Navigation – Plan du site

Éditions littéraires et linguistiques de l'université de Grenoble

AccueilNuméros61VariaContrastive vs Non-Contrastive Me...

Varia

Contrastive vs Non-Contrastive Meta-Phonetic Input in Teaching Foreign Language Pronunciation

Apport de l’enseignement métaphonétique contrastif ou non contrastif dans l’enseignement de la prononciation des langues étrangères
Zdena Kralova, Katarina Nemcokova et Jana Birova

Résumés

Presque toutes les études sur l’enseignement métaphonétique explicite de la prononciation des langues étrangères (L2) confirment son effet facilitateur. Cependant, jusqu’à présent, il n’existe pratiquement aucune étude sur l’enseignement métaphonétique contrastif (L1-L2). La présente étude examine l’efficacité de ce type d’enseignement par rapport à l’efficacité de l’enseignement métaphonétique non contrastif (L2), en analysant expérimentalement la prononciation anglaise de 80 adultes slovaques. La qualité de leur prononciation anglaise, telle qu’elle se reflète dans la structure des voyelles, est mesurée avant et après l’input contrastif dans le groupe expérimental et l’input non contrastif dans le groupe contrôle. Les valeurs sont ensuite comparées aux valeurs standard des voyelles de l’anglais britannique. On fait l’hypothèse que les voyelles anglaises produites par le groupe expérimental se rapprochent davantage des valeurs standard que les voyelles produites par le groupe contrôle. Les résultats montrent une approximation plus juste dans le groupe expérimental, ce qui indique une plus grande efficacité de l’entrée métaphonétique contrastive dans l’enseignement de la prononciation des langues étrangères.

Haut de page

Texte intégral

1. Introduction

1Although the myth of the native speaker as an ideal foreign language speaker has already been deconstructed (Benke & Medgyes, 2005; Moussu, 2006), their pronunciation is often perceived as a model in foreign language communication by non-native speakers and applied as a reference standard in related research. What is more, native speakers are the evaluators of non-native speakers pronunciation in most studies on foreign accent (e.g. Anderson-Hsieh, Johnson & Koehler, 1992; Carmichael, 2000; Flege, 1988; Flege & Hillenbrandt, 1984; Munro & Derwing, 2001; Sheppard, Hayashi & Ohmori, 2007). The same was true in foreign language teaching methodology, where learners’ errors used to be considered in a negative light, and foreign language sounds were evaluated either as correctly or incorrectly produced. Today, the communicative value of foreign language pronunciation is emphasized and the terms “deviation” or “approximation” are regarded as more adequate than “error”.

2According to interlanguage theory, foreign language (L2) learning is a process of autonomous code formation that gradually approximates to the L2 quality (Peltola, Rautaoja, Alku & Peltola, 2017). Despite the upper-limit theory of approximation to foreign language pronunciation (Flege & Hillenbrandt, 1984), several researchers (e.g. Dickerson, 1974) posit the continuous improvement of non-native speakers’ pronunciation. For Dickerson (1974), the first elements to be eliminated in this process are the most obvious pronunciation deviations, while closer approximations are typically more persistent. On the other hand, according to Weinreich’s (1953) concept of language interference, approximation is significantly aided by interlingual identification of elements. Several longitudinal experiments (e.g. Meador, Flege & MacKay, 2000), have shown that the most obvious deviations in foreign language pronunciation were eliminated after phonetic training, while closer (yet phonetically still imprecise) approximations of L2 sounds tended to persist. Minor deviations have more significant persistence, although they have less impact on speech comprehensibility (Kralova, 2016).

3In learning foreign language pronunciation, learners have to acquire and automatize a complex set of articulatory gestures or modify the existing articulatory models, with each individual using their own strategies. A high level of automatization of L2 pronunciation is necessary for effective and economical oral communication. When teaching an L2, it is important to realize that auditory-articulatory engrams are not innate and that the only way of creating or storing new memory engrams (both receptive and productive) in the human brain is repeated reception and production (Malikova, 1993). As learners become aware of the differences between their own output and an authentic output in the L2, they may attempt to modify their articulation strategy.

