Masaryk UniversityFaculty of ArtsDepartment of Englishand American StudiesTeaching English Language and Literature forSecondary SchoolsGabriela TomkováError Correction in Spoken PracticeMaster’s Diploma ThesisSupervisor: Mgr. Linda Nepivodová2013

I declare that I have worked on this thesis independently,using only the primary and secondary sources listed in the bibliography.……………………………………………..Author’s signature

AcknowledgementI would like to thank to my supervisor, Mgr. Linda Nepivodová, for her valuable advice, supportand eternal patience. I am also greatly indebted to all the teachers and students of English whotook part in my research and shared their thoughts and observations with me.

If to err and to speak are each uniquely human, then to err at speaking, or to commitlanguage errors, must mark the very pinnacle of human uniqueness.Carl James

Table of Contents1. INTRODUCTION.................................................................................................................72. HISTORICAL PERSPECTIVE..................................................................................................92.1. BEHAVIOURISM.......................................................................................................102.2. CONTRASTIVE ANALYSIS ..........................................................................................122.3. ERROR ANALYSIS .....................................................................................................162.4. KRASHEN’S NATURAL APPROACH AND INPUT HYPOTHESIS......................................192.5. COMMUNICATIVE APPROACH .................................................................................222.5.1. ERROR AND ERROR CORRECTION IN CLT ..........................................................243. ERROR IN LANGUAGE LEARNING.....................................................................................283.1. ERROR IN FIRST LANGUAGE ACQUISITION................................................................283.2. ERROR IN NATIVE SPEECH........................................................................................303.3. ERROR IN SECOND LANGUAGE ACQUISTION............................................................323.3.1. INTERLANGUAGE AND ERRORS ........................................................................354. ERROR AND MISTAKE IN SECOND LANGUAGE LEARNING.................................................384.1. DEFINITION OF ERROR.............................................................................................394.1.1. BACKGROUND..................................................................................................394.1.2. ACCURACY AND CORRECTNESS ........................................................................404.1.3. DEFINITION OF ERROR .....................................................................................414.2. ERROR VERSUS MISTAKE .........................................................................................454.3. SOURCES OF ERRORS...............................................................................................484.3.1. CONTROLING SOURCES OF ERRORS .................................................................52

4.4. TYPES OF ERROR......................................................................................................544.4.1. GENERAL CLASSIFICATIONS..............................................................................554.4.2. ERROR TYPOLOGIES .........................................................................................564.4.3. WAYS OF CATEGORIZING ERRORS....................................................................594.4.4. PSEUDO ERRORS AND LEARNING STRATEGIES..................................................625. ERROR CORRECTION .......................................................................................................655.1. DEFINITION OF ERROR CORRECTION........................................................................655.1.1. FEEDBACK ON ERROR.......................................................................................675.2. TO CORRECT OR NOT TO CORRECT?.........................................................................685.2.1. OPONENTS OF CORRECTION ............................................................................705.2.2. PROBLEMS OF HEAVY-CORRECTORS AND NON-CORRECTORS ..........................72

5.3. ERROR CORRECTION STAGES ...................................................................................735.3.1. CORRECTION OPTIONS.....................................................................................745.4. CORRECTION TECHNIQUES – HOWTO CORRECT ERRORS .........................................755.5. WHICH ERRORS TO CORRECT...................................................................................775.6. WHEN TO CORRECT .................................................................................................785.6.1. ACCURACY AND FLUENCY DICHOTOMY............................................................795.7. WHO SHOULD CORRECT ..........................................................................................835.7.1. SELF-CORRECTION............................................................................................845.7.2. PEER CORRECTION...........................................................................................865.8. WAYS OF INDICATING ERRORS.................................................................................876. BACKGROUND TO RESEARCH ..........................................................................................926.1. ANALYSIS OF COLLECTED DATA................................................................................926.2. STUDENT QUESTIONNAIRE ......................................................................................936.3. TEACHER QUESTIONNAIRE.....................................................................................1017. CONCLUSION.................................................................................................................1118. BIBLIOGRAPHY ..............................................................................................................1149. APPENDICES..................................................................................................................1239.1. Appendix 1 – Student questionnaire ......................................................................1239.2. Appendix 2 – Teacher Questionnaire .....................................................................1269.3. Appendix 3 – Students’ responses..........................................................................1299.4. Appendix 4 – Teachers’ responses .........................................................................137

9.5. Appendix 5 – List of Abbreviations.........................................................................14910. Summary...................................................................................................................15011. Resumé.....................................................................................................................151

71. INTRODUCTIONEvery day in every English lesson, teachers perform error correction. Since nohuman learning is perfect, it comes as no surprise that students of English make a lot oferrors in the process of acquiring the new language. In reaction to that, their teachersmust often provide them with some kind of feedback, which often takes the form ofcorrection. Although the process of correcting might seem straightforward at first sight,it is in fact a very complex issue that involves many decisions on the part of the teacherbefore any correction as such is actually carried out.This diploma thesis is devoted to error and error correction in oral practiceduring English language lessons. The goal is to map the circumstances of various errorsthat learners make during speaking as well as possible ways of correcting such errors. Iwill try to evaluate and analyse individual aspects of error and correction and try tocome up with some conclusions that might help teachers in their decisions on errorcorrection. The research part will try to gain information from both teachers andlearners of English regarding their opinions on error and correction. These will then beanalysed and compared with theory.The thesis is divided into five chapters. The first chapter introduces severalapproaches towards error and correction that have been dominant in second languageacquisition theory since the mid-20th century until today. These theories are introducedand critically analysed. The communicative language approach to teaching is alsopresented here.The second chapter introduces the notion of error in language learning ingeneral. Firstly I discuss error and correction in first language acquisition and nativelanguage conversations. Subsequently, error in second language acquisition is dealtwith. A short section on learners’ interlanguage is also included here.

8In the following chapter a closer look at error and mistake in second languageacquisition is taken. After discussing accuracy and correctness, I try to come up with asuitable definition of error. Subsequently, error is contrasted with mistake and a strictdivision is drawn between the two. Various sources of error in speaking as well as theirpossible ways of limitation are discussed. Finally, several classifications and typologiesof error are analysed and evaluated; a short section is devoted to pseudo errors andlearning strategies, too.The fourth chapter discusses in detail error correction in oral practice. First of allcorrection is defined and contrasted with feedback or treatment of error. Afterwards, thequestion of effectiveness and usefulness of correction is discussed; arguments ofcritiques of correction are presented and analysed; problems of too much and too littlecorrection are dealt with, too. Subsequently, stages of error correction and possibletechniques for it are presented. Individual decisions regarding when to correct, whicherrors to correct, who should correct, and how are then discussed and evaluated. Finally,individual ways of indicating errors are presented and analysed.The last chapter of the thesis discusses the results of my research. The researchis shortly introduced, after which I present results of individual questionnaires andcompare them both with each other and theoretical observations. To make it morecomprehensible, the findings are supported by several charts and graphs. A conclusionis drawn from the practical findings and the theoretical background.Throughout the paper, I often refer to teachers of English in singular and in thosecases I use the pronoun “she”. The reason for this is that most teachers of Englishlanguage in the Czech Republic are female. That does not mean, however, that theobservations do not relate to male-teachers, too.

92. HISTORICAL PERSPECTIVELanguage, as one of the typically human characteristics, has been subject togreat scientific study for over a long time. Already at the time when Latin was thedominant language in Europe theorists studied the language from several perspectives.At that time, however, only receptive skills were studied, as people learned the languagethrough literature and writing only, which was then followed by a detailed study ofgrammar. In this grammar-translation method, as Lightbound (2005) observes, “themain activity was focused on written translation, and no attention was given to theactual communicative use of the language in question” (p. 64). Later on, when Englishtook over as the most widely used language worldwide; it adopted the same methods forstudying and teaching language as Latin. It was assumed that grammar-translationmethod proved its efficiency through time thanks to its dominant status and thus had tobe a high-quality method.With an increased need for actual communication in English, the 20thcenturysaw the grammar-translation method ceasing to be sufficient and therefore new methodsof approaching language learning were needed. As a result, numerous approaches tolanguage acquisition emerged explaining the process of language acquisition andsuggesting new methods of learning and teaching. As Christison & Krahnke (1983)point out, perhaps no aspect of language pedagogy has been the subject of more interestthan the study of learner error. The fact that sources of error cannot be observed in anyway made it a feature of language learning that has gained considerable attention, aseach method was trying to come up with a reasonable explanation. This part of thepaper presents those approaches to second language learning that deal with error anderror correction the most and pay close attention to it. In order not to make it too

10lengthy, I decided to concentrate on those approaches that were prominent in the mid20th century and later.2.1. BEHAVIOURISMTeachers are often worried when their students make errors, because theybelieve the students may learn the errors and keep using them. Therefore teachers thinkthey must avoid this happening and make sure that everything students say must beperfect. This belief originally comes from behaviourism – an approach to languagelearning, which was dominant in the 1950s and 1960s. It was one of the first schools ofthought – if not the very first –which dealt with learners’ errors in great depth.Behaviourists believed that learning a language is a habit formation – a purelymechanical process. They viewed language learning as a process of acquiring skills,similar to learning to do something practical, like cooking or driving a car. “Thecomplex skill was broken down into a series of habits, which were drilled until theybecame automatic and unthinking.” (Hubbard et al., 1983, p. 133). As Hubbard furtherobserves, the behaviourist would not say that a speaker of a language knows hislanguage, but rather that he is able to perform in it. From this point of view, repeatingand imitating incorrect forms would be harmful and thus had to be avoided.If an error really occurred, it was a sign of bad teaching methods. As Corder(1984) points out, “if we were to achieve a perfect teaching method the errors wouldnever be committed in the first place” (p. 20).The occurrence of errors was thus solely asign of one’s inadequate teaching techniques. Errors were generally seen as somethingnegative, signalling a failure. In lessons, they were corrected immediately after theyappeared: “When errors do occur, they are to be remedied by a bombardment of correct

11forms. This bombardment is achieved by the use of intensive drilling or over-teaching”(Hubbard et al., 1983, p. 144). The drilling tasks were designed so that the studentswere always speaking or writing correct English only. In behaviourist lessons, there wasthus no space for free practice or improvisation, as in such cases it was impossible toprevent students from making errors and to create ideal learning conditions.The notion of errors as something harmful was even strengthened by scientificresearch. One of the first was Kay (1951), who claimed that it is extremely difficult forthe organism to rid itself of errors. This difficulty increases with age and therefore olderlearners will take longer to unlearn their errors. Another study by Belbin, Downs, andMoore (1970) provided support for this, when it concluded that errors greatly hinder thelearning process, as “the older learner cannot readily accept the notion that his originalresponse was wrong” (Singer, 1977, p. 484). It was thus believed that the older thelearner, the greater the necessity to prevent errors from happening.To some extent, the Behaviouristic school was right. Structural drills can beuseful, if one wishes to get rid of an error, but they cannot be the only means to do so.As Hubbard et al. (1983) stress, teachers should always remember that “the learner isdoing his best to sort things out for himself and requires intellectual, as well asmechanical help” (p. 145). They believe it is not enough to provide learners withmechanical drills only, as they need other types of guiding, apart from mechanical, too.For many years, the behaviourist view was widely accepted. As Broughton et al.(2003) note, “At first sight it appears self-evident that errors are a very bad thing andsignal a breakdown in the teaching and learning situation” (p. 134). Continuously,however, it was realised that drills and mechanical learning were not enough forstudents to actually learn a language and be able to freely communicate in it. Thusbehaviourism was gradually replaced by new dominant linguistic methods, mainly

12contrastive analysis and error analysis, which will be discussed in the followingsections.2.2. CONTRASTIVE ANALYSISFrom the 1950s till the early 1970s, contrastive analysis (CA) was a widespreadmethod in the field of second language acquisition (SLA). It was used for explainingerrors made by language learners. Additionally, the method was trying to say why somelanguage features were more difficult for learners to acquire than others. Manyresearchers at that time carried out contrastive analyses between pairs of languages (seee.g. MacDonald Lightbound, 2005, p. 65).The contrastive analysis theory was built on the behaviourist view of languagelearning as a mechanical process of habit formation. It was believed that the aspects ofthe target language (L2) which were different from one’s mother tongue (L1) wereconsequently to cause problems in learning, as learners had no habits to follow from. AsMacDonald Lightbound (2005) states, the logic was that “when learning a secondlanguage a person will tend to use mother tongue structures in second languageproduction, and where L1 structures differ from the L2, mistakes will be made” (p. 66).Finding out the differences and similarities of individual pairs of languages was thusbelieved to be enough to handle the problems arising in teaching. Based on that view, itwas believed that “contrastive analysis can highlight and predict the difficulties ofpupils” (Richards, 1984, p. 172). The method therefore concentrated on predictingplaces of error before any were actually made.An extreme claim of contrastive analysis stated that all errors in languagelearning were caused by interference, or transfer, from L1 into L2. “If two languages

13were similar, positive transfer occurred, if they were different the transfer would benegative” (MacDonald Lightbound, 2005, p. 66). Interference causing learners to makeerrors was thus negative transfer. The amount of interference in one’s language learningwas something CA also predicted: “the less of a bilingual the speaker is, the moreinterference there will be when he attempts to communicate with speakers of the targetlanguage” (Dulay & Burt, 1984, p. 104). On the basis of this knowledge, contrastiveanalysts believed to be able to predict what and how many errors each learner wouldproduce.These notions about interference were, however, challenged by growingempirical research in the 1970s. Theoretical expectations raised by CA failed to explainall types of error that in reality occurred. As Richards and Sampson (1984) point out,“errors which did not fit systematically into the native language or target language were,for the most part, ignored” (p. 4). It was often the case that teachers were dissatisfied byfindings of CA, as it in fact did not tell them much what they did not know from theirpractice already. Moreover, many errors noted in classroom were not even predicted bythe analysis anyway (Nemser, 1984). As Richards and Sampson (1984) stressed, “thesecond language learner tries to derive the rules behind the data to which he has beenexposed, and may develop hypotheses that correspond neither to the mother tongue nortarget language” (p. 6). Even more puzzling was the observation that a large number oflearners, no matter what their L1 was, tended to make the same errors. As Dušková(1969) indicated, L1 interference explains only some of the errors involved in learnerproduction, and it does not take into account the interference observed between theforms of the language being learnt, both in grammar and lexis. Therefore it becameevident that CA could not embrace all learning problems and places of making errors.

14Contrastive analysis often predicted interference errors only. Apart from these,there are also other types of error; these, however, cannot be explained by interference.Richards (1984) calls these intralingual and developmental errors. As he explains, these“reflect the learner’s competence at a particular stage, and illustrate some of the generalcharacteristics of language acquisition. Their origins are found within the structure ofEnglish itself, and through reference to the strategy by which a second language isacquired and taught” (p. 173). Such errors can be expected from anyone learningEnglish as their second language, irrespective of their mother tongue.Another shortcoming of CA is seen in its preoccupation with merely theoreticaldescriptions, which often did not even reflect reality. Its close attention to the analysisof two grammars is by some linguists (e.g. Richards and Sampson, 1984) even seen asits major defect. A closer study of performance of actual learners would be preferredhere. Logically, language instructors are generally more interested in real difficultiesrather than those predicted by theory. It was also claimed that “too much attention wasbeing paid to hypothesising about what the learner may do, to the detriment of studyingwhat s/he actually does” (MacDonald Lightbound, 2005, p. 68). Teachers therefore sawCA as impractical and over-theoretical, as it failed to answer questions concerningdiffering levels of success in individual language learners.It would, however, be wrong to claim that findings of contrastive analysis arenot predictive at all. As Richards (1984) admits, “Studies of SLA, however, have tendedto imply that contrastive analysis may be most predictive at the level of phonology, andleast predictive at the syntactic level” (p. 172). As Warsi (n.d.) notes, it may be due tothe “heavy reliance on structuralistic comparison why contrastive analysis fails topredict such mistakes” (p. 11). It may thus be advisable to rely on CA with respect to

15problems with pronunciation, but not such a good idea in the case of word order,sentence formation and such.In general, contribution of contrastive analysis to the study of learner languagecannot be denied. S. P. Corder (1984) sums it up in his paper on “The Significance ofLearners’ Errors”:The major contribution of the linguist to language teaching was seen as anintensive contrastive study of the systems of the second language and the mothertongue of the learner; out of this would come an inventory of the areas ofdifficulty which the learner would encounter and the value of this inventorywould be to direct the teacher’s attention to these areas so that he might devotespecial care and emphasis in his teaching to the overcoming, or even avoiding,of these predicted difficulties. (p.19)He also adds that thanks to CA, the emphasis was shifted away from a “preoccupationwith teaching towards a study of learning” (p. 20), although this was rather a side effectof CA than an aim. Bartram and Walton (1991) see contrastive analysis beneficial in theway it shed light on the relationship between learners’ L1 and L2. Nevertheless, theyseem to interpret the relationship in a more modern way: “the L1 is a resource which thestudent uses when, for some reason, the L2 form eludes them. Equally, other languages- an L3 - can be called upon, a habit which seems to us to be natural, intelligent andresourceful, and so to be encouraged” (p. 16). However, they still acknowledge that CAwas the first method to shift attention of linguists towards this relationship, althoughwith a slightly different explanation.Today CA is seen as an explanatory method rather than a predictive one. AsShastri (2010) notes, it “helps us to decide the extent of the interference of L1 in L2” (p.26). Broughton et al. (2003) have come to the same conclusion and subsequently add:

16“It [L1 interference] is one of the possible causes for error which the teacher mustconsider, not a basis on which stands all his teaching” (p. 136). Richards and Renandya(2002) in their paper on “Methodology in Language Teaching” observe that CA hasrecently given way to theories of equivalence classification, which are moresophisticated. In these theories, “learners approach a new sound system by mapping itonto their L1 sound system, using existing categories where similarities exist and createnew categories for unfamiliar features” (p. 181). This view of language learning is morecomplicated than CA and seems to embrace the complexity of the process moreprecisely.2.3. ERROR ANALYSISIn the late 1960s and early 1970s emphasis was shifted from contrastive analysisto error analysis (EA). One of the main reasons, as mentioned in the previous section,was the fact that CA was unable to explain all errors in language learning, as it is notpossible to ascribe them all to differences between L1 and L2. Moreover, some of thepredicted errors did not even appear in practice (see e.g. Allwright & Bailey, 1991).Thus there was a need for a new linguistic method, which consequently took form oferror analysis. One of the major proponents of error analysis was Stephen Pit Corder.Error analysis was established as an alternative to contrastive analysis. Bothmethods explain sources of error, but in different ways. CA sees errors as results of L1interference only. EA, on the other hand, “accepts many sources of errors such asintralingual interference, overgeneralization, misteaching and the role of the variables ofage, attitude, aptitude, motivation etc.” (Shastri, 2010, p. 25). An important point that

17EA made was that many learners often wrongly inferred rules of their L2, which thencaused a large number of errors.It is important to stress that EA not only explored sources of error, but also manyother aspects of error and language learning. As Yang (2010) states,Error Analysis is the process of determining the incidence, nature, causes andconsequences of unsuccessful language. The novelty of EA, distinguishing itfrom CA, was that the mother tongue was not supposed to enter the picture.Errors could be fully described in terms of the TL, without the need to refer tothe L1 of the learners. (p. 1)The shift of focus from learners’ mother tongue onto the target language is the majoraspect that distinguishes EA from the preceding methods. Another important changewas that errors were not only predicted, but mainly observed, analysed and classified.Similarly to contrastive analysis, EA also compares utterances in two“languages” – the learner’s language with the target language. In the study of languagelearning, as Corder (1981) specifies, “we are interested in the relation of what has beentaught so far with the learner’s knowledge at that same point” (p. 57). These twoutterances compared can be regarded as synonymous, or rather as equivalent. Errors areidentified by comparing the erroneous utterances with correct utterances in the L2having the meaning originally intended by the learner. As Corder explains, “We canregard the reconstructed utterances as translations of the learner’s utterances into thetarget language” (p. 37). In this aspect, EA is very much like contrastive analysis.The process of analysing an error has three stages. According to Shastri (2010),these are identification, reconstruction and description of error. First of all, one needs toidentify an utterance as being erroneous (or idiosyncratic as referred to by Corder).Such idiosyncratic utterance is then compared with the L2 and an utterance with the

18same meaning is reconstructed. The final stage of the analysis is a description of howand why the idiosyncratic utterance occurred (for more detail, see Corder (1984), pp.166-9).Error analysts frequently came up with various error typologies. The mostwidely spread division was between errors (systematic) and mistakes (accidental). Fromthat, error typologies were derived. To illustrate the point, Larsen-Freeman and Long(1991), in their summary of intralingual errors (those that are not traceable to L1interference), identified four types: overgeneralization, simplification, communicationbasederrors, and induced errors. S. P. Corder (1984) offers a similar typology, althoughwith slight changes. Each of these categories is always clearly defined and distinguishedfrom the others, which makes it easy to categorise all errors. These typologies will bepaid closer attention further in the paper in the section on Types of Error.Despite the “sophisticated” process of analyzing errors, since its very beginning,EA was said to lack methodological rigour and to be limited in its scope (Ellis, 1985).As MacDonald Lightbound (2005) claims, one of the problems was that severalresearchers “have pointed out the negative attitude reflected in its analysis of ‘what iswrong’ or deviant in the learners’ IL, while ignoring the achievements” (p. 74).Furthermore, the various typologies of error were problematic, too. In theory, it is verysimple to distinguish between different kinds of error – not so much in real practicethough. It was often impossible for teachers or researchers to find out what kind of errora learner was making. Last but not least, it was impossible to analyse errors such asavoidance of certain forms, which learners used when they needed to get by without anunfamiliar form.Finally, it is necessary to mention that the transition from CA to prevailing erroranalysis was not as linear as it might seem. As MacDonald Lightbound (2005) claims,

