Jeff Mielke

The interplay of speech perception and phonology: experimental evidence from Turkish.

This study supports claims of a relationship between speech perception and phonology with evidence from a crosslinguistic perception experiment involving /h/ deletion in Turkish. Turkish /h/ is often deleted in fast speech, but only in a specific set of segmental contexts which defy traditional explanation. It is shown that /h/ deletes in environments where lower perceptibility is predicted. The results of the perception experiment verify these predictions and further show that language background has a significant impact on speech perception. Finally, this perceptual account of Turkish /h/ deletion points to an empirical means of testing the conflicting hypotheses that perception is active in the synchronic grammar or that its influence is limited to diachrony.

Ambivalence and ambiguity in laterals and nasals

Ambivalent segments are speech sounds whose cross-linguistic patterning is especially variable, creating contradictions for theories of universal distinctive features. This paper examines lateral liquids, whose [continuant] specification has been the subject of controversy because of their ability to pattern both with continuants and with non-continuants, and because phonetically they are situated in the contested ground between two different articulatory definitions for the feature [continuant]. Evidence from a survey of sound patterns in 561 languages shows that lateral liquids, like nasals, pattern with continuants about as often as with non-continuants. Ambivalent phonological behaviour is argued to be natural and expected for phonetically ambiguous segments in a theory of emergent distinctive features where features are the result of sound patterns, rather than the other way around.

Variability in American English s-retraction suggests a solution to the actuation problem

Although formulated by Weinreich, Labov, and Herzog in 1968, the actuation problem has remained an unsolved problem in understanding sound change: if sound change is conceived as the accumulation of coarticulation, and coarticulation is widespread, how can some speech communities resist phonetic pressure to change? We present data from American English s-retraction that suggest a partial solution. S-retraction is the phenomenon in which /s/ is realized as an [ʃ]-like sound, especially when it occurs in an /stɹ/ cluster (‘ street ’ pronounced more like [ʃtɹit] than like [stɹit]). The speech of English speakers judged not to exhibit s-retraction shows a large coarticulatory bias in the direction of retraction. Further, there is also substantial interspeaker variation in the extent of this bias. We propose that this interspeaker variation, coupled with the coarticulatory bias, facilitates the initiation of sound change. In this account, sound change begins when a listener accidentally interprets an extreme case of a phonetic effect as an articulatory target and then adjusts her own speech in response. This adoption of a new target requires phonetic variation that predates the change. Thus, sound change is predicted to be biased toward phonetic effects that exhibit interspeaker variability, and if sound change requires an accident that is rare, then sound change itself is correctly predicted to be rare as well.

Looking through opacity

Abstract 1. Introduction Comparative Markedness deals with alternations which are problematic for classical Optimality Theory such as counterfeeding opacity. In Sea Dayak, for example, the distribution of nasal and oral vowels is generally predictable: after a nasal consonant, a vowel is typically nasal and after an oral consonant, the vowel is oral. However, an oral vowel also occurs after a nasal consonant just in case the consonant is optionally followed by an oral stop, as in [rambo?] ∼ [ramo?] ‘a kind of flowering plant’. The orality of the postnasal vowel in such cases is thus opaque (Scott 1957, 1964). Representative forms are shown in (1).

Discovering place and manner features—What can be learned from acoustic and articulatory data?

Theories of phonological acquisition often assume that features are innate and universal, and establishing distinctive features requires the availability of minimal pairs. In this study, we explore an inductive approach to feature discovery without resorting to minimal pairs. Our basic assumption is that learners have access to a wide variety of phonetic information, and so feature induction can proceed through recursive clustering of the data. To test our hypothesis, we used as our data acoustic and articulatory measurements made simultaneously from a set of CVC syllables. The articulatory data are obtained with ultrasound imaging, and the vocal tract is characterized by a vector of cross distances from the tongue body to a reconstructed palate. The raw articulatory data are then mapped to a lower dimensional space using principle components, on which clustering is carried out. Experiments using acoustic and articulatory data suggest that these two sources of information lead to the discovery of place an...

