syllables and phonotactics - umass · phonotactics the structure of syllables phonotactics...

Syllables andPhonotactics

SupplementaryReadings

Introduction andIPA Practice

The Basics ofSyllables


Syllabification Rule

SyllabicConsonants

Onsets OverCodas

. . . . . .

Syllables and Phonotactics







SyllabicConsonants

Onsets OverCodas

. . . . . .

Supplementary Readings

The following readings have been posted to the Moodlecourse site:

I Contemporary Linguistics: Chapter 3 (pp. 77-84)I “Syllables” by Kyle Johnson







SyllabicConsonants

Onsets OverCodas

. . . . . .

A Review of Where We Are

The Fundamental Question (for Linguists):What is the system of rules and mental representationsthat underlies out ability to speak and understand ahuman language?

I Last Unit: IPANotation that unambiguously represents the phonesof all human languages.

I This Unit: PhonologyRules that operate over those representations, andthereby affect how words are pronounced.







SyllabicConsonants

Onsets OverCodas

. . . . . .

Some Quick IPA Practice

First, let’s practice reading some IPA:

[maj b2lowni hæz @ f1ôst nejm Its ow Es si ej Aô]

(My bologna has a first name, it’s O-S-C-A-R)







SyllabicConsonants

Onsets OverCodas

. . . . . .

Some Quick IPA Practice

Here’s another one:

[mInti plEZ1ôz Inklud pEptow bIzmAl ]

(Minty pleasures include Pepto Bismol)







SyllabicConsonants

Onsets OverCodas

. . . . . .

Towards ‘Phonology’

IPA transcription represents some of our knowledge ofhow a word is pronounced:

I Phones making up the wordI Sequence in which those phones are produced

However, there is more to a word’s pronunciation thanjust this...

I There are also general rules governing Englishpronunciation.







SyllabicConsonants

Onsets OverCodas

. . . . . .

Introducing Syllables

Key ExampleWhen we pronounce ‘understanding’, we don’t pronounce(or perceive) it as a simple string of sounds...

I Rather, the phones are grouped together into ‘beats’

I These ‘beats’ making up the word give it a rhythm

I These ‘beats’ and rhythm come out naturally whenwe speak words slowly

I [2n . d1ô . stæn . diN] ‘understanding’I [pEp . tow . bIz . mAl] ‘Pepto Bismol’I [b2 . low . ni] ‘Bologna’







SyllabicConsonants

Onsets OverCodas

. . . . . .

Introducing SyllablesVocabulary:These ‘beats’ making up the sound-structure of the wordare called syllables.

Key Fact:For most English words, speakers only accept one way ofbreaking it up into syllables

I [2n . d1ô . stæn . diN] *[2 . nd1ôs . tænd . iN]

I [pEp . tow . bIz . mAl] *[pEpt . ow . bIzm . Al]

I [b2 . low . ni] *[b2l . own . i]

Conclusion:Part of knowing English is knowing how to divide wordsinto syllables

I So, this information is represented in our brains in someway.







SyllabicConsonants

Onsets OverCodas

. . . . . .

Towards Syllabification

Question:How is this information represented in our brains?

Wrong Hypothesis:Maybe we just memorize it on a word-by-word basis?

I When we learn a word like ‘bologna’, we learn:I The phones that compose it: [b2lowni]I The word’s syllabification: [b2 . low . ni]

Vocabulary:The way that a word is broken down into syllables iscalled syllabification.







SyllabicConsonants

Onsets OverCodas

. . . . . .

Towards Syllabification

Wrong Hypothesis:The syllabification of word is memorized on a case-by-casebasis.

Problem for This Hypothesis:English speakers know how to syllabify words they’ve neverheard spoken before.

Example:

I Read the following word to yourself silently: “badartogly”

I Say the word to yourself slowly, breaking it into syllables...

I I would bet that you syllabified it as follows:

I [ bæ . dAô . tA . gli ]

I I would bet that you didn’t syllabify it this way:

I [ bæd . Aôt . Ag . li ]







SyllabicConsonants

Onsets OverCodas

. . . . . .

Towards SyllabificationConclusion:Since you’d never heard “badartogly” before, yoursyllabification couldn’t have been memorized...

I So the ‘Wrong Hypothesis’ is wrong...

Right Hypothesis:English syllabification is based on a general rule.

I This rule tells you how to syllabify any imaginable word ofEnglish.

Vocabulary:Phonology = the study of the general rules that governhow words are pronounced in a language.