4The training of phonematic hearing and modification of a learner’s foreign language perception base are important steps that should precede the practice of articulation (Chebenova, 2001). Following the principles of language ontogenesis and phylogenesis, the receptive phase of training should be followed by the productive phase aimed at drill, fixation and automatization of articulatory gestures and the creation of dynamic articulatory stereotypes. At the same time, it is useful for adult learners to be aware of the differences between phonetic-phonological norms of the native and the foreign languages (Cummins, 2005). For them, conscious practice is more effective than intuitive-imitative practice, and several studies (e.g. Kralova, 2011) confirm the benefits of practical phonetic training combined with adequate theoretical information.

5Nonetheless, there is some disagreement about the extent to which it is necessary for learners to have meta-linguistic knowledge of the given system. Some authors (e.g. Peltola, Rautaoja, Alku & Peltola, 2017) assume that the automatic processing of language phenomena does not require it, while others (e.g. Carmichael, 2000) suggest that meta-linguistic context has a facilitative effect with adult learners of a foreign language. This is partially due to the fact that the conceptual-abstract memory develops intensively with age. Therefore, the process of acquiring new habits and skills in adult learners should include adequate theoretical information. When learners are cognitively mature for explicit teaching, it can significantly accelerate the whole process of learning (Wrembel, 2005), as new temporary links from the kinaesthetic analyser (Kralova, 2011) created by the learner’s own activity become subsequently connected to the theoretical system.

6Not many studies on the effectiveness of explicit phonetic-phonological instructions (meta-phonetic input) in L2 pronunciation learning have been undertaken so far, but almost all of them confirmed their positive correlation (e.g. Derwing & Munro, 2005; Kissling, 2013). However, the existing literature does not provide any experimental comparison of the effectiveness of contrastive meta-phonetic input (comparing L1 and L2 phonic systems) with the effectiveness of non-contrastive input (dealing solely with the L2 phonic system).We believe that the comparative analysis of native and foreign language phonic systems (focusing on potential interference phenomena) might be beneficial to adults when learning foreign language pronunciation, because the identification of identity is a guiding principle in foreign language learning (Kralova, 2011).

7Flege, Bohn and Jang (1997) state that cross-linguistic phonetic interference is obvious mainly with vowels and other authors join them to claim that the production of consonants plays a much less significant role in a foreign accent (e.g. Mildner & Horga, 1999). Kralova’s (2011) findings confirm both claims. Moreover, several studies have found that the correct position of vowels in the L2 formant scheme highly positively correlates with an overall level of L2 phonic competence (e.g. Mildner & Horga, 1999; Munro, Derwing & Flege, 1999). Therefore, given that vowels are relatively more variant, some (e.g. Bohn & Munro, 2007) argue that vocalic mistakes have a more significant influence on speech comprehension than consonantal mistakes.

2. Methods

2.1. Objectives

8The primary objective of the study was to compare the extent of qualitative approximation of English short vowels produced by Slovak learners after ten-weeks of meta-phonetic input (theoretical information on the phonic system of a language). The input focused on Slovak-English contrastive phonetics (Kralova, 2011) in the experimental group (EA) and solely on the English phonic system (Roach, 2009) in the control group (KA). Together with the contrastive and non-contrastive meta-phonetic input (45 minutes a week) both groups received identical pronunciation training aimed at the segmental subsystem of the English language (45 minutes a week).

2.2. Participants

9Eighty EFL university students (60 females, 20 males) participated in the experiment. Their average age was 19 years and their native language (L1) was Slovak. Their English grammatical and lexical competence was at B1-B2 level (Allan, 2005). Most of them started learning English at primary school with a non-native teacher of English and had never stayed in an English-speaking country for any length of time. Two quasi-homogeneous groups were created by random sampling on the principle of availability: the experimental group (40 participants) and the control group (40 participants).

2.3. Material

10The primary research material was the audiorecording of participants’ extemporaneous English speech (average length: 3.8 minutes) in the pre-test and in the post-test (after ten weeks) conditions. The topic of their utterances was autobiographical in order to preserve similar vocabulary and style.

