like memory and attention. The complexity of reading task is
clearly illustrated in recent computational models of reading (Coltheart, Rastle, Perry, Langdon, &
Ziegler, 2001; Perry, Ziegler, & Zorzi, 2007). Dual Route Cascaded Model (DRC) by Coltheart, Rastle,
Perry, Langdon and Ziegler (2001) is considered to be the most successful one in explaining visual
word recognition and reading aloud. The purpose of the present study was to explore the subtypes of
children with developmental dyslexia based on the dual route cascaded model and to assess each
representational level of the DRC model (Coltheart et.al 2001). Two groups i.e., age matched normal
children (40) and children with dyslexia (16) from grades III to VI were considered for the study.
Subjects were tested on tasks for each representational level of the model. The accuracy and reaction
time of their responses were measured and recorded using DMDX software. Overall, children with
dyslexia performed significantly poorer in all the tasks than normal children. .The results revealed that
on Subtyping in children with dyslexia there was no single isolation deficits observed and multiple
deficits were observed based on DRC model. However, cluster analysis and qualitative analysis of the
data could group the clinical data into two subtypes- phonological and mixed type of dyslexia. Subtyping
has further implication in selecting the appropriate strategy for these children in the management
program.
Introduction
Reading is a complex cognitive process. It involves the co-ordination of a series of
functions which include visual functions such as configurational (feature) and orthographic
(word form) analyses and verbal or language functions such as phonological, semantic and
syntactic coding and decoding and other cognitive functions like memory and attention and
motor skills. The complexity of reading task is clearly illustrated in recent computational
models of reading (Coltheart, Rastle, Perry, Langdon, & Ziegler, 2001; Perry, Ziegler, & Zorzi,
2007). Dual Route Cascaded Model (DRC) by Coltheart, Rastle, Perry, Langdon and Ziegler
(2001) is considered the most successful one in explaining visual word recognition and
reading aloud. As the name suggests, the DRC model has two core assumptions. First,
processing throughout the model is cascaded. That is, any activation in earlier modules starts
flowing to later modules immediately. Second, there are two routes for translating print into
sound: a lexical route, which utilizes word-specific knowledge, and a non-lexical Grapheme-
to-Phoneme Conversion (GPC) route, which utilizes a sub-lexical spelling-sound
correspondence rule system.
________________________________
* Lecturer in Language Pathology, All India Institute of Speech and Hearing, Mysore, India
email: [email protected]
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
100
Figure1: Architecture of DRC model by Coltheart et al. (2001)
The lexical route translates the pronunciation of a word based on word specific
knowledge. The route consists of components: the orthographic lexicon, semantic and the
phonological lexicon, as seen in the left part of Figure1.The semantic system computes the
meaning of a word, whereas the lexicons compute the words‟ orthographic and phonological
form. Representations of a word in the orthographic lexicon and the phonological lexicon are
linked so that activation in one leads to activation of the other. For instance, the letters “c,”
“a” and “t” will activate the orthographic representation of “cat,” which will then activate its
phonological representation of /kæt/. Frequency scaling is also applied to each orthographic
and phonological lexicon. Thus, a high frequency word such as “the” will be named faster
than a low frequency word such as “quench.”
The non-lexical route differs from the lexical route in both the knowledge base and
the type of processing it employs. The non-lexical route generates the pronunciation of letter
string (be it a word or a non-word) via a set of sub-lexical spelling-sound correspondence
rules. The set of rules is encapsulated in the GPC module. One important feature of the GPC
module is that its processing is serial. The GPC module applies rules serially left to right to a
letter. That is, letters activate phonemes in a serial, left to right fashion. Activation of the
second phoneme does not start until a constant number of cycles after the start of activation
of the first letter. For example, given a non-word like „bant‟, the corresponding translation
would be: B -> /b/, A -> /æ/, N ->/n/, and T -> /t/. Coltheart et al. (2001) argue that the non-
word letter length effect produced by DRC model is a direct consequence of serial processing
in the GPC module. That is, because GPC processes letters serially, the time to name a non-
word increases as the length of non-word increases.
According to the DRC model, successful reading depends on the interaction of
sublexical and lexical procedures. Only when both of these procedures are functioning
adequately, is then an individual able to read all forms of text. The sublexical procedure
decodes novel letter strings via grapheme/ phoneme correspondence rules that exist in
alphabetic writing systems. A break within the lexical or the sublexical routes can affect
reading ability in a few children often referred to as children with dyslexia. Although several
different subtypes of acquired dyslexia have been proposed to equate to developmental
dyslexia counterparts (Rayner, Murphy, Henderson & Pollatsek, 1989), the most influential
corresponds to the auditory versus visual dichotomy. Castles and Coltheart (1993) proposed
Subtyping of Children with Developmental Dyslexia
101
that these phonological and surface subtypes of acquired dyslexia also existed within the
developmental dyslexic population. A break within the sublexical route results in the subtype
of acquired dyslexia referred to as phonological dyslexia. Individuals who have acquired
phonological dyslexia experience difficulties decoding unfamiliar words since the only way
to read a novel letter string that is not represented in sight vocabulary is to implement some
process of decoding (Funnell, 1983). The symptom most often associated with phonological
dyslexia is, therefore, a difficulty with the reading of non-words.
A break within the lexical procedure results in a subtype of acquired dyslexia referred
to as surface dyslexia (Behrmann & Bub, 1992). The defining characteristic of surface
dyslexia is, therefore, a difficulty with reading irregular words. The aim of the present study
is to investigating the subtypes of children with developmental dyslexia based on the dual
route cascaded model and to assess each of representational level of the DRC model.
Method
Participants
Two groups of subjects were considered for the present study, an experimental group
and a control group. The experimental group consisted of 16 dyslexic children (8-12yrs). And
the control group consisted of 40 age matched normal children and they were 10 children in
each grade from III, IV, V and VI with equal number of males and females. And none of the
children in both the groups had any known or reported hearing, neurological, developmental
or emotional problems.
Subject selection criteria
Age range of both Experimental and Control group was 8-12yrs.
Native language of both Experimental and Control group was Kannada and studied
English as the medium of instruction.
A WHO Ten –Question Disability Screening Checklist (cited in Singhi, Kumar,
Prabhjot & Kumar, 2007 ) and Developmental Screening Test (Bharath Raj, 1972),
was used to screen both the groups in terms of hearing, intelligence, motor and other
factors like school performance, emotional or behavioral factors.
Early reading skills (Loomba, 1995) was used in the selection of dyslexic children.
Computerized Linguistic Protocol for Screening (CLIPS; Anitha, 2004) was used for
screening children for language abilities.
All the children with dyslexia were assessed by a clinical psychologist for their
intelligence quotient (IQ), and were reported to be average or above average.
Test materials
Stimuli included 50 regular words and 50 irregular words taken from the class books of
grade III to VI. Pictures included 100 common line drawings taken from UNICEF picture cards
and from hundred picture naming test. For the familiarity check, from these groups of pictures and
words, 10 most familiar regular and irregular words which were matched in length were taken for
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
102
the present study. Twenty pronounceable pseudo words were also taken for the present study and
these selected words were presented through DMDX software 3.13.0 (Forster & Forster, 2003)
for about 250 ms.
Procedure
The participants were seated in a comfortable position facing the 15-inch screen of the
laptop in a quiet room.
The responses were recorded through a high quality microphone placed at a distance of 10
cm from the participant‟s mouth.
The following tasks were carried out in the present study
1. Reading task
The participants were instructed to read the stimuli aloud. The stimuli included 20 non
words, 10 regular words, and 10 irregular words .Regular and irregular words were matched in
terms of length and word familiarity. Non-words were created by changing the onset, the vowel or
the coda of an existing word that was matched in terms of frequency and length to the regular and
irregular words. The items were presented at the centre of the computer screen. DMDX
experimental software version 3.1.3.0 (Forster & Forster, 2003) was used for the experiment.
Participants‟ responses were recorded with a voice key and saved as separate wave files. Reaction
time was measured from the appearance of the stimulus on the screen until the participant
begins to utter the response and for accuracy the number of correct responses were assessed
2. Letter search task
The task was to search for a target letter embedded in a letter string. Following an initial
fixation point, a target letter (e.g., „„A‟‟) which will appear on the computer screen for 500 ms
(milliseconds) followed by the stimulus (word or unpronounceable letter string), which stays
on the screen until the participant presses one of the two response buttons to indicate whether
the target letter was present or not in the stimulus. The stimuli include 20 five-letter words and
20 five-letter nonwords (i.e., unpronounceable letter strings). Identity and position of the
target letter were matched across words and nonwords (e.g., „„R‟‟ in „„boire‟‟ versus
„„ghyrc‟‟). To avoid visual matching strategies, target letters were presented in upper case and
letter strings were presented in lower case. The accuracy and reaction time were assessed.
3. Picture naming task
Two sets of five line drawings of familiar objects were selected from UNICEF picture
cards a database for picture naming. All pictures names had a consonant-vowel-consonant
(CVC) structure. There was no phonological overlap between them. The pictures were checked
for familiarity and name agreement.
The objects were displayed in the centre of the computer screen one per trial.
The participant‟s task was to name the object as quickly as possible. The two lists
of five objects were repeated in pseudo-random order 10 times each (i.e., a total
Subtyping of Children with Developmental Dyslexia
103
of 50 naming responses per list). During training, participants were, first presented
with a sheet that contained the five objects in an unspeeded naming task. Following
training, participants were asked to perform the picture naming task twice, once using the
items of list 1 and once using the items of list 2 (counterbalanced across participants). During
the test, participants‟ responses were recorded with a voice key. Each response was saved as a
sound file. The accuracy and reaction time were assessed.
4. Phoneme matching
Participants were asked to assess the phonological similarity of spoken words either
for the initial or the final phoneme. On each trial, three spoken CVC words were presented.
Two of them shared either the initial or the final phoneme. The participants‟ task was to indicate
which item did not share the initial or the final phoneme. To facilitate the task, phoneme
position was blocked (first position block versus final position block). The order of blocks
was counterbalanced across participants. The accuracy and reaction time were assessed
Results
Broadly the results were analyzed for accuracy and reaction time (RT) measurements
for performance of normal children and children with dyslexia on all the tasks. The results are
described below in the following sections,
I. Comparison of performance of normal children and children with dyslexia
(CWD) on reading tasks
Accuracy measurements on reading tasks
Mixed ANOVA was carried out for accuracy measures across grades in normal children
and the results showed an overall significant main effect on reading tasks across grades i.e.,
F (2, 72) = 93.22, p<0.001.
Reading Tasks
RNWIWRW
Mea
n Ac
cura
cy (i
n %
)
110
100
90
80
70
60
50
Grades
III
IV
V
VI
Figure 2: Comparison of mean accuracy of reading tasks across grades in normal children.
One- way ANOVA was carried out for accuracy measures for reading tasks (RW, IW
and NW) across grades. The results revealed that there was a significant difference in reading
RW, F (3, 36) = 2.874, p< 0.05, reading NW, F (3, 36) = 3.06, p< 0.05 in normal children and
overall reading performance F (3, 36) = 3.021, p< 0.05.
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
104
Reading Task
RNWIWRW
Mea
n A
ccur
acy
( in
% )
110
100
90
80
70
60
50
40
30
20
10
0
Grades
III
IV
V
VI
Figure 3: Comparison of mean accuracy for reading tasks across grades in CWD
Mann- Whitney test revealed that this was statistically significant across groups and
across tasks (RW, IW & NW).The results showed that there was a significant difference
between CWD and normal children.
Repeated measures ANOVA was computed for within subject effects on each grade
separately for accuracy measures of reading tasks in normal children. In grade III, there was
an overall significant main effect on reading tasks i.e., F (2, 18) = 22.22, p< 0.001. In grade
IV, results showed that there was an overall significant main effect on reading tasks i.e., F (2,
18) = 17.96, p<0.001. In grade V, there was an overall significant main effect on reading
tasks i.e., F (2, 18) = 10.07, p<0.05. In grade VI, results showed there was an overall
significant main effect on reading tasks i.e., F (2, 18) = 15.7, p<0.001.
Wilcoxon singed rank test results revealed that there was a significant difference found
for reading RW and NW and for reading IW and NW at 0.05 levels. Mann- Whitney test
revealed that there was any significant difference on accuracy measures of reading tasks
within each grade for normal children and children with dyslexia
Reaction Time measurements on reading tasks
Mixed ANOVA was carried out across grades in normal children. The statistical
results showed an overall significant main effect in the reading tasks across grades i.e.,
F (2, 72) = 56.61, p<0.001. The results of one-way ANOVA for RT on reading tasks across
grades revealed that there was a significant difference in reading RW, F (3, 36) = 2.87,
p<0.05 and reading IW, F (3, 36) = 2.77, p<0.05 in normal children.
Subtyping of Children with Developmental Dyslexia
105
Reading Tasks
RNWIWRW
Mea
n R
eact
ion
Tim
e (R
T) (
in m
s)
2400
2200
2000
1800
1600
1400
1200
1000
800
600
Grades
III
IV
V
VI
Figure 4: Comparison of mean reaction time for reading tasks across grades in normal
children.
Mann- Whitney test results showed that there was a significant difference for reading
RW, IW and for the overall reading performance between CWD and normal children. But
there was no significant difference for reading NW between the groups.
Reading Tasks
RNWIWRW
Mea
n R
eact
ion
Tim
e (R
T) (
in m
s)
2400
2200
2000
1800
1600
1400
1200
1000
800
600
Grades
III
IV
V
VI
Figure 5: Comparison of mean reaction time for reading tasks across grades in CWD
Repeated measures ANOVA was computed for within subject effects on each grade
separately for reaction time measures of reading Tasks. In grade III, there was an overall
significant main effect on reading tasks i.e., F (2, 18) = 22.22, p<0.001. In grade IV,
Repeated measures ANOVA results showed there was an overall significant main effect on
reading tasks i.e., F (2, 18) = 17.96, p<0.001. In grade V, Repeated measures ANOVA results
showed there was an overall significant main effect on reading tasks i.e., F (2, 18) = 10.07,
p<0.05. In grade VI, Repeated measures ANOVA results showed there was an overall
significant main effect on reading tasks i.e., F (2, 18) = 15.7, p< 0.001.
II. Comparison of performance of normal children and children with dyslexia on
different components of the DRC model.
Accuracy measurements for DRC component tasks
Mixed ANOVA was carried out across grades in normal children and the results
showed an overall significant main effect in the DRC component tasks across grades i.e., F
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
106
(3, 108) =52.43, p<0.001. One- way ANOVA was carried find out is there any significant
difference in the performance of normal children in terms of accuracy measures of reading
tasks across grades. The results revealed that there was a significant difference in PN, F (3,
36) = 4.93, p<0.05
Tasks
PMPNLSLSNWLSW
Mea
n Ac
cura
cy (i
n %
)
110
100
90
80
70
60
50
Grades
III
IV
V
VI
Figure 6: Comparison of mean accuracy of DRC component tasks across grades in normals.
Mann- Whitney test was done for analyzing the accuracy measures across groups. The
results showed that there was a significant difference at 0.05 levels all the tasks which check
the components of the model between dyslexic and normal children.
Tasks
PMPNLSLSNWLSW
Mea
n Ac
cura
cy (
in %
)
110
100
90
80
70
60
50
40
30
20
10
Grades
III
IV
V
VI
Figure 7: Comparison of mean accuracy of DRC component tasks across grades in CWD
Repeated measures ANOVA showed that performance in terms of accuracy measures
revealed that in grade III, there was an overall significant main effect on reading tasks i.e., F
(3, 27) = 13.102, p<0.001. In grade IV, results showed there was an overall significant main
effect on reading tasks i.e., F (3, 27) = 14.766 p<0.001. In grade V, results showed there was
an overall significant main effect on reading tasks i.e., F (3, 27) = 19.439 p<0.05. In grade VI
normal children, results showed there was an overall significant main effect on reading tasks
i.e., F (3, 27) = 8.486, p< 0.001.
Wilcoxon singed rank test results revealed that there was a significant difference
found for the PM and PN and for PM and LSW, PM and LSNW. Mann- Whitney test was
done for each grade across both the groups. The results showed that in grade III, PN, LSNW
and over all letter search showed a significant difference across groups (CWD and Normal
children) at 0.05 levels. In grade VI, LSNW and over all LS showed a significant difference
across groups at 0.05 levels. In grade V all the component tasks showed a significant
difference across groups at 0.05 levels. In grade VI, PN, LSNW and PM tasks showed a
significant difference across groups at 0.05 levels.
Subtyping of Children with Developmental Dyslexia
107
Reaction Time measurements of DRC component tasks
Mixed ANOVA was carried out across grades in normal children, the results showed
an overall significant main effect in the reading tasks across grades i.e.,
F (3, 108) = 10.71, p<0.001. One- way ANOVA was carried out in normal children to find
out is there any significant difference in the RT measures of component tasks across grades.
The results revealed that there was a significant difference in letter search words, F (3, 36) =
5.1, p< 0.05, letter search nonwords words, F (3, 36) = 4.36, p<0.05 in normal children and
overall letter search performance F (3, 36) = 5.07, p<0.05
Tasks
PMPNLSLSNWLSW
Mea
n R
eact
ion
Tim
e ( R
T)( i
n m
s)
2400
2200
2000
1800
1600
1400
1200
1000
800
600
Grades
III
IV
V
VI
Figure 8: Comparison of mean reaction time (RT) of DRC component tasks across grades in
normal children.
Repeated measures ANOVA was computed for within subject effects on each grade
separately for RT measures in normal children on component Tasks. In grade III, there was
an overall significant main effect on component tasks i.e., F (3, 27) = 3.58, p< 0.05. In grade
V normal children, there was an overall significant main effect on reading tasks i.e., F (3, 27)
= 5.02, p< 0.05.
Wilcoxon singed rank results revealed that there was a significant difference in
performance for the pairs PN and PM, for PN and LSW, PN and LSNW at 0.05 levels.
Mann- Whitney test was done for each grade across both the groups. In grade IV there was a
significant difference for LSNW and overall performance on LS at 0.05 levels. For grader V,
there was a significant difference for PN task at 0.05 levels and not for LSNW, LSW and PM
tasks.
Tasks
PMPNLSLSNWLSW
Mea
n R
eact
ion
Tim
e ( R
T) (
in m
s)
2400
2200
2000
1800
1600
1400
1200
1000
800
600
Grades
III
IV
V
VI
Figure 9: Comparison of mean reaction time (RT) of DRC component tasks across grades in
children with dyslexia.
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
108
To summarize, the performance in terms of accuracy and RT measures in CWD was
significantly poor in comparison with normal children on all the component tasks. CWD
performed poorly on PM tasks compared to LSW and LSNW, PN tasks.
III. Subtypes of dyslexia derived from the DRC model
Hierarchical cluster analysis was done for the purpose of Subtyping children with
developmental dyslexia. In this method the clusters were represented in dendrograms for the
tasks which tests the component in the nonlexical route/sublexical route (phonological tasks)
and also for the lexical route (non phonological tasks) in the dual route cascaded model
(Coltheart et al., 2001)
Phonological Tasks
Hierarchical Cluster analysis for phonological tasks on children with dyslexia
revealed that cluster I {4, 15, 10} performed poorly on phoneme matching and reading
nonwords, cluster II {8, 16, and 12} also showed poor performance in phoneme matching and
nonword reading and cluster III{6, 11}and IV {5, 14} also showed poor performance on the
above same tasks. Cluster V {2, 3} and cluster VI {7} showed poor performance only on
nonword reading and these cluster was later joined by subjects 1 and 9. And this cluster V
was less severe when compared to the above four clusters (I, II, III, IV) on phonological
tasks.
Non-Phonological Tasks
Hierarchical cluster analysis non-phonological tasks in CWD revealed that the Cluster
I {1, 12} and cluster IV {3} which had poor performance in reading irregular words and V
{6, 13} had poor performance in reading irregular words only. Cluster II {14, 15} had poor
performance in letter search words and reading regular words. Cluster III {4, 16} both had
the poor performance in letter search words and reading irregular words. Here clusters (I, IV,
V) are very less severe than cluster II and III because they had poor performance in only one
task (reading irregular words). Clusters VII {2} and VIII {9, 10} had poor performance in all
of the non-phonological tasks and so they are more severe in type than the other clusters.
Table 1. Subtyping of children with dyslexia based on phonological and non-phonological
tasks.
subjects
(CWD)
Phonological tasks
accuracy (%)
Non-phonological tasks
accuracy (%)
subtypes of
dyslexia
1 RNW (20%) RIW (40%) Mixed
2 RNW (10%) RRW (30%), RIW (30%), PN (13%) Mixed
3 RNW (0%) RIW (0%) Mixed
4 PM (20%), RNW (0%) LSW (45%), RIW (30%) Mixed
5 PM (40 %), RNW (25%) RIW (30%) Mixed
6 PM (30 %), RNW (20%) RRW (40%) Mixed
7 RNW (15%) - Phonological
8 PM (20 %), RNW 10%) RIW (20%) Mixed
9 RNW (0%) RRW (0%), RIW (0%) Mixed
10 PM (30 %), RNW (0%) LSW (45%), RRW (0%), RIW (0%), Mixed
Subtyping of Children with Developmental Dyslexia
109
11 PM (30 %), RNW (15%) - Phonological
12 PM (20 %), RNW (10%) RIW (30%) Mixed
13 PM (30 %), RNW (5%) RRW (40%) Mixed
14 PM (30 %), RNW (20%) RRW (50%) Mixed
15 PM (40 %), RNW (25%) LSW (30%), RRW (30%), Mixed
16 PM (10 %), RNW (10%) - Phonological
Table 1 summarizes the Subtypes of children with developmental dyslexia based on
the lexical and nonlexical route; this table revealed that subjects {7, 11, and 16} are purely
phonological dyslexics because they performed poorly on phonological tasks and better on
non-phonological tasks. And all other subjects performed poorly on both the phonological
and non-phonological tasks so they are grouped into mixed dyslexics {1, 2, 3, 4, 5, 6, 8, 9,
10, 12, 13, 14, 15}. There were no pure surface dyslexics found in the present study.
Discussion
The results and the findings of the present study are discussed below in the following
sections,
Performance of normal children and CWD on reading tasks
The performance of normal children on accuracy measures improved from lower
grades to higher grades thus indicating a developmental trend on irregular word reading tasks
and on reaction time measures this trend was found on reading irregular words and nonwords
but not for regular words. This developmental trend could be explained using dual route
cascaded (DRC) model (Coltheart, Rastle, Perry, Langdon, & Ziegler, 2001). In the present
study there was a significant difference across the tasks [regular words (RW), irregular words
(IRW) and non-words (NW)]. Among the reading tasks, accuracy and reaction time measures
was poorer on nonword reading task compared to irregular and regular words in normal
children. These results are in consonance with the study by Burani et al., (2002) who derived
reaction times (RTs) for naming non-words in third- to fifth-grade children. They attributed
the difference in performance effect of younger graders to the older one due to
morphological structure ,that nonwords made up of roots and derivational suffixes which
looks similar to words are read quickly than the nonwords no morphological constituency.
And the reaction time for naming reduced as the grade levels increased.
In the present study, children with dyslexia (CWD) performed poorly on all reading
tasks compared to normal children on both accuracy and reaction time measures. On nonword
reading task CWD performed poorly compared to irregular and regular word reading. This
poorer performance on nonword reading in CWD could be explained using the DRC model
(Coltheart et.al, 2001). When there is a problem in the GPC module and proper phoneme to
grapheme correspondence doesn‟t take place this leads to the impaired sublexical route, so
more errors were seen in the nonword reading. These results are in consonance with the study
by Landerl, Wimmer and Firth (1997) who studied nonword reading in 12 year old CWD and
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
110
reading level matched normals. Their results revealed that CWD performed poorer on
nonword reading compared to reading level matched controls.
Results also revealed that poorer performance on irregular word reading in CWD
compared to normals and this could be explained using DRC model (Coltheart et.al 2001).
For reading irregular words, there should be an intact lexical nonsemantic route which
doesn‟t involve the semantic lexicon and there is a direct connection between the
orthographic lexicon and phonological lexicon (see fig1). These results were in consonance
with the study by Coltheart and Leahy (1996) who reported that learning to read in English is
characterized by a slow increase in accuracy, although this is quite apparent for irregular
words, many errors in reading regular words can still be expected even after several years of
schooling.
Performance of normal children and CWD on different components of the DRC model.
In the present study, poor performance on letter search task for nonwords were seen in
normal children for both latency and accuracy measures. This could be explained through
DRC model (Coltheart et.al, 2001) (see Figure 1). According to this model, the letter search
task involves letter processing i.e. the letter unit component of the model (Ziegler, Van
Orden, &Jacobs, 1997), where each letter in the word is processed visually and sent to the
orthographic lexicon. If there is a problem in the letter unit processing it can lead to delay in
the letter search task or can indicate reduced accuracy and latency measures. The results also
revealed that accuracy and reaction time measures on letter search words and nonwords were
poorer compared to normal children. This poorer performance could be explained through the
DRC model (Coltheart et.al, 2001). There may be impaired connection between the letter unit
and the orthographic lexicon or problem at the level of letter unit processing. And problem at
the level of letter units causes visual processing errors. Such errors were reported in a number
of recent studies as well (Bosse, Tainturier, & Valdois, 2007)
On picture naming tasks (PN), the results of the present study showed that in normal
children the performance was better in terms of both accuracy and reaction time measures.
This could be explained taking support from the DRC model (Coltheart et al, 2001). Picture
naming tasks are supposed to involve access to the phonological lexicon in the lexical route
(Wolf & Bowers, 1999). Here there is no involvement of the letter units and orthographic
lexicon because of no written form. So it directly accesses the phonological lexicon for
naming the picture (see Figure 1).
The results also revealed that there was no significant difference in accuracy measures
on rapid picture naming tasks for both normal and CWD group. However, there was a
significant difference in the reaction time measures for phoneme matching tasks for both the
groups. Longer reaction time in CWD may be due to the delayed processing that takes place
in the phonological lexicon which results in the longer processing time. Wolf and Bowers
(1999) opined that deficits in rapid naming are most strongly associated with deficits in the
development of orthographic representations for words.
Subtyping of Children with Developmental Dyslexia
111
On phoneme matching tasks, the results of the present study revealed that both
normal children and CWD performed poorer on phoneme matching (PM) tasks. When
compared to normal children, CWD performed poorer on this task. The reason for this
trend in n could be explained by DRC model (Coltheart et al, 2001). Phoneme matching
task measures the capacity to detect and manipulate phonemes while not requiring
orthographic or visual-attentional processes. Although phoneme matching does not directly
measure the GPC procedure, the claim is that meta-linguistic awareness of individual
phonemes is necessary to create grapheme–phoneme mappings (Hulme, Caravolas, Malkova,
& Brigstocke, 2005).
Subtypes of dyslexia derived from the DRC model.
Hierarchical cluster analysis results showed that the performance of the children with
dyslexia on reading tasks and tasks related to the components of the model did not find a
convincing interpretation of the subtypes in terms of single dissociated deficits. Rather than
having a single deficit on either the lexical or nonlexical route, surface and phonological
dyslexics seem to have multiple deficits in both the lexical and sublexical route. The findings
are in consonance with the study by Pennington (2006) who has discussed the existence of
heterogeneity in developmental dyslexia. From the cluster analysis and the qualitative
analysis two major subtypes of dyslexia were derived which included phonological and
mixed subtypes. Literature review suggests broadly three subtypes of developmental dyslexia-
phonological, surface and mixed type or unclassified (Castles & Coltheart, 1993; Ziegler,
Castel, George, & Perry, 2008).
The existence of phonological and mixed subtypes of dyslexia in the present study
can be explained taking support from the DRC model (Coltheart, Rastle & Perry, 2001). In
the present study, three out of sixteen children were found to be phonological subtype {7, 11,
and 16} of dyslexia (see Table). And these children have been found to perform poorly on
phonological related tasks like nonword reading (NW), phoneme matching (PM) and letter
search nonwords (LSNW). These results were in consonance with the study by Castles and
Coltheart (1993) identified pure developmental phonological dyslexics among CWD and
their nonword reading was found to be poor compared to chronological age- matched
controls, but their exception word reading was within normal range. Castles and Coltheart
(1993) concluded that these results were best interpreted in terms of a dual route model, with
the subtype profiles representing different levels of development of the lexical and nonlexical
procedures.
Thirteen out of sixteen were found to be of mixed subtype
{1,2,3,4,5,6,8,9,10,12,13,14,15} of dyslexia(see Table 5). And these children have been
found to perform poorly on phonological tasks like nonword reading (NW), phoneme
matching (PM) and letter search nonwords. (LSNW) and non-phonological related tasks like
picture naming (PN), letter search words, reading regular (RW) and irregular words (IRW).
These results could be explained using the DRC model (Coltheart et.al 2001). When both the
lexical route and sublexical route are affected, it may show up errors in both phonological
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
112
tasks and non phonological tasks. Most of the errors that were seen when both the routes are
affected are of irregular word and nonword reading tasks.
These results of the present study are in consonance with the study by Castles and
Coltheart (1993) who also observed that many poor readers are impaired on both irregular
word and nonword reading tasks. They hypothesized that the reason why a correlation
between irregular and nonword reading performance might also be expected on a dual-route
account is related to the structure of the model itself, as the computational Dual Route
Cascaded (DRC) model of Coltheart and colleagues (Coltheart, Rastle, Perry, Langdon, &
Ziegler, 2001).
Conclusions
Thus, the present study explained development of various component processes of
reading in normal children where these processes develop from lower grades to higher grades
for different components of reading. Also, the study supports the existence of subtypes of
dyslexia even in Indian children learning to read and write the alphabetic language, English.
Implications of the study
The present study has investigated individually the components of the DRC model.
This gives us an idea at which representation level the individual is having problem or
whether a single component is affected or multiple components are affected. This will
further help the clinician to plan intervention on those levels or components which
have deficits in children with dyslexia.
Studying the subtypes of developmental dyslexia (surface or phonological or mixed)
based on DRC model has made way for us to explore whether only lexical route is
affected or only the non-lexical route is affected or both. This would aid us in
understanding that subtypes could exist in developmental dyslexia and thus emphasize
on the need to design an Individualized education program (IEP) for intervention
depending on the subtypes of dyslexia.
The present study is only an initial attempt to investigate the reading performance and
subtypes of children with developmental dyslexia using DRC model. Future studies
are warranted in Indian population to study processing in semisyllabic Indian
languages (like Kannada, Hindi, etc.). Also, studies related to cross language
influences is also warranted to see whether there is facilitation or interference of an
alphabetic language with a non-alphabetic language can also be explored.
Limitations of the study
The number of subjects considered for each grade in the study is very limited and
hence difficult to generalize to normal population or children with developmental
dyslexia.
Subtyping of Children with Developmental Dyslexia
113
The present study is only an initial attempt to investigate the reading performance and
subtypes of children with developmental dyslexia using DRC model. Other factors
like the influence of native language or mother tongue or socio –economic factors,
literacy level of parents, etc. have not been explored and explained in the present
study.
References
Anitha, R. (2004). Computerized Linguistic Protocol for Screening (CLIPS). Unpublished
Master‟s dissertation submitted to University of Mysore.
Behrmann, M., & Bub, D. (1992). Surface dyslexia and dysgraphia: Dual routes: A single
lexicon. Cognitive Neuropsychology, 9, 209-258.
Bosse, M.L., Tainturier, M.J., & Valdois, S. (2007). Developmental dyslexia: The visual
attention span deficit hypothesis. Cognition, 104 (2), 198-230.
Burani, C., Marcolini, S., & Stella, G. (2002). How early does morpholexical reading develop
in readers of a shallow orthography? Brain and Language, 81, 568-586.
Castles, A., & Coltheart, M. (1993). Varieties of developmental dyslexia. Cognition, 47, 149-
180.Coltheart, M., & Leahy, J. (1996). Assessment of lexical and nonlexical reading
abilities in children: Some normative data. Australian Journal of Psychology, 48, 136-
140.
Coltheart, M., Rastle, K., Perry, C., Langdon, R., & Ziegler, J. C. (2001). DRC: A dual route
cascaded model of visual word recognition and reading aloud. Psychological Review, 108,
204-256.
Forster, K. I., & Forster, J. C. (2003). DMDX: A windows display program with millisecond
accuracy. Behavior, Research Methods, Instruments & Computers, 35, 116-124.
Funnell, E. (1983). Phonological processes in reading: New evidence from acquired dyslexia.
British Journal of Psychology, 74, 159-180.
Hulme, C., Caravolas, M., Malkova, G., & Brigstocke, S. (2005). Phoneme isolation ability is
not simply a consequence of letter-sound knowledge. Cognition, 97, B1-B11.
Landerl, K., Wimmer, H., & Firth,U. (1997). The impact of orthographic consistency on
dyslexia: a German –English comparison. Cognition, 63,315-334.
Loomba, M. (1995). Descriptive analysis of the sequential progression of early reading skills
among Indian children. Unpublished Master‟s dissertation submitted to University of
Mysore.
Pennington, B. F. (2006). From single to multiple deficit models of developmental disorders.
Cognition, 101, 385-413.
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
114
Perry, C., Ziegler, J. C., & Zorzi, M. (2007). Nested incremental modeling in the
development of computational theories: The CDP+ model of reading aloud.
Psychological Review, 114(2), 273-315.
Rayner, K., Murphy, L.A., Henderson, J.M., & Pollatsek, A. (1989). Selective attentional
dyslexia. Cognitive Neuropsychology, 6, 357-378.
Singhi, P., Kumar, M., Malhi, P., & Kumar, R. (2007). Utility of the WHO Ten Questions
Screen for Disability Detection in a Rural Community - the North Indian Experience.
Journal of Tropical Pediatrics. 53, 6, 383-387.
Wolf, M., & Bowers, P. (1999). The question of naming-speed deficits in developmental
reading abilities: An introduction to the double deficit hypothesis. Journal of
Educational Psychology, 91, 1-24.
Ziegler, J. C., Castel, C., George, C., & Perry, C. (2008). Developmental dyslexia and the
dual route model of reading: Simulating individual differences and subtypes.
Cognition, 107,151-178.
Ziegler, J. C., Van Orden, G. C., & Jacobs, A. M. (1997). Phonology can help or hurt the
perception of print. Journal of Experimental Psychology: Human Perception and
Performance, 23, 845-860.
Dyslexia Assessment Profile for Indian Children
115
Dyslexia Assessment Profile for Indian Children (DAPIC)
Kuppuraj. S & Jayashree. C. Shanbal*
Abstract
Reading is a complex process which includes several underlying cognitive processes. The
successful assessment of reading has been challenged by factors like heterogeneity of the dyslexic
population. There is a need to assess children with dyslexia considering the high prevalence of
dyslexic population even in the Indian context. The present study aimed to develop an assessment tool
that can be used to study the development of reading and writing skills in Indian children. The
present study also focused on profiling reading and writing characteristics of children with dyslexia.
Comparisons were made between normal and dyslexic group and findings revealed that tasks like
rapid naming (RN), alliteration, sound discrimination (SDis) and spelling were found to be good
predictors for reading ability. Further, analysis revealed three major subtypes of dyslexia-
phonological, surface and mixed types of dyslexia.
Introduction
Reading is a complex cognitive process that involves multiple skills. Components of
reading ability include phonological awareness, phonological decoding, reading
comprehension, spelling, orthographic knowledge and rapid automatized naming (Gayan &
Olsen, 2003). Before one becomes completely literate, he or she must develop early reading
skills which comprised of steps like literacy awareness, syntactic awareness, word
recognition, phonological awareness, orthographic awareness (Strommen & Mates,
1997).Reading process is the conversion of print into auditory equivalents and the subsequent
interpretation of those equivalents into meanings based on previously learned language.
Comprehension is the next step which is the ability to interpret and understand the decoded
words. The next component of literacy is writing. Written language is characterized by rules
of phoneme grapheme correspondence (Nagy, Berninger, Abbott, Vaughan & Vermaulen,
2003). The rules of orthographic knowledge (processing written language letters and letter
patterns), phonological knowledge (processing or manipulating oral language sounds), and
morphological awareness in spelling performance have been well documented. As a result
spelling is related to reading and written expression. There are several stages through which a
child travels before master spelling skill. Gentry (1982) proposes stages like
precommunicative (where the child uses symbols from the alphabet but shows no knowledge
of letter-sound correspondences), semi phonetic (sounds are assigned to letters), phonetic
(The child uses a letter or group of letters to represent every speech sound that they hear in a
word), transitional (the speller begins to assimilate the conventional alternative for
representing sounds),and correct stage (the speller knows the English orthographic system
and its basic rules).
____________________________________ * Lecturer in Speech Pathology, All India Institute of Speech and Hearing, Mysore, India
email: [email protected]
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
116
Fletcher, Lyon, Fuchs and Barnes (2006) have quoted that reading is not natural, and
everyone is taught to read. Those children who lack the ability to learn literacy skills are
considered „dyslexic‟. Smythe and Everatt (2000) define dyslexia is a difficulty in the
acquisition of literacy skills that may be caused by combination of phonological processing,
visual and auditory system deficits. Lexical confusions and speed of processing difficulties
may also be present. The manifestation of dyslexia in any individual will depend upon not
only individual cognitive differences, but also the language used.
There are several causes proposed in the literature for dyslexia. They are heredity
(Pennington, 1989), brain differences (Galaburda, 1991), defects in rapid temporal
information processing (Tallal, Miller & Fitch 1993), selective attention and attention deficit
disorder (Zentall, 1993), middle ear problems (Roberts & Medley, 1995), cognitive rigidity
and learned helplessness (Clay, 1984). Similar to numerosity in causes of dyslexia, there exist
numerous types of learning difficulty (dyslexia). Theoretically, there are as many types of
dyslexics making the dyslexic group a heterogeneous one.
Several models have been proposed in order to hypothesize at what level the difficulty
could exist in children with dyslexia. Newcombe and Marshall (1984) proposed a reading
model which explains normal visual word reading. This model composed of two routes
namely lexical (direct route) through which words and irregular words are read. Another route
of the model is the sublexical route (indirect route) through which non words are read. Several
researchers have attempted to understand dyslexia based on the processing along these two
routes.
Castles and Coltheart (1993) broadly classified dyslexia in to two main types called
phonological and surface dyslexics. He opined that with difficulties in sub lexical skills shall
come under phonological dyslexics and if the children with dyslexia have difficulties in
lexical skills, he/she may fall under surface dyslexia. Edwards and Hogben (1999) included
another type of dyslexia called mixed dyslexia. Dyslexia of this subtype tends to find
difficulties in both lexical and sublexical route of reading, along with already proposed two
types.
There are several standardized western assessment tools available to assess children
with dyslexia. However, developing tools for children with dyslexia in Indian multilingual
context has been challenging. Moreover, assessment tools for dyslexia in India is the need for
the hour as the estimated prevalence rate of learning disability has been found to range from
3% to 10 % (Ramaa, 1985) in India. The high prevalence rate of learning disability indicates
the need for early identification and tailoring individualized intervention programs for such
children once they have been identified using the tools. Reviewing the available literature, it
has been found that dyslexic population is vastly heterogeneous and this heterogeneous nature
in the dyslexic group poses the requirement to profile the learning disabilities on individual
based performances of these children. It is important that the children with dyslexia are sub
grouped under existing subtypes based on their individual performances after profiling so that
an appropriate treatment program can be developed for these children.
Dyslexia Assessment Profile for Indian Children
117
Objectives of the study
The objectives of the study were-
I. To study the development of reading and writing skills in Indian children.
II. To identify subtypes of dyslexia based on the profiles established on children
with dyslexia.
Method
The participants of the present study included two groups. One group consisted of 60
school going normal children from grades I, II, III, IV and V. Each grade consisted of twelve
children. The other group consisted of 16 children with dyslexia (CWD) who were identified
as dyslexia at the All India Institute of Speech and Hearing, Mysore, using Test for Early
Reading Skills (ERS) (Loomba, 1995) in the clinical set up. All the participants were native
Kannada speakers with English as the medium of instruction.
A WHO Ten –Question Disability Screening Checklist (cited in Singhi, Kumar,
Prabhjot and Kumar, 2007 ) and Developmental Screening Test (Bharathraj, 1972) was used
to screen for normal children in terms of hearing, intelligence, motor and other factors like
school performance, emotional or behavioral factors.
The study was conducted in two phases.
Phase I: Test Development
The test was adopted from International Dyslexia Test (IDT) developed by Ian
Smythe (2000). This test was further modified so that it was culturally relevant to our
population. The final test that was derived out of the IDT after revisions was called the
Dyslexia assessment profile in Indian children (DAPIC).The subtests adopted from the IDT
(Smythe, 2000) included, Alphabet ,Shape copying, Spelling, Reading words and non words,
Phonological awareness skills, Word and non word repetition, Rapid naming and Sound
discrimination.
The tasks of the material have been divided under phonological and non phonological
tasks. The tasks those are considered as phonological are Alphabet, sound
discrimination(SDis), non word reading (NWreading), non word repetition(NWrep),
alliteration and rhyming. And the tasks those are considered as non phonological are Word
repetition(Wrep), word reading(Wreading), handwriting quality(HQ),shape copying(
SC),rapid naming(RN) and spelling. The following Table shows illustrations of tasks and
scoring.
Phase II: Test administration
Testing environment.
The test was administered in a quiet, noise free set up using a paper and pencil only.
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
118
Table 1: Details of tasks and scoring for each of the subtests
Sl.No. Subtest Task Scoring 1. Alphabet To write the alphabet that is named. Alphabets
those are visually similar are considered.
Which includes b,d,n,u,m,w,p,9,q
1- Correct
0- Incorrect
Total score: 9
2. Shape copying(SC) Copy the shapes that are given.
There are totally four shapes
(The shape that is displayed supposedly the
most complicated)
7- Most approximating
shape
1- Least clear shape.
Score:7
1 for correct and 0 for
incorrect shapes for
another three simple
shapes.
Score:3
Maximum score:10
3. Written Language
a)Spelling
Has to write down the spelling for words and
non words that are dictated. There are 30
words, 10 non words. However, the non words
are also following the phonotactics of English
1- Correct
0- Incorrect
Total score:40
b)Handwriting
quality(HQ)
Scored from Spelling performances 5- Very good
handwriting
1- Bad handwriting.
Total score: 5
4. Reading
word(Wreading)
Has to read the words that are given(list
consists of few irregular words)
1- Correct
0- Incorrect
Total score: 70
5. Reading non
word(NWreading)
Has to read the non words that are given 1- Correct
0- Incorrect
Total score: 10
6. Repetition of word
(Wrep)
Has to repeat the word after examiner. 1- Correct
0- Incorrect
Total score: 7
7. Repetition of non
word(NWrep)
Has to repeat the nonword after examiner 1- Correct
0- Incorrect
Total score: 8
8. Rhyming and
Alliteration
Has to find out the word which are in rhyme
E.g. Bat, Mat, Wall- Here “Bat” and “Mat” are
in rhyme with each other where as the „Wall”
is not
He has to find out the words which are in
alliteration with other words.
E.g. Shine, Shoe, Shop, Monsoon. Here all the
words but Monsoon begin with different letter,
or not in alliteration with other words.
1- Correct
0- Incorrect
Total score: 30
Rhyming:20
Alliteration:10
9. Rapid naming(RN) Has to name the pictures those are given. Time taken to
completely name all the
pictures.
10. Sound discrimination Has to say whether presented two words are
same or different.
E.g.
Cat, Rat- Different
Pin, Pin- Same
1- Correct
0- Incorrect
Total score: 20
Dyslexia Assessment Profile for Indian Children
119
The scoring was done as given in the Table 1. The data was subjected to quantitative
(details of statistical procedure are mentioned in the results section) analysis and qualitative
analysis. The results of this study have been discussed in the following sections.
Results
The aims of the study were
I. To study the development of reading and writing skills in Indian children.
II. To identify subtypes of dyslexia based on the profiles established on children with
dyslexia
Performance of normal children and children with dyslexia (CWD) across various tasks
Table 2 shows mean and standard deviation (SD) for normal children from Grades I to
V on all the tasks.
Table 2 shows that the performance of normal children improved from lower grades
to higher grades on tasks like HQ, SC, spelling and Wreading. The results revealed a
developmental trend on these tasks. It is evident from Table 2, that , on the alphabet subset,
all children in Grade I itself were able to score the maximum. A ceiling effect was seen for
alphabet task in Grade I itself. This indicates that the development of alphabet writing has
already taken place by Grade I itself.
Table 2: Mean and Standard Deviation (SD) of scores of Grades I to V.
Tasks
Grades
I II III IV V
Mean SD Mean SD Mean SD Mean SD Mean SD
Alphabet 9 0 9 0 9 0 9 0 9 0
HQ 3.83 0.94 4.00 0.74 4.17 0.72 4.67 0.89 4.83 0.39
RN 37.67 4.12 40.92 7.66 39.50 5.27 31.50 4.96 25.75 4.00
SC 8.08 2.07 8.75 1.22 8.92 1.24 9.42 .90 9.92 0.29
NW reading 5.17 1.75 5.08 2.78 7.00 2.26 7.33 1.83 7.58 1.97
Alliteration 2.92 2.27 3.50 2.71 6.42 2.15 6.17 2.40 9.08 1.24
Rhyming 3.92 3.39 0.42 0.99 7.33 2.35 10.33 4.70 14.50 3.00
Wrep 4.17 1.11 3.75 1.06 4.92 0.90 5.58 0.99 5.25 0.96
NWrep 4.33 0.89 4.17 1.53 5.17 1.11 6.17 0.72 6.50 0.67
SDis 17.42 1.88 17.08 1.56 18.67 1.15 19.58 1.44 20.00 0
Spelling 8.67 3.85 8.75 2.14 16.00 3.36 18.42 7.09 29.25 3.16
Wreading 15.58 7.23 20.08 4.17 26.67 4.03 31.25 10.97 54.67 6.36
Note: Handwriting Quality (HQ), Rapid Naming (RN), Shape Copying (SC), Non Word Reading (NW reading),
Word Repetition (Wrep), Non Word repetition (NWrep), Sound Discrimination (SDis),Word Reading (W
reading).
On handwriting quality (HQ), results showed that children in the higher grades
showed better performance than the lower grades (see Table 2).On shape copying (SC) task,
results revealed that children in the higher grades showed better performance than the lower
grades (see Table 2). On both spelling and reading tasks, results revealed that children in the
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
120
higher grades showed better performance than the lower grades (see table 2). Overall, there
was a developmental pattern observed on handwriting quality (HQ), shape copying (SC),
spelling and reading words (Wreading) (see Figure 1). Results in Table 2 further showed that
for NWreading, rhyming, Wrep, NWrep and SDis tasks, showed that the performance of
children was significantly better only in the higher grades (III to V). Performance of children
on alliteration task was found to be significant in the lower grades (I to III). On rapid naming
(RN) task, no significant difference was found among Grades I through III, however grade V
participants performed significantly better than Grade IV.
To summarize, a developmental trned was found for performance of normal children
across the tasks- HQ, SC, alphabet, spelling,Wreading, NWreading, alliteration,
rhyming,Wrep, NWrep, RN, SDis. A multiple regression analysis revealed that rapid naming,
alliteration, sound discrimination and spelling were the four significant predictors for the
reading score (r2=0.89; Rapid Naming β =-.469; p<0.05; Alliteration β =1.016; p<0.05;
Sound Discrimination β =-1.54; p<0.05; Spelling: β=1.3 03; p<0.001.
Comparison of performance of normal children and children with dyslexia (CWD)
The CWD group consisted of sixteen children with dyslexia (CWD) ranging from
grade I to V who participated in the study. An individual profile of each child with dyslexia
was derived after a detailed qualitative analysis of all the samples. There was no significant
difference found in the performance of normal children and CWD on tasks of alphabet, SC,
NWreading, and Wrep were not significantly different.
Table 3 displays the mean and standard deviation values for normal children and
CWD. Results showed that normal group were significantly better than CWD group on most
of the tasks but CWD performed as good as normal children on phonological awareness tasks
like alliteration, rhyming and Dis. Figure 4 compares the performances of normal and CWD
groups across all the tasks across Grades I to V.
Table 3: Mean and SD for overall performance of normal children and CWD.
Subtests
Groups
Normal CWD
N Mean SD N Mean SD
HQ 60 4.30 .83 16 2.44 1.26
RN 60 35.07 7.71 16 50.63 20.30
Alphabet 60 9.00 .00 16 7.94 1.91
SC 60 9.017 1.38 16 6.56 2.37
NWreading 60 6.43 2.35 16 2.75 1.39
Alliteration 60 5.62 3.10 16 3.94 3.13
Rhyme 60 7.30 5.79 16 4.12 4.21
Wrep 60 4.73 1.19 16 4.00 1.09
NWrep 60 5.267 1.38 16 4.12 1.26
SD 60 18.55 1.76 16 16.94 5.94
Spelling 60 16.22 8.68 16 6.50 5.39
WReading 60 29.65 15.31 16 12.38 8.71
Dyslexia Assessment Profile for Indian Children
121
Note: Handwriting Quality (HQ), Rapid Naming (RN), Shape Copying (SC), Non Word Reading (NWreading),
Word Repetition (Wrep), Non Word repetition (NWrep), Sound Discrimination (SDis), Word Reading (W
reading).
From the Figure 4, it is evident that CWD were found to perform poorly than the
normal children on all the tasks.
[HQ-Handwriting Quality, RN-Rapid Naming, SC-Shape Copying, NW reading-Non Word reading, Wrep-
Word repetition, NWrep-Non Word repetition, SDis -Sound Discrimination, WReading-Word reading]
Figure 1: Mean scores for performance of normal children and CWD across tasks.
The performance of CWD was compared with normal children qualitatively, the
CWD showed,
Regularization error, inappropriate spacing between letters, unfinished or omitted
words. mirror images of the alphabet
CWD who performed better on shape copying performed better on handwriting
quality.
Working memory deficits were masking the real performance on alliteration and
rhyming tasks.
Implementing lesser lexical cue on the Wreading task.
Inappropriate phonological working memory was evident on the repetition (words and
nonwords) tasks.
Sub typing of dyslexia based on profiling and cluster analysis
Cluster analysis was carried out as part of quantitative analysis for the CWD
group in order to classify them in to different clusters based on homogeneity among subjects
on various tasks. The tasks those were considered to arrive at overall Dendrogram were SDis,
Wrep, NWrep, Wreading, NWreading, alliteration, rhyming, RN and spelling. Analysis
revealed that Subjects 3, 2, 5, 6 performed similarly thereby forming a cluster (Cluster I)
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
122
themselves, Subject 1 (Ia) also performed similar to this cluster but it fell slightly apart, since
this subject couldn‟t manage any score on Spelling task, Subject 4(Ib) Having taken
maximum time for performing the Rapid Naming task falls slightly apart from Cluster the
cluster I and Ia ,Subjects 7 and 9 formed the cluster II. Cluster III was formed by the subjects
8, 12 &13, Subject 11 (IIIa) was slightly different from the cluster III, since its performance
on spelling was better compared to subjects in cluster III. Subjects 14, 15 and 16 fall in same
region of the Dendrogram forming its own cluster (Cluster IV).Ultimately, lonely subject that
didn‟t seem to fall near any of the cluster was the subject 10. This was the only subject who
read half of the words correctly in reading task (i.e.35 words), and scored maximum in this
task compared to all the other tasks.