19…until the proponents of two different views come to a mutual agreement thatrecognizes the importance of taking into account both ends of the pendulum, i.e.L1 interference and intralingual interference, there have been dramatic swingsfrom one extreme to the other, which serve at best, to cause confusion amongapplied linguists working in the field of SLA. (p. 107-8)It would thus be misleading to assume that EA globally took over the previous methods.As it seems, it took some time to establish a balance between the two methods (CA andEA), with error analysis gradually dominating the field. MacDonald Lightbound (2005)sees the reasons for its resulting dominance in its greater universality: “EA adopted amore balanced view with regard to the role of the MT in second language learningenvironments. Intralingual errors are liable to be committed by learners from all aroundthe globe regardless of their L1” (p. 108). Despite its initial struggles, it thus seems thatEA prevailed thanks to its more general applicability.Error analysis was at its peak in the 1970s and 1980s; after that, however, itsinfluence began to fade. Since then, “researchers working more on matters related toerror gravity, and causes of error, in particular, transfer or cross-linguistic influence. Inthis sense, it has never totally disappeared from the language learning scenario,continuing to be a more widespread practice than it is duly given credit for” (James,1998, p. 18). Continually, it has given way to the communicative approach, which keepsprevailing still today and is a largely dominant practice overall.2.4. KRASHEN’S NATURAL APPROACH AND INPUT HYPOTHESISSo far I have described approaches to language learning that deal with error anderror correction in one way or another. There are a few approaches to second language

20learning (SLL), however, which advocate no error correction at all and consider error inlanguage learning as irrelevant. Perhaps the most prominent of these is Stephen Krashenand his hypotheses regarding SLA.There are very few language-teaching experts today who promote no focus onform at all. A prime advocate of this view is Stephen Krashen with his NaturalApproach. This theory is strongly against any error correction in L2 acquisition anderrors are not treated at all. Together with Tracy Terrell (1983) they developed a secondlanguage hypothesis claiming that learners should never be forced into speaking in L2before they are ready for it. They introduced the notion of “silent period” (Krashen,1982) during which learners only listen and decide for themselves when they are readyto perform in L2, too. After the silent period an early production stage comes whenteachers are supposed to focus on meaning and not to correct errors during this stage(Brown, 2007, p. 31). It is only during the last stage of extended production whenteachers are allowed to correct some errors, though very sparsely and only in utmosturgency. The reason is that Krashen believes that when students are relaxed, acquisitiontakes place. And since he believes that language acquisition which occurs outside ofconscious awareness is more effective than conscious learning (Krashen, 1982), anyfocus on form is undesirable and may even hinder the process of acquiring the learner’sL2.The core thought behind refusing any error correction is Krashen’s inputhypothesis. This hypothesis claims that “We acquire by understanding language thatcontains structure a bit beyond our current level of competence (i + 1). This is done withthe help of context or extra-linguistic information” (Krashen, 2009, p. 21). Krashen thusbelieves that if comprehensible input i + 1 is provided to the learner, it is all he or sheneeds to develop his interlanguage further. According to him, learners do not need to be

21aware of the gap between their knowledge and the input – that is, they do not need toknow that they are learning something new. Focus on form is undesirable here and thusno correction or error treatment is needed whatsoever. As he claims, acquisition is aresult of comprehensible input, not error correction.Krashen (2009) calls error correction “a serious mistake” (p. 74) and sees it asunfortunate that correction is teacher’s typical reaction to error. In his view, correctingerrors of form is in fact wasting time and effort; moreover, the teacher is doingdisservice to his or her students. As students make many meaning errors already, itwould be harmful to intervene with form-based errors, too. Students would then feeloffended, refuse to speak in class and thus the atmosphere would get tense. The reasonis Krashen believes that correction puts students on the defensive and thus encourages astrategy of avoiding any possible errors, which brings about more focus on form ratherthan on meaning. This, as a result, “may disrupt the entire communicative focus of anexchange” (Krashen, 2009, p. 75). By no means would such reaction be wanted andthus he advocates abandoning error correction. As Krashen (2009) believes,improvement may even come faster without correction, since the input will more easilyget in, “the [affective] filter will be lower, and students will be off the defensive” (p.76). Generally, it is thus more beneficial for learners if their teachers decide to give upform-based error correction at all.Stephen Krashen with his refusal of error correction and lack of attention paid toerrors in general belongs to the more extreme and controversial language-teachingtheorists today. His belief that errors of form should be completely overlooked and thatteachers should not correct their students is in conflict with most other SLA theories.The vast majority of them indeed see error correction as facilitative and useful (Ellis,2009). Nevertheless, some of Krashen’s ideas are regarded of use, such as leaving

22learners enough time to listen before forcing them to speak, or avoiding over-correctionso as not to discourage students from speaking. The next section will move to discussthe currently most widespread approach to learning L2 – one that deals with error andcorrection much more than Krashen does – and that is the communicative languageapproach.2.5. COMMUNICATIVE APPROACHThe dominant approach to language teaching today is communicative approach.It is often thought of as a general approach to teaching, rather than a method withclearly defined rules and practices. This approach started to gain significance already inthe 1970s and early 1980s. At that time, commonly used methods such as the grammartranslationone started to appear inadequate when students needed to learn a languagequickly or when they were not particularly talented. A new approach that would aim atfluency and interaction was thus needed.Many aspects of communicative language teaching (CLT) are different fromthose of the earlier approaches. Teachers are no longer seen as dominant in the learningprocess, but rather as co-communicators with learners, which makes them equal. Thisshould lower tension and break down barriers between them (Littlewood, 1994). Agreater emphasis is placed on learners’ independent learning. Communicativeinteraction in classes allows for cooperative relationships to be created among allparticipants, including the teacher. Regarding errors, not all of them are dealt with andlearners are not being perpetually corrected. Errors are more tolerated; they are seen as a“completely normal phenomenon in the development of communicative skills”(Littlewood, 1994, p. 94). This is in short the core idea of CLT. The learners are given

23more space and they enter the learning process with their unique personalities. Theteachers, on the other hand, are to step down from their pedestals, where the earlierapproaches used to place them, and become more of “humans among humans”(Littlewood, 1994, p. 94).CLT presupposes a different aim of students, too. It believes that what studentsseek when learning an L2 is not a native-like perfection, but rather an ability to freelycommunicate and interact in the L2. As Bartram & Walton (1991) observe, very fewstudents are interested in the language for its own sake. Vast majority of them need touse the language as a practical skill for communication. This implies that a goodlanguage user is not one who completes drill exercises without an error, but rathersomeone who is able to communicate well. The aim thus is not perfection, butcomprehensibility. As Bartram and Walton add, “When a person says someone’sEnglish is good, because s/he makes very few mistakes, you can be sure that the personmaking the judgment is a language teacher” (p. 5). Hopefully, the results of my surveywill show that there are fewer and fewer of such teachers who see the criterion ofnumber of mistakes as the crucial one.The attitude towards error gradually started to change already in the 1970s.Although the dominant belief still held the status of an error as something negative, as asign of failure in SLA, there were linguists who already saw advantage in makingerrors, too. Already in 1970 Holding observed that[…] learning by making errors may occur in circumstances where gaining extrainformation about the task depends upon executing a wide range of responses. Inthis case, making or being forced to make only the correct responses may denythe subject knowledge of alternatives, with detrimental effects upon theacquisition of skill. (p. 733)

24He thus admitted the existence of circumstances when it might be beneficial for learnersto make errors. Almost fifteen years later, Richards (1984) raises a strong claim againsterror-free learning, when he states:[…] the conventional approach, which proscribes the interlingua as a “corpus oferror” either stifles the learner’s communication drives altogether, or requiresthat the linguistically mature student becomes as a little child, practisingperfectly well-formed native-speaker’s sentences, which are, however, oftenidealized and usually trivial. (p. 89)From these viewpoints, it is possible to see that the behaviourist approach was alreadylosing its significance and the more modern approaches, from which the CLT approachderives, were gaining prominence. It was mainly thanks to the mentalist approach,which saw errors as an inevitable part of SLL: “a learner must make errors as anunavoidable and necessary part of the learning process, so errors are not the bad thingonce thought but visible proof that learning is taking place” (Broughton et al., 2003, p.135). It is thus evident that the attitude to errors was gradually changing towards a morefavourable status in SLL. The same can be said about error correction: “The correctionof errors provides precisely the sort of negative evidence which is necessary todiscovery of the correct concept or rule” (Corder, 1984, p. 170). Correction will be dealtwith in more detail in the subsequent chapters.2.5.1. ERROR AND ERROR CORRECTION IN CLTIt was already suggested that error in language learning has a different place inCLT than it used to have in the earlier methods, such as the grammar-translation one.As Hendrich et al. (1988) observed, the key criterion in the contemporary approach toSLL in deciding whether something is an error or not is the extent of disruption of

25communication. Learning anL2 is viewed today as a system of rules that the learner hasto acquire, and that “trying out language and making errors are a natural andunavoidable part of this process” (Doff, 1993, p. 187). As Doff adds, this should be farfrom surprising when we take into account the acquisition of other complex skills, suchas mastering chess or learning a musical instrument, where errors are accepted as anatural part of the process. Further on, Doff makes a crucial claim:Students’ errors are a very useful way of showing what they have and have notlearnt. So instead of seeing errors negatively, as a sign of failure, we can seethem positively as an indication of what we still need to teach. Obviously, if wetry to prevent students from making errors we can never find out what they donot know. (p. 188)This belief, although pronounced twenty years ago, is still valid today. Errors aregenerally seen as a sign of learning taking place and students experimenting withlanguage, rather than as signs of their failure or inabilities. Broughton et al. (2003)describe the likely process that goes on in learners’ heads when experimenting withtheir L2 as follows: First of all, the learner makes a guess based on his knowledge of hismother tongue as well as his partial knowledge of the target language. As hiscompetence in the L2 language grows, he keeps refining the originally formulatedhypotheses. He thus moves “from ignorance to mastery of the language throughtransitional stages, and the errors he makes are to be seen as a sign that learning istaking place” (p. 135). It is thus clear that in CLT, errors are seen as an inevitable partof the learning process and that they are far from being viewed as something harmful ornegative, which was the case in the earlier approaches to error.One might ask what implications this new concept of error has for teachers. Themain demand on teachers is that they should not correct all errors of their students. As

26Hubbard et al. (1983) note, “Teachers will also have to allow errors to go uncorrectedon many occasions – something which the behaviourist would not feel happy about” (p.135). This is true of fluency activities mainly, as theory says. At some points, it mighteven be impossible for teachers to spot all errors, such as in group or pair work tasks.This, however, might create an internal conflict in some teachers. Their pedagogicalrole traditionally required from them evaluating learners’ performance on the basis ofclearly defined criteria. It thus seems that adopting the communicative approach forcesthe teacher to redefine this traditional role – at least to some extent. This requiresteachers to adopt a certain amount of flexibility and open mind. If we, however, wantour students to gain confidence, be creative, test out new grammar hypotheses, inventvocabulary, and practise new structures (Bartram & Walton, 1991), then we inevitablyneed to adopt such approach and let some errors be uncorrected.This section has introduced the communicative approach towards languagelearning, which is the prevailing approach used in SLL in general. It has shown itsdifferences from the older approaches and outlined its beneficial aspects. I have pointedout that teachers are regarded as equal to learners, somebody who is a partner inlearning to them. Learners are learning more independently and are responsible for theirlearning more than before. CLT generally aims at being able to communicate freely in alanguage rather than being perfect. A greater tolerance of error is proposed, too, sinceerrors are understood as a normal and inevitable, sometimes even beneficial part oflearning. Hand in hand with this goes the concept of limited error correction – not allerrors should be corrected; the key in deciding on error treatment is the level ofcomprehensibility or, the level of disruption of communication. This is because theprocess of learning a foreign language is seen as constant hypotheses (re)formulationduring which errors work as signs of learning taking place. The next part of the paper

27will deal with error in language learning and present the various views held on it. I willalso describe the differing approaches towards error in learning first and secondlanguages. A short section will be dedicated to errors in native speech, too.

283. ERROR IN LANGUAGE LEARNINGAs the motto of this paper already prefaced, it is typically human to make errorsin using and learning a language. Part of our experience with error in language learningis in witnessing small children learning their mother tongue. This acquisition of one’sL1 has been thoroughly studied for a long time. Before moving to consider errors inSLL, it may thus be instructive to compare the situation with what theory has to sayabout the “errors” children make when learning to talk. Furthermore, I will have a lookat errors made in native language conversations, since these are often forgotten incomparison to learning an L2 despite having something in common.3.1. ERROR IN FIRST LANGUAGE ACQUISITIONFirst language acquisition (FLA) is in many ways similar to second or foreignlanguage acquisition in general. It thus suggests drawing parallels between variousaspects of the process, such as making errors in the process of learning and errorcorrection. It was probably S. Pit Corder who first established an analogy betweenerrors made by children learning their L1 and those of a second language learner. Manyerrors typically uttered by children, such as overgeneralizing the past tense rule into“goed” instead of went, are also typical of second language learners. It is thus clear thatthere must be some parallel between the two processes.There can be no doubt that children’s language is very different from the one ofadults. As Allwright & Bailey (1991) observe, it is precisely the errors the child makesand the simplified, erroneous utterances that characterise children talk the most. In thissense, children language is very close to the language of second language learners.

29Their utterances can also be often described in terms of the errors they make – although,as Allwright and Bailey (1991) point out, it feels somewhat strange to talk aboutchildren language as being erroneous.Occurrence of errors in FLA is not usually seen as a negative aspect – somethingdeviant, ill-formed, or faulty. On the contrary, it is thought to be a good evidence of thechild obtaining new construction rules of the language: “We do not regard it as an errorin any sense at all, but rather as a normal childlike communication which providesevidence of the state of his linguistic development at that moment” (Corder, 1984, p.22). If, on the other hand, the child was producing correct utterances only, it would bequite likely that he or she was only repeating or imitating something he or she heard.Errors are thus viewed as a positive sign of learning taking place.Adults’ reactions to children’s error in speaking are rather different from whatwe call error correction in SLA. Parents or older and more competent speakers ingeneral tend to focus on the content of the utterance rather than on the form. AsAllwright and Bailey (1991) point out, “factual content and effectiveness areemphasised over form in such interactions” (p. 86). It is common to focus ontruthfulness and appropriateness, not on linguistic perfectness. As they further observed,negative feedback is not usually provided in any systematic way; when it is provided,however, it does not seem to have much effect on the child.The most common type of feedback children get on their erroneous utterances isrepetition and expansion. Corder (1984) states that when a child produces a sentencesuch as This is mummy chair the common reaction of its caretakers is to repeat andimprove the utterance in an adult way – something like “Yes, dear, that’s Mummy’schair” (p. 22). The aim of such reaction is to add words or morphemes and thus to create

30a grammatical utterance. It is not, however, intended as criticism or a reproof, but theadult usually utilizes the child’s utterance, praises it and suggests little improvement.From observing children experimenting with language it is clear that they learnmainly through the method of trial and error. Gradually, children accumulate more dataand rules about their L1 and the process is characteristic by a gradual disappearance oferrors in their speech. This process is rather complex and far from just memorizingindividual sentences pronounced by adults, as the behaviourist theory believed. Duringthe process, children make many errors, which are, however, seen as a positive sign ofexperimenting with language and gradually acquiring new rules. As Bartram andWalton (1991) stress, this fact is in a way important for second language learners andtheir teachers: it tells us that errors are a part of the learning process and absolutelynatural: “not wrong turnings on the road towards mature language use, but actually partof the road itself” (p. 11). This is an important notion to bear in mind when dealing witherrors of second language learners. Before I examine these errors in SLL, I would liketo briefly comment on errors in adult native speech and error correction connected to it.3.2. ERROR IN NATIVE SPEECHWhile we do not usually think about it that way, it is necessary to stress thaterrors occur even in native language conversations of adults. These errors are not asapparent as those of first or second language learners and thus do not usually come toone’s mind when thinking about errors in language use. When such errors occur, theytend to be overlooked unless they cause a serious breakdown in communication.Everybody tends to make errors in their L1. It was Noam Chomsky who claimedthat natives make a lot of errors when speaking (performing), “even though a native

31speaker has, by definition, a perfect command of his language – perfect knowledge ofgrammatical rules, lexis and the sound system” (Hubbard et al., 1983, p. 133). Theseerrors can be made due to a memory lapse, tiredness or perhaps mental stress. They areoften called performance errors(in Chomskyan terms) or slips of the tongue. In suchcases, meaning is clear from the context and in most cases there is no seriouscommunication breakdown. Typically, speakers are able to self-correct.Once a serious error in native conversation occurs there is a need for pointing itout and fixing the breakdown. The issue of who will point out the existence of the error,be it the speaker or someone else in the conversation, is referred to as self-initiation orother-initiation (Allwright and Bailey, 1991). Such initiation does not need to be verbal.The actual fixing of the error is referred to as “repair” (rather than correction); this, too,can be carried out either by the speaker (self-repair) or someone else (other-repair).Based on these options, Schegloff, Jefferson and Sacks (1977) offer four possiblecombinations that may occur in native speech conversations:1. self-initiated other-repair, when speakers note errors, but ask for help withfixing it (such as searching for a word they wish to use);2. self-initiated self-repair, when speakers are able to both notice and repair theerrors;3. other-initiated self-repair, when someone else signals the communicationbreakdown, but the speaker is able to self-correct; and4. other-initiated other-repair, when the interlocutors both note and repair theerrors.Research into native English conversations has so far shown that speakers have a strongtendency for self-initiated self-repair. On the other hand, the least frequent combinationof initiation and repair is other-initiated other-repair (Allwright and Bailey, 1991). This

32makes it very different from both FLA and SLA. As was stated in the preceding section,it is predominantly other-initiated other-repair that takes place in learning one’s L1. Thesame is true for classes of SLL, where there has traditionally been a strong preferencefor other-initiated other-repair, too. As Allwright & Bailey (1991) note, thischaracteristic marks second language classes “as somehow different from “real life”” (p.90). It would certainly feel awkward to receive evaluative feedback on the linguisticform of one’s utterance in normal conversations outside classroom. While bearing thisissue in mind, teachers should always consider whether students are not able to carryout self-initiated self-repair before they step in and correct them. In doing so, theconversations in second language classes would resemble real-life native talk muchmore.3.3. ERROR IN SECOND LANGUAGE ACQUISTIONFor the time being, we will limit ourselves to saying that error in SLL is usuallyidentified as a form that deviates in a certain way from the typical form produced bynative speakers. This comparison to native speaker’s language is called the nativespeaker norm (Allwright & Bailey, 1991). Although such comparison is not ideal, it isthe most widespread one.There are several problems that the native speaker norm has to face. Firstly, avast majority of language learners all over the world are taught by non-native speakersof English. Thus the model they receive might actually deviate from the native speakernorm, too, depending on the proficiency level of the teacher. This makes trying toachieve the native-speaker norm even more difficult for the learners, since the inputthey are exposed to is already different from the norm as well. Secondly, it is often the

33case that even native speakers of a certain dialect produce and use the same linguisticforms differently. The question is whether all the different forms should then beregarded as the native speaker forms, or just some of them, and how we should decide.Lastly, the situation in some parts of the world – such as India, Hong Kong, or WestAfrica – is even more complicated. As Allwright and Bailey (1991) observe there arealways alternative norms in these regions that are perfectly acceptable (an examplebeing West African English). These and other problems make it difficult to simplycompare learners’ utterances with native speakers’ ones and base the notion of errors onthis comparison. It will be useful to bring this to mind later on when I discuss theproblems of defining error in more detail. For now it is sufficient to keep this in mindwhile we discuss other aspects of error in SLL.It is indeed impossible to avoid making errors in any occupation we have – notexcluding language learning. In SLL, however, errors can be a very positive aspect inthat they show us that “learning is taking place and that learners are taking risks withthe language” (Spratt et al., 2010, p. 143). Errors are thus not to be seen as signs offailure or inhibition, but rather as evidence of students experimenting with language.The fact is that if learners, on the other hand, speak correctly only, we cannot be surewhether they have learned new forms or not. Learners may, after all, be merelyrepeating utterances they heard, in which case we cannot speak of acquiring newlanguage, but of “language-like behaviour” (Spolsky, 1966, cited in Corder, 1984, p.26). Errors can thus be regarded as useful indicators of learners advancing in their SLL,since they provide us with feedback on their mental processing of new rules andinformation.Communicative language approach to SLL plays a significant role in the recentview of errors as a positive aspect of learning, not as something harmful. In

34communicative teaching less emphasis is placed on formal accuracy than it used to be,since more importance is “given to the question of communicative effectiveness”(Allwright & Bailey, 1991, p. 84). Teachers are more focused on the learners’ ability toexpress themselves and convey their ideas than on their ability to form grammaticallycorrect utterances. Thus the more commonly asked question is whether somethingworks in the communicative sense rather than if it is grammatically correct. It isimportant to stress though that error has not become less important– it is just vieweddifferently and perhaps in a more complex way. Teachers, as well as learners, are askedto consider parts of utterances in more ways than just by considering grammaticalaccuracy. What is or is not an error has become a more complex decision and placesgreater demands on teachers in their decision making regarding error correction.Nevertheless, it is crucial to remember that in general, communicative effectiveness isbeing valued over grammatical accuracy and that errors do not need to be somethingnegative at all.In his book Mistakes and correction (1989), Julian Edge suggests dropping theterm errors and replacing it with learning steps (p. 14). In doing so, we will get a newpoint of view and start seeing student’s language as “something developing in thestudent” (p. 15). The role of teachers is not to compare the students’ language withStandard English and point out differences between the two; they should ratherencourage the growth and development of the language by appreciating the learningsteps. Teachers should thus not require error-free learning: if they insist on accuracyonly, students will not make any errors; if we look at it from the other side, however, wecan say that students will stop taking learning steps. From this point of view, thesituation looks very different – avoiding learning steps is rather negative. Moreover, itwas Corder already in 1984 who suggested that errors are beneficial for everyone within

35the learning process: it is a message for the learner about how far he or she has got inlearning; it is a chance for other students to learn from other people’s errors; and finally,it carries information for the teacher to see how effective her teaching methods are andwhether any remedial teaching is necessary. As Edge (1989) concludes, “the importance[of errors] is that they should often be ignored” (p. 20). Not only is it necessary to viewlearners as people with things to say, who are being listened to, but errors can beregarded as learning steps and be overall beneficial for everyone involved in the processof learning.3.3.1. INTERLANGUAGE AND ERRORSThe language of language learners is frequently referred to as interlanguage – aterm coined by Selinker in 1972. It is marked by several stages of development throughwhich learners have to go through the process towards language proficiency. AsAllwright and Bailey (1991) observe, each stage of this process is characterised byerrors learners typically make at the given stage. The development through these stagesis not linear, but rather fluctuating. Learning an L2 takes time and learners’ performancefluctuates every day. Bartram and Walton (1991) stress that it is “wrong to correct astudent on Wednesday for what they were able to get right on Tuesday” (p. 18). It is atypical characteristic of learners to have seemingly forgotten what they already knew afew days before. The important point here is, however, that interlanguage consists ofthose stages that frequently overlap, and thus learners tend to make errors even if we donot expect them to do so anymore. It is then necessary to remember to view errors aslearning steps from one developmental stage to another and not to hinder learning bytoo much criticism.