2.4. Data Analysis

11The recordings were experimentally analysed in the phonetic laboratory in the Speech Analyser system (version 2.7) which displays the oscillogram, broadband sonogram and LPC spectrum (Figure 1). Seven short English vowels /I/, /e/, /æ/, /Λ/, /D/, /ʊ/, /ə/ were segmented from each recording both in the pre-test and the post-test. The location of the most significant change in amplitude, frequency and shape of the acoustic wave on the oscillogram was determined and the corresponding vowel was manually segmented on the basis of audio-correlational and visual-correlational methods.

12Then, the spectral analysis of the vowel was done and the first (F1) and the second formant (F2) of the vowel was read from the LPC spectrum. The average F1 and F2 values for every vowel were calculated from five different measurements due to high individual and contextual variation of formants. The formant values of the vowels produced by five British English native speakers (A) were used as reference values for the approximation analysis. The standard F1 and F2 values of English short vowels (A0) (Gimson, 1989) and the standard F1 and F2 values of corresponding Slovak short vowels (S0) (Kral & Sabol, 1989) stated in relevant phonetic publications were used as the basis for comparison to balance the potential variance of native speakers production.

13Statistical analyses were carried out using the OpenStat program to identify the relationship between the variables: the experimental analysis of pronunciation (dependent variable) and the contrastive meta-phonetic input (independent variable). The inter-group differences between the pronounced and the referential formant values were analysed in the pre-test and the post-test.

Figure 1. – Experimental analysis.

Figure 1. – Experimental analysis.

2.5. Hypotheses

14The following hypotheses were formulated:

  1. English short vowels produced by the participants in the post-test will approximate to the referential values more than the vowels produced in the pre-test.

  2. English short vowels produced by the experimental group in the post-test will approximate to the referential values more than vowels produced by the control group in the post-test.

3. Results

15Table 1 contains the F1 and F2 values of vowels produced by the experimental group (E1A) and the control group (K1A) in the pre-test and by the experimental group (E2A) and the control group (K2A) in the post-test, the English and Slovak standard formant values (A0, S0) and the reference formant values of vowels produced by British English native speakers (A). The variance for individual formants is in most cases lower than 50%, with the exception of F2 [ə] values in the E1A group (Vx  = 55%).

Table 1. – F1 and F2 mean values.

Group

/I/

/e/

/æ/

/Λ/

/D/

/ʊ/

/ə/

F1

F2

F1

F2

F1

F2

F1

F2

F1

F2

F1

  F2

F1

F2

S0

285

1916

452

1718

700

1510

682

1315

481

1084

326

  967

 

 

A0

360

2220

600

2060

800

1760

760

1320

560

  920

380

  940

560

1480

A

351

2114

699

2021

771

1700

758

1367

600

  970

394

  980

525

1479

E1A

315

1914

499

1818

630

1571

642

1352

491

1116

386

  955

492

1854

E2A

353

2110

644

1966

682

1689

708

1396

542

1091

385

  996

519

1546

K1A

321

1909

473

1801

694

1568

656

1348

482

1099

380

  962

486

1866

K2A

342

2050

569

1927

744

1689

704

1383

525

1067

395

1002

508

1653

16The formant scheme (Figure 2) illustrates a high degree of proximity between the standard formant values (A0) and the reference formant values produced by the five British English native speakers (A), as well as their distance from the standard formant values of Slovak vowels (S0).

Figure 2. – F1 and F2 reference and standard values.

Figure 2. – F1 and F2 reference and standard values.

17The formant scheme (Figure 3) comparing the English formant values produced by the participants in both the experimental and control groups in the pre-test (E1A and K1A) shows their significant deviation from the reference values (A). The formant values are in the positions closer to the Slovak vowels than to the English vowels (cf. Figure 2). On the contrary, the formant scheme (Figure 4) showing the relationship of formant values produced by the participants in both groups in the post-test (E2A and K2A) indicates closer approximation of formant values to the English reference values (A) and increased distance from the Slovak values (S0) (cf. Figure 2). It is thus possible to confirm more significant approximation of formant values in the post-test in comparison with the pre-test.