Articulation of the Kagayanen interdental approximant: An ultrasound study

This paper documents the articulation of the interdental approximant, an unusual speech sound that occurs in several languages spoken in the Philippines and Western Australia. This sound is notable for the fact that the tongue protrudes from the mouth and contacts the lower lip, and it seems to have a lateral perceptual quality, but documentation of the other details of the sound have been sketchy. We use ultrasound imaging to study the sound produced by a speaker of Kagayanen. We show that the only constriction is interdental, that the degree of tongue protrusion is related to vowel context and focus, and that the sound does not involve tongue raising. Coronal section images indicate that the sound involves the lowering of at least one side of the tongue, making it articulatorily lateral. We also discuss the implications for theories of tongue movement.

Greater than noise: frequency effects in Bantu height harmony

A major concern of phonological theory is how complex phonological patterns can be acquired simply through exposure to imperfect language data. Central to this topic is the nature of the cognitive apparatus brought to bear on this problem of acquisition. The “nativist” extreme in this regard is to postulate a highly articulated Universal Grammar (UG), with a rich set of structures that are specific to language, hard-wired into the language-learning infant. The “nurture” extreme is to assume that there is very little cognitive infrastructure that is specific to language. Instead, general purpose learning mechanisms are applied to language data, resulting in an Emergent Grammar (EG) that develops from exposure to the patterns of an adult grammar. These two proposals make different predictions about the nature of sound patterns. In this study, our basic question is whether the distribution of sounds within languages with vowel harmony exhibits the patterns predicted by UG or those predicted by EG. To do so, we explore the frequency of words that violate height harmony rules, and as we will see, in several languages they are less frequent than would be expected by chance, but more frequent than would be expected on the basis of a categorical phonological rule, i.e., the robustness of the patterns is greater than the noise in the data, but less than exceptionless.

The phonetic status of the (inter)dental approximant

The (inter)dental approximant is a little-studied speech sound in the Philippines and Western Australia. In this paper, we document the articulation of the sound, providing acoustic and video data from Kagayanen and Limos Kalinga, respectively. The sound is attested in at least fifteen languages. It is contrastive in five Western Australian languages, while in the Philippines it generally patterns as an allophone of / l / but has emerged recently as a separate phoneme due to contact. It arose independently in the two regions. The sound is easily describable in terms of values of phonological features or phonetic parameters. All of these factors argue for the inclusion of the sound in the International Phonetic Alphabet.

Acoustic properties of the interdental approximant in Kagayanen

This paper describes the acoustic properties of the interdental approximant, a rare speech sound reported to date only in a dozen languages in the Philippines. The measurements analyzed here are based on recordings of one female speaker of Kagayanen. The interdental approximant exhibits acoustic characteristics typical of semi‐vowels: the formant pattern is similar to that of vowels, and the formant transitions with adjoining vowels have a long duration, usually 35 ms or more. The values of F2 and F3 are analogous to those of the Kagayanen [l], a prototypical voiced alveolar lateral; the only significant formant value difference between the two sounds involves F1: the interdental approximant has a mean of 508 Hz (n=9), whereas the mean for [l] is 368 Hz (n=9). On the other hand, while these two segments sound impressionistically similar, the interdental approximant differs acoustically from [l] in several other respects as well: there is no abrupt change in the formant pattern at the junctures with adjoin...

The articulatory dynamics of pre-velar and pre-nasal /æ/-raising in English: An ultrasound studya)

Most dialects of North American English exhibit /æ/-raising in some phonological contexts. Both the conditioning environments and the temporal dynamics of the raising vary from region to region. To explore the articulatory basis of /æ/-raising across North American English dialects, acoustic and articulatory data were collected from a regionally diverse group of 24 English speakers from the United States, Canada, and the United Kingdom. A method for examining the temporal dynamics of speech directly from ultrasound video using EigenTongues decomposition [Hueber, Aversano, Chollet, Denby, Dreyfus, Oussar, Roussel, and Stone (2007). in IEEE International Conference on Acoustics, Speech and Signal Processing (Cascadilla, Honolulu, HI)] was applied to extract principal components of filtered images and linear regression to relate articulatory variation to its acoustic consequences. This technique was used to investigate the tongue movements involved in /æ/ production, in order to compare the tongue gestures involved in the various /æ/-raising patterns, and to relate them to their apparent phonetic motivations (nasalization, voicing, and tongue position).