Language and Speech

The historical provenance of a minimum cut-off point (of about 0.25 sec) for pauses in temporal analyses of speech production is associated with Goldman-Eisler's usage. Her rationale was the predominance of articulatory pauses at lengths shorter than 0.25 sec. Both phonotactic facts and empirical analysis of several corpora of readings disconfirm this predominance with respect to pauses 0.13-0.25 sec in length. The vast majority of these pauses are found to be psychological; they are determined by syntax, punctuation, rhetorical and expressive emphasis, poetic format, and stylistic pecularities.

The effects of rater's accent and amount of linguistic experience on semantic differential ratings of samples of three varieties of Irish accent, Standard, Rural and Dublin, were explored. Principal components analysis revealed two major components. The first, labelled Prestige, accounted for 43.5% of the variance, while the second, labelled Solidarity, accounted for a further 21.4% of the variance. The analysis of variance showed no significant experience effect or accent loyalty effect. For both Prestige and Solidarity scales the differences between the mean scores given to the speech varieties were statistically significant. The differences between the mean scores given to individual tapes within speech varieties and between those given by individual subjects within accent/experience groups were also highly significant statistically. For Solidarity the results suggest an interaction between the effects of the individual subjects and the accent varieties. The dimensionality of the results is discussed in relation to the findings of Edwards (1979), Williams (1970), Gallois and Callan (1981), Carranza and Ryan (1975) and Brennan and Brennan (1981).

Perceptual judgments, acoustic measurements and electromyographic (EMG) records were obtained for one deaf speaker producing the vowels [i, I, æ, a, u, u] in an [hVd] frame. Overall listener judgments were consistent with spectral measurements. In general, front vowels were perceived as more similar to targets than back vowels, and high vowels were perceived correctly more often than low vowels. Experienced and inexperienced listeners were found to differ significantly in their categorization of the point vowels [i, æ, a, and u] but not for [I and u] . The vowel space, as determined by the formant frequency measures, was reduced with respect to normal values particularly in the region appropriate to high back vowels. However, EMG records of genioglossus and orbicularis oris do not entirely account for the perceptual and acoustic data. In particular, genioglossus activity is relatively undifferentiated across all vowels when compared to data from normals. The results of this study generally support the widespread notion of reduced vowel space secondary to a reduced range of tongue movement in this deaf speaker. The physiological records were also characterized by a significant degree of variability from token to token. In this regard, these data are different from acoustic and physiological patterns that have been previously reported for vowels produced by deaf speakers.

It is shown that the use of colour words in poetry and literature correlates closely with the order of evolution of colour words, as described by Berlin and Kay. This relationship is independent of the overall frequency of colour word usage. A review of other literature suggests that a number of the psychological aspects of colour naming correlate closely with the Berlin and Kay ordering.

Two studies are presented which investigated whether spontaneous speech is characterized by the presence of a cognitive rhythm that consists of alternating phases of hesitancy and fluency. In the first study, cumulative phonation-silence graphs were generated from the tape recordings of 54 one-minute monologues. These phonation-silence graphs were presented to two judges who were asked to rate the graphs for the presence or absence of cognitive rhythm. The results showed that the judges considered cognitive rhythm to be absent. A similar procedure was used in a second study in which the graphs were produced from a further set of 48 one-minute monlogues. As before, the two judges considered cognitive rhythm to be absent. In addition, in the second study the judges were unable to distinguish a set of computer-generated random walks from the speaker-generated graphs. It is concluded that cognitive rhythms are no more typical of normal speech than they are of random walks, and that any claims about their presence in speech must be treated with great caution.

This study examined the development of idiomatic expression. Sixty subjects were presented with a series of drawings corresponding to sentences exhibiting dual semantic representation (i.e., a literal and an idiomatic component). A literalization strategy appeared to be operating between the ages of 6 and 9 years and an idiomatization strategy predominated from the age of 9 through adulthood. Comprehension of idiomatic meaning appeared to develop prior to the ability to verbally interpret idioms.

Previous research identifies a number of factors associated with a socially stigmatized speech style known initially as "women's language" and subsequently as "powerless language." This study examines the effects of speaker sex, socio-economic status, ability, communication apprehension, rigidity, and question type on the incidence of 16 style markers and on verbosity in simulated job interviews. Results reveal limited evidence of the impact of social and psychological factors, but pervasive potency for question type. Open questions constitute an element of interaction structure that engenders expressions of tentativeness and uncertainty. Under these circumstances, it is more reasonable to change dysfunctional linguistic stereotypes than common, contextually induced language behaviors.

A silent interval that cues the /š/-/č/ distinction in many contexts is less likely to do so when it coincides with certain boundaries. In natural speech these boundaries are generally marked by both prosody and syntax. We independently varied syntax and prosody to assess their contributions to the phonetic interpretation of silences occurring at these boundaries. We used a set of four sentences, four durations of silence, and two prosodic patterns (Experiment I). We constructed sentences using three techniques that differed in the amount of prosodic control and in naturalness: synthesis by rule, concatenation of naturally produced syllables, and cross-splicing of naturally produced utterances. Silence duration had a strong effect on the perception of the /š/-/č/ contrast in all conditions. For the Synthetic Condition, we also found a strong effect of the prosodic pattern. We found no evidence of any purely syntactic effect. In Experiment II, the two syllables surrounding the silence were excised from the sentences of Experiment I and presented to listeners for labeling. Prosody had a significant effect in the Synthetic Condition and in the Natural Condition. The results indicate that the local prosodic pattern (one syllable with a pitch fall and a greater length) can be sufficient to influence listeners' perception of the /š/-/č/ contrast. There is also evidence that the prosodic information may be subject to context effects.