Robert Allen Fox

Effect of Lexical Status on Phonetic Categorization

To investigate the interaction in speech perception between lexical knowledge (in particular, whether a stimulus token makes a word or nonword) and phonetic categorization, sets of [bVC]-[dVC] place-of-articulation continua were constructed so that the endpoint tokens represented word-word, word-nonword, nonword-word, and nonword-nonword combinations. Experiment 1 demonstrated that ambiguous tokens were perceived in favor of the word token and supported the contention that lexical knowledge can affect the process of phonetic categorization. Experiment 2 utilized a reaction time procedure with the same stimuli and demonstrated that the effect of lexical status on phonetic categorization increased with response latency, suggesting that the lexical effect represents a perceptual process that is separate from and follows phonetic categorization. Experiment 3 utilized a different set of [b-d] continua to separate the effects of final consonant contrast and lexical status that were confounded in Experiments 1 and 2. Results demonstrated that both lexical status and contextual contrast separately affected the identification of the initial stop. Data from these three experiments support a perceptual model wherein phonetic categorization can operate separately from higher levels of analysis.

VOWEL DURATION IN THREE AMERICAN ENGLISH DIALECTS

The article reports on an acoustic investigation into the duration of five American English vowels, those found in hid, head, had, hayed, and hide. We compare duration across three major dialect areas: the Inland North, Midlands, and South. The results show systematic differences across all vowels studied, with the longest durations in the South and the shortest in the Inland North, with the Midlands in an intermediate but distinct position. More generally, the sample differs from and complements other work on this question by including detailed evidence from relatively small, cohesive areas, each within a different established dialect region.

Prosodic prominence effects on vowels in chain shifts

This study examines synchronic variation in vowels in an effort to advance our understanding of the “transmission problem” in language change, in particular, the cross-generational perseverance of vowel shifts. Seeking a connection to patterns anddirectionsofshiftsinvowelsystemsovertime,weexaminetheroleofalargely neglectedparameterofstructuredheterogeneity:prosodicprominence.ExperimentaldatafromtwoMidwesterndialectsofAmericanEnglish—centralOhioandsouthcentral Wisconsin—show that, for the vowels studied here, the changes in vowel characteristics observed under higher degrees of prosodic prominence (or greater emphasis) correspond to the changes predicted by well-established principles of chainshifting.Anacousticstudyassessesvariationinprosodicprominencebyexamining formant frequencies at multiple locations in the course of vowel duration, whichprovidesinformationaboutvowelqualitydynamics.Aperceptualstudydetermines listeners’sensitivity to the obtained acoustic variation, as manifested in specific patterns of vowel identification, confusions, and category goodness ratings. Finally, a prosodically based explanation of the transmission of sound change is described, which offers new connections between structural and social factors in sound change, notably the roles of “social affect” and speaker gender.

The effect of vowel quality variations on stress‐beat location

This study examines the effect which vowel quality variations have upon the location of vowel onsets—indicators of stress‐beat location—in rhythmically produced lists of monosyllables. Three talkers produced 120 sequences consisting of seven monosyllabic tokens of the form /s___t/. These token sequences were either homogeneous (identical components) or alternating (stimulus syllables alternated with the token “sight”). Talkers produced the sequences in both a metronome and nonmetronome condition. Analysis of the obtained results shows that vowel quality variations had a significant effect upon the location of the vowel onset. Regression analysis demonstrated that the differences in vowel onset‐to‐vowel onset intervals were significantly related to vowel duration differences between the tokens. There was also a significant difference between talkers in terms of the location of the vowel onsets, suggesting either the use of different production strategies or differential capabilities in doing the rhythmic task. No evidence of preboundary lengthening was obtained. These data support the suggestions [R. A. Fox and I. Lehiste, J. Acoust. Soc. Am. Suppl. 1 77, S54 (1985)] that the location of the “stress beat” is determined on the basis of the acoustic structure of the entire syllable, rather than strictly upon the articulatory onset of the vowel and that subjects in such a task do not necessarily impose a hierarchical structure on the syllabic sequences. Perceptual data relevant to the importance of vowel durations to the stress‐beat location will also be presented.

Agrammatism in Jordanian-Arabic speakers

The studies of agrammatism show that not all morpho-syntactic elements are impaired to the same degree and that some of this variation may be due to language-specific differences. This study investigated the production of morpho-syntactic elements in 15 Jordanian-Arabic (JA) speaking individuals with agrammatism and 15 age-matched neurologically healthy individuals. Two experiments were conducted to examine the production of complementizer, tense, agreement and negation morphology in JA. The results indicated that the speakers of JA with agrammatism had marked dissociations in producing specific morpho-syntactic elements. The observed impairment patterns overlapped, in many respects, with those observed in other linguistic groups. The findings are discussed with respect to current theories of agrammatism, including both morpho-syntactic and computational accounts.