In order to classify the CWD into phonological and non phonological group the data
was further analyzed considering phonological and non phonological tasks separately. The
phonological tasks considered were SDis, NWrep, NWreading, alliteration and rhyming. The
non phonological tasks considered were Wreading, NWreading, RN and spelling.
Phonological clusters
Subjects 2, 3, 5 and 6 formed cluster a, Subjects 1 and 4 formed the cluster (b),
Cluster(c) was formed by subjects 7 and 9, Cluster d was formed by the subjects 8, 11 &13.
Subject 12 formed the Cluster (d1), since this subject performed fair in rhyming task, Cluster
e was formed by the subject 14 since this subject had salient feature of very poor performance
on rhyming and good performance on alliteration task, Cluster (f) was formed by the subject
15, the salient phonological feature it had from other clusters was that it had very good
performance on alliteration and fair performance on rhyming task, Cluster (g) is formed by
subjects 10 and 16 and the common features these subjects shared were that all the
phonological tasks were performed fair to good. Even very good performance noticed in
alliteration tasks of subject 10
Non-phonological clusters
Subjects 8, 12 and 13 forms the cluster (a), the common non phonological features of
this group were, that all the subjects performed poorly on Spelling, WReading and Wrep.
Subject 7 and 9 forms the cluster (b), the common non phonological features they shared
were that they had good performance on Wrep and the correct responses in Wreading task
was over ten in both the subjects (poor). Subjects 11 formed cluster (b1) performed fairly on
spelling, fairly on Wreading ad very poor on Wrep thus forming a different cluster .Cluster
(c) formed by 15 and 16, the features of this cluster were their fair performance on Wrep and
Spelling tasks and Very poor to poor performance in Wreading task. Subject 10 formed the
cluster (d) performed good on Wreading and spelling and poor on Wrep. Subject 14 formed
the cluster (d1), since they had very poor performance in Wreading and spelling. These
clusters also had fair performance in RN and Wrep. Subjects 2, 3, 5 and 6 forms the Cluster
(e) the common non phonological features were, They performed poorly on spelling and
Wreading but fair performance on Wrep task. Subject 1(e1) Spelled none of the words and
read none of the words. The subject repeated repeat only the first series of stimulus which
Dyslexia Assessment Profile for Indian Children
123
had only two words in it. Subject 4 formed the cluster (f) since they showed Very poor
performance on spelling and Wreading and fair performance on Wrep task. Table 4 displays
the classification of subgroups of dyslexics based on phonological and non phonological
tasks.
Table 4: Sub grouping of CWD under phonological, non phonological and mixed type.
Phonological Surface Mixed
Subjects 1,4,16,14 & 10 Subject 15 Subjects 2,3,5,6,7,8,9,11, 12 and 13
Subjects were classified as phonological dyslexics {1, 4, 16, 14, 10} if they showed
difficulties on tasks of NWreading, NWrep, alliteration and rhyming. Subjects were classified
as surface dyslexics {15} if they find difficulties in tasks of Wreading, Wrep, RN and
Spelling. Subjects were classified as Mixed {2, 3, 5, 6, 8, 9, 11, 12, 13} if they show
difficulty in both phonological and non phonological tasks.
Discussion
The results of the study are discussed in terms of,
Development of reading and writing skills based on the adopted IDT (Smythe, 2000)
in normal children from grade I through grade V. This section will also include
discussion over the performance of children with dyslexia in comparison to the
normal grade level children.
Subtypes of dyslexia based on the profile
Development of reading and writing skills in normal children and performance of CWD
On the alphabet task all the normal children performed excellently. The mirror
writing errors shown by lower grades of CWD disappeared as grades increased. Study by
Terepocki, Kruk, and Willows (2002) who compared 10-year-old average readers and
children with reading disability concluded that the difficulties of reading disabled group in
discriminating similar looking items could be due to poorly specified representations of
letters.
It is evident from the results of the present study performance of normal children
improved from grade I to grade V on both the visuo-motor tasks such as handwriting quality
and shape copying. This may be attributed to better fine motor control over this skill with
development. These findings of the present study evidenced that these skills are acquired
much earlier that they were ought to be as mentioned in the Gessel and Amatruda (1947)
study. With respect to children with dyslexia, the results of the present study showed that
CWD children were found to have poorer hand writing quality and shape copying in
comparison to normal children. Literature suggests that children with developmental dyslexia
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
124
have been found to have poor fine motor abilities and often poor writing skills (Denckla,
1985). Motor problems are frequently observed in dyslexic children (Snowling, 2000), thus
reflecting on their poor writing skills.
Results of the present study also showed that the performance on spelling task
increased from lower to higher grades. It was found that children in the earlier grades used
symbols from the alphabet but showed no knowledge of letter-sound correspondences
(indicating the precommunicative stage of spelling development according to Gentry (1982).
In the later grades children were found to make better letter-sound correspondences and
further higher grades (grade IV and grade V), children were able to understand the basic
spelling rules of English and use them appropriately for reading purpose. On spelling tasks,
CWD were found to perform poorer than the normal children in the present study. Qualitative
analysis revealed that CWD even in the higher grades showed more errors in comparison to
young normal children. The errors revealed that CWD in higher grades who were supposed to
have reached the orthographic stage of spelling (Frith, 1985) are still in logographic or
alphabetic or in a transition from logographic stage to alphabetic stage.
The present study showed improved performance on word reading and non word
reading tasks from lower to higher grades in normal children. It can be hypothesized that
these lower grades normal children are still in logographic to alphabetic stage of Frith‟s
model. Typically, according to Frith this alphabetical stage is followed by orthographic stage.
This refers to the mastery of the alphabetic principle of phonography according to which
written words may be segregated into left to right series of letters, each of which can be
decoded as standing for as segment of speech. Poor performance of normal children lower
grades on non word reading task can be explained using the developmental stage models of
reading like the dual route models suggested in literature (Castles & Coltheart, 1993) .To read
a non word, it is the sub lexical processing which is important. Results of the present study
are indicative that probably in the lower grades, children have still not developed the
component process (GPC buffer in case reading a non word) that requires sub lexical
processing to read a non word correctly. The performance of CWD was significantly poorer
than normal in both word reading and non word reading tasks. The poor performance of
CWD on Wreading can be explained through Frith‟s model (1985) of reading development.
According to this model, in order to read a word, it is significant to have an orthographic
stage of reading development. In the present study we can hypothesize that, this stage of
development is deficient in CWD, thus this population find difficulties reading word. CWD
even in the higher grades were found to perform poorly in the present study. It could be that
CWD have still not developed the component process (GPC buffer in case reading a non
word) that requires sub lexical processing to read a non word correctly (Castles & Coltheart,
1993).
On the word repetition task, normal children in the present study, showed no
significant improvement in the lower grades (grade I and II), whereas showed significant
improvement in the higher grades (grades III, IV and V) only. Gathercole & Baddeley
(1989); Hoff, Cotre and Bridges (2008) have reported that word repetition in very young
Dyslexia Assessment Profile for Indian Children
125
children is related to their vocabulary size of the child. This means that children in the higher
grades have better and greater vocabulary size than the lower grades and hence the former
perform better than the latter on word repetition task indicative of an obvious developmental
trend. This vocabulary delays might have caused poor vocabulary repertoire available for the
CWD. Similar to word repetition task, on non word repetition task, the performance found to
have improved significantly in the higher grades (III, IV and V) only. It has been found to be
indicative of phonological working memory capacity (Coady & Evans, 2008) which means
developmentally older children have a better capacity of phonological working memory
which in turn help them repeat or read non words appropriately compared to younger children
who are still building up their phonological repertoire and/or knowledge. The CWD of the
present study performed poorly compared to normal children on NWrep task suggesting poor
phonological working memory capacity, lexical (the degree of accuracy correspondence to
lexical intactness)and phonological knowledge. Study by Snowling (1981) showed that
children with dyslexia experienced a greater degree of difficulty with non-words repetition,
especially at longer lengths. They hypothesized that speech-motor programs would be used
for familiar words, while such programs would be unavailable for non-words. Successful
repetition of non-words would require „subjects to process the auditory stimulus, to decode
the sound segments, and to recode these as instructions in the form of a speech-motor
program .This decoding and recoding deficits reported by Snowling could be given as a
reason for poor performance of CWD on NWrep task.
On rapid naming task, children in the present study showed increasing performance
from lower to higher grades suggesting a developmental pattern. This could be because the
time taken to access a lexicon from their repertoire is faster with the development of children.
Study by Wagner, Torgesen, Roshotte, Hecht, Baker and Burgess (1997) investigated the
relative contribution of RN to later reading ability in three developmental periods from
kindergarten to grade II, from grade I to grade III and from grade II to grade IV. They
concluded that RN was a significant variation for word reading till kindergarten to grade III
only. Findings of present study contradict Wagner et al., (1997) study as in the present study
RN was found to be improving significantly in higher grades (grade IV and V) only. In other
words RN was also found to be influencing reading in children of higher grades.
On RN task, the performance of CWD was significantly poorer than normal children
suggesting poor phonological accessing speed in CWD. Our findings are in consonance with
study by study by Bowers and Wolf (1993) who also admitted the same findings in CWD.
On rhyming task, the results revealed that the performance improved significantly
from grade III till V. On the alliteration task the significant improvement in performance was
noticed only for higher grades (grade III, IV and V). This supports findings of Wagner et al.,
(1997) concludes after his study that phoneme awareness tasks are significant contributors to
reading in higher grades only (grade III and IV). Further, results of the present study showed
that on rhyming and alliteration task, CWD performed poorly than the normal children and
performance are varying across grades. Studies by Moats and Foorman (1997) and Adams
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
126
(1990) also suggests that children who learn to read alphabetic language system such as
English have been found to have poor phonemic awareness tasks like rhyming and letter
identification.
On SDis the performance on normal children were improving from lower through
higher grades. It was not found to be significant in lower and higher ends whereas only grade
III found to be significantly better than grade II. In other words there was no improvement in
sound discrimination skill as the child grows older. The insignificant improvement in scores
of auditory discrimination findings are in consonance with study by Neff (2006) who
concluded that both spectral and temporal cues attain adult like by the age of six. In the
present study the CWD showed poor performance compared to normal children, though it
was not statistically significant. Tallal (1980) proposed that children with specific reading
difficulties are deficient in processing brief and rapidly changing acoustic information like
brief acoustic events in stop consonants
Subtypes of dyslexia based on the profile
One of the major aims of the present study was to subtype children with dyslexia
based on the individual profiles obtained after administering the IDT (Smythe, 2000) and
cluster analysis done on the data obtained. It was evident from cluster analysis in the present
study that dyslexia is not a homogenous group but a heterogeneous group with existence of
three major subtypes of dyslexia including phonological, surface and mixed types. Children
with dyslexia (CWD) who fell under the phonological subtype in the study included {1, 4, 14,
16, 10}.The cluster that fell under the surface subtype of dyslexia included {13}. The cluster
under the mixed subtype included CWD {2, 3, 5, 6, 7, 8, 9, 11, 12, 15} (See Table 5).
The phonological subgroup
CWD of phonological subtype could have performed poorly on phonological related
tasks in comparison to the others due to deficit in the sub lexical processing of phonemes
(Castles & Coltheart, 1993). In this group it was found that CWD performed poor mostly on
nonword reading tasks. By theory, to read a non word, it is crucial to have phoneme
grapheme conversion system intact. However, children with phonological dyslexia are unable
to read non words as it is hypothesized that the sub lexical route which is composed of
grapheme to phoneme conversion system is grossly affected.
The surface subgroup
The reason for the performance of the surface dyslexic profile could be due to a
nonphonological deficit in the lexical route as suggested also by Castles and Coltheart,
(1993). In surface dyslexics the disconnection is in between the semantics and phonological
functions leaving only the sub lexical route in operation .So, the reading of non words is
relatively intact in this population.
The mixed subgroup
The performance of mixed group in the present study could be due to the deficits in
both the routes of reading. i.e., sublexical and lexical route as suggested by Edwards &
Dyslexia Assessment Profile for Indian Children
127
Hogben, 1999) It can be hypothesized that the combined deficits in orthographic lexicon and
in Grapheme Phoneme (GPC) rules could have resulted in mixed dyslexia .In other words
they have difficulty in reading stimuli that requires both lexical and sub lexical processing
(Edwards & Hogben, 1999).
Conclusions
The results of the present study revealed that DAPIC can be used as a tool to profile
those children who show difficulties in phonological and non phonological tasks of literacy.
The results of the present study reveal that differentiating among dyslexia subtypes with
specific impairments allows a more fine grained understanding of disorder than simply
comparing dyslexics with normal. The profiling and sub typing of the present study also lead
us to understand and plan for individualized education program (IEP) for children with
developmental dyslexia .The predictors those are mentioned as predictors of reading can be
used to make a screening tool. The tool needs to be further administered on a larger sample to
generalize the findings and use regularly for the assessment of children with reading
problems. This study included a small sample of children for each grade. Administering it on
larger sample would help in standardization of the tool for assessing children with dyslexia.
Other domain specific tasks of phoneme awareness skills like phoneme stripping, word attack
tasks could have been included in the test to give a complete profile of individual‟s phonemic
awareness skills.
References
Adams, M. J. (1990). Beginning to read: Thinking and learning about print. Cambridge,
MA: MIT Press.
Bowers, P.G., & Wolf, M. (1993).Theoretical links among naming speed, precise timing
mechanisms and orthographic skill in dyslexia. Reading and Writing, 5, 69-85.
Castles, A., & Coltheart, M. (1993). Varieties of developmental dyslexia. Cognition, 47, 149-
180.
Clay, M. M. (1984). (Ed). Reading: The patterning of complex behavior (2nd
Edn.).Auckland,
New Zealand: Heinemann Educational books.
Coady, J. A., & Evans, J. L. (2008). Uses and interpretations of non-word repetition tasks.
Dyslexia, 7, 197-216.
Edwards, V., & Hogben, J. (1999). New norms for comparing children‟s lexical and
nonlexical reading: A further look at subtyping dyslexia. Australian Journal of
Psychology, 5, 37-49.
Fletcher, J. M., Lyon, G. R., Fuchs, L. S., & Barnes, M. A. (2006). Learning disabilities.
New York, NY: Guilford. A.91, 8010-8013.
Dissertation Vol. VII, Part – B, SLP, AIISH, Mysore
128
Frith, U. (1985). Beneath the surface of developmental dyslexia. In K. E. Patterson, J. C.
Marshall, & M. Coltheart (Eds.), Surface Dyslexia: Neuropsychological and cognitive
studies of phonological reading. (pp. 301-330). Hillsdale, NJ: Erlbaum.
Galaburda, A. M. (1991). Anatomy of dyslexia: Argument against phrenology .In D.D
Duane&D.B. Gray, (Eds.), The reading brain: The biological basis of dyslexia
(pp.119-131). Parkton, MD: York press.
Gentry, R. (1982). An analysis of developmental spelling in GNYS AT WRK. The Reading
Teacher, 36, 192-200.
Gessel, A., & Amatruda, C.S. (1947). Developmental Diagnosis, 2nd
Ed. New York; Paul B.
Hoeber.
Hoff. E.,Core. C., Briges. K. (2008). Non-Word Repetition Assesses Phonological Memory
and Is Related to Vocabulary Development in 20- to 24-Month-Olds. Journal of
Child Language, 35, 4, 903-916.
Moats, L. C. & Foorman, B. R. (1997). “Introduction to the special issue of SSR:
Components of effective reading instruction.” Scientific Studies of Reading 1 (3),
p. 187-189.
Nagy, W., Berninger, V., Abbott, R., Vaughan, K., & Vermeulen, K. (2003). Relationship of
morphology and other language skills to literacy skills in at-risk second grade readers
and at-risk fourth grade writers. Journal of Educational Psychology, 95(4), 730-742.
Newcombe, F & Marshall, J.C. (1984). Varieties of acquired dyslexia:A linguistic approach.
Seminars in Neurology 4:2:181-95.
Ramaa,S. (1985). Diagnosis and remediation of dyslexia-An attempt. Ph.D dissertation,
University of Mysore, Mysore.
Roberts, J. E & Medley, L. P. (1995). Ottitis media and speech language sequale in young
children: Current issue ion measurement. American Journal of Speech Language
Pathology, 4, 15-24.
Smythe, I., (2000). International Dyslexia Test. Retrieved January 21, 2008 from E-mail
communication.
Snowling, M. J. (1981). Phonemic deficits in developmental dyslexia. Psychological
Research, 43, 219-234.
Snowling, M. J. (2000). Dyslexia (2nd Ed.). Oxford, England: Blackwell.
Strommen, L.T., Mates. B. F. (2004). Learning to love reading. Interviews with older
children and teens. Journal of Adolescent and adult literacy, 48: 3,1 18-200
Tallal, P. (1980). Auditory temporal perception, phonics and reading disabilities in Children.
Brain and Language. 9,182–98.
Tallal. P., Miller, S., & Fitch, H. (1993). Neurobiological basis of speech : A case for the
preeminence of temporal processing. In Tallal, A.M. Gulbarda, R. R. Llinas, &
C.Von Euler.(Eds.), Annals of the New York Academy of Sciences, 682, 27-47.
Dyslexia Assessment Profile for Indian Children
129
Terepocki, M., Kruk, R. S., & Willows, D. M. (2002). The incidence and nature of letter
orientation errors in reading disability. Journal of Learning Disabilities, 35, 214-233.
Wagner, R. K., Torgesen, J. K., Rashotte, C. A., Hecht, S. A., Barker, T. A., Burgess, S. R.,
Donahue, J., & Garon, T. (1997). Changing relations between phonological
processing abilities and word-level reading as children develop from beginning to
skilled readers: A 5-year longitudinal study. Developmental Psychology, 33(3), 468-
479.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
130
Clinical Protocol for Assessment of Swallowing in Adults - (CP-ASA)
Meera Priya. C. S. & R. Manjula*
Abstract
The objective of the study was to develop a clinical protocol for the assessment of swallowing
disorders in adults, check for item and content validity and test the sensitivity of protocol developed,
by administering the same on clients with swallowing disorders. The protocol was compiled,
developed and tested for its sensitivity in 3 stages: Stage 1: Compilation and development of the
protocol and test kit. Stage 2: Checking for item & content validity of the protocol. Stage 3: Testing
the sensitivity of protocol developed by administering the same on clients with swallowing disorders.
Stage 4: Finalizing the protocol after incorporating modifications based on the feedback obtained
from stage 2 and 3. The sensitivity of the protocol was tested by administering it on three subjects
with swallowing disorders due to various impairments. Video recording of the assessment session of
the subject were randomly presented to 3 judges who were speech-language pathologists by
profession. The results of the swallowing assessment revealed that S1 presented difficulty in oral
preparatory phase, oral phase with an accompanying delay in pharyngeal phase; S2 presented
difficulty in oral preparatory phase & oral phases of swallow; S3 presented difficulty in oral
preparatory, oral and pharyngeal phases of swallowing. The outcome of these with the clients helped
to improvise some of the items in the protocol by rephrasing, substituting some of the items and
deleting some of the items. The protocol requires to be standardized after administration on many
more clients with swallowing difficulties.
Introduction
Breathing and eating are most basic physiologic functions that define life’s beginning
outside the mother’s womb for newborn infant. Breathing is reflexive and life sustaining but
provides no other intrinsic pleasure. Eating on the other hand is partly instinct and partly a
learned response. It requires ingestion of foods, which in newborn must be provided by an
outward source. Sucking and swallowing requires a complex series of events and
coordination of the neurological, respiratory and gastrointestinal systems. Normal GI function
must occur in digestion of foods to provide nutrients. All of this function occurs within the
framework of developing physical and emotional maturity. The pleasure of eating extends
beyond the feeling of satiety to the pleasure, gained through food ingested. Feeding requires a
well functioning sensorimotor swallowing mechanism, overall adequate health (including
respiratory, gastrointestinal and neurology), appropriate nutrition, central nervous integration
and adequate musculoskeletal tone. The successful emergence of communication depends
heavily on feeding and swallowing.
According to American gastroenterological association, (1999), Dysphagia, or
difficulty in swallowing, is not a disease in itself but a condition that can be brought on by
_____________________________
* Professor of Speech Pathology, All India Institute of Speech and Hearing, Mysore, India
email: [email protected]
Clinical Protocol for Assessment of Swallowing in Adults
131
many different causes because swallowing is a delicate process, easily disturbed. Dysphagia
is defined as a dysfunction or impairment of the stages of swallowing. It is defined by its
clinical features because it is symptom, or collection of symptoms of one or a number of
possible underlying disorders. In patients with dysphagia, various aspects of the anatomy and
neurophysiology in any or all of the stages in swallowing may be impaired, resulting in the
diagnosis of swallowing disorder.
Dysphagia could be caused due to multiple etiological factors Table 1 shows some of
the common causes of dysphagia in its various stages.
Table 1: Common causes of dysphagia
Types Causes
Oro
pharyngeal
dysphagia
Obstructive/
Mechanical
Neuro-
genic
Neuro-
muscular
Junction
Muscular
Upper
Esophageal
Sphincter (UES)
Esophageal Obstructive
Lesion within
the Esophagus
Obstructive
Lesion
Outside the
Esophagus
Neurogenic
Neuro-
muscular
The reported high incidence of dysphagia following stroke and consequent risk
associated emphasizes the need for early identification and evaluation of dysphagia in adult
population. American stroke association (ASA) management guidelines, (2001)
recommended the completion of a comprehensive clinical assessment for any patient
suspected with dysphagia. Identifying patients at risk for developing dysphagia remains a
difficult task. Evaluation of swallowing difficulties / dysphagia in adults is hence a preferred
practice for the professional speech - language pathologists. An extremely wide range of
assessment techniques and procedures are employed by individual therapists and institutions
for most of adult dysphasic clients, more so in India.
Several methods have been proposed for the evaluation of dysphagia. No consensus
currently exists on a standard method of assessment. Regardless of method, an appropriate
and valid assessment tool should be readily accessible for the assessment of adult dysphagic
population and should demonstrate both inter and intra subject reliability. Majority of the
specialized clinics which cater to the assessment and management of clients with dysphagia
across the world base their clinical intervention on the data available through instrumental
analysis, especially for the assessment of pharyngeal and esophageal stages of swallow. The
advantages of instrumental analysis are many, including the fact that specific information
such as site of dysfunction during swallow, data related to duration in bolus transition and the
response to different types of food consistencies are available to the clinician as first hand
information, based on which a management program specific to the client’s needs can be
developed. Many clinics, especially so in India cannot afford to procure instruments which
are often costly and also require professionals such as radiologists and gastroenterologists to
help in interpretation of the data. Hence, most of the clinics adopt the subjective and
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
132
behavioral analysis procedures for clients with dysphagia. This is very relevant to most of the
speech clinics in India where adoption of instrumental procedures for assessment of
swallowing is a far fetched goal at present due to lack of qualified manpower, financial
constraints and cost effectiveness.
Even in those centers which adopt instrumental procedures, the common practice in
assessment is that the oral phase of swallowing is most often assessed through direct
observation as the structures are visible and accessible to the examiner. The abnormalities in
the pharyngeal and esophageal phases of swallowing are often based on instrumental
evaluation. However, the abnormalities in the pharyngeal and esophageal phases of
swallowing can also be assessed based on the inferred observations made through
information available from history and documenting the behavioral correlates of the process
of swallowing as related to the function of pharynx and esophagus.
Matrino, Pron and Diamant, (2003) reported survey data which indicated that 71% of
the respondent dysphagia clinicians (speech language pathologists) performed a complete
evaluation of dysphagia. Conversely only 36% of these clinicians completed an instrumental
swallowing examination. Moreover instrumental swallowing examinations were never
completed in the absence of full clinical examination. Thus these results strongly suggest that
clinical examination is the primary method of swallowing examination among the practicing
clinicians.
The clinical examination of swallowing is designed to provide the clinician with the
following data for use in diagnosis and treatment planning:
Information on the current medical diagnosis and medical history of the patients
swallowing disorder, including the patients awareness of his/her swallowing disorder
and indications of the localizations and nature of the disorder
Patients medical status, including nutritional and respiratory status (presence of
nasogastric feeding tube or gastrostomy and placement of a cuffed or uncuffed
tracheotomy tube)
Patients oral anatomy
Patients respiratory function and its relationship to swallow
Patients labial control, as this may affect keeping food in his/her mouth
Patients lingual control, as it may affect oral manipulation of food and posterior
transit of food through the oral cavity
Patients palatal function. As it may affect entrance of food into his or her nose during
the swallow.
Patients pharyngeal wall contraction as it may affect movement of food through the
pharynx and may cause aspiration after the swallow
Patients laryngeal control, as it may affect airway protection and aspiration during
swallow
Patients general ability to follow directions and monitor and control his or her
behavior
Clinical Protocol for Assessment of Swallowing in Adults
133
Patients reaction to oral sensory stimulation including taste, temperature and texture
and
Patient’s reactions and symptoms during attempts to swallow (Griffin, 1974).
Protocols available for swallowing evaluation: Many protocols which are available and
reported are ones which are developed in the western countries. Assessment of swallowing in
children and adults assumes different dimensions. This study aims to develop a protocol for
the assessment of swallowing disorders in adult population in India.
Any protocol / test / scale developed for the assessment of swallowing difficulties is
required to meet certain features to account for its credibility as a clinical tool to aid in the
evaluation of swallowing difficulties as well as its further application in the management of
the client. These factors as listed by Murry (2001) include the following:
Appropriate definition of the anatomical structure that is involved in swallowing
dysfunction
Ability to detect and quantify aspiration
Ability to detect etiology for the swallowing dysfunction
Comprehensively assess all the phases of swallowing
Ability to test for oral reflexes
Type of test (screening and diagnostic)
Critical evaluation of the western protocols / scales:
Table 2 shows the critical evaluation of the western protocols / tests / scales.
Defines
anatomy
Detects
aspiration
Quantifies
aspiration
Detects
etiology in
swallowing
dysfunction
Comprehen-
sive assess-
ment of all
phases of
swallowing
Quantifies
reflexes
Screening/
Diagnostic
BDST - + - - - - Screening
MASA - + - + - partial Diagnostic
SAFE - - - + - partial Diagnostic
TORBSST - + - - - - Screening
EDAT - - - Screening
SWAL-
QOL
- - - - - - Patients
self
assessment
SWAL-
CARE - - - - - - Patients
self
assessment
MDADI - - - - - - Diagnostic
RSI - - - - - - Self
assessment
Critical evaluation of the western protocols / scales used for assessment of
swallowing in adults as listed in Table 2 reveals the following:
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
134
All the scales do not assess all the phases of swallowing at a time.
Majority of the parameters addressed in these scales advocate use of stimuli, eg food
articles and their consistencies which are not culture friendly to Indian context
The scales do not promote a cause – effect relationship in terms of associating the
dysfunction with the structural deficit.
Although the BDST and MASA scales allow for detection of aspiration, there is no
scope for quantifying the extent of aspiration. The SAFE scale does not allow for
detection or quantification of aspiration.
There is no scope for inferring the specific etiology in the different phases of swallow in
BDST.
Oro pharyngeal reflexes which play a major role in the swallowing function is not
accounted for in BDST, whereas it is only partially accounted for in MASA and SAFE.
BDST is only a screening tool unlike MASA and SAFE which are diagnostic tools.
In comparison to the assessment tools available in the western countries, there are no
comprehensive protocols available to suit the Indian population and hence this is attempted in
the study.
Need for a assessment protocol in India
There are no formal protocols / tests / scales developed in the Indian context. Most
of the centers which cater to the clientele with swallowing disorders have developed their
own indigenous questionnaires and proformae to assess the swallowing disorders in
children as well as adults. However, it is known that the anatomical and physiological
differences in the swallowing mechanism call for the use of different scales in children as
compared to adults. Most of the clinical tools developed in India are not comprehensive,
specifically in terms of testing with direct stimuli (different consistencies of solid, semi
solid and liquid items and the quantity of the same) and indirect stimuli (assessment of
posture, oral sensory testing, assessment of motor functions etc). Given the multicultural
issues in India, selection of direct stimuli requires to be done with care if one attempts to
develop a test kit for the assessment of swallowing.
Keeping these in mind, the present study aims to develop a clinical protocol for
comprehensive assessment of all the three phases of swallowing, that is, oral, pharyngeal
and esophageal phase, based on behavioral paradigm. Assessment of oral and pharyngeal
stages will be aimed at using direct observation procedures, but the esophageal stage will be
assessed based on history and indirect observation procedures.
Aims of the study
To develop a clinical protocol for the assessment of swallowing disorders in adults.
To check for item and content validity of the protocol developed
To test the sensitivity of protocol developed, by administering the same on clients with
swallowing disorders.
Clinical Protocol for Assessment of Swallowing in Adults
135
Method
The protocol was compiled, developed and tested for its sensitivity in 3 stages:
Stage 1: Compilation and development of the protocol and test kit
Stage 2: Checking for item & content validity of the protocol
Stage 3: Testing the sensitivity of protocol developed by administering the same on clients
with swallowing disorders
Stage 4: Finalizing the protocol after incorporating modifications based on the feedback
obtained from stage 2 and 3.
Stage 1: Compilation and development of the protocol and test kit
A comprehensive review of different assessment scales/tools/literature and web based
search was employed. Items that contributed to the assessment of a gamut of swallowing
disorders in adults were pooled in order to develop a comprehensive protocol addressing
various aspects of the disorder. The items were then verified one by one to look for its
relevance to the adult population as well as its application in Indian context. The items which
were not relevant to the Indian context were modified and / or replaced as per the need; the
material was further classified under different domains, sections and subsections. This
exercise led to the development of a protocol comprising of four main domains, with sections
and subsections within them.
The structure of the protocol included the following:
Part A
A) Demographic details of the client
B) Client history
C) Family and medical history
Part B
Section Title of the Section & Subsections Number
of items
Scoring
Section A Assessment of Posture 14 4 point rating scale
3 = Within functional
limits
2 = Mild impairment
1 = Moderate impairment
0 = Severe impairment
Section B
Assessment of respiratory abilities
At rest
During speech
Coordination of
Respiration and swallowing
Assessment of sensory abilities
Taste
Temperature
Oral sensation
1
1
3
5
2
10
4 point rating scale
3 = Within functional
limits
2 = Mild impairment
1 = Moderate impairment
0 = Severe impairment
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
136
Assessment of cognitive status
General status
Specific dysfunction
11
9
Section C
Assessment of Cranial nerve
functions and Oral reflexes
Assessment of cranial nerve
functions
Assessment of oral reflexes
4
4
4 point rating scale
3 = Within functional
limits
2 = Mild impairment
1 = Moderate impairment
0 = Severe impairment
2 point rating scale
1 = Normal reflex for the
age
0 = Primitive/immature/ no
response/exaggerated
response to the stimulus
Section D
Physical Examination of the oral
mechanism
Lips
Tongue
Velum
Teeth
Jaw
5
7
9
2
4
4 point rating scale
3 = Within functional
limits
2 = Mild impairment
1 = Moderate impairment
0 = Severe impairment
Section E
Assessment of swallow in different
stages
Oral preparatory stage
of swallow
Oral stage of swallow
Pharyngeal stage of
swallow
Esophageal stage of
swallow
25
14
18
10
4 point rating scale
3 = Within functional
limits
2 = Mild impairment
1 = Moderate impairment
0 = Severe impairment
The protocol is supplemented with 2 Appendices to aid the clinician in preparation
of material required for the administration of food as well as a ready reckoner. They
include:
Scoring
A 4 point rating scale was used in all the sections except for the subsection on oral
reflexes which utilized a 2 point rating scale. The scales depict the descending order of
normal behavior of swallow in various sections. Under each section and subsection, the
specific behavior for which the rating scales need to be given is described.
Clinical Protocol for Assessment of Swallowing in Adults
137
Apart from the quantitative assessment using the rating scales, scope for qualitative
description, noting the behaviors, comments and other remarks is made available under the
column of remarks in every section and subsection.
Stage 2: Checking for item and content validity of the protocol
Once the protocol was developed, it was verified by four speech language
pathologists who had working and clinical knowledge in dysphagia management. The items
in the protocol was verified by these specialists individually and independent of each other.
They were asked to rate each item within the sections and subsections of Part A and B on a
4 point rating scale to identify whether the items were worthy of inclusion (or not) in a
clinical assessment of dysphagia. The rating scale was as follows:
1 = is poor and needs to be deleted or substituted
2 = needs major alteration
3 = needs minor alteration
4 = is adequate
Those items which were rated as 1, 2 and 3 were subject to omission or modification
as per the suggestion and the same was incorporated in the protocol.
Stage 3: Testing the sensitivity of protocol developed by administering the same on
clients with swallowing disorders
The sensitivity of the test protocol was tested by administering it on three subjects
with swallowing disorders due to various impairments. During the physical testing, it was
ensured that the client is kept in most comfortable position possible in a well-lit room with
no distracters. The session was video recorded after desensitizing the clients, in order to
facilitate observation and verification of the responses at a later time to ensure that the
behavior was rated appropriately.
The testing session was spread over 2-3 sittings depending on the clients cooperation
and fatigue level. Written consent was obtained from the clients and the purpose of the test
was also explained. Suitable instructions appropriate for each of the section and subsection
was given. The responses of the clients were recorded on the score sheets of the CP-ASA.
Whenever possible, family members were allowed to be present during the assessment and
they were also informed of the process, results and recommendations. The investigator noted
in detail the performance of each of the client on the protocol during the live testing of the
clients. This was further verified with the video recordings in order to confirm, observe in
depth, reflect and verify whether the assessment proceeded in the right way.
The results of the swallowing assessment revealed that S1 presented difficulty in oral
preparatory phase, oral phase with an accompanying delay in pharyngeal phase; S2 presented
difficulty in oral preparatory phase & oral phases of swallow; S3 presented difficulty in oral
preparatory, oral and pharyngeal phases of swallowing. Based on the observation and noting
made, suitable modifications were made in the instructions, scoring, definition of a behavior
during rating and others as was indicated during the observations. The outcome of these with
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
138
the clients helped to improvise some of the items the protocol by rephrasing, substituting
some of the items and deleting some of the items.
Stage 4: Finalizing the protocol after incorporating modifications based on the
feedback obtained from stage 2 and 3.
The protocol was subjected to modification based on the suggestions/outcome
obtained from stages 2 and 3.
Summary and Conclusions
This was a preliminary attempt made to aid in the diagnosis and therapeutic
intervention for adults with dysphagia. The protocol provides a qualitative as well
quantitative (by the use of rating scales) understanding of the elements of dysphagia in the
client. The protocol partially fulfills the need for a clinical tool required by clinicians which
can be administered based on subjective assessment of the client in the absence of
sophisticated objective tools.
The administration of the protocol takes approximately 30-45 minutes. The protocol
includes 4 main domains as follows:
Domain I: Demographic details of the client
Domain II: Client history
Domain III: Family and Medical history
Domain IV: Assessment section.
The protocol requires to be standardized after administration on many more clients
with swallowing difficulties in order to:
Establish the cut off criteria (for quantitative scoring of errors) for the diagnosis of
dysphagia.
Elaborate on specific distinguishing characteristics of different stages of swallow
dysphagia and identify clearly based on the behavioral cues, the dysfunction in different
phases of swallowing disorder.
References
American Speech-Language-Hearing Association. (1999). Definition of Swallowing
Disorder. 35. (Suppl. 10), 40-41.
American stroke Association (ASA). (2001). Comprehensive clinical assessment for any
patient suspected with dysphagia. Ad hoc committee on dysphagia report, 32 (2), 53-
57.
Clinical Protocol for Assessment of Swallowing in Adults
139
Breumelhof, R., & Smout, A. J. (1991). The symptom sensitivity index: A valuable additional
parameter in 24-hour esophageal pH recording. American Journal of
Gastroenterology, 86, 160-4.
Chen, A. Y. (1996). MD Anderson Dysphagia Inventory. In Adaptation and validation of the
Italian MD Anderson dysphagia inventory (MDADI). Revue de laryngologie -
otologie - rhinologie .(2008). 129, 2, 97-100.
Depippo, K. L., Holas, M. A., & Reding, M. J. (1994). The Bruke dysphagia screening test:
validation of its use in patients with stroke. Archives of psychiatric medical
rehabilitation. 75, 1284-1286.
Griffin, K. (1974). Clinical examination of swallowing. Archives of Physical Medicine and
Rehabilitation. 55 (3), 467-570.
Kipping, P., & Ross, D. (2003). Swallowing ability and functional evaluation (SAFE), Pro-
Ed. Austin.
Martino, R., Pron, G. & Diamant, N. (2000). Screening for oropharyngeal dysphagia in
stroke: Insufficient evidence for guidelines. Dysphagia, 15, 19-3
Mann, G. D. (2002). The Mann’s assessment of swallowing ability, Austin, TX: Pro-ed.
Matrino, R., Damant, N. (2007). Toronto bed side screening test. Austin, TX: Pro-ed.
Mc Horney., Colleen, A., Harris, M. (2006). Validity and sensitivity of SWAL-QOL and AL-
CARE for patients with swallowing disorder. Dysphagia, Volume 21, No. 3, pp. 141-
148(8)).
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
140
Comprehensive Language Assessment Tool for Children (3-6 years)
Navitha U. & K. C. Shyamala*
Abstract
Normative approaches to assessment have historically evolved in the context of the need to
screen, identify, isolate and diagnose low achieving children from others. The valid assessment of
language is somewhat of an enigma for clinicians. Eventhough there are multitude of resources
available still assessment of language continues to be one of the most challenging. The present study
was designed to construct a tool for assessment of language and cognitive abilities in typically
developing preschool children in the age range of 3 to 6 years. The study was carried out in three
phases. The phase I incorporated the test construction, phase II consisted of pilot study where initial
filed try out of all the selected items was administered to 25 parents of typically developing preschool
children in the age range of 3- 6 years. Thus, the test device developed based on the performance of
the typically developing preschool children in the pilot study had three main sections, i.e., reception,
expression and cognition. The Phase III included the standardization procedure where the test
developed based on pilot study was administered on 150 typically developing preschool children for
obtaining the normative data. Hence, there were 5 groups ( 3 to 3.6 years, 3.6 to 4 years, 4 to 4.6
years, 4.6 to 5 years and 5 to 6 years ) consisting of 30 children in each group. The data was obtained
by an interview approach. Both boys and girls were considered for the study. The data for each group
was statistically analysed. Mean and standard deviation were obtained for each group. A general
significant correlation was obtained between the development of language and cognition. The
performances of males and females was compared which showed mixed results. After the
administration of the test, the child’s score had to be compared with the normatives.
Introduction
Communication is an essential aspect of human beings. Language is unique to man
especially in the form of speech. Language starts developing from birth through several years
of life. The speech of the child changes with different stages of language acquisition. Child‟s
expression through speech serves as one of the important avenue for language testing. The
development of speech and language in particular is a dynamic constructive process (Thelen
2005).
Cognition in simple words can be described as understanding of our experiences
through mental processes such as perception, recall, and reasoning, and it provides an
important element for the development of language. Piagetion theory has emphasized the
importance of the developing cognitive abilities of the child (Piaget, 1962). In Piaget‟s view
language is dependent on and shaped by underlying cognitive structures, and it reflects the
thought processes made possible by those structures at different stages of development.
Vygotsky (1978) claimed that linguistic growths are facilitated by cognitive development. A
child‟s development of language is dependent upon his knowledge of concepts, which in turn
is dependent upon his level of cognitive development. As a result, there is a positive
__________________________________________ Professor of Language Pathology, All India Institute of Speech and Hearing, Mysore, India
e-mail: [email protected]
Comprehensive Language Assessment Tools for Children
141
correlation between the degree of retardation and the level of language development in the
retarded child (Karlin & Strazzula, 1952). Chomsky (1957) described language on the
psychological perspective of a language user‟s ability to produce and comprehends language.
By the nature of cognitive hypothesis, language development is rooted in early cognitive
development, prior to the appearance of the first word. A particular level of cognitive
achievement is necessary before language can be used expressively.
Perception and the subsequent development of an understanding of semantic
relationships are the result of human cognitive functioning (Spelke & Tsivkin, 2001).
Language development is the product of strategies and process of general cognitive
development, although not a direct manifestation of it. Parents could identify their preschool
children‟s language skills. The literature has reported of high correlation between the
language ages obtained from a parental interview and those obtained from the tests
administered directly on the preschool children. An interview method of obtaining
information could be used by speech-language pathologists with confidence when
information could not be obtained by directly testing the child due to physical, emotional, or
intellectual disabilities. Obtaining information from an informant could be an easier, quicker
way of screening large numbers of children.
Need for the study
A number of tests have been developed to assess the language skills of preschool and
school going children in Indian context but it is very detailed and more time consuming.
Many studies available in India have focused on language acquisition (Roopa, 1981; Sudha,
1981; Vijayalakshmi, 1981) etc.
There are few assessments tools available for children with higher age in Indian
context, but they are very restricted and more time consuming. So there is a need for quick
and comprehensive standardized test for different age group of school going children. There
is a lack of tests available for assessing the acquisition of language and cognitive abilities of
the growing children especially as screening tools in the Indian context to predict the age of
their performance particularly in for difficult to test population.
Aim of the study
The main aim of the study is to develop assessment tool for children in the age range
of 3- 12years for testing their language and cognitive ability to standardize the assessment
tool on Indian population and to obtain normative data for the same.
Method
Participants
The participants in the present study were the parents or the caretakers of typically
developing preschool going children in the age range of 3-6 years. The study was carried out
in three different phases. Phase I consisted of test construction, Phase II included the pilot
study and Phase III consisted of the standardization procedure. Two groups of participants
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
142
were involved in the present study. The first group included a total of 25 participants who
served as subjects for the pilot study. The second group included a total of 150 participants,
who served as subjects for the standardization of the test material with 30 subjects per age
group. Hence, a total of 175 parents whose children were in the age range of 3-6 years were
considered. Language was not considered as criteria for the selection of the subjects. Based
on the report obtained from the school teachers, the children with an average or above
average performance were considered. The inclusion criteria for the selection of the
participants also included children with no history of any known organic and sensory deficits.
Both males and females were considered.
Phase I: Construction of the test material
Selection of test items
The assessment tool was designed to elicit systematic information based on the skills
present in typically developing preschool children. The construction started with the
formation of a comprehensive item pool of activities related to children below the age of
seven years. An indulgent list of approximately ten to fifteen items in each section for each
age group was selected from various sources used for the assessment of children. Majority of
the items selected were obtained from the checklist developed by Venkatesan (2004) entitled
“Activity Checklist for Preschool Children with Developmental Disability”. During initial
formation of the item pool, care was taken to see that the test items were placed in a
hierarchical order of increasing performance difficulty according to the chronological ages of
the children. The easier and lower chronologically aged test items were placed at the
beginning of the checklist, and the more difficult and higher- aged items were placed towards
the end of the checklist.
The specific domains included under Comprehensive Language Assessment Tool for
children were reception, expression and cognition. It was seen that the domains included in
the test were both exhaustive and mutually exclusive. In order to check for the content
validity of the question selection, the selected questions were given to five speech language
pathologists who had a clinical experience of more than five years. They were asked to rate
as R – if the question comes under reception, E – if under expression, and C – if under
cognition section. This was to eliminate the repetition of test items within a domain as well as
between domains. Thus, the modified questions consisted of fifty items in reception, fifty in
expression and fifty five in cognition section.
Phase II: Pilot study
The initial filed trial of selected items was carried out with 25 parents of the typically
developing preschool children based on interview session and direct observation. A minimum
of five subjects were considered in each age group which consisted of both males and
females. The entire item from each section such as reception, expression and cognition were
administered for the 25 parents of the children. An appropriate example was provided for
each age item and was explained in a simpler form in their native language to further clarify
the question for the parent. Therefore, based on the information obtained from the parents
regarding the performance of the children, the questions in which the respective age group
Comprehensive Language Assessment Tools for Children
143
children performed well were considered for the development of the test in tune with the
suggestions of a psychologist and speech language pathologists. Thus, after addition, deletion
and modification of the test item appropriate for each age group, the test finally contained six
items in each section of Reception, Expression and Cognition. The ambiguous words and
phrases were simplified and thus a final modified test material was made with a total of 90
items, six items in each section under each age group.
Phase III: Standardization for normative data
It is essential to obtain normative data on the test constructed. Standardization can be
explained with respect to the following:
a) The population for standardization: The subject population included 150 parents of
normal children between 3 years – 6 years. The native language of the child was not
considered in grouping as the test was meant for all language groups. Only those
parents whose children had no history of any complications, either pre-natal or peri-
natal and free from any known functional or organic disabilities were taken for the
study.
b) Age groups: Totally five age groups were considered. The first four age groups
covered a time span of six months each, since this is considered as the most crucial
period of language and cognitive development ( 3-3.6, 3.6-4, 4-4.6 & 4.6-5 years ).
The last group covered an age range of one year (5-6 years).
c) Procedure: The parents of the children or a care taker (familiar with the child‟s
behavior) were interviewed to obtain the data. Parental meetings were carried out for
the same by arranging parental meetings in the school. The interview was carried out in
a closed and noise free environment free from distractors. They were told about the
purpose of the study in brief and about the kind of information that was required
regarding their children. The parents were distributed the scoring sheet and were
initially asked to fill the details of their child. The chronological age of the child in
terms of both years and months was noted along with the date of birth. All the ethical
considerations were met.
The instruction for each task was given differently based upon the type of ability
tested such as reception, expression and cognition. Items from the corresponding age groups
was asked first and their performance on other items above their age group and below was
also evaluated under each section. The parents were provided with appropriate examples
when they did not understand the purpose of the question. Responses of the child were
checked for two consecutive age groups above and below that of the child‟s chronological
age. It was continued in the lower age groups till 3 „+‟ was recorded and in the higher ones
till 3 „_‟ was obtained within the age group.
Scoring and Reliability: The responses were recorded in the response sheet. The
informants were instructed to mark the response on each item as a „plus‟ (+) when they feel
that the behavior of the child was established, a „minus‟ (-) when it had not yet emerged and a
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
144
„plus-minus (+/-) whenever the given language behaviors were only partially exhibited or
inconsistently noted (Remarks were also noted).
The Reliability of test scores was examined, in part, using test – retest reliability. In
this case, the consistency of performance as a function of test administration was
investigated. The test – retest reliability was conducted for 10% of the subjects in the age
range of 3-6 years selected randomly from each age group and tested using the same tool
within a span of 3 days.
Analysis of data: The data obtained from the above procedures was subjected to quantitative
analysis using Statistical Package for the Social Sciences (SPSS) (16.0 version) software for
both the pilot study and standardization population.
Results and Discussion
The raw scores were obtained by scoring the performance of every child on the test
items. A credit of one point for a response (+) and zero for no response (-) was given and the
total scores for each child on the test within each dimension (Reception, Expression and
Cognition) were obtained. The data, subjected to statistical analysis, are discussed for both
phase II (pilot study) and phase III (standardization process).
Mean scores and standard deviations
The raw scores obtained were used to determine the mean and standard deviations.
The scores achieved by each child through the above scoring method were cumulated to
calculate a total for each age group along each dimension. The total score for the boys and
girls of different groups were also found separately. Thus, the mean and standard deviation
scores for both Phase II and III in each group were calculated from the total score.
Phase I: Pilot study
Table 1: Mean and Standard deviation for the pilot group in each section of Reception,
Expression and Cognition for different age groups.
Age
Reception Expression Cognition
Mean Std.
Deviation Mean
Std.
Deviation Mean
Std.
Deviation
3-3.6 yrs
3.6-4 yrs
4-4.6 yrs
4.6-5 yrs
5 - 6 yrs
8.00 0.71 8.40 0.89 10.20 0.84
16.40 1.14 16.00 1.22 15.80 0.84
25.00 0.71 23.20 0.84 27.80 4.44
30.20 1.30 29.20 1.30 34.40 0.55
41.60 1.14 41.20 1.64 45.80 0.84
The above data represented in table 1, gives the overall mean and the standard
deviation scores. It indicated that there was not much difference present between reception,
expression and cognition within each group, but cognition was found to be better when
compared with reception and expression. The mean and standard deviation was greater for
the higher age group when compared with the lower age group. The difference in the skills
Comprehensive Language Assessment Tools for Children
145
increased as the age progressed which could be interpreted from the increase in mean score
and standard deviation with increase in age group.
Kruskal Wallis test was carried out to compare the responses across the five groups
for the items in the three sections (reception, expression and cognition). It revealed that there
was a significant difference present between age groups in reception, expression and
cognition, reception at (2(4) = 23.175, p<0.001), expression at (
2(4) = 23.166, p<0.001) and
cognition at (2(4) = 23.166, p<0.001). From Mann Whitney U test, significant differences
were found to exist across all the age groups for all the three sections at 0.001 level of
significance. Friedman test was done to compare the scores of reception, expression and
cognition within age groups in the pilot study and the following results were achieved. If
significance was found to exist, then Wilcoxon signed rank test was done to compare items in
each section of each age group.
a) Within age group 1 (3-3.6 years)
It was ascertained that significant difference existed between reception, expression
and cognition at (2
(2) =6.000, p<0.05). Wilcoxon test revealed that there was no significant
difference found between reception and expression, expression and cognition but there was a
significant difference present between cognition and reception at 5% level of significance.
The present research is in agreement with the study done by Ogura (1991) who found a
significant relationship between young normal children‟s play and language comprehension,
but not expressive language.
b) Within age group 2 (3.6-4 years)
Friedman test showed a significant difference in three section such as reception,
expression and cognition within age group 2 at (2
(2) =7.600, p<0.05) and thus Wilcoxon
signed rank test was carried out. This also showed that there was a significant difference
present between cognition and expression (z=2.03, p<.0.05) and cognition and reception
(z=2.02, p<.0.05) but no significant difference were found between reception and expression.
Mc Cathren, Warren and Yoder (1996) reported that symbolic play competences are seen as
being among the most influential predictors of early language development during the
preschool year. Therefore parallel developments in play and language could be explained as
deriving from a common underlying capacity for cognitive representation.
c) Within age group 3 (4-4.6 years)
For the group of children who participated in the age range of 4-4.6 years, Friedman
test revealed no significant difference across the three sections such as reception, expression
and cognition at 5 % level of significance. Cognitive skills and language abilities are
associated; they develop in parallel fashion. New and increased cognitive ability may enable
a child to function differently, but it does not cause language change. Rather, cognition and
language are strongly related with underlying factors. Attainment of a skill in either area is
reflected in the other although no direct overall relationship exists, specific relationships are
evident during development (Gopnik & Meltzoff, 1986).
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
146
d) Within age group 4 (4.6-5 years)
It was determined from the Friedman test that there was significant difference
prevailing between all the section at (2
(2) =7.111, p<0.05). Hence pair wise analysis using
Wilcoxon signed ranks test was used to find the significant pairs between items. Results
demonstrated that there was significance differences existing for cognition and expression
and for reception and expression at 5% level of significance but no such significant difference
were found for reception and cognition. The present study is in accordance with the research
done by Chapman (2000) where the authors studied the comprehension and production of
reversible active sentences in children. He reported that the children performed better in the
production task than in the comprehension task, leading to conclusion that production
precedes comprehension.
e) Within age group 5 (5-6) years
The statistical analysis using Friedman test showed that there was no significant
difference found between all the three sections at (2 (2)=0.73, p>0.05). Thus the present
study supports one such research where it was found that language-play correlations were
strongest in early language development, and that the domains did not develop in parallel as
the child matured (Kennedy, Sheridan Radlinski & Beeghly, 1991; Ogura 1991). Ogura
(1991) also highlight that cognitive domains become increasingly modular with development
and may not be strictly „innate‟ from birth, which is in line with the neuroconstructivist
approach to development. Thus, by 5-6 years the children develop most skills when
compared to younger children.