The Burning Question:What are the rules for syllabifying a word of English?





Syllables andPhonotacticsThe Structure of Syllables

Phonotactics


SyllabicConsonants

Onsets OverCodas

. . . . . .

Before we can state the rule, we need to talk a bit moreabout the internal structure of syllables...






Phonotactics


SyllabicConsonants

Onsets OverCodas

. . . . . .

The Structure of Syllables

I Every syllable is made up of (at most) three parts:

I Onset: the consonants that begin the syllable

I Nucleus: the sound in the middle of the syllable(usually a vowel)

I Coda: the consonants the end the syllable

I Syllables can differ in size:

I Some syllables do not have onsets (e.g. [it])

I Some syllables do not have codas (e.g. [ti])

I But, every syllable has a nucleus






Phonotactics


SyllabicConsonants

Onsets OverCodas

. . . . . .

The Structure of Syllables

Here are some syllables, broken down into their parts:

Syllable Onset Nucleus Coda[bæn] [b] [æ] [n][bi] [b] [i] NONE[æn] NONE [æ] [n][stænd] [st] [æ] [nd][plæNk] [pl] [æ] [Nk][bOj] [b] [Oj] NONE[spôej] [spô] [ej] NONE[aj] NONE [aj] NONE

Note:Both parts of a diphthong count as being in the ‘nucleus’of the syllable.






Phonotactics


SyllabicConsonants

Onsets OverCodas

. . . . . .

Towards ‘Phonotactics’

Key Fact:Not every sequence of phones is a possible onset orcoda in English.

I Consider the following sequences of phones:I [mba . ka]I [Na . to]I [tsA]I [nukÙ]I [tæg . nIsp]

I Which one sounds most like an English word?

I You probably said [tæg . nIsp]I After all, it’s the only one that English speakers can

easily pronounce.






Phonotactics


SyllabicConsonants

Onsets OverCodas

. . . . . .

Towards ‘Phonotactics’

Key Fact:Not every sequence of phones is a possible onset orcoda in English.

I Consider the following sequences of phones:I [mba . ka]I [Na . to]I [tsA]I [nukÙ]I [tæg . nIsp]

I Which one sounds most like an English word?I You probably said [tæg . nIsp]

I After all, it’s the only one that English speakers caneasily pronounce.






Phonotactics


SyllabicConsonants

Onsets OverCodas

. . . . . .

Towards ‘Phonotactics’But why don’t those other words sound like English?

I [mba . ka]

I The first syllable starts with [mb]I No English word or syllable can start with [mb]I The sequence [mb] is not a possible onset

I [Na . to]

I The first syllable starts with [N]I No English word or syllable can start with [N]I The sequence [N] is not a possible onset

I [tsA]

I The first syllable starts with [ts]I No English word or syllable can start with [ts]I The sequence [ts] is not a possible onset

I [nukÙ]

I The first syllable ends with [kÙ]I No English word or syllable can end with [kÙ]I The sequence [kÙ] is not a possible coda






Phonotactics


SyllabicConsonants

Onsets OverCodas

. . . . . .


I [mba . ka]I The first syllable starts with [mb]I No English word or syllable can start with [mb]I The sequence [mb] is not a possible onset

I [Na . to]

I The first syllable starts with [N]I No English word or syllable can start with [N]I The sequence [N] is not a possible onset

I [tsA]


I [nukÙ]







Phonotactics


SyllabicConsonants

Onsets OverCodas

. . . . . .



I [Na . to]I The first syllable starts with [N]I No English word or syllable can start with [N]I The sequence [N] is not a possible onset

I [tsA]


I [nukÙ]







Phonotactics


SyllabicConsonants

Onsets OverCodas

. . . . . .




I [tsA]I The first syllable starts with [ts]I No English word or syllable can start with [ts]I The sequence [ts] is not a possible onset

I [nukÙ]







Phonotactics


SyllabicConsonants

Onsets OverCodas

. . . . . .




I [tsA]I The first syllable starts with [ts]I No English word or syllable can start with [ts]I The sequence [ts] is not a possible onset

I [nukÙ]I The first syllable ends with [kÙ]I No English word or syllable can end with [kÙ]I The sequence [kÙ] is not a possible coda






Phonotactics


SyllabicConsonants

Onsets OverCodas

. . . . . .

Phonotactic ConstraintsVocabulary:The rules that determine the possible onsets or codas arecalled phonotactics constraints (a.k.a phonotactics).

