Mona Lindau

The story of /r/

In describing sound changes, sound patterns and alternations in languages one clearly needs to refer to a natural class of r sounds, or rhotics. This phonological class encompasses sounds with a wide variety of manners and places of articulation, and it has been suggested that its “phonetic correlate” is acoustic in nature, namely a lowered F3. Acoustic properties of phonetically different /r/s were investigated in languages like American English, Yoruba, French, Southern Swedish, Hausa, and Edo, using several speakers of each language. The results do not show any single acoustic parameter underlying the phonological class of [+rhotic]. The third formant does not lower for all types of /r/s. There is often a decreased intensity associated with the rhotic, but this is by no means consistent. The natural class of [+rhotic] sounds is a phonologically convenient class, but it is not controlled by any single articulatory or acoustic correlate. Instead, this phonological class is associated with complex combina...

Individual differences in vowel production

It is often assumed that a relatively small set of articulatory features are universally used in language sound systems. This paper presents a study which tests this assumption. The data are x-ray microbeam pellet trajectories during the production of the vowels of American English by five speakers. Speakers were consistent with themselves from one production of a word to the next, but the articulatory patterns manifested by this homogeneous group were speaker specific. Striking individual differences were found in speaking rate, the production of the tense/lax distinction of English, and in global patterns of articulation. In terms of a task-dynamic model of speech production, these differences suggested that the speakers used different gestural target and stiffness values, and employed different patterns of interarticulator coordination to produce the vowels of American English. The data thus suggest that, at some level of speech motor control, speech production tasks are specified in terms of acoustic output rather than spatiotemporal targets or gestures.

Phonetic differences in glottalic consonants

Data on glottalic consonants in a number of languages were examined. Data on implosives come from Hausa, Degema, and three Eastern Ijo languages, all spoken in Nigeria. Ejective data come from Hausa and Navaho. The implosives in Degema and the Ijo languages show a strong positive correlation (r = 0.9; p = 0.025) between the closure duration and the intensity of implosion. However, Hausa implosives do not follow this correlation in that they have a high intensity implosion with a relatively brief closure duration. This is attributed to the fact that Hausa implosives are also laryngealized, and the resulting lower airflow into the vocal tract makes it possible to have a more efficient implosion in Hausa. Due to the different phonation type the implosive mechanism is thus considerably different in Hausa than in the other languages. The ejectives in Hausa and Navaho differed along several parameters: total duration, ratio closure duration/open interval, and phonation type in the vowel onset. Thus what appears to be the same sound nevertheless has to be described in different ways for different languages. [This work was supported by NSF.]

Testing a model of intonation in a tone language

Schematic fundamental frequency curves of simple statements and questions are generated for Hausa, a two-tone language of Nigeria, using a modified version of an intonational model developed by Gårding and Bruce [Nordic Prosody II, edited by T. Fretheim (Tapir, Trondheim, 1981), pp. 33-39]. In this model, rules for intonation and tones are separated. Intonation is represented as sloping grids of (near) parallel lines, inside which tones are placed. The tones are associated with turning points of the fundamental frequency contour. Local rules may also modify the exact placement of a tone within the grid. The continuous fundamental frequency contour is modeled by concatenating the tonal points using polynomial equations. Thus the final pitch contour is modeled as an interaction between global and local factors. The slope of the intonational grid lines depends at least on sentence type (statement or question), sentence length, and tone pattern. The model is tested by reference to data from nine speakers of Kano Hausa.

Some cross-linguistic differences in diphthongs

In this paper we will discuss some phonetic differences that are found in the production of diphthongs. The study is of a limited scope and may be considered a pilot study. The kinds of questions that will be addressed have to do with the specification of timing in speech production. Are the vowel elements of diphthongs joined in language-specific ways, or can they be joined by more general principles, such as the assumption of a fairly constant transition duration, so that the transition rate will be faster if the distance between onset - and offset - vowels is greater? This kind of principle was proposed by Kent and Moll (1972) for some types of vowel-to-vowel sequences. Or does the duration of the transition lengthen as the distance of the transition increases, keeping the transition rate fairly constant? If any of these principles are at work, then their effects will show up acoustically in the transition duration, particularly of F2. If the Kent and Moll principle is applicable, it would for example operate as a tendency towards similar F2 transition durations in different languages for the ‘same’ diphthongs.