Figure 1: Graphical representation shows the mean percentage scores of Reception,
Expression and Cognition across different level of age groups (3-6 years).
From the graph, it‟s observed that there is an increasing pattern in all the sections
(Reception, Expression and Cognition). Thus, it can be clearly interpreted that with an
increase in age, the performance on language and cognitive abilities are also found to be
increased. The results for the pilot study across the age groups and their performance
revealed that the differences in performances were observed across the groups with increase
in age of the children. There was a clear pattern of hierarchy noticed. The items with more
difficulty were found to be performed better by higher age group children when compared
with lower age group children. Result based on the statistical analysis highlights that there
was a well defined difference found on the performance of cognitive items when compared
with reception and expression.
Comprehensive Language Assessment Tools for Children
147
According to the above results, receptive as well as expressive language skills and
cognitive skills found to be improved across the years; seemingly there are highly significant
differences in the skills acquired between 3-6 years of age. This could be attributed to the fact
that language acts as another form of behavior which is acquired as a response to the stimuli
in the environment and then it is learnt. Parental stimulation and environment exposure are
seemingly the important factors for the increased linguistic development for the present
generation. Vijayalakshmi (1981), Sudha (1981) and Santhi (2008) reported “significantly
better performance in the older age group compared to the younger age group particularly for
the expression of children in the age range of 1 to 5 years and 2 to 5 years, which supports the
present study. This could be attributed to the fact that with increase in age there was
advancement in neuromuscular maturity and motor skills as well as the linguistic and
cognitive abilities. Many researchers have also found a correlation between various aspects of
language with play. Ogura (1991) found a significant relationship between young normal
children‟s play and language comprehension, but not expressive language. Thus indirectly
supports the fact that plays acts as the basis for cognitive and linguistic development
especially in preschool children. Thus, the above mentioned studies are in accordance with
the present findings thereby highlighting the importance of play on cognition and language
acquisition during the preoperational stage in the preschoolers.
Phase III: Standardization
a) Age group 1(3-3.6 years)
Table 2: Mean and Standard deviation for the age group 3-3.6 years according to gender for
three sections (Reception, Expression and Cognition).
Subject
Number of
subjects Mean Std. Deviation
Reception
Males 13 5.00 0.913
Females 17 4.88 1.27
Expression
Males 13 4.31 1.25
Females 17 5.00 1.37
Cognition
Males 13 4.69 1.32
Females 17 5.29 1.16
The above table 2 gives the mean and standard deviation for children in the age range
of 3- 3.6 years, which included 13 males and 17 females, obtained by quantitative analysis of
the raw scores. On comparison of the performance of the male and female children across the
items, it was found that there was a wide difference between males and females children
within the age group and across the items too. The present findings contrasted with the study
done by Griffin & Norris (1967) which showed no significant differences in speech of boys
and girls. `The independent t- test was carried out to find differences between males and
females on three sections such as reception, expression and cognition. Results revealed that
there was no significant difference between the male and female groups. Further, Pearson‟s
Product Moment correlation was applied on the data to study the relationship between
reception, expression and cognition scores. It showed was a positive correlation between
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
148
expression and cognition (r=0.810, p<0.01), but reception was not significantly correlated
with cognition and expression. Repeated measure ANOVA was done to compare the
performance of children in the age range of 3-3.6 years across the three different sections
such as reception, expression and cognition. However, the result obtained in the present study
contrasts with the result obtained by Vijayalakshmi (1981) which revealed performance of
females to be better than the males.
Figure 2: Graphical representation of the mean scores for the age range 3-3.6 years for
Reception, Expression and Cognition between males and females.
From the graph i.e. figure: 2, it can be construed that the performance of males and
females is almost same for reception whereas expression and cognition illustrates difference,
thought not statistically significant. Researchers have found that children demonstrate certain
cognitive abilities at the approximate time as corresponding language behavior emerges
(Clark, 1980). However, Clark (1980) opines that correlation is not the same as causation;
because if two events occur at about the same time does not necessarily mean that one caused
the other. The present study concurs with the same.
b) Age group 2 (3.6- 4) years
The computation of the scores resulted in the below table 3 depicts the mean and
standard deviation for children in the age range of 3.6-4 years for both males and females (15
males and 15 females). The result revealed that there was no significant difference found
between male and female children within the age group as observed from the mean and
standard deviation scores.
Table 3: Mean and Standard deviation for the age group 3.6- 4 years according to gender for
three sections (Reception, Expression and Cognition).
Subject Number of
subjects Mean Std. Deviation
Reception
Males 15 4.53 1.46
Females 15 4.87 1.36
Expression
Males 15 5.07 1.16
Females 15 5.27 0.96
Cognition
Males 15 4.93 1.44
Females 15 4.67 1.18
The Independent t- test which was done to compare the performance of male and female
children also showed no significant difference between them. Further, Pearson‟s product
moment correlation exemplified that there was a significant correlation existing between
Comprehensive Language Assessment Tools for Children
149
reception, expression and cognition leading to language development in a preschooler.
Repeated measure ANOVA was done to compare the performance of children in the age
range of 3.6-4 years across the three different sections such as reception, expression and
cognition. No significant difference was present across the sections. The present findings are
in support with the study done by Gopnik and Meltzoff (1997), where the author reported a
relationship between cognitive mechanisms and language through analysis of linguistic
components.
Figure3: Graphical representation of the mean scores for 3.6-4 years for reception, expression
and cognition between males and females.
From the above graph it can be interpreted that the performance of males is better
when compared to females though not statistically significant. Additionally, the performance
expression was considerably better followed by cognition and then reception. Recent
advances in cognitive psychology, neuroscience and linguistics support an embodied view of
cognition, i.e. the fact that cognitive functions (perception, categorization, reasoning and
language) are strictly interwined with sensorimotor and emotional processes (Ogura,1991).
This is particularly evident in recent studies on the grounding of language in action and
perception (Casby, 2003). This indirectly supports the relationship between cognition and
language, as observed in the present study.
A) Age group 3 (4-4.6 years)
The below table indicates the mean and standard deviation for children in the age
range of 4-4.6 years which included 12 males and 18 females for all the three sections i.e.
reception, expression and cognition. The score obtained by descriptive statistical procedure
reveals that there were no evident differences noticed between from the scores.
Table 4: Mean and Standard deviation scores for the age group 4-4.6 years for both male and
female children.
Subject
Number of
subjects Mean Std. Deviation
Reception
Males 12 4.50 1.09
Females 18 4.72 1.02
Expression
Males 12 4.92 1.08
Females 18 5.00 1.14
Cognition
Males 12 5.00 1.04
Females 18 4.78 1.22
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
150
There was no significant difference between the performance of male and female
children in the age range of 4-4.6 years as indicated by Independent t- test. Pearson‟s Product
Moment correlation was applied on the data to study the relationship between comprehension
and production and cognition scores. It showed no correlation between reception, expression
and cognitive. Repeated measure ANOVA was done to compare the performance of children
in the age range of 4-4.6 years across the three different sections such as reception,
expression and cognition. No significant difference was demonstrated across the sections in
the age range of 4-4.6 years.
Figure 4: Graphical representation of the mean scores for Reception, Expression and
Cognition between males and females in the age range of 4-4.6 years.
It can be inferred from the above graph that there was no significant difference
observed between the performance of males and females. Reception , Expression and
Cognition also do not demostrate a wide variation across them.The present study concurs
with the study done by Sudha (1981) where it was reported that the comprehension and
expression of children in the age range of 2 to 5 years showed no significant difference
between males and females in their performance. Slobin (1973) recognized the important
roles that the language being learned may play both in at the time of development of the
linguistic structures in that language and in the development of cognitive concepts that run
parallel to those structures, which support the present study where reception, expression and
cognition are found to be acquired paralelly.
B) Age group 4 (4.6-5 years)
Table 5: Mean and Standard deviation scores for Reception, Expression and Cognition of
males and females in the age group of 4.6-5 years.
Subject
Number of
subjects Mean Std. Deviation
Reception Males 17 4.82 1.24
Females 13 4.77 0.93
Expression Males 17 4.88 1.27
Females 13 5.31 1.03
Cognition Males 17 4.82 1.55
Females 13 4.54 0.78
Comprehensive Language Assessment Tools for Children
151
The mean and standard deviation scores obtained by the male and female preschool
children in the age range of 4.6- 5 years for the three different sections of reception,
expression and cognition are represented in table 5. It can be understood from the table that
there are not much variation in the performance between the two genders in the same section
neither even across the sections such as reception, expression and cognition. Independent„t‟
test showed no significant difference between males and females. Pearson‟s product moment
correlation which was applied showed revealed that reception correlates more with cognition
and expression than, expression with cognition. Repeated measure ANOVA showed that
there was no significant difference found in the age range of 3-3.6 years across the three
different sections of reception, expression and cognition.
Figure 5: Graphical representation of the mean scores for Reception, Expression and
Cognition between males and females for 4.6-5 years.
It can be observed from the graph that reception, expression and cognition do not show
obvious differences and there was no evident variation found in the performance of males
and females except for expression. Thus, the present study finds support from the study done
by Jerperon (1922) who reported that girls talk earlier than boys; they articulate better and
acquire a more extensive vocabulary than boys of comparable age.
E) Age group 5 (5-6 years)
The mean and standard deviation for children in the age range of 5-6 years which
included 12 males and 18 females are shown in the table 6 below. There were not much
difference between males and females across the three sections.
Table 6: Mean and Standard deviation scores of males and females for Reception, Expression
and Cognition (5-6 years).
Subject N Mean Std. Deviation
Reception Males 12 4.67 1.37
Females 18 5.06 1.06
Expression Males 12 4.83 1.11
Females 18 5.00 1.03
Cognition Males 12 4.58 1.08
Females 18 5.00 1.19
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
152
Mc. Neil (1970) and Brown (1973) has indicated that the period from 18 months to 4
years is the most active period of language acquisition and distinct level in language
development can be made out. After the age of five years the rate of acquisition decreases
markedly and differences between adult and child specially are not so obvious. The
immaturity of a child‟s language after the age of five years is revealed only if a depth analysis
of the language structure is done (Chomsky, 1969). This can be correlated with the above
fact, demonstrated by the present study. However, it would be premature to draw any
conclusion since the age range considered was wide.
Independent„t‟ test which was carried out to compare the male and female
performances in the age range of 5- 6 years showed no significant difference between males
and females. Pearson‟s product moment correlation which was performed revealed that there
exists a close association between reception, expression and cognition. Repeated measure
ANOVA illustrates that there was no significant difference found in the age range of 5-6
years across the three different tasks such as reception, expression and cognition. The results
of the present study are in agreement with the research done by Piaget ,1962; Ogura, 1991;
McCune, 1995; Lyytenin, Laakso, 1997 & Tomasello et al., 1999). These studies also
demonstrated a strong relationship between play, cognitive development, and early
communication and language which tells the importance of cognition and language
development in preschoolers.
Figure 6: Graphical Representation of the mean scores for 5-6 years for Reception,
Expression and Cognition between males and females.
Bliss, Allen & Wrasse (1977) in their study found that males and females performed
alike, holding age constant, but the males require more prompting and structuring before they
produced the correct response. However, in the present study from the above graph, it is
understood that reception, expression and cognition the 5- 6 years old group of children. In
other words, the rate of emergence of various linguistic expressions in child language more or
less directly indexes the degree of their conceptual complexity (Huttenlocher, Smiley, &
Charney, 1983). According to some researchers, linguistic effects on cognition are more
likely to be found in domains removed from perception, involving higher-level cognitive
representations where human cognition appears to differ from other species (Spelke &
Tsivkin, 2001) thus supporting the importance of relationship between cognition and
language as discussed in the present study. Comprehension prior to production was
previously considered a universal stage of language acquisition. Data from young Thai
children suggests, however, that they may employ a distributional (Location & frequency)
Comprehensive Language Assessment Tools for Children
153
strategy for production of certain language forms before they comprehend these forms
(Carpenter, 1991).
In general however, the current study discovered mixed results where few groups
demonstrated correlation between cognition, reception and expression and few only between
reception or expression and cognition which could be due to factors such as linguistic, social
and environmental influences experienced by the child in their day to day life. Therefore,
during the preschool years, the relationship between comprehension, production, and
cognition changes frequently and inconsistently as the child develops. In general, linguistic
developments parallel much of the cognition growth of the preschool child, although there is
no one- to- one relationship. Such heterogeneity suggests that development is very complex.
Similarly, the present study also found inconsistent female superiority in the age range
studied.
Conclusions
The present study was mainly designed to construct a tool for assessment of language
and cognitive abilities in typically developing preschool children in the age range of 3 to 6
years where language was not considered as a hurdle for testing especially in multilingual and
multicultural Indian context. A general significant correlation was obtained between the
development of language and cognition. The performances of males and females was
compared which showed mixed results. After the administration of the test, the child‟s score
had to be compared with the normatives. This would indicate whether the child is deficient in
speech, language and cognitive aspects. It can be concluded that the present assessment tool
tests the development of language and cognitive abilities in typically developing preschool
children appropriately. The normative data obtained reveals that the performance of the
children increases as a function of age.
Implications and further research
The current study would help to assess and identify the language delayed/deviant
children between 3-6 years based on their receptive, expressive and cognitive abilities. The
age range of the present test can be extended further considering the cognitive development.
Further, the test tools can be standardized for different language speakers keeping in
consideration the social and environmental needs.
References
Bliss, S.L., Allen, V.D., Wrasse, W.K. (1977). A Story Completion Approach as a Measure
of Language Development in Children. Journal of Speech and Hearing Research, 20
(2), 358-372.
Brown, R. (1973). A first language: The early stages. Cambridge, MA: Harvard University
Press
Carpenter, K. (1991). Later rather than sooner: Extra linguistic categories in the acquisition of
Thai classifiers. Journal of Child Language, 18 (1), 93-113.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
154
Casby, W.M. (2003). Developmental Assessment of Play: A Model for Early Intervention.
Communication Disorders Quarterly. 24, 163-174.
Chapman, R. (2000). Children's language learning: An interactionist perspective. Journal of
Child Psychology and Psychiatry, 41, 33-54.
Chomsky, N. (1957). Syntactic structures, Cambridge, MA: MIT Press.
Chomsky, N. (1969). Acquisition of Syntax in Children from 5 - 10. Cambridge, MA: MIT
Press.
Clark, (1980). Owens, R. (2005). Language development: an Introduction, 7th
Ed., Merill
Publication Co.
Gopnik, A., & Meltzoff, A. (1986). Relations between semantic and cognitive development
in the one word stage: the specificity hypothesis. Child Development, 57, 1040-
1053.
Gopnik, A. & Meltzoff, A. (1997). Words, Thoughts, and Theories. Cambridge, MA: MIT
Press.
Griffin & Norris (1967). In Santhi, S.N. (2008). Development of Test for Syntax in
Malayalam. Unpublished master‟s dissertation. University of Mysore, Mysore,
India.
Huttenlocher, J., Smiley, P., and Ratner, H. (1983). In Part 5. Seiler, Th.B., and
Wannermacher, W. (Eds). Concept development and the development of word
meaning. (pp.210-234). New York: Springer Verlag.
Jesperson, O. (1922). “The Woman." In Language: Its Nature, Development and
Origin. London: Allen & Unwin.
Karlin, I.W. & Strazzula, M. (1952). Speech and language problems of mentally deficient
children. Journal of Speech Hearing Disorders, 17, 286-294.
Kennedy, M., Sheridan, M., Radlinski, S., & Beeghly, M. (1991). Play-language relationships
in young children with developmental delays: Implications for assessment. Journal
of Speech and Hearing Research, 34, 112-122.
Lyytinen, P., & Laakso, M. (1997). Language and symbolic play in toddlers. Journal of
Behavioural Development, 21(2), 289-302.
McCathren, R.B., Warren, S.F. and Yoder, P. (1996). Prelinguistic predictors of later
language development. In N. Cole, P. Dale and D. Thal (Ed), Assessment of
Communication and Language, pp. 57-75, Baltimore: Paul H. Brooks.
McCune (1995). In Vicky, L., & Jill, B. (2000). Relationship between symbolic play,
functional play, verbal and non-verbal ability in young children. International
Journal of Language and Communication Disorders.35 (1), 117-127.
McNeil, O. (1970). The Acquisition of Language: The study of developmental
Psycholinguistics. New York: Harper and Row.
Comprehensive Language Assessment Tools for Children
155
Ogura, T. (1991). A longitudinal study of relationship between early language development
and play development. Journal of Child Language, 18, 273-294.
Piaget, J. (1962). Play, dreams, and imitation in childhood. New York: Norton and Company.
Roopa, N., (1980). “Some aspects of syntax in 4-5 year old children: A descriptive study in
Hindi”. Unpublished master‟s dissertation, University of Mysore, Mysore, India.
Santhi, S.N. (2008). Development of test for syntax in Malayalam. Unpublished master‟s
dissertation, University of Mysore, Mysore, India.
Slobin, D. (1973). Cognitive prerequisites for the development of grammar. In C. Ferguson &
D. Slobin (Eds), Studies of child language development. New York: Holt, Rinehart
& Winston.
Spelke, E. & Tsivkin, S. (2001). Initial knowledge and conceptual change: space and
number. In M. Bowerman & S. Levinson (Eds.), Language acquisition and
conceptual development (pp. 70-100). Cambridge: Cambridge University Press.
Sudha, K.M. (1981). A syntax screening test in Tamil. Unpublished master‟s dissertation,
University of Mysore, Mysore, India.
Thelen (2005). In Owens, R. (2005). Language development: an Introduction, 7th
Ed.,
Merill Publication Co.
Tomasello, M., Striano, T., Rochat, P. (1999). Do young children use objects as symbols?
British Journal of Developmental Psychology, 17 (4), 563-584.
Venkatesan, S. (2004). Children with developmental disabilities: A training guide for
parents, teachers and caregivers. India: Sage Publication.
Vijayalakshmi, A.R. (1981). Development of a test in Kannada in assessing Language
acquisition in children, Unpublished doctoral thesis, University of Mysore, Mysore,
India.
Vygotsky, L. (1978). Mind in Society. The development of higher psychological processes.
Cambridge, MA: Harvard University Press.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
156
The Order of S, O, V Structures in Sign Language Users with Hearing
Impairment: Influence of Verbal Native Language?
Pallavi Malik & Dr. R. Manjula*
Abstract
Many investigators have expressed that structure of sign languages is essentially different
from that of the spoken languages. Very few studies have addressed the influence of verbal language
on the word order of ISL. The study aimed at investigating the sequence in which the Subject, Object,
Verb (SOV) signs occur in the sign language produced by individuals with hearing impairment
belonging to five verbal native languages (Kannada, Malayalam, Telugu, Tamil and Hindi). It also
aimed to investigate the influence of mode of communication by the signers and others in the family
on the use of SOV sequence of signs as tapped through a questionnaire. Ninety hearing impaired sign
language users belonging to five language backgrounds of Kannada, Malayalam, Telugu, Tamil and
Hindi participated in the study. They were asked to produce signs for a three picture story sequence
cards which was video recorded. The signs used were analyzed by three judges for the order of the
‘base units’ which was operationally defined for the study. Results indicated that SV was the most
commonly occurring structure in all the language groups suggesting the occurrence of an overly
simple structure which is not similar to the verbal language word order i.e. SOV. Signed expressions
beginning with S or O or V as a whole were taken into consideration for analysis. Results revealed
that signed expressions beginning with ‘S’ occurred maximally followed by ‘V’ and lastly ‘O’ for all
the language groups. The results of the questionnaire indicated that ‘only signs’ was the primary
mode of communication used by the sign language users with Hearing Impairment. A combination of
‘sign and speech’ was the primary mode of communication used by the family members except for the
Malayalam language group.
Introduction
Sign language is a language which uses manual communication, body language and
lip patterns instead of sounds to convey meaning—simultaneously combining hand shapes,
orientation and movement of the hands, arms or body, and facial expressions to express
fluidly a speaker's thoughts. In linguistic terms, sign languages are as rich and complex as
any oral language, despite the common misconception that they are not "real languages".
They have complex grammars of their own. Signs are conventional in nature, often arbitrary
and do not necessarily have a visual relationship to their referent, as in most of the spoken
languages. Sign languages are used most commonly by those individuals who have hearing
impairment. Woodward (1977) identified several language families based on hypothesized
relationship between known sign languages. According to Woodward (1977), American Sign
Language (ASL) belongs to the French sign language family and British Sign language
(BSL) belongs to the British sign language family. Other families identified include Asian,
South American, Egyptian, Indian, and Malaysian.
__________________________________
* Professor of Speech Pathology, All India Institute of Speech and Hearing, Mysore, India
email: [email protected]
Influence of Verbal Native Language on Grammatical Structure in SL users
157
The structure of sign languages as used in different countries varies. Sign languages
are natural languages that arise spontaneously wherever there is a community of
communicators; they effectively fulfill all the social and mental functions of spoken
languages; and they are acquired without instruction by children, given normal exposure and
interaction. These characteristics have led many linguists to postulate that sign languages
could be similar to spoken languages in significant ways. But sign languages are also
reported different from spoken languages. Sign languages exploit a completely different
physical medium as compared to the vocal-auditory system of spoken languages. Thus, sign
languages have their own unique phonology, morphology and syntax; which is essentially not
similar to the spoken languages.
Syntax, whether in verbal language or in sign language, begins with phrase structure.
It includes a set of properties which determine the construction of sentences in that language
(Hawkins, 2001). In spoken languages, there are conventional markings in the beginning and
end of the sentences. The order of words in a language is as important as is the meaning. If
the words or signs are not combined to the syntactic rules of that language, then the meaning
is either changed, lost or becomes unclear. The same holds good for sign languages as well.
Sign language syntax
For a long time, sign languages have been regarded as somewhat disreputable related
of spoken language. Sapir (1921) described it as a derivative of spoken language The major
breakthrough in the area of sign language syntax research began with studies in American
Sign Language (ASL). Amongst the 121 existing sign languages in the world (Gordon Jr.,
2005), American Sign Language (ASL) has been studied most extensively.
American Sign Language (ASL) is the dominant sign language used by the Deaf
community in the United States, Canada, and in some parts of Mexico. The proposal that
basic word order in ASL is Subject-Verb-Object (SVO) was supported by Fischer (1975)
who stated that (SVO) order is found in signed expressions with reversible subject and object.
Fischer (1975) also suggested the possibility of other word orders as well which generally
accompanied intonation breaks. Friedman (1976) argued against Fischer‘s (1975) analysis
claiming that several word orders occur in ASL and that the word order is of no grammatical
significance. Friedman (1976) claimed that word order is relatively free, with the exception
of the tendency for the verb to occur at last. Liddell (1980) critically evaluated both Fischer‘s
(1975) and Friedman‘s (1976) accounts and argued that Friedman‘s (1976) proposals were
not convincing. Although there was some variability in word order for declaratives, the only
grammatical word order corresponding to the yes/no question was SVO.
British Sign Language (BSL) is the sign language used in United Kingdom (UK).A
terminology known as ―Proform‖ is identified in BSL. Proforms are similar to pronouns. It is
anything that refers to, and stands in the place of something that is previously identified
(Sutton-Spence & Woll, 1999). In case of sign language, the term pronoun is used to mean I,
you, he, she, it, we, them etc., and proform is more specific only to BSL. The sign order of
BSL has its own rules which are very different from the rules of English. To use the proform,
it is necessary to sign the full sign and then the proform. BSL usually puts adjective second
whereas English puts adjective first. The time frame in BSL is marked at the beginning of the
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
158
signer‘s first sentence. The question and verb sign may be placed at the beginning as well as
at the end of the sentence. In BSL, the topic is established first followed by the comment on
the topic. Thus, the topic is the subject of the sentence. It does not have fixed role. It can also
be the focus, some old information, theme of the discourse till the next theme is established,
or the person or thing upon which the conversation is centered. On the other hand, comment
is the predicate. It is the new information about the topic. In BSL, the topic is marked in
several ways:
a) Topic comes first and is followed by a pause.
b) Eyes are widened during the topic and can be accompanied by a head nod.
c) The topic may be signed and held with one hand while producing the comment with
the other hand.
Major contribution in the area of word order of BSL was made by Deuchar (1983)
who supported the analysis of Friedman (1976) and stated that topic- prominence is the
characteristic of both BSL and ASL. Deuchar (1983) explained topic-prominence in BSL. On
the basis of these evidences, Deuchar (1983) questioned the variation reported in studies.
Some studies as that of Fischer (1975) and Liddell (1980) identified SVO as the basic word
order whereas others (McIntire, 1980) identified topic-comment as the basic structure of sign
language sentences. In fact, the studies which reported a topic-comment structure have
generally used recordings of spontaneous signing, while other studies reporting SVO
structure have used elicited sentences. Thus, Deuchar (1983) suggested that communication
situation may have an important effect on sentence type and thus analysis of the data in terms
of the topic-comment nature may give a useful insight into the syntactic structure of a Sign
language.
In contrast to ASL and BSL, Indian Sign Language (ISL) or Indo-Pakistani Sign
Language (IPSL) is the sign language variety that is predominantly used in the Indian sub-
continent. One of the earliest attempts to study IPSL was made by Vasishta, Woodward and
Wilson (1978). They studied the sign languages from the four major geographical divisions
of India which are expected to have some variations in the pattern of signing within the
country, namely, Delhi, Bombay, Bangalore and Calcutta. The results that they obtained were
as follows:
1. In simple sentences, predominant word order is SV.
2. Sentences comprising of Subject, Object, Verb relation were expressed using sign
order as well as various grammatical operations. Subject always preceded the object
and verb occupied the final position in 95% of the sentences
3. The expression of Subject, Verb, Object and Indirect Object relations gave different
results for different cities.
Influence of Verbal Native Language on Grammatical Structure in SL users
159
Table 1: Expression of Subject (S), Verb (V), Object (O) and Indirect Object (IO).
4. Pronominal forms are used when a context has been established earlier. But no specific
classifiers were observed to designate Pronouns in ISL unlike The American Sign
Language. Instead, effective use of space was seen which eliminated the need for
classifiers.
5. It was noticed that adjectives were always placed after nouns that they modified, but
with one exception. Color adjectives were always placed before the noun, particularly,
when modifying object nouns.
6. Negation was always placed at the end of the sentence.
7. The single past marker occurred in the sentence final position. Like the ASL, even in
Indian Sign Language, a time frame is generally set and no mention of it is made till
shift is made.
8. Non-manual markers are seen in case of Interrogative sentences. In an attempt to
produce the interrogative equivalent of a declarative sentence, forward and sideward
tilting of the head, raised eyebrows or holding the hands in the position of the last sign
is found.
Zeshan (2003) reported that sentences are always predicate final and generally has a
verb ending. She also reported that one-word sentences are common in IPSL and that there is
a strong preference for sentences with only one lexical argument. Further, she commented
that constituent order does not play any role in the marking of grammatical relations.
Topicalization of constituents was reported as a common strategy in IPSL (Zeshan 2003;
Aboh, Pfau & Zeshan 2005).
Verbal language syntax
English is the most widely spoken verbal language. English has an unmistakable word
order of Subject – Verb – Object (SVO) (Tomlin, 1986a). It is observed that English
maintains a rigid SVO word order mainly used to signal grammatical roles.
Kannada is one of the major Dravidian languages of India. Hiremath (1961) reported
that there is freedom of occurrence in the expression of various words in the sentence in
Kannada language, implying that the SOV usage is flexible. Schiffman (1979) also stated that
the basic word order in a Kannada sentence is SOV, but other orders can be found in the
language, due to stylistic variations. Malayalam is a Dravidian language used predominantly
in the state of Kerala. Peet (1980) stated that noun precedes its governing particle and the
finite verb always closes the sentence in Malayalam language. Adverbs are placed before the
verbs and adjectives or sometimes in any part of the sentence. Telugu is a Dravidian language
used predominantly in the state of Andhra Pradesh. Subhramanyam (1974) reported that verb
Cities Word order
Delhi S – O – IO – V (Incorporation of O and IO in V)
Bangalore S – IO – O – V (Incorporation of O in V)
Calcutta S – IO – O –V
IO – S – O – V (Incorporation of O and IO in V)
Bombay S – IO – O – V
IO – S – O – V (Incorporation of O and IO in V)
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
160
occupies the final position in the sentence, and there is an object in the sentence that usually
precedes the verb in Telugu language and the subject occupies the first position in the
sentence. Krishnamurti and Cwynn (1985) stated that when there is a transitive verb in the
sentence, then the natural word order of words is noun (subject) - noun (object) – verb. Tamil
is a Dravidian language used predominantly in the state of Tamil Nadu. Kerslake and Aiyar
(1953) reported that verb always comes in the last and the object always precedes the verb in
Tamil sentences. So the unmarked word order of Tamil language is Subject – Object – Verb
(SOV). In general, the unmarked word order of constituents in a sentence is subject-object-
verb (SOV) in the Dravidian languages (Krishnamurti, 2003).
Hindi belongs to Indo-Aryan Language family. Sharma, Chatterji, Satyanarayana,
Saksena and Nene (1994) suggested that the word order in Hindi is normally as follows:
1. When a sentence has an Intransitive Verb, then the word order found is SV.
2. When a sentence has a Transitive Verb, then the word order found is SOV.
3. When a sentence consists of a Transitive verb with two objects then the word order
followed is Subject – Secondary Object – Primary Object – Verb ( S 20O 1
0O V).
4. Adjectives are placed immediately before the noun that they qualify.
5. Adverbs are usually placed before the verb.
6. Negation denoting words are placed before the verb.
It is noteworthy that all these varied cultures have individuals with hearing
impairment who would use Sign Language. Thus, it seems natural that there would be an
influence of these spoken languages on the sign language used by these individuals as a result
of their interaction with the hearing individuals.
Need for the study
A major issue in the area of research that is less addressed is the influence of word
order of the verbal language to which the individual with Hearing Impairment who uses sign
language belongs, on the order expressed in the sign language of that individual. There are no
studies which have addressed this topic and it is quite understandable because most of the
well studied sign languages such as ASL and BSL originate from countries which are
essentially monolingual. The implication of this is that it does not facilitate cross language
comparisons in terms of the influence of native verbal language on sign language or the
influence of one or more verbal language over the sign language usage. India is a multilingual
country with 114 languages belonging to four distinct linguistic families. Thus, the scope to
address the issue of influence of verbal language word order on the order of emergence of
structure in the sign language is abundant in India.
This study attempts to look into the influence of the word order in the verbal language
on that of the order of emergence of structure in the sign language in all the four Dravidian
Languages (Kannada, Malayalam, Telugu and Tamil) and Indo-Aryan Language (Hindi). All
these languages have been found to have SOV as their predominant word order. Thus, it is
interesting to address the issue of the influence of word order of verbal language on the
Influence of Verbal Native Language on Grammatical Structure in SL users
161
signed expressions in individuals with Hearing Impairment using sign language belonging to
various native language backgrounds. The outcome of the study will facilitate comparison of
the emerging word order in the sign language expressed by persons with Hearing Impairment
and will throw light on the influence of the verbal language if any, on the signed order by
individuals with hearing impairment. It can indirectly contribute towards an understanding of
the factors which have led to the finding of different types of signs used by persons with
Hearing Impairment residing in different states in India, even different regions/ districts
within a state of India. It will help in understanding the linguistic constraints if any in the
variations from region to region seen in the sign languages of individuals with hearing
impairment in India.
Aims of the study
To compare the sequence in which the Subject, Object, Verb (SOV) signs occur in the
sign language produced by individuals with hearing impairment belonging to five verbal
native languages (Kannada, Malayalam, Telugu, Tamil and Hindi) and investigate if there
is any similarity between the sequences of SOV in signs used and
To understand the influence of mode of communication by the signers and others in the
family on the use of SOV sequence of signs as tapped through a questionnaire.
Method
Subjects
90 sign language users with hearing impairment participated in the study. The
subjects were selected randomly for the study. The demographic details of the subjects and
the mean ages across groups are provided in Table 2.
Table 2: Demographic details of the subjects
Subject selection criteria
They belonged to the verbal native language background, of Kannada, Malayalam,
Telugu, Tamil and Hindi languages (not necessarily spoken by the subjects, but the
subjects could be exposed to these languages through parents / caregivers / teachers or
medium of instruction through writing or reading).
They were in the age range of 15 - 25 years.
The minimum educational qualification of the subjects was secondary education in a
special school for the individuals with hearing impairment.
Age Range Verbal Native Language
Kannada Malayalam Telugu Tamil Hindi
M F M F M F M F M F
15 to 25 yrs 12 8 17 3 18 2 14 1 15 -
Total 20 20 20 15 15
Mean Age 18.75 20.47 19.4 20.06 21
M = Males
F = Females
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
162
The subjects were not exposed to any other verbal language other than their native
verbal language under which they were grouped.
Procedure
The subjects were grouped according to their verbal native language into five separate
groups. Each subject participated in the task individually. All the subjects were kept blind to
the purpose of the study. They were not allowed to discuss the task mutually before the
completion of the task.
Material
Two types of material were used in the experiment:
a) Questionnaire: A questionnaire was prepared to collect the following information from
each subject:
Family history of hearing loss
Type of school attended
Predominant mode of communication of the subject – sign / speech / sign and speech.
Predominant mode of communication used by subject‘s parents, teachers and other
communication partners, with the subject – sign / speech / sign and speech.
b) Picture stimulus cards: The subjects were provided with three sequence picture cards
which consisted of a story sequence. The three-picture story sequence cards depicted a
theme and it was such selected, that it provided a chance for the occurrence of a subject,
object and a verb in each card.
Setting and instruction: Initially, each subject was asked to fill the questionnaire. Then, each
subject was made to sit on a comfortable chair in a well-lit room with no external distracters.
Each subject was asked to express through signs, the content that was seen in the three
picture sequence story cards, as slowly and as clearly as possible.
Recording and storing
Before the commencement of the task, the three sequence cards were presented to the
subject for about 2 minutes to facilitate familiarization of the theme, the picture cards and
formulation of thoughts to express the contents through signs. Then, the cards were placed in
a row in front of the subject and the subject was asked to narrate the content using signs.
Recording of each individual‘s performance was done using a video camera and the data was
saved in compact disks for later analysis.
Analysis
Analysis of the data was done by three judges (sign language interpreters by
profession). The judges carried out the coding task independently without mutual
consultation / discussion. A ‘base unit’ was identified in the signs produced by the subjects
for the ease of analysis and to identify the word order. The ‗base unit‘ was operationally
defined in this study as one identified by each judge on the basis of following criteria:
Influence of Verbal Native Language on Grammatical Structure in SL users
163
The unit selected should include a meaningful (semantic) unit.
The selected base unit should be related semantically to the preceding and following base
units.
The video recordings were shown to each judge separately. Each video was paused at
appropriate ‗base unit‘ locations identified by the judge. Then the judge was asked to write
down the order of occurrence of Subject (S), Object (O) and Verb (V) produced in the video
in that particular ‗base unit‘. In this manner, all the videos were coded by all the three judges.
The various word orders, in which S, V and O occurred in these five languages was then
coded according to the ‗Base units‘ identified by the judges and tabulated separately. The
order identified by the judges was then grouped as the signs beginning with ‗S‘ as the initial
structure, or ‗V‘ as the initial structure, or ‗O‘ as the initial structure. As the number of ‗base
units‘ identified in each video by the three judges was varied; the data obtained was
converted into percentages for later statistical analysis.
Inter-judge reliability: Item-by-item comparison (agreement- disagreement) of the judges
was carried out. To treat the coding as valid, a reliability coefficient of 0.8-0.9 between the
judges per subject was considered.
Results and Discussion
I) Subject, Object, Verb (SOV) order occurrence in signs
A. Various SOV patterns
With the identification of the base units, it was observed that different patterns of
SOV orders emerged in the signs used by individuals with hearing impairment belonging to
different native verbal language groups. In depth analysis of the various sentence structures
revealed that there were many structures that were exhibited less frequently than others. A
score of ‗>/= 25%‘ of occurrence was considered to include a sentence structure for further
statistical analysis. Thus, ‗V(+O)‘ was considered for statistical analysis only for Malayalam
language group, ‗O‘ for Tamil language group and ‗V (+S, O)‘ for Telugu, Tamil and Hindi
language groups. The structure ‗SOV‘, which is the reported predominant word order pattern
of all the verbal native languages of the selected languages by many investigators [Kerslake
and Aiyar (1953); Hiremath (1961); Subhramanyam (1974); Schiffman (1979); Peet (1980);
Mohanan (1982); Krishnamurti and Cwynn (1985); Sharma, Chatterji, Satyanarayana,
Saksena and Nene (1994); Asher and Kumari (1997); and Krishnamurti (2003)], was also
exhibited by all the language groups in their signed expressions. But this structure was still
considered for statistical analysis to find out if it was significantly present in these languages
or not, as it is significantly present in all the verbal native languages of the subjects of the
study.
Thus, a total of seven Subject, Object, Verb (SOV) order patterns in the signed
expressions were considered for further analysis which included SV, SOV, SV (+O), O, V
(+S), V (+O) and V (+S, O). The exclusion criteria followed for this selection was based on
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
164
Reliability coefficient α for these word order patterns across all the language groups as coded
by all the three judges. High reliability values were obtained (α < 0.8). Further, the means and
standard deviations of these structures were computed and are tabulated in Table 3.
Table 3: Mean and Standard Deviation (SD) of the various SOV order patterns in all the
language groups.
Language groups Mean & SD SV SOV SV(+O) O V(+S) V(+O) V(+S,O)
Kannada Mean 35.34 4.45 15.72 1.30 29.84 7.93 .66
SD 21.11 8.21 15.05 3.67 30.67 17.56 2.05
Malayalam Mean 39.95 4.60 23.70 2.32 13.81 15.40 4.12
SD 15.02 7.42 13.39 4.82 16.89 28.41 7.34
Telugu Mean 45.20 3.44 19.33 3.10 17.26 .00 5.48
SD 20.47 6.61 15.57 5.61 23.36 .00 9.95
Tamil Mean 36.49 3.17 10.79 5.25 24.78 .00 13.38
SD 25.44 7.90 14.10 7.17 27.30 .00 20.26
Hindi Mean 36.17 5.33 20.65 1.48 19.12 3.57 13.66
SD 25.76 11.93 19.15 4.43 15.46 13.83 19.10
Total Mean 38.89 4.19 18.29 2.61 20.85 5.78 6.79
SD 21.29 8.26 15.69 5.26 23.84 17.49 13.46
It is observed that the ‗SV‘ structure emerged as the most predominant SOV order
pattern in all the language groups. This finding suggests that an overly simple sentence
structure was used by the subjects belonging to all the language groups predominantly which
was not similar to the one observed in verbal languages i.e. SV was found to be the most
predominant structure in all the language groups whereas SOV was identified as the principal
word order in verbal languages Kannada, Malayalam, Tamil, Telugu and Hindi languages
[Kerslake and Aiyar (1953); Hiremath (1961); Subhramanyam (1974); Schiffman (1979);
Peet (1980); Mohanan (1982); Krishnamurti and Cwynn (1985); Sharma, Chatterji,
Satyanarayana, Saksena and Nene (1994); Asher and Kumari (1997); and Krishnamurti
(2003)]. Thus, the order pattern of occurrence of Subject, Object, Verb (SOV) in sign
language expression is not similar to the verbal language word order pattern. This finding is
in line with the findings of Zeshan (2003) and Aboh, Pfau and Zeshan (2005), who stated that
ISL is a verb - final language. Vasishta, Woodward and Wilson (1978) also observed that in
simple sign expressions, subject always precedes the verb.
Multivariate measure ANOVA was carried out to compare all the five language
groups for the occurrence of predominant order patterns namely SV, SOV, SV (+O) and V
(+S) in the signed expression; which were present in all the language groups. Results of the
test revealed no significant difference for all the SOV order patterns across all the language
groups at p<0.05. This finding signifies that there exists identical pattern of occurrence of
these structures for all the language groups. Thus, it can be deduced that there is no difference
Influence of Verbal Native Language on Grammatical Structure in SL users
165
in sign expressions among these groups further suggesting the yet to be proved possibility
that there are no dialects in ISL as observed by Gordon Jr. (2005).
One way ANOVA was also carried out to compare the occurrence of the structure ‗V
(+S, O)‘ in Telugu, Tamil and Hindi Language groups and no significant difference was
revealed at p<0.05 which indicates that the structure ‗V (+S, O)‘ occurs in the same manner
in these language groups on a statistical basis despite percentage variations. This variation in
the percentage of occurrence might be because of the sample size wherein Telugu language
group had a sample size of 20 subjects and both Tamil and Hindi language groups had a
sample size of 15 subjects only. One way ANOVA was also carried out to compare the
occurrence of the order structure ‗V (+O)‘ in Malayalam, Kannada and Hindi language
groups, ‗O‘ across Tamil language group and ‗SOV‘ across all the language groups and the
results revealed no significant difference at p<0.05.
Repeated measure ANOVA was carried out to compare the occurrence of
predominant structures i.e., SV, SV (+O) and V (+S); within all the language groups.
Repeated measure ANOVA for the Kannada language group revealed significant difference
for SV and SV (+O) structure at p<0.1 but no significant difference was observed for V (+S)
structure. This finding suggests that SV is the most predominantly occurring structure
followed by V (+S) and SV (+O) which have equal probability of occurrence in the signed
expressions of the Kannada language group. Repeated Measure ANOVA for the Tamil
language group revealed significant difference for SV and SV (+O) at p<0.05 but no
significant difference was observed for V (+S) structure These results are in agreement with
the results obtained for the Kannada language group but the results for the Tamil language
group were highly significant at p < 0.05. This pattern [i.e. SV followed by SV (+O) and V
(+S)] was more well-defined in Tamil language group than Kannada language group is
evident from the higher significance obtained. The similarity between these two language
groups might be because these languages are reported to have similar features in their
linguistic structure including the phonetic and the syntactic structure (Krishnamurti, 2003).
Repeated Measure ANOVA for the Malayalam language group revealed significant
difference for SV & SV (+O) and SV & V (+S) at p<0.001 but no significant difference was
observed for SV (+O) and V (+S). Thus, SV was the most predominantly occurring structure
followed by SV (+O) and V (+S) in the Malayalam language group. Repeated Measure
ANOVA for the Telugu language group also revealed similar results as the Malayalam
language group at p<0.01. This implies that SV is the most predominantly occurring structure
followed by SV (+O) and V (+S) in descending order. But this pattern (i.e. SV followed by
SV (+O) and V (+S) in descending order) was more well-defined in Malayalam language
group than Telugu language group as evident from the statistical analysis. This similarity can
be attributed to the fact that Kannada, Malayalam and Telugu verbal languages are reported
to be similar in their grammar, syntax and vocabulary (Varma, 1999; Kunjamma, 1993). But
here, only Malayalam and Telugu were found to share a similar hierarchy and Kannada
shared a similar hierarchy with Tamil language group.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
166
Repeated Measure ANOVA for the Hindi language group revealed different results as
compared to the other language groups. The results showed no significant difference for any
of the word order patterns at p < 0.05. Thus, this finding shows that there exists equal
probability of occurrence of all the three structures under comparison namely, SV, SV (+O)
and V (+S) in the Hindi language group. This unique pattern obtained might be because of
the fact that Hindi language belongs to Indo – Aryan language family which has a different
origin than other language groups which belong to the Dravidian language family.
From the above findings, it is deduced that sign language used by the signers
belonging to all the language groups showed a predominant SV word order. In the literature,
it has been reported that the SOV structure is the most predominant word order in Kannada,
Malayalam, Telugu, Tamil and Hindi languages [Kerslake and Aiyar (1953); Hiremath
(1961); Subhramanyam (1974); Schiffman (1979); Peet (1980); Mohanan (1982);
Krishnamurti and Cwynn (1985); Sharma, Chatterji, Satyanarayana, Saksena and Nene
(1994); Asher and Kumari (1997); and Krishnamurti (2003)]. However, it is apparent that the
order that emerged in signed expressions of all the verbal native language groups (i.e. SV)
was not the same as reported for the word order in verbal language expressions of these
languages (i.e. SOV). These findings support the fact that even though these language groups
do not replicate the predominant word order pattern of the respective verbal native languages,
they show some similarity in terms of depicting similar hierarchy of predominant word order
pattern occurrence for Kannada & Telugu language groups and Malayalam & Tamil language
groups; and Hindi language group follows a totally different hierarchy which is not similar to
any of the other groups.
B. Sentence initial structures
After the analysis of the various word order patterns, signed expressions beginning
with S or O or V as a whole were taken into consideration for further analysis. Reliability
coefficient α for the signed expressions beginning with S, O or V; across all the language
groups was computed for all expressions as coded by the three judges. The results revealed a
high reliability value of α > 0.8. The means and standard deviations (SD) for the occurrence
of the signed expressions beginning with S, O and V were calculated and are represented in
Table 4.
Table 4: Means and SD for the structures beginning with S, O and V
Language Mean and SD Subject (S) Object (O) Verb (V)
Kannada Mean 62.41 4.90 37.66
SD 29.13 8.24 30.76
Malayalam Mean 76.29 3.32 20.38
SD 21.25 7.16 21.45
Telugu Mean 69.66 5.11 25.22
SD 25.92 8.49 26.79
Tamil Mean 52.92 6.92 40.21
SD 34.13 8.62 32.34
Hindi Mean 62.54 3.48 33.97
SD 23.30 6.43 24.74
Total Mean 65.55 4.69 30.86
SD 27.38 7.78 27.80
Influence of Verbal Native Language on Grammatical Structure in SL users
167
It was observed that the standard deviation for the structures beginning with O was
greater than the mean percentage of occurrence suggesting a scatter. From Table 4, it is
evident that structures beginning with S occurred more frequently in the signed expressions
of all the verbal native language groups (65.55%), followed by structures beginning with V
(30.86%) and structures beginning with O (4.69%) in the descending order. This finding is in
accordance with previous finding that ‗SV‘ is the predominant sign pattern in all the five
language groups which has contributed to the high percentage of occurrence of structures
beginning with ‗S‘ as well. Other structures found to be most predominantly occurring is ‗SV
(+O)‘ which is also a structure beginning with ‗S‘. Structures beginning with ‗V‘ follow next
as structures like V (+S), V (+O) and V (+S, O) were predominantly occurring in various
language groups. The pattern of occurrence of these structures has been discussed earlier. Out
of these structures, V (+S) was found to occur predominantly in all the language groups
whereas, other structures i.e. V (+O) and V (+S, O) were found to occur predominantly only
in Malayalam language group and Telugu, Tamil & Hindi language group respectively.
Consequently, these factors have contributed to similar hierarchy in all the language groups.
Mixed ANOVA was carried out to investigate for interaction effect, if any among all
the sentence initial structures namely S, O or V for all the language groups namely Kannada,
Malayalam, Telugu, Tamil and Hindi. The results revealed no significant difference for the
occurrence of the sentence initial structure (S or O or V) among the five language groups at p
< 0.05 i.e. there was no interaction between the sentence initial structures and the language
groups. However, a significant difference was observed for the sentence initial structures
alone at p < 0.001]. Further, Bonferroni‘s test for pair-wise comparison revealed significant
difference for all the three combinations. On the other hand, comparison of the languages
alone revealed no significant difference at p < 0.05. This finding reflects that the pattern of
occurrence of these structures are similar across all the language groups.
Further, one-way ANOVA was carried out to find out the difference, if any, across the
sentence initial structure categories between the language groups. The results of the test
suggested no significant difference for the structures beginning with S, V and O at p<0.05.
This indicates that individuals with hearing impairment in all the verbal language groups
showed a similar pattern of word order in signing when only sentence initial structure was
taken into consideration.
These findings thus suggest that though there were some differences found in the
pattern of occurrence of various signed order patterns, there was no evident difference
obtained when sentence initial structures i.e. all the structures beginning with S, O and V,
were taken up for investigation as a whole. This observation indicates that all the language
groups follow a similar trend and there are no significant variations between signing patterns
of hearing impaired individuals. This finding is in line with the findings of Vashista,
Woodward and Wilson (1978) who also stated similar findings. But their investigation
reported similarity of the signs for a list of words and not in terms of the emergence of signed
order for S, O or V. Thus, the findings of this study may be viewed as an extension to the
findings of Vashista, Woodward and Wilson (1978) with respect to emergence of S, V, O
order in signs. Although substantial support is required from other similar studies, it may be
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
168
stated that at a preliminary level, the results probably point towards the existence of a single
sign language in India. Further, it may also imply that ISL as a sign language may have less
likelihood of presenting with different dialects. In other words, it may be inferred that at least
in term of emergence of the order of signs for Subject, Verb and Object as syntactic
constituents of ISL, there may not be much variation across deaf communities spread across
Indian subcontinent. This observation falls in line with that of Vashista, Woodward and
Wilson (1978); Woodward (1993); Zeshan (2000) and Jepson (1991).
Furthermore, Repeated measure ANOVA was carried out to compare the occurrence
of order patterns beginning with S, O or V within the various language groups. Repeated
measure ANOVA for Kannada group revealed no significant difference between S and V as
initial structure at p<0.05. But there was a significant difference between S & O and O & V
as the initial structure. Repeated measure ANOVA for Tamil and Hindi groups revealed
similar results as the Kannada group. Hence, for Tamil and Hindi language group, there was
no significant difference between S and V as initial structure. But there was a significant
difference between S & O and O & V as the initial structure at p=0.05. Thus, the results of
the present study show that Kannada, Hindi and Tamil language groups follow a similar trend
i.e. for these language groups, there exists similar probability of occurrence for sentences
beginning with S and sentences beginning with V in spite of the huge variation of the mean
percentage occurrence of S as initial structure and V as a initial structure. The percentage of
occurrence of structures beginning with S were significantly more than the structures
beginning with O; and the structures beginning with V were significantly more than the
structures beginning with O. The similarity of the trend in Hindi language group with
Kannada and Tamil language group is not understandable as Hindi (Indo – Aryan language
family) has different origin from Kannada and Tamil languages (Dravidian language family).
Repeated measure ANOVA for Malayalam language group revealed significant
difference for all the combinations i.e. between S & V, O & V and S & O at p<0.05. Thus,
Malayalam and Telugu language groups seem to follow a similar trend of presenting
significant difference for all the sentence initial structures. Thus, sentences beginning with S
occurred more frequently followed by sentences beginning with V and O in the descending
order. This can be accredited to the piece of information that Malayalam and Telugu verbal
languages have similar syntax structure (Varma, 1999; Kunjamma, 1993) which has resulted
in this similarity in signed expression as well.
II) Response to the questionnaire to understand the mode of communication
The responses to the questionnaire obtained from the subjects were tabulated for
presence or absence of the factor sought for. The percentage of positive responses across the
subjects in the group is represented in Table 5.
Influence of Verbal Native Language on Grammatical Structure in SL users
169
Table 5: Percentage responses of the subjects to the questionnaire
It was also observed that all the signers belonging to different language groups used
‗only signs‘ followed by ‗both signs and speech‘ as their predominant mode of
communication. However, individuals with hearing impairment using ‗both signs and speech‘
as the predominant mode of communication were more in Kannada and Malayalam language
group followed by Hindi, Telugu and Tamil language group in the descending order. This
shows that the subjects fulfilled the criterion of being a signer more predominantly than being
a user who used a ‗combination of speech and signs‘ to communicate.