I ‘phono’ = sounds ; ‘tactic’ = arrangement(how the sounds can be arranged)

Key Fact:Different languages have different phonotacticconstraints.

I In Dholuo, [mb] and [N] are possible onsets:I [mba . ka] = storyI [Na . to] = person

I In Tlingit, [ts] is a possible onset, and [kÙ] a possiblecoda:

I [tsA] = sealI [nukÙ] = to do






Phonotactics


SyllabicConsonants

Onsets OverCodas

. . . . . .

Phonotactics and Syllabification

Fun Fact:Phonotactic constraints play a major role in syllabification.

I What’s wrong with: *[2 . nd1ôs . tænd . iN]I [nd] is not a possible onset of English

(No word of English begins with [nd])

I Why not this: *[pEpt . ow . bIzm . Al]I [zm] is not a possible coda of English

(No word of English ends in [zm]])






Phonotactics


SyllabicConsonants

Onsets OverCodas

. . . . . .

Toward the Syllabification Rule

But, phonotactic constraints don’t explain everything...

I Why don’t we syllabify like this: *[b2l . own . i]I This syllabification is totally consistent with English

phonotactics.I [b2l] is a possible English syllableI [own] is a possible English syllableI [i] is a possible English syllable

I The answer will come from a concrete rule(algorithm) for syllabifying an English word...






Syllabification RuleStep 1: Label Nuclei

Step 2: Label Onsets

Step 3: Label Codas

Step 4: Group Into Syllables

SyllabicConsonants

Onsets OverCodas

. . . . . .

The Syllabification Algorithm

I There are four main steps to the ‘syllabification rule’(‘syllabification procedure’, ‘syllabification algorithm’)

I We will illustrate each step with our made-up words[bædAôtAgli] and [tægnIsp].








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .

Step 1: Label the Nuclei

I Every syllable has a nucleus.I Key Fact: All vowels serve as syllabic nuclei.I So, Step One of our syllabification procedure is:

I Identify all the vowels in the word, and...I Label those vowels as ‘nuclei’ (N)

b æ d A ô t A g l i t æ g n I s p








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .

Step 1: Label the Nuclei

I Every syllable has a nucleus.I Key Fact: All vowels serve as syllabic nuclei.I So, Step One of our syllabification procedure is:

I Identify all the vowels in the word, and...I Label those vowels as ‘nuclei’ (N)

b

N

æ d

N

A ô t

N

A g l

N

i t

N

æ g n

N

I s p








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .

Step 2: Label the Onsets

Step 2:To the extent allowed by the phonotactics, gather up all thephones preceding a nucleus, and label them as an onset.

I Look at the consonants preceding each ‘N’

I Find the largest continuous sequence that Englishphonotactics allows to be an onset.

I Write an ‘O’ above that sequence and connect eachphone to that ‘O’ with a line.

b

N

æ d

N

A ô t

N

A g l

N

i t

N

æ g n

N

I s p








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .






O

b

N

æ d

N

A ô t

N

A g l

N

i t

N

æ g n

N

I s p








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .






O

b

N

æ

O

d

N

A ô t

N

A g l

N

i t

N

æ g n

N

I s p








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .






O

b

N

æ

O

d

N

A ô

O

t

N

A g l

N

i t

N

æ g n

N

I s p








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .






O

b

N

æ

O

d

N

A ô

O

t

N

A g l

N

i t

N

æ g n

N

I s p

I [ôt] is not a possible onset of English (no word starts with [ôt])








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .






O

b

N

æ

O

d

N

A ô

O

t

N

A g

O

l

N

i t

N

æ g n

N

I s p









Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .






O

b

N

æ

O

d

N

A ô

O

t

N

A

O

g l

N

i t

N

æ g n

N

I s p


I [gl] is a possible onset of English (e.g. [glow])








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .






O

b

N

æ

O

d

N

A ô

O

t

N

A

O

g l

N

i

O

t

N

æ g n

N

I s p










Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .






O

b

N

æ

O

d

N

A ô

O

t

N

A

O

g l

N

i

O

t

N

æ g

O

n

N

I s p










Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .

Step 2: Label the OnsetsStep 2:To the extent allowed by the phonotactics, gather up all thephones preceding a nucleus, and label them as an onset.




O

b

N

æ

O

d

N

A ô

O

t

N

A

O

g l

N

i

O

t

N

æ g

O

n

N

I s p



I [gn] is not a possible onset of English (no word starts with [gn])








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .

Finding the Onsets

I A major part of ‘Step 2’ is figuring out “Can thissequence of phones be an onset in English?”

I To answer this, either:I Use your own ‘intuitions’ as a native speaker

(if you are one)I Team up with a friend who is a native speaker

(if you aren’t one)

I Either way, you ask “can a word of English start withthis sequence of phones?”








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .

Step 3: Label the Codas

Step 3:To the extent allowed by the phonotactics, gather up theremaining unlabeled phones following a nucleus, and labelthem as a coda.

I Look at the consonants following each ‘N’ (that aren’talready labeled as ‘O’s)

I Find the largest continuous sequence that Englishphonotactics allows to be a coda.

I Write a ‘C’ above that sequence and connect each phoneto that ‘C’ with a line.

O

b

N

æ

O

d

N

A ô

O

t

N

A

O

g l

N

i

O

t

N

æ g

O

n

N

I s p








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .






O

b

N

æ

O

d

N

A

C

ô

O

t

N

A

O

g l

N

i

O

t

N

æ g

O

n

N

I s p








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .






O

b

N

æ

O

d

N

A

C

ô

O

t

N

A

O

g l

N

i

O

t

N

æ

C

g

O

n

N

I s p








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .






O

b

N

æ

O

d

N

A

C

ô

O

t

N

A

O

g l

N

i

O

t

N

æ

C

g

O

n

N

I

C

s p








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .

Step 3: Label the CodasStep 3:To the extent allowed by the phonotactics, gather up theremaining unlabeled phones following a nucleus, and labelthem as a coda.




O

b

N

æ

O

d

N

A

C

ô

O

t

N

A

O

g l

N

i

O

t

N

æ

C

g

O

n

N

I

C

s p

I [sp] is a possible coda of English (e.g. [k2sp])








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .

Finding the Codas

I As with Step 2, a major part of ‘Step 3’ is thequestion “Can this sequence of phones be a coda inEnglish?”

I To answer this question, either:I Use your own intuitions as a native speaker

(if you are one)I Team up with a friend who is a native speaker

(if you aren’t one)

I Either way, you ask “can a word of English end withthis sequence of phones?”








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .

Step 4: Add the Syllable Lablels

Step 4:Group together the ‘O’s, ‘N’s and ‘C’s into syllables.

I Above each N, write a ‘σ ’ (for syllable), and draw a lineconnecting them.

I If there is an O preceding an N, connect that O to the σthat N is connected to.

I If there is a C following an N, connect that C to the σ thatthe N is connected to

O

b

N

æ

O

d

N

A

C

ô

O

t

N

A

O

g l

N

i

O

t

N

æ

C

g

O

n

N

I

C

s p








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .






O

b

σ

N

æ

O

d

σ

N

A

C

ô

O

t

σ

N

A

O

g l

σ

N

i

O

t

σ

N

æ

C

g

O

n

σ

N

I

C

s p








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .






σ

O

b

N

æ

σ

O

d

N

A

C

ô

σ

O

t

N

A

σ

O

g l

N

i

σ

O

t

N

æ

C

g

σ

O

n

N

I

C

s p








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .

Study Exercise:

I At home, try running this procedure on the followingwords:

I [2nd1ôstændiN] (“understanding”)I [pEptow bIzmAl] (“Pepto Bismol”)I [b2lowni] (“Bologna”)

I Confirm that our rule will predict the correctsyllabification for these words.








Step 3: Label Codas


SyllabicConsonants

Onsets OverCodas

. . . . . .

Summing Up

What We’ve Done So Far:

I We’ve laid out a rule that will correctly syllabifyEnglish words.

I This rule, then, is one (very small) part of an answerto our ‘fundamental question’.

The Fundamental Question (for Linguists):What is the system of rules and mental representationsthat underlies out ability to speak and understand ahuman language?







SyllabicConsonants

Onsets OverCodas

. . . . . .

Problem: Syllabic Consonants

But, there’s a problem for our syllabification rule...

I If we follow this rule, the only possible nuclei we willget will be vowels

I However, although every vowel is a nucleus...

I ...some nuclei are not vowels.







SyllabicConsonants

Onsets OverCodas

. . . . . .

Syllabic Consonants in EnglishKey Fact:The consonants [l], [n], and [ô], can be nuclei in English.

Words With Consonants as Syllabic Nuclei:

I “kitten” [kI . tn]

I “cuddle” [k2 . dl]

I “banner” [bæ . nô]

Question: Why do we say they are the nucleus of a syllable?

