Ocke-Schwen Bohn

Asymmetries in vowel perception

Asymmetries in vowel perception occur such that discrimination of a vowel change presented in one direction is easier compared to the same change presented in the reverse direction. Although such effects have been repeatedly reported in the literature there has been little effort to explain when or why they occur. We review studies that report asymmetries in vowel perception in infants and propose that these data indicate that babies are predisposed to respond differently to vowels that occupy different positions in the articulatory/acoustic vowel space (defined by F1-F2) such that the more peripheral vowel within a contrast serves as a reference or perceptual anchor. As such, these asymmetries reveal a language-universal perceptual bias that infants bring to the task of vowel discrimination. We present some new data that support our peripherality hypothesis and then compare the data on asymmetries in human infants with findings obtained with birds and cats. This comparison suggests that asymmetries evident in humans are unlikely to reflect general auditory mechanisms. Several important directions for further research are outlined and some potential implications of these asymmetries for understanding speech development are discussed.

An Instrumental Study of Vowel Reduction and Stress Placement in Spanish-Accented English.

Morphophonological alternations in English words such as able versus ability involve changes in both stress and vowel quality. This study examined how native speakers of Spanish and English produced four such morphologically related English word pairs. Degree of stress and vowel quality was assessed auditorily and instrumentally. Stress placement generally seemed to constitute less of a learning problem for the native Spanish speakers than vowel reduction. The results suggest that Englishlike stress placement is acquired earlier than vowel reduction and that the ability to unstress vowels is a necessary, but not sufficient, condition for vowel reduction. The magnitude of stress and vowel quality differences for the four word pairs suggests that L2 learners acquire stress placement and vowel reduction in English on a word-by-word basis.

Language Experience in Second Language Speech Learning: In Honor of James Emil Flege

Published to mark James Flege’s retirement, this collection of papers is a fitting tribute to one of the most influential and probably the most prolific scientists in the field of second language (L2) speech research. In 365 pages 20 papers offer discussion and present empirical findings on a wide variety of topics from the field of secondor foreign-language research, and a further 40 pages of bibliography, name and subject indexes complete a fascinating picture of present-day L2 research. Inevitably, since the book is a festschrift for an American scholar, there is a preponderance of North American or North-America-based authors among the 30 contributors (22, with 3 each based in Scandinavia and Australia, respectively, 1 in Japan and 1 in Germany). Flege’s Speech Learning Model (SLM) is at the centre – or at least is the shared focus – of a large majority of the articles. This no doubt reflects both the authors’ respect for his work and the extent of his influence in the field. As the editors write, all the central themes of pronunciation research receive attention from one or more papers. However, it is not our intention to duplicate the useful theme-orientated commentary offered by the editors in their introductory overview. We therefore try to capture the main thrust, with comments on the merits and points of special interest, of the individual papers in the order they are presented in five thematic sections, which form the structuring framework for the 20 papers. These sections are: Part I – The Nature of L2 Speech Learning, Part II – The Concept of Foreign Accent, Part III – Consonants and Vowels, Part IV – Beyond Consonants and Vowels, and Part V – Emerging Issues. It must be said, however, that the subsection titles do not necessarily capture a particular thread running through the component papers. The book is not an a priori structured work, but rather a chance collection by established scientists and colleagues of James Flege. The titles of Parts I–V must therefore be taken as only rough guides to the orientation of the papers they contain. Part I is opened by the editors’ introduction, ‘The Study of Second Language Speech: A Brief Overview’, which serves both as a laudation and as a theoretical orientation to the 19 other contributions. It adds a historical perspective to the picture of L2 research that emerges in the course of the volume and provides some hints at possible future work for those looking to continue the effort. The four other papers in Part I consider foreign-language perception from different angles. Catherine Best and Michael Tyler (‘NonNative and Second-Language Speech Perception’) provide a theoretically concentrated discussion of the similarities and differences between Flege’s SLM and their own Perceptual Assimilation Model (PAM). Stressing the basically different orientation – SLM being concerned with L2 learning while PAM is a model to explain nonnative perception of a foreign language – they discuss the possible extension of PAM concepts to cover L2 phenomena. For those already familiar to some extent with the two models (though for the uninitiated the account is probably set at too high a level of abstraction), there is an illuminating contrapuntal explication of SLM postulates and PAM principles. Many of the differences in viewpoint, and possibly in the understanding of terms such as ‘phonetic’ and ‘phonological’, stem from a basic difference in as yet non-disprovable tenets, whether perception operates on concrete distal events (articulatory gestures) or mental representations (categories). The chapter by Winifred Strange (‘CrossLanguage Phonetic Similarity of Vowels’) is much more concrete in its approach, providing a discussion of different approaches to cross-language phonetic comparison – articulatory, acoustic and perceptual – with examples of acoustic and perceptual analyses. The example data are both pertinent and convincing, and underline the two main messages, namely that analyses – whether of production or of perception – have to be

Natural Referent Vowel (NRV) framework: An emerging view of early phonetic development

The aim of this paper is to provide an overview of an emerging new framework for understanding early phonetic development—the Natural Referent Vowel (NRV) framework. The initial support for this framework was the finding that directional asymmetries occur often in infant vowel discrimination. The asymmetries point to an underlying perceptual bias favoring vowels that fall closer to the periphery of the F1/F2 vowel space. In Polka and Bohn (2003) we reviewed the data on asymmetries in infant vowel perception and proposed that certain vowels act as natural referent vowels and play an important role in shaping vowel perception. In this paper we review findings from studies of infant and adult vowel perception that emerged since Polka and Bohn (2003), from other labs and from our own work, and we formally introduce the NRV framework. We outline how this framework connects with linguistic typology and other models of speech perception and discuss the challenges and promise of NRV as a conceptual tool for advancing our understanding of phonetic development.