Regarding the mode of communication used predominantly by the family members
with the individuals with hearing impairment, it is evident that a combination of ‗sign and
speech‘ was used in all the language groups except the Malayalam language group.
Conclusions
The study provides an insight into the SOV order pattern used by sign language users
with hearing impairment belonging to various native verbal language background and this
study supports the findings of the previous studies by Vashishta, Woodward and Wilson
(1978), Zeshan (2003) and Aboh, Pfau and Zeshan (2005); that word order pattern of the
various sign language groups is not same as the verbal native languages of those groups.
Signers with hearing impairment tend to use an overly simplified sentence structure of SV
pattern most of the time whereas verbal native languages have a predominant structure of
SOV. Presence of embedded structures reflects upon the simultaneity feature of the sign
languages. The similarities obtained between some of the sign language groups are
suggestive of the fact that they have similar origin and similar syntactical structure. Thus,
these results reveal that the various sign language groups, at least in India may not be using
different sign languages but probably various dialects of the Indian Sign language attributing
to the fact that all sign languages show same structure with subtle variations when examined
at a gross level.
Parameters Kannada Malayalam Telugu Tamil Hindi
1. Positive family history of
hearing loss
20% 10% 5% 26.6% 20%
2. Predominant mode of communication
i) Speech - - - - -
ii) Signs 55% 55% 75% 80% 66.66%
iii) Signs and Speech 45% 45% 25% 20% 33.33%
3. Mode used by the family members
i) Speech 35% 50% 30% 33.33% 13.33%
ii) Signs 10% 15% 15% - 26.66%
iii) Signs and Speech 55% 35% 55% 66.66% 53.33%
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
170
References
Aboh, E., Pfau, R. & Zeshan, U. (2005). When a Wh-word is not a Wh-word: The case of
Indian Sign Language. In T. Bhattacharya (Ed.), The Yearbook of South Asian
languages and linguistics 2005 (pp. 11-43). Berlin: Mouton de Gruyter.
Deuchar, M. (1983). Is BSL an SVO language? In J. Kyle and B.Woll (Eds.) Language in
Sign: International Perspectives on Sign Language (pp. 69-76). London: Croom
Helm.
Friedman, L. (1976). The manifestation of subject, object and topic in American Sign
Language. In C. Li (Ed.), Subject and Topic. New York: Academic Press.
Fischer, S. (1975). Influences on word order change in ASL. In C. Li (Ed.), Word order and
word order change. Austin: University of Texas Press.
Gordon, R., Jr. (2005). Ethnologue: Languages of the World, Fifteenth edition. Dallas, Tex.:
SIL International. Online version: http://www.ethnologue.com/.
Hawkins, 2001 Hawkins, R. (2001). Second Language Syntax: a Generative Introduction.
Blackwell Publishers.
Hiremath (1961 Hiremath, R.C. (1961). The Structure of Kannada. Prasaranga, Karnatak
University, Dharwad.
Jepson, J. (1991). Urban and rural sign language in India. Language in Society, 1 (2): 37-57.
Kerslake, P. C. & Aiyar, N. C. R. (1953). Tamil course for European Schools (Book Four).
Christian Literature Society.
Krishnamurti, B. H. & Cwynn, J. P. L. (1985). A Grammar of Modern Telugu. Oxford
University Press.
Krishnamurti, B. (2003). The Dravidian languages. Cambridge University Press.
Kunjamma, S. (1993). Syntactic patterns of Malayalam: A Diachronic Study. Unpublished
Ph. D thesis. Department of Linguistics, University of Kerala, Thiruvananthapuram.
Liddell, S.K. (1980). American Sign Language syntax. The Hague: Mouton.
McIntire, M. (1980). Locatives in American Sign Language. Unpublished Ph.D dissertation,
University of California, Los Angeles.
Peet, J. (1980). A Grammar of Malayalam Language. Christian Missionary Society.
Sapir, E. (1921). Language: An Introduction to the Study of Speech. New York: Harcourt.
Schiffman, H. (1979). A Reference Grammar of spoken Kannada. U.S. Department of Health,
Education and Welfare.
Sharma, A., Chatterji, S. K., Satyanarayana, M. P., Saksena, B. R. & Nene, G. P. (1994). A
basic grammar of Modern Hindi. Central Hindi Directorate, Ministry of Human
Resource and Development, Government of India.
Influence of Verbal Native Language on Grammatical Structure in SL users
171
Subhramanyam, P. S. (1974). An Introduction to Modern Telugu. Sivakami Printers,
Annamalai University.
Sutton-Spence, R. & Woll, B. (1999). The Linguistics of British Sign Language: An
Introduction. Cambridge: University Press.
Tomlin, R. S. (1986a). Basic word order: Functional principles. London: Croom Helm
Varma, R. A. R. (1999). Kērala pāniniyam. International school of Dravidian Linguistics,
Thiruvananthapuram.
Vashishta, M. M, Woodward, J. C & Wilson, K. L. (1978). Sign language in India: regional
variation in deaf population. Indian journal of Applied Linguistics, 4(2): 66-74.
Woodward, J. C. (1977). All in the family: Kinship Lexicalization across sign languages.
Georgetown Roundtable on Language and Linguistics, Washington, DC.
Zeshan, U. (2003). Indo-Pakistani Sign Language grammar: A Typological Outline. Sign
Language Studies, 3: 157-212.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
172
Articulatory Acquisition in Kannada Speaking Urban Children: 3-4 years
Prathima. S & N. Sreedevi
*
Abstract
The purpose of the present study was to re-examine the articulatory norms for the
Diagnostic Kannada Articulation test (Babu, Rathna, and Bettagiri, 1972). This test was
administered to 120 typically developing children in the age range of 3-4 years. They were
divided into 2 age groups (3-3.6 years and 3.6-4 years) with six months age interval. 62
target words were picturized and were used to elicit the target response from the subjects.
The responses were audio recorded and the data obtained were transcribed using IPA
transcription. The test was scored individually and the data was statistically treated by
obtaining the mean and standard deviation. In this study, articulatory development for a
particular speech sound was assumed to be complete/mastered, if 90% of the subjects
articulated them correctly in the initial and medial positions of the target word. Two-way
ANOVA was carried out to determine the significant difference in articulatory scores between
the 2 age groups as well as across gender. In general it was found that there was no
significant difference in articulatory acquisition scores between the 2 age groups considered
as well as across gender. Results indicated that all the speech sounds in Kannada were
mastered by the age of 3-3.6 years except /r/ in the initial position and /h/ in both initial and
medial positions. However among the clusters, only /ski/ was mastered by the age of 4 years
and rest of the clusters probably continues beyond the age of 4 years.
Key Words: Articulatory acquisition, Kannada, initial/medial positions.
Introduction
Over the years, many investigators have tried to determine approximate ages of
individual sound “mastery” in specific groups of children. (Wellman‟31, Poole‟34,
Templin‟57 and Prather‟76). The results of all these studies (shown in Table 1) suggest that
the articulation scores, which is a measure of the sounds, produced correctly, increases with
age until maturity is reached by 8 years. They have also found sex and socio-economic status
differences in the age of acquisition of articulation. These guidelines regarding the speech
sound development is important when trying to separate normal from abnormal phonological
development and when determining therapy goals.
Irwin and Wong (1983) conducted a study on 100 children in the age range of 18-72
months. 10 boys and 10 girls in each of the following groups were considered for the study:
18 months, 2 years, 3 years, 4 years and 6 years. Results indicated that- At 18 months of age
only /a/, /u/, /i/ and /Λ/ were produced with atleast 70% accuracy. At 2 years of age all the
vowels and diphthongs were produced with atleast 80% accuracy with the exception of /ɚ/
_________________________________________
* Lecturer in Speech Sciences, All India Institute of Speech and Hearing, Mysore, India
email: [email protected]
Articulatory Acquisition in Kannada Speaking Children
173
and /ɝ/. At 3 years of age all the subjects produced all the vowels and diphthongs with 99-
100% accuracy.
Table 1: Age levels for the speech sound development according to different authors. Empty
space indicates sound not tested. * Criteria for the sound to be considered as acquired.
Speech
sounds
Wellman‟
31 75%*
Poole‟34
100%*
Templin‟57
75%*
Sander‟72
75%*
Prather‟75
75%*
Arlt‟76
75%*
Fudala‟86
90%*
IP FP
m 3 3 ½ 3 Below 2 2 3 2 2
n 3 4 ½ 3 Below 2 2 3 2 2
h 3 3 ½ 3 Below 2 2 3 1 ½
p 4 3 ½ 3 Below 2 2 3 2 3
f 3 5 ½ 3 3 2-4 3 2 ½ 3
w 3 3 ½ 3 Below 2 2-8 3 1 ½
b 3 3 ½ 4 Below 2 2-8 3 2 3
4 ½ 3 2 2 3 3
j 4 4 ½ 3 ½ 3 2-4 3
k 4 4 ½ 4 2 2-4 3 2 ½ 3
g 4 4 ½ 4 2 2-4 3 2 ½ 3
l 4 6 ½ 6 3 3-4 4 5 5
d 5 4 ½ 4 2 2-4 3 2 ½ 2 ½
t 5 4 ½ 6 2 2-8 3 3 4
s 5 7 ½ 4 ½ 3 3 4 11 11
r 5 7 ½ 4 3 3-4 5 5 ½
t 5 4 ½ 4 3-8 4 5 ½ 5 ½
v 5 6 ½ 6 4 4 3 ½ 5 ½ 5 ½
z 5 7 ½ 7 4 4 4 11 11
Ʒ 6 6 ½ 7 6 4 4
θ 7 ½ 6 5 4 5 6 6
ʃ 6 ½ 4 ½ 4 3-8 4 ½ 5 ½ 5 ½
Stoel-Gammon (1985), in her longitudinal study using spontaneous speech sample on
34 children, 19 boys and 15 girls in the age range of 15 to 24 months. The samples were
collected every 3 months, from 9 months to 24 months. Only the meaningful speech
production of at least 10 identifiable words during a 1- hour recording session was analyzed
and the results indicated that- By 15 months of age- voiced stops, nasals and glides were
produced in the word initial position. By 24 months of age- voiceless stops, velars, and few
fricatives were produced in the initial position. Voiceless stops and alveolar consonants were
produced in the word final position. Voiced stops appeared first in the initial position. /t/ and
/r/ appeared first in the word final position.
Fudala and Reynolds (1986) studied the acquisition of single phonemes and the
results indicated that, the age of acquisition of /s/ and /z/ appears to be quite late compared to
earlier studies i.e. from 6 to 6.5 years. As the group got older the percentage of children who
produced correctly got decreased to less than 90% (as low as 62% for the 7 to 7.11 years)
then the percentage began to gradually increase again at 8 to 8.11 years and by the age of 11
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
174
to 11.11 years 98.4% of children produced it correctly. Similar results were obtained for final
/-z/ and /-s/.
Extensive normative data on the articulatory acquisition in Indian languages are
limited. However, a number of attempts have been made in some of the Indian languages also
which are summarized in Tables 2 & 3 below. In Kannada, the articulation test was
developed by Babu, Rathna and Bettagiri in 1972, but the norms for the same was established
by Tasneem Banu in 1977.
Table 2: Speech sound development according to different Indian studies.
Thirumalai ‟72- Tamil Kumudavalli‟73- Kannada Sreedevi ‟76- Kannada
No. of subjects
considered: one
Age: 4.4 years old.
Results-
All the stop consonants
were produced and
Substituted alveolars
for retroflex sounds
No. of subjects considered:
105 school going children
Age range: 4 to 8 years.
Results-
Alveolar and retroflex
sounds were acquired last.
No. of subjects considered:4
children
Age range: 2 to 2.6 years.
Results-
All the stop consonants
were acquired by the age
of 2.6 years.
The laterals, sibilants and
trills were not fully
established.
Table 3: Age levels for the speech sound development according to different Indian studies.
Speech
sounds
Tasneem Banu
‟77 (Kannada)
75%*
Usha‟86
(Tamil)
75%*
Padmaja‟88
(Telugu)
75%*
Arun Banik‟88
(Bengali)
90%*
Maya‟90
(Malayalam)
75%*
m 3 3 2.6 2.5 3-3.6
n 3 3 2.6 2.5 3-3.6
ŋ + + + 2.5 3-3.6
p 3 3 2.6 2.5 3-3.6
f + + 2.9 + 3-3.6
h + + 2.6 3 3-3.6
w + + + + +
y 3 3 2.6 2.5 3-3.6
k 3 3 2.6 2.7 3-3.6
b 3 3 2.6 2.5 3-3.6
d + 3 2.6 3 3-3.6
g 3 3 2.6 3 3-3.6
r 4.6 + 3.9 4 3.7-4
s 3 3 3.3 + 3.6-4
ʃ 5.1 6 3.6 3 5-5.6
tʃ 3.7 3 2.6 3 3-3.6
t + 3 2.6 3 3-3.6
Articulatory Acquisition in Kannada Speaking Children
175
t 3 + + 3 3-3.6
v + 3 2.6 + 3-3.6
l 3 3 2.6 3 3-3.6
t + + + 3 +
z + + + + +
z + + + 4 +
j 3 3 2.6 3 3-3.6
hw + + + + +
’+’ indicates sound not tested. * Criteria for the sound to be considered as acquired
More recently, researchers have been concerned with the validity of the „classical‟
data with respect to modern day living with its earlier and more extensive educational
programs and the expanding effects of television on children. In view of the fact that the
study on articulatory acquisition in Kannada was conducted 30 years back (Tasneem Banu,
1977) and that recent studies focusing on phonological processes have revealed suppression
of most of the processes by the age of 3- 4 years, there is an immediate need to study the
articulatory acquisition in the present day children to verify this observation and apply it in
the clinical set up.
Method
The objectives of the study were:
1. To revise the norms for the Kannada Articulation Test (1972) and to establish the ages
at which 75% and 90% of the children produce the phonemes of Kannada correctly.
2. To compare the articulatory skills across age and gender.
3. To compare the difference in the articulatory acquisition of phonemes in the initial
and medial positions of the words.
4. To compare the data obtained with that of the earlier reported studies in both English
and Kannada.
Subjects
Kannada speaking typically developing urban children in the age range of 3-4 years
were selected randomly from different localities in Mysore city as subjects. They were sub
divided into two groups with an inter age interval of six months (3-3.6 and 3.6-4 years). Each
group comprised a total of 60 subjects including 30 boys and 30 girls. So a total of 120
subjects were considered for the present study. The subjects were selected based on the
following criteria:
1. Native speakers of Kannada, being reared in an urban ambient environment of Kannada
and belonging to middle socio economic status. The subjects were exposed to some
amount of English language also in the kindergarten set up.
2. Parents/teachers report and informal chronological age-level performance on selected
speech, language, hearing and developmental tests were used to establish typical
development. The Assessment Checklist for Speech and Language skills (Geetha,
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
176
Rajkumar and Divya, 2006) and Developmental Screening Test (Bharathraj, 1981) were
utilized for this purpose.
Test material
The Diagnostic Kannada Articulation Test by Babu, Ratna and Bettagiri (1972) was
used as the test material. (The recording sheet of the test is given in Appendix I). This test
comprises of four parts; Part I includes items to test 10 vowels, 2 diphthongs and 22
consonants in Kannada. Part II is similar to Part I but has different words with the same
sounds being tested for test-retest reliability. Part III tests 10 clusters and Part IV includes a
paragraph for older children. In the present study Part I and Part III were utilized. Vowels and
diphthongs were tested only in the initial position. All the consonants were tested in both
initial and medial positions except the phoneme /ʃ/ which was tested only in initial position
and phonemes // and // which were tested only in the medial position. Four clusters were
tested in the initial position and six in the medial position. The target words were picturized
on white cards of 4 x 6” size in color. The picture cards were arranged in order as per the
sequence of Kannada alphabets. One target phoneme in one position i.e. either in the initial or
medial position was tested with each picture card.
Data Collection Procedure
Each individual subject was brought into a noise free room and seated opposite to the
examiner. Once the rapport was established, the examiner presented the target pictures one
after another. The subjects were encouraged to name the item in the picture card. The
response obtained was audio recorded using Olympus digital recorder WS-100. If any of the
subjects failed to identify a target word, additional cues were presented by the examiner. In
spite of the additional cues, if the child failed to name the target picture, the child was asked
to repeat the target word after the examiner.
Data Scoring
The data obtained from all the 120 subjects were transcribed using broad and narrow
IPA transcription. All the responses of each subject were analyzed sound-by-sound on a
response sheet. Correct responses (CR), substitutions (S), indicating the substituted
phoneme, omissions (O), distortions (D), additions (A) or any other type of articulatory
deviation (Ao) was recorded on the response sheet.
Test –Retest-Reliability
Reliability of the responses was tested by “test retest” method. 10% of the subjects
from each age group were selected randomly and tested using the same material on the same
day or with in a span of three days.
Data Analysis
The detailed analysis for obtaining the articulatory norms was carried out in the
following manner:
Articulatory Acquisition in Kannada Speaking Children
177
1. Identification of the phonemes uttered correctly by each subject in both the positions
and a score of “one‟ was assigned to each correct item
2. From the scores obtained for the correct articulatory production, mean, standard
deviation and range of scores were calculated for each age group and independently
for boys and girls in initial and medial positions
3. Identification of phonemes which are produced correctly by 75% and 90% of the
subjects in each age group in the initial and medial positions separately. An analysis
of the phonemes substituted or distorted in production is discussed in detail.
Results and Discussion
The present study aimed at re-establishing the norms for „The Diagnostic Kannada
Articulation Test‟ by Babu, Ratna and Bettagiri (1972). Two-way ANOVA was carried out to
find the significance difference in articulatory scores between the 2 age groups as well as
across gender. In general it was found that there was no significant difference in articulatory
acquisition scores between the 2 age groups considered as well as across gender.
The results are discussed under the following 5 main headings:
1. Age vs. acquisition.
2. Gender vs. acquisition.
3. Order of acquisition of the sounds.
4. Speech sound acquisition vs. word position.
5. Test-retest reliability.
1. Age vs. Articulatory acquisition
Two-way ANOVA was carried out to find the significant difference in articulatory scores
between different age groups (3-3.6 years and 3.6-4 years). The results revealed that there
was no significant difference (p>0.05) in articulation scores across the two age groups tested.
This can be attributed to the acquisition of most of the consonants by the age of 3-3.6 years
itself, and the inter group interval considered was broad (6 months). Table 4 and Graph 1
show the mean articulation scores in the two age groups considered in the study.
Table 4: Mean articulation scores & SD in different age groups.
Gender Age Mean Std. Deviation N
Boys
3 - 3.6 yrs. 56.3917 4.38340 30
3.6 - 4 yrs 57.8000 3.36974 30
Total 57.0958 3.94079 60
Girls
3 - 3.6 yrs. 56.9167 2.96992 30
3.6 - 4 yrs 57.9500 3.37358 30
Total 57.4333 3.19390 60
Combined
Scores
3 - 3.6 yrs. 56.6542 3.72153 60
3.6 - 4 yrs 57.8750 3.34382 60
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
178
The findings of Wellman (1931) Poole (1934) and Templin (1957) indicated that the
“phoneme” development correlated with age, that is some sounds are mastered earlier than
others. The results of the present study are also in consonance with the above reports. It is
evident that with neuromuscular maturity all motor skills increased as the age advanced and
so also the articulation skills.
2. Gender vs. Articulatory articulation
Two-way ANOVA revealed no significant difference (p>0.05) between boys and girls
in terms of articulatory acquisition in both the age groups (shown in Table 4 & Graph 1).
Tables 5 and 6 show the development of vowels, diphthongs and consonants in 3-3.6 and 3.6-
4 year old Kannada speaking children in boys and girls respectively. Articulatory acquisition
in terms of 75% and 90% of acquisition by the children was considered as carried out in the
Western studies. In the present study, articulatory development for a particular speech sound
was assumed to be completed, if 90% of the children articulated them correctly in the initial
and medial positions of the target word.
In accordance with the present study, Roe and Milisen (1942) reported no significant
difference between articulatory skills of males and females. Perkins (1977) states that, the sex
of a child does not appear to be a significantly major factor. In the present study, the
difference between boys and girls was not significant and this may be accounted to the equal
opportunity in the speech environment for both boys and girls in the present day scenario
especially in the urban population.
3. Order of acquisition of sounds
It was generally observed that all the vowels and most of the consonants except /r/, /h/
// /d/ and // were acquired by the age of 3-3.6 years. The results of the present study
were compared with the studies of (Templin‟57 and Wellman ‟31 in English) to observe
whether the order of acquisition was similar. At the outset it appears that the age of
acquisition of different speech sounds in Indian languages are relatively faster compared to
Western studies. However this observation needs to be interpreted with much caution because
all the reported studies in the western context that are available are carried out from early
thirties to the seventies or so.
Articulatory Acquisition in Kannada Speaking Children
179
Comparing the present study with that of study by Tasneem Banu (1977) in Kannada
language, there are certain discrepancies noticed. She reports that the fricative /ʃ/ was
acquired by 75% of children at the age of 5.1 years and in the present study, considering 90%
criteria it is acquired by 3.6-4 years among boys and by 3-3.6 years in girls. Another salient
observation is that the semivowel /v/ is not acquired by 75% of children even at the age of 6.6
years in the earlier report, where as in the present study it is acquired by 90% of girls and
boys by the age of 3-3.6 years.
Table 5: Age of articulatory acquisition by 75%, 90% and obtained % of the boys. ‘+’ indicates
Sound acquired, ‘-‘indicates sound is not acquired. Empty space indicates sound not tested.
Boys
Speech
sound
3-3.6 years 3.6-4 years
Initial position Medial position Initial position Medial position
75% 90% obtained 75% 90% obtained 75% 90% obtained 75% 90% obtained
a + + 100% + + 100%
a + + 100% + + 100%
i + + 100% + + 100%
i + + 100% + + 100%
u + + 100% + + 100%
u + + 100% + + 100%
e + + 100% + + 100%
e + + 100% + + 100%
ai + + 100% + + 100%
o + + 100% + + 100%
o + + 100% + + 100%
ou + - 86.6% + + 90%
k + + 93.3% + + 93.3% + + 100% + + 100%
g + + 93.3% + + 93.3% + + 100% + + 100%
t + + 86.6% + + 83.3% + + 96.6% + + 100%
d + + 83.3% + + 90% + + 93.3% + + 100%
t + + 90% + + 96.6% + + 93.3% + + 96.6%
d + + 93.3% + + 93.3% + + 90% + + 90%
n + + 93.3% + + 93.3%
t + + 100% + + 100% + + 100% + + 100%
d + + 100% + + 100% + + 96.6% + + 100%
n + + 100% + + 100% + + 100% + + 100%
p + + 100% + + 100% + + 100% + + 100%
b + + 100% + + 100% + + 100% + + 100%
m + + 100% + + 100% + + 100% + + 100%
j + + 100% + + 100% + + 100% + + 100%
r - - 46.6% + - 76.6% + - 83.3% + + 90%
l + + 96.6% + + 100% + + 100% + + 100%
v + + 73.3% + + 100% + + 83.3% + + 86.6%
+ + 96.6% + - 86.6% + + 100% + + 93.3%
s + + 93.3% + + 93.3% + + 100% + + 100%
h - - 36.6% - - 43.3% - - 66.6% - - 70%
+ + 100% + + 93.3%
sta + - 83.3% - - 70% + - 80% - - 73.3%
sku + - 80% + - 76.6%
ski + - 86.6% + + 96.6%
dra - - 60% - - 60%
rti - - 66.6% + - 83.3%
kra - - 70% + + 90%
ksa + - 70% - - 73.3%
ble + - 76.6% - - 73.3%
skru - - 43.3% - - 26.6%
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
180
Table 6: Age of articulatory acquisition by 75%, 90% and obtained % of the girls. ‘+’ indicates
sound acquired, ‘-‘indicates sound is not acquired. Empty space indicates sound not tested.
Girls
Speech
sound
3-3.6 years 3.6-4 years
Initial position Medial position Initial position Medial position
75% 90% obtained 75% 90% obtained 75% 90% obtained 75% 90% obtained
a + + 100% + + 100%
a + + 100% + + 100%
i + + 100% + + 100%
i + + 100% + + 100%
u + + 100% + + 100%
u + + 100% + + 100%
e + + 100% + + 100%
e + + 100% + + 100%
ai + + 100% + + 100%
o + + 100% + + 100%
o + + 100% + + 100%
ou + + 96.6% + + 90%
k + + 100% + + 100% + + 100% + + 100%
g + + 100% + + 100% + + 100% + + 96.6%
t + + 96.6% + + 96.6% + + 100% + + 100%
d + + 96.6% + + 96.6% + + 96.6% + + 96.6%
t + + 96.6% + + 96.6% + + 96.6% + + 93.3%
d + - 83.3% + - 86.6% + + 90% + + 86.6%
n + + 93.3% + + 96.6%
t + + 96.6% + + 100% + + 96.6% + + 100%
d + + 100% + + 100% + + 100% + + 96.6%
n + + 100% + + 100% + + 100% + + 100%
p + + 100% + + 100% + + 100% + + 100%
b + + 100% + + 100% + + 100% + + 100%
m + + 100% + + 100% + + 100% + + 100%
j + + 100% + + 100% + + 100% + + 100%
r - - 53.3% + - 80% - - 70% + - 80%
l + + 100% + + 100% + + 100% + + 100%
v + + 90% + + 100% + + 93.3% + + 100%
+ + 96.6% + + 100% + + 100% + + 100%
s + + 96.6% + + 100% + + 100% + + 100%
h - - 53.3% - - 30% - - 66.6% - - 43.3%
+ - 86.6% + + 93.3%
sta - - 70% - - 63.3% + - 80% - - 66.6%
sku + - 86.6% - - 70%
ski + + 100% + + 96.6%
dra - - 46.6% - - 60%
rti - - 60% - - 56.6%
kra - - 63.3% + - 76.6%
ksa + - 80% - - 73.3%
ble + - 83.3% + - 80%
skru - - 43.3% - - 46.6%
When comparing the present study with that of other Indian studies, there is no much
discrepancy observed. In Tamil // is acquired by the age of 6 years and this late acquisition
is attributed to the phonetic system of the language itself as /s/ is a borrowed phoneme (Usha,
1986). Most of the speech sounds of Malayalam language are acquired by the age of 3 years
itself (Maya, 1990). In Malayalam, /s/ is acquired by 75% of children by the age of 3.6-4
years and // by the age of 5-5.6 years which is comparatively late compared to the present
study (3- 3.6 years). Padmaja‟s study (1988) in Telugu reveals that most of the speech sounds
Articulatory Acquisition in Kannada Speaking Children
181
are acquired by 75% of children by the age of 2.6 years and in Bengali (Arun Banik, 1988)
90% of children had acquired most of the speech sounds by 2.5-3 years of age.
The speech sounds acquired are also discussed separately in terms of Place, Voicing
and Manner features.
Place feature: According to place of articulation, un aspirated speech sounds of Kannada
language can be classified as bilabials (/p/, /b/, /m/), labiodentals (/v/), dentals (/t/, /d/,
/n/,), retroflex (//, //, //, //), palatals (/t/, /d/), velars (/k/, /g/) and glottal (/h/)
sounds. In the present study, all the bilabials, labiodentals, dentals, palatals, velars and
retroflex except //and // were acquired by the age 3-3.6 years; glottal /h/ was not
acquired even by the age of 4 years.
Voicing feature: In the present study, voiced sounds such as // and // were acquired
late i.e by the age of 3.6- 4 years considering the 90% criteria. Sounds such as /h/ and /r/ were
not acquired by 90% of the children even by the age of 4 years. /r/ was achieved in the medial
position by 90% of the boys. Where in /r/ remained not acquired even by 75% of children in
the initial position in the older group (3.6-4 years)
Manner of articulation
Vowels and Diphthong: All the vowels and diphthong /ai/ were mastered by 90% of the
children by the age of 3-3.6 years. Here there is a great possibility that all the vowels were
achieved by 3 years or soon after 3 years itself. But because of the relatively large age
interval (6 months) considered in this study, it is only possible to say vowels were mastered
by 3-3.6 years. Diphthong /ou/ was acquired by the age of 3.6-4 years. This is similar to the
reports in Western studies.
Plosives: In general, all the stops in this study were acquired by the age of 3-3.6 years.
Similar results are reported in Templin‟s (1957) and Wellman‟s (1931) studies. However,
among the stops, the voiceless retroflex // was acquired at an earlier age of 3-3.6 years in
the present study when compared to the studies of Templin (1957; 4.6 years) and Wellman
(1931; 5 years).
Fricatives: Dental /s/ (initial and medial positions) and palatal /ʃ/ (initial position) was
acquired by 90% of the children by the age of 3-3.6 years. Retroflex // (medial position)
was acquired by the age of 3.6-4 years by 90% of the girls and 87% of the boys. It is
observed that the acquisition of fricatives is much earlier compared to the studies of Templin
(1957) at 4.5 years
However, another salient observation was that the glottal fricative /h/ was not
mastered by even 75% of the children by 4 years also. This is because in colloquial Kannada,
usage of /h/ is minimal. For example, the word /hasu/ (meaning cow), is generally uttered as
/asu/ and it is considered as normal production. Earlier also in Kannada, Tasneen Banu
(1977) reports that /h/ is not acquired by 75% of children even by the age of 6.6 years owing
to the same reason.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
182
Affricates: There are two affricates in Kannada language /t/ and /d/. It is observed that
both the sounds are acquired by 90% of the children by the age of 3-3.6 years, which is much
earlier compared to the western studies, as per Sander‟72, 4 years, Prather‟75, by 3-3.8 years
and Fudala‟86 , by the age of 5.6 years.
Nasals: In general it was noted that all the nasals were acquired by 3-3.6 years, which is
similar to the Western studies.
Laterals: There are two laterals in Kannada language, voiced alveolar lateral /l/ and voiced
retroflex lateral /l/. /l/ occurs in initial and medial positions and // occurs only in medial
position in Kannada language. /l/ was found to be acquired earlier (3-3.6 years) when
compared to English speaking children. In the present investigation, retroflex // was
mastered in girls earlier (3- 3.6 years) than boys (3.6- 4 years).
Trills: Voiced alveolar trill /r/ was not mastered till 4 years of age which is relatively late
compared to the other phonemes in Kannada. Also the observation was that the percent
acquisition of trills was higher in the medial position than in the initial position for both boys
and girls. The late mastery is due to the difficulty in production of /r/.
Semivowels: In the present study, /v/ and /j/ were acquired by 3-3.6 years which was much
earlier compared to that of Templin‟s (1957), by 6 years, Wellman‟s (1931) by 5 years and
in Tasneem Banu‟s (1977) data, it was not acquired till the age of 6.6 years.
Clusters: In the present study, 4 clusters (/sta/, /sku/, /ble/ and /skru/) were tested in the
initial position and other six (/sta/, /ski/, /dra/, /rti/, /kra/ and /ksa/) were tested in the
word medial position. In girls, /ski/ had 90% acquisition, /ksa/ and /ble/ had 75% acquisition
by 3-3.6 years. There were also few unusual observations in this group, i.e. clusters /sku/ and
/ksa/ was found to be acquired by 75% of children in the younger age group (3-3.6 years)
and the percentage reduced in the older age group (3.6-4 years), /sku/ (70%), /ksa/
(73.33%).
Similarly among boys, the clusters /ksa/ and /ble/ were acquired by the age of 3-3.6
years by 75% of the children and reduced in percentage by 3.6-4 years, /ksa/ (73.33%) and
/ble/ (73.33%). Clusters like /ski/ and /kra/ was acquired by 90% of the boys. The finding that
/kra/ and /ski/ are mastered by 90% of the children is in accordance with the study by Curtis
& Hardy, 1957 where in /r/ is produced correctly more often in stop blends than in fricative
blends. So stops facilitate correct production of /r/ /s/ and /z/ but only when they precede /r/
and follow /z/ and /s/.
Based on the scoring obtained from the present study, typically developing Kannada
speaking urban children should get a score of 56.39 + 4.3 (boys) and 56.91 + 2.9 (girls) in
the age range of 3-3.6 years. 57.80 + 3.36 (boys) and 57.95 + 3.37 (girls) in the age range of
3.6-4 years. Maximum score is 62. Articulation scores expected for typically developing
children in the age range of 3-4 years in boys and girls are depicted in the Table 7.
Articulatory Acquisition in Kannada Speaking Children
183
Table 7: Articulation scores expected for typically developing children in the age range of 3-
4 years
Age Gender Scores expected for typically
developing children
3-3.6 years Boys 56.39 + 4.3
3-3.6 years Girls 56.91 + 2.9
3.6-4 years Boys 57.80 + 3.36
3.6-4 years Girls 57.95 + 3.37
Speech sound acquisition vs. word position
In the present study, boys acquired voiced affricate /d/ earlier in the medial position
by the age of 3-3.6 years and later in the initial position. Trill sound /r/ was first acquired in
the medial position by 90% of the boys by the age of 4 years and was not acquired by 90% in
the initial position even by the age of 4 years. Among girls, there was no such variation
present in terms of position however only /r/ was acquired by 75% of girls in word medial
position first (3-3.6 years) and less than 75% of children produced it in the initial position.
Such positional variations are reported in other studies also.
Studies have suggested that fricatives (Farwell, 1977) and velars (Ingram, 1974) may
appear first in final position. Stoel-Gammon‟s (1985) study reveals one consistent difference
between initial and final position: the phoneme /r/ appeared word finally well before it
occurred word initially. This finding is similar to the present study, where in /r/ appeared
earlier in medial position than in the initial position of words.
In the present study, medial clusters (/ski/ and /kra/) were acquired earlier compared
to initial clusters. This finding is in accordance with the study by Vani Rupela and Manjula
(2006) where in medial clusters were first to be acquired and appeared by the age of 18-24
months and, more frequent and predominant at the age of 30-36 months.
5. Test-retest reliability
There was 98% reliability found between the scores of the first and the second test
administration. Table 8 shows the mean articulation scores between first and second
administration scores.
Table 8: Shows the mean articulation scores in the first and second
administration of Diagnostic Kannada Articulation Test.
3-3.6 years 3.6-4 years
First
administration
Second
administration
First
administration
Second
administration
54.50 54.00 60.00 59.50
56.00 56.00 59.50 59.50
56.00 56.00 57.00 55.50
57.50 56.00 60.00 60.00
60.50 61.00 59.50 59.00
60.00 60.00 57.00 56.50
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
184
Conclusions
It can be concluded that, all the vowels and diphthong /ai/ are acquired by the age of
3-3.6 years, diphthong /ou/ is acquired by the age of 3.6-4 years and most of the consonants
are acquired by the age of 3.6- 4 years except /r/ and /h/. /r/ was acquired by 90% of the
children in medial position but not in the initial position and /h/ was not acquired in both the
positions even by the age of 4 years. Among clusters, medial clusters (/ski/ and /dra/) are
acquired by the age of 3.6-4 years. Another feature observed was that, the children in the
present study acquire most of the sounds at a younger age compared to the earlier reports in
English and Kannada.
References
Arlt, P. B., & Goodban, M. T. (1976). A comparative study of articulation acquisition as
based on a study of 240 normals, aged three to six. Language, Speech, and Hearing
Services in Schools, 7, 173-180.
Arun Banik (1988). Articulation test in Bengali. Unpublished master‟s dissertation submitted
in part fulfillment for the master‟s degree in Speech and Hearing. University of
Mysore.
Babu, R. M., Bettagiri, R, & Rathna, N. (1972). Test of articulation in Kannada, Journal of
AIISH, 3, 64-79.
Bharathraj, (1981). Developmental Screening Test, Swayam Sidha Prakashana, Mysore
Curtis, M. E. & Hardey, W. M. (1957). Cited in Padmaja, B. (1989). Telugu articulation test.
Unpublished master‟s dissertation submitted in part-fulfillment for the master‟s
degree in Speech and Hearing. University of Mysore.
Farwell, C. B. (1976). Some strategies in the early production of fricatives. Papers and
Reports on Child Language Development, 12, 97-104.
Fudala, J. B., & Reynolds, W. M. (1986). Arizona articulation proficiency scales (2nd Ed.).
Los Angeles: Western psychological Press.
Geetha, Y. V., & Rajkumar (2006). The Assessment Checklist for Speech and Language
Skills. Unpublished ARF project, AIISH, Mysore.
Ingram, D. (1974). Phonological rules in young children. Journal of Child Language, 1, 49-
64.
Irwin, J. W., & Wong, S. P. (Eds.). (1983). Phonological development in children 18 to 72
months. Journal of Speech and Hearing Disorders, 12, 402-404.
Kumudavalli. (1973). The relationship between articulation and discrimination of Kannada
speech sounds in terms of distinctive features. Unpublished master‟s dissertation
submitted in part-fulfillment for the master‟s degree in Speech and Hearing.
University of Mysore.
Articulatory Acquisition in Kannada Speaking Children
185
Maya. (1990). An articulation test battery in Malayalam. Unpublished master‟s dissertation
submitted in part-fulfillment for the master‟s degree in Speech and Hearing.
University of Mysore.
Padmaja, B. (1988). Telugu articulation test. Unpublished master‟s dissertation submitted in
part-fulfillment for the master‟s degree in Speech and Hearing. University of Mysore.
Perkins. (1977). Cited in Padmaja, B. (1989). Telugu articulation test. Unpublished master‟s
dissertation submitted in part-fulfillment for the master‟s degree in Speech and
Hearing. University of Mysore.
Prather, E., Hedrick, D., & Kern, C. (1975). Articulation development in children aged two to
four years. Journal of Speech and Hearing Research, 40, 55-63.
Poole, E. (1934). Genetic development of articulation of consonant sounds in speech.
Elementary English Review, 11, 159-161.
Roe & Milisen. (1942). Cited in Padmaja, B. (1989). Telugu articulation test. Unpublished
master‟s dissertation submitted in part-fulfillment for the master‟s degree in Speech
and Hearing. University of Mysore.
Sander, E. (1972). When are speech sounds learned? Journal of Speech and Hearing
Disorders, 37, 55-63.
Sreedevi. (1976). The acquisition aspects of Kannada language in 2+ year old children.
Unpublished master‟s dissertation submitted in part-fulfillment for the master‟s
degree in Speech and Hearing. University of Mysore.
Stoel-Gammon, C. (1985). Phonetic inventories, 15-24 months: A longitudinal study. Journal
of Speech and Hearing Research, 28, 505-512.
Tasneem, B. (1977). Articulatory acquisition in Kannada A study of normal children 3-6.6
years. Unpublished master‟s dissertation submitted in part-fulfillment for the master‟s
degree in Speech and Hearing. University of Mysore.
Templin, M. C. (1957). Spontaneous versus imitated verbalization in testing articulation in
preschool children. Journal of Speech and Hearing Disorders, 12, 293-300.
Thirumalai (1972). Acquisition of phonology of a child. Unpublished master‟s dissertation
submitted in part-fulfillment for the master‟s degree in Speech and Hearing.
University of Mysore.
Usha, D. (1986). Tamil Articulation Test. Unpublished master‟s dissertation submitted in
part-fulfillment for the master‟s degree in Speech and Hearing. University of Mysore.
Vani, R., & Manjula, R. (2006). The Phonotactic Development in Kannada speaking children
in the age range of 0-5 years. International Journal of Communication Disorders,
Wellman, B., Case, I., Mengert, I., & Bradbury, D. (1931). Speech sounds of young children.
University of Iowa Studies in Child Welfare, 5 (2).
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
186
APPENDIX-I (Sample of the scoring sheet)
Name: XXX Age/Gender: 3.4 years/Male
Name of the School: The Play House, Kuvempunagar, Mysore.
Sl.
No
Ph
on
eme
Check word
Initial position
S
core
Check word
Medial position
S
core
C
R
S O D A A
o
C
R
S O D A A
o
1 a alilu √ 1
2 a ane √ 1
3 i ili √ 1
4 i iligeman
e
√ 1
5 u ugura √ 1
6 u uta √ 1
7 e ele √ 1
8 e elu √ 1
9 ai aidu √ 1
10 o ondu √ 1
11 o ole √ 1
12 ou ouada √ 1
13 k kannadaka √ 1
14 k pustaka √ 1
15 g gadijara √ 1
16 g mugu √ 1
17 t tamata √ 1
18 t batanige √ 1
19 d dade √ 1
20 d sudi √ 1
21 t topi √ 1
22 t kitaki √ 1
23 d dabbi √ 1
24 d bledu √ 1
25 n 1 gini √ 1
26 t takkadi √ 1
27 t koti √ 1
28 d dalimbe √ 1
29 d kudure √ 1
30 n naji √ 1
31 n kannadi √ 1
32 p pennu √ 1
33 p tappali √ 1
34 b bagilu √ 1
35 b ombattu √ 1
36 m muru √ 1
Articulatory Acquisition in Kannada Speaking Children
187
37 m emme √ 1
38 j jama √ 1
39 j rupaji √ 1
40 r redijo le ½
41 r karu lu ½
42 l lari √ 1
43 l kalu √ 1
44 v vimana √ 0
45 v bavuta √ 1
46 artu √ 1
47 bra √ 1
48 s surja √ 1
49 s bassu √ 1
50 h hattu √ 0
51 h simha √ 0
52 l 1 koli √ 1
53 sta stampu √ 1
54 sta postabaksu √ 0
55 sku skutar √ 1
56 ski biskittu √ 1
57 dra tandra √ 1
58 rti kurti √ 1
59 kra takra √ 1
60 ks
a
atoriksa √ 1
61 ble bledu √ 1
62 skr skru √ 0
Correct responses (CR-score 1), substitutions (S-score 1/2) indicating the substituted
phoneme in the respective block, omissions (O- score 0), distortions (D-score 3/4), additions (A-
score 0) any other type of articulatory deviation (Ao).
This subject obtained a Total score of 56/62.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
188
Normative Nasalance Value in Hindi Language
Pravesh Arya & Dr. M. Pushpavathi*
Abstract
Nasalance is a commonly used measure that allows the speech-language pathologist to
validate and quantify a perceptual assessment. The nasalance score is calculated as a ratio of the
nasal sound pressure level to the combined nasal and oral sound pressure level (Fletcher, 1978). The
present study was aimed to develop normative data on nasalance value across gender and nasalance
deviation in Hindi language for oral and nasal sentences and paragraphs for adults and to study the
effect of stimuli and gender on nasalance value. One hundred normal Hindi speaking adults including
fifty males and fifty females with age range 18 to 35 years were taken as subjects for the study. The
Nasal View (version 4) was used for the data collection. The subjects were instructed to read or
repeat 5 oral and 5 nasal sentences & 1 oral and 1 nasal paragraphs prepared by an experienced
speech and language pathologist whose native language is Hindi. Results showed that there is a
significant difference across stimuli and gender. Females exhibited higher nasalance value than
males across gender and nasalance value for nasal stimuli was higher than oral stimuli. Clinically the
normative data reported in the present study may help to identify and treating individuals with
resonance disorders. However, normative nasalance value across dialect variation in Hindi language
and other Indian languages needs to be investigated.
Introduction
Speech refers to the processes associated with the production and perception of
sounds used in spoken language. “Speech is the form of communication in which the
transmission of information takes place by means of speech waves which are in the form of
acoustic energy. The speech waveforms are a result of interaction of one or more source with
the vocal tract filter system” (Fant, 1960).
Nasalization may be defined as the existence of significant communication between
the nasal cavity and the rest of vocal tract. Two of the most characteristic clinical components
of resonance disorders are hypernasality and nasal emission. Hypernasality may be defined as
the presence of excessive nasal resonance during the production of vowels or vowel-like
consonants. Nasal emission deals with the presence of turbulent noise production during the
production of high pressure consonants. This turbulent noise is often detected as an audible
“puff” of air emitted via nostrils. Nasality is a perceptual attribute whose detection requires
the judgment of a listener (Moll, 1964; McWilliams et al. 1981; Haapanen, 1991a).
Perceptual judgments have been and will be continued to be used in the clinical
setting because of the apparent face validity of these judgments (Dalston, 1997). However,
perceptual judgment scale has a number of significant drawbacks.
___________________________
* Reader in Speech Pathology, All India Institute of Speech and Hearing, Mysore, India
email: [email protected]
Normative Nasalance Value in Hindi
189
In an effort to address the limitations of perceptual judgments by themselves,
numerous methods of objectively measuring various aspects of speech disorder have been
developed. In the area of resonance disorders, a number of methods have been developed to
objectively evaluate the characteristics of hypernasality and nasal emission. One of the
stronger objective assessment methods of the nasal speech signal is through the measurement
of nasalance. Nasalance has been defined as the ratio of nasal (n) to oral (o)
sound pressure level and is commonly derived via the following formula:
Nasalance is a commonly used measure that allows the speech-language pathologist
to validate and quantify a perceptual assessment. The nasalance score is calculated as a ratio
of the nasal sound pressure level to the combined nasal and oral sound pressure level
(Fletcher, 1978). There are some instruments like Nasometer, NasalView and Oronasal
system which are being use to obtain nasalance value for normal as well as disordered
individuals in all population.
NasalView (Awan, 1997) is a new PC-based system for the computerized
measurement of nasalance (Tiger Electronics, Seattle, WA). NasalView provides mean,
minimum and maximum nasalance values (in percentage) for different stimuli like syllable,
sentences and paragraphs i.e. oral, nasal and oro-nasal. The sampled sound signal or signal
selections can be played back and edited with the NasalView program. Oscillograms of the
nasal and oral signal and a nasalance curve are displayed together with nasalance statistics so
that speech segments can be identified accurately and the envelope of the nasalance curve
over time can be related to particular sounds.
Very few studies have been conducted to establish normative data across gender and
across stimuli using NasalView system and there is no published study done in Indian
language to establish normative nasalance value and to measure the effect of nasalance across
gender and across stimuli using NasalView system.
Awan, S. (1998) conducted a study to find the nasalance across two instruments i.e.
Nasometer and NasalView and to measure the effect of mean nasalance across stimuli Total
181 subjects were considered for the study, among which 161 were children and 20 were
male and female adults (age range 18-30 years). Three common passages i.e. Zoo passage
(oral paragraph), Rainbow passage (oro-nasal paragraph) and Nasal Sentence (nasal
paragraph) were used as stimuli. Results revealed that mean nasalance values (in percentage)
obtained by NasalView system were as; nasal paragraph (48.44%), Rainbow passage (34.19
%), Zoo passage (24.67 %) for normal adult subjects. Author reported that, nasalance values
obtained with the NasalView tend to be higher for oral stimuli and lower for nasal stimuli,
compare with the mean nasalance values obtained using Nasometer. Author concluded that
NasalView system appears effectively to separate varying degrees of nasal speech and
although actual nasalance values differ from those of the Nasometer, NasalView provides
high levels of both validity and reliability in its ability to measure RMS nasalance.
Nasalance = {Nasal(n) / Nasal(n) + Oral(o)} X 100
mmmmf==========+=++Oral(o)
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
190
Awan, S. Daniel, Z. H., Jordan, R. G. (2001) established a preliminary normative data
for both children and adults. Data was collected from total 255 subject among them 203 were
children and 52 were adult subjects (29 females and 23 males) using NasalView instrument.
Authors considered 29 normal subjects without any communication impairment and had no
history of clefting or velopharyngeal incompetence/insufficiency and no complaint of upper
respiratory infection and/or nasal congestion. To assess varying degrees of normal nasal
resonance, subjects were asked to read the following three passages while wearing the Nasal
View headgear. The Zoo Passage (Fletcher, 1972), Rainbow Passage (Fairbanks, 1960) and
Nasal Sentences (Fletcher, 1972). For those very young subjects (ages 5 to 6 years) who had
difficulty reading the three passages, subjects asked to repeat the following: The first two
sentences of the Zoo Passage; the 2nd
sentence of the Rainbow Passage and the first sentence
of the Nasal Sentence. The authors reported that mean nasalance exhibited by female subjects
was 26.71 % and mean nasalance exhibited by male subjects is 26.02 % for oral paragraph
(Zoo Passage). Similarly, mean nasalance exhibited by female subjects was 51.18 % and
mean nasalance exhibited by male subjects was 50.81% for nasal paragraph (Nasal
Sentences). Although authors have not mentioned about nasalance difference across gender
but the results revealed that significant difference was not evident across gender.
Bressmann, T. et al. (2006) conducted a study to compare the nasalance values
obtained by fifty normal subjects (Thirty-one female and nineteen male) with age range 23 to
44 years and nineteen hypernasal patients with cleft palate (Eight female and 11 male) with
age range 11 to 19 years. There mean nasalance were measured across three instruments i.e.
the Nasometer, the NasalView and the OroNasal system. Two passages were taken as
material i.e. Zoo passage (oral paragraph) and Nasal Sentence (nasal paragraph). Authors
reported that mean nasalance exhibited by fifty normal subjects (male and female combined
group) on NasalView instrument was 21.09 % for oral passage and 55.74 % for the nasal
paragraph. On the other hand, mean nasalance exhibited by eight subjects with cleft palate
with hypernasality was 25.72 % for oral paragraph and 47.55 % for nasal paragraph. Subjects
with cleft palate with moderate hypernasality exhibited mean nasalance of 27.71 % for oral
paragraph and 50.21 % for nasal paragraph using NasalView instrument. Authors concluded
that mean nasalance value exhibited by cleft palate patient with hypernasality (mild or
moderate) is higher than mean nasalance exhibited by normal subjects for both oral as well as
nasal paragraph. Their results revealed the significant difference in mean nasalance across
stimuli. They conclude that the nasalance scores from the Nasometer, the NasalView and the
Oronasal system are not interchangeable and that nasalance magnitudes from the three
systems cannot be compared directly.
A need for a reliable, objective measure of speech nasality with high level of content
validity was largely met with the NasalView system to address the limitations of perceptual
judgment in the area of assessment of resonance disorders. Studies have shown that nasalance
of normal speech is sensitive to the phonetic composition of the speech stimuli, native
language, regional dialect, age and gender. There are limited data concerning nasalance
values in Indian languages. This makes the strong need for the establishment of regional
norms as there are very few standardized normal nasalance values for normal speakers in
Indian languages.
Normative Nasalance Value in Hindi
191
Aims of the study
To develop normative data on nasalance value across gender and nasalance deviation in
Hindi language for oral and nasal sentences and paragraphs for adults.
To study the effect of stimuli on nasalance value across gender.
To study the effect of gender on nasalance value across stimuli.
Method
Subjects
One hundred normal Hindi speaking adults participated in the present study. All the
participants were native speakers of Hindi language. Each subject was evaluated by an
experienced speech and language pathologist to assess oral structure and function. Subjects
were divided into two groups. First group consist of 50 females in the age range of 18-35
years and second group was consisted of 50 males in the age range of 18-35 years. Table 1
depicts the subjects‟ details.
Table 1: Subjects details
Gender Age range (Mean age) No. of subjects
Male 18-35 years (26.5years) 50
Female 18-35 years (26.5years) 50
Total 100
Stimuli
Two sets of stimuli were prepared by an experienced speech language pathologist
whose mother tongue is Hindi. One set consisted of oral sentences, which had predominantly
oral consonants and the other set was nasal sentences which consisted of predominantly nasal
sentences. Each category consisted of ten sentences. Sentences were made simple, short, easy
to remember and meaningful. The sentences selected were ranged in length from three to four
words (five to six syllables). The paragraphs were ranged in length from six to seven
sentences.