36In order to provide students with sufficient space for progressing through theindividual stages, teachers should allow their students to make guesses, experiment andbe creative in using the language. As Bartram and Walton (1991) claim, students “haveto have the opportunity to make [errors]” (p. 14). Only then will students be able tomake hypotheses about language and not be limited to copying what they have heardsomeone else say (often that implies the teacher only). As a result, students will makemore progress, as they will be able to create their own sentences.The idea that errors are a natural part of language learning should by all meansbe transmitted to students. As Edge (1989) observes, it is, however, necessary for theteacher to believe the same – otherwise the students will not believe it either. Onlywhen students will not be interrupted after each error will they start believing that errorsare perfectly normal. Otherwise they might stop trying to use language creatively if theysee they are being criticised for it. As Bartram and Walton (1991) stress, it might beuseful for teachers to praise the good, successful tries, rather than criticise errors. Byconcentrating on what the students manage right, teachers can encourage their languagelearning progress. Since learners usually do not make errors on purpose, it is indeedmore beneficial to aid the learning process by praising the success and leaving enoughspace for learners trying out new pieces of language.Finally, there is one more aspect of error in SLL that I would like to bringforward. I believe it is an important factor one needs to consider when deciding onerrors that in cases when learners produce erroneous utterances by breaking some rulesthey are not familiar with yet, we cannot call this an error. If learners do not know that acertain rule exists in language, how can we punish them for not applying it? As Corder(1984) points out, “While it is true that [erroneous utterances] cannot be accounted forby the rules of the target dialect, they are in fact grammatical in terms of the learner’s

37language” (p. 163). Corder takes this even further when he compares SLL with learningone’s L1 and claims that we should not call learners’ utterances deviant in general, sincewe do not call children’s utterances erroneous either. Indeed calling them deviant orerroneous evokes intentional breaching of rules, which is not generally the case. For thesame reason, I believe it is beneficial to think of errors as the above mentioned learningsteps, as this notion avoids seeing errors as something undesirable. On the other hand,for educational purposes it seems to me practical to call utterances erroneous, despiteCorder’s claims, in the cases when learners should already be familiar with certainrules, but their performance fails (although here we need to bear in mind indeed thatinterlanguage does not develop linearly as discussed above). Only if learners cannotcorrect themselves precisely because they operate within the only rules known to them,we should not call this an error and take it into account when considering errorcorrection.This section has introduced error in SLL. It has discussed the problems ofcomparing learners’ utterances with the so-called native speaker norm. It has beenshown that errors can be very beneficial to SLL – not only to the learners themselves,but to other people, too. For this reason, it is useful to think of errors as learning stepsthrough which learners advance from one stage of their interlanguage into another.Since one’s interlanguage is fluctuating and constantly developing, teachers shouldleave enough space for their learners to experiment and be creative with language. Onlythen will they be able to take learning steps and make progress. Finally, it was pointedout that we should not criticise learners for breaking rules they are not familiar with yet,and always first consider whether correction is in place at all.

384. ERROR AND MISTAKE IN SECOND LANGUAGELEARNINGThe previous chapters have introduced the notion of error in SLL compared toerror in children’s mother tongue learning and native conversations. It was observedthat although there are some unified aspects, many are radically different – particularlythe way errors are perceived and dealt with. This chapter will deal with error andmistake in SLL in much more detail. Error indeed plays an important role in SLL, as itis an inseparable part of learners’ production. As MacDonald Lightbound (2010)observes, “comprehension is, to most learners, an easier task due to the way theincoming information is processed, […] whereas production seems to involve greaterprocessing constraints” (p. 97). It is thus not surprising that dealing with error, whichcomes about as a result of the constraints, is an important part of teachers’ work. Thischapter will try to define error and contrast it with the notion of correctness oraccuracy. Furthermore, I will distinguish error from mistake and discuss severalproblems regarding this distinction. I will also introduce possible sources of error as themost prominent linguistic theorists approach it. Finally, a section will also be devoted todifferent typologies of error and mistake, which will be analysed and compared to eachother.

394.1. DEFINITION OF ERROR4.1.1. BACKGROUNDThe view of error in SLL has been shifting over time. In early psychologicalresearch, error in learning was considered as a factor signalling lower quality oflearning. Traditionally, pedagogy has seen error as a measure of the extent of failureand as a criterion for evaluating students. As Hendrich et al. (1988) observe, forlearners, the notion of error was indeed connected to a failure resulting in a series ofundesirable consequences. Today, the perspective has changed. Research into error is avaluable source of the process of learning. As Hendrich et al. (1988) further note, itallows the teacher to step into the individual phases of the learning process and toregulate or direct it. Additionally, errors serve as good feedback for teachers onefficiency of the chosen teaching methods, or perhaps on their methodologicalimperfections, too.When talking about errors made by learners of an L2 we are traditionallycomparing their utterances with the already mentioned native speaker norm. Thiscomparison is problematic especially due to one aspect: as Corder (1981) states, “whenwe talk about errors made by learners we are already applying to their languageintuitions about grammaticality possessed by speakers of the target language” (p. 57).For one thing, there may be good reasons in doing so: after all, the learners aresupposed to be acquiring the rules of the L2. This theoretical model of native-speakercompetence is, however, idealised. Research into native speakers’ language use hasshown that they often make a lot of errors of all kinds. As MacDonald Lightbound(2005) lists, the following distribution of native speaker errors is generally found:o 15% - grammatical errorso 12% - lexical errors

40o 33% - errors of styleo 40% - errors of spellingThis fact makes the native speaker model problematic and calls into question the nativespeaker norm. It would thus be appropriate now to ask ourselves what an error is and,alongside with that what accurate forms are.4.1.2. ACCURACY AND CORRECTNESSLet us consider the question of accuracy or correctness first. In some cases, it isperfectly clear to decide what is correct:o your not *yuoro she is not *she areo a blue car not *a car blueIn other cases, however, there may already be some arguments regarding correctness. Inthe following examples, some forms could be referred to as variants rather than errors:1. He ain’t there.2. She come.3. They spent a fortnight away.4. I am waiting here for ages.5. The woman which …(Adapted from Ur, 2012)The question for us as teachers is whether we should accept such forms, or rathercorrect them? Jenkins (2006) claims that this issue is becoming a political one: “withforms like she goes, the people who and so on being identified with a ‘native speaker’model and native speaker hegemony” (p. 43). When forms of Standard English areimposed on L2 speakers, Jenkins calls this ‘native speakerism’, or ‘linguistic

41imperialism’. She believes that the future of English will be dominated and led by nonnativespeakers, who vastly outnumber the natives already. In this case, however, Iagree with conclusions of Penny Ur who claims that “We should continue to teachconventional standard grammar and vocabulary as a basis for our students’ ownproduction” (Ur, 2012). However, learners should also be aware of the existence ofvariants in speech and writing and should respect them. In general, we should thus teachthose forms that are actively accepted worldwide rather than local varieties. Studentsshould be aware that the latter ones do exist, but should not be encouraged to use themin their own production. Regarding correction, we should certainly keep correcting nonstandardforms in students’ performance and only forms of Standard English should beaccepted as the accurate ones.4.1.3. DEFINITION OF ERRORCountless language-teaching theorists have come up with various definitions oferror, always referring to different aspects concerning the error, namely native speakers’competence, target language norms, intentions of the learner and many others. In somecases, the definitions were not readily accepted, as they were referring to idealised,vague models the learners were supposed to imitate, an example being Liski andPutnanen’s definition from 1983: “An error occurs where the speaker fails to follow thepattern or manner of the speech of educated people in English speaking countriestoday” (cited in MacDonald Lightbound, 2005, p. 77; my italics). Other linguists, suchas Hendrickson (1978), define error with reference to error correction and teachers:“an utterance, form or structure that a particular language teacher deems unacceptablebecause of its inappropriate use or its absence in real- life discourse” (p. 387). The mainflaw of this type of definition lies in the underlying fact that the teacher is considered

42the “ultimate decider of what is right or wrong, and thus implies a high degree ofsubjectivity” (MacDonald Lightbound, 2005, p. 77). A more objective and unbiaseddefinition of error is thus needed.In the most general terms, language error can be described as an “unsuccessfulbit of language” (James, 1998, p. 1). Such definition is, however, too broad, and thusinsufficient for our purposes. Hendrich et al. (1988) speak about error as a sort ofdeviation from a prescribed performance norm or a solution leading to a given aim.Such description could indeed be attributed to error in all fields of human activity;although it is more specific than the previous one, it still does not tell much about SLL.When defining error in SLL, Carl James (1998) stresses that an error arises “only whenthere was no intention to commit one” (p. 77). If, on the other hand, any deviance wasincorporated to an utterance intentionally, we would not say it is erroneous, but ratherdeviant; examples of such use are poetic language or jingles in advertising.Additionally, as Shastri (2010) points out, it is crucial to stress that an error is notcorrigible by the learner him/herself and suggests certain lack of linguistic competence.Following James’s and Shastri’s observations, I would thus define error here as:A form or structure in the learner’s utterance which is unintentionally deviantfrom the system of language and is not self-corrigible by the learner.This definition is bound to SLL only and is suitable enough to cover all aspects of theproblem. It is better than the previously mentioned definitions for two reasons inparticular: neither does it refer to any expectations of the teacher; nor it tries to assumewhat the learner’s intentions were and what s/he wanted to communicate. It is necessarythough to explain what is meant by deviant. At this point it is impossible to avoidcertain comparisons to native speakers of the L2 or the system of the L2 in general. In

43order to avoid referring to any native-speaker norms, I will adopt P. Lennon’s (1991)definition of deviant utterances as:Linguistic form […] which, in the same context […] would in all likelihood notbe produced by the learner’s native speaker counterparts. (p. 64)By context Lennon refers to the linguistic and the communicative context of anutterance (MacDonald Lightbound, 2005). Together with the above stated definition,this settles the notion of error clear enough for the purposes of this work.Finally, it must be pointed out that there are several types of linguistic deviance.As Littlewood (1994) observes, “an utterance may be formally correct but fail to conveythe intended meaning” (p. 90). The same could be said the other way round: anutterance may be communicatively successful while being formally incorrect. This factthus needs to be born in mind when speaking of linguistic deviance. James (1998)comes up with four categories of this – grammaticality, acceptability, correctness, andstrangeness and infelicity:o Grammaticality – the well-formedness of an utterance; it is context-free and thusobjective. It is grammar itself that decides whether an utterance is grammaticalor not.o Acceptability – a practical notion that comes into play when non-linguisticfactors violate the use of a form. It is merely subjective as the users of thelanguage themselves decide on it and needs to be contextualized. The problemof this aspect is the question whether acceptability is in the eye of the beholderor not. Besides, it is speculative whether it takes a native speaker to decide whatis acceptable or not.o Correctness – the most problematic aspect. It includes recourse to normativestandards and reflection about learnt cannons. Decisions about correctness are

44metalinguistic in its nature. Something can be acceptable, although not correct.Different demands regarding correctness are placed upon native speakers andlanguage learners: although an incorrect piece of language may be perfectlyacceptable from a native speaker, it would not be the case with languagelearners.o Strangeness and infelicity – linguistically strange combinations. As Jamesobserves, it is typically poets who make use of this aspect of language in theirworks.(Adapted from James, 1998, pp. 68-75)When deciding on deviance of a piece of utterance, it is thus important to consider allthe above mentioned categories. Grammatical deviance is usually the most obvious one;teachers should not, however, forget about the other three aspects either, as an utterancemay be deviant in any of the four ways.This section has discussed the notion of error in SLA in more detail. It hasdiscussed the problems of the so-called native-speaker norm, which is in fact idealisedand rather vague. Furthermore, I have shown several problems of deciding on accuracyand correctness; I have also claimed that teachers should keep correcting non-standardforms, although learners should be aware that in some parts of the world such formswould be accepted. Additionally, I have presented several difficulties of defining errorin SLL and have come up with a definition I regard as an appropriate one. Finally, itwas observed that deviancy of learners’ utterances must be judged from four differentperspectives, not only the most common (grammatical) one.Having set up a firm definition of error in language learning, it is now necessaryto contrast it with a term frequently regarded as synonymous: mistake.

454.2. ERROR VERSUS MISTAKEIt is a common practice to regard error and mistake as synonymous terms. SLLtheory, however, distinguishes between the two terms, although such distinctions arenot identical in all books. In this section, I will thus present definitions of error andmistake in several prominent dictionaries of English language and show how literatureon SLL deals with this problem. I will also provide a clear distinction between the twoterms, which is used in this paper.Dictionaries seem to regard mistake as a more general term than error.Cambridge Online Dictionary of British English provides us with a simple definition oferror: “error: a mistake” (Cambridge Dictionary Online). Mistake, on the other hand, isexplained in more detail: “an action, decision or judgment which produces an unwantedor unintentional result” (Cambridge Dictionary Online). Exactly the same situationoccurs in the Oxford Dictionary – error is simply defined as a mistake. A mistake isagain given more space and is defined as “an act or judgement that is misguided orwrong“(Oxford Online Dictionary). Not surprisingly, Macmillan Dictionary follows thesame pattern and only adds that error is a “mistake, for example in a calculation or adecision” (Macmillan Dictionary). Mistake is then defined as “something that you havenot done correctly, or something you say or think that is not correct” (MacmillanDictionary). From these definitions it is clear that these dictionaries regard error andmistake as synonymous. The only aspect in which they differ is the precise wording ofindividual definitions of mistake. At its core, however, all are more or less the same andtoo vague for our purposes. It is thus necessary to turn to literature on SLA.It is a commonly acknowledged fact that it was S. P. Corder (1984) who firstintroduced the error vs. mistake distinction. In his paper The significance of learners’errors he puts down the distinction as follows:

46It will be useful therefore hereafter to refer to errors of performance as mistakes,reserving the term error to refer to the systematic errors of the learner fromwhich we are able to reconstruct his knowledge of the language to date, i.e. histransitional competence. (p. 25)He also adds that the difference between the two lies in the fact that mistakes are of nosignificance to the language learning process. This distinction between the two is themost commonly used one: Allwright & Bailey (1991), Hubbard et al. (1983), orBartram & Walton (1991) all adopt Corder’s distinction. We thus refer to errors in casethat learners lack knowledge about a certain aspect of the target language or by creatingwrong hypotheses about it; when they try out something they have not learned yet andget it wrong then we speak about errors. Mistake, on the contrary, are caused byconfusions, temporary lapses of memory, slips of the tongue etc. Thus when learners donot put into practice something they have already learned, mistakes occur.In the distinction between errors and mistakes it is vital not to forget about thecriterion of correction, which makes a difference, too. As Allwright & Bailey (1991)point out, “L2 learners can often correct their own mistakes, but the errors they make …are part of their current system of interlanguage rules and hence are not recognisable (tothe learners themselves) as wrong” (p. 92). If the learner is inclined or able to selfcorrect,we assume that the deviant utterance is a mistake. When, however, the learner isnot able or not inclined to perform self-correction, we shall assume that it is an error.The same criterion of correctability is applied by Carl James (1998) and Hubbard et al.(1983). As James (1998) observes, the test of auto-correctability is, however, ratherproblematic to apply in real situations. Determining what is a mistake or an error is aproblem of some difficulty and as Corder (1984) noted, it requires a “much moresophisticated study and analysis of errors than is usually accorded them” (p. 25). It is

47indeed true that teachers cannot carry out any extensive analyses of the students’deviant utterances in their lessons. Although we can never be absolutely sure, the autocorrectabilitytest can tell us to a certain extent whether we are dealing with an error or amistake.Error and mistake are frequently distinguished on the basis of systematicness,too: when there is a random occurrence of a deviance only, we call it a mistake; when itis of a more systematic nature, then we regard it as an error (MacDonald Lightbound,2010). This distinction, however, seems to be rather unclear. Already more than fiftyyears ago, in the sphere of philosophy, Wittgenstein (1958) argued that there is no sharpline between a random mistake and a systematic error. In the case of SLL, it is ratherdifficult to distinguish the nature of a deviance. According to MacDonald Lightbound(2010),The mere repetition of a particular deviant form does not always mean that it is asystematic error in the learner’s IL – it indicates that under certain conditions,where students have to comprehend and produce language at high speed, theymay be paying more attention to meaning than to form (p. 81).Under different conditions, learners will behave differently and apply differentstrategies; as a result, they might be making errors more or less systematically. Duringan exam, students will try hard not to make any errors; under other conditions, however,such as in a conversation with foreigners outside classroom, they might be makingmuch more errors: in this case, “errors may be due more to the mode of communication(e.g. casual conversation) than to any deficiency in their linguistic competence”(MacDonald Lightbound, 2005, p. 81). The question of systematicness is thus ratherproblematic, too. To a certain extent, however, it can signal whether we are dealing withan error or a mistake, and thus may be useful after all. When we also consider whether

48we are dealing with something unknown to learners, or just temporary lapses ofmemory and putting into practice something learned, we should be more or less able todecide whether we are dealing with errors or mistakes. This knowledge will then becrucial for deciding on correction and dealing with the deviant utterance as a whole.4.3. SOURCES OF ERRORSAn essential part of the study of errors is searching for their possible origins. Asthe American psychologist F. Barron (1957) states, it is not sufficient to only ask howmany errors a learner makes, but also what types of error and why. Harmer (2007) statesthat by trying to work out why something has gone wrong, “students learn more aboutthe language they are studying” (p. 96). Realizing sources of error can thus be a richsource of knowledge not only for teachers, but the learners themselves as well. Thissection is going to introduce the notion of reasons of error in SLL and present severaldistinctions of the sources made by prominent language theorists.Although the question of sources of error, as Choděra (2006) observes, belongsto the study of psychology, it has mainly been subject to linguistic theory of SLL.Different schools of thought assigned different sources to learner errors. Behaviourismwholly ascribed errors to human fallibility: “there has not been enough effort on the partof the learner or enough explanation or practice on the part of the teacher” (Corder,1981, p. 65). Contrastive analysis, on the other hand, saw errors as a result of theinfluence of one’s mother tongue on the process of learning an L2 – L1 interference (ortransfer). Broughton et al. (2003), however, state that recent experimental evidenceshows that “only a minority of errors are attributable to mother tongue interference” (p.136). This finding, together with the insight that errors are a natural part of the learning

49process – as promoted by the communicative approach today – are very important indealing with sources of error.It is needless to say that no human learning is perfect. Mostly, humans learn bytrial and error – and SLA is not an exception. The extent to which people experimentwith language and to which they are likely to make errors is rather individual. AsRichards (1984) claims, the individual limitations in the process of learning reflect“personal differences in motivation, perseverance, aptitude and so on” (p. 88). Since thecircumstances and conditions for individual L2 learning are never identical, theacquisition of new items varies from person to person, and so does the type and amountof errors. Due to these different contexts of individual learning, sources of errors vary,too. Although the reasons might be innumerable, we will limit ourselves to the mostcommon ones and the traceable ones – such as those resulting from differences betweenone’s L1 and L2 or those attributable to the process of learning.From the perspective of learners, sources of error can be classified into internaland external. Faerch et al. (1984) explain the internal factors as “motivation of thelearners, the capacity they have for studying, and their attitude to the target languageand culture” (cited in MacDonald Lightbound, 2005, p. 98). The external sourcesinclude teaching materials or teaching methods (such as causing confusion by providingtoo much information). This distinction, however, is not a generally adopted one; amore common classification is based on the languages acquired and the process oflanguage learning.Several authors have created different classifications of reasons causing errorsconcerning one’s L1 and the L2; however, on the most general level, they agree on twomain sources: interlingual and intralingual. The former are those that “may be tracedback to the learner’s knowledge of the L1 or another language previously acquired”

