James G. Martin

On Judging Pauses in Spontaneous Speech.

Spectrograms of spontaneous speech revealed that syllables preceding a judged-pause location were usually longer than those following, whether or not a silent interval was present. Most judged-pause locations were junctures, but syllable length governed judgments independently of juncture cues.

The perception of hesitation in spontaneous speech

The issue in this paper was whether attending to acoustic elements and to message elements in a speech signal were compatible operations. In four experiments Ss listened for pauses and other hesitation phenomena in spontaneous speech; in three the task was reproduction of heard speech to include hesitations; in one the task was simply the marking of heard hesitations on transcripts. Experimental variables were instructions, degree of “ungrammaticality” of hesitations in speech inputs, time interval between listening and reproduction, and task manipulations along a continuum between simple hesitation detection and hesitation detection plus simultaneous speech decoding. Results were: (I) In all experiments Ss displaced within-constituent hesitations to constituent boundaries, suggesting a grammatical organization between input and output. (2) Instructional set to reproduce hesitations increased hesitations and words but at the expense of per cent words correct, suggesting that attending to acoustic elements such as hesitations was an interfering task during speech decoding. (3) The hesitation shift persisted in the hesitation-marking task when simultaneous speech decoding was required by the nature of the task, indicating that speaking (encoding) characteristics may not completely account for the shift. (4) The distribution of hesitation marking errors toward grammatical organization seemed to require an account in terms of perceptual processes during listening.

Hesitations in the speaker's production and listener's reproduction of utterances

Twenty-four college Ss (encoders) described TAT pictures in short utterances. Each was yoked unsystematically with one of 24 listener Ss (decoders) who heard his recorded utterances and attempted to reproduce them. Words were classified as content or function. While encoders and decoders yielded about the same proportion of content words (41%), encoders yielded a relatively higher proportion of repeats, unfilled pauses, and total hesitations before content words (which have greater uncertainty) than did decoders. Decoders placed relatively more of their hesitations at sentence breaks than did encoders. Apparently, while encoder pauses reflect uncertainty, decoder pauses tend more to mark grammatical boundaries. The selection of semantic-syntactic structure precedes selection of individual words during encoding but follows during decoding.

Rhythm-induced judgments of word stress in sentences

The hypothesis was suggested that the successive sounds of speech have a hierarchical structure that is rhythmic, and that the hierarchical structure is revealed by the rhythmic patterning of the prosodic features in speech such as stress. Some rules for characterizing relative duration and stress were discussed. In the experiment Ss heard two versions of sentences and judged relative stress on the last two words. Sentence versions were copies differing only in rhythmic structure determined by a tape-spliced interval. Judged words were thus acoustically identical in the two versions. Stress judgments varied as predicted by rhythmic structure. Discussion concerned music and speech rhythms, the potential relation between rhythmic structure and the transformational cycle, and the perceptual reality of stress.

Associative strength and word frequency in paired-associate learning

Summary Lists of normatively associated word pairs were learned. Three levels of stimulus-word frequency (Thorndike-Lorge) and three levels of associative strength (relative frequency in word association) were used. Results were that for pairs with high-frequency stimulus words associative strenth had no effect on learning. With lower stimulus-word frequency learning was directly related to associative strength. Learning of pairs not normatively associated was directly related to stimulus-word frequency.

Temporal word spacing and the perception of ordinary, anomalous, and scrambled strings

Ordinary sentences, semantically anomalous sentences, and scrambled word strings were heard behind masking noise in a speech-perception experiment. Time interval between words was .5, 1, or 2 sec. Results were: (a) ordinary sentences were increasingly difficult to hear with increasing intervals between words; (b) anomalous-sentence scores were higher than scrambled-string scores but did not decrease with time interval; and (c) beginning words were heard better than end words in ordinary sentences and scrambled strings but not in anomalous sentences.

Visual Rhythms: Report on a Method for Facilitating the Teaching of Reading

Sentences were prepared such that they appeared on a TV screen with each syllable timed, syllable-by-syllable, as though it were spoken. This was accomplished by synchronizing the onset timing of each syllable as it appeared on the screen with the onset timing of the same syllable as it was heard through the speaker. The result was a sentence which “grew” left-to-right across the screen and which could be presented either visually by itself or in combination with its auditory counterpart. Primary-school children were exposed in training sessions to these sentences in “visual rhythm.” Their pre-to-post-experimental change in reading fluency was compared with that of other children in a control condition whose training was the same except that the sentences they saw were not rhythmic, but static. Positive results encouraged the discussion and rationale for further applications of the method.

Anticipatory Coarticulation and Reaction Time to Phoneme Targets in Spontaneous Speech

The subjects task was to listen to continuous utterances and to press a reaction-time (RT) button upon hearing an assigned phoneme target. Stimulus materials were utterances of 7--13 syllables in length excised from tape-recorded spontaneous speech; each contained a target located in early, middle or late positions in the utterance. Either 200 msec of silence was tape-spliced into the utterance 33, 67, or 100 msec prior to the target, or the utterance was left intact (as spoken). The main results were faster RT to targets following the silent interval by 33 msec than to targets in the intact version or targets following the silent interval by 10 msec, that is, facilitative effects on target RT varied directly with proximity to target of silent interval. As with previous results using citation-form stimulus materials, these results were interpreted in terms of information from anticipatory coarticulation provided to the subject in advance, permitting extra time to allow the target to be anticipated across the silent interval.

Word-association frequency and the proximity effect

Summary Twenty KR stimulus words and a normative response to each were aranged on four word-association tests such that the response word appeared either 5, 10, or 20 words before, or 20 words after, the stimulus word. Appearance of response words at a distance of 10 words or less before the stimulus words increased the frequency of these responses.

Perception of anticipatory coarticulation effects in /stri, stru/ sequences

These experiments assess the identity and temporal reach of anticipatory coarticulation effects in /stri, stru/ sequences which can influence perception. Spoken /stri, stru/ pairs were cross‐spliced after /s/ and after /t/ so that any coarticulatory information preceding the crosspoint incorrectly predicted the remainder of the sequence. Results were that recognition time (RT) to final vowel targets /i/ or /u/ was slower in crossed compared to intact (as spoken) sequences. Differences in RT between crossed and intact sequences were much smaller when crossing after /s/ than after /t/. Further work examines the relative contributions of acoustic information throughout the /str/ region. These experiments and others reported earlier on vowel‐to‐vowel effects show that the perceptual effects of anticipatory coarticulation are not limited to information in adjacent segments. Theories may treat context effects from nonadjacent segments as unwanted or unaccounted‐for variance, but there is little reason to believ...