Procedure
To find the content validity of the stimulus materials, sentences were given for
content judgment to ten speech language pathologists who had at least one year experience in
the field. Judges were asked to read the given instructions carefully before rating. A five
point perceptual rating scale was used. Rating of „0‟ indicates fully oral or no nasality and „4‟
indicated highly nasalized for both the categories. Ten sentences and two paragraphs were
given to them in each category and were provided with a scoring sheet to rate for each
sentence and paragraphs i.e. oral and nasal (scoring sheet: Appendix- I). Most appropriate
five sentences in each category out of ten were selected on the basis of rating given by
judges. The material and scoring sheet is provided in the Appendix-I (A).
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
192
Instrumentation
The Nasal View system is a PC/Windows-based system which provides for the
recording of high resolution speech signals using windows-compatible sound cards (sampling
at up to 4410 Hz at 8 or 16 bits of resolution). The hardware components found in the Nasal
View system include headgear and a portable custom dual-channel pre-amplification unit.
The key component of the headgear is a rigid plate constructed of 5mm thick styrene straps.
The sound separator plate is suspended from a Jackson Model 170 headgear (Jackson
Products, Belmont, MI) by styrene straps. The sound separation characteristics of this plate
are augmented by the addition of a light weight acoustic barrier material. (Fig 1.)
Figure 1: PC based NasalView system with headgear
Data Collection
The NasalView was setup in a suitable quite recording room. The instrument was
calibrated prior to the experiment based on the instructions provided in the manual. The
speech sample was recorded individually. After selecting the subjects they were seated
comfortably, and the nasal view head set was placed on subjects head (Fig 1 (c)). The
position of the nasal view head set was adjusted and secured firmly in accordance with the
manufacturer‟s instructions. Once the nasal view headset correctly positioned, subjects were
asked to read the stimuli if he/she is a literate. In case of non literate, they were asked to
repeat the speech stimuli after the speech pathologist for a reliable output. This was repeated
after 30 minutes on the same subjects for the purpose of test re -test reliability. This reliability
measures was followed for both the oral and nasal category (sentences and paragraph).
Figure 2: Subject undergoing NasalView assessment.
Normative Nasalance Value in Hindi
193
Statistical Analysis
Four types of tests i.e. descriptive analysis, mixed ANOVA (Analysis Of Variance),
independent t-test, paired t-test were administered to perform the statistic analysis. Mixed
ANOVA was used to study the effect of type, nasality and gender on nasalance value,
independent t-test was used to measure the effect of gender on nasalance value, paired t-test
was used to find the effect of stimuli on nasalance value and to compare the nasalance value
within and across stimuli.
Results
I. Effect of Stimuli on Nasalance Value in Females
The nasalance mean for oral sentences was 22.58 % with the range from 6.56 % of
minimum to 65.55 % of maximum. Similarly, nasalance mean obtained for oral paragraph
was 22.23 % with the range from 4.0 % of minimum to 85.22 % of maximum. The nasalance
mean for nasal sentences were 50.41 % with the range from of 11.31% of minimum to
87.80% of maximum. Similarly, nasalance mean for nasal paragraph was 49.74 % with the
range from 6.29% of minimum to 90.02% of maximum.
Table 2: Mean Nasalance Value (in percentage) of Oral Stimuli
Stimuli Mean (S.D) Min. (S.D) Max. (S.D)
Oral sentences 22.58 (4.61) 6.56 (1.65) 65.55 (8.42)
Oral paragraph 22.23 (4.33) 4.00 (2.13) 85.22 (6.71)
Nasal sentences 50.41 (3.94) 11.31 (2.56) 87.80 (4.62)
Nasal paragraph 49.74 (4.55) 6.29 (2.90) 90.02 (4.38)
II. Effect of Stimuli on Nasalance Value in Males
The nasalance value obtained for oral sentences was 22.48 % with the range from
6.51% of minimum to 62.83% of maximum. Similarly, nasalance values obtained for oral
paragraph was 21.59% with the range from 3.46 % of minimum to 83.22 % of maximum. It
shows that mean nasalance value obtained for Nasal sentences was 47.03 % with the range
from 9.75% of minimum to 85.74 % of maximum. Similarly, mean nasalance value obtained
for Nasal Paragraph was 46.84% with the range from 5.59% of minimum to 88.30 % of
maximum.
Table 3: Mean Nasalance Value (in percentage) of oral stimuli
Stimuli Mean (S.D) Min. (S.D) Max. (S.D)
Oral sentences 22.48 (4.79) 6.51 (1.78) 62.83 (11.40)
Oral paragraph 21.59 (4.85) 3.46 (1.23) 83.22 (8.45)
Nasal sentences 47.03 (4.86) 9.75 (2.52 ) 85.74 (6.37)
Nasal paragraph 46.84 (5.23) 5.59 (2.14) 88.30 (5.58)
The following figure depicts that the mean nasalance value obtained for oral sentences
and oral paragraphs.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
194
Figure 3: Nasalance Value for Oral sentence and oral paragraph across gender.
III. Effect of Gender on Nasalance Value: Oral Stimuli
Results shows that both female and male subjects exhibited approximately same mean
nasalance value i.e.22.58 % for females and 22.48 % for males. Independent t-test was done
to find the significant difference and the results indicated that significant difference was not
evident across gender.
Table 4: Nasalance Value (in percentage) for oral sentences
Gender Mean (S.D) Min (S.D) Max (S.D)
Male 22.48 (4.79) 6.51 (1.78) 62.83 (11.40)
Female 22.58 (4.61) 6.56 (1.65) 65.55 (8.42)
IV. Effect of Gender on Nasalance Value: Nasal Stimuli
Results show that female subjects exhibited higher mean nasalance value i.e. 50.41 %
compared to mean nasalance value obtained by male subjects i.e. 47.03 % for nasal
sentences. Similarly, female subjects exhibited higher minimum nasalance value i.e. 11.31 %
as compared to minimum nasalance value obtained by male subjects i.e. 9.75 %. Maximum
nasalance value obtained by female subjects was 87.80 % which is higher than maximum
nasalance value obtained by male subjects i.e.85.74 % for nasal sentences. Independent t-test
was done to find the significant difference across gender and the results revealed significant
difference was evident. Female subjects exhibited higher mean nasalance Value for nasal
paragraph i.e. 49.74% as compared to mean nasalance values obtained by male subjects i.e.
46.84 %. Female subjects exhibited higher minimum as well as maximum nasalance values
i.e. 6.29 % and 90.02 % respectively as compared to minimum and maximum nasalance
values obtained by male subjects i.e. 5.59 % and 88.30 % respectively. Independent t-test was
used to find the significant difference across the two groups. The results revealed a
significant difference for the nasal paragraph across gender.
Normative Nasalance Value in Hindi
195
Table 5: Nasalance values for Nasal Sentences across Gender
Gender Mean (S.D) Min(S.D) Max(S.D)
Male 47.03 (4.86) 9.75 (2.52) 85.74 (6.37)
Female 50.41 (3.94) 11.31 (2.56) 87.80 (4.62)
Male 46.84 (5.23) 5.59 (2.14) 88.30 (5.58)
Female 49.74 (4.55) 6.29 (2.90) 90.02 (4.38)
The following figures depict the mean nasalance value obtained for nasal sentences and nasal
paragraphs.
Figure 3: Nasalance Value for Nasal sentence and Nasal paragraph across gender.
Results of the present study are supports the finding of Bressmann et.al, (2000), who
conducted a study to find the relationship between nasalance ratio for nasal and nonnasal
sentences using NasalView instrument. Results from their study reveal that there is a
significant difference present across stimuli. Mean nasalance obtained from nasal sentences
showed higher values than oral sentences which supports the results obtained from the
present study.
Daniel Zaoming Huang (2001) conducted a study to develop preliminary normative
data for NasalView instrument using three passages i.e. Zoo passage, Rainbow passage and
Nasal sentences. Authors reported that mean nasalance exhibited by female subjects and male
subjects are approximately same for oral paragraph (Zoo Passage), oro-nasal paragraph
(Rainbow Passage) as well as for nasal paragraph (Nasal Sentences) which reveals a
significant difference across stimuli. This could be attributed to the reason that these reports
were made on the basis of English or languages with similar phonological characteristics to
English.
Similar results were obtained from study done by Tim Bressman (2005), which
revealed that there is a significant difference present across three passages i.e. Zoo Passage,
Rainbow Passage, and Nasal Sentences i.e. mean nasalance obtained from oral and nasal
sentences, whereas there is no significant difference present within stimuli i.e. mean
nasalance obtained from oral sentences and oral paragraph and mean nasalance obtained from
nasal sentences and nasal paragraph.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
196
III. Clinical Interpretation of Normative Data
Establishing the nasalance values for clinically significant abnormalities is important
in many areas of medical epidemiology. It can be approached from a clinical or statistical
perspective (Baker and Rose, 1984). Initially clinical perspective had been widely used
Dalston et al. (1991a, 1993) who used clinical rather than statistical approach. Perceptual
ratings on a numerical scale were predetermined to be clinically significant at a particular
value. Then, nasalance scores for nasalance were set as those that give the best overall correct
prediction of the presence or absence of abnormal nasality. Regardless of the methodology
differences that have led to the development of cut off scores, it is quite clear that cut off
values determines that there are some speakers whose resonance is judged to be normal and
who have abnormal resonance. Table (6) shows cut off values for normal adult male and
female.
Table 6: Cut off values for across gender and stimuli
Stimuli Female Male
Oral sentence 13.36-30.18 12.90-32.06
Oral paragraph 13.57-30.89 11.89-31.29
Nasal sentence 42.53-58.29 37.31-56.75
Nasal paragraph 40.64-58.84 36.38-57.30
A significant difference was evident across gender in adults. This difference may be
attributed to basic structural and functional differences across gender. An adult female has
difference in size, shape and resonating cavity of vocal system than an adult male. This
difference might lead to show significantly high mean nasalance value exhibited by female
subjects as compare to male subjects.
The possible reason could be the underlying anatomical and physiological differences
related to velopharyngeal closure across gender. But the present study does not support the
findings of Trindade et al., 1997; Van Doorn & Purcell, 1998; Sweeney et al., 2004, who
reported no significant difference across gender.
Summary and Conclusions
The present study primarily aimed to establish normative data for Hindi speaking
male and female adults. The subjects for the study considered were one hundred normal
subjects with normal oral structure and functions. All the participants were native speakers of
Hindi language. Subjects were divided into two groups. First group consisted of fifty females
and the second group consisted of fifty male adults. Both group had equal number of males
and females. For the purpose of the stimuli, experienced speech language pathologists whose
mother tongue was Hindi prepared two sets of Hindi sentences. Ten sentences and two
paragraphs were given to them in each category and were provided with a scoring sheet to
rate for each sentence and paragraphs i.e. oral and nasal (scoring sheet provided in Appendix-
I). Most appropriate five sentences in each category out of ten were selected on the basis of
Normative Nasalance Value in Hindi
197
rating given by judges. The material and scoring sheet is provided in the Appendix-II. The
Nasal View (version 4) was used to for the data collection. The instrument was calibrated
prior to the data collection. Nasal View head gear was placed on the subjects head. Once the
head set is positioned properly, the subjects were instructed to read or repeat sentences. After
the completion of each speech sample, the nasalance trace was stored on computer file for
latest analysis. The data was analyzed for adults (male and female separately) using
descriptive statistics, independent t-test, mixed ANOVA and paired t-test using SPSS
software version 16.0 package.
Following Table (12) shows the mean normative nasalance values with the range of
minimum nasalance value to the maximum nasalance value for adults across gender and
across stimuli.
Table 7: Normative mean nasalance value across gender and across stimuli
Stimuli Gender Mean (S.D) Min (S.D) Max (S.D)
Oral Sentences Male 22.48 (4.79) 6.51 (1.78) 62.83 (11.40)
Female 22.58 (4.61) 6.56 (1.65) 65.55 (8.42)
Oral paragraph Male 21.59 (4.85) 3.46 (1.23) 83.22 (8.45)
Female 22.23 (4.33) 4.00 (2.13) 85.22 (6.71)
Nasal sentences Male 47.03 (4.86) 9.75 (2.52) 85.74 (6.37)
Female 50.41 (3.94) 11.31 (2.56) 87.80 (4.62)
Nasal paragraph Male 46.84 (5.23) 5.59 (2.14) 88.30 (5.58)
Female 49.74 (4.55) 6.29 (2.90) 90.02 (4.38)
The mean nasalance value exhibited by normal adult male and female subjects for
oral sentences was 22.48 % (4.79) and 22.58 % (4.61) respectively. The mean nasalance
value exhibited by male and female subjects for oral paragraph was 21.59 % (4.85) and
22.23% (4.33) respectively. The mean nasalance values exhibited by adult male and female
subjects for nasal sentences were 47.03% (4.86) and 50.41% (3.94) respectively. The mean
nasalance value exhibited by male and female subject for nasal paragraph was 46.84% (5.23)
and 49.74 (4.55) respectively. A significant difference was evident across gender and across
stimuli.
This difference may be attributed to the basic structural and functional differences
across genders. The resonance of voice is influenced by the size, shape, and surface of the
intraglottal and supraglottal resonating structures and cavities (Shprintzen and Bardach,
1995). The mechanism for velopharyngeal valving has been found to be different for men and
women. Mckerns and Bzoch (1970) suggested that velar length is greater in men, the height
of elevation is greater and the inferior point or contact is most usually above palatal plane. In
the female similar results are not found. The other finding that supports the present result is
the acoustic transmission of palate. As the age increases, the sympathetic transfer of acoustic
energy from oral cavity to the nasal cavity also increases in females (Hoit et al., 1994). It has
been found that females had higher nasalance value in both categories of stimuli. The result
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
198
can also be attributed to increased respiratory effort and increased nasal cross-sectional area
in female (Seaver et al.,1991; Van Lierde et al., Fletcher,1978; Hutchinson,1978).
Overall the present study adds to the body of evidence that there are gender
differences in nasalance values. Clinically the normative data reported in the present study
may help to identify and treating individuals with resonance disorders.
References
Awan, S. N., Daniel, Z. H. & Jordan, R.G. (2001). Dr.Speech User‟s Guide, Version 4.Tiger
DRS, Inc.
Awan, S.N. (1997). Analysis of nasalance: NasalView (the Nasalance Acquisition System)
In: Ziegler W, Deger K, eds. Clinical Phonetics and Linguistics, (London: Whurr-
Publishers), pp. 522-531.
Awan, S.N. (1998). Analysis of nasalance: NasalView. In: Ziegler W, Deger K, eds. Clinical
Phonetics and Linguistics. London: Whurr-Publishers. 1997; 518-525.
Baker, D. J. P., & Rose, G. (1984). Epidemiology in Medical Practice. 3rd
ed. Edinburgh:
Churchill Livingstone.
Bressmann, T. (2005) Comparison of nasalance scores obtained with the nasometer, the
nasalview and the oronasal system. Cleft Palate Craniofacial Journal 42, 423-433.
Bressmann, T., Klaiman, P., Fischbach, S. (2006). Same noses, different nasalance scores:
Data from normal subjects and cleft palate speakers for three systems for nasalance
analysis. Clinical Linguistics & Phonetics, 20, 163-17.
Dalston, R. M., Neiman, G. S., & Gonzalez-Landa, G. (1993). Nasometric sensitivity and
specificity: a cross- Dialect and cross-culture study. Cleft Palate Craniofacial
Journal, 30, 285-291.
Dalston, R. M., Warren, D., & Dalston, E. (1991a). A primary investigation concerning use
of Nasometery in identifying patients with hypanasality and/or nasal airway
impairment. Journal of Speech and Hearing Research, 34, 11-18.
Fairbanks, G., 1960, Voice and Articulation Drillbook, 2nd Edition (New York: Harper and
Bros.).
Fant, G. (1960). “Acoustic theory of speech production”. Mournton and Co., S. Gravehage.
Fletcher, S. G., & frost, S. D. (1974). Qualitative and graphical analysis of prosthetic
treatment for “nasalance” in speech. Journal of Prosthet. Dent. 32, 284-249.
Fletcher, S.G. (1978). Diagnosing speech disorder from the cleft palate. New York : Grune &
Stratton.
Haapanen, M.L. (1991a), “A simple clinical method of evaluating perceived nasality", Folia
Phoniatr. 43 (3), 122-132.
Normative Nasalance Value in Hindi
199
Hoit, J. D., Watson, P. J., Hoxon, K. E., McMahon, P., & Johnson, C. L. (1994). Age and
velopharyngeal function during speech production. Journal of Speech and Hearing
Research, 37, 295-302.
Hutchinson, J. M. , Robinson, K.L., & Nerbonne, M.A. (1978). Pattern of nasalnace in a
sample of normal gerontologic subjects. Journal of Communication Disorders, 11,
469-481.
McWilliams, B.J., Glaser, E.R., Philips, B. J., Lawrence, C., Lavorato, A. S., Eery, B. C., &
Skolnick, M. L. (1981). A comparative study of four methods of evaluating
velopharyngeal adequacy. Plastic and Reconstructive Surgery, 68, 1-9.
Moll, K. L. (1964). „Objective‟ measure of Nasality. Cleft Palate Journal, 1, 371-374.
Nasometer Manual, Kay Elemetrics, Pine Brook, New Jersy, 2003 edition.
Seaver, E.J., Dalston, R.M., & Leeper, H.A. (1991). A study of nasometric value for normal
nasal resonance. Journal of Speech and Hearing Research, 34, 715-721.
Shprintzen, R., & Bardach, J., (1995). Cleft palate speech management: A multidisciplinary
Approach. Missouri: Mosby.
Sweeney, T., Sell, D., & O‟ Regan, M (2004). Nasalance scores for normal speaking Irish
children. Cleft Palate Craniofacial Journal, 41 (2), 168-174.
Trindade, I. E. K., Genero, K. F., & Dalston, R. M.(1997). Nasalance score for normal
Brazilizn Portuguese Speakers. Brazil Journal of Dysmorphollogy Speech & Hearing
Disorders, 1, 23-34.
Van Doorn, J., & Parcell, A., (1998). Nasalance levels in the speech of normal Australian.
Van Lierde, K. M., Wuyts, F. L., De Bodt, M., & Van Cauwenberge, P. (2001). Normative
value for normal nasal resonance in the speech of young Flemish adults. Cleft Palate
Craniofacial Journal, 38, 112-118.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
200
Metalinguistic Abilities in Children with Developmental Dyslexia:
Implications for Reading and Writing
Priya M. B. & R. Manjula*
Abstract
The contribution of metalinguistic abilities to reading and writing is well established. These
skills are reported to be poor in children with developmental disabilities, particularly children with
developmental dyslexia. The study aimed to investigate the metalinguistic abilities and its contribution
to reading and writing in bilingual-biliterate (Kannada-English) typically developing children and
children with developmental dyslexia. 20 children with developmental dyslexia in the age range of 8-
13 years and equal number of typically developing children matched to the language age and gender
of those in the experimental group participated in the study. Subsections from Linguistic Profile Test
in Kannada (Karanth, 1980) and Reading Acquisition Profile in Kannada (Prema, 1997) were used to
assess metaphonology, metasemantics, metasyntax, reading and writing abilities in the two groups of
subjects. Results revealed poor performance of children with developmental dyslexia on all the
metalinguistic, reading and writing tasks. The metalinguistic skill that significantly contributed to
reading and writing skills in typically developing children was metaphonology whereas
metasemantics was identified as the skill that contributed significantly to reading and writing in
children with developmental dyslexia. The pattern of errors on all tasks was similar in the two groups
of children. However, the performance of children with DD was found to be similar to the younger TD
children on most of the tasks.
Introduction
Metalinguistic ability is the ability to reflect upon and manipulate the structural
features of spoken language, treating language itself as an object of thought (Tunmer, Pratt &
Herriman, 1984). Tunmer and Bowey (1984) identified four levels of metalinguistic
awareness: word awareness, phonological awareness, form awareness and pragmatic
awareness. They hypothesized that these levels play a vital role at different stages of reading
acquisition. Gombert (1992) categorized metalinguistic awareness into six groups:
metaphonological, metasyntactic, metalexical, metasemantic, metapragmatic, and
metatextual.
The development of metalinguistic ability in children is a metacognitive skill that
emerges towards the end of preschool period and is characterized by a cognitive shift in
intellectual functioning when a child can begin to treat language as an object of thought. This
ability to reflect on language has been attributed to emergence of the Piagetian stage of
concrete operations that begins to develop between 5 and 7 years of age (Van Kleeck, 1984).
Middle elementary school seems to be a pivotal period in both learning to read and
developing metalinguistic skill. Jarmulowicz, Hay, Taran and Ethington (2008) proposed a
developmental sequence beginning with receptive language followed by phonological
_________________________________ * Professor of Speech Pathology, All India Institute of Speech and Hearing, Mysore, India
email: [email protected]
Metalinguistic Abilities in Children with Developmental Dyslexia
201
awareness, morphological awareness, and a new metalinguistic task measuring oral
morphophonological accuracy, followed by decoding and culminating in reading
comprehension.
Research has shown that phonemic awareness, a crucial part of phonological
awareness is both a prerequisite and a consequence of learning to read. Grammatical
awareness has also been reported to influence reading comprehension (Bentin, Deutsch, &
Liberman, 1990). There have been several studies documenting the metalinguistic abilities in
children with various language impairments (Kamhi, Lee & Nelson, 1985; Boudreau &
Hedberg, 1999 and others). Mattingly (1972), Hodgson (1992) and others emphasize that the
metalinguistic processes, especially the metaphonological skills need to be paid more
attention to in the identification and management of reading disabled children. Children with
language disorders have shown a lack of syntactic awareness (Nation & Snowling, 2000) and
morphological awareness (Carlisle, 1987).
It has been reported that children with dyslexia appeared to be primarily impaired in
phonological and orthographic processing, rapid automatized naming, and executive
functions but to have intact oral language skills for morphology and syntax, that is, good
metalinguistic awareness at those levels of language (Berninger, 2006). However, children
with language learning disability appeared to be impaired in oral language skills,
phonological skills and reading comprehension than the dyslexics (Butler & Silliman, 2002).
Their impaired metalinguistic awareness of morphology and syntax was accounted for the
lower verbal IQs. Siegel and Ryan (1988) reported that reading disabled children scored
lower on measures of syntactic awareness than age-matched normal readers.
Bilingualism, which has been shown to affect metalinguistic abilities, influences
reading performance. The consensus in the field is that learning a second language permits
children to view their language as one system among others, thereby enhancing their
linguistic awareness. Bilingual children also outperform monolingual children on some
metalinguistic and emergent literacy tasks (Bialystok, Shenfield, & Codd, 2000).
Metalinguistic and reading abilities are language and script specific in nature and
hence investigations in the respective language and scripts are called for. Indian scripts
developed from Brahmi that are semi-syllabic in nature, is said to have highly transparent
orthographies. Studies in Kannada language on metaphonology and reading abilities
contradicted the hitherto accepted notion that metaphonological abilities are prerequisites for
the acquisition of reading (Rekha, 1996; Prakash, Rekha, Nigam & Karanth, 1993; Prakash,
Chandana & Suma, 2001). Prema (1997) has profiled the reading acquisition of children from
grade III to grade VII and reported that the hierarchy of predictors of reading disability in
Kannada are metasemantic, metasyntactic and metaphonological skills. Sharma (2000),
Cutinho (2000) and George (2001) studied children with learning disability in Hindi, English
(native language being Kannada) and Malayalam respectively and reported poor
metalinguistic abilities in these children.
Schwartz, Geva, Share and Leikin (2007) studied the cross linguistic transfer of
phonological processing skills in learning to read English as a third language (L1 and L2
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
202
being Russian and Hebrew respectively). The results revealed that cross-linguistic transfer of
early literacy skills can be found even in the context of different alphabetic orthographies.
The authors further reported that the outcomes of the study suggest that the actual mechanism
of transfer of early literacy skills across alphabetic orthographies is the interaction between
the generalized insight into the alphabetic principle and the specific benefits of knowledge of
an orthography characterized by fully-fledged alphabet with letters representing consonants
and vowels (i.e. Russian) in the acquisition of the another alphabet such as English.
Need for the study
It has been well documented that children with developmental dyslexia have reading
as one of their core deficits and various components that are necessary for the acquisition of
reading are affected. Studies in the area of metalinguistic abilities in bilingual-biliterate
(Kannada-English) children with developmental dyslexia are limited. Further, the relationship
between different metalinguistic skills (metasemantics, metasyntax, metaphonology) and
literacy have not been addressed in this population.
Thus, it is of interest to study the hierarchy of skills that contribute to the acquisition
of reading and writing in Kannada-English bilingual-biliterate typically developing children
and children with developmental dyslexia, which will have implications in the management
of these children.
Aims of the study
The study was undertaken with the following aims:
To compare the performance of bilingual-biliterate (Kannada-English) typically
developing children and children with developmental dyslexia across the major
domains of Metalinguistics, Reading and Writing.
To compare the performance of bilingual-biliterate typically developing children and
children with developmental dyslexia across the sub-domains of Metalinguistic
components, Reading and Writing.
To study the correlation of Reading and Writing with the Components of
Metalinguistic Skills.
To determine the Metalinguistic Skills that contributes significantly to the acquisition
of Reading and Writing Abilities in bilingual-biliterate typically developing children
and children with developmental dyslexia.
To compare the pattern of errors on Metalinguistic, Reading and Writing tasks based
on Qualitative Analysis in the two groups of children.
Method
Participants
The participants were classified into experimental and control groups.
Metalinguistic Abilities in Children with Developmental Dyslexia
203
Experimental group: A total of twenty children (18 males & 2 females) with developmental
dyslexia in the age range of 8-13 years (mean age: 10 years 6 months) constituted the
experimental group.
Control group: Equal number of typically developing children (mean age: 9 years 2 months),
matched for gender and language age of the subjects in the experimental group constituted
the control group.
Subject Selection Criteria
All the participants spoke Kannada as their native language and were studying in
schools with English as the medium of instruction. There was no change in the medium of
instruction at any time for any of the participants. Participants and/or parents were explained
about the purpose of the study and an informed written consent was taken.
Experimental group
Children in the age range of 8-13 years diagnosed as having developmental dyslexia.
The diagnosis of developmental dyslexia was based on the performance in the test of
Early Reading Skills (norms developed by Prema & Jayaram, 2002) as assessed by a
qualified Speech Language Pathologist and assessment by a Clinical Psychologist.
Children with developmental dyslexia who have attended therapy for not more than 6
months.
Children with additional disabilities like ADHD, stuttering, misarticulation or any
other neurological deficits were excluded from the study.
Control group
Participants in this group were screened using the WHO Ten Question Disability
Screening Checklist (cited in Singhi, Kumar, Prabhjot & Kumar, 2007 - Refer Appendix 1) to
rule out:
Learning disability
Language deficits
Delayed speech and language milestones,
Hearing impairment
Mental retardation
Behavioural and emotional disorders
Neurological deficits.
Procedure
The following tests were carried out individually for all the participants (experimental
and control subjects) in a quiet environment (Refer to Table 1).
Table 1: Tests administered on the participants of the study
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
204
Tests Purpose
WHO Ten Question Disability
Screening checklist
(cited in Singhi, Kumar, Prabhjot &
Kumar, 2007)
Linguistic Profile Test in Kannada
(Karanth, 1980)
Reading Acquisition Profile in
Kannada (RAP-K) (Prema, 1997)
Test for Metaphonological skills
Reading Tests
Writing Tests
To rule out any disability in control
group
To rule out disabilities in areas other
than language in experimental group
To assess the language age,
metasemantics and metasyntactic
abilities.
To assess the metaphonological, reading
and writing skills.
The tests were administered on all the participants by the investigator. The
participants in the control group were tested individually in a quiet environment in the school
setting and the participants in the experimental group were tested in a quiet environment in
the clinical setting. The testing was carried out in 2-3 sittings (on consecutive days)
depending on the comfort level of the participants. The total time taken for testing each
participant ranged between 2 ½ to 3 hours. The order of the tasks was randomised across
subjects and across groups to rule out order effect.
Data Analysis
The scores obtained in each of the domains were computed and tabulated. The data
was then subjected to suitable statistical measures and analyzed quantitatively. Qualitative
analysis of the data was also carried out to determine the pattern of errors in both the groups
of subjects. The results of the analysis are presented and discussed in the sections that follow.
Results and Discussion
The performance of children with DD and the language age matched TD children will
be presented under the following sections:
I. Performance of children across the major domains of Metalinguistics, Reading and
Writing.
II. Performance of children across the sub-domains of Metalinguistic components,
Reading and Writing.
III. Correlation of Reading and Writing with the Components of Metalinguistic Skills.
IV. Metalinguistic Skills that contribute to the acquisition of Reading and Writing
Abilities.
V. Qualitative Analysis
Metalinguistic Abilities in Children with Developmental Dyslexia
205
I. Performance of children across the major domains of Metalinguistics, Reading and
Writing
The mean percent and standard deviation (SD) values for tasks assessing
metaphonology, metasemantics, metasyntax, reading and writing skills for the two groups of
subjects are shown in Table 2 and figure 1. The results from Table 2 and figure 1 reveal that
the TD group performed better than DD on all the tasks under study.
Figure 1: Performance of the two groups of subjects on metalinguistic, reading and writing
tasks.
Table 2: Percent Mean and Standard Deviation (SD) for the two groups of subjects on
metalinguistic, reading and writing tasks
Groups Percent Mean SD
Metaphonology TD 79.16 7.89
DD 49.65 13.70
Metasemantics TD 89.52 6.04
DD 70.67 8.33
Metasyntax TD 88.60 5.82
DD 56.90 11.89
Reading TD 85.38 8.62
DD 40.25 20.25
Writing TD 82.13 9.26
DD 24.33 14.64
Statistical analysis using mixed ANOVA revealed a significant effect of test [F (4,
152) = 82.034, p < 0.001] and groups [F (1, 38) = 147.369, p < 0.001] in the performance of
the two groups of subjects. Pair-wise comparisons using Bonferroni‘s multiple comparison
showed a significant difference between all the domains (p < 0.001) except for
metaphonology and reading (p > 0.05). The results of independent ‗t‘ test carried out to
compare the performance across the two groups of subjects revealed significant differences
between TD and DD groups for all the five domains (p < 0.001). This suggests that the
performances of the two groups of subjects on the metalinguistic, reading and writing tasks
are different.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
206
The finding that children with DD perform poorly than the TD children on
metalinguistic tasks are in consonance with literature reported on metalinguistic abilities in
children with language disorders (Kamhi at al., 1985; Carlisle, 1987; Cutinho, 2000). Cutinho
(2000) reported poor metalinguistic skills in English in children with Learning Disability
whose native language was Kannada whereas the present study revealed poor metalinguistic
skills in Kannada language in Kannada-English bilingual-biliterates. Thus, children with DD
may be thought to have metalinguistic deficits in both the languages. However, the nature of
the tasks used in both the studies was different.
The findings of the present study are also in consonance with that of Prakash et al.
(1993), and Karanth and Suchitra (1993), whose findings indicate that acquisition of reading
(literacy) itself facilitates metaphonological and metasyntactic abilities, rather metalinguistic
abilities. The finding of significantly poor performance of children with DD on reading and
writing tasks are in agreement with Prakash et al. (2001) who reported that the dyslexic
children in Kannada were very poor on their orthographic awareness. This could be attributed
to the poor mental representation in these children of how phonology and orthography are
interrelated and expressed in the script.
II. Performance across the sub-domains of Metalinguistic components, Reading and
Writing.
The performance across the sub-domains for the two groups of subjects will be
presented under the following headings:
1. Metaphonology
2. Metasemantics
3. Metasyntax
4. Reading
5. Writing
For each of the domains, statistical analyses were carried out using mixed ANOVA
for the sub-domains with groups as the independent variable. This was followed by
independent ‗t‘ test to compare the performance across the two groups of subjects and
repeated measures ANOVA was carried out across the tests within each subject group. The
results of the statistical analysis reveal that the TD group performed better than DD on all the
subdomains of metaphonology, metasemantics, metasyntax, reading and writing. These are
shown in figures 2, 3, 4, 5 and 6 respectively.
Metalinguistic Abilities in Children with Developmental Dyslexia
207
Figure 2: Performance of the two groups of subjects on the sub-domains of metaphonology.
Figure 3: Performance of the two groups of subjects on the sub-domains of metasemantics.
Figure 4: Performance of the two groups of subjects on the sub-domains of metasyntax.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
208
Figure 5: Performance of the two groups of subjects on the sub-domains of reading.
Figure 6: Performance of the two groups of subjects on the sub-domains of writing.
The metaphonological deficits observed in children with DD are supported by a vast
majority of studies in the literature which document poor phonological awareness in children
with reading disability (Tunmer & Bowey, 1984; Kamhi et al., 1985 and others).
Metaphonological deficits in these children have also been documented in the syllabic scripts
(Padakannaya, Rekha, Vaid & Joshi, 2002). Of the metaphonological tasks, poor performance
was found on phoneme deletion in both the groups of children. This supports the notion that
phoneme awareness develops at a later age when compared to the other metaphonological
tasks. Exposure to English facilitated better performance on phoneme awareness tasks
(phoneme deletion and phoneme oddity) in TD children which is in consonance to that
reported by Prakash and Rekha (1992). However, the same was not observed in children with
DD which might suggest that these children were unable to integrate the principles of
alphabetic script and the knowledge of orthographic principles of semi-syllabic script, which
is considered to be essential for cross linguistic transfer of metaphonological tasks (Schwartz
et. al, 2007).
The performance of both groups of children on tasks assessing metasemantic and
metasyntactic skills draws support from the results of investigations on these abilities in
Metalinguistic Abilities in Children with Developmental Dyslexia
209
alphabetic scripts (Bentin et al., 1990) as well as semi-syllabic scripts (Sharma, 2000;
George, 2001). While the reported literature is for monolingual-monoliterates, the present
study revealed similar findings in bilingual-biliterate children for the native language. The
performance on the sub-domains of metasemantics and metasyntax of the two groups of
children in Kannada is similar to that reported by Sharma (2000) and George (2001) in Hindi
and Malayalam languages respectively.
The results also support the view that judgment tasks are easier than revision tasks
and is the first of the metasyntactic skills to develop since the younger typically developing
children were able to judge the grammaticality of the sentence but unable to revise them.
III. Correlation of Reading and Writing with the Components of Metalinguistic Skills
Pearson‘s correlation was done to determine the correlation of combined reading and
writing abilities with that of metaphonology, metasemantics and metasyntax independently in
TD and DD groups. In the TD group, the correlation of reading and writing was found to be
significant with metaphonology and metasemantics but not for metasyntax. The correlation
co-efficients for the combined reading and writing abilities with that of the metalinguistic
components for the TD group is shown in Table 3.
Table 3: Correlation co-efficients for TD group
Total
Reading &
Writing
Meta-
phonology
Meta-
semantics
Meta-
syntax
Total
Reading
& Writing
Pearson Correlation 1.000 0.591** 0.489* 0.440
Sig. (2-tailed) - 0.006 0.029 0.052
N 20 20 20 20
Note: **- p<0 .01 (2-tailed), *- p<0 .05 (2-tailed).
Table 4: Correlation co-efficients for DD group
Note: **- p<0 .01 (2-tailed), *- p<0 .05 (2-tailed).
The trend of correlation in the DD group was found to be different from that of the
TD group. In the DD group, the correlation of the combined reading and writing skills was
found to be significant with metasemantics, metaphonology and metasyntax in that order. The
correlation co-efficients for the combined reading and writing abilities with that of the
metalinguistic components for the DD group is shown in Table 4.
Total
Reading &
Writing
Meta-
phonology
Meta-
semantics
Meta-
syntax
Total
Reading &
Writing
Pearson Correlation 1.000 0.660** 0.704** 0.558*
Sig. (2-tailed) - 0.002 0.001 0.011
N 20 20 20 20
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
210
IV. Metalinguistic Skills that contribute to the acquisition of Reading and Writing
Abilities.
A step-wise multiple regression analysis was performed separately for each subject
group to determine the potential variables which contribute to the acquisition of reading and
writing skills. The results of step-wise multiple regression for the TD group revealed a
regression equation for the combined reading and writing abilities which was significant with
F (1, 18) = 9.641, p < 0.01. The regression equation included only metaphonology whereas
metasemantics and metasyntax were excluded. Of the metalinguistic skills, metaphonology
was identified as the potential contributor to the acquisition of reading and writing in the TD
group (r2=0.349; Metaphonology: β =3.111; p < 0.01) and constant was significant with value
124.237 (p < 0.05).
Step-wise multiple regression for the DD group resulted in a regression equation for
the combined reading and writing abilities which was significant with F (1, 18) = 17.736, p <
0.001. The regression equation included only metasemantics whereas metaphonology and
metasyntax were excluded. The results revealed that metasemantics was the potential
contributor to the acquisition of reading in the DD group (r2=0.496; Metasemantics: β
=6.119; p < 0.001) and constant was significant with value -251.337 (p < 0.05).
Prema (1997) reported that the hierarchy of predictors of reading abilities in Kannada
monolingual–monoliterates was found to be metasemantics, metasyntactic and
metaphonology. In the present study, the contribution of metaphonology to reading and
writing abilities in bilingual-biliterate (Kannada- English) TD children was significant.
Metasemantic and metasyntactic abilities may not have a significant contribution to reading
and writing in this population. Thus, the skills contributing to reading and writing in the
native language are different in monolingual-monoliterate and bilingual-biliterate children.
The exposure to an alphabetic script can be attributed to this finding in children who can
speak as well as read and write in two languages.
Alternatively, the metalinguistic skill which contributed significantly to reading and
writing Kannada in children with DD was found to be metasemantics, rather than
metaphonology as was the case in TD children. This finding in children with DD is similar to
that reported by Prema (1997) on TD children for monolingual-monoliterate children. This
might suggest that bilingual-biliterate children with DD perform similar to monolingual-
monoliterate children and there may be no significant cross linguistic transfer of
metalinguistic skills in this population. Thus, the metalinguistic abilities that contribute
significantly to reading and writing may be different in TD and DD groups.
V. Qualitative Analysis
The performance on the metasyntactic tasks was found to be highly influenced by the
spoken form of the language. The differences in the colloquial and pedantic forms of
Kannada, particularly with the minute morphosyntactic violations being acceptable in the
colloquial form, could be attributed to the inability of the subjects to identify certain
grammatical violations in the pedantic form.
Metalinguistic Abilities in Children with Developmental Dyslexia
211
It could thus be speculated that the DD group exhibited a persisting logographic type
of whole word reading, relying on semantic cues when non-word stimuli are presented.
However, they also tended to ignore the semantic cues while reading words resulting in
erroneous reading of words. The errors on writing were similar to those observed for reading
including substitution of unaspirated sounds for aspirated sounds and auditorily similar
sounds.
Thus, on the whole, no significant qualitative differences were observed on these
tasks between TD children and children with DD. However, the performance of children with
DD was found to be similar to the younger TD children on most of the tasks. These findings
are in agreement with that of Ramaa (1985), Karanth (1990) and Share (1996) who reported
that the error patterns of older poor readers suggest that they use the same strategies as
younger, normally achieving children. Children with DD were found to have greater
problems on non words and also a general delay along the developmental sequence of reading
and writing. Their errors ranged from inability to identify syllables and words to misreading,
slow reading, substitution of visually and/or auditorily similar syllables and words.
Conclusions
Thus, the present study revealed significant differences between bilingual-biliterate
typically developing children and children with developmental dyslexia on metalinguistic and
literacy skills. The significant contribution of metaphonology to the acquisition of reading
and writing and also a higher correlation between the two in a semi-syllabic script in typically
developing children exposed to an alphabetic script simultaneously, implies the cross
linguistic transfer of these skills across different orthographic systems (alphabetic and semi-
syllabic in the present study). In contrast, the contribution of metasemantics was higher in
children with developmental dyslexia leading to the speculation of persistent logographic
type of reading in this population. Thus, the metalinguistic skills were found to have a
significant role in the acquisition of reading and writing in both the groups of children.
Overall, the performance of children with developmental dyslexia resembled that of younger
typically developing children on all tasks of metalinguistics, reading and writing.
References
Bentin, S., Deutsch, A., & Liberman, I. Y. (1990). Syntactic competence and reading ability
in children. Journal of Experimental Child Psychology, 48, 147-172.
Berninger, V. W. (2006). A developmental approach to learning disabilities. In W. Damon, &
R. M. Lerner (Eds.), Handbook of child psychology (6th
ed., Vol. 4: Child Psychology
in Practice, pp. 420-452). New York, NY: John Wiley & Sons, Inc.
Bialystok, E., Shenfield, T., & Codd, J. (2000). Languages, scripts, and the environment:
Factors in developing concepts of print. Developmental Psychology, 36, 66-76.
Boudreau, D., & Hedberg, N. (1999). A comparison of early literacy skills in children with
specific language impairment and their typically developing peers. American Journal
of Speech-Language Pathology, 8, 249-260.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
212
Butler, K., & Silliman, E. (2002). Speaking, reading and writing in children with language
learning disabilities. Mahwah NJ: Erlbaum.
Carlisle, J. F. (1987). The use of morphological knowledge in spelling derived forms by
learning-disabled and normal students. Annals of Dyslexia, 37, 90-108.
Cutinho, J. (2000). Early metalinguistic skills: A comparative study between normal and
learning disabled children. Unpublished Master‘s Dissertation submitted to the
University of Mangalore, Mangalore.
George, N. (2001). Language skills in Malayalam speaking learning disabled children.
Unpublished Master's Dissertation submitted to the University of Mangalore,
Mangalore.
Gombert, J. E. (1992). Metalinguistic development. Chicago: University of Chicago Press.
Hodgson, J. (1992). The status of metalinguistic skills in reading development. Journal of
Learning Disabilities, 25(2), 96-101.
Jarmulowicz, L., Hay, S. E., Taran, V. L., & Ethington, C. A. (2008). Fitting derivational
morphophonology into a developmental model of reading. Reading and Writing: An
Interdisciplinary Journal, 21, 275-297.
Kamhi, A., Lee, R., & Nelson, L. (1985). Word, syllable, and sound awareness in language
disordered children. Journal of Speech and Hearing Disorders, 50, 207-212.
Karanth, P. (1980). Linguistic Profiles Test in Kannada. Indian Council of Medical Research
Project, India.
Karanth, P., & Suchitra, M. G. (1993). Literacy acquisition and grammaticality judgement in
children. In R. J. Scholes (Ed.), Literacy and language analysis (pp. 143-155).
London: Lawrence - Erlbaum Associates.
Mattingly, I. (1972). Reading, the linguistic process, and linguistic awareness. In J.
Kavanagh, & I. Mattingly (Eds.), Language by ear and by eye. Cambridge,
Massachusetts: MIT Press.
Nation, K., & Snowling, M. J. (2000). Factors influencing syntactic awareness skills in
normal readers and poor comprehenders. Applied Psycholinguistics, 21, 229-241.
Padakannaya, P., Rekha, D., Vaid, J., & Joshi, M. (2002). Simultaneous acquisition of
literacy skills in English and Kannada: A longitudinal study. Paper presented at 13th
World Congress of Applied Psycholinguistics (AILA), Singapore (December).
Prakash, P., Rekha, D., Nigam, R., & Karanth, P. (1993). Phonological awareness,
orthography, and literacy. In R. J. Scholes (Ed.), Literacy and language analysis (pp.
55–70). Hillsdale, New Jersey: Lawrence Erlbaum Associates.
Prakash, P., Chandana, M. V., & Suma, S. (2001). Orthographic awareness, phonemic
awareness and developmental dyslexia. Paper presented at the Society for Scientific
Study of Reading (SSSR) conference (May 29-June 3), Colorado, U.S.
Metalinguistic Abilities in Children with Developmental Dyslexia
213
Prakash, P., & Rekha, B. (1992). Phonological awareness and reading acquisition in
Kannada. In A. K. Srivastava (Ed.), Researches in child and adolescent psychology
(pp. 47-52). New Delhi, India: NCERT.
Prema, K. S. (1997). Reading Acquisition Profile in Kannada. Unpublished Doctoral
dissertation submitted to the University of Mysore, Mysore.
Prema K. S. & Jayaram, M. (2002). Investigation of reading skills in children with reading
disability: Indian perspective. Unpublished project report. Project funded by AIISH
Research Fund (ARF/Sp. Path-2/KSP/15-09-2001 to 26-07-2002), All India Institute
of Speech and Hearing, Mysore, India.
Rekha, D. (1996). Reading acquisition and metaphonological awareness: A longitudinal
study. Unpublished Doctoral dissertation submitted to the University of Mysore,
Mysore.
Schwartz, M., Geva, E., Share, D. L., & Leikin, M. (2007). Learning to read in English as
third language: The cross-linguistic transfer of phonological processing skills. Written
Language & Literacy, 10, 25-52.
Sharma, M. (2000). Language skills in children with learning disability. Unpublished
Master‘s Dissertation submitted to the University of Mangalore, Mangalore.
Siegel, L. S., & Ryan, E. B. (1988). Development of grammatical-sensitivity, phonological
and short-term memory skills in normally achieving and subtypes of learning disabled
children. Developmental Psychology, 24, 28-37.
Tunmer, W. E., & Bowey, J. A. (1984). Metalinguistic awareness and reading acquisition. In
W. E. Tunmer, C. Pratt, & M. L. Herriman (Eds.), Metalinguistic awareness in
children: Theory, research, and implications (pp. 144-168). Berlin: Springer-Verlag.
Tunmer, W. E., Pratt, C., & Herriman, M. L. (1984). Metalinguistic awareness in children.
New York: Springer – Verlag.
Van Kleeck, A. (1984). Assessment and intervention. Does ―meta‖ matter? In E. Wallach, &
K. G. Butler (Eds.), Language learning disabilities in school-age children. (pp. 179-
199). Baltimore: Williams & Wilkins.
WHO Ten Questions Screen. In Singhi, P., Kumar, M., Prabhjot, M, & Kumar, R. (2007).
Utility of the WHO Ten Questions Screen for Disability Detection in a rural
community—the north Indian experience. Journal of Tropical Pediatrics, 53(6): 383-
387.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
214
Language Proficiency Questionnaire:
An Adaptation of LEAP-Q in Indian Context
Ramya Maitreyee & S.P.Goswami*
Abstract
The prime focus of the study was to modify and adapt the LEAP-Q to the Indian scenario. The
modified version of the questionnaire was made and given to five SLPs working in the area of
bilingualism for their suggestions. It was piloted on a group of 10 Kannada- English bilinguals.
Based on their difficulty level to respond to the questions, further modifications were incorporated
and the final questionnaire were administered on 60 Kannada/Hindi- English bi/multilinguals. The
responses were tabulated and subjected to statistical analysis using the SPSS software. The results
indicated that L1 was the primary language of communication during childhood. All the participants
had acquired L1 prior to L2 and then followed by L3. The understanding and speaking skills were
acquired earlier than the literacy skills in L1. However, all the four skills developed in a simultaneous
manner for L2 and L3. The attainment of proficiency levels in all the language followed a similar
pattern as in the acquisition. The proficiency ratings for understanding, speaking, reading and
writing obtained did not indicate a one to one correlation between the competence and performance
levels for L1. However, it was seen that if the participants rated themselves to be proficient like in
using the language then there actual performance was either native-like or good proficiency, but not
below that. Hence, this finding indicates that a correction factor needs to be employed while judging
the actual performance. Conversely, not much discrepancy was obtained between the competence and
performance ratings for L2. The findings of the various sub-sections indicate that this tool will help us
in assessing the proficiency level of a bi/multilingual participant in any of the languages in four skills,
i.e., understanding, speaking, reading and writing. It would serve as a quick measure to assess the
competence levels of an individual.
Introduction
India is a multilingual country and houses a number of languages, culture, religion
and society. Bi/multilingualism is widely spread in India making it the tower of babbles. In
India, bilingualism is one of the least understood phenomenons and there lies a dearth of
studies in this area. It has been viewed as an area too complex to investigate, considering the
multilingual mosaic of India. At times it becomes difficult on part of the researchers to
adequately assess the proficiency level of an individual in the skills of understanding,
speaking, reading and writing especially in a language they do not speak and thus, have to
rely on self-assessed information. A number of factors have been found to be responsible for
attaining proficiency in a language. The age of acquisition, the environment where it is learnt,
the culture and socioeconomic status of an individual plays a vital role in the proficiency of
languages. The complex nature of bilingualism in India has led us to adapt easily to the
western models blindly and as a result have failed to appreciate the multiplicity of Indian
languages as a beautiful aspect of this country.
______________________________________
* Reader in Speech Pathology, All India Institute of Speech and Hearing, Mysore, India email: [email protected]
Language Proficiency Questionnaire in Indian Context
215
Thus, the tests developed in western context may not be viable to be used in the Indian
context. Hence, there is a need to modify the tests according to our diversified multilingual
situation which would further prove to be a more valuable tool.
Aim of the study
The aim of the present study was to modify and adapt the LEAP-Questionnaire to the
Indian multilingual situation.
Review of Literature
Bilingualism has been extensively viewed as the equal mastery of two languages.
Different researchers have defined it in different ways. However, the core aspect remains the
same. Bloomfield (1933) defined it as the native-like control of two languages, Haugen
(1953) as the ability to produce complete meaningful utterances in the other language.
Innumerable definitions have led to the emergence of a number of typologies or classification
systems to describe various types of bilingualism.
Types of bilingualism
The different classificatory system as cited in Thirumalai and Shyamala (1986) are
listed below:
Compound and Coordinate Bilingualism
Simultaneous versus Sequential/Successive acquisition of two or more languages
Second language acquisition and Second language learning
Dominant versus Balanced bilingualism
Additive and Subtractive Bilingualism
All these classifications mentioned above which are either based on developmental or
contextual parameters of bilingualism indicate the necessity of emphasizing the process by
which bilingual proficiency is acquired in order to fully understand the nature of individual
bilingualism.
Bilingual proficiency
Bilingual proficiency refers to an individual’s ability in their two languages. The
following four skills form a core for attainment of bilingual proficiency:
Understanding ● Reading
Speaking ● Writing
Thus, a bilingual person needs to have proficiency in all the four dimensions of each
of the languages (L1 & L2). The four language abilities form an approximate ladder of
complexity. Listening would be the easiest to acquire, followed by speaking, reading and
writing. However, these four aspects are neither independent nor different skills. The
development of proficiency in one of the skills has an effect on the other.
Any individual might have an adequate knowledge of the language but while
performing might not be able to use this language to the fullest extent. Chomsky (1965) made
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
216
a fundamental distinction between these two concepts, which he referred to as competence
and performance. According to him, competence refers to the speakers-hearer’s knowledge of
language and performance to the actual usage of language in concrete situations.
Factors affecting language acquisition and proficiency
Extensive research has been carried out by the researchers on the influence of several
factors which affect the language proficiency of a person. These include:
a. Developmental Factors: Lenneberg (1967); Johnson and Newport (1989); Flege,
Mackay and Piske (2002); Marian, Blumenfeld and Kaushanskaya (2007); Li, 2009.
b. Educational and Environmental factors
Educational context: Collier (1995); Durgonuglu and Oney (2000); Flege and Liu
(2001).