50(MacDonald Lightbound, 2005, p. 99). These are what CA called interference errors.The latter ones are those that are caused by incorrect applications of the rules of the L2,which then causes conflicts and “‘non-target-language-like’ behaviour” (MacDonaldLightbound, 2005, p. 99). Apart from these, P. D. Shastri (2010) lists fossilization asanother source of error. It occurs when some errors remain within the language systemof proficient learners, who then become unable to correct them. As she observes, theseerrors “can be due to the internalisation of the wrong elements of the language. […]They become part of the personality and cannot be corrected” (Shastri, 2010, p. 30).Personally, I do not believe that fossilization as such could be regarded as a source oferror. More likely, it is a general problem of majority of proficient learners, which ischaracterised by certain features in their language use. Since Shastri herself states thatfossilization can happen due to internalisations of wrong elements, it seems wrong tostate that fossilization itself is a source, too. I will thus stick to the classification thatcounts with intralingual and interlingual sources and leave fossilization aside.Another, more detailed classification of sources comes from Julian Edge (1989),who sees four different types of sources: firstly, the influence of the speaker’s firstlanguage (L1 interference); secondly, when learners assume having mastered a rule, butin fact they do not know enough yet (overgeneralization); thirdly, using knowinglyincorrect forms in order to be able to get a message across; and finally, due to being in ahurry, tired, or not concentrated (these are, however, sources of what we call mistakes,not errors). Edge’s classification does not consist of any specific categories; rather helists the four above mentioned causes without giving them any names. Out of these onlythe first three are relevant to us, since the fourth one describes what I have calledmistakes. The third one, which does not carry any title, could be called a

51communication-strategy source, as the speaker makes an error as a part of a strategy tocommunicate.A similar classification is presented by Hubbard et al. (1983).The authors statethat they have adapted S. P. Corder’s classification of sources of error; what theyactually did was using the same categories with different labels. Thus their categoriesare mother-tongue interference (originally called transfer by Corder),overgeneralization (originally analogical source), and errors induced by teachingmaterial or method (originally teacher-induced). Beside these three sources, Hubbard etal. point out that teachers may induce errors of overgeneralization by providingmisleading information. This may happen in cases when teachers state that 3rdpersonsingular always takes an –s ending and after some time introduce the verb can.Compared to Julian Edge’s classification, this comes up with errors induced by teachingmaterial, method, or teachers themselves, which is certainly a frequent source of errors,too. On the other hand, they leave out the big group of communication-strategy errors,which are indeed a frequent source, too. This classification is thus not ideal either,although it introduces another source not mentioned in any of the previousclassifications.Probably the best classification, and at the same time the most complex one, isone introduced by Carl James (1998). When considering causality of language learningerrors, he comes up with six possible reasons of errors that learners make:1. Ignorance and avoidance – lack of knowledge leading to using an L1 item andthus creating L1 transfer or interference; avoidance happens when learners donot know the L1 and L2 items and have to avoid them somehow;2. L1 influence – the interlingual errors caused by L1 interference;

523. L2 causes – the already mentioned intralingual errors – happen when learnersare ignorant of an L2 item, an example can be overgeneralization of a rule;4. Communication strategy-based errors – using near-equivalent forms, synonyms,superordinate terms, or antonyms instead of the unknown target language form5. Induced errors – result from classroom situations, learners get misled by the wayteachers give examples, definitions, explanations; there are also teacher-inducedor exercise-based induced errors;6. Compound and ambiguous errors – errors ascribable to more than one cause,which are complementary; ambiguous errors – happen when there are twocompeting diagnoses of errors and we cannot be sure which source is the trueone.This classification is an exhaustive one and as such conceives the variability of sourcesof error the best. One should always be able to classify errors made by learners to one ofthe above mentioned categories, as they embrace all those listed in the incompleteclassifications mentioned earlier.4.3.1. CONTROLING SOURCES OF ERRORSThe question for teachers is whether they can prevent some type of errors fromhappening. Theoretically, the only source of error that could be controlled is errorsinduced by teaching materials or methods. According to Corder (1981), if we “attemptto teach [a learner] something before he is ready for it, the result will be confusion, falsehypotheses, and what we could call redundant errors” (p. 58). In order to decrease theamount of induced errors, teachers should only teach what the interlanguage of theirlearners permits them to learn at any specific moment. Meyer (1986) observes thatstudents make the most errors when “teachers present new or difficult material” (p.

53230). What is therefore needed is a “sequence of progressive complication” (p. 77)which shall make sure that the number of material-induced errors will be limited tominimum. It is remarkable that already in 1981 Corder expressed a core idea of CLT,when he claimed a need for grading of demands on learners rather than of the teachingmaterial itself:If we can control the level of these [more complex communicative tasks]correctly, the grammar will look after itself. Instead, then, of grading thelinguistic material that we expose the learner to, we should consider grading thecommunicative demands we make on him. (p. 78)This strategy is widely promoted by the CLT today. In order to bring down possibleerrors induced by teaching material, it is thus vital to grade the communicative demandsteachers place on their students. Nevertheless, this might only solve the case of inducederrors, and the other sources remain unchanged anyway. Although it is thus impossiblefor teachers to prevent most errors from happening, since the sources cannot beeliminated, it is important for them to be aware of the various sources, since “the reasonwhy [an error] is made influences the way we correct it” (Spratt et al., 2010, p. 143).Moreover, reasons also decide whether to correct or not at all, considering thedifference between errors and mistakes. This issue will, however, be dealt with in moredetail in the section on error correction.While various typologies of sources of error provide linguists and teachers witha lot of descriptive information on the categories, there has been little agreementregarding psychological background behind the individual types of sources. AsMacDonald Lightbound (2010) argues, there is still little known about the weight ofeach of these main sources. Although on the most general level it is commonly observedthat L1 interference is not as fatal as CA thought it to be, linguists do not agree on the

54actual weight that should be assigned to it or other sources. Individual studies, such asMacDonald Lightbound (2010), Dulay & Burt (1974), or Nickel (1981) bringcompletely different results regarding the size of the share of interlingual errors on thetotal number of errors. Additionally, the question about the nature of the mental andlinguistic processes employed by learners when producing errors still remainsunresolved. Christison & Krahnke (1983) observed that “Transfer and interference areterms that are more metaphoric than empirical” (p. 634), since very little is actuallyknown about processes behind them. These aspects of sources of error are only poorlyunderstood even today and might thus be subject to more scientific research in future.This section has presented various typologies of sources of error and pointed outthe importance of being aware of the sources. The individual typologies were subjectedto comparison and evaluation; the classification by Carl James was found to be the mostsuitable and comprehensive one. Additionally, it has also considered the possibility ofreducing potential sources, which turned out to be rather minimal – only induced errorsmight be eliminated to a certain extent by teachers; the rest needs to be dealt withappropriately through error treatment. It has also been argued that some aspects ofsources of errors, such as the nature of psychological processes behind them, are stillonly poorly known today and that more research will be needed to understand thembetter.4.4. TYPES OF ERRORAs Dulay, Burt and Krashen (1982) observe, an accurate description of errors iscompletely different “from the task of inferring the sources of those errors” (cited in

55Yang, 2010, p. 2). One reason for this is surely the differing seriousness of variouserrors: some of them completely impede easy comprehension and are thus more serious;others may cause only minor difficulties in comprehending one’s message, and theirimpact is thus not as grave as in the first case. When considering different types of errorand mistake, it is thus necessary to bear in mind the degree to which they transgresslinguistic rules, as well as their social tolerability.4.4.1. GENERAL CLASSIFICATIONSConsidering seriousness of error, on the most general level, we distinguishbetween global and local errors. The distinction is one of the amounts of affectednesswithin a sentence or an utterance: global errors affect the whole sentence organization,while local ones only single elements in a sentence. As Burt and Kiparsky (1974)explain, local errors “are to be found within a limited linguistic domain – i.e. involvingthe word or clause and affecting specific items such as articles, verb and nounagreement” (cited in MacDonald Lightbound, 2005, p. 85). Global errors, on the otherhand, “tend to be located within the relations between clauses, or sentences, or overlonger stretches of discourse” (MacDonald Lightbound, 2005, p. 85). Typically, theseinclude word order, wrongly placed sentence connectors, tenses, relative pronouns, etc.Burt (1975) stated that teachers should concentrate on global rather than local errors.These, rather than simple misplaced articles or spelling errors, should then more likelybe subject to error correction.Another general classification of errors comes from S. P. Corder (1981), whodistinguishes between overt and covert error. This distinction deals with superficialcorrectness and erroneousness, as Corder explains:

56Superficial well-formedness is not guarantee of freedom from error. It is for thisreason that we have to distinguish between sentences which are overtlyerroneous, i.e. are superficially erroneous, and those which are covertlyerroneous, i.e. apparently acceptable, but so by chance, or which areinappropriate in one way or another. (Corder, 1981, p. 42)Thus a covert error occurs when a learner’s utterance is superficially correct, but isnevertheless erroneous. Bartram and Walton (1991) describe it as situations whenlearners say “something right by accident” (p. 21). An overt error, on the contrary,appears in cases of superficially ill-formed utterances in case of the target-languagerules, that is, when an utterance is clearly erroneous. Regarding seriousness of the twotypes, Bartram and Walton (1991) believe that covert errors are as serious or asunimportant as any other kind of error: “if they block communication, they do matter,but otherwise perhaps not” (p. 39). The same argument of the level ofcomprehensiveness and successful communication, as with global and local errors, isapplied here, too.4.4.2. ERROR TYPOLOGIESSeveral linguists, such as Harmer or Edge, see mistake as the superordinate termof all kinds of deviant utterances of learners. Harmer (2007) distinguishes between threetypes of mistakes: slips, errors and attempts. Slips are what we call mistakes here:students can correct themselves after the mistake has been pointed out to them. Errorsare those types of mistakes that need explanation, as learners are unable to self-correct.Finally, attempts are mistakes that “students make when they try to say something butdo not yet know how to say it” (Harmer, 2007, p. 96). Harmer probably took thisclassification over from Edge (1989), who lists exactly the same categories, although he

57does not mention it anywhere. Apart from these, Harmer adds one more type ofmistakes, which somehow stands aside and does not belong under any of the abovementioned categories, but at the same type is not listed as a subcategory of mistakeseither. This category is called developmental errors, which according to Harmer “occurnaturally as the students’ language knowledge develops, and are the result of thestudents making apparently sensible (but mistaken) assumptions about the waylanguage works” (Harmer, 2007, p. 96). This type of error, however, considers thesource of error rather than its nature or linguistic character. It should most probably notbe mentioned there at all, as Harmer seems to be comparing incomparable, when headds a category based on a source of error to three types of mistakes based on thelearners’ ability to self-correct.Another, even simpler division with mistakes as the hyperordinate comes fromSpratt et al. (2010), who divide mistakes into errors and slips only. Here slips are againa result of tiredness or temporary memory lapse, what we here call mistakes in general.Errors then occur when learners try to say something beyond their current level oflanguage knowledge (Spratt et al, 2010, p. 44). These correspond to what Harmer andEdge called attempts. What is different though is that the authors claim that learners areable to correct both types of mistakes, “once they realise they have made them” (p. 44).This assumption seems to be, however, incorrect, as learners could hardly correctsomething that is beyond their current knowledge. Moreover, since Spratt et al. list onlythe two above mentioned types of mistakes and see both of them as self-corrigible, thenthey must be assuming that learners are able to self-correct all kinds of deviances theymake: and that is certainly not the case. This classification by Spratt et al. (2010) thusappears to be rather inaccurate and incomplete.

58Bartram and Walton (1991) are using two terms only: mistakes and slips. Slipskeep the same definition as in the previous cases; mistakes are seen as “wrong languagewhich a native speaker would not usually produce” (p. 21). That is, mistakes aresomething that only learners of the L2 produce. Although this division is ratherstraightforward and clear, it is rather insufficient for our purposes: it does not take intoconsideration any aspects of self-corrigibility, linguistic nature of these, or seriousnessof the error. Thus a more detailed and precise classification is needed.Carl James (1998) uses four-way taxonomy of errors, where error is thehyperordinate term for all types of linguistic deviations, and at the same time acategory:o Slips – or lapses, can be detected and self-correctedo Mistakes – can be corrected only if their deviance is pointed out to the learner. Ifa simple indication that there is some deviance is a sufficient prompt for selfcorrection,then we have a first-order mistake. If additional information isneeded, in the form of the exact location and some hint as to the nature of thedeviance, then we have a second-order mistake.o Errors – cannot be self-corrected until further relevant input has been providedand converted into intake by the learner. In other words, errors require furtherrelevant learning to take place before they can be self-corrected.o Solecisms – breaches of the rules of correctness as lay down by purists andusually taught in schools: split infinitive and dangling participles, for example.Although it might be slightly misleading at first sight to use the term error for both atype of category and a global term for all deviations, this classification seems to be themost complete one and the clearest one, too. It does not leave anything out, nor does it

59simplify the possible types of occurrences of errors. It also takes into consideration thein/ability of the learner to self-correct and the amount of correction needed.4.4.3. WAYS OF CATEGORIZING ERRORSDulay, Burt and Krashen (1982) claim that there are two basic ways ofcategorising learners’ errors and that is into either linguistics categories (morphology,phonology, etc.) or surface structure taxonomies (errors of addition, omission, etc.).Corder (1981) argues that the most frequent classification of errors lists errors ofomission (where some element is omitted that should actually be present); errors ofaddition (where some element is present though it should not be there); errors ofselection (where a wrong form was selected in place of the right one); and errors ofordering (where the items presented are selected correctly, but placed in a wrong order).According to Corder, however, such classification is too superficial and thus a moreadequate one “in terms of systems, such as tense, number, mood, gender, case, and soon” (p. 37) is needed. This belief is in agreement with what Christison & Krahnke(1983) observe when they state that a classification based on linguistic types “is a morefamiliar way of categorizing error and is possibly of more immediate use to classroomteachers” (p. 635) and thus a more practical one is needed. In this approach, sources oferror are ignored in favour of “the part of the linguistic system which is ill-formed”(Christison & Krahnke, 1983, p. 635); examples can be article omission, third personsingular verb endings, etc. The authors also explain when such analysis can be to a gooduse:Classification by linguistic type can be a useful analytic procedure and canprovide a useful basis for instructional intervention as long as the classificationis not mistaken for a psychologically real analysis of the process by which the

60errors are produced or for a hierarchy of the communicative effect of errors.(p. 635)When such conditions are met, then linguistic classifications can be useful not only tolinguists, but more importantly to classroom practice, too.Hubbard et al. (1983) come up with a simple linguistic classification of errors,where they list only two categories – grammatical (She told [them] she was on holiday)and lexical (She told said she was on holiday). Within the lexical category, there are twotypes: errors in the chain of language (To school should have gone Maria) and errors inchoice of individual words or word forms (Mary is knowing the answer). Out of these,as Yang (2010) points out, the lexical errors are considered by native speakers as moredisruptive and irritating than the grammatical ones. Others, such as Politzer & Ramírez(1973) include categories of morphology, syntax, and vocabulary, while Green & Hecht(1985) list grammar, vocabulary, spelling and style as categories of errors.Choděra (2006) lists four categories from the linguistic aspect, too, although hementions slightly different ones: grammatical, lexical, spelling and phonetic errors. It isinteresting to note that all these classifications agree on the group of lexical errors(errors of vocabulary) and otherwise employ differing angles of view and thus differentcategories. Based on a corpus of spoken language, Chun et al. (1982) distinguishedfive categories:o Discourse errors (errors beyond the sentence level and includingstructural/pragmatic aspects including inappropriate openings and closings,incorrect topic switches, inappropriate refusals, etc.)o Factual errors (including those concerning the factual knowledge or truth valueof an utterance)

61o Word choice errors (incorrect choice or addition of a noun, verb, adjective,adverb, preposition, etc.)o Syntactic errors (tense agreement, morphology, word order, etc.)o Omissions (involving the incorrect omission of nouns, verbs, auxiliaries,articles, etc.)(cited in MacDonald Lightbound, 2005, p. 88)This classification seems to be, as MacDonald Lightbound (2010) observes, a hybridone, as it is not built up around a single classifying aspect, but rather combines severalinto a mixed classification (i.e. is not purely linguistic). There are, however, quite a fewsuch classifications that do not consider one single factor only, but rather mix more ofthem together – Chun et al. are not at all an exception.A detailed classification comes from Dulay, Burt and Krashen (1982) whoproposed to classify errors based on how they differ from the L2 structures. Similarly toCorder (1981), they state the categories of omission, addition, misordering, andmisformation (Corder calls it errors of selection). Besides these, the authors identifyblends (also called contamination, or cross-association or hybridization errors) whichoccur when two alternative grammatical forms are combined to produce anungrammatical blend, i.e. *according to her opinion. This type of taxonomy is asurface-structure one; as such, it is more useful to immediate use to teachers, since suchcategories are much easier to recognize (James 1998). Although, as Corder (1981)claims, it might be less academic. On the other hand, it is not a hybrid classification andseems to cover the whole range of possible deviations; as such, it could turn out to be ofa good use to classroom practice. James himself uses an even more detailedclassification that takes into consideration several factors – but it is far too complicatedto be mentioned here. For reference, see James (1998), chapter 5.

624.4.4. PSEUDO ERRORS AND LEARNING STRATEGIESA type of error that stands slightly aside all the above mentioned classifications iswhat Choděra (2006) calls pseudo errors or quasi errors. He explains it as a case ofavoiding a more adequate form in favour of a less suitable, but familiar one. This oftenhappens when learners experience a language deficit and resort to saying somethingdifferent in order not to say something wrong. These pseudo errors are, however, asChoděra (2006) observes, often rather difficult to recognise, since the same strategy isoften employed in one’s L1, too, and some individuals may not see anything wrongabout it. Corder (1981) mentions the same as Choděra; he, however, regards these aslearning strategies rather than a type of error. Corder lists four different strategies:o topic avoidance – a refusal to enter into or continue a discourse within somefiled or topic because of a feeling of total linguistic inadequacy;o message abandonment – trying but giving up; it is a less extreme form of theformer;o semantic avoidance – saying something slightly different from what youintended but still broadly relevant to the topic of discourse; ando message reduction – saying less, or less precisely what you intended to say;often see as rather vague general talk (Corder, 1981, p. 105).The inclination towards the use of individual strategies largely depends on the type oflearners. As Corder explains, “some are determined risk-takers, others value socialfactors of interaction above the communication of ideas” (p. 106). In general, we may,however, assume that people generally prefer to maintain their intended message andresort to these strategies in case of the utmost necessity. The level of risking and tryingvaries with the speaker’s personality and the speech situation itself.

63Another erroneous behaviour that is typically left out from the taxonomies iswhat Choděra (2006) calls an inadequately long latent time of reaction. Unnatural,slow, and barely fluent speech, although without any errors, may also be considerederroneous. These aspects of learner utterances are, however, generally neglected by bothteachers evaluating learners’ performance, and linguists creating error typologies(Choděra 2006). The author sees this as a serious problem and calls for more attentionon the side of teachers and language theorists.As has been shown above, there is no single agreement on the typology of error;there are several factors that can be taken into consideration or left out, and eachclassification is always slightly different. It is not easy to decide which typology isbetter or worse – unless there is something obviously wrong or missing. It has also beenshown that learners’ linguistic deviancies can be categorised on many levels – on themost general level, such as Corder’s global or local errors, on the level of linguisticcategories (such as errors of grammar and lexis), on the surface-structure level (such aserrors of omission, addition, etc.), or based on several other factors, such as the abilityto self-correct or making errors in language the learners are already familiar with or not.All these classifications complement each other rather than compete. Overall, theycreate a broad picture showing the vast possibilities of errors learners can make. It wasalso observed that it is hardly possible to select a single best typology that wouldencompass the whole problem; rather it is necessary to bear in mind the whole complexproblem and approach it from several aspects. Despite this complexity, taxonomies oferror are of great importance to both teachers and learners in general. According toHarmer (2007), “the way we give feedback and correct such [errors] will be heavilyinfluenced by which type [of error] we think the students are making” (p. 96). When

64dealing with errors and deciding on error treatment, types of error thus play a crucialpart in the teacher’s decisions and the provided feedback.

655. ERROR CORRECTIONAlthough a vast majority of teachers would agree that error correction isnecessary for learners’ development of interlanguage, they would most probablydisagree on many aspects of correction – when to correct, what types of error tocorrect, who should be correcting, what technique to use and how to indicate that anerror has occurred. The process of correction is not straightforward, although it mightseem so at first sight. Teachers have to consider several factors influencing it and decideon many partial steps leading towards correction.5.1. DEFINITION OF ERROR CORRECTIONError treatment is a complex phenomenon and it is important to note that it is notsynonymous to error correction. Before I move to individual types of error correctionand the various decisions teachers have to make before performing any correction, it isnecessary to clearly define what is meant by error correction in general.As Brown (2007) explains, not all error treatment automatically includes errorcorrection: “Error treatment encompasses a wide range of options, one of which – at theextreme end of a continuum – may be considered to be a correction” (p. 348).Correction is thus only a subcategory of treatment of error. Apart from treatment, somelinguists, such as Christison & Krahnke (1983) use the term repair, which they defineas “any conversational move which occurs because there is some real or perceiveddifficulty in the conversation” (p. 633). Like treatment, repair is a more general termthan correction; correction is therefore one type of repair. In this paper, we will dealwith error correction only.