How to organize a fairly large vowel inventory: the vowels of Fering (North Frisian)

This paper presents parts of the results of thefirst stage in the phonetic documentation of Fering, a dialect of North Frisian. Descriptions of Fering indicate that it has a large vowel inventory (15 monophthongs, 7 diphthongs, and 3 triphthongs). Ten older male native speakers were recorded, producing the vowel and consonant inventory at different speaking rates, in different phonetic contexts, and in different speaking styles. The results of the acoustic analyses of the vowels of Fering throw light on the questions of a) how the vowel space is exploited in Fering, b) how Fering differentiates its large vowel inventory, c) to what extent consonant‐vowel coarticulation affects the acoustic differentiation of vowel categories, and d) to what extent the implementation of the long‐short vowel contrasts is affected by speaking rate. The results also provide a more accurate description of the vowel system of an endangered language than do previous auditory accounts.

Revisiting vocal perception in non-human animals: a review of vowel discrimination, speaker voice recognition, and speaker normalization.

The extent to which human speech perception evolved by taking advantage of predispositions and pre-existing features of vertebrate auditory and cognitive systems remains a central question in the evolution of speech. This paper reviews asymmetries in vowel perception, speaker voice recognition, and speaker normalization in non-human animals – topics that have not been thoroughly discussed in relation to the abilities of non-human animals, but are nonetheless important aspects of vocal perception. Throughout this paper we demonstrate that addressing these issues in non-human animals is relevant and worthwhile because many non-human animals must deal with similar issues in their natural environment. That is, they must also discriminate between similar-sounding vocalizations, determine signaler identity from vocalizations, and resolve signaler-dependent variation in vocalizations from conspecifics. Overall, we find that, although plausible, the current evidence is insufficiently strong to conclude that directional asymmetries in vowel perception are specific to humans, or that non-human animals can use voice characteristics to recognize human individuals. However, we do find some indication that non-human animals can normalize speaker differences. Accordingly, we identify avenues for future research that would greatly improve and advance our understanding of these topics.

The influence of different native language systems on vowel discrimination and identification

The ability to identify the vowel sounds of a language reliably is dependent on the ability to discriminate between vowels at a more sensory level. This study examined how the complexity of the vowel systems of three native languages (L1) influenced listeners perception of American English (AE) vowels. AE has a fairly complex vowel system with 11 monophthongs. In contrast, Japanese has only 5 spectrally different vowels, while Swedish has 9 and Danish has 12. Six listeners, with exposure of less than 4 months in English speaking environments, participated from each L1. Their performance in two tasks was compared to 6 AE listeners. As expected, there were large differences in a linguistic identification task using 4 confusable AE low vowels. Japanese listeners performed quite poorly compared to listeners with more complex L1 vowel systems. Thresholds for formant discrimination for the 3 groups were very similar to those of native AE listeners. Thus it appears that sensory abilities for discriminating vowel...

Natural referent vowels guide the development of vowel perception

Certain vowels are favored across languages of the world. This selection bias has received a great deal of attention in linguistic theories seeking to explain vowel system typologies. In comparison, the role that specific vowels might play in the ontogeny of vowel perception has been more implicit. In this talk we will summarize recent findings that elucidate the functional significance of peripheral vowels in the development of vowel perception. Data from cross‐language studies of infant vowel discrimination and vowel preference will be presented. This work shows that peripheral vowels have a perceptual priority for young infants and that this bias is independent of the phonemic status of the vowels presented in the perceptual task. Findings from cross‐language experiments with adults reveal that language experience builds on the natural vowel biases observed in infancy. Adult data suggest that the natural bias remains in place in mature listeners unless the perceiver needs to override the bias to optimi...

Exploring the natural referent vowel hypothesis in infant perception experiments

Directional asymmetries in infant vowel perception studies have led Polka and Bohn [Speech Commun. 41, 221–231 (2003)] to posit the Natural Referent Vowel hypothesis (NRV) according to which vowels that are relatively more peripheral in the F1/F2 space are perceptually privileged vis‐a‐vis less peripheral vowels. This bias has been observed in preference and in change/no change experiments with infants and with adult non‐native listeners. NRV was further tested in headturn (change/no change) experiments with Danish‐learning infants. With one important exception, results support the NRV by showing that the predicted asymmetries exist in areas of the vowel space that have not been previously explored. The exception was observed for the Danish front vowel pair /e/‐/o//, where NRV incorrectly predicts that a change from the more peripheral /e/ to the less peripheral /o// is harder to discriminate than a change in the opposite direction. Interestingly, the unexpected asymmetry was observed only with infants up...

Acoustic studies comparing Danish vowels, British English vowels, and Danish‐accented British English vowels

Few studies have examined the production of English vowels by native speakers of languages with vowel inventories as large as, or larger than, English. Danish is such a language, whose vowels are unevenly distributed in the vowel space, with a densely populated upper portion and a sparsely populated lower portion of the vowel space. This paper reports on acoustic comparisons of British English vowels as produced by ten native speakers of British English and of Danish, and of Danish vowels as produced by ten native speakers of Danish. Danish and English vowels were produced in CVC syllables in a variety of consonantal contexts. Acoustic analyses revealed temporal, spectral, and dynamic differences between the vowels produced by the speaker groups. The results of this study suggest that comparisons of vowels across languages, as well as analyses of the productions of non‐native speakers, are incomplete and may even be misleading unless the effects of consonantal context are taken into account. The results o...