Declarative Utterances Research Articles

Declination of fundamental frequency (F) in speech production

Declarative utterances in English are typically accompanied by a gradual long-term fall in F0. It has been suggested in previous literature that this fall may be represented as a linear function of time, on which various inflections are superimposed. This study was designed to determine the form of F0 declination in phonetically matched sentence materials. Ten practiced speakers read the sentences, which were subsequently analyzed for F0 and amplitude using computerized techniques. The results indicated that, for a variety of declarative sentences, the declination could be approximated by a piecewise linear function in which two-thirds of the F0 fall occurred during the first one-half of the utterance's duration. This piecewise linear function provided a close fit to the data for each of the ten speakers, including both males and females. In additional experiments, F0 declination was examined as a function of systematic variation in syntactic structure, constituent length, and speaking rate. The results of these experiments permit inferences about the organization of the speaker's coding operations, including look-ahead schemes. Implications for speech synthesis-by-rule will also be discussed. [Supported by NIH Grant NS-13028 and an NIH postdoctoral fellowship.]

On the “Constant Total Amplitude” Theory of Final Syllable Lengthening

It has been suggested [B. Lindblom, “Temporal Organization of Syllable Production,” Royal Inst. Technol., Stockholm, Sweden, STL-QPSR 2, No. 3, 1–5 (1968)] that the greater duration of final syllables over nonfinal syllables is caused by a tendency on the part of speakers to maintain a constant total energy (amplitude) output on syllables of the same degree of stress. Since utterance-final syllables characteristically show drops in amplitude per unit of time, they are lengthened, according to the theory, in order to compensate for the amplitude drop. In the present study, it is shown that no amplitude drop occurs on the final syllables of interrogative utterances in English. It is further shown that final-syllable lengthening is just as great for interrogative as for declarative utterances. The theory cannot account for this result. Alternative explanations of finalsyllable lengthening are discussed.

Interrogatives, Imperatives, Truth, Falsity and Lies

This paper aims to establish three major theses: (1) Not only declarative sentences, but also interrogatives and imperatives, may be classified as true or as false. (2) Declarative, imperative, and interrogative utterances may also be classified as honest or as dishonest. (3) Whether an utterance is honest or dishonest is logically independent of whether it is true or is false.The establishment of the above theses follows upon the adoption of a principle for identifying what is meant by any sentence, declarative, interrogative, or imperative. The analysis aims to show that meaning is to be attributed to the uttered or written sentence-token, rather to the thereby exhibited sentence-type. Further, the meaning of the sentential token is to be identified with a purpose of the speaker, that the speaker would reveal to the addressee by uttering the sentence.The to be revealed purpose is analysed into two components: an ultimate concern (that the addressee stand in such and such a relation—e.g., of believing, or informing the speaker about, or making it true that) and an ultimate topic of concern (the state of affairs, i.e., proposition, relative to which the speaker would have the addressee stand in the specified relation).Sentential utterances “signify” different purposes by “expressing” different ultimate concerns and “indicating ”different ultimate topics of concern. Variations in expressed concern are correlated with variations in sentential form, such as declarative, interrogative and imperative. Variations in indicated topic of concern are correlated with variations in the subject and predicate of the uttered sentence. Thus, for example, utterances of “Johnny will jump in the lake,” “Will Johnny jump in the lake?” and “Johnny, go jump in the lake! ” all indicate one and the same ultimate topic of concern but express different ultimate concerns with this topic.A sentential utterance is true or false according as its indicated topic of concern is true or false. Hence, declaratives, interrogatives and imperatives may all be classified as true or as false. But honesty or dishonesty is a function (explained in the paper) of the expressed concern, rather than of the topic of concern. Hence, although utterances of all sentential forms are honest or dishonest, their honesty or dishonesty is logically independent of their truth or falsity.