66The term error correction is generally used in SLL in several meanings.Typically, it refers to a corrective activity aimed at elimination of a recognizeddiscrepancy (Kulič, 1971). Jefferson et al. (1977) state that the term usually refers to the“replacement of an 'error' or 'mistake' by what is 'correct'” (p. 363). Carl James (1998)observes that there are three main meanings of the term correction:1. Informing the learners that there is an error and leaving them to discover it andrepair it themselves.2. Providing treatment or information that leads to the revision and correction ofthe specific error without aiming to prevent the same error from recurring later.In addition to indicating that the present attempt is wrong, the corrector canspecify how and where, suggest an alternative, give a hint.3. Providing learners with information that allows them to revise or reject thewrong rule they were operating with when they produced the error.Although this three-way definition looks concise, I would object to the first meaningmentioned. When the teacher lets learners to correct themselves and only signals that anerror or a mistake has been made, then we cannot talk about error correction, but rathererror treatment or feedback in general.Correction should by no means be understood as criticism or punishment oferrors. As Julian Edge (1989) points out, “Correction is a way of reminding students ofthe forms of Standard English” (p. 20). When correcting, teachers help students becomemore accurate in their L2 performance. It should not mean “insisting on everythingbeing absolutely correct” (Edge, 1989, p. 33); on the contrary, correction should beperceived as a means of helping learners on their way to mastering the L2, not as a toolfor reaching perfection. For that purpose, Bartram & Walton (1991) prefer the termmanaging to correcting, as they believe that the outcome of an erroneous utterance

67should be the same as of a correct one: the students will make progress (Bartram &Walton, 1991).5.1.1. FEEDBACK ON ERRORThere are several types of error feedback (or treatment), error correction beingjust one of them. On the most general level, feedback on error can be positive ornegative. According to Ellis (2009),Positive feedback affirms that a learner response to an activity is correct. It maysignal the veracity of the content of a learner utterance or the linguisticcorrectness of the utterance. In pedagogical theory positive feedback is viewedas important because it provides affective support to the learner and fostersmotivation to continue learning. (p. 3)The opposite is negative feedback. As Ellis (2009) observes, a lot of careful attentionhas been paid to negative feedback from the side of language teachers and L2 theorists,much more than to the positive feedback. Negative feedback signals that the learner’sutterance is linguistically deviant or lacks enough veracity. Error correction is thena type of negative feedback; as Ellis (2009) points out, it “takes the form of a responseto a learner utterance containing a linguistic error” (p. 3). The individual parts whicherror correction can consist of will be dealt with in the following sections.Corrective feedback in SLL is a very complex phenomenon. Although it isaddressed in majority of handbooks for teachers, has been subject of many studies (in2006, Russell & Spada talked about fifty-six) and it commonly occurs in languageclassrooms, there are still no simple conclusions regarding successful ways of providingcorrection. On the whole, it is not possible to form clear instructions for teachers thatwould form informed advice (Ellis, 2009). The reason for this, as Ellis states, is the

68complexity of error correction “as an instructional and interactive phenomenon and as apotential tool for acquisition” (p. 11-2). Bartram & Walton (1991) further support thisopinion. As Julian Edge (1989) argues, teachers’ job is much more difficult andinteresting than just passive listening for learners’ deviations.The complexity of error correction is manifested in the many decisions teachershave to make (Ellis, 2009). Although it is not possible to give teachers any clearprescriptions for handling errors and mistakes, awareness of the possible steps that canbe taken within each decision can work as guidelines for teachers in their deciding. It istrue indeed that their decisions on correction must be led by specific situations and thelearners needs; however, to a certain extent, rules regarding the above mentionedchoices can be applied in general and thus can be of good use to teachers in theirclassroom practice.5.2. TO CORRECT OR NOT TO CORRECT?It is clear today that error correction has a substantial effect on the acquisition ofone’s target language (Russell & Spada, 2006). The teachers’ attitude towards error andcorrection is thus of crucial importance.In general, it is possible to say that learners expect their teachers to correct them,because “that is the traditional view of what a language teacher does” (Bartram &Walton, 1991, p. 28). Most often, learners come from educational backgrounds whereteachers correct extensively; one could hardly find a secondary-school language teacherwho does not correct at all. Due to this fact, learners will expect to be corrected.According to Bartram & Walton (1991), this, however, does not mean that all studentswant to be corrected. The authors distinguish two groups of students: those who believe

69that correction is good for them and thus want to be corrected a lot; and those who feelfrustrated by correction. As the authors observe, these two groups “often findthemselves in the same class” (p. 29) and this is where the question of correction comesto play.Although it may seem that at a single moment a teacher corrects one error of oneparticular learner, the reality is slightly different. In the situations of group learning, asis generally the case in basic and secondary schools, the fact is that the output of onelearner may serve as input of the other learners (Allwright & Bailey, 1991). When alearner uses a deviant form of the target language, the teacher’s decision whether tocorrect will affect more people at the same time. As Allwright & Bailey (1991) explain,“If a teacher chooses not to treat an error in one learner's utterance, the other learnersmay assume that the form or function was correct as it stood” (p. 102). In deciding on asingle error, teachers can influence the learning process of more than one student at atime; and she can both support the process and hinder it – all depends on the decisionwhether to correct.There is growing evidence that error correction is overall useful and can behelpful in L2 learning (Russell & Spada, 2006). Results of classroom research, such asLightbown & Spada (1990), show that “when teachers corrected learners' errors duringcommunicative lessons, the frequency of at least some errors […] was reduced” (Ellis,1998, p. 53). As Kulič (1971) points out, correction is nearly always more effective thannon-correction. Although, as Kulič stresses, correction makes sense only when it isaccounted for – simple mechanic information on the correct form does not work assatisfactory correction. The feedback should thus include some explanation or commenton the error. When doing so, then such “feedback after wrong responses may have thegreatest positive effect on student learning” (Meyer, 1986, p. 228). Such feedback can

70accelerate the learners’ interlanguage development and thus influence it positively. Infact, it serves as a signal showing that there is a gap between their knowledge and whatis regarded as acceptable or correct (Ellis, 1998). Showing learners that their utterancesare erroneous can thus be of great importance to their L2 development.5.2.1. OPONENTS OF CORRECTIONDespite the growing evidence of correction being useful, there are still languagetheorists who find error correction completely useless (Truscott 1996, 1999, 2007;Krashen 1982) or claim that it influences only the immediate performance but leavestheir general competence untouched (Schwartz 1993). As Allwright & Bailey (1991)put it: “No matter how hard a teacher tries to correct errors, in the long run, only thelearner can do the learning necessary to improve performance, regardless of how muchtreatment is provided” (p. 99). The authors state that one of the main reasons for theirscepticism about correction is the fact that treatment is not the same as cure, which theysee as a major problem. Just because teachers correct errors and mistakes does not meanthat students have learned to use the problematic rule correctly. Truscott (1996, 1999,2007) claims that error correction in oral practice does not work because of thefollowing five reasons:o Teachers correct inconsistently, sometimes wronglyo Students are sometimes hurt by being correctedo Students may not take corrections seriouslyo Correction may interfere with fluencyo Learners do not learn from the correctionAlthough these comments may be valid, the arguments contradicting these statementsare even stronger. Penny Ur (2012) strongly argues against these statements and

71supports her claims with results of several researches. Although teachers correctinconsistently, the same can be said of doctors treating patients; in the latter case,however, nobody requests doctors to stop treating patients. Secondly, although it is truethat some students might feel discouraged by correction (and my research will show thatthe percentage is low), students see correction as a part of language learning and acceptit. Thirdly, a big part of teaching might not be understood or taken seriously (i.e. givinginstructions), but that does not mean teachers should stop doing it. Fourthly, teachersare professionals and should know when to correct. Moreover, there is strong evidencethat students want to be corrected in general –as Cathcart and Olsen (1976) state,“learners say they want more correction than is typically offered by their teachers”(cited in Allwright & Bailey, 1991, p. 103). Finally, research is inconsistent in the lastpoint; however, it is more in favour of correction than against it. Ur adds that learningfrom correction is by all means gradual, not immediate – that does not mean though thatit is ineffective. Overall, Ur admits that correction may only work partially or gradually;however, if there is anything even less effective, it is not correcting. If learners receivetoo much neutral feedback on erroneous utterances, their errors will be reinforced andmay gradually become fossilized. If teachers employ professional competence andcorrect with tact (Ur suggests using phrases such as ‘What a lovely mistake, I’m so gladyou made it’), then error correction will be more effective than no correction at all. AsAllwright & Bailey (1991) conclude, “teachers must provide learners with appropriatecognitive feedback as well as affective support” (p. 99) to secure effective errorcorrection.

725.2.2. PROBLEMS OF HEAVY-CORRECTORS AND NON-CORRECTORSBesides keeping a supportive attitude, it is necessary to find the right balancebetween over correcting and correcting too little. Both extremes are problematic, asBartram & Walton (1991) argue, and therefore a right balance is needed. In the case of aheavy corrector, the main problems are that the class is teacher-dominant and there islack of space for students’ creativity. Moreover, learners become too cautious and tense,as they are worried to make errors. Bartram & Walton (1991) point out that one of thecommonest problems is that teachers feel obliged to do all correcting themselves and tocorrect every mistake or error that occurs. A non-correcting teacher is problematic, too:such situation often rises from teachers feeling guilty to correct or feeling thatcorrection is useless. As a result, students often complain, since they expect to becorrected, as was already mentioned above. Additionally, learners tend to see suchteachers as incompetent or lazy and begin to worry if such teacher is actually aprofessional at all (Bartram & Walton, 1991). Since teachers often argue that they spendmost of the time during their lessons deciding on correction (Bartram & Walton, 1991)it is surprising to point out that David Nunan found in a classroom research that only 4% of all decisions teachers make are to do with correction. As most linguists agree, it isnecessary to find an optimal level of correction in order to make it effective. As Bartram& Walton (1991) stress, it is not proven that heavy correctors or non-correctors are themost successful teachers; what is clear, however, is that the most important factorinfluencing the class is the teachers’ attitude towards correction. For that reason, it iscrucial for teachers to consider the appropriateness of correction in each individualsituation before rushing into correction or deciding to abandon it.

735.3. ERROR CORRECTION STAGESThe process of error correction in oral practice consists of several stages. AsMacDonald Lightbound (2005) observes, the simple process of locating an error ismuch more difficult in oral production than it is in writing. Before analysing individualdecisions teachers have to make during corrections, I will thus have a brief look at theindividual stages of error correction.Linguists are not at one in identifying individual stages of working with error.Hendrich et al. (1988) identify three stages: identification (an error has occurred),interpretation (type of the error) and correction (suggesting an acceptable solution).Although Broughton et al. (2003) identify three stages, too, they are slightly different:In the first stage, teachers establish what the error is; subsequently, they establishpossible sources of the error in order to be able to work out an effective strategy ofdealing with it; in the final stage, teachers decide on the seriousness of the error anddecide whether correction should be provided (for more detail, see Broughton et al.,2003, p. 136-7). The most detailed staging of the process of dealing with an error isprovided by Choděra (2006), who identifies five separate phases: detection (noting anerror has been made), identification (identifying the type of error), interpretation ofsources (identifying possible reasons for error), correction (dealing with the error) andfinally prevention (remedial teaching, more practice etc.). The last stage is only optionaland does not necessarily need to be present; teachers may also decide in the fourth stageagainst correction. Overall, this classification of stages seems to be the most fitting withrespect to the complicated process leading to correction and the correction itself.According to Corder (1981), locating errors in learners’ utterances is logicallyachieved by comparing what is said with what the teacher believes the learner wanted tosay. Corder calls those correct utterances reconstructed utterances (p. 37). This is done,

74as Corder explains, by comparing the learner’s performance with “the target language’scode and identifying the differences” (p. 54). I must, however, partially disagree withCorder on this: we cannot compare learners’ utterances with the target language code ingeneral, since then we would have to be correcting not having used present perfect tenseeven though the learner might not be familiar with the tense yet. More likely, wecompare the learner’s performance with an ideal model of the learner’s interlanguage –that is, we compare the utterance only with the part of the L2 the learner should ideallybe familiar with. In doing so, we get a more reliable image of the learner’s utterancethan by comparing it with the whole L2.5.3.1. CORRECTION OPTIONSKathleen Bailey (1985) came up with a useful taxonomy of seven basic optionsand eight complementing features for each option regarding the choices teachers facewhen deciding on correction:Basic options:1. To treat or to ignore2. To treat immediately or to delay3.To transfer treatment (to, say, other learners) or not4. To transfer to another individual, a subgroup, or the whole class5. To return, or not, to the original error maker after treatment6.To permit other learners to initiate treatment7. To test for the efficacy of the treatmentPossible features:1. Fact of error indicated

752. Location indicated3. Opportunity for new attempt given4. Model provided5. Error type indicated6. Remedy indicated7. Improvement indicated8. Praise indicated (cited in Brown, 2007)It is clear from this taxonomy that error correction is far from being straightforward: infact, it is a rather complicated process involving lots of important decisions on the sideof the teacher. To be able to do so, Brown (2007) observes that teachers need to“develop the intuition, through experience and established theoretical foundations, forascertaining which option or combination of options is appropriate at given moments”(p. 348). These decisions, regarding who should treat an error, how and when will bedealt with in the next sections.5.4. CORRECTION TECHNIQUES – HOWTO CORRECT ERRORSMost teachers would agree that it is necessary to use error correction in a waysupporting the learning process (Edge, 1989). What they would probably not agree on,however, is how to correct errors learners make. The reason for this is probably the factthat “there is no single best technique for correcting errors” (Doff, 1993, p. 190).Despite that, theory suggests several strategies that teachers should be using generally inorder to make their correction techniques efficient.Many SLL theorists agree that teachers should always suit correction to theirlearners’ needs and individualities: “the most important thing is for the teacher to be

76flexible and to be aware of the effect on each individual learner of correcting errors”(Doff, 1993, p. 190). As a result, teachers should be using different strategies andtechniques according to the type of error, the learner’s personality and the situation;teachers should “make a series of instant judgments about the learner’s language egofragility; anxiety level, confidence, and willingness to accept correction” (Brown, 2007,p. 350). Teachers may therefore decide to abandon correction if the error was made byan unconfident, highly anxious learner. It is thus always necessary, as Bartram &Walton (1991) point out, to listen to our students carefully, observe them and evaluateeach situation individually.Research into classroom practice has found some techniques more efficient thanothers. According to Meyer (1986), “significant effects were found for teachers usingsustained feedback, staying with the student who first made the error” (p. 228). Thistype of feedback led to most improved answers and seemed to be the most effective. AsMeyer states, teachers used process feedback –giving students hints leading to selfcorrection– only very rarely and thus it was impossible at that time to observe itseffectiveness.Besides the strategies already mentioned, Adrian Doff (1993) adds several thatare generally found useful. In his book Correcting errors, Doff argues that teachersshould focus on what is right instead of busying themselves over errors. Additionally,teachers should praise students for correct answers, since they will get a feeling ofprogress in their interlanguage development and become more willing to risk andexperiment with language. Finally, Doff observes that teachers should perform errorcorrection quickly: “if too much time is spent over correcting errors, it gives them toomuch importance and holds up the lesson” (p. 190).

77What teachers must avoid at all costs, on the contrary, is corrective feedback thatis derogatory or punitive in any way. Such feedback is viewed by learners asdevaluating or insulting (Brown, 2007). Hubbard et al. (1983) claim the same when theyobserve that “Nothing will undermine a learner’s confidence as much as a series ofderogatory comments on his language performance” (p. 143). Such attitude of teacherswould be particularly damaging and would encourage withdrawal on the part oflearners. It is thus necessary for teachers to maintain positive attitude towards correctionand try to avoid such humiliating feedback in general.5.5. WHICH ERRORS TO CORRECTAs Christison & Krahnke (1983) observed, methodological recommendationsrange from correcting all spotted errors to not correcting at all. With the preference forCLT today, the directions for teachers are not as variable as before.Language theorists are at one regarding correction of global and local errors. AsBrown (2007) explains,[…] local errors usually need not be corrected since the message is clear andcorrection might interrupt a learner in the flow of productive communication.Global errors need to be treated in some way since the message may otherwiseremain garbled. (p. 347)Global errors, or those that impede intelligibility or interfere with communication,should thus be given priority in correction. Hendrickson (1978) points out that errorscausing the most unfavourable reactions, since they are the most stigmatised, shouldalso become high candidates for correction. Besides, errors that occur most frequentlyshould also be seen as needing correction the most (Hendrickson, 1983).

78CLT evaluates seriousness of error from the point of communication: errors thatcause a breakdown of communication should be considered the most serious, whereaslittle details not disturbing easy understanding do not need much treatment. As BarbaraSeidlhofer (2004) explains, many errors typically regarded by teachers as gravest do notin fact cause much communicative trouble:In particular, typical "errors" that most English teachers would consider inurgent need of correction and remediation, and that consequently often getallotted a great deal of time and effort in English lessons, appear to be generallyunproblematic and no obstacle to communicative success. (p. 220)As Seidlhofer explains, among those errors are dropping the third person present tense –s, confusing relative pronouns who and which, omitting articles, or failing to use correctquestion tag forms (Seidlhofer, 2004). Other forms, which obviously cause much moretrouble in communication, such as lack of vocabulary or cases of unilateral idiomaticity(using idiomatic expression from one’s L1 literally translated into the L2), are not givenmuch attention in correction. Seidlhofer (2004) sees this as a highly problematic featureof error correction and claims that teachers should correct more in line with CLT. Thisobservation can be compared with what teachers stated as the most problematic errors inmy research.5.6. WHEN TO CORRECTAnother question which teachers have to answer for themselves when decidingon correction is when to correct. As Bartram & Walton (1991) claim, “often thespontaneous reaction on hearing an [error] is to correct immediately” (p. 4). That doesnot mean that it is the only option possible, or the best one. The possibilities regarding

79the timing of correction include immediately, after a while (when the learner finishes asentence/whole utterance etc.), at the end of an activity, later in the lesson, at the end ofthe lesson, in the following lesson, later in the course, or never (Scrivener, 2005); thedecision for specific timing is also influenced by the type of activity the learners areconcerned with.The basic options regarding timing are either delayed or immediate correction.The problem of immediate correction is that “it often involves interrupting the learner inmid-sentence – a practice which can certainly be disruptive and could eventually inhibitthe learner's willingness to speak in class at all” (Allwright & Bailey, 1991, p. 103). Inother words, the affective feedback would be negative. On the contrary, psychologyliterature shows that “feedback becomes less effective as the time between theperformance of the skill and the feedback increases” (Allwright & Bailey, 1991, p. 103).This observation makes the decision even more complicated to teachers. Moreover,classroom research does not claim general preference for delayed or immediatefeedback regarding its value for language acquisition. Despite all that, teachers have tocorrect their learners and make decisions regarding timing of correction on everydaybasis. For that reason, it might be useful to have a closer look at what theoryrecommends regarding individual timing options.5.6.1. ACCURACY AND FLUENCY DICHOTOMYAccording to Broughton et al. (2003), immediate feedback is of extreme value tolearners. This type of feedback is generally preferred by methodologists in activitiesaimed at accuracy (Ellis, 2009). Accuracy activities are such where “students areencouraged to make their utterances as near to a native-speaker’s as possible – which isusually taken as necessitating more intense correction” (Bartram & Walton, 1991, p.