Web-based learning: Sanaoui and Lapkin (1992); Chiswick and Miller (n.d.).
Exposure to language in the family/society: Chomsky (1957); Rosenberg (1996);
Delgado, Guerroro, Goggin and Ellis (1999), Cummins (2000) Hasson (2006),
Harniess (2008); Schmitt (2008); Tahta, Wood & Loewenthal (1981); Piske,
Mackay & Flege (2002).
Length of Residence (LOR)
Media-based learning: Johannessen and Lopez (2002)
Formal Methods for Assessing Language Proficiency
Many tests have been used to measure bilingualism. Macnamara (1969) grouped these
tests as:
I. Rating scales include:
● Self-rating scales ● Language background questionnaires
● Language usage rating scales ● Experimenter interviews
In self-rating, the bilingual is asked to rate his/her proficiency in each of the basic
skills in each language. A balance score is then calculated by subtracting the ratings of one
language from those of the other. If the difference is zero or close to zero, the bilingual is
considered to be equally fluent in both the languages.
The Language Efficiency and Proficiency Questionnaire (LEAP-Q) developed by Marian,
Blumenfeld and Kaushanskaya (2007), is a self-assessment tool that includes relevant
proficiency and experience variables in a single instrument. The internal validity of the
Language Experience and Proficiency Questionnaire (LEAP-Q) was examined in study 1
which was based on the self-report data. In Study 2 they established that self-reports were
reliable indicators of language performance. L1 proficiency was reported to be better than L2
proficiency. Family based experiences, years spent in a L1 country contributed to L1
competence and proficiency. Although, this was one of the first questionnaires which
evaluated proficiency in both the languages, yet it also has some limitations. It does not
include a detailed language history during the childhood and all the possible factors that
Language Proficiency Questionnaire in Indian Context
217
would contribute to language acquisition and proficiency directly or indirectly. It uses a wide
rating scale which can give us varied data. Keeping these shortcomings in mind the present
questionnaire is being modified and adapted to be used in the Indian scenario for adult
bi/multilinguals.
Method
Participants: A total of 60 participants were included in the present study.30 Hindi-English
bi/multilinguals (14 males & 16 females) and 30 Kannada-English bi/multilinguals (15 males
& 15 females)
Ethical procedure
Participants were selected by ethical procedures. They were explained the purpose and
procedures of the study, and an informed verbal consent was obtained from them.
Inclusion criteria
The educational qualification level of all the participants ranged from a minimum of 12 years
of education or higher in at least one of the languages. All the participants were native
speakers of Northern/Southern India and had acquired the first/second language (L1-
Hindi/English and Kannada/English in the present study) both for academic and
communicative purposes and had the knowledge of third language (L3- either Kannada or
Hindi). No obvious deficits of any neurological, psychological and or sensory were either
present or reported.
Age groups
Adult bilinguals in the age range of 18-25 years were included in the present study.
Procedure
The study was carried out in three stages. The three stages are as follows.
Stage I
A modified version of the already existing Language Efficiency and Proficiency
Questionnaire (LEAP-Q) was made. The questionnaire (Appendix I) contained a total of 18
questions. The first question was intended to collect details about the number of languages
known by each of the participants. The rest 17 questions with their sub-sections were more
relevant to assess proficiency in a language. The following domains were included in the
questionnaire:
Language history during childhood (3 questions)
Acquisition of language (3 questions)
Proficiency of language (4 questions)
Usage of language in different context (2 question)
Exposure to different languages (4 questions)
Native/non-native speaker (1 question)
The participants were asked to encircle the required language used with specific
people and situations. In obtaining information about the language acquisition age for each of
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
218
the languages (L1, L2 & L3), the participants were asked to give the approximate age of
acquisition and proficiency level reached for each of the languages.
Stage II
For the appropriateness of the questionnaire, it was distributed to five speech
language pathologists experienced in working with issues concerned with bilingualism for
their suggestions and the questionnaire was altered accordingly. A pilot study was also
carried out with ten adult bilinguals in the age range of 18-25 years recruited from All India
Institute of Speech and Hearing campus and based on their difficulty of answering the
questions the questionnaire was modified accordingly.
Stage III
At this stage the participants were given to fill in the final questionnaire. Along with
filling in the questionnaire the participants performed similar tasks in both the languages for
the four skills of understanding, speaking, reading and writing.
The following samples were collected:
1. Speaking
a. Picture description: The picture stimulus from the Western Aphasia Battery (Kertesz,
1982) was used in the present study. The participants were asked to describe the picture for
duration of 1 minute in both the languages, i.e., L1 and L2.
b. A spontaneous speech sample of 1 minute was also recorded from each participant.
2. Reading: Three standardized passages were used in the following study. The
standardized Kannada passage containing both voiced and unvoiced sounds was selected for
the study. The Hindi passage was selected from the AIISH Research Fund project entitled
“Speech Rhythm in Indo-Aryan and Dravidian languages”. The “Rainbow passage
(Fairbanks, 1960)” was used in order to study their proficiency in reading in L2 (English).
The samples were recorded on Wavesurfer 6.0 using a microphone, placed six inches
from the mouth of the participant, which was connected to a laptop.
3. Comprehension: A passage from the XII Standard English textbook of Karnataka
state board was chosen for the study. The original passage was retained and then it was
translated to both Kannada and Hindi. 5 questions were framed in order to assess the
comprehension abilities. The participants were asked to read the passage and answer the
questions. The same procedure was followed for both the languages, i.e., L1 and L2
4. Writing: The participants of the study were asked to write on “India- My country” in
both the languages (L1 and L2) and their performance was evaluated by three competent
speech language pathologists.
The samples collected were given to three competent speech language pathologists to
assess their proficiency levels in each of the four skills. The judges were asked to rate on 4-
point rating scale (1- Zero proficiency, 2- Low proficiency, 3- Good proficiency, 4- perfect
proficiency). This was carried out to find a co-relation between the competence and the actual
performance levels of each of the participants in the various language abilities.
Language Proficiency Questionnaire in Indian Context
219
Task
The participants were seated comfortably and they were instructed to read the
questions thoroughly and select the most appropriate option that best suits the question. The
instructions were also given in written form in the questionnaire.
The samples were collected in a quiet environment after taking consent from the
individual participants in the study.
Analysis of data
The data collected from 60 participants was subjected to quantitative analysis using
SPSS (16.0 version) software.
Results and Discussion
The results of the study are being presented under the following domains:
I. Language history during childhood
The overall statistical analysis revealed that in both the groups, i.e., Hindi and
Kannada participants used L1 to communicate with family members and neighbors in
childhood. A small percentage also accounted of using a combination of L1 and L2. These
findings could be due to the proficient and frequent usage of L1 by the family members. This
suggests that the nurture influences the acquisition of a language. Thus, this study advocates
that the stimulation provided by the environment has an impact during the childhood. Similar
views have been reported by the empiricists Delgado, Guerroro, Goggin and Ellis (1999) and
Li (2009).
The use of a combination of language was more prevalent in the Hindi group which
could be due to the environmental demands that vary from region to region in India. The
usage of English by the mothers of Hindi speaking participants may be due to the need of
adapting to the common language for globalization or higher educational achievements.
II. Acquisition of language
Learning of language
It was seen that L1 was the first language learnt for understanding and speaking. L2
was the first language used for reading and writing in Hindi group. L1 alone or a combination
of L1 and L2 was used by the Kannada-English bilinguals for reading and writing.
These findings could be attributed to the fact that the exposure of the participants
from the surrounding environment was maximum for L1. Thus, the usage of L1 in the early
stages of life for fulfilling the basic needs of an individual or for functional communication,
i.e. for understanding and speaking is widely used and accepted. Similar findings were
reported by Delgado, Guerroro, Goggin and Ellis (1999) and Li (2009).
However, the role of the second language comes into effect when an individual starts
acquiring literacy skills (reading and writing) during childhood. The findings of the present
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
220
investigation are in agreement with the study by Marian, Blumenfeld and Kaushanskaya
(2007).
Age of acquisition of the above skills in L1, L2 and L3
The acquisition of L1 for understanding and speaking is learnt first followed by
reading and writing for all the participants. It is obvious from the results that the basic skills
(understanding and speaking) to communicate are acquired at an earlier age (1 year),
however, the skills which require more precise integration of language and cognition for
literacy skills (reading and writing) are acquired at the age of 4 in both the groups.
All the skills develop in a parallel fashion in the acquisition of L2 and L3. L1 is
acquired first followed by L2 and L3. A significant difference was obtained in Hindi {2
(2) =
120, p<0.05}, and Kannada group {2
(2) = 202.31, p<0.05}, the data was further subjected
to Wilcoxon signed rank test. Results of this test indicated a significant difference between
L1 and L2 (z= 7.88, p<0.05); L2 and L3 (z= 9.14, p<0.05) and L1 and L3 (z= 9.53,
p<0.05) in the Kannada group. Similar findings were observed in Hindi group.
A significant difference was obtained on Mann-Whitney U test between Hindi and
Kannada groups for L1 writing (z= 2.876, p<0.05), L3 understanding (z= 5.873, p<0.05)
and L3 speaking (z= 5.875, p<0.05). The Hindi group attained understanding and speaking
L3 at a later age than the Kannada group.
The results are in accord with Durgonuglu and Oney (2000) who considered listening
skills to be developed prior to the acquisition of reading and writing in English speaking
beginning readers. Similar reports have been put forth by Li (2009). In the present study L1
was acquired within the age range of 1-6 years and L2 within 3-15 years. These findings are
in consonance with the report of Lenneberg (1967) who stated the acquisition of L2 by 12
years or puberty, while Johnson and Newport (1989) considered the time of closure to be 15
years of age.
The participants in the Hindi group also reported of never acquiring L3 for reading
and writing purposes. They had learnt Kannada in order to understand and communicate for
routine activities and for communicating to the general public.
Age of attainment of proficiency of the above skills in L1, L2 and L3
A self-reported age for attainment of proficiency in each of the four skills
(understanding, speaking, reading and writing) for L1, L2 and L3 was obtained and subjected
to descriptive statistical analysis.
(i) For L1: It can be noted that understanding and speaking which were mostly used
for functional communication reached a level of native like proficiency earlier than reading
and writing skills. However, the minimum age for acquiring proficiency in all the skills in the
Kannada group was lesser than that of the Hindi group.
(ii) For L2 and L3: It was seen that the proficiency in L2 and L3 was achieved
later than L1 and proficiency in all the skills were achieved at around the similar ages.
Language Proficiency Questionnaire in Indian Context
221
Comparison across the Languages: The overall mean and standard deviation was calculated
using descriptive statistical analysis and it was seen that the attainment of proficiency
followed the same pattern as the age of acquisition; L1 was mastered first followed by L2 and
L3. Friedman test revealed a significant difference across languages in both the groups [{2
(2) = 32.00, p<0.05}-Hindi and {2
(2) = 174.20, p<0.05}-Kannada]. Pair wise comparison
was done using the Wilcoxon signed rank test and significant difference was obtained across
all the languages in both the groups.
ProficiencyAcquisition
Me
an
Ag
e (
yrs
.)
20
18
16
14
12
10
8
6
4
2
0
Lang. & Group
L1 KAN
L1 HIN
L2 KAN
L2 HIN
L3 KAN
L3 HIN
Figure 2: Mean age of acquisition and proficiency across languages and groups.
Evidences from the literature have shown that interactions with family members
posed to be one of the major contributors for attaining proficiency in L1 (Hasson, 2006;
Marian, Blumenfeld & Kaushanskaya, 2007). Apart from the innate ability to learn a
language, the current results shed a light of importance on the variables that facilitates
language proficiency.
The upshots of the current investigation corroborated the earlier findings of Marian,
Blumenfeld and Kaushanskaya (2007). Hence, it can be concluded that all bi/multilinguals
have the same pattern of language acquisition and mastery in spite of regional, social,
cultural, environmental and attitudinal differences among varied regions of the world. This
supports the role of the nature or the innate language device as stated by Chomsky (1957).
III. Proficiency of language
a. Competence across languages: The mean values indicated that the proficiency in L1
was considered to be the best in both Hindi-English and Kannada-English bilinguals.
The participants rated themselves to be better competent in L2 [{Hindi group: 3.11,
SD=0.37} & {Kannada group: 3.12, SD=0.47}] than in L3. The Friedman test was
further carried out which revealed a significant difference across languages for both
the groups [Hindi: {2
(2) = 193.47, p<0.05}, Kannada: {2
(2) = 154.64, p<0.05}].
Additionally, the Wilcoxon signed rank test was done to analyze pair-wise difference
and a statistically significant difference was obtained across all the languages (L1, L2
and L3) in both the groups.
b. Comparison between the competence and performance ratings: A comparison of the
self-ratings and the ratings by the professional was carried out. The reliability
coefficient (Cronbach’s alpha) indicated a good reliability among the judges.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
222
The self reports indicated a better L1 competence than L2 and L3. The ratings of
participants and the judges were in concordance for L2. However, it was seen that on an
average the Hindi-English bilinguals rated themselves to be more competent in L1 than their
actual performance. In contrast, Kannada-English bilinguals felt that they were less
competent than they actually were. The same has been illustrated in figure 4 & 5 (Hindi
group) and figure 6 & 7 (Kannada group). As L1 was the native language of all the
participants, they considered themselves to be native like competent in using the language.
With reference to L1, participants received either little or no feedback about their skills.
Hence, leading to either an over- or underestimation of their skills in L1. Review of earlier
findings by Shameem (1998) suggests similar results as of the present study.
Figure 4: Comparison of competence and Figure 5: Comparison of competence and
performance in L1 for Hindi group performance in L2 for Hindi group
Figure 6: Comparison of competence and Figure 7: Comparison of competence and
performance in L1 for Kannada group performance in L2 for Kannada group
There was accordance among the judges and the self-ratings for L2. The results
replicate and extend previous research on self ratings for L2, where researchers have reported
a correlation between competence and performance levels, thus validating the use of self
reports as a quick tool to measure proficiency. The present outcomes are consistent with the
Language Proficiency Questionnaire in Indian Context
223
results of MacIntyre, Noels and Clement (1997). One possible explanation for this result is
that participants receive more practice and feedback in English (L2) compared to L1. One
may contemplate that bilingual students in schools, where English is the lingua franca,
engage in more use of English either for literary activities or communication than L1.
However, it refuted the findings of Delegado, Guerrero, Goggin and Ellis (1999).
The overall ratings revealed that the competence ratings for L2 almost coincided with
the performance ratings. In contrast, this was not the case for L1 in both the groups. They
either under or over-estimated themselves by one level of their actual performing abilities.
Hence, the ratings need to be inferred with caution in order to assess the degree of bilingual
proficiency in an individual. Thus, a correction factor of level-1 below the reported rating is
advocated.
IV. Usage of language in different contexts
School environment: Descriptive statistical analysis was done and frequencies were
calculated through cross-tabulations. L2 was found to be the medium of instruction in schools
and was the language used maximally when interacting with teachers. However, L1 formed
the language of communication between peers. Some of the participants even reported of
using both the languages. The mean years of education were 17.26 (SD=1.74) years [Hindi-
17.36 (SD=1.99); Kannada-17.16 (SD=1.48)] and all the participants had a range of 15-21
years of education.
The results indicated that L2 was the most widely used language during formal
teaching situations. This would further help to contribute to the development of L2 in an
individual. This finding is supported by the study of Collier (1995); Johannessen and Lopez
(2002) and Li (2009). However, the use of L1 or a combination of languages while
interacting with friends could be attributed to the fact of language choice or proficiency of
language use of the listener/ communication partner.
The longer duration of education also contributes to the attainment of language
proficiency. This finding is in harmony with the view of Flege and Liu (2001).
Interaction with family, friends, market places, media and web based learning
situations:
L1 was the preferred language while interacting with family members and neighbors
L2 was most preferred in the educational set-up and in listening to language
instruction tapes, reading books, using internet sources.
A combination of languages (L1, L2 and L3) was preferred for media based learning,
talking to friends and in market places.
Results of the present study contribute to the existing body of literature which also
emphasizes the usage of L1 while interacting with family members (Hasson, 2006; Marian,
Blumenfeld & Kaushanskaya, 2007). The reasons of language choice could be due to the
interlocutor’s proficiency or choice of language. Cultural and ethnic factors also play a role in
the choice of language according to situations.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
224
The usage of L2 especially for internet sources has also been reported by Chiswick
and Miller (n.d.) and Sanaoui and Lapkin (1992). A combination of languages was most
preferred for media-based learning and while interacting with friends. Similar evidences were
given by Hasson (2006).
Exposure to different languages
For L1: The participants in both the groups reported of being exposed to L1 for all
the 7 (SD=0.00) days in a week. The overall average hours of exposure to L1 ranged from 8-
18 hours with a mean of 14.69 (SD=1.94) hours.
For L2: The total average number of days and hours of exposure to L2 ranged from
5-7 {mean=6.79 (SD=0.58)} and 4-16 {mean=9.52 (SD=3.01)} respectively for all the
participants.
For L3: On the whole, the average number of days and hours all the participants were
exposed to L3 was 6.16 (SD=1.46, range: 0-7) days and 3.64 (SD=2.21, range: 0-10) hours.
The data for duration of exposure of each of the languages in different settings
(family, school and state) was collected and subjected to statistical analysis and the mean,
standard deviation and the range were obtained. The total mean hours of exposure to L1 for
both the groups by the family was 21.15 (SD=2.02, range: 18-25) years; school was 16.76
(SD=1.62, range: 13-20) years and state was 19.62 (SD=1.97, range: 17-23) years. However,
it was seen that participant’s family was never exposed to L2 (English) and L3
(Kannada/Hindi). Similarly the participants did not visit an English speaking country.
The Mann Whitney U-test was carried out and a significant difference among both the
groups was noted for the number of years of exposure to schooling in L3 (z=6.98, p<0.05)
and the duration of stay in a L1 (z=5.16, p<0.05) and L3 (z=6.54, p<0.05) state. The years
of exposure has been represented graphically in figure 8 (Hindi group) and figure 9 (Kannada
group). The participants showed maximum exposure for L1 with respect to the number of
days, hours, interaction with family members and residing in the L1 native-state. These
factors were same for both the groups. This shows that the exposure has played a significant
role in acquiring and making the participants native-like speakers. Similar reports have been
documented in the literature (Marian, Blumenfeld and Kaushanskaya, 2007).
Language
L3L2L1
Me
an
Nu
mb
er
of
Ye
ars
- H
ind
i G
rou
p
25
20
15
10
5
0
Exposure
Education
Family
State
3
18
21
1717
Language
L3L2L1
Me
an
Nu
mb
er
of
Ye
ars
- K
an
na
da
Gro
up
25
20
15
10
5
0
Exposure
Education
Family
State
2121
8
1617
Mean years of exposure to education, family and state for L1, L2 and L3
Figure 8: Hindi group Figure 9: Kannada group
Language Proficiency Questionnaire in Indian Context
225
Acquiring of proficiency in L2 for both the groups could be attributed to the medium
of instruction in schooling. In the Indian scenario the schools affiliated to state or central
board mostly have English as medium of instruction. As most of the participants studied in
private institutions where the medium of instruction was English, it is evident from the results
that these participants were exposed to L2 for 5-7 days with an average mean of 10 hours per
day. Thus, the duration of schooling was an important factor in learning L2. Further, all the
participants belonged to literate families where parents used L2 also for communication.
(Collier, 1995; Marian, Blumenfeld & Kaushanskaya, 2007).
The Kannada group participants had an exposure to L3 at the school level, whereas
the Hindi participants began learning L3 only when they moved from their own states to
Karnataka for higher education. This again shows that classroom teaching at school did play
an active role in learning L3. This was evident from the performance of Kannada group who
learnt L3 (Hindi) during initial stages and continued till 12th
standard of education. Thus,
most of the participants had an exposure for academic purposes for L3 for more than 7-8
years. However, such exposure was not seen for Hindi group. This further strengthens that
classroom exposure is an important contributing factor for acquiring L3. Thus, the results of
the study get support from Marian, Blumenfeld and Kaushanskaya (2007).
As a result, it can be stated from the findings that it is not the only the factors but the
duration of exposure which acts as a significant variable to learn a language/s.
Duration of exposure
a. For L1: Interaction with family, neighbors, market places, radio, watching television and
interacting with friends were the different settings where the participants were maximally
exposed for most of the time to L1. The next major important factors that contributed to
L1 learning were education writing newspapers story books and textbooks.
Exposure
Mkt
Radio TVNbFrn
dWrt
NetHBNPSBDicTBLTEd
n.Fly.
Mean
Rating
- L 1
4.0
3.0
2.0
1.0
Groups
Hindi
Kannada
Figure 10: Mean ratings of duration of exposure to the factors contributing to L1
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
226
Exposure
Mkt
Radio TVNbFrn
d
Wrt
Net
HBNPSBDicTBLTEdn.Fly.
Mean
Ratin
g - L
2
4.0
3.0
2.0
1.0
Groups
Hindi
Kannada
Figure 10: Mean ratings of duration of exposure to the factors contributing to L1
b. For L2: The results revealed two categories of contributors, i.e. most important and the
moderately important contributors to the learning of L2. The most important contributors,
i.e., the factors to which all the participants were exposed for most of the time or always were
internet, textbooks, dictionary, writing, newspapers, story & historical books, instructional
tapes and educational set-up.
c. For L3: The mean scores indicate that the participants were exposed to L3 maximally
when watching television (1.96, SD=0.88), followed by educational/workplaces (1.95,
SD=0.72), interaction with friends (1.73, SD=0.66), listening to radio (1.60, SD=0.69) and
market areas. These factors serve as catalyst to improve either the learning and/or proficiency
of language. Thus, all the above variables are unique and contribute in a unique manner to
learn the language. This study receives support from the findings of Harniess (2008).
Exposure
Mkt
Radio TVNb
FrndWrt
Net
HBNPSBDicTBLT
Edn.
Fly.
Mea
n Ra
ting
- L 3
4.0
3.0
2.0
1.0
Groups
Hindi
Kannada
Figure 12: Mean ratings of duration of exposure to the factors contributing to L3
V. Native/non-native speaker
The participants were asked to rate themselves on a four point rating scale about their
identification as a native speaker of the language based on their accent and pronunciation.
95% (Hindi-96.7% and Kannada- 93.3%) of the total participants reported of being identified
as a native speaker of L1 the entire time. However, majority of the participants though
competent in using L2 were rarely identified as the native speakers of that language.
The participants of the present study identified themselves as native speakers of L1
(Hindi/Kannada) as they had acquired the language since childhood and were maximally
exposed to it in all kind of situations and contexts. This finding is supported by Piske,
Mackay & Flege (2002).
Language Proficiency Questionnaire in Indian Context
227
Though they had attained good proficiency in the usage of L2 yet the degree of
foreign accent was not achieved as the participants were never exposed to the native speakers
of English. The review of literature also suggests that if a child acquires L2 before puberty,
then they acquire little or no foreign accent (Tahta, Wood & Lowenthal, 1981). All the
participants in the present study had acquired L2 before puberty and had not stayed in any of
the English speaking countries, hence this could have attributed to the development of no
foreign accent. However, few of the participants self-reported of being identified sometimes
as the native speaker of L3. This finding could be due to a longer and intensive exposure to
the language by native speakers compared to L2.
Summary and Conclusions
From the findings of the present study it can be concluded that this is a viable tool in
assessing the language proficiency of an individual and the factors contributing to it. This
questionnaire does not replace the earlier assessment batteries but can serve as an adjunct and
quick measure for assessing a bi/multilingual for his proficiency level. It would assist a
Speech Language Pathologist to assess the language levels of an individual in the language
that s/he does not know. Hence, it can be stated that the present questionnaire is not
supplementing rather complementing the existing tools.
Implications of the study
This questionnaire will help the professionals to find out the level of proficiency of an
individual in a language on four basic skills, i.e., understanding, speaking, reading and
writing.
The present study further corroborates the evidences to research in language
proficiency and the factors contributing to it and opines to carry out extensive
research in this area.
Evaluation of first and second language factors will give an insight into the transfer
skills between L1 and L2 and should be considered during routine speech and
language assessments in bilinguals.
In therapeutic intervention, for individuals with aphasia, factors related to second
language exposure become crucial while deciding on the selection of language for
intervention. Hence, an evaluation of the proficiency level in the pre- and post-morbid
condition will help to choose the appropriate language for speech and language
remediation program.
References
Bloomfield, L. (1933). Language. New York: Henry Holt & Co.
Chiswick, B. R., & Miller, P.W. (n.d.). Computer usage, destination language proficiency and
the earnings of natives and immigrants. KMI Working Paper Series, 9, 1-41
Chomsky, N. (1957). Syntactic Structures. The Hague:Mouton.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
228
Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge, MA: MIT Press.
Collier, V. P. (1995). Acquiring a second language for school. Directions in Language and
Education, 1(4), 1-10.
Cummins, J. (2000). Language, power, and pedagogy: Bilingual children in the crossfire.
Clevedon, England: Multilingual Matters.
Delgado, P., Guerrero, G., Goggin, J. P., & Ellis, B. B. (1999). Self-assessment of linguistic
skills by bilingual Hispanics. Hispanic Journal of Behavioral Sciences, 21, 31-46.
Durgunoglu, A. Y. & Oney, B. (2000). Literacy Development in Two Languages: Cognitive
and Sociocultural Dimensions of Cross-Language Transfer. Paper presented at a
research symposium on High standards in reading for students from diverse language
groups: Research, practice & policy, Washington D.C.: U.S.
Flege, J. E., & Liu, S. (2001). The effect of experience on adults’ acquisition of a second
language. Studies in Second Language Acquisition, 23, 527-552.
Flege, J. E., MacKay, I. A., & Piske, T. (2002). Assessing bilingual dominance. Applied
Psycholinguistics, 23, 567-598.
Harniess, J. (2008). Approaches to second language English proficiency: Parents’
perspective. Retrieved March 23, 2009 from
http://www.hkugac.edu.hk/news/EnglishAcquisition- A Parents Perspective-
November 2008.pdf
Hasson, D. J. (2006). Bilingual language use in Hispanic young adults: Did elementary
bilingual programs help? Bilingual Research Journal, 30, 45-64.
Haugen, E. (1953). The Norwegian Language in America. Cambridge: Cambridge University
Press.
Johnson, J. S., & Newport, E. L. (1989). Critical periods effect in second language learning:
The influence of maturational state on the acquisition of English as a second
language. Cognitive Psychology, 21, 60-99.
Johannessen, B. G. G., & Bustamante-López, I. (2002). Bilingual Academic Spanish
Proficiency Tests: Assessment of bilingual cross-cultural and language and academic
development teacher candidates. Bilingual Research Journal, 26, 563-574.
Kertesz, A. (1982). The Western Aphasia Battery. Philadelphia, PA: Grune & Stratton.
Lenneberg, E. (1967). Biological foundations of language. New York: John Wiley.
Li, Z. (2009). A probe into classroom teaching and second language acquisition.
International Education Studies. 2(1), 124-128.
MacIntyre, P. D., Noels, K. A., & Clement, R. (1997). Biases in self-ratings of second
language proficiency: The role of language anxiety. Language Learning, 47, 265–287.
Marian, V., Blumenfeld, H. K., & Kaushanskaya, M. (2007). The Language Experience and
Proficiency Questionnaire (LEAP-Q): Assessing language profiles in bilinguals and
multilinguals. Journal of Speech, Language, and Hearing Research, 50, 940–967.
Language Proficiency Questionnaire in Indian Context
229
Piske, T., MacKay, I. R. A., & Flege, J. E. (2001). Factors affecting degree of foreign accent
in an L2: A review. Journal of Phonetics, 29(2), 191-215.
Rosenberg, M. (1996). Raising Bilingual Children. The Internet TESL Journal, 2(6).
Retrieved March 12, 2009 from http://iteslj.org/Articles/Rosenberg-Bilingual.html
Sanaoui, R., & Lapkin, S. (1992). A case study of an FSL senior secondary course integrating
computer networking. [Abstract] The Canadian Modern Language Review, 48 (3),
525-552.
Schmitt, N. (2008). Review article: Instructed second language vocabulary learning.
Language Teaching Research, 12(3), 329-363.
Shameem, N. (1998). Validating self-reported language proficiency by testing performance in
an immigrant community: The Wellington Indo-Fijians. Language Testing, 15, 86-
108.
Tahta, S., Wood, M., & Loewenthal, K. (1981). Foreign accents: factors relating to transfer of
accent from the first language to a second language. Language and Speech, 24(3),
265-272.
Thirumalai, M. S. & Chengappa, S. (1986). Simultaneous acquisition of two languages: An
overview. Central Institute of Indian Languages, Mysore.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
230
APPENDIX I
LANGUAGE PROFICIENCY QUESTIONNAIRE
Name: Age: Gender: Male / Female
Instructions:
Please read the questions carefully and choose the most appropriate choice wherever applicable.
1. Name all the languages you know beginning with the language that you learnt first.
Using the below mentioned scale, answer the questions below.
(1- L1, 2-L2, 3-L3, 4- Combination of any of the languages)
L1- First language that you learnt, L2- Second language that you learnt in your life,
L3- Third language.
2. When you were a child, which language did you speak
At Home 1 2 3 4
With your father 1 2 3 4
With your mother 1 2 3 4
With siblings 1 2 3 4
With guardians 1 2 3 4
With neighbors 1 2 3 4
3. Native Language of
Father 1 2 3 4
Mother 1 2 3 4
Sibling’s 1 2 3 4
Guardians 1 2 3 4
4. Language spoken with you by your
Father 1 2 3 4
Mother 1 2 3 4
Sibling’s 1 2 3 4
Guardians 1 2 3 4
Neighbors 1 2 3 4
5. Which language did you learn first for
Understanding 1 2 3 4
Speaking 1 2 3 4
Reading 1 2 3 4
Writing 1 2 3 4
6. Mention the age when you first started using each of the languages for each of the following
parameters:
Understanding Speaking Reading Writing
L1
L2
L3
7. Mention the age when you became proficient for each of the following parameters:
Understanding Speaking Reading Writing
L1
L2
L3
Language Proficiency Questionnaire in Indian Context
231
8. How many years of formal education do you have? (please specify your qualification)
What was the medium of instruction? 1 2 3 4
Which language was used maximally? 1 2 3 4
Which language did you speak with teachers 1 2 3 4
Which language did you speak with classmates 1 2 3 4
Which language was spoken by your teachers with you 1 2 3 4
Which language was spoken by your classmates with you 1 2 3 4
Did you change your medium of instruction? Yes No
If yes, specify the changed medium of instruction. At what age
did you change your medium of instruction?
1 2 3 4
9. Have you changed your state? If yes, which language do you
use to communicate?
1 2 3 4
10. On a scale from one to five, mark your level of proficiency in each of the skill
(1-Zero proficiency, 2- Low, 3- Good, 4- Native like/perfect)
Language Understanding Speaking Reading Writing
L1
L2
L3
11. How many dialects can you speak in each of the languages?
L1: L2: L3:
12. On a scale from one to five, mark your level of proficiency in each of the skill for each of the dialects
in L1, L2, L3.(1-Zero proficiency, 2- Low, 3- Good, 4- Native like/perfect)
L1 L2 L3
Dialect D1 D2 D3 D1 D2 D3 D1 D2 D3
Understanding
Speaking
13. On a scale from one to five, mark your level of proficiency in shifting from one language to the other
1-Zero proficiency 2- Low
3- Good 4- Perfect
14. Use the rating scale mentioned below, indicate which language you used maximum for the following:
(1- L1 , 2- L2, 3- L3, 4- Combination of any of the languages)
Interaction with family 1 2 3 4
Education/ work 1 2 3 4
Listening to instruction tapes at school 1 2 3 4
Text books 1 2 3 4
Dictionary 1 2 3 4
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
232
Story books 1 2 3 4
Newspapers 1 2 3 4
Historical books 1 2 3 4
Internet source 1 2 3 4
Writing 1 2 3 4
Interacting with friends 1 2 3 4
Interacting with neighbors 1 2 3 4
Watching TV 1 2 3 4
Listening to the radio 1 2 3 4
Market places 1 2 3 4
15. On an average, mention below the time you are exposed to each of the languages.
Languages Number of days per week Number of hours per day
L1
L2
L3
16. Mention the number of years you spent in each language environment:
Family School State Work place
L1
L2
L3
17. Using the rating scale mentioned below, indicate the extent to which you are currently exposed to
each of the languages in the following contexts in a day.
(1- never, 2- sometimes, 3- most of the time, 4- always)
L1 L2 L3
Interaction with family
Schooling/ work
Listening to instruction tapes at school
Text books
Dictionary
Story books
Newspapers
Historical books
Internet source
Writing
Interacting with friends
Interacting with neighbors
Watching television
Listening to the radio
Market places
18. Rate how frequently others identify you as a native speaker based on your accent or pronunciation in
the language (1- Never, 2- Sometimes, 3- Most of the time, 4- Always).
1. L1 2. L2 3. L3
Paraphasias in Bilingual Aphasia
233
Paraphasias in Bilingual Aphasia
Ridhima Batra & Dr. K.C. Shyamala*
Abstract
The language abilities of aphasia yields deficit in phonological, semantic, syntactic and
pragmatic systems. During their attempt to produce a word, aphasics tend to substitute an incorrect
word for the intended or the target word which are termed as paraphasias. Bilingual aphasics do not
necessarily show the same language disorders with the same degree of severity in both languages, so
it becomes ethically important to do a detailed evaluation of all the languages known by an aphasic
patient. The present study aimed at investigating the paraphasias in monolingual and bilingual
aphasics and highlights the variation/correlation of paraphasias across languages in bilingual
aphasics. Twenty four individuals with aphasia (twelve monolinguals and twelve bilinguals) with a
mean age of 55 years participated in the study. Naming, repetition and picture description sections of
the WAB test were audio recorded, transcribed and analyzed for the six categories of paraphasias
namely, semantic, formal, mixed, unrelated, phonemic and neologism. The results of the study
revealed a better performance of the bilingual individuals with aphasia on all the three tasks.
Amongst the aphasic subgroups, anomic aphasics showed more semantic paraphasias; conduction
aphasics produced abundant phonemic paraphasias; Wernicke’s aphasia presented with more
number of semantic and unrelated paraphasias and Broca’s and global aphasics exhibited a high
incidence of phonemic and neologistic paraphasias. The bilingual individuals with aphasia performed
better in L2 than in their mother tongue. Different types of paraphasias were observed in the two
languages in the bilingual aphasics.
Introduction
Bilingualism is a phenomenon which refers to use of two or more languages by
individuals in their everyday lives. With increasing globalization, the number of people using
two or more languages i.e., bilingual individuals are also increasing. There have been
different views of the phenomenon of bilingualism in the literature. Some researchers point
out that bilingual individuals are not two monolinguals in one person, but rather speakers
using different languages in different domains or situations, for different purposes, and with
different interlocutors. Statistics reveal an increase in bilingual population in the world. India
being a multilingual country has abundant bilingual/multilingual population with various
permutations and combinations of the languages paired.
Since most people in India know more than one language, bi/multilingualism is a very
prominent phenomenon in India. On the basis of these considerations, a large population of
individuals with aphasia in India would therefore show “bilingual aphasia”. Bilingual
aphasics do not necessarily show the same language disorders with the same degree of
severity in both languages, so it becomes ethically important to do a detailed evaluation of all
the languages known by an aphasic patient.
____________________________________________
Professor of Language Pathology, All India Institute of Speech and Hearing, Mysore, India
e-mail: [email protected]
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
234
The characteristic features of the language abilities of aphasia yields deficit in
phonological, semantic, syntactic and pragmatic systems. During their attempt to produce a
word, aphasics tend to substitute an incorrect word for the intended or the target word which
are termed as paraphasias. Paraphasias can appear in naming, repetition, spontaneous speech,
reading or writing tasks. Paraphasias are common in aphasia and can help differentiate fluent
from non-fluent subgroups of aphasia.
Paraphasias can be of different kinds and have a good localization value. Different
kinds of paraphasias described by Goodglass (1993) are as follows:
Verbal paraphasia: It refers to the unintended use of another word in lieu of the target.
Phonemic paraphasia: These are also called as “literal paraphasia”. It is the production
of unintended sounds or syllables in the utterance of a partially recognizable word
(e.g., “paker” for “paper”, “sisperos” for “rhinoceros”).
Phonosemantic blends: it is often the case that a phonemic sound substitution results in
another real word, related in sound but not the meaning. E.g. „table‟ becomes „cable‟;
„telephone‟ becomes „television‟.
Neologistic paraphasia: it is the production of a non-sense word or words, usually
without recognition of error. E.g. „table‟ becomes „tilto‟.
Paraphasias have been classified differently by various authors. Lesser (1978) classified
based on the word forms, if they belonged to the language used or not. She also identified
whether the spoken word is sufficiently similar to the actual word form phonologically,
morphologically, or semantically. Li and Williams (1990) gave a checklist to examine the
repetition errors made across various aphasic syndromes and divided the errors into seven
categories (word substitution errors, addition errors, omission errors, revision errors, jargon,
paraphrase error and inadequate response). Paraphasic errors were also divided into lexical
where a real word is substituted for another, or sub-lexical when a non-word is produced
(Dell, Schwartz, Martin, Saffran & Gagnon, 1997).
Paraphasias in individuals with aphasia has been researched upon extensively, using
different tasks like naming (e.g., Goodglass, Kaplan et al., 1976; Goodglass, 1981; Kohn &
Goodglass, 1985; Martin & Saffran 1992; Gagnon et. al, 1997), repetition (e.g., Gardner &
Winner, 1978; Goodglass & Kaplan, 1983; Li & Williams, 1990) and picture description
(Williams & Canter, 1982, 1987). The naming task helps to find paraphasias in single object
confrontation naming; the repetition task taps the deficits in the transfer of information
between the input and output pathways and the picture description task is used to find the
paraphasias in a narrative context.
Studies on naming task have revealed high frequency of semantic paraphasias in
anomics as an index of word-finding difficulty (Kohn & Goodglass, 1985). Li and Williams
(1990) reported that aphasics tend to exhibit significantly more indefinite terms, extended
circumlocutions and perseverations in the naming conditions. Gagnon et. al (1997) reported
of presence of formal paraphasias in aphasics in a naming task. Using the repetition task,
Paraphasias in Bilingual Aphasia
235
Gardner and Winner (1978) reported that conduction aphasics make more meaning errors or
verbal paraphasias, whereas, Li and Williams (1990) found phonemic attempts and revisions
to be more prominent in the repetition of conduction aphasics. Studies on picture description
task found that Broca‟s aphasics performed significantly better when naming objects on
confrontation naming task than on picture description task, whereas, a reverse trend was seen
in Wernicke‟s aphasics (Williams & Canter, 1982). Williams and Canter (1987) found that
anomics produced more of delayed responses and extended circumlocutions; Wernicke‟s
produced more neologisms and the Broca‟s produced significantly more phonemic errors and
semantic-phonemic errors on picture description task.
However, in the Indian scenario, there have been just a handful of studies to examine
the type of paraphasias produced in various subgroups of aphasia. Shantala (1997) studied
naming deficits in confrontation naming, responsive naming and generative naming tasks and
reported of neologisms and phonemic errors in Broca‟s aphasics; semantic and phonemic
errors in the anomics and conduction aphasics exhibited neologisms and gestural responses.
Similar tasks were used by Arpita (1997) to tap the naming deficits in Kannada-English
bilingual aphasics. Results revealed parallel deficits in L1 and L2 on responsive naming and
generative naming task, however, in confrontation naming task, performance was better in
L1. Error analysis revealed a difference in the performance of the bilingual aphasics in the
two languages.
Chengappa, Bhat and Damle (2003) investigated paraphasias on repetition tasks in a
multilingual Wernicke‟s aphasic patient and highlighted the variation of these across the four
languages known by the patient. Hegde and Bhat (2007) also highlighted the variation of
paraphasias across four languages known by a multilingual conduction aphasic on a repetition
section.
However, there has been limited research done to find out the type of paraphasias
produced by various bilingual aphasic syndromes and explore as to whether it differs from
monolingual aphasics. Also, there is a scarcity of research where a comparison has been
made for the types of paraphasias produced by the bilingual aphasics in the different
languages known to them.
Aims of the study
The present study aimed at investigating:
1. The paraphasias in monolingual and bilingual aphasics.
2. Highlight the variation/correlation of paraphasias across languages in bilingual
aphasics.
3. Describe the type of paraphasias in different subgroups of aphasia.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
236
Method
Participants
Twenty four individuals with aphasia in the age range of 30-80 years (mean age of 55 years),
identified through various sources like institutes, hospital records were taken for the study.
These participants were divided into two groups: twelve monolinguals and twelve bilinguals.
The following criteria were considered in the selection of the participants in the study.
All participants were diagnosed as having aphasia by a Speech Language Pathologist
and/or neurologist.
All monolingual individuals with aphasia were native Kannada speakers. The bilingual
individuals with aphasia had Kannada as their mother tongue and had learnt English as
second language before the age of 15 years. The bilingual individuals were identified
using Australian Second Language Proficiency Rating (ASLPR, Ingram, 1985). On the
basis of self report and information from significant others, the individuals who passed
fourth level (Vocational Proficiency) in the ASLPR (Appendix) in the second language
i.e English in at least speaking and listening domains of the rating scale premorbidly
were considered as bilinguals.
Different aphasic syndromes were considered. This was determined on the basis of
clinical observation and Western Aphasia Battery (WAB, Kertesz & Poole, 1974)
findings. The participants were grouped into fluent and non-fluent aphasia types as per
the classification system of Goodglass and Kaplan (1972). The various types of aphasic
syndromes identified were anomia (5), conduction (3), Wernicke‟s (2), trans-cortical
motor (TCM) (1), Broca‟s (8) and global (5).
Both male and female aphasics were considered for the study. Participants included five
females and nineteen males.
The participants suffered a left hemisphere stroke revealed by MRI/CT scan reports. The
time from the onset of stroke varied from two to twenty four months.
All participants were right handed. This was determined using self-report and
information from significant others.
None of the participants had any auditory or visual deficit as assessed informally.
Ethical considerations were met.
Procedure
Subjects were seated comfortably. Before starting the evaluation and recording, the subjects
were informed about the entire procedure and an informed consent was taken. The
environment was made as distraction free as possible by carrying out the procedure in a quiet
room and by removal of any potential visual distracters.
Paraphasias in Bilingual Aphasia
237
Tests administered
Western Aphasia Battery (WAB, Kertesz & Poole, 1974) was administered for all the
participants for diagnosing and identification of the type of aphasia.
The following three sections of the WAB test were audio recorded, transcribed and
analyzed for the presence of paraphasias:
Repetition- It comprises of 20 stimulus items consisting of words, phrases and
sentences. The subjects were instructed to repeat after the examiner.
Naming- This section consists of 20 common objects. The objects were presented to
the subjects in their visual field one after the other and they were instructed to name
the objects in a single word as soon as possible. If the subjects were unable to name
the object in 30 sec, it was considered as a no response.
Picture description: The “picnic” picture served as the test stimuli to look for
paraphasias in narrative context. The subjects were instructed to describe the picture
in their own words. No time limit or word limit was given for this task.
All the three sections of the WAB were administered in Kannada for the monolingual
individuals with aphasia and both in Kannada (L1) and English (L2) for the bilingual
individuals with aphasia for comparing the type of paraphasias between monolinguals and
bilinguals and also between the two languages (L1 & L2) for bilingual individuals with
aphasia.
Scoring
The subjects‟ first response was evaluated in terms of absolute correctness for all the
three sections; repetition, naming and picture description. Any deviation from the target was
analyzed for the presence of paraphasias.
Analysis
The repetition, naming and picture description sections of the WAB test were audio
recorded, transcribed and the responses which were not appropriate to the target word
were analyzed for the presence of paraphasias and comparison was made across:
Monolinguals and bilinguals
Kannada (L1) and English (L2) in bilingual individuals with aphasia
Qualitative analysis of the different kinds of paraphasias
A list was prepared to classify the type of paraphasias shown by the participants based on
Li and Williams (1990) checklist and Dell‟s classification system of errors (1997).
The paraphasias were divided into two broad categories, lexical and sublexical. A
lexical paraphasia is a real word substituted for another whereas; a non-word produced falls
into sublexical category.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
238
a) At lexical level, paraphasias can be:
(i) Semantic- word related to target in meaning.
(ii) Formal- word related to target in sound.
(iii) Mixed- word with sound and meaning relationship.
(iv) Unrelated- word with no apparent relation to target.
b) At sublexical level, paraphasias can be:
i) Phonemic- non-word related in sound.
ii) Neologistic- non-word with a remote relationship to target.
Statistical Analysis
Following statistical measures were used for the analysis of the data using SPSS software
(Version-16).
a) Mann Whitney U test was done to observe whether there exists a difference between
monolingual and bilingual individuals with aphasia on the various kinds of responses
(paraphasias and other responses) given on the three tasks namely naming, repetition
and picture description.
b) Wilcoxon Signed Rank test was done for the bilingual group to see variation of the
kinds of paraphasias and other responses across the two languages, Kannada and
English on all the three tasks.
Results
The primary aim of the study was to observe the type of paraphasias in monolingual
and bilingual individuals with aphasia. A total of 24 individuals with aphasia were evaluated
on naming, repetition and picture description task. The results of the study have been
presented with reference to the performance of the participants on:
4.1 Naming task
4.2 Repetition task
4.3 Picture description task
The results of all the three tasks i.e., naming, repetition, and picture description in this
section are presented under following headings:
Comparison of monolingual group and bilingual group
Across language comparison for the bilinguals participants
Qualitative analysis of all the paraphasias
Paraphasias in Bilingual Aphasia
239
4.1 Naming task
4.1.1 Monolinguals and bilinguals
The naming task consisted of 20 stimuli which were scored by the investigator as
either correct or incorrect response. The total mean percentages for the correct and incorrect
responses for both the fluent and non-fluent groups in monolinguals and bilinguals were
calculated. It was seen that, in the overall scores, both fluent and non-fluent bilinguals
performed better (14.58%) than the fluent and non-fluent monolinguals (9.16%). The
incorrect responses of all the participants were subjected to further analysis. Based on the
results of the analysis, the responses of the participants were classified under six categories of
paraphasias namely, semantic, formal, mixed, unrelated, phonemic and neologism.
The means and standard deviations (SD) of all the types of responses for all the groups
were measured. Figures 4.1 and 4.2 graphically represent the percentage occurrence of the six
types of paraphasias in both monolingual and bilingual group of aphasics.
Figure 4.1 Figure 4.2
Paraphasia in Naming (Kannada)
NeoPhonUnrelMixedFormSem
Me
an
Pe
rce
nta
ge
Re
sp
on
se
s -
Mo
no
lin
gu
als
55
50
45
40
35
30
25
20
15
10
5
0
Groups
Fluent
Non-fluent
Paraphasia in Naming (Kannada)
NeoPhonUnrelMixedFormSem
Me
an
Pe
rce
nta
ge
Re
sp
on
se
s -
Bilin
gu
als
55
50
45
40
35
30
25
20
15
10
5
0
Groups
Fluent
Non-fluent
Sem- semantic paraphasia, Form- formal paraphasia, Mixed- mixed paraphasia, Unrel-
unrelated paraphasia, Phon- phonological paraphasia, Neo- neologisms
As seen from the above results, in the naming task, the bilinguals performed better
than the monolinguals in both fluent and non-fluent groups. The most prominent paraphasias
seen were semantic paraphasias followed by phonemic paraphasias in the fluent group
whereas neologisms and phonemic paraphasias were more common in the non-fluent group
in both monolingual and bilingual aphasics. Further, Mann Whitney U test was carried out to
check the difference in the performance of fluent and non-fluent monolinguals and bilinguals
on the six types of paraphasias. The result of the Mann Whitney U test revealed no significant
difference between fluent and non-fluent monolinguals and bilinguals on any of the
paraphasias.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
240
4.1.2 Across language comparison
The bilingual individuals with aphasia were compared for their performance in
naming task across both Kannada and English language. The total mean percentage for the
correct and incorrect responses for both the fluent and non-fluent groups of bilinguals in both
the languages was done. Results show that the performance of both fluent and non-fluent
bilinguals is much superior in English (L2) than in Kannada (L1). Wilcoxon signed rank test
was carried to find out if there was any significant difference in the performance of fluent and
non-fluent bilinguals. The results of the test also revealed that L2 is significantly better than
L1 (z=2.371 at p<0.05).
Figure-4.3 depicts graphical representation of the percentage occurrence of the
various paraphasias for the bilingual group in English language.
Figure 4.3: Percentage occurrence of the six kinds of paraphasias in the bilingual group in the
naming task: English
Paraphasia in Naming (English)
NeoPhonUnrelMixedFormSem
Me
an
Pe
rce
nta
ge
Re
sp
on
se
s -
Bili
ng
ua
ls
55
50
45
40
35
30
25
20
15
10
5
0
Groups
Fluent
Non-fluent
Sem-semantic paraphasia, Form- formal paraphasia, Mixed- mixed paraphasia, Unrel-
unrelated paraphasia, Phon- phonological paraphasia, Neo- neologisms
The results in across language comparison state that the performance of the
participants in L2 was better than in the mother tongue (L1). The fluent aphasic group
showed more semantic paraphasias in L1 and the non-fluent group showed more of
neologisms and phonemic paraphasias in both L1 and L2. Wilcoxon Signed Rank test was
conducted to find out if the performance of the aphasics in the two languages differed in
various types of paraphasias. The semantic paraphasias were found to be significantly more
in L1 than in L2 (z=2.032, p<0.05) in the fluent group of aphasia. In contrast, it was noted
that no significant difference was obtained for the non-fluent bilingual group across the two
languages in the six types of paraphasias.
4.1.3 Qualitative analysis
With reference to the paraphasias present in the various types of aphasias, semantic
paraphasias were maximally present in anomics followed by conduction aphasia. The non-
fluent individuals with aphasia in both monolingual and bilingual group exhibited phonemic
Paraphasias in Bilingual Aphasia
241
and neologistic paraphasias predominantly. Formal paraphasias (real word related to target in
sound) were also commonly seen in Broca‟s aphasia.
4.2 Repetition task
The repetition task consisted of 50 target words presented in isolated words, phrases
and sentence forms to monolingual and bilingual aphasics.
4.2.1 Monolinguals and bilinguals
By calculating the total mean percentage of correct and incorrect responses obtained
in repetition task for both the fluent and non-fluent monolingual and bilingual aphasics, it
was seen that the overall scores of both fluent and non-fluent bilinguals (34%) were better
than the fluent and non-fluent monolinguals (13.66 %). The error responses were divided
into six paraphasias. Figures 4.4 and 4.5 show the graphical representation of the percentage
occurrence of the six types of paraphasias in monolingual and bilingual participants in the
repetition task.