Figure 3. – F1 and F2 referential and produced values (pre-test).

Figure 3. – F1 and F2 referential and produced values (pre-test).

Figure 4. – F1 and F2 referential and produced values (post-test).

Figure 4. – F1 and F2 referential and produced values (post-test).

18The difference of formant values of the vowels produced by the participants and the reference values were calculated. The degree of approximation not the direction of approximation (positive or negative) was relevant, therefore all values were treated as positive. An overall difference of formant values for each participant and an overall difference of formant values within both groups (in the pre-test and the post-test) were calculated (Table 2).

19The average difference of the produced and the reference values was higher in the pre-test (i.e. the average approximation to the reference values was lower) than in the post-test. The average difference in the post-test was significantly lower (i.e. the approximation was higher) in the experimental group than in the control group.

Table 2. – The difference of the produced and the reference values.

/I/

/e/

/æ/

/Λ/

/D/

/ʊ/

/ə/

Mean

F1

F2

F1

F2

F1

F2

F1

F2

F1

F2

F1

F2

F1

F2

E1A

45

316

210

244

169

183

172

144

134

189

49

133

118

370

176.9

E2A

49

248

101

161

  97

126

108

193

105

150

61

149

  73

141

125.8

K1A

37

289

224

223

  86

173

139

156

127

165

40

98

  98

387

160.2

K2A

40

363

161

190

  67

153

90

208

106

150

45

157

  68

188

141.9

4. Discussion

20The experiment attempted to synthesize a theoretical analysis as well as to establish causal relationships of the explored parameters, using an intentional manipulation of the dependent variable along with the analysis of variable causal relationships. We used a covariance analysis in the experimental plan to measure the dependent variable before and after the experimental intervention. Under the conditions of the current experiment, a complete randomization of participants was not possible, so we applied the principle of availability. The research design can thus be characterized as a quasi-experiment.

21The validity of the measurement should be confirmed by the justified conclusions drawn on the basis of measurement. In order to ensure the internal measurement validity, we used the statistical operations of analysis of variance. We tried to strengthen the external measurement validity, i.e. the possibility of the generalization of results beyond the scope of this experiment, by experimenting in conditions reflecting natural communication. The content validity was derived from the prerequisite that the measurements of vocalic formants represent an overall level of participants pronunciation. The criterion validity was evaluated from the point of view of measurement agreement (experimental analysis) with a criterion variable—standard values of English vocalic formant values. The construct validity was verified with the given theoretical context and the prognostic validity was verified by formulating and verifying the hypotheses.

22One of the primary aims of the research was to compare the effectiveness of contrastive and non-contrastive metaphonetic input in teaching English phonetics and phonology at a Slovak university. In adult learners, an analytical (cognitive) type of pronunciation training is considered to be more effective than an imitative type of training (Chebenova, 2001). It follows that explicit awareness of the differences, similarities and potential possibilities of pronunciation mistakes (or deviations) resulting from the differences between the sound systems of the native and the foreign language would significantly contribute to improving the foreign language phonic performance of an individual.

23Our study confirmed the higher effectiveness of contrastive meta-phonetic input, as reflected in the degree of vocalic approximation to the target formant values. The general rule of the psychology of learning—that by becoming aware, the resolution capacity of the analyser becomes significantly refined—has been confirmed.

24The research results confirm both hypotheses:

  1. English short vowels produced by the participants in the post-test will approximate to the referential values more than vowels produced in the pre-test.
    We can confirm this because the average approximation in both groups in the pre-test was 168.55 and in the post-test 133.85.

  2. English short vowels produced by the experimental group in the post-test will approximate to the referential values more than vowels produced by the control group in the post-test.
    We can confirm this because the approximation in the experimental group in the post-test is 125.8 and in the control group in the post-test 141.9.