8032). Since the aim of such activity is practising newly learned forms and improvingone’s interlanguage, immediate correction is at place, since it gives the learners instantfeedback on how well they are absorbing the new information.A completely different situation comes with fluency activities, such as roleplayingor free discussions. These are such where learners “work on their capacity tocommunicate within the language” (Bartram & Walton, 1991, p. 32). In these activities,methodologists generally dismiss correction at all, since that would mean insisting onaccuracy, whereas the aim of such activity is communication and “developing studentmotivation to speak and increasing fluency and ease of expression” (Valette, 1973).Immediate correction might, on the other hand, inhibit students’ desire to speak and thusthe whole learning process. In order to bring about fluency students need to experience“uninterrupted, meaningful communication if they are to learn to use the language”(Edge, 1989, p. 37). When we are paying attention to what the learners say, we shouldthus not correct them immediately. In such cases, encouragement is more importantthan correction and more probably will lead to greater communicative independence.The only exception to this, as Edge (1989) observes, is errors which affectcommunication and make information exchange impossible – in such cases, correctionshould be applied even in fluency activities.A special case of communicative activities aimed at fluency is tasks carried outin pairs or small groups. In such cases, teachers are unable to hear all students at thesame time and as Littlewood (1994) observes, “unnoticed errors are the price to pay forthe more intensive and individualised practice” (p. 54). The aim of such activities iscommunication rather than accuracy; therefore teachers should not indulge inimmediate correction. Such type of activities naturally calls for delayed correction,where teacher notes down errors she hears and deals with them later – after the activity,

81at the end of the lesson or even later (Brown, 2007). In doing so, the activity is notdisrupted and all learners have an opportunity to hear the feedback on the error, whichis especially useful with repeated, frequent errors.Apparently, the crucial factor in deciding on timing is the aim of the givenactivity. In the case of normal conversation, there is no space for overt correction: “Ifstudents are to say anything meaningful, they need to feel that people are listening towhat they are saying, not to how they are saying it” (Edge, 1989, p. 37). Nevertheless,there are some actions that teachers can take, such as asking for clarification orrepetition (Sorry I didn’t catch that) or repeating something in a slightly different way(Really? Do you mean that…?) in order to ensure what the speaker means (Edge, 1989).In doing so, teachers use corrective feedback but at the same time learners feel involvedin conversation, not just an exercise. This type of correction could be regarded as nonintrusiveimmediate feedback.There are language theorists, however, who see the binary opposition ofaccuracy-fluency activities problematic. As Bartram & Walton (1991) argue, teachersshould not be deciding on the basis of the aim of an activity, but rather based on thestudents’ aims. As they observe, it is “very unlikely that the student is thinking in termsof fluency or accuracy” (p. 33). What they also see problematic is the two-sided view ofcorrection and language which are in fact multi-sided (Bartram & Walton, 1991).According to them, teachers should rather consider whether the error is caused whileworking on the aim of a lesson or not: “If the students make [an error] when working onthe aim of the lesson, the [error] should be dealt with. If the [error] is on another point,it can be left” (p.37). Such approach places the students’ aims above the type of activity,which might be more useful particularly in cases when the accuracy/fluency distinctionis not clear cut: as Valette (1973) observes, sometimes accuracy activities can turn into

82fluency activities and vice versa, because the learners felt the need to practice the latter,no matter what the intended aim was supposed to be. Other language theorists, such asDoughty (2001), claim a need for immediate correction even during fluency activities.In his opinion, delayed correction “leads to focal attention on form resulting in explicitrather than implicit L2 knowledge” (cited in Ellis, 2009, p. 11), which is indeednegative. Such position is in sharp contrast to the generally accepted inclination towardsdelayed correction in fluency activities.From the opinions of theorists such as Doughty or Bartram and Walton it ispossible to see that it is difficult to arrive at any general satisfactory conclusionregarding delayed and immediate correction. Despite that, it is clear that a vast majorityof language theorists adopts the fluency-accuracy dichotomy and the preference fordelayed or immediate feedback respectively. The implications of SLL theory are clear:during fluency activities aimed at communication errors should be tolerated and if needbe, corrected only afterwards; the only exceptions would be errors preventing successfulcommunication. In accuracy activities where the aim is learning, on the contrary, thefocus is on form and thus immediate error correction is in place. A special case wouldbe normal conversation, where there is generally no space for correction and teacherscan resort to phrases seeking clarification or repetition only. It is thus clear that decisionabout timing is rather a complex one. As with the other aspects already mentioned,every decision needs to be tailored to each specific situation and learner’s needs in orderto be most effective.

835.7. WHO SHOULD CORRECTWhen Richards and Renandya were observing conversation classes in HongKong in 2002 they found out that all correction was carried out by the teacher only. Theteacher was there the sole source of input as well as feedback to the learners’ utterances.This kind of pattern is the more traditional one (Richards & Renandya, 2002). AsBroughton et al. (2003) point out correction may, however, also come from other peoplethan the teacher –from the learner as self-correction or other learners in the classroom.This section will deal with individual types of correction regarding the source personand will discuss their positives and negatives.Teachers have traditionally been the most common source of feedback tolanguage learners in general (Allwright & Bailey, 1991). This was due to the dominantposition they occupied in the classroom; with CLT, this has slightly changed andteachers have become more of one possible sources of correction. Scrivener (2005)explains that the reason for this is that “It may be that being over-helpful as a teachercould get in the way of learning. I cannot learn for my students. The more I do myself,the less space there will be for the learners to do things” (p. 21). The reason is thus theintention to give learners more space for learning. As Julian Edge (1989) claims, “Themore the students are involved in correction, the more they have to think about thelanguage used in the classroom” (p. 27). Since thinking about language supports thelearning process, it is necessary to give such opportunities to learners.Most SLA theorists argue today that teachers do not need to correct straightaway, but should rather signal that an error has occurred by focusing attention on theplace of the error, or just explain what has gone wrong. Julian Edge (1989) explains thesituation as follows:

84If we think that [an error] needs to be corrected, and if neither the student whomade [it], nor any other student can correct it, then the teacher has to give morehelp. This still doesn’t mean that the teacher has to give the correct form straightaway. (p. 27)It is more beneficial for learners if teachers lead them towards the correct solution,which they reach more or less individually, than being told the correct form straightaway. Carl James (1998) reports on a research where learners evaluated their teacherson the basis of the ability to explain language and relates it to error correction:Learners who are asked to evaluate their teachers often refer to the good teacheras someone who can “explain” the TL and its difficulties clearly. If explanationis perceived as a sign of good teaching, it must also be a sign of good correctingof error. (p. 263)This observation supports the idea of teachers not carrying on the correction themselvesbut rather giving their learners time to self-correct or leading them towards self-correction.Holley & King (1974) concluded from their classroom research that what learnersneeded most was extra wait-time, when the teachers did not intervene and let thelearners answer themselves: “They reported that in over fifty per cent of the cases theyvideotaped no corrective efforts from the teacher were needed” (Allwright & Bailey,1991, p. 108). Given the extra wait-time, teachers mostly did not need to intervene at alland learners were able to self-correct.5.7.1. SELF-CORRECTIONSelf-correction is thus another option in the question of who should correct andaccording to some theorists it is also the best form of correction (Edge, 1989). The

85reason for this Edge states is that people usually prefer to sort out their errorsthemselves rather than being corrected by someone else – which is true not for languagelearning only. Kulič (1971) also stresses preference for independent self correction,especially in cases of higher, more complicated learning processes, since self-correctionnot only corrects the result, but also the process which led to such result, and increasesactive participation of the learner in this phase of learning, too. As the author claims,error can become a vehicle of valuable correcting and improving information for thelearner, particularly in the case of the more complicated learning processes (Kulič,1971). As such, self-correction is a valuable help in the development of IL and shouldbe given enough space in classrooms.According to Allwright & Bailey (1991), self-correction should be our long-termgoal. Since learners only are capable of progressing with their IL, as teachers cannotlearn for them, self-correction is the ideal goal for them to achieve: “students will repairtheir own communication breakdowns and produce the target language accurately andfluently without guidance from us, and that the correct forms will be internalised” (p.107). We thus want our learners to become capable of self-initiated self-repairs, as theauthors add. To achieve that, it is advisable to provide them with the extra wait-time, asHolley & King (1974) suggested. This will make learners more independent and preparethem for real-life situations better.A big advantage of self-correction, as James (1998) points out, is that it issubject-sensitive and non-threatening. James stresses that correction needs to be madewith respect to the type of learners: “The more sensitive the learner, the more gentleshould be the correction” (James, 1998, p. 251). It is clear that the most sensitive type ofcorrection possible is the already mentioned self-initiated self-repair, where studentscannot experience humiliation or feelings of shame in front of others. James also argues

86for an extra wait-time with teachers signalling that an error has occurred: “Given anextra second or two, accompanied by a slightly disapproving or puzzled gesture on theteacher’s part and the pupil will often be able to self-correct, with little or no loss offace” (James, 1998, p. 250). It is thus obvious that self-correction is highlyadvantageous to the learning process of our learners; before leaping into correction,teachers should thus try to give their learners the chance to self-correct by signallingoccurrence of an error and providing them with extra time to think.5.7.2. PEER CORRECTIONThe other option for teachers transferring correction onto someone else is peercorrection. This usually comes into play when students are unable to self-correct (Edge,1989). Learners can help each other especially with accuracy and form, as Russell &Spada (2006) observe. With peer correction, however, it is important to go back to thelearner who first made the error (Edge, 1989) to make sure s/he has understood whatwas wrong at first place. Furthermore, it is necessary to ensure that it is not the samefew students who do correction always: “The idea of peer correction is to encouragecooperation, not to put one or two students in the traditional place of the teacher” (Edge,1989, p. 26). If these conditions are fulfilled, then peer correction can be beneficial, too.In his book Mistakes and Correction, Julian Edge (1989) states four mainadvantages of peer correction. Firstly, both the learner who made the error and the onewho corrects it are involved in thinking about language and listening. Secondly, teacherthus gets a lot of important information regarding students’ in/ability to hear and correcterrors that occur. Thirdly, peer correction supports cooperation and mutual help andmakes learners more independent from teachers. Finally, getting used to peer correctionimproves mutual help during pair or small group work when teachers cannot hear

87everybody. Besides these advantages, Carl James (1998) mentions the minimal threat toface which peer correction means to learners – if possible, James argues for peercorrection rather than teacher correction to minimize the threat.5.8. WAYS OF INDICATING ERRORSResearch into classroom practice has increasingly been showing that correctioncan help learning, as was already discussed in the previous sections. The focus ofresearch has thus shifted from asking whether correction works into what kind ofcorrection works best (Ellis, 2009). Numerous studies reveal that the uptake of learnersgreatly varies depending on the type of corrective strategies used. The findings alsoshow that teachers have a large variety of techniques available for error correction,although they do not generally make full use of the whole range and tend to prefer oneor two strategies (Allwright & Bailey, 1991).More recently, as Ellis (2009) states, researchers have started developinghierarchical taxonomies based on theoretical assumptions regarding correction andlanguage acquisition. The reason for this is perhaps the disagreement regarding efficacyof individual ways of indicating errors. The most widely used taxonomy of oralcorrection forms distinguishes individual types of feedback as explicit vs. implicit, andinput-providing or output-prompting types (Ellis, 2009). This typology probably comesfrom Lyster & Ranta (1997) who identified the following types of corrective feedbackin their classroom research and categorised them as follows:Implicit ExplicitInput-providing Recast Explicit correctionOutput-prompting Repetition; ClarificationrequestElicitation; Paralinguisticsignal; Metalinguisticfeedback

88(Ellis, 2009, p. 8)In explicit correction, the teacher “overtly states that a learner's output was not part ofthe language-to-be-learned” (Carroll & Swain, 1993; cited in Russell & Spada, 2006, p.137), whereas in implicit feedback, “learners must infer that the form of their utteranceis responsible for the interlocutor's comprehension problems” (Russell & Spada, 2006,p. 137). Lyster & Ranta (1997) defined the above mentioned types of correctivefeedback as follows:o recast – saying the correct expression without commenting on it or explaining ito repetition – repeating the error, usually with rising intonation or stresso clarification request – indicating that the utterance was not understoodo explicit correction – providing the correction straight awayo elicitation – attempt to elicit the correct form from the speakero metalinguistic feedback – using technical linguistic explanation of an errorLyster and Ranta found in their research that recast was the most frequently used type,although it is also the “least likely type of feedback to elicit student repair” (Ellis, 1998,p. 52). As the authors observed, recasts may not be perceived as correction at all. This isin keeping with an observation by Egi (2007), who found that learners did notsometimes perceive recasts as a type of correction, but rather a reaction to the content. Iflearners indeed fail to realize they are being corrected, then such way of indicating erroroccurrence is seemingly an inadequate one. On the contrary, elicitation andmetalinguistic feedback turned out to be the most effective ones, although theydramatically lacked behind recast in the frequency of use (Lyster & Ranta, 1997). Thisresearch thus brought clear results in favour of explicit ways of indicating errors.Other researchers into strategies of indicating error showed similar results. Ellis,Loewen & Erlam (2006) concluded that both implicit and explicit types of feedback

89assist language acquisition, but explicit feedback is more effective in general than theimplicit one (cited in Ellis, 2009). In an experimental laboratory study, Carroll andSwain (1993) also reported preference for explicit types of feedback as the mosteffective one, although some positive effects were observed with recast, too. Studies ofMackey & Philp (1998); Iwashita (2003); and Philp (2003) uniformly reported that lessadvanced students needed more explicit feedback, whereas more developed learnersresponded positively to recasts, too (taken from Russell & Spada, 2006, p. 138). Allresearch more or less uniformly shows explicit ways of indicating errors as the moreeffective ones, with recasts being acceptable with more advanced students.Penny Ur (2012) conducted a research into the needs of young Israeli learners ofEnglish regarding error correction and the preferred types of feedback. The researchshowed that learners want to be corrected in general and see correction as valuable totheir language development. Regarding the type of feedback, they showed a strongpreference for explicit correction. It is therefore interesting to observe that learnersvalue the most the type of correction that has turned out to be the most effective, too,although they might not be doing it knowingly. It will be possible to compare theseresults with what learners in my research have claimed to value the most.This chapter has discussed error correction, its complex structure, positives andnegatives. Although students may have different attitudes towards correction, theyexpect to be corrected in general. Correction influences not only the learner who haserred, but rather everybody present in the learning process. Research shows thatcorrection in overall is useful, as it can speed up the interlanguage development.Although some language theorists still stand against correction, their arguments are notstrong enough. Even though correction might be only partially effective, it is still more

90effective that no correction at all. Moreover, non-correction might be harmful, as it doesnot help to modify learners’ wrong hypotheses and may even lead to fossilization.Furthermore, teachers should always try to find the right balance between noncorrectingand over-correcting.The following section has briefly presented the individual stages leading towardserror correction and including correction itself. It has showed that the process is not assimple as it might seem and includes several important decisions on the side of theteacher. In detecting errors, teachers must be comparing their learners’ utterances notwith the whole of the target language code, but rather only with the part the learnersshould already be familiar with.Regarding the question of how to carry out error correction, theory suggests thatthere is no single uniform technique that works best in general. On the contrary,teachers should try to fit the type of correction to individual learners and situations.Overall, there is an inclination towards sustained feedback over process feedback.Correction should never be humiliating and teachers should always try to praise correctanswers apart from correcting errors.With respect to what errors should be corrected, CLT places emphasis on thosethat impede communication and easy understanding than little grammatical details.Global errors are generally seen as more disturbing than local ones; errors that occurmost frequently should also be given priority in treatment. Overall, teachers shouldalways consider the extent towards which communication is disrupted and based onthat, decide on correction of individual types of error.Considering when to correct, the SLL theory suggests that during fluencyactivities aimed at communication errors should be tolerated and if need be, correctedonly afterwards; the only exceptions would be errors preventing successful

91communication. In accuracy activities where the aim is learning, on the contrary, thefocus is on form and thus immediate error correction is in place. Nevertheless, teachersshould always bear in mind the learners’ aims and needs and should place them abovethe activity aims.The choice of individual types of correction regarding who should correctlargely depends on the type of learner as well as the situation. Some learners mightprefer teacher correction, since they perceive it as more valuable; other learners mayfeel annoyed by their peers correcting them due to a certain lack of authority.Nevertheless, the generally prevailing pattern calls for self-correction supported bysufficient wait-time first, followed by either peer correction, or, if other learners areunable to correct, resort to teacher correction in the end. This pattern may, however,differ based on the individual needs of learners as well as time possibilities andatmosphere in the class.Regarding indicating errors, a great variety of techniques is available to teachers.Despite it, teachers have shown a strong preference for recasts, which are, however,probably the least efficient type regarding language acquisition. On the contrary,explicit feedback has turned out to be the most effective (elicitation and metalinguisticfeedback in particular) and at the same time also preferred by learners (as far as researchcan tell now). These findings, together with the observations from the previous sections,can now be compared with the results of my research among teachers and students ofEnglish.

926. BACKGROUND TO RESEARCHFor my research I created questionnaires for both teachers and students tocompare their opinions about error and error correction. Both questionnaires were madein an e-form in order to be able to reach as many respondents as possible. Thequestionnaires were spread by email and via social networks among teachers of Englishin the Czech Republic and students of English aged 13 to 20 (lower and uppersecondary school). The respondents were aware of the fact that their answers would bekept anonymous.Both the questionnaires consisted of eleven questions regarding error and errorcorrection. The following types of questions were used in both of them: open-endedquestions, multiple choice, rating scale questions and two-option questions. In the caseof rating scale questions, respondents were ordering the options based on preference,frequency of occurrence, or seriousness.Each questionnaire was introduced by a short text explaining the purpose of theresearch and giving short instructions on how to fill the questionnaires in (for moredetails, see Appendices 1 and 2). Respondents were asked to state their gender and age.Besides, teachers were asked whether they had studied any methodology courses andasked about the number of years of teaching English. This data were collected in orderto see if the answers differed based on any of the above mentioned aspects.6.1. ANALYSIS OF COLLECTED DATAThe data were collected during February and March 2013. In total, 48 teachersfilled in the teacher questionnaire, whereas 150 students filled the other one. The

93questionnaires were running on the web site Moje anketa (www.mojeanketa.cz), whichis a Czech portal designed for creating online surveys.6.2. STUDENT QUESTIONNAIREIn total, there were 150 students aged 13 to 20 who filled in the questionnaire.Out of that there were exactly 50 men and 100 women. The average age of therespondents was 16.2. After the results were gathered they were analysed to see ifdifferent groups of respondents prefer different aspects of correction or are ratherunified. Remarkable differences were observed in three questions between men andwomen as well as the respondents from lower and upper secondary school (whichgenerally correlated with the reached level of English).The whole questionnaire with all the questions can be seen in Appendix 1, whilethe results are in Appendix 3. It should be, however, noted that students were given thequestionnaire in Czech in order to make it accessible for students of all levels ofEnglish. For the purposes of the thesis and reference, the survey was translated intoEnglish.The first chart shows that more than 50% of students feel to be corrected by theirteachers always, and a vast majority of the rest feel to be corrected only sometimes. It isthus clear that students notice correction and realize they are being corrected in general,which is in keeping with the theoretical assumptions stated in the previous chapter.

94Fig. 1: Students’ responses to Question 1Regarding students’ satisfaction with the amount of correction they receive,almost two thirds of students feel satisfied, and 41 students feel their answers arecorrected too rarely. This observation confirms Bartram and Walton (1991) whoclaimed that many students require more correction from their teachers than is usuallyprovided to them.Fig. 2: Students’ overall responses to Question 2However, when I compared responses of the younger and older students, I alreadyfound slight differences. It turned out that many younger students felt they werecorrected too often, whereas only roughly 20% of respondents claimed not enoughcorrection of oral errors. On the contrary, more than a third of the older studentsresponded that they were corrected only too rarely. Only 6.5% of older students felt to

95be corrected too often. The responses show that the fact that a lot of students wish to becorrected more often is true for the older students only, while the opposite is true withthe younger ones.The responses also show that most students notice that the amount of errorcorrection in speaking depends on the type of activity they are involved in. One third ofstudents felt correction is done equally regardless of an activity, whereas one in fivestudents thought that the amount of correction students get differs from one student toanother regardless of an activity. This may suggest that teachers adapt their correctivestrategies not on the basis of the activity only, but also to individual student’s needs.The most frequently corrected types of error as perceived by the students aregrammar and pronunciation errors. Vocabulary and word order were not felt to becorrected as often, and the least frequently corrected error was error concerning thewhole utterance or sentence. It is interesting to see that since many theoreticians believethat global errors affecting the whole sentence or utterance are the most serious onesand should be the hottest candidates for correction in general. It will be interesting tocompare these answers with what types of error teachers feel they correct mostfrequently.The question about students’ typical reactions to error turned out as one of themost interesting ones, as it generated various answers from both different ages andgenders. Overall, nearly half the respondents selected option two, where they admitcorrection and then continue talking. One in five students stated some kind ofconsulting the error with the teacher, whereas only 14% stated they carry on withoutthinking about the correction and the error (see App. 1, Fig. 7). The remaining10% ofstudents uniformly responded that they correct themselves and then continue talking.This shows that a vast majority of students is by no means depressed by correction,

96which is an argument some theoreticians use against correction. This survey shows,however, that 92% of students have no problems with correction and do not getdistracted by it.Fig. 3: Students’ overall responses to Question 5With respect to gender, much more boys than girls either self-correct or do notnotice correction and go on talking without distraction. It is interesting to observe thatonly girls felt frustrated by correction or reported getting out of balance (only one boyreported getting out of balance). The results show that boys seem to be more selfassuredand less affected by correction than girls, who turned out to be more sensitive toit.Regarding the age of respondents, younger students preferred the second option(see Fig. 3) and often reported consulting teacher about the error. They also statedhigher frustration level than the older students. Respondents from upper high school, onthe contrary, showed greater resistance to correction, when more than 20% of themstated they do not notice correction and continue talking without correcting. Thissuggests that teachers should be more sensitive towards younger students whencorrecting due to the learners’ higher susceptibility to it.