Figure 4.4 Figure 4.5
Paraphasia in Repitition (Kannada)
NeoPhonUnrelMixedFormSem
Me
an
Pe
rce
nta
ge
Re
spo
nse
s -
Mo
no
ling
ua
ls
55
50
45
40
35
30
25
20
15
10
5
0
Groups
Fluent
Non-fluent
Paraphasia in Repitition (Kannada)
NeoPhonUnrelMixedFormSem
Me
an
Pe
rce
nta
ge
Re
spo
nse
s -
Bili
ng
ua
ls
55
50
45
40
35
30
25
20
15
10
5
0
Groups
Fluent
Non-fluent
Sem- semantic paraphasia, Form- formal paraphasia, Mixed- mixed paraphasia, Unrel-
unrelated paraphasia, Phon- phonological paraphasia, Neo- neologisms
As depicted in the above results, the bilinguals performed better than the
monolinguals in the repetition task. It was also noted that less number of paraphasias and
more number of no responses were obtained from all the individuals with aphasia who
participated in the study. Amongst the type of paraphasias present, formal paraphasias were
most prominent in the fluent group whereas neologisms and phonemic paraphasias were more
often seen in the non-fluent group. Mann Whitney U test was done to check whether there
exists a difference between fluent and non-fluent monolinguals and bilinguals on the six
types of paraphasias in the repetition task. As per the result of the Mann Whitney U test,
monolinguals showed significantly more mixed paraphasias in the fluent group (z=2.372 at
p<0.05) and, unrelated paraphasias in the non-fluent group (z=2.318 at p<0.05) than the
bilinguals.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
242
4.2.2 Across language comparison
The bilingual individuals with aphasia were also examined whether their performance
on repetition task differed across languages. Results showed that the performance of fluent
bilinguals is much superior in English (L2) than in Kannada (L1). Wilcoxon signed rank test
also revealed L2 to be significantly better than L1 (z=1.992 at p<0.05) in the fluent group.
However, in the non-fluent group, the performance of the participants was not significantly
different in the two languages which was also confirmed by Wilcoxon signed rank test (z=
1.069 at p>0.05).The incorrect responses for the aphasics were again classified into six
paraphasias (graphically shown in figure-4.6).
Figure 4.6: Percentage occurrence of the six kinds of paraphasias in the bilingual
group in the repetition task: English
Paraphasia in Repitition (English)
NeoPhonUnrelMixedFormSem
Me
an
Pe
rce
nta
ge
Re
spo
nse
s -
Bili
ng
ua
ls
55
50
45
40
35
30
25
20
15
10
5
0
Groups
Fluent
Non-fluent
Sem- semantic paraphasia, Form- formal paraphasia, Mixed- mixed paraphasia, Unrel-
unrelated paraphasia, Phon- phonological paraphasia, Neo- neologisms
To sum up, the repetition ability of the bilingual individuals with aphasia was better in
L2 than in L1. The fluent group produced significantly more phonemic paraphasias in L1 and
more formal paraphasias in L2. On the other hand, the non-fluent group produced more
neologisms in both L1 and L2. Wilcoxon Signed Rank test was done for the fluent and non-
fluent bilingual aphasics to see variation across the two languages, Kannada and English and
the results of the Wilcoxon Signed Rank revealed significantly more phonemic paraphasias in
L1 in the fluent group (z=2.220 at p<0.05) while, significantly more neologisms in L2 in the
non-fluent group (z=2.023 at p<0.05).
4.2.3 Qualitative analysis
Results of the study revealed that in the repetition task, conduction aphasics exhibited
a greater number of phonemic paraphasias; Broca‟s aphasics demonstrated more phonemic
errors and neologisms; and Wernicke‟s aphasics exhibited more unrelated paraphasias and
jargon.
Across language comparison revealed a significant difference in the performance of
the fluent group. The fluent bilingual group produced more number of formal paraphasias in
L2. Contrastively, phonemic paraphasias (non words related to the target in sound) were
Paraphasias in Bilingual Aphasia
243
prominent in L1. Extensive number of neologisms was also seen in the repetition task by the
non-fluent group in both monolingual and bilingual group. It was also noted that the semantic
and mixed paraphasias were scantily present in the repetition task.
4.3 Picture description task
The „picnic‟ picture from the Western Aphasia Battery (WAB) was used to elicit
spontaneous speech sample of the 24 aphasics who participated in the study. This task was
chosen to examine paraphasias in the narrative context.
4.3.1 Monolinguals and bilinguals
In the picture description task, the total number of words uttered and the number of
correct words in that were calculated from the speech sample. Results showed that, both the
fluent and non-fluent bilinguals uttered more words (32% & 21% respectively); and their
mean percentage of the correct responses was also slightly higher than the monolinguals. This
finding was confirmed by Mann Whitney U test which also revealed that the bilinguals at
large uttered significantly more number of words than the monolinguals (z=2.170 at p<0.05).
However, the total number of correct words was not significantly different in the two groups
as per the results of the test (z=0.322 at p>0.05).
Figures 4.7 and 4.8 graphically represent the percentage occurrence of the six types of
paraphasias in both monolingual and bilingual group of aphasics in the picture description
task.
Figure 4.7 Figure 4.8
Paraphasia in Picture Description (Kannada)
NeoPhonUnrelMixedFormSem
Me
an
Pe
rce
nta
ge
Re
sp
on
se
s -
Mo
no
lin
gu
als
55
50
45
40
35
30
25
20
15
10
5
0
Groups
Fluent
Non-fluent
Paraphasia in Picture Description (Kannada)
NeoPhonUnrelMixedFormSem
Me
an
Pe
rce
nta
ge
Re
sp
on
se
s -
Bili
ng
ua
ls
55
50
45
40
35
30
25
20
15
10
5
0
Groups
Fluent
Non-fluent
Sem- semantic paraphasia, Form- formal paraphasia, Mixed- mixed paraphasia, Unrel-
unrelated paraphasia, Phon- phonological paraphasia, Neo- neologisms
As depicted in the results of monolingual and bilingual comparison for the picture
description task, bilinguals elicited more number of words than the monolinguals. Also, the
total number of correct responses were more in the bilingual group. With reference to the
type of paraphasias present, in the fluent aphasic group, both monolinguals and bilinguals
illustrated more number of formal and phonemic paraphasias. However, semantic paraphasias
were more commonly seen only in the fluent bilingual group. The non-fluent aphasics on the
other hand produced neologisms and phonemic paraphasias maximally. Mann Whitney U test
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
244
was done to support the above findings. The test results showed that in the non-fluent group,
monolinguals showed significantly more neologistic paraphasias (z=2.588 at p<0.05) and the
bilinguals showed significantly more semantic paraphasias (z=2.156 at p<0.05).
4.3.2 Across language comparison
The bilingual individuals with aphasia were studied further to see if their performance
in picture description task differed across Kannada (L1) and English (L2) language. Figure-
4.9 illustrates the graphical representation of the percentage of occurrence of the various
paraphasias for the bilingual group in English language in the picture description task.
It was seen that the mean percentage of total words for both fluent and non-fluent
bilinguals is almost equivalent in both L1 and L2. However, the mean percentage of total
correct responses was slightly greater for L2 than L1 in both fluent and non-fluent groups.
This difference of performance of bilinguals in L1 and L2 was not significantly different as
shown by Wilcoxon signed rank test (z=1.735 at p<0.05).
Figure 4.9: Percentage occurrence of the six kinds of paraphasias in the bilingual group in the
picture description task: English
Paraphasia in Picture Description (English)
NeoPhonUnrelMixedFormSem
Me
an
Pe
rce
nta
ge
Re
spo
nse
s -
Bili
ng
ua
ls
55
50
45
40
35
30
25
20
15
10
5
0
Groups
Fluent
Non-fluent
Sem- semantic paraphasia, Form- formal paraphasia, Mixed- mixed paraphasia, Unrel-
unrelated paraphasia, Phon- phonological paraphasia, Neo- neologisms
The results of across language comparison show that the performance of the
bilinguals did not differ much across the two languages. The type of paraphasias seen in the
two languages were also not very different. The fluent group exhibited more of semantic,
formal and phonemic paraphasias whereas, the non-fluent group showed more evidence of
neologisms and phonemic paraphasias. Wilcoxon Signed Rank test was conducted to know if
the two languages were significantly different in the type of paraphasias exhibited. Results
revealed no significant difference between the languages in the fluent group for all kinds of
responses. In the non-fluent group, no significant difference was seen for six types of
paraphasias.
Paraphasias in Bilingual Aphasia
245
4.3.3 Qualitative analysis
Amongst the type of paraphasias present, the semantic paraphasias were most
commonly seen in the fluent aphasic group. The monolingual and bilingual aphasic group
differed slightly in their performance. Individuals with Wernicke‟s aphasia presented with
most number of semantic and unrelated paraphasias in a narrative context. Anomics also
exhibited a high number of semantic paraphasias in the bilingual group.
In the non-fluent group, the Broca‟s and global aphasics produced more of non words
with either sound relation to the target word (phonemic paraphasias) or with no apparent
relation to the target word (neologism) in the picture description task.
Thus, the results of the present study establish the fact that paraphasias exist in all the
types of aphasia and across all the languages. The paraphasias can be similar or may vary in
the different languages of a bilingual individual with aphasia. The types of paraphasias also
varies across different language tasks namely, naming, repetition and picture description task
among the various subtypes of aphasia. However, the generalization of the results would be
difficult unless variables like severity of language impairment, large and equal sample size of
all the subtypes of aphasia, the literacy level and the pre-morbid language proficiency of the
different languages known by the bilingual aphasics etc are controlled and studied.
Discussion
Comparison of monolingual and bilingual group
The results of the study revealed a better performance of the bilingual individuals
with aphasia on all the three tasks studied namely, naming, repetition and picture description.
Higher scores obtained by the bilingual participants throw light on the fact that the word
retrieval ability required for naming and picture description task, and repetition abilities for
the repetition task after stroke are better retained in individuals knowing two languages rather
than one. This can be explained by the presence of two lexicons, one for each language i.e.,
dual representation of language in bilinguals which results in a better vocabulary.
With reference to the paraphasias present in the various types of aphasias, semantic
paraphasias were produced frequently by the anomics followed by conduction aphasics. The
high frequency of semantic paraphasias in anomics is an index of word-finding difficulty
(Kohn & Goodglass, 1985). This can be due to a breakdown in the semantic boundaries
between meaning-related words that were pre-morbidly undoubtedly well-known by the
anomics.
Comparison across tasks
Across tasks, it was observed that the semantic paraphasias were more common in the
naming and picture description task while being meagerly present in the repetition task. The
current study supports the views of Li and Williams (1990) who explain that aphasics tend to
exhibit significantly more indefinite terms, extended circumlocutions and perseverations in
the naming conditions. These behaviors are verbal strategies to compensate for inability to
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
246
produce the specific target word. These strategies are less employed during the repetition task
which is more target bound.
The results of the current study supports the findings of Li and Williams (1990) who
found phonemic attempts and revisions to be the most prominent in the repetition of
conduction aphasics.
Predominance of phonemic paraphasias in conduction aphasia was also seen in the
picture description task which is in accordance with the results of Kohn and Goodglass
(1985).
The results of the picture description task for the anomics is in accordance with
Williams and Canter (1987) who reported delayed responses and extended circumlocutions in
anomics. Williams and Canter (1987) also reported more neologisms in Wernicke‟s aphasia
which supports the finding of the current study with reference to the monolingual individual
with Wernicke‟s aphasia. However, the paraphasias seen in the bilingual individual with
Wernicke‟s aphasia disputes the results of William and Canter (1985).
Earlier findings by Gardner and Winner (1978) reported more meaning errors
whereas, Li and Williams (1990) reported of more linguistic errors (unrelated words and
jargon) in Wernicke‟s aphasics. However, in the current study, monolingual individuals with
Wernicke‟s aphasia produced unrelated, formal and phonemic paraphasias in the repetition
task. In contrast, the bilingual person with Wernicke‟s aphasia did show a reasonable
percentage of semantic and mixed paraphasias along with unrelated, formal and phonemic
paraphasias. These findings draw support from the interpretation of Goodglass and Kaplan
(1983) who explained the deficient repetition skills of Wernicke‟s aphasics on the basis of
their poor comprehension ability, resulting in a partial or complete distortion of auditory
image. The disparity in the results of Wernicke‟s aphasics can be attributed to the small
sample size in the present study and thus, the results of the present study cannot be
generalized to the entire Wernicke‟s aphasic population.
In comparing the present study with Gardner and Winner‟s 1978 study, similarities
appear in the Broca‟s aphasic group. Gardner and Winner (1978) found a majority of sound
errors, consisting of literal paraphasias, elaborations, simplifications and articulation errors,
in their Broca‟s and mixed anterior aphasics. Li and Williams (1990) also reported of
phonemic errors and omissions in Broca‟s aphasic group. The current study is in accordance
with the findings of Gardner and Winner (1978) and Li and Williams (1990). Results of the
current study revealed fewer phonemic errors in anomic aphasia which is in accordance with
the findings of Williams and Canter (1982).
Across language comparison
Across language comparison revealed a significant difference in the performance of
the fluent group. Hegde and Bhat (2007) studied paraphasias in a multilingual conduction
aphasia person. They reported phonemic errors in Kannada language which was more
familiar to the conduction aphasic subject and neologisms and real word jargon in the less
Paraphasias in Bilingual Aphasia
247
familiar languages (i.e., Hindi and English). In the current study, since the second language
(i.e., English) was more familiar to the participants than their mother tongue, the findings of
Hegde and Bhat (2007) get supported.
Overall, the results of the current study draw support from literature in the type of
paraphasias produced predominantly by the different subgroups of aphasia in both
monolingual and bilingual population. The generalization of the results is however guarded
as the number of subjects under each subgroup of aphasia was limited.
Conclusions
The present study was undertaken to investigate and compare the type of paraphasias
in monolingual and bilingual individuals with aphasia. Within the bilingual aphasia group, a
comparison of the type of paraphasias present across the two languages i.e., the mother
tongue (L1) and the second language (L2) was also studied.
To conclude, the influence of bilingualism was observed in the better performance of
individuals with aphasia knowing two languages over monolingual individuals with aphasia.
The type of paraphasias did not differ significantly between the monolingual and bilingual
individuals with aphasia in naming and repetition tasks. However, in the picture description
task, monolinguals produced more formal and phonemic paraphasias and the bilinguals
produced more of semantic paraphasias. Amongst the aphasic subgroups, anomic aphasics
were found to produce more semantic paraphasias; conduction aphasics produced abundant
phonemic paraphasias; Wernicke‟s aphasia presented with more number of semantic and
unrelated paraphasias and Broca‟s and global aphasics exhibited a high incidence of
phonemic and neologistic paraphasias. The bilingual individuals with aphasia performed
better in L2 than in their mother tongue. Different types of paraphasias were observed in the
two languages in the bilingual aphasics.
Thus, the results of the study show a lot of variation indicating that there could be
several parameters interacting differently in each case. Specific pattern of paraphasias is
precluded from emerging. The small number of sample may be accounting for this wide
variation. Considering the Indian scenario where the majority of the urban population is
either bi/multilingual, future research on the language aspects of bilingual individuals with
aphasia would add to both theoretical and clinical implications.
References
Arpita, B. (1997). Naming deficits in bilingual aphasics. Unpublished master‟s dissertation,
University of Mysore, India.
Chengappa, S. K., Bhat, S., & Damle, M. (2003). Paraphasias in multilingual Aphasia- A
single case study of Wernicke‟s Aphasia. Journal of the Indian Speech and Hearing
Association, 17, 66-70.
Dell, G.S., Schwartz, M.F., Martin, N., Saffran, E.M., & Gagnon, D.A. (1997). Lexical
access in aphasic and non-aphasic speakers. Psychological Review, 104, 801-838.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
248
Gagnon, D.A., Schwartz, M.F., Martin, N., Dell, G.S., & Saffran, E.M., (1997). The Origins
of Formal Paraphasias in Aphasics‟ Picture Naming. Brain and Language, 59, 450-
472.
Gardner, H., & Winner, E. (1978). A Study of Repetition in Aphasic Patients. Brain and
Language, 6, 168-178.
Goodglass, H. (1981). The syndromes of aphasia: Similarities and differences in
neurolinguistic features. Topics in Language Disorders, 1, 1-14.
Goodglass, H., & Kaplan, E. (1983). Boston Diagnostic Aphasia Examination. Philadelphia:
Lea & Febiger.
Goodglass, H., Kaplan, E., Weintraub, S., & Ackerman, N. (1976). The “tip of the tongue”
Phenomenon in Aphasia. Cortex, 12, 145-153.
Goodglass, H. (1993). Understanding Aphasia. Academic Press, Inc.
Hegde, M., & Bhat, S. (2007). Paraphasias in Multilingual Conduction Aphasia: A single
case study. Indian Journal of Applied Linguistics, 33 (2), 45-52.
Kohn, S. E., & Goodglass, H. (1985). Picture-Naming in Aphasia. Brain and Language, 24,
266-283.
Lesser, R. (1978). Linguistic investigation of aphasia. London: Edward.
Li, C.E., & Williams, S.E. (1990). Repetition deficits in three aphasic syndromes. Journal of
communication Disorders, 23, 77-88.
Martin, N., & Saffran, E. M. (1992). A computational account of deep dyslexia: Evidence
from a single case study. Brain and Language, 43, 240-274.
Shantala, M.S. (1997). Naming deficits in aphasics. Unpublished master‟s dissertation,
University of Mysore, India.
Williams, S. & Canter, G. (1982). The influence of situational context on naming
performance in aphasic syndromes. Brain and Language, 17, 92-106.
Williams, S. & Canter, G. (1987). Action-naming performance in four syndromes of aphasia.
Brain and Language, 32, 124-136.
Meta-Phonological Abilities in Monolingual & Bilingual Children
249
Meta-Phonological Abilities in Monolingual and Bilingual Children: A
Comparative Study
Samasthitha. S & S.P. Goswami1*
Abstract
Meta-linguistic ability is the ability to reflect upon and manipulate the structural features of
spoken language, treating language itself as an object of thought (Tunmer & Harriman, 1984). Meta-
linguistic abilities play vital roles in different stages of reading acquisition in which meta-
phonological skills are one of the facets of meta-linguistic ability. Meta-phonological abilities refer to
an individual’s explicit awareness and the ability to process and manipulate the speech sound
segments of words. This study was taken up to compare the meta-phonological and reading abilities
in monolingual and bilingual children. A total of 60 participants participated in the study (30 in
monolingual and 30 in bilingual group) in the age range of 8-9 years. All the participants were
screened to rule out signs and symptoms of learning disability, language deficits and academic
failures. RAP-K (Meta-phonological and reading sub-tests) was administered and the responses were
scored according to the instructions in the test manual. Results of the study revealed better
performances on meta-phonological and reading tasks by the bilingual group in comparison to their
monolingual counterparts. This study supports the view that the bilingual children appear to have an
advantage over the monolingual children in regard to phonological awareness and reading skills.
Introduction
“Meta-linguistic ability is the ability to reflect upon and manipulate the structural
features of spoken language, treating language itself as an object of thought” (Tunmer &
Harriman, 1984). This term was first used by Cazden (1974) to describe and explain the
transfer of linguistic knowledge and skills across languages. The nature of meta-linguistic
awareness is not defined clearly. Over the last few decades the studies in the area of meta-
linguistic skills have been attempted due to its relationship with reading acquisition.
The three main theoretical perspectives that are held in reference to reading acquisition
and meta-linguistic skills are:
1. Meta-linguistic skills are required for reading acquisition.
2. Meta-linguistic and reading skills are interactive.
3. Meta-linguistic skills are as a consequence of literacy acquisition.
Meta-linguistic abilities play vital roles in different stages of reading acquisition in which
meta-phonological skills are one of the facets of meta-linguistic ability. Meta-phonological
abilities refer to an individual‟s explicit awareness and the ability to process and manipulate
the speech sound segments of words. This requires non-lexical processing which has to look
beyond the meaning of a word to focus on the sound structure of the word. In the recent
years, predominantly the meta-phonological skills have been crucially linked to reading and
*Reader in Speech Pathology, All India Institute of Speech and Hearing, Mysore, India
Email:[email protected]
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
250
acquisition of reading. The relationship between meta-phonological skills, reading and
spelling ability has been well documented (Torgeson, Wagner & Rashotte 1994; Macdonald
& Cornwall, 1995; Troia, 2004). Meta-phonological ability or phonological awareness
encompasses many different skills, but all of these skills develop from the same underlying
ability (Anthony & Lonigan, 2004; Schatschneider, Francis, Foorman, Fletcher & Mehta,
1999).
Morphological awareness on the other hand reflects on the ability to understand and
correctly use small words, letters, and letter combinations that change the meaning of a word.
Strong correlations between phonological and morphological awareness are apparent
throughout the elementary years (Deacon & Kirby, 2004).
Hence, research in this field shows that both phonological and morphological
awareness are important for the acquisition of reading and spelling skills during the early
years of education.
The development of phonological awareness skills across children are learnt in the
same order. Awareness of rhyme and breaking words into syllables are two of the early meta-
phonological skills to emerge. There appears to be several influential factors affecting the
development of these skills in children. These factors include the environmental and
experiential factors. They also have a „protective influence‟ in the development of meta-
phonological and reading skills. Protective influences such as social factors and print
exposure may have an effect in the child‟s early literacy development (Scarbrough &
Dobrich, 1994). Linguistic aspects such syntax, semantics, learning to read does play key
roles in the development of meta-phonological skills. Tunmer (1989) reports that the
syntactic awareness, seems to be an important independent contributor to early reading skills.
Cognitive (verbal IQ such as abstract concepts, verbal factual knowledge, vocabulary and
verbal reasoning and inference, attention and verbal memory skills etc...) and non-cognitive
(home environment, social and family factors) predictors influence development of oral and
written language skills.
Apart from these factors print exposure i.e. in the form of alphabet books, storybooks
comics or magazines etc. influence early literacy experiences. An added relevant factor,
clearly strongly related to print exposure is how often parents read to their children.
(Scarbrough & Dobrich, 1994; Senechal et.al, 1998).
There are many documented empirical research reports on metaphonological and
reading skills and its relationship with bilingualism.
Meta-phonological abilities and reading achievement
Meta-phonological abilities begin to emerge before children start to learn to read, and
is a powerful and consistent predictor in beginning readers. Though phonological awareness
is a powerful predictor of reading achievement, both in the short and long term, acquisition of
alphabetic principles forms the foundation for children‟s rapid expansion of literacy skills.
Over the decades researchers have investigated the meta-phonological awareness and its
Meta-Phonological Abilities in Monolingual & Bilingual Children
251
association with reading skills. The findings have been equivocal. Few reports suggest that it
is a requirement for reading (Elkonin, 1973; Liberman et.al 1977; Bradley & Bryant, 1983).
Other researchers like Morais et.al (1979), Liberman and Mattingly (1985), Read et.al (1986)
report that these skills are as a consequence of learning to read. Few other researchers report
that meta-phonological skills and reading are interactive i.e. they are reciprocal. (Ehri, 1979;
Bryant & Goswami, 1987; Prefetti, Beck, Bell & Hughes, 1987). Studies have also been
documented keeping in mind the nature of othrography like alphabetic and non-alphabetic
scripts (Morais 1991; Prakash, 1993; Prakash, Rekha, Nigam & Karanth, 1993)
Bilingualism and meta-linguistic abilities
Meta-linguistic awareness is the primary variable mediating the positive effects of
bilingualism on academic achievement. There are reports suggesting that fluent bilingualism
results in increased meta-cognitive/meta-linguistic abilities which in turn facilitate reading
acquisition. Thus, resulting in higher levels of academic achievement. It seems reasonable to
assume that not all thought processes are enhanced through a bilingual experience and that
those cognitive tasks, which rely more on language, will benefit most from that experience
(Hamers & Blanc, 1983). The most commonly studied phenomenon in bi-literacy learning
that transfers across languages and enhances literacy learning among bilingual learners is
“meta-linguistic awareness” (Koda, 2008).
Meta-phonological abilities in bilinguals have been researched and there are several
reports suggesting that bilinguals have enhanced skills compared to the monolingual
counterparts. (Bruck & Genesee, 1995; Loizou, Stuart,2003; Surabhi Bharati,2004; Vanasse,
Begin-Bertrand, Courc, Lassonde & Beland,2005; Ibrahim, Eviatar, Judith ,2007; Dodd, So
& Lam,2008)
Various studies have been documented in the area of meta-phonological abilities and
reading skills. These studies have focused on the development and factors influencing the
acquisition of the meta-phonological abilities and reading in monolingual and bilingual
children. The literature basically refers to Western studies. Hence, in the Indian context, this
domain provides the motivational grounds for this study. Thus, the present study was
undertaken with the need to compare the meta-phonological abilities and reading skills on
various meta-phonological and reading tasks in Kannada speaking monolingual and
Kannada-English speaking bilingual children in the age range of 8-9 years.
Aim of the study
The current study primarily intended to investigate the meta-phonological abilities in
monolingual and bilingual children between the age ranges of 8-9 years.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
252
The objectives of the study were to:
Quantitatively study the gender differences across the two groups on the meta-
phonological and reading tasks.
Quantitatively study the performances of the two groups on the meta-phonological
and reading tasks.
Qualitatively analyze the performances of the meta-phonological and reading tasks.
Method
Participants
A total of 60 participants were included in the study. These participants were divided
in two groups:
Group- I:
30 (15 females and 15 males) school going children, with mother-tongue and exposure to
Kannada (L1) only. (Monolingual).
Group- II:
30 (15 females and 15 males) school going children with mother-tongue and exposure to
Kannada (L1) and exposure to English as L2. (Bilingual).
Age range
8.0 - 9.0 years
Ethical Standards used in the study for the selection of participants
Participants were selected by adhering to the appropriate ethical procedures.
Participants and/or parents were explained the purpose and procedures of the study, and an
informed verbal and/or written consent were taken. They were randomly selected based on
the inclusionary criteria/s.
Inclusionary Criteria
Participants were screened to rule out:
Signs and symptoms of learning disability.
Academic failures (based on teacher‟s report).
Language deficits, delayed speech and language milestones, history of any ear
pathology, emotional, behavioral and/or neurological deficits.
Participants were from middle socio-economic status.
Tools
Reading Acquisition Profile -Kannada (RAP-K) test, (Prema, 1997).
Purpose:
- To assess the meta-phonological and reading skills.
Meta-Phonological Abilities in Monolingual & Bilingual Children
253
The following sub-sections of the test were used in the current study:
a. Meta-phonological Test
i. Rhyme recognition.
ii. Syllable stripping.
iii. Syllable oddity (words).
iv. Syllable oddity (non-words).
v. Phoneme stripping.
vi. Phoneme oddity.
b. Reading Test
i. Syllable inventory [consonant-vowel (CV), consonant-consonant-
vowel (CCV), and consonant-consonant-consonant-vowel (CCCV)].
ii. Words and non-words.
iii. Geminates – words and non-words.
iv. Polysyllabic – words and non-words.
v. Arka – words and non-words.
vi. Anuswara– words and non-words.
Data collection (Measures and Materials)
Testing was conducted in a quite surrounding (room set up). The test was
administered individually with minimal distractors around. Table-1 shows the tests
administered on the two groups with their purposes.
Table 1: Tests administered and the purpose of the tests for the two groups.
Scoring: The scores were allotted based on the performance of the participants in the two
groups.
In the meta-phonological test
All the sub-sections‟ maximum score is 12 and participants received a score of 1 for every
correct response.
In the reading test
The scoring for sub-tests were done according to the instructions given in the test
manual for the reading tasks, each correctly read word was marked with a tick mark. Each
correctly read word was given a score of 1.
Groups Tests Purpose
Group-I
(Monolinguals)
Meta-phonological and reading
tests (sub-sections of Reading
Acquisition Profile Kannada
(RAP-K) test.
To assess the meta-
phonological and reading
skills.
Group-II
(Bilinguals)
Meta-phonological and reading
tests (sub-sections of Reading
Acquisition Profile Kannada
(RAP-K) test.
To assess the meta-
phonological and reading
skills.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
254
The obtained data were appropriately tabulated and subjected to statistical measures.
SPSS software (version 16.0.) package was used for statistical analysis. The tabulated scores
were used for obtaining the mean (M) and standard deviation (SD). Parametric tests were
utilized to obtain the significant difference measures. MANOVA was used to compare the
performance of female and male participants across the various meta-phonological and
reading tasks for both the groups. Repeated measures ANOVA were used to obtain the
significant difference scores between the two groups across the various meta-phonological
and reading tasks. The data was further subjected to Bonferroni test to find out significant
difference within each group, if any.
Results and Discussion
The results of the present study are summarized in the following tables,
Table 2: Mean and SD values for female and male participants of the two groups across the
meta-phonological tasks
Females Males Total
Tasks Groups Mean
SD
Mean
SD
Mean
SD
RR Monolingual
Bilingual
11.73
12.00
0.79
0.00
11.80
11.93
0.56
0.25
11.76
11.96
0.67
0.18
SS Monolingual
Bilingual
11.13
11.86
0.99
0.35
11.73
11.60
0.79
0.82
11. 43
11.73
0.93
0.63
SOW Monolingual
Bilingual
7.73
9.46
1.22
1.18
8.46
10.00
1.68
1.46
8.10
9.73
1.49
1.33
SONW Monolingual
Bilingual
7.46
9.33
1.30
1.23
7.80
9.33
1.93
1.39
7.63
9.33
1.62
1.29
PS Monolingual
Bilingual
7.80
9.13
1.42
1.50
7.73
8.53
1.38
1.18
7.76
8.83
1.38
1.36
PO Monolingual
Bilingual
5.46
7.33
1.06
1.23
6.13
7.93
1.64
1.53
5.80
7.63
1.39
1.40
Total Monolingual
Bilingual
51.33
59.13
5.21
3.94
53.66
59.33
6.56
5.39
52.50
59.23
5.94
4.64
Table 2 shows the mean and SD values for female and male participants of the two
groups across the meta-phonological tasks. The mean scores put forth clearly show that both
female and male participants of the two groups performed equally well on the meta-
phonological tasks. The scores obtained were not statistically significant at p < 0.05.
Meta-Phonological Abilities in Monolingual & Bilingual Children
255
Table 3: Mean and SD values of female and male participants of the two groups across the
reading tasks
Females Males Total
Tasks Groups Mean SD Mean SD Mean SD
CV Monolingual
Bilingual 36.60
36.33
2.58
2.19
35.66
36.73
3.57
2.65
36.13
36.53
3.10
2.40
CCV Monolingual
Bilingual 7.53
7.80
1.30
1.14
6.93
8.20
1.43
1.08
7.23
8.00
1.38
1.11
CCCV Monolingual
Bilingual 6.13
6.66
1.68
0.81
6.20
7.06
1.74
1.22
6.16
6.86
1.68
1.04
W Monolingual
Bilingual 18.66
19.06
1.23
1.03
18.53
18.80
1.59
1.47
18.60
18.93
1.40
1.25
NW Monolingual
Bilingual 17.60
18.66
1.54
0.97
17.73
18.53
1.66
1.40
17.66
18.60
1.58
1.19
GW Monolingual
Bilingual 9.20
9.33
1.01
0.72
8.73
9.13
1.53
1.06
8.96
9.23
1.29
0.89
GNW Monolingual
Bilingual 8.40
8.66
1.12
1.11
8.20
8.53
1.65
1.35
8.30
8.60
1.39
1.22
PW Monolingual
Bilingual 17.33
17.46
1.95
1.59
16.80
17.33
1.97
2.79
17.06
17.40
1.94
2.23
PNW Monolingual
Bilingual 16.93
17.46
2.25
1.76
16.53
17.06
3.41
2.60
16.73
17.26
2.85
2.19
ARW Monolingual
Bilingual 7.66
7.46
1.79
1.18
6.73
8.13
2.37
1.45
7.20
7.80
2.12
1.34
ARNW Monolingual
Bilingual 5.86
6.40
1.95
1.63
5.60
7.13
3.08
1.68
5.73
6.76
2.54
1.67
AW Monolingual
Bilingual 7.00
8.00
1.85
1.46
6.40
8.20
2.97
1.61
6.70
8.10
2.45
1.51
ANW Monolingual
Bilingual 8.00
9.20
2.59
1.20
7.20
8.66
3.12
1.87
7.60
8.93
2.84
1.57
Total Monolingual
Bilingual 166.93
172.53
18.11
14.33
161.26
173.53
25.99
19.73
164.10
173.03
22.20
16.95
Table 3 shows the mean and SD values of female and male participants of the two
groups across the reading tasks. It is apparent that both female and male participants of the
two groups performed equally well on the reading tasks. The scores were not statistically
significant at p < 0.05 between the performance of the female and male participants in both
groups on reading tasks.
This leads to the contemplation that both female and male participants were equally
competent in learning the meta-phonological skills and reading skills in the age range of 8-9
years. This could be attributable to fact that the variables like language learning environment,
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
256
literacy instruction, socio-economic status, age, nurturing, parent scaffolding, may not be an
influential factor across females and males. Thus, these factors do enhance learning skills
equally across gender.
In support with the literature, there seems to be no gender-differences across the two
tasks and the two groups. This is in accord with the research work done by Burt, Holm and
Dodd (1999) who reported that females and males performed equally well on the meta-
phonological tasks. However, the socio-economic status (SES) affected the performances on
majority of the tasks. Further the older children exhibited better phonological awareness in
contrast to younger peers.
The present study also showed similar findings where female and male participants
performed like-wise on the meta-phonological tasks. This result is supported by the findings
of Dodd and Carr (2003). Evidence also comes from the research work by McDowell,
Lonigan, Goldstein (2007) who reported that age, speech sound accuracy, and vocabulary
each contributed unique variance to the prediction of phonological awareness in both females
and males.
However, the results of the current study contraindicated the results by Moura,
Mezzomo and Cielo (2009) in which they reported that on the phonemic segmentation of
words with six phonemes and phoneme reversion of words with two or three phonemes,
female participants performed better than male participants. However, this suggests that the
phonological awareness stimulation program might have lead to better learning of specific
skills like segmentation, reversion. Yet such a learning experience might not be generalized
to other skills like stripping, oddity.
Table 4: Mean and SD for the total scores obtained by the groups on meta- phonological
tasks
Tasks
Monolinguals Bilinguals Total
Mean SD Mean SD Mean SD
RR 98.05 5.65 99.72 1.52 98.88 4.19
SS 95.27 7.79 97.77 5.33 96.52 6.73
SOW 67.50 12.44 81.11 11.14 74.30 13.57
SONW 63.61 13.57 77.77 10.79 70.69 14.10
PS 64.72 11.51 73.61 11.38 69.16 12.20
PO 48.33 11.66 63.61 11.67 55.97 13.90
Total 72.91 8.25 82.26 6.44 77.59 8.72
Table 4 shows the mean and standard deviation obtained by the monolingual and
bilingual groups on meta-phonological tasks. The table-4 clearly exemplifies that the
bilingual group performed better than the monolingual group. They demonstrated better
scores particularly on the SOW, SONW, PS and PO tasks. The results also showed a
statistically significant difference at {Mono= [F= (5,145) = 197.668, p < 0.05], Bi= [F=
(5,145) = 103.093, p < 0.05]} on meta-phonological tasks in both the groups.
Meta-Phonological Abilities in Monolingual & Bilingual Children
257
The research findings in equivalence and in disparity with that of the current study are
discussed in conjunction to the Western literature and also bearing in mind the Indian
perspective. The results of the study indicated that bilingual children performed better on
meta-phonological compared to their monolingual peers. This finding is supported by
research reports by Karanth and Prakash (1996) and Rekha (1997) reported that beginning
readers of a non-alphabetic script (Kannada) revealed that phonological awareness is greatly
influenced by alphabet like features present in the orthography and not by rhymes. Also,
Read, Zhang, Nie, and Ding (1986) reported that some aspects of phonological awareness
may not be because of maturational factors but may be a consequence of learning an
alphabetic orthography. The current study also throws insight into findings suggesting that
bilingual children perform better on the meta-phonological tasks which in turn promotes
learning to read alphabetic script which is further supported by a study reported by Prakash
et.al (1993). There also seems to be a bilingual enhancement effect. Analogous research has
been supported by researchers like Loizou, Stuart 2003; Shwartz, Geva, Share and Leikin,
2005, Ching-ning Chien, Kao and Li Wei, 2008; Dodd, So and Lam, 2008.
Table 5: Mean and SD for the total scores obtained by the groups on reading tasks
Tasks
Monolinguals Bilinguals Total
Mean SD Mean SD Mean SD
CV 90.33 7.76 91.33 6.00 90.83 6.89
CCV 72.33 13.81 80.00 11.14 76.16 13.03
CCCV 61.66 16.83 68.66 10.41 65.16 14.31
W 93.00 7.02 94.66 6.28 93.83 6.66
NW 88.33 7.91 93.00 5.95 90.66 7.33
GW 89.66 12.99 92.33 8.97 91.00 11.15
GNW 83.00 13.93 86.00 12.20 84.50 13.07
PW 85.33 9.73 87.00 11.18 86.16 10.43
PNW 83.66 14.25 86.33 10.98 85.00 12.68
ARW 72.00 21.23 78.00 13.49 75.00 17.89
ARNW 57.33 25.45 67.66 16.75 62.50 21.98
AW 67.00 24.51 81.00 15.16 74.00 21.40
ANW 76.00 28.47 89.33 15.74 82.66 23.78
Total 82.05 11.10 86.51 8.476 84.28 10.04
Table 5 shows the mean and standard deviation obtained by the monolingual and
bilingual groups on reading tasks. The overall total mean percentage scores across the tasks
for the monolingual group was 82.05 (SD=11.10); bilingual group got a total of 86.51
(SD=8.47). It is also apparent from table-5 that the mean performances of the participants in
CV, W, NW, GW, GNW, PW and PNW were almost same with not much difference in their
mean values. Alternatively the mean values for CCV, CCCV, ARW, ARNW, AW and ANW
showed higher differences. The scores also were statistically significant at {Mono= [F=
(12,348) =31.605, p < 0.05], Bi= [F= (12,348) =50.041, p < 0.05]}.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
258
Figure 1: Mean total scores of the monolingual and bilingual group across the meta-
phonological and reading tasks
Figure 1 depicts the performances of the two groups across the meta-phonological and
reading tasks.
With respect to reading abilities, it is evident from the mean scores that across the
reading tasks, the bilingual children performed better than monolingual children. This finding
has been supported by Morais et.al (1979, 1986) and Read (1986) reported that individuals
who did not have direct experience with alphabetic orthography were unable to carry out
phonological segmentation tasks. Also, research reports on Indian children by Prakash and
Mohanty (1989), Malini (1996), Rekha (1996) indicate that alphabetic script may facilitate
phonological awareness. Bilingual children revealed slightly better scores when compared to
monolingual children on the non-word reading tasks, this finding is supported by the research
report Prema (!997) where she reported that children might show the absence of logographic
reading in Kannada. Investigations by Vasanta, (2004); Hamilton and Gillon (2006) and the
result of the present study suggests that phonological awareness knowledge interacts during
the reading words and non-words in such a way that children exposed to more than one
language and formal literacy instruction are more affluent in accessing phonological
knowledge as compared to children with less able-bodied orthographies.
However, there are contrastive research findings. Mann (1986). He documented that
Japanese children who were not exposed to alphabetic script were able to successfully
complete the phonological segmentation tasks by the time they reached IV grades. Similarly
Morais (1991) claims the fact that entire writing system need not be alphabetic for
development of phonological awareness. A non-alphabetic script would allow development
of phonological awareness skills to certain degree depending on specific orthographic
features of the script.
Apart from these quantitative measures, qualitative analysis was carried out to study
the nature of performances exhibited by both the groups across the two tests. Qualitative
analysis gives an insight about the performances and errors which were observed in both the
groups. The errors which were not tapped out by the quantitative analysis.
For the meta-phonological tasks, it was observed that on the oddity tasks, children in
both the groups used strategies like reading the articulatory movements, counting the number
of syllables in the word, in order to retrieve the memorized syllables to pick the odd one out.
Meta-Phonological Abilities in Monolingual & Bilingual Children
259
A consistent finding in both the groups was that they had difficulty in deleting the middle
syllable on the syllable stripping tasks. This is further supported by Goswami (1994). She
attributed to the inability to perceive the intrasyllabic difference.
On the reading tasks, it was observed that as the complexity of the stimuli increased
there was deterioration in the performances was observed. They simplified the articulation of
the target words. It was also observed that the children (in the monolingual group) did not
exhibit realization for „arka‟ where, it is uttered first, but written following the syllable across
the words and non-words, but managed to read the frequently occurring words. For the
„anuswara‟ word and non-word reading tasks, participants of the monolingual group read the
„anuswara‟ (O) as zero when it preceded the syllable instead of reading it as /m/ or /n/or /ŋ/
depending on the word context. This could be attributed to the fact that the frequency of
occurrence and exposure to „anuswara words‟ between 8-9 years is comparatively less. It as
noticed that some participants read non-words with ease. This can be supported by the results
reported by Prema (1997) where she stated that children might show the absence of
logographic reading in Kannada.
Conclusions
The female and male participants performed equally on the meta-phonological and
reading tasks however there were subtle differences in the mean percentage scores. This
shows that SES, literacy instruction, language learning environment etc. are all independent
of the gender variable between 8-9 years. There appears to be developmental trend in the
acquisition of meta-phonological skills. The findings of the current study also suggest that the
cognitive load entailed by the children is equal amongst female and male participants in the
acquisition of meta-phonological skills. Reading skills on the other hand, should be viewed
holistically, taking into consideration other aspects like communicative abilities,
environmental factors, mental status of the child, literacy level of parents, media, apart from
the meta-phonological skills which does play a vital role in reading acquisition.
Bilingual group performed better than the monolingual group on the meta-
phonological and reading tasks suggesting that they have an advantage in fine-tuning and
growth of meta-phonological and reading skills.
On a concluding remark, it is apparent from the findings that though there are
differences in the performances of the monolingual and bilingual children on meta-
phonological tasks, these differences no where hamper but promote literacy skills.
Implications
The results of the present study are of clinical significance as it has been observed that
rhyme recognition and syllable awareness are the earliest indicators of reading abilities in
both the groups. This trend has to be kept in mind while assessing and planning treatment
program for children with developmental disorders such as in children with learning
disability, specific language disorders, and phonological disorders.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
260
This study also gives scope for professionals assessing and planning intervention
programs for children with developmental disorders who are exposed to more than one
language such as Kannada, a semi-syllabic script (L1), and English, alphabetic script (L2).
Thus, it is imperative in the Indian context where bilingualism is a common phenomenon,
assessment of children in L1 and L2 becomes crucial. Hence, it is proposed that separate
assessment protocols should be made in L1 and L2 in order to arrive at clear consensus about
the assets of bilingual children.
References
Bialystok, E. (2001). Bilingualism in Development: Language, Literacy, and Cognition. U.K:
Cambridge University Press.
Bruck, M., Genesee, F. (1995). Phonological awareness in young second language learners.
Journal of Child Language, 22(2), 307-324.
Bryant, P.E., & Bradley, L. (1980). Why children sometimes write words which they do not
read. In U.Firth (Ed.), Cognitive processes in spelling (pp. 495-516). London:
Academic Press.
Bryant, P.E., & Bradley, L. (1981). Visual memory and phonological skills in reading and
spelling backwardness. In Catts, H.W. & Kamhi, A.G.(Ed.) The Linguistic basis of
reading disorders, LSHSS, 17, (1986) (pp. 329-534).
Bryant, P.E., & Goswami, U. (1987). Beyond Grapheme-Phoneme Correspondence. In
B.Gelder & J. Morais (Ed.), Speech and reading- A comparative approach (pp.249-
269). U.K: Erlbaum.
Burt, L., Holm, A., & Dodd, B. (1999). Phonological awareness skills of 4 -year -old British
children: an assessment and developmental data. International Journal of Language
& Communication Disorders, 34, 311-335.
Cazden, C.B. (1972). Child Language and Education. New York: Holt, Rinchart &
Winston.
Chien, C., Kao, L., & Wei, L. (2008). The Role of Phonological Awareness Development in
Young Chinese EFL Learners. Language Awareness, 17(4), 271-288.
Dodd, B., & Carr, A. (2003). Young Children's Letter-Sound Knowledge. Language,
Speech, and Hearing Services in Schools, 34, 128-137.
Hakes, D.T. (1980). The development of meta-linguistic abilities in children. Berlin:
Springer - Verlag.
Hamilton, E., Gillon, G. (2006). The phonological awareness skills of school-aged children
who are bilingual in Samoan and English. Advances in Speech Language Pathology,
8(2), 57-68.
Meta-Phonological Abilities in Monolingual & Bilingual Children
261
Hogan, T.P., Catts, H. W., & Little, T.D. (2005). The Relationship between Phonological
Awareness and Reading. Language, Speech, and Hearing Services in Schools, 36,
285-293.
Ibrahim, R., Eviatar, Z., Judith, A. (2007). Meta-linguistic Awareness and Reading
Performance: A Cross Language Comparison. Journal of Psycholinguistic Research,
36 (4), 297-317.
Karanth, P., & Prakash, P. (1996). A Developmental Investigation of Onset, Progress and
Stages in the acquisition of literacy, India. Research project (F.2-17/189/eric/1147),
National Council of Education, Research and Training, New Delhi, India.
Karanth, P. (2003). Cross-linguistic study of acquired reading disorders: Implications for
reading models, disorders, acquisition and teaching. New York: Kluwer Academic.
Loizou, M., Stuart, M. (2003). Phonological awareness in monolingual and bilingual English
and Greek five year olds. Journal of Research in Reading, 26, 3-18.
McDowell, K.D., Lonigan, C.J., Goldstein, H. (2007). Relations among Socioeconomic
Status, Age, and Predictors of Phonological Awareness. Language, Speech, and
Hearing Services in Schools, 50, 1079-1092.
Moura, S.R., Mezzomo, C.L., Cielo, C.A. (2009). Phonemic awareness stimulation and its
effects regarding the variable gender. Pro Fono, 21, 51-56.
Pence, K.L. (2007). Assessment in Emergent Literacy. San Diego, CA: Plural Publishing
Inc.
Prema, K.S. (1997). Reading Acquisition Profile in Kannada- (RAP-K). Unpublished
doctoral thesis, University of Mysore, Mysore, India.
Tunmer, W.E., Pratt, C., Herriman, M.C. (1984). Metalinguistic awareness in children.
Berlin: Springer-Verlag.
Van Kleeck, A. (1982). The Emergence of linguistic awareness and cognitive framework.
Merrill-Palmer Quarterly, 28(2), 237-265.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
262
Dementia Assessment Battery – Kannada
Sunil Kumar Ravi & K.C. Shyamala*
Abstract
Dementia is a common clinical syndrome characterized by a decline in cognitive function and
memory from previously attained intellectual levels, which is sustained over a period of months or
years. The core symptoms of cognitive dysfunction in dementia include memory, language, praxis,
gnosis, executive functioning, and personality and behavioral changes. Deficits in cognitive
functioning result in various speech and language impairments in these individuals. Very limited
number of test batteries is available for speech – language pathologists to assess the speech and
language skills in individuals with dementia. Therefore, there is a great need to develop a test battery
to assess this population especially in Indian context. The present study was taken with the aim of
developing a test battery in Kannada to assess individuals with dementia. The test battery includes
four domains such as memory, linguistic expression, linguistic comprehension, and visuospatial
construction skills. This test battery was administered on 30 normal subjects in each of the age ranges
of 20 – 40 years, 40 – 60 years and above 60 years, and also administered on 10 dementics in the each
severity category of mild, moderate and severe. The normative data of all subtests included the mean
and standard deviation (S.D) scores was calculated. Results of the present study also revealed
significant deficits in individuals with dementia on all domains. Performance of mild dementics was
near normal in most of the subtests like spontaneous speech, reading comprehension, following
commands, etc. Moderate dementics had more difficulty in tasks like working memory, episodic
memory, picture naming, generative naming, and generative drawing, etc. Severe dementics’
performance was poor in all the subtests especially due to poor spontaneous speech skills, and poor
intelligibility of speech and these subjects had more difficulty in all the memory tasks and language
expression tasks like naming. These results showed that there is a significant difference in the mean
and S.D of normal groups and dementia groups and the performance of dementia group deteriorated
as the severity increased.
Keywords: dementia, Alzheimer’s disease, memory assessment, language assessment, visuospatial
construction.
Introduction
Dementia is a common clinical syndrome characterized by a decline in cognitive
function and memory from previously attained intellectual levels, which is sustained over a
period of months or years. The deterioration is of such severity that it impairs the affected
individual’s ability to work and to perform activities of daily living, including
communication. Cummings and Benson (1992) state that at least three of the following five
areas of mental activity must be involved: 1. Language; 2. Memory; 3.Visuospatial skills; 4.
Emotion or personality; and 5. Cognition (ex: abstraction, calculation, and judgment).
The Diagnostic and Statistical Manual of Mental Disorders—Fourth Edition (DSM –
IV); American Psychiatric Association, 1994 states that the essential feature in dementia is
impairment in short term and long term memory. This deficit in memory may also be
____________________________________________
Professor of Language Pathology, All India Institute of Speech and Hearing, Mysore, India
e-mail: [email protected]
Dementia Assessment Battery - Kannada
263
associated with one or more features like, aphasia, apraxia, agnosia, impairment in abstract
thinking, impaired judgment, and personality changes.
The nature and course of dementia will vary depending upon the etiology. Most
dementias are progressive, but some are static. Dementia can be caused by a variety of
conditions like diseases, infections and infarcts. The most commonly occurring cause is
Alzheimer’s disease accounting for 50 to 60% of all the patients with dementia. Vascular
dementias (dementias caused by multiple infarcts) are seen in 20% of the dementia patients.
Alzheimer’s dementia and vascular dementia co-occur in approximately 15% of this sample,
and other conditions such as Pick’s disease, Parkinson’s disease (PD), Progressive
Supranuclear Palsy (PSP), and Creutzfeldt-Jacob disease (CJD), account for the reminder of
the irreversible dementias.
Although the core features are the same for all dementias, the onset and course may
vary. Alzheimer’s disease is usually insidious in onset. The time from the onset of clinical
features to presentation for evaluation varies considerably and depends on the etiology of the
dementia, as well as personal and social factors, including individual and cultural attitudes
and beliefs about aging, premorbid personality, and intelligence. The core symptoms of
cognitive dysfunction in dementia, as defined in the DSM-IV-TR (2004) include memory,
language, praxis, gnosis, executive functioning, and personality and behavioral changes.
The progression of dementia maybe complicated by the associated medical illness
such as stroke, TBI, etc. complicating the course of Alzheimer’s disease. In general,
degenerative dementias have an insidious onset and are gradually progressive. The pattern
may initially include periods of more gradual decline, followed by a more rapid progression.
Vascular dementia tends to have an abrupt onset and a more stepwise pattern, associated with
further vascular insults, but may have a gradual and progressive course. Radiation induced
dementia may present months after radiation exposure and may have a progressive course.
Considering the incidence of dementia related communication disorders is increasing
in India, there is a need to develop test batteries for identification and diagnose dementia by
speech language pathologists (SLPs). SLPs need to have a test battery which will be used for
differential diagnosis between normal aging and several types of dementia. Currently, there
are no specific treatment programs available for individuals with dementia as there is a lack
of information on language deficits in individuals with various types of dementia. Therefore,
there is a need to develop a test battery on which we can plan individual treatment
programmes depending upon the type of dementia and severity of dementia in early stages
itself. There are some assessment tools available in western countries but no suitable tests are
available for dementia in Indian context.
Changes in communicative function with advancing age may signal the beginning of
serious neurological conditions (e.g., dementia) that significantly impact functional
independence (Bayles & Kaszniak, 1987). The ultimate purpose of the research studies are to
develop sensitive and reliable measures of cognitive linguistic change that can be used to
periodically assess speech, language, and cognitive abilities as part of a comprehensive adult
illness prevention and health maintenance protocol. Early detection of cognitive-linguistic
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
264
disorders with sensitive, reliable assessment tools may be expected to result in timely
intervention thereby reducing disability and enhancing rehabilitation.