Cross‐linguistic differences in diphthongs

Similar diphthongs may vary cross‐linguistically in terms of the relative duration and slope of the transition between the onset and offset of the diphthong. At moderate rates of speech this transition will constitute the major part of the diphthong in some languages, while in other languages it will be quite fast. In /ai/ the transition occupies about 60% of the total duration in American English; about 50% in Peking Chinese; but about 15% in Egyptian Arabic. In the diphthong /au/ the transition takes up the major part (about 75%) of the total duration in American English, while in Arabic and Hausa it is only about 25% of the total duration. The different transitional durations do not correlate with the total duration of the diphthong. Nor is the explanation to be sought in the phonologies of the different languages. We have also found that in Hausa and Arabic, where the diphthongs have sharp transitions, the movements of the first and second formants are not synchronous. The transition of the first form...

Larynx height in Kwa

The vertical positions of the larynx in various vowels in Akan, Igbo, and Ijo were measured from radiographic data. These languages have a vowel harmony such that one set of vowels [i, e, u, o] is produced with an expanded pharynx, and another set [ι, ɛ, o, ɔ] with a constricted pharynx. The major function of vertical larynx displacement is to contribute to the variation in pharyngal size for categorizing vowels into harmonizing sets. The larger pharynx is produced as a combined effect of advancing the tongue root and lowering the larynx, and the smaller pharynx is produced by retracting the root and raising the larynx. The vertical position of the larynx is reasonably well correlated with the position of the tongue root. But within each harmony set there is no consistent relationship between larynx height and vowel height or between larynx height and backness. Sometimes phonologically high vowels have a higher larynx than nonhigh vowels, sometimes not. The only possible exception is the vowel [a]. In those languages where [a] occurs, there is a strong tendency for it to have the highest larynx position of all the vowels.

Factor Analyses of Format Frequencies

The exact relationship between the phonetic quality of a vowel and its formant frequencies has never been stated, because it has not been possible to account for the variations due to the personal quality of a speaker. We have analyzed the fundamental and first three formant frequencies of the eight primary cardinal vowels as spoken by 11 phoneticians using the PARAFAC procedure (Harshman, this meeting). The results were first interpreted as demonstrating three factors which corresponded to the three traditional phonetic (auditory) parameters “tongue height, tongue fronting, lip rounding.” In these vowels, “tongue fronting” and “lip rounding” co‐vary. To separate these two factors, an analysis was conducted of a subset of Swedish vowels which included front rounded vowels. A factor corresponding to “tongue height” was still evident. The other two factors were not so easily interpretable. A possible interpretation of one of these two factors related it to to the Jakobsonian feature “compact.”

The degrees of freedom in controlling articulations

What do speakers control in the production of vowels? Data from two groups of five speakers saying words containing ten different American English vowels indicate that in saying the same vowel there is a great deal of variation between speakers in the articulatory gestures used, but considerable consistency within speakers in the articulations used. All ten speakers had constrictions in similar regions of the vocal tract for each vowel; but the relation between tongue movement and jaw movement was variable. Principal components analyses of the locations of six pellets showed that, for one group, two of the five subjects used the jaw and tongue in combination, and their vocal tract shapes could be described in terms of only two components. For the other three subjects, three components were needed to account for the same amount of variance, with the third component distinguishing tense and lax vowels. The other group of five subjects was analyzed separately. Two subjects showed good correlation between ton...

Interarticulatory relationships in vowel production

In earlier studies it has been shown that, within a language (English), the jaw and the tongue interact in different ways for different speakers to produce differences between vowels. These studies were, however, limited to front vowels. Using the x‐ray microbeam system, articulatory movements were recorded from pellets placed on the lips, mandible, and the tongue, in five speakers saying utterances illustrating all the English vowels. The results show some rather complex relationships. For example, for some speakers tongue height, as measured by the height of the tongue dorsum, does not always distinguish between the high front /i/ and mid front /e/ (e.g., in dee and day); but these vowels are distinguished by the height of the mandible. The height of the dorsum does separate the tense /i/ and the lax /ι/ (e.g., in dee and did) and the back /u/ and /o/ (e.g., in do and doe). The results also show the degree of consistency within individuals in favoring certain trading relationships between the lips, the ...