5. Conclusions

25From the point of view of the percipient, the auditory impression of “good” or “bad” foreign language pronunciation is co-created by several subsegmental, segmental, plurisegmental and suprasegmental phonic phenomena (Kralova, 2016). Some studies (e.g. Kralova, 2011) have shown that the amount of segmental sound substitutions significantly correlates with the evaluation of speech as non-native. However, this does not mean that substitutions are the only criterion. They are likely the easiest to be identified by the ear, and the listener constructs an overall impression of foreign language speech combining several factors.

26The current study provides several departure points for possible future research. It would be possible to carry out a similar analysis of suprasegmental level phenomena, or to explore the influence of segmental training of English pronunciation to an overall English phonic competence in comparison with a suprasegmentally focused training. Future work could also attempt to establish the retention rate of the phenomena after training, as well as other lingual or extra-lingual determinants of foreign language pronunciation.

27Foreign language pronunciation is a complex and complicated phenomenon. It is not always possible to atomize elements and study foreign language pronunciation as a whole. Nevertheless, the difficulties and complexities involved should not prevent researchers from seeking appropriate generalizations, in the pursuit of findings which are applicable to a variety of contexts of teaching and learning foreign language pronunciation.

Haut de page

Bibliographie

Allan, Dave. (2005). Oxford Placement Tests 1. Oxford: Oxford University Press.

Anderson-Hsieh, Janet, Johnson, Ruth & Koehler, Kenneth. (1992). The Relationship between Native Speaker Judgements of Non-Native Pronunciation and Deviance in Segmentals, Prosody and Syllable Structure. Language Learning, 42(4), 529–55.

Benke, Eszter & Medgyes, Peter. (2005). Differences in Teaching Behaviour between Native and Non-Native Speaker Teachers: As Seen by the Learners. In E. Llurda (dir.), Non-Native Language Teachers. Educational Linguistics (vol. 5, pp. 195–216). Boston, MA: Springer.

Bohn, Ocke-Schwen & Munro, Murray J. (2007). Language Experience in Second Language Speech Learning: In Honor of James Emil Flege. Amsterdam-Philadelphia: John Benjamins.

Carmichael, Lesley. (2000). Measurable Degrees of Foreign Accent: A Correlational Study of Production, Perception, and Acquisition (Diploma thesis). Washington: University of Washington.

Chebenova, Viera. (2001). Methodological Procedures of German Pronunciation Training (I). Cizi jazyky, 45(1), 6–8.

Cummins, Jim. (2005). Teaching for Cross-Language Transfer in Dual Language Education: Possibilities and Pitfalls. In TESOL Symposium on Dual Language Education (pp. 1–16). Istanbul: Bogazici University.

Derwing, Tracey M. & Munro, Murray J. (2005). Second Language Accent and Pronunciation Teaching: A Research-Based Approach. TESOL Quarterly, 39(3), 379–97.

Dickerson, Lonna J. (1974). Internal and External Patterning of Phonological Variability in the Speech of Japanese Learners of English: Toward a Theory of Second-Language Acquisition (Doctoral thesis). Chicago: University of Illinois.

Flege, James E. (1988). Factors Affecting Degree of Perceived Foreign Accent in English Sentences. Journal of the Acoustical Society of America, 84, 70–9.

Flege, James E., Bohn, Ocke-Schwen & Jang, Sunyoung. (1997). The Effect of Experience on Non-Native Subjects Production and Perception of English Vowels. Journal of Phonetics, 25(4), 437–70.

Flege, James E. & Hillenbrandt, James. (1984). Limits on Pronunciation Accuracy in Adult Foreign Language Speech Production. Journal of the Acoustical Society of America, 76(3), 708–21.

Gimson, Alfred C. (1989). An Introduction to the Pronunciation of English. Newcastle upon Tyne: Athenaeum Press.

Kissling, Elizabeth M. (2013). Teaching Pronunciation: Is Explicit Phonetics Instruction Beneficial for FL Learners? Modern Language Journal, 97(3), 720–44.

Kral, Abel & Sabol, Jan. (1989). Phonetics and Phonology. Bratislava: SPN.

Kralova, Zdena. (2011). Slovak-English Phonic Interference. Zilina: EDIS.