Numerous studies have investigated diagnostic markers for early identification of
dementia and mild cognitive decline with advancing age, (Albert, Blacker, Moss, Tanzi, &
McArdle, 2007; Cunje, Molloy, Standish, & Lewis, 2007). Mainly, these protocols focus on
dementia rating scales and neuropsychological test batteries to assess normally functioning
older adults and individuals with cognitive deficits. The long-term goal of these studies was
to identify measures sensitive to cognitive decline.
Language measures such as verbal fluency may be sensitive to cognitive decline
(Ostberg, Fernaeus, Hellstrom, Bogdanovic, & Wahlund, 2005), and studies have explored
language as well as cognitive variables for signs of pathological aging. For example, the
Barnes Language Assessment (Bryan, Binder, Dann, Funnell, Ramsey, & Stevens, 2001) was
developed to assess language and associated cognitive abilities in older individuals. It
includes tasks that examine language expression, comprehension, reading, writing, executive
function, and memory that were adapted from existing measures.
Confrontation naming deficits may or may not be evident early in the course of DAT,
but are invariably present by the later stages. When asked to name real items or items that
are pictured in outline drawings, DAT patients often are completely unable to name them, or
they commit semantic errors, such as producing the name of the super ordinate category to
which the item belongs (e.g. animal for horse) or an incorrect name from the same semantic
category (e.g. cow for horse). This impairment in naming ability becomes more severe as
dementia progresses. In contrast to DAT, confrontation naming remains relatively unaffected
in some other dementing disorders. Bayles and Tomoeda (1983) reported that both mildly
and moderately demented HD (Huntington’s disease) patients were unimpaired on the Boston
naming Test. Difficulty in naming or word retrieval has been observed to be the most obvious
early symptom of dementia, regardless of cause, and has been found to occur before other
language changes associated with the syndrome are measurable (Bayles, Tomoeda, Kaszniak
& Troset, 1990).
In the early stages of probable Alzheimer’s disease (PAD), semantic difficulties are
thought to be more influential on naming responses than perceptual responses and have
attracted more investigation. Semantically-related errors predominate in naming, suggesting
that there is an underlying semantic disorder (Bayles and Tomoeda 1983). Explanations for
this have followed those proposed by Warrington and Shallice (1979) of either impaired
access to semantic information or a degraded semantic store. Both the impaired access and
impaired storage accounts have been applied to naming behavior in PAD.
Nicholas, Obler, Au, & Albert, (1996) compared DAT patients performance on the
Boston naming test (BNT) with elements of empty speech in narrative description of the
cookie theft picture in an attempt to evaluate the claim that discourse incoherence could be
attributed to anomia. The authors reported a significant negative correlation between the
score of DAT patients on the Boston Naming test and the use of indefinite terms (e.g. thing
and stuff) and significant positive correlation between the BNT and the production of content
Dementia Assessment Battery - Kannada
265
elements (i.e. references to characters and actives in cookie theft stories). They concluded
that the naming deficit did not underlie the emptiness of discourse presumably because many
other measures of discourse emptiness (e.g. paraphasias pronouns with accidents and deictic
terms) did not correlate with the BNT scores.
Freedman and Berman (1986) have reported that the impaired performance of the
DAT patients on these tasks presumably results from deficiency in cognitive flexibility that is
required to alternate responses and shift mental set.
An overview of the appropriate case history, medical laboratory studies and
neuropsychological tests and behavioral ratings as well as language and communication
measures will be presented based on this collaborative perspective. As specialists in
communication disorders, SLPS are often asked to provide consultation regarding
communication competence of dementia patients. Although the results of the communication
assessment may be used for differential diagnosis, it is more likely that results will be used to
evaluate the patient’s progression in the course of the dementia.
Neuropsychological performance tests provide objective and precise measurement of
cognitive function; however, a major problem with their use is that they do not relate directly
to functioning in practical situations, i.e., they lack ecological validity. For this reason,
behavioral ranting scales that illustrates/actual behavioral features are an important addition
to the assessment protocol. A multimodal assessment provides various perspectives of the
patient’s behavioral abilities necessary for a comprehensive description of functional
behavior.
The Brief Cognitive Rating Scale (BCRS) (Reisberg, 1983) is a rapid, structured
instrument for assessing cognitive decline, regardless of etiology. Items are organized into
five categories or axes. These include concentration and calculating ability, recent memory,
remote memory, orientation and Functioning and self care. Within each axis behaviors are
scored from one to seven, with one being the least impaired and seven being the most
impaired. Each score is related to distinguishable levels of functioning within the category.
The Functional Assessment Stages (FAST) distinguishes 15 distinct progressive
characteristics of the disease (Reisberg, et al, 1984). These characteristics can be related to
the seven stages within Axis V, functioning and self care, of the brief cognitive rating scale.
It is proposed that patients with uncomplicated DAT typically proceed on a linear course
through the characteristics of decline. Recognition of these distinct stages of DAT is clearly
an advance in enabling clinicians to identify the precise magnitude of the impairment as well
as help in the differential diagnosis of DAT. FAST is particularly useful in the later stages of
the disease when other measures may not carefully identify the magnitude of the breakdown.
The Working Party on Dementia Report (College of Speech and Language Therapists,
1993) reported that therapists working with people who have dementia use assessments
designed for other client groups, such as people with aphasia. Therefore, the working party
suggested that there exists a ‘clinical need for the development of conceptually strong and
clinically appropriate evaluations of a diagnostic and/or descriptive nature’.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
266
Such an assessment has already been developed and standardized in the USA and UK
– Arizona Battery for Communication Disorders of Dementia (ABCD) (Bayles & Tomoeda,
1993) which can be used by SLPs to diagnose and assess the severity of dementia. This
fulfills the criteria of working party in its conceptual strength and clinical appropriateness and
in its primary aim to quantify the linguistic communication deficits associated with AD.
Furthermore, the ABCD was designed to allow a differential diagnosis between normal aging
and mild AD and between mild AD and moderate AD.
The ABCD comprises 14 subtests which comprehensively assess linguistic
comprehension and expression, verbal episodic memory, visuospatial construction and
mental status. It also contains four tasks for identifying individuals with problems that might
invalidate the test results, such as a hearing problem, illiteracy, a visual field defect, and
visual agnosia. Tests of memory, mental status and visual perceptual and construction skills
were included in the battery because by definition, dementing diseases produce deficits in
memory, mental status and perception (Reisberg, Ferris, DeLeon & Crook, 1982; Cummings
& Benson, 1992). Integrity of these processes and types of knowledge are essential for
normal communicative function to occur.
Need for the study
There are some assessment tools available in western countries but no suitable tests
are available for dementia in Indian context. So, we need to have appropriate tests to identify
the persons with dementia in Indian context. As the review of literature suggests, individuals
with dementia have language deficits along with other cognitive deficits. Therefore, there is a
need to develop a test battery to assess the language deficits and other cognitive deficits such
as memory, etc… in these subjects. Considering that the incidence of dementia related
disorders is increasing in India, there is a need to develop specific test batteries for
identification and diagnosis of dementia. We need to have a test battery which will be used
for differential diagnosis between normal aging and the several types of dementia. Currently,
there are no specific treatment programs available for individuals with dementia as there is a
lack of information on language deficits in individuals with various types of dementia.
Therefore, there is a need to develop a test battery on which we can plan initiation of
individual treatment programmes depending upon the type of dementia and severity of
dementia in early stages itself.
Aim of the study
1. To develop an assessment test battery for Kannada speaking individuals with
dementia.
2. To standardize the test material on normal population and in individuals with various
types of dementia.
3. To profile the language deficits of individuals with various types of dementia.
Dementia Assessment Battery - Kannada
267
Method
The present study was undertaken to develop and standardize a test battery on normal
population and also on dementia population. Four groups of subjects were considered in
which three groups of subjects were normal individuals (young adults, adults and geriatric
groups) and another group included individuals with dementia.
Normal subjects were tested to form a baseline which will be considered as normative
for this test. Normal 30 young adult subjects in the age range of 20 – 40 years, normal 30
adults in the age range of 40 – 60 years and normal 30 old subjects aged above 60 years were
selected based on the following inclusion criteria: native Kannada speakers, no history of
major neurological or psychiatric illness or of alcoholism or drug abuse, all the subjects were
evaluated for their mental status on Mini Mental Status Examination (Folstein, Folstein &
McHugh, 1975) and all passed with a score of above 23. Finally, individuals who received
rating of 1 or 2 on Brief Cognitive Rating Scale (Riesberg, 1983) and rating of 1 (normal/ no
cognitive decline) on Functional Assessment Stages (Reisberg, Ferris, Anand, 1984) were
included in this group.
30 subjects with various types of dementia, 10 subjects in each group, mild, moderate
and severe dementia based on Functional Assessment Stages (Reisberg, Ferris, Anand, 1984)
and Brief Cognitive Rating Scale (Riesberg, 1983) were taken for the present study. The
subjects with dementia were identified through local hospitals, dementia associations and
dementia clinics in Mysore and Bangalore cities. The selection criteria for this group were:
age over 50 years, diagnosed as having dementia by a neurologist/ psychiatrist according to
NINCDS – ADRDA, native Kannada speaker, adequate hearing and vision, reported history
of gradual deterioration in cognitive abilities. All the subjects were evaluated for their mental
status on Mini Mental Status Examination (Folstein, Folstein & McHugh, 1975) and all failed
with a score of below 23. Finally, individuals who received rating above 2 on Brief Cognitive
Rating Scale (Riesberg, 1983) and rating above 1 (normal/ no cognitive decline) on
Functional Assessment Stages (Reisberg, Ferris, Anand, 1984) were included in this group.
Tests/ tools
1. Mini Mental Status Examination (MMSE) (Folstein, Folstein & McHugh, 1975)
2. Brief Cognitive Rating Scale (BCRS) (Reisberg, 1983)
3. Functional Assessment Stages (FAST) (Reisberg, Ferris & Anand, 1984)
4. Dementia Assessment Battery – Kannada.
Dementia Assessment Battery was developed in Kannada. The DAB – K (Dementia
Assessment Battery – Kannada) comprised of 17 subtests which comprehensively assessed
memory, linguistic expression, linguistic comprehension, and visuospatial construction. The
subtests were selected from different language tests, mainly from Arizona Battery of
Communication Disorders of Dementia, which was standardized on USA and UK population.
The overview of DAB – K (various domains, subtests under each domain) is given in Table
1.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
268
Table 1: Domains and subtests of DAB – K.
Memory Linguistic
Expression
Linguistic
Comprehension
Visuospatial
construction
Episodic Memory Picture Naming Comparative
Questions
Generative
Drawing
Working Memory Generative
Naming
Following
Commands
Figure Copying
Semantic
Memory
Sentence
completion
Reading
Comprehension
of Sentences
Delayed Story
Telling
Responsive
Speech
Reading
Commands
Spontaneous
Speech
Repetition
Results and Discussion
The main objective of the study was to develop and standardize the Dementia
Assessment Battery – Kannada test on normal population and also on individuals with
dementia. For this, the test material was administered on different groups of subjects and the
results of the each group on each subtest are given in this section.
The Dementia Assessment Battery – Kannada consists of four domains, memory,
linguistic expression, linguistic comprehension, and visuospatial skills with several subtests
in each domain. The mean and standard deviation (S.D) of each subtest and also for each
domain are calculated and they are given in this section.
Overall Results of each domain
Domain – 1: Memory
This domain consists of four subtests namely episodic memory, working memory,
semantic memory and delayed story recall subtests. The maximum total score of this domain
is 100. The mean and standard deviation scores of each group of subjects are given in Table
2.
Table 2: Mean and S.D of each group on memory domain.
Group
Young
adults
(20 – 40
years)
Adults
(40 – 60
years)
Normal
Geriatrics
(above 60
years)
Mild
dementia
Moderate
dementia
Severe
dementia
N 30 30 30 10 10 10
Mean 93.9 92 82.3 50.7 32.6 14.8
S. D. 2.49 3.04 5.13 5.83 2.98 4.02
As it can be seen from the above table the mean and S. D of young adults; adults and
geriatrics are 93.9, 2.49; 92, 3.04 and 82.3, 5.13 respectively. The mean scores of normal
Dementia Assessment Battery - Kannada
269
groups ranged from 93.9 to 82.3. This shows that there is not much difference between the
three normal groups ranging in age from 20 – 60 years on memory domain. The mean and
S.D of mild dementia, moderate dementia and severe dementia groups are 50.7, 5.83; 32.6,
2.98 and 14.8, 4.02 respectively. The scores of dementia population ranged from 50.7 to 14.8.
This shows that there is a significant difference in the mean and S.D of normal groups and
dementia groups and the performance of dementia group deteriorated as the severity
increased.
All the dementia groups performed similarly on all the tasks of memory domain, i.e.,
on episodic, working and semantic memory tasks. As mentioned above, mild dementics
performed better compared to moderate and severe dementic groups. These results are in
concurrence with the results of studies by Holm et al (2005), Lamar, et al (2007), Baddeley,
et al (1991), which reported several memory deficits in individuals with dementia.
Domain – 2: Linguistic Expression
The second domain in DAB – K is linguistic expression. This domain consists of six
subtests namely picture naming, generative naming, Sentence completion, responsive speech,
spontaneous speech and repetition subtests. The maximum total score of this domain is 250.
The mean and standard deviation scores of each group of subjects are given in Table – 3.
Table 3: Mean and S.D of each group on linguistic expression domain.
Group
Young
adults
(20 – 40
years)
Adults
(40 – 60
years)
Normal
Geriatrics
(above 60
years)
Mild
dementia
Moderate
dementia
Severe
dementia
N 30 30 30 10 10 10
Mean 242.26 233.1 217.6 149.3 110.4 55.8
S. D. 4.21 6.05 10.25 9.91 6.44 7.68
As it can be seen from the above table the mean and S. D of young adults; adults and
geriatrics are 242.26, 4.21; 233.1, 6.05 and 217.6, 10.25 respectively. The mean scores of
normal groups ranged from 242.26 to 217.6. This shows that there is not much difference
between the three normal groups on linguistic expression domain. The mean and S.D of mild
dementia, moderate dementia and severe dementia groups are 149.3, 9.91; 110.4, 6.44 and
55.8, 7.68 respectively. The scores of dementia population ranged from 149.3 to 55.8. This
shows that there is a significant difference in the mean and S.D of normal groups and
dementia groups and the performance of dementia group was deteriorating as the severity
increased.
All the dementia groups performed well on the tasks like spontaneous speech,
repetition etc. and performed poorly on tasks like picture naming and generative naming
tasks. As mentioned above, mild dementics performed well compared to moderate and severe
dementic groups. These results are in correlation with the results of studies by Small et al
(2008), Thomas, et al (2005), Rosselli, et al (2000) which reported linguistic expression
deficits in individuals with dementia.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
270
Domain - 3: Linguistic Comprehension
This domain consists of four subtests within it. They are comparative questions,
following commands, reading comprehension of sentences and reading commands. The
maximum total score of this domain is 150. The mean and standard deviation scores of each
group of subjects are given in Table – 4.
Table 4: Mean and S.D of each group on linguistic comprehension domain.
Group
Young
adults
(20 – 40
years)
Adults (40
– 60 years)
Normal
Geriatrics
(above 60
years)
Mild
dementia
Moderate
dementia
Severe
dementia
N 30 30 30 10 10 10
Mean 150 149.6 146.43 101.5 68.0 44.5
S. D. 0 1.30 5.32 13.25 2.9 5.9
As it can be seen from the above table the mean and S. D of young adults; adults and
geriatrics are 150, 0.0; 149.6, 1.30 and 146.43, 5.32 respectively. The mean scores of normal
groups ranged from 150 to 146.43. This shows that there is not much difference between the
three normal groups on linguistic comprehension domain. The mean and S.D of mild
dementia, moderate dementia and severe dementia groups are 101.5, 13.25; 68.0, 2.9 and
44.5, 5.9 respectively. The scores of dementia population ranged from 101.5 to 44.5. This
shows that there is a significant difference in the mean and S.D of normal groups and
dementia groups and the performance of dementia group deteriorated as the severity
increased.
Visuospatial Construction Domain
This domain consists of two subtests, namely, generative drawing and figure copying.
The maximum possible score is 50. The mean and standard deviation scores of each group of
subjects are given in Table – 5.
Table 5: Mean and S.D of each group on visuospatial construction domain.
Group
Young
adults
(20 – 40
years)
Adults
(40 – 60
years)
Normal
Geriatrics
(above 60
years)
Mild
dementia
Moderate
dementia
Severe
dementia
N 30 30 30 10 10 10
Mean 49.86 49 47.06 30.4 19.5 11.6
S. D. 0.73 2.19 4.47 5.03 2.95 1.83
As it can be seen from the above table the mean and S. D of young adults; adults and
geriatrics are 49.86, 0.73; 49, 2.19 and 47.06, 4.47 respectively. The mean scores of normal
groups ranged from 49.86 to 47.06. This shows that there is not much difference between the
three normal groups on visuospatial construction domain. The mean and S.D of mild
dementia, moderate dementia and severe dementia groups are 30.4, 5.03; 19.5, 2.95 and 11.6,
Dementia Assessment Battery - Kannada
271
1.83 respectively. The scores of dementia population ranged from 30.4 to 11.6. This shows
that there is a significant difference in the mean and S.D of normal groups and dementia
groups and the performance of dementia group were found deteriorating as the severity
increased.
All the dementia groups performed well on the figure copying task than generative
drawing task. As mentioned above, mild dementics performed well compared to moderate
and severe dementic groups.
Total DAB – K Scores
The Dementia Assessment Battery – Kannada consists of four domains, memory,
linguistic expression, linguistic comprehension, and visuospatial construction with several
subtests in each domain. The mean and standard deviation (S.D) of each subtest and also for
each domain are calculated and they are given in Table 6 and as well as in Figure 1.
Table 6: Mean and S.D of the subtasks and domains of DAB – K.
Group Young
adults (20
– 40 years)
Adults (40
– 60 years)
Normal
Geriatrics
(above 60
years)
Mild
dementia
Moderate
dementia
Severe
dementia
Task M S.D M S.D M S.D M S.D M S.D M S.D
Episodic
Memory
30 .00 29.8 .50 28.5 2.2 15.7 2.0 9.8 1.9 5 1.9
Working
Memory
24.6 1.9 23.6 2.0 19.4 2.9 16.2 1.6 9.4 1.6 4.4 1.5
Semantic
Memory
29.4 .89 29.0 1.4 27.2 2.5 14.0 2.8 9.4 1.5 4.8 1.8
Delayed Story
Recall
9.8 .61 9.46 1.0 7.0 1.6 4.8 2.3 4.0 1.3 1.6 .96
Memory
Domain
93.9 2.4 92 3.0 82.3 5.1 50.7 5.8 32.6 2.9 14.8 4.0
Picture Naming 89.8 .76 89.4 1.4 81.5 5.1 53.7 6.7 36.0 5.1 13.8 3.8
Generative
naming
18.1 1.7 16.5 2.2 15.1 2.2 7.5 1.9 5.8 1.1 2.5 1.4
Sentence
completion
9.7 0.5 9.0 1.3 8.4 1.7 5.3 1.0 3.9 0.7 2.3 0.6
Responsive
Speech
9.8 0.5 9.7 0.6 9.4 1.0 6 0.6 3.1 1.2 2.1 0.9
Spontaneous
Speech
20 0.0 19.9 .40 19.5 1.0 13.9 1.2 10.1 0.7 5.0 0.8
Repetition 94.7 4.0 88.5 5.6 83.7 6.8 62.9 5.8 51.5 4.0 30.1 3.7
Linguistic
Expression
Domain
242 4
233 6.0
217
10.2 149 9.91 110 6.44 55
7.68
Comparative
questions
20 0.0 19.9 .36 19.2 1.3 11.8 1.4 7.6 1.2 4.0 1.6
Following
commands
80
0.0
79.3 2.1
76.7 5.1
48.8 9.4
28.6 2.7
21.9 2.72
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
272
Reading
Comprehension
of Sentences
40
0.0
40
0.0
40
0.0
34.5 3.7
27
2.6
16
3.94
Reading
Commands
10
0.0
10
0.0
9.6
0.75 6.4
1.5
4.8
1.03 2.6
0.96
Linguistic
Comprehension
Domain
150 0 149
1.3
146 5.32 101
13.2 68.0 2.9
44.5 5.9
Generative
Drawing
30
0.0
29.5 1.4
28.6 2.9
16.6 2.9
12.1 2.0
7.1
1.2
Figure Copying 19.8 0.7 19.4 1.3 18.2 2.7 13.8 2.5 7.4 1.1 4.5 0.8
Visuospatial
construction
domain
49.8 0.7
49
2.1
47.0 4.4
30.4 5.0
19.5 2.9
11.6 1.8
Total Scores of
DAB –K
536 5.0 523 7.3 493 15 328 54.5 205 30.7 122 19.3
As it can be seen from the above table the mean and S. D of young adults; adults and
geriatrics are 536, 5.0; 523, 7.3 and 493, 15 respectively. The mean scores of normal groups
ranged from 536 to 493. This shows that there is not much difference between the three
normal groups on DAB – K. The mean and S.D of mild dementia, moderate dementia and
severe dementia groups are 328, 54.5; 205, 30.7 and 122, 19.3 respectively. The scores of
dementia population ranged from 328 to 122. This shows that there is a significant difference
in the mean and S.D of normal groups and dementia groups and the performance of dementia
group was inferior throughout and was found deteriorating as the severity increased.
Figure 1: Bar graph showing mean of all the groups on DAB – K.
The results of this study revealed that there is a significant deterioration in the
performance of dementia groups in all the tasks assessed as the severity increased. No
significant difference was found between the three normal age groups on many tasks like
episodic memory, semantic memory, responsive speech, and spontaneous speech, linguistic
comprehension, generative drawing and figure copying parameters. Thereby, indicating that
these parameters are not affected in geriatrics and are suitable to differentiate normal aging
and pathological aging. Significant difference was found on tasks of picture naming,
generative naming, working memory, repetition, etc. and indicating that these cannot be used
to differentiate normal aging and pathological aging. Results also indicate that the test can be
effectively used to differentiate between normal aging and dementics and also to assess the
severity of dementia.
Dementia Assessment Battery - Kannada
273
Summary and Conclusions
Results of the present study revealed significant deficits in individuals with dementia
in memory, linguistic expression and linguistic comprehension. Performance of mild
dementics was near normal in most of the subtests like spontaneous speech, reading
comprehension, following commands, etc. Moderate dementics had more difficulty in tasks
like working memory, episodic memory, picture naming, generative naming, and generative
drawing, etc. Severe dementics’ performance was poor in all the subtests especially due to
poor spontaneous speech skills, and poor intelligibility of speech and these subjects had more
difficulty in all the memory tasks and expression tasks like naming, etc. These results showed
that there is a significant difference in the mean and S.D of normal groups and dementia
groups and the performance of dementia group deteriorated as the severity increased.
Results also revealed significant deterioration in the performance of normal
individuals as the age increased. Performance of geriatric group was poor compared to young
adults and adults. No significant difference was found between the three normal age groups
on many tasks like episodic memory, semantic memory, responsive speech, and spontaneous
speech, linguistic comprehension, generative drawing and figure copying parameters.
Thereby, indicating that these parameters are not affected in geriatrics and are suitable to
differentiate normal aging and pathological aging. Significant difference was found on tasks
of picture naming, generative naming, working memory, repetition, etc. and indicating that
these cannot be used to differentiate normal aging and pathological aging.
Implications of the study
This test battery can be used by Speech language Pathologists to assess and diagnose
the individuals with dementia of mild, moderate and sever stages. This test battery can be
used to differentiate between normal aging and pathological aging. This test battery can also
be used to plan therapy programs for individuals with dementia and related disorders.
References
Albert, M., Blacker, D., Moss, M. B., Tanzi, R. & McArdle, J. J. (2007). Longitudinal change
in cognitive performance among individuals with mild cognitive impairment.
Neuropsychology, 21 (2), 158-169.
Baddeley, A. D., Bressi, S., Della Sala, S., Logie, R. Spinnler, H. (1991). The decline of
working memory in Alzheimer’s disease: A longitudinal study. Brain, 114 (6), 2521-
2542.
Bayles, K. A., & Kaszniak, A. W. (1987). Communication and cognition in normal aging and
dementia. Boston: College Hill Press.
Bayles, K. A., & Tomoeda, C. K. (1993). Arizona Battery for Communication Disorders of
Dementia. Tucson, AZ: Canyonlands Publishing.
Bayles, K. A., & Tomoeda, C. K. (1993). Arizona Battery for Communication Disorders of
Dementia. Tucson, AZ: Canyonlands Publishing.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
274
Bayles, K. A., Tomoeda, C. K., & Trosset, M. W. (1990). Naming and categorical knowledge
in Alzheimer’s Disease: The process of semantic memory deterioration. Brain &
Language, 39, 498-510.
Bryan, K., Binder, J., Dann, C., Funnell, E., Ramsey, V., & Stevens, S. (2001). Development
of a screening instrument for language in older people (Barnes Language
Assessment). Ageing & Mental Health, 5, 371-378.
College of Speech and Language Therapists (1993). Speech and Language Therapy and the
Elderly with dementia: Position Paper. London: College of Speech and Language
Therapists.
Cumnings, J. L., & Benson, D. F. (1992). Dementia: A Clinical Approach. Boston:
Butterworth-Heinemann.
Cunje, A., Molloy, D. W., Standish, T. I., & Lewis, D. L. (2007). Alternate forms of logical
memory and verbal fluency tasks for repeated testing in early cognitive changes.
International Psychogeriatrics, 19, 65-75.
Diagnostic and statistical manual of mental disorders, 4th
ed. (1994). Washington, D. C.
American Psychological Association.
Diagnostic and Statistical manual of mental disorders, 4th
edn, Text Revision (DSM-IV-TR).
(2004). Washington, D. C. American Psychological Association.
Folstein, M. F., Folstein, S. E, & McHugh, P. R. (1975). Mini Mental State: A practical
method for grading the cognitive state of patients for the clinician. Journal of
Psychiatry Research, 12: 189-198.
Freedman, M. & Oscar-Berman, M. (1986). Selective delayed response deficits in
Parkinson’s and Alzheimer’s disease. Archives of Neurology, 43, 886-890.
Heindel, W. C., Salmon, D. P., Shults, C. W., Walicke, P. A., & Butters, N. (1989).
Neuropsychological evidence for multiple implicit memory systems: a comparison of
Alzheimer's, Huntington's, and Parkinson's disease patients. Journal of Neuroscience,
Vol. 9, 582-587.
Holm, R. N., Lepp, R. N., & Ringsberg, K. C. (2005). Dementia: involving patients in story
telling – a caring intervention. A pilot study. Journal of Clinical Nursing, 14 (2), 256-
263.
Huppert, F. A., & Piercy, M. (1979). Normal and abnormal forgetting in organic amnesia:
effect of locus of lesion. Cortex.15 (3): 385-390.
Lamar, M., Price, C. C., Libon, D. J., Penney, D. L., Kaplan, E., Grossman, M., & Heilman,
K. M. (2007). Alterations in working memory as a function of leukoaraiosis in
dementia. Neuropsychologia, 45 (2), 245-254.
Nicholas, M., Obler, L., Au, R., & Albert, M. (1996). On the nature of naming errors in aging
and dementia: A study of semantic relatedness. Brain and Language, 54, 184-195.
Dementia Assessment Battery - Kannada
275
Ostberg, P., Fernaeus, S. E., Hellström, K., Bogdanovic, N., & Wahlund, L. O. (2005).
Impaired verb fluency: a sign of mild cognitive impairment. Brain Language, 95(2):
273-9.
Reisberg, B. (1983). Alzheimer’s disease: The standard reference. New York: The Free
Press, 178-179.
Reisberg, G., Ferris, S. H., & Anand, R. (1984). Functional staging of dementia of the
Alzheimer’s type. Annals of New York academy Sciences, 435: 481-486.
Reisberg, G., Ferris, S. H., & Anand, R. (1984). Functional staging of dementia of the
Alzheimer’s type. Annals of New York academy Sciences, 435: 481-486.
Reisberg, G., Ferris, S. H., & Anand, R. (1984). Functional staging of dementia of the
Alzheimer’s type. Annals of New York academy Sciences, 435: 481-486.
Roselli, M., Ardila, A., Araujo. K., Weekes, V. A., Caracciolo, V., Padilla, M., & Ostrosky-
Sol, F. (2000). Verbal fluency and repetition skills in healthy older Spanish – English
bilinguals. Applied Neuropsychology, 7 (1), 17-24.
Small, J. A., & Sandhu, N. (2008). Episodic and semantic memory influences on picture
naming in Alzheimer’s disease. Brain and Language, 104 (1), 1-9.
Thomas, C., Keselj, V., Cercone, N., Rockwood, K., & Asp, E. (2005). Automatic detection
and rating of dementia of Alzheimer type through lexical analysis of spontaneous
speech. Mechatronics and Automation, 2005 IEEE International Conference, Vol. 3,
1569-1574.
Warrington, E. K., & Shallice, T. (1979). Semantic access dyslexia. Brain. 102(1): 43-63.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
276
Vocal Registers in Classical Carnatic Singers: An Inquiry using
Electroglottography
Sweety Joy & K. Yeshoda*
Abstract
The present study investigated glottal behaviour using some of the EGG parameters in
different vocal registers in singers. Purpose of the study was to analyze EGG waveforms and
parameters and compare the within and across registers for both Carnatic vocal singers and non
singers. The subjects participated in the study consisted of 20 female professional Carnatic singers
with 3 years of formal training in Carnatic classical musical form. Control group consisted of 20
female non singers in the age range of 18-50 years with a mean age of 34 years. Subjects were
instructed to sing vowel /a/ in ascending scale in a single breath reaching the three sthayis (registers).
And the non singers were instructed to sing vowel /a/ in the lowest pitch, habitual pitch and the highest
pitch simultaneously in continuous manner and sustain the pitch for at least three seconds. The
following EGG parameters were extracted CQ, OQ, SQ, and F0. Summary of results a) F0: The mean
F0 and CQ were higher for Carnatic singers when compared to the non singers in falsetto register. B)
CQ: CQ mean values were less in head or modal register compared to falsetto register in singers. C)
OQ: mean values were higher in head register for singers when compared to the other two registers.
And for non singers mean OQ were higher in falsetto register D)SQ: mean SQ values were
significantly different for falsetto register in singers and mean SQ values were higher in non singers in
head and falsetto register. These results revealed that except for F0, the OQ, SQ, CQ were not
significantly different though they varied across registers.
Key words: Carnatic classical music, Chest register, Head register, Falsetto register, EGG
Introduction
Singing is a sensory motor phenomenon that requires particular balanced physical
skill. A sensitive performer achieves singing as an art when these skills are developed.
„Singing‟ is such a human and moving act, both for those who sing and for those who listen.
Hence singers are considered artists in the truest sense, as they combine concepts, melody,
text and stage movement, making it all seem effortless and yet capable of winning audience
appreciation (Bunch, 1982).
It has been reported in literature that there are different registers for singing and
speech. In speech, registers are recognized in ascending scale of pitch and they are: pulse
(vocal fry), chest/modal, falsetto and whistle. In singing: they are denoted as the chest, head
and falsetto for males; and chest, middle and head for females, (Thurman, Welch, Theimer, &
Klitzke, 2004). Typical speaking registers are pulse, modal and falsetto; typical singing
registers are chest, head and falsetto (Hollien, 1974).
_____________________________ * Lecturer in Speech Sciences, All India Institute of Speech and Hearing, Mysore, India
email:[email protected]
Vocal Registers in Classical Carnatic Singers
277
The concept of register is understood to be a series of consecutive similar vocal tones
which the musically trained can differentiate at specific places from another adjoining series
of likewise internally similar tones. Its homogenous sounds depend on a definite, invariable
behaviour of the harmonics. These rows of tones correspond to definite objectively and
subjectively perceptible vibratory regions in the head, neck and chest. The position of the
larynx changes more in natural singer during the transition from one series of tones to
another than in a well trained singer. The registers are caused by a definite mechanism
(belonging to that register) of tone production (vocal fold vibration, glottal shape, air
consumption), which allows for a gradual transition however from one to an adjoining
register. A number of these tones can actually be produced in two overlapping registers but
not always with the same intensity, (Nair, 1999).
The registers have been named by singers according to the subjective sensations they
produce. The low register is referred to as “chest”, because singing in that register produces a
feeling of vibration in the upper chest and lower neck. The upper register is called “head”
(female) or “falsetto”(male) because vibration are sensed high in head. The middle register is
often called “mixed”/modal because it has the high and low qualities in it.
Different registers sound different, but variant acoustic impression derived from
changes in the way the vocal source signal is moulded by the vocal tract as well as from
differences in the vocal source signal itself. A laryngeal register reflects a specific and distinct
mode of the laryngeal action. Vocal tract contributions are irrelevant. A laryngeal register is
produced across a contiguous range of fundamental frequencies. The F0 range of any given
laryngeal register has little overlap with the F0 range of any other register.
Nair (1999) opined that production of voice registers could be described from
acoustical or physiological point of view. In particular there are three main factors influencing
the voice register production: (a) Configuration of larynx – phonatory settings (mechanism of
oscillation of the vocal folds), (b) Vocal tract resonance (formants setting influence resultant
spectrum) and (c) Interaction of the sub glottal and supra glottal resonances with the vocal
fold oscillations.
Changes in vocal register may lead to differences in control of frequency. The
description of laryngeal control of frequency relate most closely to modal register, in which a
rise in vocal pitch is principally achieved by contractions of both the cricothroid and
vocalis/thyroarytenoid. The most noticeable difference in muscle control between the modal
and falsetto register is in the activity of the vocalis/ thyroarytenoid, Aikin (1902). Closed
quotient varies in falsetto as well as in chest register (Nair, 1999).
Aronson (1973) explained the physiological mechanism of falsetto in term of both
intrinsic and extrinsic muscles of larynx as follows: a) through the action of the thyrohyoid
and the supra hyoid musculature the larynx would be elevated high in to the neck, b) through
action of the stlyohyoid muscles the larynx would be tilted downwards, maintaining the vocal
folds in a state of laxity, c) although the vocal folds would be in a flabby state, the contraction
of the crico thyroid muscles will causes the vocal folds to be stretched, d) as a result the
vocal fold mass reduces owing to loss resistance to sub glottal air pressure, e) and hence only
the medical edges of the folds vibrate will vibrate because of reduced sub glottal air pressure.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
278
Measurement of Vocal registers
Vilkman, Alku, Laukkanen (1995) studied the shifts from chest to falsetto registers in
trained and untrained male and female subjects. The shift from breathy to “falsetto” phonation
to normal chest voice phonation was studied in normal female (pitch range 170–180 Hz) and
male (pitch range 94–110 Hz) subjects. The phonations gliding from falsetto to chest register
were analyzed using adaptive inverse filtering and electroglottography. The results revealed
that differences noticed in the trained and untrained subjects during the register shift were not
significant and register shift were interpreted in terms of critical mass concept.
Chen, Robb & Gilbert (2002) studied the vocal fry register in men and women using
EGG. EGG parameters noted were vocal F0, duration of opening and closing phase and SQ.
Results revealed significantly higher F0, further; it also revealed that female speakers
demonstrated greater increase in SQ in fry register, indicating longer opening-phase duration
per glottal cycle. The results confirmed that the general notion the gender difference exist
even in vocal fry register.
Henrich, (2003) measured EGG open quotient in 18 classically trained male and female
singers in different registers and results revealed that open quotient values were usually lowest
in modal and chest when compared to falsetto and head. It was concluded that OQ depended
on the functioning of laryngeal mechanism and the laryngeal mechanize were similar for
males and females.
Henrich, N., d‟Alessandro, C., Castellengo, M., & Doval, B. (2004) explored the
relationship between
open quotient and laryngeal mechanisms: vocal intensity and
fundamental frequency using EGG in 18 classically trained male and female singers in
different vocal registers. It was found that open quotient (OQ) ranged from 0.3 to 0.8 in chest
voice and from 0.5 to 0.95 in head and falsetto voices. The OQ was
strongly related to vocal
intensity in chest voice and to fundamental frequency in head and falsetto voices. Henrich,
d'Alessandro, Doval and Castellengo (2005) conducted a study and measured open quotient
using EGG during speech and singing in three singers. EGG signals were analyzed, under
various speech and singing conditions: sustained vowels, spoken and sung sentences,
crescendos and decrescendos, glissandos. They reported that open quotient was different in
male and female voices and depended on underlying laryngeal mechanism. The dynamic
variations produced were much pronounced during speaking when compared to singing. And
they also observed correlation between OQ and vocal intensity in chest register.
Howard (1995) studied the upper registers in trained singers using EGG. The results
revealed increased open phase duration and increased open quotient. It was reported that
EGG wave morphology revealed sharp peaks during upper register production reflecting; the
fact that only the superior area of the VF mucosa was in contact. Archana (1987) in her study
on five trained female singers investigated the efficacy of EGG parameters and LTAS to
differentiate the musical notes within and across the registers in carnatic vocal music:
Mandhra (low) sthayi, Madhya sthayi (mid), tara sthayi (high). And the results revealed that
all EGG parameters: the open time, closed time, opening time, closing time, open phase,
closed phase, open quotient, speed quotient, speed index and total period were significantly
different across notes and registers.
Vocal Registers in Classical Carnatic Singers
279
The present study was planned to understand the vocal fold closure or laryngeal
behaviour in singers during different register production. The aims of the study were as
follows:
1. To compare the EGG parameters across vocal registers in classical Carnatic singers.
2. To compare the EGG parameters between Carnatic singers and non singers across
registers.
Method
Subjects
40 females in the age range of 18-50 years with a mean age of 34 years participated in
the study. Subjects had no complaints of vocal, speech or hearing problems at the time of
recording. The subjects were divided into control group and experimental group.
Experimental group consisted of 20 professional Carnatic singers with a minimum of 3 years
of formal training in Carnatic classical musical form. Control group consisted of 20 female
non-singers who had no formal training in any form of singing.
Instrumentation
Kay Pentax Electroglottograph Model 6103 was used for acquisition of the EGG data.
The acquired data was represented as (a) selected parameter, (b) F0 in Hz and (c) a glottal
wave. The glottal wave was represented as time (in seconds) on X axis and amplitude of the
signal (in volts) on Y axis. One glottal cycle at any given point could be measured by moving
the cursor horizontally.
Instruction
All the subjects were explained about the purpose of the study. Familiarization was
done for all subjects.
Singers: Singers were asked to sing song or a particular „Raaga‟ wherein, transitions across
the three registers occurred for the familiarization. Then they were asked to sing vowel /a/ in
ascending scale in a single breath reaching the three sthayis (registers).
Non singers: were familiarized with suitable examples wherein the investigator provided pre
recorded audio sung samples sung the required task. They were asked to model the task and
practice. Then they were instructed to sing vowel /a/ in the lowest pitch, habitual pitch and
the highest pitch possible simultaneously in continuous manner and sustain the pitch for at
least three seconds. They were asked to maintain a comfortable loudness level while singing.
Recording
The recordings were done individually. Subjects were seated comfortably in front of
the instrument neckbands with the electrodes were tied to the neck and securely positioned on
either sides of the thyroid prominence. They were asked to sit quiet without moving head and
the positions of the electrodes were adjusted until clear EGG sine waveforms appeared on the
screen when the subjects phonated. The gain of the instrument was also manually adjusted
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
280
and the task was captured directly on to the instrument. Subjects were asked to sustain the
pitch for at least 3 seconds. Each subject was asked to sing the vowel /a/ thrice and all the
data were saved on to the computer memory. The recording that yielded best EGG waveform
morphology was retained for extraction of the parameters. The cursor was positioned at the
steady state of F0 and the following EGG parameters were extracted CQ, OQ, SQ, and F0.
The measured EGG parameters were tabulated for further statistical analysis. The data
was then subjected to suitable statistical analysis using SPSS software (version 15). Mean
and standard deviation were extracted. Repeated measures ANOVA was done to analyze the
variation of glottal parameters across registers within singers and independent t test was done
for pair wise comparison of Carnatic singers and non singers for each parameter.
Results and Discussion
The study aimed to investigate the vocal fold closure during singing at different
registers in Carnatic singers and non singers. The EGG parameters used in the current study
were fundamental frequency (F0), contact quotient (CQ), open quotient (OQ), and speed
quotient (SQ). Results are discussed under the following headings. A) Mean and standard
deviations for the Carnatic singers and non singers across registers B) Comparison of Carnatic
singers across registers. C) Comparison of non singers across registers. D) Comparison of
Carnatic singers and non singers across registers
A) Mean and standard deviations for the Carnatic singers and non singers across
registers
Mean and standard deviations for all the EGG parameters in the Carnatic singers in
three different register (Chest, head and falsetto) are tabulated in Table 1. EGG values across
registers showed no much difference except in fundamental frequency. Mean CQ and SQ
values were higher in falsetto register and mean OQ scores were higher for head register.
Table 1: Mean and SD of EGG parameters across registers for Carnatic singers
Parameter Chest Head Falsetto
Mean SD Mean SD Mean SD
F0 169 31.40 223 31.72 351 39.58
CQ 46 4.29 45 5.16 47 3.54
OQ 53 4.29 54 5.15 52 3.54
SQ 116 91.20 102 81.19 127 95.40
Similarly, Mean and standard deviations were tabulated for non singers in three
different registers in Table 2. The mean values EGG parameters for the non singers across
registers do not show much difference except in fundamental frequency. Mean CQ and SQ
values were higher in head register and mean OQ values were lower for head register.
Vocal Registers in Classical Carnatic Singers
281
Table 2: Mean and SD of EGG parameters across registers for non singers
Parameter Chest Head Falsetto
Mean SD Mean SD Mean SD
F0 193 12.43 210 21.74 341 47.14
CQ 41 7.55 43 3.45 41.21 6.71
OQ 58 7.56 56 3.44 58.78 6.71
SQ 233 115.96 258 84.50 210.91 122.51
Comparison of tables 1 and 2 gives an impression that non singers were not able to
match the F0 of Carnatic singers across the registers. Especially in chest register, non singer
obtained higher mean F0 values. Also the standard deviations of non singers were more
compared to Carnatic singers in chest and falsetto registers. This indicates that non singers
could not control their F0 production in chest and falsetto registers.
B) Comparison of Carnatic singers across registers
For the comparison of EGG parameters across three register, repeated measures
ANOVA were used as shown in Table 3. It shows there is no significant difference across
registers in any of the EGG parameters, except fundamental frequency which increased with
the registers and was highest for falsetto and lowest for chest registers. Pair wise comparison
for F0 showed significant difference across the three conditions.
Table 3: f value across registers in Carnatic singers.
Parameter Number of Subjects f Significance
F0 20 227.36 0.000**
CQ 20 0.721 0.493
OQ 20 0.737 0.485
SQ 20 0.490 0.616
**p<0.001
The findings of the current study results are in contrast with Archana (1987) and
Henrich, et. Al., (2005). Archana (1987) reported significane difference in EGG parameters
across registers. Henrich et al (2005) found that OQ was higher in falsetto register. The glottal
closure decreases as the rate of vocal fold vibration increases. This could explain the
significant increase in F0 as register changes. The Graph 1 shows comparison for F0, OQ, SQ
and SQ in singers across three registers. Although from the graph it is clear that only F0 had
greater difference between the registers, even SQ was different between the registers but was
not statistically significant.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
282
Graph 1: Comparison of Carnatic singers across 3 registers
C) Comparison of EGG parameters across three registers in non singers
For the comparison of EGG parameters across three register in non singer (chest, head
and falsetto) repeated measures ANOVA was used. Table 4 shows the results of repeated
measure ANOVA across registers. It shows there is no significant difference across registers
in any of the EGG parameters, except fundamental frequency which shows statistically high
significant difference across register. Pair wise comparison for F0 showed significant
difference across the three conditions.
Table 4: f value across registers in non singers.
Parameter Number of subjects f Significance
F0 20 184.07 0.000**
CQ 20 1.114 0.389
OQ 20 1.087 0.348
SQ 20 1.038 0.364
**p<0.001
Graph 2 shows the comparison F0, CQ, OQ and SQ for all the three register. It can be
noticed that only F0 had greater difference between the registers, SQ was different across the
registers but not statistically significant.
Graph 2: Comparison of non singers across 3 registers
D) Comparison of Carnatic singers & non singers across registers.
The results are pooled for the two groups in Table 5 and it reveals that singers had a wider
pitch range compared to the non singers.
Vocal Registers in Classical Carnatic Singers
283
Table 5: Mean, SD, t values for Carnatic singers and non singers.
Parameter Groups Mean SD t
F0
Falsetto Singers 351 39.58 0.488
Non-singers 341 47.14
Chest Singers 169 31.40 0.003**
Non-singers 193 12.43
Head
Singers 223 31.72 0.148
Non-singers 210 21.74
CQ
Falsetto
Singers 47 3.54 0.001**
Non-singers 41 6.71
Chest
Singers 46 4.29 0.012*
Non-singers 41 7.55
Head
Singers 45 5.15 0.117
Non-singers 43 3.45
OQ
Falsetto
Singers 52 3.54 0.001**
Non-singers 58 6.71
Chest Singers 53 4.29 0.012*
Non-singers 58 7.56
Head
Singers 54 5.15 0.117
Non-singers 56 3.44
SQ
Falsetto
Singers 127 146.40 0.058*
Non-singers 210 122.51
Chest
Singers 116 91.20 0.001**
Non-singers 233 115.96
Head
Singers 102 81.19 0.000**
Non-singers 258 84.50
* p< 0.05, **p<0.01
The mean F0 and CQ values were higher for Carnatic singers when compared to the
non singers in falsetto register. These verdicts are in consonance with the findings of Howard
(1995) and Nair (1999). Trained singers are able to produce upper register with a CQ that is
slightly above 50% (Howard, et. al., 1995), presumably because they have the ability to
strongly adduct their vocal folds. Mc Coy & Scott, (2007) opined that CQ values are higher in
heavy mechanism like chest voice; it‟s more than 50%. And CQ is less in light mechanism
like head and falsetto. In this present study CQ values are less in head or modal register
compared to falsetto register. Increased CQ values shows increased glottal adduction and
stronger contraction of inter arytenoids and lateral cricoarytenoid (LCA) increased medial
compression from activity of the thyroaryteniod muscles. Graph 3 shows comparison of EGG
parameters between the Carnatic singers and non singers. Results also shows that mean OQ
values are higher in head register for singers when compared to the other two registers. And
for non singers OQ scores are higher in falsetto register which is in consonance with Henrich
(2005). But according to Henrich (2005) the open quotient depends on the laryngeal
mechanisms in singers. The open quotient is strongly related to vocal intensity in chest,
modal,
and male head register. And OQ related to fundamental frequency in falsetto for male and
head register for female singers. SQ values were significantly different for falsetto register in
singers and mean SQ values were higher in non singers in falsetto register. SQ reflects the
asymmetry of the glottal pulse; since the F0 is higher in falsetto, it suggests more
asymmetrical vibrations in falsetto register both in Carnatic singers and non singers and
especially in non singers as they have not mastered production of those registers.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
284
Registers
PulseModalLoft
Ave
rag
e P
itch
(H
z.)
400
300
200
100
Groups
Singers
Non-singers
a) F0: Carnatic singers vs
non-singers.
Registers
PulseModalLoft
Me
an
CQ
(%
)
100
80
60
40
20
0
Groups
Singers
Non-singers
c) CQ: Carnatic singers vs
non-singers.
Registers
PulseModalLoft
Me
an
OQ
(%
)
100
80
60
40
20
0
Groups
Singers
Non-singers
b) OQ: Carnatic singers vs
non-singers.
Registers
PulseModalLoft
Me
an
SQ
(H
z)
300
200
100
0
Groups
Singers
Non-singers
d) SQ: Carnatic singers vs.
non-singers
Fal Head Chest Fal Head Chest
Fal Head Chest Fal Head Chest
Graph 3: Comparison of EGG parameters between the Carnatic singers and non singers.
Vocal Registers in Classical Carnatic Singers
285
Conclusions
In general the results revealed that except for F0, the OQ, SQ, CQ were not
significantly different though they varied across registers. On comparison, singers win though
they had minimum of 3 years of experience, EGG parameters were significant by different
especially in chest and falsetto registers. This would be attributed to the training in voice use.
Limitations
Most of the singers had a minimum of 3 years of formal practice they were
still under the formal training.
Numbers of subjects were limited.
Future direction
Comparison of classical Carnatic singers and Hindustani classical singers to
know more about their laryngeal mechanism. Comparison between male and
female classical singers.
References
Aikin, J. (1902). The Effect on Vocal Pitch of the Contracture of the Thyro - Arytænoid
Muscles. The Journal of Laryngology Rhinology and Otology, 23, 643-648.
Aronson, C. (1973). Isotonic Vocalis Contraction as a Means of Producing Rapid Decreases
in F0. Journal of Speech and Hearing Research, 23, 576-592.
Archana, M. (1987). Acoustic Parameters in Singing Registers. Unpublished master‟s
dissertation submitted in part fulfillment for the master‟s degree in speech and hearing.
University of Mysore.
Bunch, M. (1982). Dynamics of the singing voice. New York Springer; 1982:156.
Chen, Y., Robb, M., Gilbert, H. (2002). Electroglottographic evaluation of gender and vowel
effects during modal and vocal fry phonation Journal of Speech, Language, and
Hearing Research, 45, 821-829.
Henrich, Roubeau, Castellengo. (2003). On the use of electroglottography for characterisation
of the laryngeal mechanisms. Proceedings of the Stockholm Music Acoustics
Conference, August 6-9, 2003 (smac 03), Stockholm, Sweden.
Henrich, Roubeau, Castellengo & Doval, B. (2004). On the use of electroglottography for
characterisation of the laryngeal mechanisms. Journal of Acoustic Society of
America, 115, 1321-1332.
Henrich, N., d‟Alessandro, C., Castellengo, M., & Doval, B. (2005). Glottal open quotient in
singing: Measurements and correlation with laryngeal mechanisms, vocal intensity,
and fundamental frequency. The Journal of the Acoustical Society of America. 117,
1417-1430.
Hollien, H. (1974). On vocal registers. Journal of Phonetics, 2, 125-143.
Dissertation Vol. VII, 2008-09, Part – B, SLP, AIISH, Mysore
286
Howard, M. D. (1995). Variation of electrolaryngographically derived closed quotient for
trained and untrained adult female singers, Journal of Voice, 9, (2), 163-172.
McCoy, Scott (2007). A classical pedagogue explores belting. Journal of Singing. 1.
Nair, G. (1999). Voice tradition & technology. Singular Publishing Group. San Diego.
Thurman, L., Welch G., Theimer, A., Klitzke C. (2004). Addressing vocal register
discrepancies: an alternative, science-based theory of register phenomena. Second
International Conference. The Physiology and Acoustics of Singing National Center
for Voice and Speech, 6–9 October 2004. Denver, Colorado, USA.
Vilkman, E., Alku, P., & Laukkanen, A. M. (1995). Vocal-fold collision mass as a
differentiator between registers in the low-pitch range. Journal of Voice, 9(1), 66-73.