Auditory contrast and speaker quality variation in vowel perception

Selective adaption and anchoring effects in speech perception have generated several different hypotheses regarding the nature of contextual contrast, including auditory/phonetic feature detector fatigue, response bias, and auditory contrast. In the present study three different seven-step [hId]-[h epsilon d] continua were constructed to represent a low F0 (long vocal tract source), a high F0 (long vocal tract source), and a high F0 (short vocal tract source), respectively. Subjects identified the tokens from each of the stimulus continua under two conditions: an equiprobable control and an anchoring condition which included an endpoint stimulus from one of the three continua occurring at least three times more often than any other single stimulus. Differential contrast effects were found depending on whether the anchor differed from the test stimuli in terms of F0, absolute formant frequencies, or both. Results were inconsistent with both the feature detector fatigue and response bias hypothesis. Rather, the obtained data suggest that vowel contrast occurs on the basis of normalized formant values, thus supporting a version of the auditory-contrast theory.

Reconceptualizing the vowel space in analyzing regional variation

Vowel space area calculated on the basis of the corner vowels has emerged as a metric for the study of regional variation, speech intelligibility, and speech development. We verify the basic assumptions underlying both the concept of the vowel space and the utility of the vowel space area in making speaker, dialect, or language comparisons. Undeniably, the traditional vowel triangle and vowel quadrilateral both fail as a metric in the context of dialect variation because substantial parts of the actual working space are excluded from analysis. Utilizing the formant values at a number of different locations for a wider range of individual vowels has significant implications for the size and shape of the resulting vowel space. Indeed, dialectal variations in vowel production can best be characterized in terms of formant density regions in the formant space and not as locations of individual vowel categories. The formant density approach is based on the assumption that vowel sounds are dynamically changing m...

Speech intelligibility in cross‐dialectal multi‐talker babble.

This study examines speech perception in multi‐talker babble by varying speaker dialect as a masker. Studies varying the language of the target speech and the babble masker show that the target is most intelligible when the masker is constructed from a non‐native language and target represents exemplars from the listener’s native language. Does variation in speaker and listener dialect has a similar effect on speech intelligibility? In this study, central Ohio listeners listen to target sentences spoken by either native Ohio or native North Carolina talker. The sentences are masked by multi‐talker babble consisting of speech from either Ohio or North Carolina talkers presented at several sound‐to‐babble (S/B) ratios. The same material is also presented to listeners in North Carolina. The results for Ohio listeners show significant effects of dialect at 0 dB S/B ratio. Intelligibility improved when the background babble featured the dialect that was different from the dialect of the target speaker. The per...

Generational and dialectal effects on children’s vowel identification.

This study examines vowel identification by 8–13 years old children who grew up in either Southeastern Wisconsin (whose regional variant is affected by the Northern Cities Shift) or Western North Carolina (affected by the Southern Vowel Shift). In the first identification task, the children responded to words edited from sentences which elicited both stressed and unstressed vowel exemplars. This speech material was produced by multiple talkers representing two generations (children and older adults who represent their grandparents’ generation). In the second identification task, the children were presented with citation‐form tokens produced by three generations of talkers (children, their possible parents, and their possible greatgrandparents). Both within‐ and across‐dialect vowel identification was examined. The cross‐generational results showed that some vowels were identified more accurately when spoken by children, some when spoken by adults and for others there were no cross‐generational differences...

Salience of dynamic virtual formants in diphthongs

Variations in the quality of a synthesized vowel can be produced by changing the frequency of its steady‐state formants or the amplitude of closely‐spaced static spectral components—the latter effect a function of auditory spectral integration. Recent studies have demonstrated that dynamic modification of the amplitude ratios of these spectral components can give rise to the perception of virtual diphthongs. This study examines the extent to which spectral integration of these dynamic components is uniform across a range of frequency separations of these spectral components (previous work [Fox et al., ‘‘Title,’’ J. Acoust. Soc. Am. 120, 3252 (2006)] demonstrated the integration of static components beyond 3.5 bark). Two synthetic vowel series were created by modifying the rise or fall of F2 producing [■i]‐[■:]‐[■u] and [i■]‐[■:]‐[u■] series. Next, several different series were created in which this dynamic F2 was replaced by two static spectral components whose amplitude ratios were varied over time (virt...