Kralova, Zdena. (2016). Classification of Factors of Foreign Language Phonic Competence. XLinguae, 9(4), 66–75.

Malikova, Maria. (1993). To the Issues of Foreign Language Abilities. Phonematic Ear. Nitra: VSP.

Meador, Diane, Flege, James E. & MacKay, Ian R. A. (2000). Factors Affecting the Recognition of Words in a Second Language. Bilingualism: Language Cognition, 3(1), 55–67.

Mildner, Vesna & Horga, Damir. (1999). Relations between Second Language Proficiency and Formant-Defined Vowel Space. In Proceedings of the 14th International Congress of Phonetic Sciences (pp. 1455–8).

Moussu, Lucie. (2006). Native and Non-Native English-Speaking English As a Second Language Teachers: Student Attitudes, Teacher Self-Perceptions, and Intensive English Program Administrator Beliefs and Practices (Doctoral thesis). West Lafayette, IN: Purdue University.

Munro, Murray J. & Derwing, Tracey M. (2001). Modeling Perceptions of the Accentedness and Comprehensibility of L2 Speech. The Role of Speaking Rate. Studies in Second Language Acquisition, 23(4), 451–68.

Munro, Murray J., Derwing, Tracey M. & Flege, James E. (1999). Canadians in Alabama: A Perceptual Study of Dialect Acquisition in Adults. Journal of Phonetics, 27, 385–403.

Peltola, Kimmo U., Rautaoja, Tomi, Alku, Paavo & Peltola, Maija S. (2017). Adult Learners and One-Day Production Training — Small Changes but the Native Language Sound System Prevails. Journal of Language Teaching and Research, 8(1), 1–7.

Roach, Peter. (2009). Eng1ish Phonetics and Phonology. Cambridge: Cambridge University Press.

Sheppard, Chris, Hayashi, Chiyo & Ohmori, Ai. (2007). Factors Accounting for Attainment in Foreign Language Phonological Competence. In Proceedings of the 16th International Congress of Phonetic Sciences (pp. 1597–600).

Weinreich, Uriel. (1953). Languages in Contact. New York: Linguistic Circle of New York.

Wrembel, Magdalena. (2005). Metacompetence-Oriented Model of Phonological Acquisition: Implications for the Teaching and Learning of Second Language Pronunciation. In PTLC2015 (pp. 1–4). London: UCL.

Haut de page

Table des illustrations

Titre Figure 1. – Experimental analysis.
URL http://journals.openedition.org/lidil/docannexe/image/7377/img-1.png
Fichier image/png, 1,1M
Titre Figure 2. – F1 and F2 reference and standard values.
URL http://journals.openedition.org/lidil/docannexe/image/7377/img-2.png
Fichier image/png, 633k
Titre Figure 3. – F1 and F2 referential and produced values (pre-test).
URL http://journals.openedition.org/lidil/docannexe/image/7377/img-3.png
Fichier image/png, 635k
Titre Figure 4. – F1 and F2 referential and produced values (post-test).
URL http://journals.openedition.org/lidil/docannexe/image/7377/img-4.png
Fichier image/png, 633k
Haut de page

Pour citer cet article

Référence électronique

Zdena Kralova, Katarina Nemcokova et Jana Birova, « Contrastive vs Non-Contrastive Meta-Phonetic Input in Teaching Foreign Language Pronunciation », Lidil [En ligne], 61 | 2020, mis en ligne le 02 mai 2020, consulté le 28 septembre 2020. URL : http://journals.openedition.org/lidil/7377 ; DOI : https://doi.org/10.4000/lidil.7377

Haut de page

Auteurs

Zdena Kralova

(corresponding author)
Tomas Bata University, Zlin, Czech Republic
zkralova@utb.cz

Katarina Nemcokova

Tomas Bata University, Zlin, Czech Republic

Jana Birova

University of Ss. Cyril and Methodius in Trnava, Trnava, Slovakia

Haut de page

Droits d’auteur

© Lidil

Haut de page
  • OpenEdition Journals
Search OpenEdition Search

You will be redirected to OpenEdition Search