Answer: We hear these words as having two syllables, butthere is no ‘second vowel’ before the final consonant.

I “kitten” and “cuddle”: tongue never leaves alveolar ridgeafter [t]/[d].

I “banner”: you go immediately from [n] to [ô], with nointervening vowel







SyllabicConsonants

Onsets OverCodas

. . . . . .

Syllabic Consonants in English

Vocabulary:When a consonant serves as a syllabic nucleus, it is saidto be a syllabic consonant

IPA Representation:In IPA, syllabic consonants are indicated by putting a littlevertical line beneath them:

I “kitten” [ kI . tn"

]

I “cuddle” [ k2 . dl"

]

I “banner” [ bæ . nô"

]







SyllabicConsonants

Onsets OverCodas

. . . . . .

Syllabic Consonants in English

It seems, then, that the real syllabification rule for Englishoutputs representations like these:

σ

O

k

N

I

σ

O

t

N

n"

σ

O

k

N

2

σ

O

d

N

l"

σ

O

b

N

æ

σ

O

n

N

ô"

The Problem:

I Our current rule won’t make representations like thisI so, we have to fix our rule...I ... but it goes beyond ‘201’ to explain how.







SyllabicConsonants

Onsets OverCodas

. . . . . .

One Last Mystery

Question:Why don’t we syllabify “bologna” like this: *[b2l . own . i]

I This syllabification is totally consistent with Englishphonotactics.

I [b2l] is a possible English syllableI [own] is a possible English syllableI [i] is a possible English syllable

The answer comes from an important feature of our rule...







SyllabicConsonants

Onsets OverCodas

. . . . . .

Onsets Over CodasA Prediction of Our Rule:If phonotactics allow a consonant to be either an onset or acoda, then it will be an onset.

I [b2 . low . ni], not *[b2l . own . i]

Question: How, exactly, does our rule predict this?

Answer: Because our rule determines the onsets first.

I Step 2:Gather up all the consonants that can form an onset, andlabel them as such.

I Step 3:Of the remaining consonants, gather up all those that canform a coda, and label them as such.

Since our algorithm creates the onsets first......the codas are made out of only those consonantsthat couldn’t form onsets...







SyllabicConsonants

Onsets OverCodas

. . . . . .

Illustrative Derivation

Let’s see in detail how this works, by using our algorithmto syllabify “bologna”

Step 1:Label every vowel as a nucleus.

b 2 l ow n i







SyllabicConsonants

Onsets OverCodas

. . . . . .



Step 1:Label every vowel as a nucleus.

b

N

2 l

N

ow n

N

i







SyllabicConsonants

Onsets OverCodas

. . . . . .




b

N

2 l

N

ow n

N

i







SyllabicConsonants

Onsets OverCodas

. . . . . .




O

b

N

2

O

l

N

ow

O

n

N

i







SyllabicConsonants

Onsets OverCodas

. . . . . .

Illustrative DerivationLet’s see in detail how this works, by using our algorithmto syllabify “bologna”


O

b

N

2

O

l

N

ow

O

n

N

i

I Notice that on this step, we are forced to group [l] and [n]

as onsetsI The rule won’t let us “wait” to Step 3 to group them as

codas







SyllabicConsonants

Onsets OverCodas

. . . . . .

Illustrative DerivationLet’s see in detail how this works, by using our algorithmto syllabify “bologna”


O

b

N

2

O

l

N

ow

O

n

N

i

I Nothing happens at this step, since there are no‘remaining unlabeled phones’.







SyllabicConsonants

Onsets OverCodas

. . . . . .




O

b

N

2

O

l

N

ow

O

n

N

i







SyllabicConsonants

Onsets OverCodas

. . . . . .




O

b

σ

N

2

O

l

σ

N

ow

O

n

σ

N

i







SyllabicConsonants

Onsets OverCodas

. . . . . .




σ

O

b

N

2

σ

O

l

N

ow

σ

O

n

N

i







SyllabicConsonants

Onsets OverCodas

. . . . . .

Summing Up:

I There is a general ‘preference’ for consonants to beonsets, rather than codas.

I That is, if the phonotactics allow a particularconsonant to be either an onset or a coda, then it willbe an onset.

I [b2 . low . ni], not *[b2l . own . i]

I Our syllabification algorithm captures this fact, byordering the labeling of onsets before the codas.

I The codas will only be made out of those consonantsthat couldn’t be onsets.

syllables and phonotactics - umass · phonotactics the structure of syllables phonotactics...

Documents