97In the questions regarding the students’ opinions on their teachers’ correction, avast majority of students (81.4%) stated satisfaction with their teacher’s corrective style.The remaining students stated they would prefer a change in their teachers’ style ofcorrection. Among these, most reported derogatory style of correction, greatinconsistency in correction, bias against some students, and too little correctionespecially of pronunciation. Not surprisingly, such responses mostly came from theyounger learners, which is in keeping with the findings in the previous question.Furthermore, students were asked to rate various corrective strategies as verygood/good/not very good/or bad. Here students showed a general preference for explicitcorrection from the side of their teachers. On the contrary, they regarded no correctionas bad (almost 65%); therefore they showed a great preference for being corrected. Theyalso showed a general preference for teacher correction, while only 12% thought thatpeer correction is very good for them. The results thus completely mirrored thosegained by Penny Ur in her survey among Israeli students (Ur, 2012). It is also good tosee that students are aware of the positive influence of correction on their learningprocess, since theory believes that error and correction are crucial parts of theinterlanguage development, as was discussed earlier.In a question about their self-confidence and comparison to other students withrespect to making errors, half of respondents believed they make errors equallyfrequently to their peers, and more than a third felt stronger than their peers. Suchresponses support the finding that students generally do not get frustrated by errors andcorrection, since they do not feel to be more prone to them than their peers.An interesting picture is created when one compares responses of boys and girls.The results correlate with those from Question 5, where boys also showed higher selfconfidence.In this case, almost half of all boys felt stronger than their peers, while only

9829% of girls stated to make errors less often than others and rather saw themselves asequally strong.Fig. 4: Boys’ responses to Question 8Fig. 5: Girls’ responses to Question 8When students were asked about the probable sources of their errors, theanswers were distributed rather evenly among all the options. Nevertheless, mostrespondents saw the major source of error in them, which might have greatconsequences for their perception of errors in general. The fact that errors are normal orthat students have too few options to talk were chosen by 20% of students each. Anunsatisfactory course book or teaching methods were the least frequent options, whichsignals that students see material-induced errors as the least typical ones.The last cloze question asked students about their opinion on what an error is.Overall, respondents showed a great liberal view, since most of them chose the third or

99fourth option (purple or dark green in Fig. 6). The responses, however, varied with ageand gender.Fig. 6: Students’ overall responses to Question 10Girls showed a greater preference for standard rules of British or American English andalso the communicative approach to error (fourth option). Boys, on the other hand,defined error more on the comparison with native speakers; similarly, they frequentlyinclined towards the communicative view of error. This shows that boys tend to be moreliberal in their opinion on error. With respect to age, the younger learners perceivederror as more in conflict with either the standard rules of English or with the expectedanswers. The older students showed greater preference for a freer definition of error.This observation suggests that the broader, more liberal concepts of error are acceptedlater in life, with more experience with the language, which is particularly true for boys.The last question in the survey was the only open-ended question. It askedstudents to add any relevant information, interesting observation, or simply anythingthey would like to comment on regarding error and correction. The collected responsesvaried greatly, but several tendencies were observed. Most students reported satisfactionwith the current style of correction of their teacher. Many students stated that correctionis beneficial for them and would welcome more of it in their classes. Regardingcorrective strategies, a majority claimed preference for explicit correction with

100additional explanation of what was wrong; others directly claimed preference forelicitation, self-correction and peer correction. Several respondents would prefercorrection only after the whole performance.Quite a few students used this open question as a space for expressing theirdissatisfaction, too. Several of them would prefer more opportunities for talking in theirclasses. Furthermore, they would welcome greater acceptability of variants, not justthose from the course book, as was the frequent claim. A few students felt that teachersshould adopt a more fair way of correcting when they would approach all studentsequally. What was striking is that quite a large number of respondents also complainedabout their teachers’ humiliating, offensive, and insulting way of correcting their errors.The number of students who were satisfied with their teachers, however, still greatlyoutnumbered those discontented ones.The results of the survey overall showed great similarities to the similar surveycarried out by Penny Ur (2012). The participants showed a great preference forcorrection in general, with stress on self- or teacher-correction. Besides, they preferredexplicit correction strategies. The results also showed frequent variety of opinions withrespect to gender and age. It was showed that only very few students get stressed by thefact that they are corrected and rather want to be corrected, with errors being explainedto them. Students also call for greater acceptance of varieties and see error from thecommunicative approach point of view. It will therefore be interesting now to comparethese findings with what teachers stated in their questionnaires.

1016.3. TEACHER QUESTIONNAIREIn total there were 48 teachers who filled in the questionnaire, out of which therewere 12 men and 36 women. The average age was 32 and the span was really large –from the age of 22 to 64. As with the previous questionnaire, the collected answers wereanalysed in order to find any possible patterns or differences in the answers. Similarly,different preferences were found between various ages, and even greater ones withgender difference. For the purpose of the analysis, respondents were divided into threegroups: those with minimal teaching experience (age 22 – 30), medium-lengthexperience (age 31 – 50) and teachers with 25 years or more of experience (age 51+).The questionnaire is attached in Appendix 2 and graphs showing responses to allquestions can be seen in Appendix 4. Unlike the one for students, this one wasdistributed in English. Since many questions were purposefully similar to or the same asthose of students, it will be interesting to compare teachers’ answers with thoseanalysed in the previous section.The first question, asking about one’s typical reaction to error, generateddifferences between different genders and ages. Overall, more than 80% of teachersdecide on correction depending on the type of activity students are involved in.Interestingly, all men uniformly responded to this question with the option itdepends on the type of activity. Women, on the contrary, differed in their answers as thegraph suggests.With respect to age, only the youngest teachers responded that they typically letstudents self-correct first. Interestingly, the option no correction was not selected byanyone. Within the middle group, respondents nearly uniformly responded they correctdepending on the activity; only 7.2% admitted that they typically do not correct.

102Teachers with the most experience reported either general correction (33%) or againdeciding based on the type of activity (64%).Fig. 7: Teachers’ overall responses to Question 1Surprisingly, only 4.2% of teachers in total stated that they usually correct regardless ofthe activity, which was an option 55% of students chose when describing typicalreaction of their teachers. It might signal that students ascribe correction more space intheir lessons than there actually is. Further research would have to be carried out inorder to resolve this discrepancy and tell what the real situation is in fact. Moreprobably, however, students do not notice their teachers’ strategies fully.The second question, regarding the type of error most frequently correctedgenerated uniform responses among teachers. It turned out that the most frequentlycorrected errors are those of pronunciation or grammar, while the least frequent ones arethose of style and register or the whole sentence. This is in keeping with the learners’responses.The following question concerning the typical corrective strategy used byteachers again generated similar responses across genders and age. Overall, theresponses were equally distributed, with slight preference towards signalling occurrenceof an error and then letting the student self-correct (see Fig. 8).

103Fig. 8: Teachers’ overall responses to Question 3The fourth question, which inquired about teachers’ frequency of errorcorrection, showed slight differences in responses between genders. Overall, there wasagain a strong tendency to decide on correction based on the type of activity (87.5% ofresponses).A vast majority of men claimed to decide on the type of activity, but 8.4% ofthem stated they correct more or less the same during all activities. None selectedcorrecting automatically, which shows that male-teachers probably think about errorcorrection in general. Women, on the contrary, again showed greater differences incorrective strategies as was the case in the first question. Although a majority of themshowed the same strategy as men, there were also such who claimed to adapt correctionto individual students. Even more women (8.4%) stated correcting automaticallywithout thinking.In comparison with the students’ answers, a similar discrepancy as withQuestion 1 was observed. One third of students stated their teachers correctirrespectively of activities. Besides, more than 20% stated correction of individualstudents differs, too.

104Considering timing of correction, which was the subject of Question 5, theresponses differed across genders and ages, too. Overall, most teachers reportedcarrying out correction after the student’s performance (43.8%). Almost a third ofteachers correct after the erroneous sentence and almost 15% of teachers correct at theend of the whole activity. Only 10% of teachers correct immediately; hardly anyonecorrects at the end of the lesson.Most men (41.7%) prefer to correct after the erroneous sentence, whereas amajority of women (50%) correct after the student’s performance. No female-teacherscarry on correction at the end of the lesson, while 8.4% of men do. Twice as manywomen as men prefer to correct at the end of the whole activity. Only men postponecorrection till the end of the lesson. Overall, women seem to prefer less disrupting styleof correction, while men mostly correct immediately or after the erroneous sentence.Different age and experience groups also brought interesting results. Theyoungest group of teachers demonstrated the greatest variety in strategies, withpreference for correction at the end of learner’s performance (see Fig. 9). They werealso the only group to correct at the end of the whole activity or lesson.Fig. 9: Young teachers’ responses to Question 5The middle group slightly preferred correcting after the student’s performance (43%) tocorrecting after the sentence (35.8%). One in five teachers prefers correctingimmediately. The most experienced teachers, unlike the two previous groups, mainly

105correct after the erroneous sentence. One third of them prefer to correct after the wholeperformance.When teachers were asked about the type of activities in which errors should becorrected, their opinions once again differed with both age and gender. Overall,however, they stated preference for correction in accuracy activities only (42%).Almost one third of teachers chose the option Other and stated their opinions. Mostfrequently, the response was both in fluency and accuracy activities, but with differenttiming. Other responses included it depends on the student and the activity, depends onthe error and the student, depends on the error, or depends on several factors at thesame time. It was a pleasant observation that so many teachers see correction as acomplex phenomenon, which cannot be simply judged by the type of activity only. Suchanswers are in keeping with the theoretical observations stated in the previous chapter,which claimed error correction to be a complex issue depending on a great number offactors at the same time, where any simple generalization is too simplifying. With thisquestion, I was hoping to provoke responses dissatisfied with the options given, whichturned out quite well. Besides these, it should be pointed out that one-fourth of teachersstated that errors should be corrected always. This is quite a large number, consideringthat the communicative approach today is strongly against correcting all errors inspeaking.It was mainly men (75%) who chose the option Other and noted on thecomplexity of the issue. Women, on the contrary, felt that errors should be corrected inaccuracy activities only. Almost one third of women correct all oral errors if possible.Such strategy, however, is not an ideal one, at least from the point of view of thecommunicative approach to teaching.

106Regarding the number of years of experience, it was mainly the youngestteachers who preferred correction in accuracy activities only (55%). One-fourth of themthink errors should be corrected always if possible. Teachers in the middle groupshowed a strong tendency towards the option Other, which signals their longerexperience with teaching and more complex view of error correction as a result. Onlyvery few stated that errors should be corrected always (7%). A completely differentsituation occurred with highly-experienced teachers, who uniformly responded thaterrors in speaking should be corrected always. Although this is against the strategies ofcommunicative approach, such error treatment might be more in keeping with students’calls for more correction in speaking. Nevertheless, teachers should not over-correct,since they might discourage their learners from speaking at all.Question No. 7 was answered similarly irrespective of gender or age. Mostteachers chose the first option, stating that their students usually accept their signal andself-correct (see Fig. 10). The remaining options were represented more or less equally.Most teachers left the options of either getting out of balance or frustrated as the leastpossible ones, which is in keeping with the students’ responses – among the respondentsto this survey, 92% of students stated having no problems with correction at all. It isthus clear that students do not have problems with being corrected in general and thatteachers are aware of it. Arguments of theorists such as Krashen who oppose correctiondue to its harmful and frustrating effects on students are thus irrelevant and invalid here.

107Fig. 10: Teachers’ overall responses to Question 7When teachers were asked whether they always know how to correct, themajority of them (almost 70%) – not surprisingly – stated that they are sometimeshesitant about the right feedback on errors. 30% of teachers claimed no problems withcorrection. Interesting differences are found when we compare responses of differentages and genders.Men turned out to be more generally sure about how to handle correction, sincemore than 40% responded they always knew how to deal with an error. On the contrary,only 25% of women claimed absolute certainty in deciding on error treatment.Not surprisingly, the number of teachers being certain about how to deal witherrors rises with age. While almost 78% of young teachers admit struggling with findingthe best way to correct, it is only 64% of teachers from the middle category, and only33% among the most experienced ones. The results thus strongly suggest that it ispossible to learn how to correct best over time and teachers find their optimal teachingstrategies during the years of their practice.

108Question nine was the same for teachers and students. Although teachersanswered similarly regardless of age and gender, their opinions generally differed fromthose of students. In the case of teachers, the vast majority believes that errors arenormal (36%), which is indeed a healthy approach to errors. The fact that students arethe core source of errors was selected by 17% of respondents only. Learners, on theother hand, saw themselves as the main source of error and the fact that errors arenormal was selected by 20% only. For the learners’ sake, it would indeed be beneficialif a higher number of learners realized that errors are an inevitable part of the learningprocess and that it is not possible to avoid them completely. Such knowledge mightlower their inhibitions to talk, raise their self-confidence and thus support the learningprocess. Further research into learners’ reasons for seeing themselves as the mainsources of error may thus be needed.Fig. 11: Teachers’ overall responses to Question 9

109The last cloze question in the teachers’ questionnaire was again identical withthe one in the learners’ survey. Similarly to students, there were also different opinionsbetween teachers of different gender regarding what an error is. Overall, teachersshowed much greater preference for the communicative approach towards error thanstudents, when 66.7% selected option four (see Fig. 53). Surprisingly, no one definederror as anything in conflict with rules of British or American English. The 10% ofrespondents who selected Other referred to error as something relative, what depends onthe level students have achieved, the student himself or herself, the type of activity, thetype of error and many other factors.The differences between men and women were not as significant here as in otherquestions. As in the case of students, more women than men saw error as anythingpreventing understanding or successful communication. Similarly, some women alsodefined error against an expected reaction. Men again showed to be more liberal in theirperception of error. As in Question 6, more men than women saw error as a complexphenomenon and selected the option Other, where they then stated one of the givendefinitions mentioned above.Fig. 12: Teachers’ overall responses to Question 10

110The last question of the survey was the only open question, where respondentswere asked to comment or add anything. Most teachers complained that thequestionnaire is too limiting to encompass such a broad and complex topic. Suchopinion was very valuable to me, since I could see that the respondents really thinkabout correction and errors and do not correct unthinkingly. Many respondents statedthat correction depends on the personality of the learner, the lesson or stage aim, andalso the student’s achieved level of English – one can thus never generalize what is anerror, since it is very relative. While some claimed that students require morecorrection, others stated that teachers often tend to over-correct. Often teachers notedthat error and correction are highly important topics, which need more space inteachers’ books or methodological guide books for teachers. One teacher suggestedconsidering the question of writing the incorrect forms on the board, which wasunfortunately not included in the survey. This will thus remain for future research to beanswered.The survey showed that a large number of teachers see both error and errorcorrection as a complex phenomenon, which is not easy to define or tackle. A majorityof respondents showed an inclination towards views of the communicative approach toteaching. Most teachers responded that they adapt error correction to the type ofactivity, the type of error and the student. A majority of respondents prefer elicitationtype of correction, and also favour self-correction. Most of them stated preference for asensitive way of correcting rather than abrupt interrupting of students in a mid-sentence.Teachers generally saw errors as an inevitable part of learning, rather than a problemthat needs to be eradicated, which was a positive sign, too. Overall, the research showedthat the respondents correct their students sensitively and wisely.

1117. CONCLUSIONThe aim of this diploma thesis was to present and analyse error and errorcorrection in oral practice in English lessons. Theoretical findings were then comparedwith responses of teachers and students in my research.It has been shown that each dominant methodological approach since the mid-20th century had a different perspective on error and correction. Overall, the notion oferror gradually changed from efforts to completely repress and eradicate error bybehaviourists, over theoretical predictions by contrastive analysts and error analysts, tofinally accepting error as an inseparable part of learners’ interlanguage development bythe communicative language approach.The thesis has also shown the differing position of error and correction in firstand second language acquisition. Regarding error, it was shown that first languageacquisition theory helped SLA theory in that it saw errors as learning steps – as anatural part of the learning process. It has been pointed out that teachers’ positiveapproach to error as to something showing learners’ willingness to experiment and becreative with language is crucial for the learners’ natural development of interlanguageand openness towards communication. With respect to correction, it has been observedthat language classes could become more real-life like if more space was devoted toself-correction, which is the typical case in natural conversations.The paper has also shown that it is rather problematic to define error, since thenotion of accuracy, correctness or native-speaker norm is rather vague. It has beenpointed out that it is necessary to distinguish between error and mistake due to theirdiffering nature, which consequently influences decisions on correction: while teachersshould consider correction of errors, it is not the case with mistakes, which are onlymomentary lapses of memory or tongue. It has been observed that the possible sources

112of error are numerous and that only very few of them can be limited by teachers;attentive correction is thus even more important, since it is hardly possible to preventerrors from happening. Furthermore, it has been argued that it is necessary for teachersto be aware of different types of error, since it heavily influences the way they shouldcorrect, too. Global errors, influencing the whole utterance and blockingcommunication, are indeed more probable candidates for correction than for exampleminor grammatical errors. In the era of communicative approach to teaching, it ismainly the aspect of successful communication that influences decisions whether tocorrect individual errors or not.It has been shown that error correction is one type of error feedback that learnerscan receive on their errors only, and that it is not necessary – or even desirable – tocorrect every error that occurs. It has also been observed that correction is a verycomplex issue which includes several decisions teachers have to make before actuallycarrying out any correction as such. Teachers should try to find the right balancebetween over-correcting and non-correcting; correction in general, however, is alwaysmore effective than no correction at all. Overall, it has been pointed out that teachersshould mainly correct errors preventing successful communication; the way they shouldcorrect should not be threatening and should try to fit the learners’ needs. All correctionshould be done with the specific learner and situation in mind. Teachers should alsogive enough space to self-correction, since it supports the learning process the most. Ithas been stressed that the most effective ways of correcting are explicit, outputpromptingstrategies (elicitation, paralinguistic feedback and others). Most importantly,correction should be perceived as a means of helping learners rather than criticisingtheir performance.

113The results of the survey have shown that most teachers’ opinions are in keepingwith the laws of CLT. Majority of teachers see correction as a complex phenomenonand adapt their corrective strategies to numerous aspects, such as learners’ individualneeds, level of English, anxiety, situation, type of activity and many others. Teachershave shown they prefer to correct sensitively and with regard to individual learners.They also see error as an inevitable part of learning rather than as something harmful. Itwas interesting to see that most of their opinions were similar to those of students’, whostated preference for the same types of correction – self-correction and explicitcorrection. Students have also stated that they want to be corrected and believe thatvarieties of English – such as both British and American – should be more widelyaccepted in schools. In some cases, such as the amount of correction students usuallyget or reasons for making error, students’ and teachers’ responses greatly varied.Further classroom research into such issues would be needed in order to get somereliable results. Overall, both students and teachers showed great tolerance towardserror and largely inclined towards the communicative approach to error and correctionand showed that they know what is beneficial for the process of learning.

1148. BIBLIOGRAPHYAllwright, D., & Bailey, K. M. (1991).Focus on the language classroom. Cambridge:CUP.Bailey, K. (1985). Classroom-centred research on language teaching and learning. InCelce-Murcia, M. Beyond basics: Issues and research in TESOL. Rowley, MA:Newbury House.Barron , F. (1957). Originality in relation to personality and intellect. Journal ofPersonality, 25, 730-742.Bartram, M. & Walton, R. (1991).Correction. A positive approach to language mistakes.Hove: LTP.Belbin, E. & Downs, S. & Moore, B. (1970).Unlearning and its relationship to age.Ergonomics, 13, 419-427.Broughton, G. & Brumfit, Ch. & Pincas, A. & Wilde, R. D. (2003). Teaching English asa foreign language. New York: Routledge.Brown, H. D. (2007). Teaching by principles. White Plains: Pearson Education.Burt, M. & Kiparsky, C. (1972).The Gooficon: A Repair Manual for English. Rowley,MA.: Newbury House.Burt, M. & Kiparsky, C. (1974).Global and local mistakes. In Schumann, J. &Stenson, N. (Eds.)New frontiers in second language learning. Rowley, MA.: Newbury House.Burt, M. (1975).Error analysis in the adult EFL classroom. TESOL Quarterly, 9, 53-63.

115Carroll, S., & Swain, M . (1993). Explicit and implicit negative feedback: An empiricalstudy of the learning of linguistic generalization. Studies in Second LanguageAcquisition, 15, 357-386.Cathcart R. L. & Olsen, J. (1976).Teachers’ and students’ preferences for correction ofclassroom conversation errors. In Fanselow, J. F. & Crymes, R. (eds.) On TESOL‘76.Washington, D.C.: TESOL.Choděra, R. (2006). Didaktika cizích jazyků. Praha: Academia.Christison, M. A. & Krahnke, K. J. (1983).Recent Language Research and SomeLanguage Teaching Principles. TESOL Quarterly, 17 (4), 625-649.Chun, A.E., Day, R.R., Chenoweth, N.A., & Luppescu, S. (1982). Errors, interaction,and correction: A Study of native-non-native conversations. TESOL Quarterly 16(4), 537-547.Corder, S. P. (1984).The significance of learners’ errors. In Richards, J. C. (Ed.) Erroranalysis. Perspectives on second language acquisition. (pp. 19-30). Harlow:Longman.Corder, S. P. (1984).Idiosyncratic Dialects and Error Analysis. In Richards, J. C. (Ed.)Error analysis. Perspectives on second language acquisition. (pp. 158-171).Harlow: Longman.Corder, S. P. (1981).Error analysis and interlanguage. Oxford: OUP.Doff, A. (1993). Teach English. Cambridge: Cambridge University Press.Doughty, C. (2001). Cognitive underpinnings of focus on form. In P. Robinson (Ed.)Cognition and second language instruction (pp. 206-57). Cambridge: CambridgeUniversity Press.

116Dulay, H. & Burt, M. (1974).Natural sequences in child’s second language acquisition.Language learning, 24 (1), 37-53.Dulay, H. & Burt, M. (1984). You can’t learn without goofing. In: Richards, J. C. (Ed.)Error analysis. Perspectives on second language acquisition. (pp. 95-123).Harlow: Longman.Dulay, H.C., Burt, M.K. & Krashen, S. (1982). Language Two. Rowley, MA.: NewburyHouse.Dušková, L. (1969). On sources of errors in foreign language learning. InternationalReview of Applied Linguistics. 7 (1), 11-36.Edge, J. (1989).Mistakes and correction. Harlow: Longman.Egi , T. (2007). Interpreting recasts as linguistic evidence: The role of linguistic target,length, and degree of change. Studies in Second Language Acquisition, 29, 511 –537.Ellis, R. (1985). Understanding second language acquisition. Oxford: OUP.Ellis, R. (1998). Teaching and Research: Options in Grammar Teaching. TESOLQuarterly, 32 (1), 39-60.Ellis, R. (2009). Corrective Feedback and Teacher Development.L2 Journal, 1 (1), 3-18. Retrieved fromhttp://www.escholarship.org/uc/item/2504d6w3Ellis, R. (2010). Epilogue – A Framework for Investigating Oral and Written CorrectiveFeedback. Studies in Second Language Acquisition,32, 335– 349.doi:10.1017/S0272263109990544

117Ellis, R., Loewen, S., & Erlam, R. (2006).Implicit and explicit corrective feedback andthe acquisition of L2 grammar. Studies in Second Language Acquisition, 28, 339-368.Error and Mistake.(n.d.).In Cambridge Dictionary Online. Retrieved fromhttp://dictionary.cambridge.org/Error and Mistake.(n.d.).In Macmillan Dictionary Online. Retrieved fromhttp://www.macmillandictionary.com/Error and Mistake.(n.d.).In Oxford Dictionary Online. Retrieved fromhttp://oxforddictionaries.com/Faerch, C., Haastrup, K., Phillipson, R. (1984). Learner language and languagelearning. Clevedon, Avon: Multilingual Matters.Green, P.S. & Hecht, K. (1985). Native and non-native evaluation of learners' errorsin written discourse. System, 13 (2), 77-97.Harmer, P. (2005). How and when should teachers correct? Research News: TheNewsletter of the IATEFL Research SIG, 15, 38-39.Harmer, J. (2007).How to teach English.Harlow: Pearson Education Limited: Longman.Hendrich, J. et al. (1988). Didaktika cizích jazyků. Praha: SPN.Hendrickson, J. (1978). Error correction in foreign language teaching: Recent theory,research, and practice. Modern Language Journal. Retrieved from:http://onlinelibrary.wiley.com/doi/10.1111/j.1540-4781.1978.tb02409.x/abstractHolley, F. M. & King, J.K. (1974).Imitation and correction in foreign language learning.New Frontiers in Second Language Learning. Newbury House.Hubbard et al. (1983).A training course for TEFL. Oxford: Oxford University Press.

118Iwashita, N . (2003). Negative feedback and positive evidence in task-based interaction:Differential effects on L2 development. Studies in Second Language Acquisition,25,1-36.James, C. (1998). Errors in language learning and use – exploring error analysis.Edinburgh: Harlow.Jefferson, G., Sacks, H., Schegloff E. A. (1977). The Preference for Self-Correction inthe Organization of Repair in Conversation. Language, 53 (2), 361-382. Retrievedfromhttp://www.jstor.org/stable/413107Jenkins, J. (2006). Current Perspectives on Teaching World Englishes and English as aLingua Franca. TESOL Quarterly, 40 (1), 157-181.Kay, H. (1951). Learning of a serial task by different age groups. Quarterly journal ofexperimental psychology, 3, 166-183.Korčáková, J. (2004). Chyba a učení cizím jazykům. Hradec Králové: Gaudeamus.Krashen, S. (1982 & 2009). Principles and practice in second language acquisition.Oxford: Pergamon Press.Krashen, S. D. & Terrell, T. D. (1983 & 1995). The natural approach: Languageacquisition in the classroom. Hernerl Hempstead: Phoenix ELT.Kulič, V. (1971).Chyba a učení. Praha: SPN.Larsen-Freeman, D. & Long, M.H. (1991). The linguistic environment for languageacquisition. In D. Larsen-Freeman & M.H. Long (Eds.), An Introduction toSecond Language Acquisition Research. Harlow, Essex: Longman.

119Lennon, P. (1991). Error: Some problems of definition, identification, and distinction.Applied Linguistics, 12 (2): 180-196.Lightbown, P. & Spada, N. (1990). Focus-on-form and corrective feedback incommunicative language teaching: Effects on second language learning. Studiesin second language acquisition, 12, 429-448.Littlewood, W. (1994).Communicative language teaching. Cambridge: CUP.Lyster, R. & Ranta, L. (1997). Corrective feedback and learner uptake: Negotiation ofform in communicative classrooms. Studies in Second Language Acquisition, 19,37-66.MacDonald Lightbound, P. (2005).An analysis of interlanguage errors in synchronous /asynchronous/ intercultural communication exchanges [Pdf file]. (Doctoraldissertation, Universitat de Valencia, 2011). Retrieved January 3, 2013 fromhttp://tdx.cat/handle/10803/9786Mackey, A., & Philp, J. (1998). Conversational interaction and second languagedevelopment: Recasts, responses, and red herrings? Modern Language Journal,82, 338-356.Meyer, L. A. (1986). Strategies for Correcting Students' Wrong Responses. TheElementary School Journal, 87 (2), 227-241.Nemser, W. (1984). Approximative systems of foreign language learners. In Richards, J.C. (Ed.).Error analysis. Perspectives on second language acquisition. (pp. 55-62).Harlow: Longman.Nickel, G. (1981).Aspects of error analysis. AILA Bulletin, 29 (1), 1-27.

120Nunan, D. (1991). The teacher as decision-maker. Paper presented at IATEFLconference, Exeter.Philp, J. (2003). Constraints on 'noticing the gap': Non-native speakers' noticing ofrecasts in NS-NNS interaction. Studies in Second Language Acquisition, 25, 99-120.Politzer, R. & Ramírez, A. (1973).An error analysis of the spoken English of MexicanAmericanpupils in a bilingual school and a monolingual school. LanguageLearning, 18, 35-53.Richards, J. C. (Ed.). (1984). Error analysis. Perspectives on second languageacquisition. Harlow: Longman.Richards, J. C. & Sampson, G. P. (1984).The study of learner English. In Richards, J. C.(Ed.).Error analysis. Perspectives on second language acquisition. (pp. 4-12).Harlow: Longman.Richards, J. C. & Renandya, Willy A. (Ed.). (2002). Methodology in languageteaching : an anthology of current practice. Cambridge: Cambridge UniversityPress.Russel, J., Spada, N. (2006). The effectiveness of corrective feedback for the acquisitionof L2 grammar. In Norris & Ortega (2006) (Ed.) Synthesizing research onlanguage learning and teaching. Hawai: John Benjamins.Schegloff, E. A., Jefferson, G., Sacks, H. (1977). “The preference for self-correction inthe organization of repair in conversation”. Language, 53, 361-82.Schwartz, B. (1993). On explicit and negative data effecting and affecting competenceand linguistic behavior. Studies in Second Language Acquisition, 15, 143-14.Scrivener, J. (2005).Learning teaching. Oxford: Macmillan.

121Seidlhofer, B. (2004). Research perspectives on teaching English as a lingua franca.Annual Review of Applied Linguistics, 24, 209-239.Seidlhofer, B. (2010). Understanding English as a lingua franca. Oxford: OUP.Selinker, L. (1984). Interlanguage. In Richards, J. C. (Ed.).Error analysis. Perspectiveson second language acquisition. (pp. 31-54). Harlow: Longman.Shastri, P. D. (2010). Communicative approach to the teaching of English as a secondlanguage. Mumbai: Himalaya Publishing House.Singer, R. N. (1977). To Err or Not to Err: A Question for the Instruction ofPsychomotor Skills. Review of Educational Research, 47 (3), 479-498. Retrievedfromhttp://www.jstor.org/stable/1170116Spratt, M. et al. (2010).The TKT course. Cambridge: CUP.Truscott, J. (1996). The case against grammar correction in L2 writing classes.Language Learning, 46 (2), 327-369.Truscott, J. (1999). What's wrong with oral grammar correction? The Canadian ModernLanguage Review, 55(4), 437-56.Truscott, J. (2007). The effect of error correction on learners’ ability to write accurately.Journal of Second Language Writing, 16(4), 255-272.Ur, P. (2012). Accuracy and error correction [Powerpoint slides].Autumn teachers’conference, Brno.Valette, R. M. (1973). Developing and Evaluating Communication Skills in theClassroom. TESOL Quarterly, 7 (4), 407-424. Retrieved fromhttp://www.jstor.org/stable/3585871

122Warsi, J. (n.d.). A review of Selinker’s rediscovering interlanguage [Unpublishedreview of the book Rediscovering interlanguage (Applied linguistics and languagestudy)]. Retrieved January 5, 2013 fromhttp://jilaniwarsi.tripod.com/selinker.pdfYang, W. (May 2010). A Tentative Analysis of Errors in Language Learning and Use.Journal of Language Teaching and Research, 1 (3), 266-268.doi:10.4304/jltr.1.3.266-268

1239. APPENDICES9.1. Appendix 1 – Student questionnaireError correction in spoken practice in English classes – Student questionnaireBy filling in the questionnaire you help to monitor students’ opinions on errorcorrection in speaking activities. The survey was created to find out students’ opinionsand be able to compare them with what teachers think. Ideally, the results should helpteachers in deciding about error correction.The results of the survey should primarily help future teachers in deciding on errorcorrection, but could be to some use to practicing teachers, too.Filling the questionnaire takes less than 5 minutes. Your opinion is extremely useful tome, other teacher trainees, as well as students of English in general.Thank you in advance!Gabriela TomkováIf you want to be informed about the results of the survey, fill in your contact emailaddress at the end of the survey.Please enter your information:GenderAgeYears of studying English:1. When I make an error during speaking, my teacher’s typical reaction is:a) correctionb) no correctionc) sometimes correction, sometimes no reactiond) no correction, other students correct mee) no correction, I must self-correct2. My teacher corrects my errors:a) too oftenb) adequatelyc) only rarelyd) never3. If my teacher corrects my error, it is typically (order from the least = 1 to themost frequent = 5):pronunciation

124vocabularygrammarword orderwhole sentence/utterance4. In my opinion, my teacher corrects errors...:a) more or less the same with all activitiesb) sometimes less, sometimes more - depends on the activityc) errors of some students less and of others more, regardless of the activityd) Other (please specify):5. When I realize I have made an error, my typical reaction is:a) nothing at all, I continue speaking, I cannot be distracted from the thoughtb) I admit the error, think about it, and then continuec) I ask my teacher about the error and the correct solutiond) I am frustrated because of it and do not want to go on speakinge) I get out of balance so much that I forget what I was sayingf) Other (please specify):6. I would appreciate if my teacher...:a) kept correcting the way (s)he doesb) changed her/his way of correcting (how?):7. When I make an error during speaking, it is very good / good / not very good /bad when my teacher ...(select the most appropriate option for each):VerygoodGoodNotverygoodBad…does not correct me at all.…tells me that I have made an error but I mustself-correct.…tells me about the error and corrects me.…tells me the correct form and lets me repeat it.…lets other students correct my error.…corrects my error and explains what was wrongand why.

1258. In comparison with students of the same level of English, I make errors...:a) more oftenb) similarly frequentlyc) less often9. I believe that the most common sources of errors in speaking are (order fromthe crucial ones = 1 to the least important ones = 5)a) unsatisfactory course bookb) unsatisfactory teaching methodsc) the source is the student (insufficient preparation, lack of concentration etc.)d) neither, errors are normal, everybody makes theme) we have too few opportunities to talk10. In my opinion, an error is:a) anything in conflict with an expected reactionb) anything not included in rules of British/American Englishc) anything that a native speaker would not say (=slang, informal words etc are notconsidered as error)d) anything preventing understanding the sense or successful communicatione) Other definition (please specify):11. Is there anything I would change about error correction? Do I have anythinginteresting to add or comment on?

1269.2. Appendix 2 – Teacher QuestionnaireError correction in spoken practice in English classes - Teacher questionnaireBy filling in the questionnaire you help to monitor teachers’ opinions on errorcorrection in speaking activities. The survey was created to find out teachers’ opinionsand be able to compare them with what students think. Ideally, the results should helpteachers in deciding about error correction.The results of the survey should primarily help future teachers in deciding on errorcorrection, but could be to some use to practicing teachers, too.Filling the questionnaire takes less than 5 minutes. Your opinion is extremely useful tome, other teacher trainees, as well as students of English in general.Thank you in advance!Gabriela TomkováIf you want to be informed about the results of the survey, fill in your contact emailaddress at the end of the survey.Please enter your information:GenderAgeYears of teaching:1. My typical reaction to a student’s error in speaking is:correctionno correctionit depends on the type of activityI let the student correct him/herselfI let other students to correct the error2. If I decide to correct an error, the most frequent type of error is (order from themost frequent = 1 to the least ones = 6):1 2 3 4 5 6pronunciation, intonation, rhythmindividual wordsgrammarword orderthe whole sentence/utterance

127style and register (=acceptability in the given situation)3. When I realize a student made an error in speaking, my typical reaction is(order from the most frequent = 1 to the least = 5):I do not correct it.I say there was an error but let the student correct him/herself.I say there was an error, correct it, and let the student repeat the correct version.I say there was an error and let other students to correct it.I correct the error and explain what was wrong and why.4. Regarding frequency, I correct student errors in speaking:more or less the same with all activitiessometimes less, sometimes more - depends on the activityerrors of some students less and of others more, regardless of the activityI do not know, I correct automatically5. With respect to timing, I usually correct my students’ errors:immediatelyafter the sentence containing the errorafter the student has stopped talkingat the end of the whole activityat the end of the lesson6. In my opinion, errors in speaking should be corrected:neverin fluency activities onlyin accuracy activities onlyalways, if possibleOther (please specify):7. The most common reaction of my students to my signalling of an error inspeaking is (order from the most common = 1 to the least = 5):nothing at all, they continue speaking, cannot be distracted from the thoughtthey accept my signal, think about it, correct the error and go on talkingthey are unable to self-correct, it is necessary to interrupt them and discuss the errorthey are frustrated because of the error and unwilling to continue talking

128they get out of balance so much that they forget what they were saying8. Considering error correction in general, I can say that:I always know how to deal with an errorI am sometimes hesitant whether to correct or not, and if so, howI often experience trouble with error correction, as I am worried about how mystudents react to itI do not correct errors; it deprives my students too much9. I believe that the most common sources of errors in speaking are (order fromthe crucial ones = 1 to the least important ones = 5)unsatisfactory course bookunsatisfactory teaching methodsneither, errors are normal, everybody makes themthe source is the students (insufficient preparation, lack of concentration etc.)students have too few opportunities to talk10. In my opinion, an error is:anything in conflict with an expected reactionanything not included in rules of British/American Englishanything that a native speaker would not say (=slang, informal words etc are notconsidered as error)anything preventing understanding the sense or successful communicationOther definition (specify):11. Is there anything I would change about error correction? Do I have anythinginteresting to add or comment on?

1299.3. Appendix 3 – Students’ responsesFig. 1: Students’ responses to Question 1Fig. 2: Students’ overall responses to Question 2Fig. 3: Question 2 – Lower secondary school students

130Fig. 4: Question 2 – Higher secondary school studentsFig. 5: Students’ responses to Question 3

131Fig. 6: Students’ responses to Question 4Fig. 7: Students’ overall responses to Question 5Fig. 8: Boys’ responses to Question 5

132Fig. 9: Girls’ responses to Question 5Fig. 10: Younger students’ responses to Question 5Fig. 11: Older students’ responses to Question 5

133Fig. 12: Students’ responses to Question 6Fig. 13: Students’ responses to Question 7

134Fig. 14: Students’ overall responses to Question 8Fig. 15: Boys’ responses to Question 8Fig. 16: Girls’ responses to Question 8

135Fig. 17: Students’ responses to Question 9Fig. 18: Students’ overall responses to Question 10

136Fig. 19: Boys’ responses to Question 10Fig. 20: Girls’ responses to Question 10Fig. 21: Younger learners’ responses to Question 10Fig. 22: Older learners’ responses to Question 10

1379.4. Appendix 4 – Teachers’ responsesFig. 23: Teachers’ overall responses to Question 1Fig. 24: Men’s responses to Question 1Fig. 25: Women’s responses to Question 1

138Fig. 26: Young teachers’ responses to Question 1Fig. 27: Medium-experienced teachers’ responses to Question 1Fig. 28: Highly-experienced teachers’ responses to Question 1

139Fig. 29: Teachers’ overall responses to Question 2

140Fig. 30: Teachers’ overall responses to Question 3Fig. 31: Teachers’ overall responses to Question 4Fig. 32: Men’s responses to Question 4

141Fig. 33: Women’s responses to Question 4Fig. 34: Teachers’ overall responses to Question 5Fig. 35: Men’s responses to Question 5

142Fig. 36: Women’s responses to Question 5Fig. 37: Young teachers’ responses to Question 5Fig. 38: Medium-experienced teachers’ responses to Question 5

143Fig. 39: Highly-experienced teachers’ responses to Question 5Fig. 40: Teachers’ overall responses to Question 6Fig. 41: Men’s responses to Question 6

144Fig. 42: Women’s responses to Question 6Fig. 43: Young teachers’ responses to Question 6Fig. 44: Medium-experienced teachers’ responses to Question 6

145Fig. 45: Highly-experienced teachers’ responses to Question 6Fig. 46: Teachers’ overall responses to Question 7Fig. 47: Teachers’ overall responses to Question 8

146Fig. 48: Men’s responses to Question 8Fig. 49: Women’s responses to Question 8Fig. 50: Young teachers’ responses to Question 8

147Fig. 51: Middle-experienced teachers’ responses to Question 8Fig. 52: Highly-experienced teachers’ responses to Question 8

148Fig. 53: Teachers’ overall responses to Question 9Fig. 54: Teachers’ overall responses to Question 10Fig. 55: Women’s responses to Question 10Fig. 56: Men’s responses to Question 10

1499.5. Appendix 5 – List of Abbreviations(in order of appearance)CA – contrastive analysisSLA – second language acquisitionL2 – second languageL1 – first languageEA – error analysisCLT – communicative language teachingTL – target languageMT – mother tongueSLL – second language learningFLA – first language acquisitionIL – interlanguage

15010. SummaryThis diploma thesis deals with error correction in oral practice in English languageclasses. It presents and evaluates theoretical findings on error in language learning –particularly in second language acquisition – and the subsequent correction of errors madeby learners in speaking. It compares these observations with the results of an onlineresearch among teachers and students of English in Brno carried out in February and March2013.The thesis mainly focuses on error and correction as perceived from the perspective ofthe communicative approach to language teaching; however, it introduces contraryapproaches to correction (such as Stephen Krashen’s) and evaluates them, too. Afterintroducing different perspectives on error and correction in the most prominentmethodological approaches since the mid 20th century until today (behaviourism,contrastive analysis, error analysis, Krashen’s natural approach, and communicativelanguage teaching), the thesis discusses the position of error and correction in secondlanguage acquisition compared to the position in first language acquisition and nativespeech. It points out similarities and differences between them and stresses the importanceof the comparison for second language theory. The thesis distinguishes error from mistakeand presents the main sources and types of error. It also discusses the importance andeffectiveness of correction and presents all the complex decisions leading to a successfulcorrection – that is, deciding on whether to correct, which errors to correct, when to correct,how to correct, and who should do the correction. In doing so, it tries to answer thequestion whether there is a universal ideal way of correcting learners’ errors and if so, whatit is. Besides, it presents and evaluates individual ways of indicating errors that teachers useto signal that an error has been made. In the final part, the author presents results from theresearch among teachers and students of English, analyses them and compares them withthe theoretical background. Suggestions for further research into classroom practice are alsogiven.

15111. ResuméTato diplomová práce se zabývá opravováním chyb při mluvení studentův hodinách angličtiny. Práce představuje teoretické poznatky o chybě v učení jazyka –převážně cizích jazyků – a následně i opravování chyb, kterých se žáci při mluvenídopouští. Tyto poznatky jsou srovnány s výsledky internetového výzkumu mezi učitelia studenty anglického jazyka v Brně, který probíhal v únoru a březnu roku 2013. Prácese zaměřuje na chyby a opravování chyb z hlediska komunikativního přístupu k výucejazyka; představuje však i názory protichůdných přístupů k opravování chyb (napříkladStephena Krashena) a hodnotí i je. Po úvodním představení jednotlivých pohledůnejdůležitějších metodologických teorií druhé poloviny 20. století a dneška(behaviorismu, srovnávací analýzy, analýzy chyb, Krashenovy teorie Natural Approacha komunikativního učení jazyka) na opravování chyb se práce zaměřuje na pozici chybya jejího opravování v teorii akvizice cizího jazyka v porovnání s pozicí v teorii akvizicemateřského jazyka a konverzacích rodilých mluvčích. Práce poukazuje na rozdíly ashody mezi výše zmíněnými a zdůrazňuje přínos takového srovnání pro teorii akvizicecizího jazyka. Práce představuje hlavní příčiny a typy chyb a rozlišuje mezi termínemerror a mistake. Práce také projednává důležitost a účinnost korekce chyb a představujevšechna komplexní rozhodnutí, která ke korekci vedou – zda chybu opravit, jak jiopravit, kdy chybu opravit, kdo by ji měl opravit a jakým způsobem. Snaží se takzodpovědět otázku, zda existuje univerzální ideální způsob na opravování chyb, a pokudano, tak jaký. Dále práce představuje a hodnotí jednotlivé způsoby signalizováníchybných výpovědí. V závěrečné části autorka představuje výsledky výzkumu meziučiteli a studenty anglického jazyka, analyzuje je a srovnává s teoretickými poznatky.Navrhuje také podněty pro další výzkum v této oblasti, který je třeba pro hlubšípochopení vnímání chyb a opravování mezi učiteli a studenty.