No abstract

No abstract

No abstract

MANY STUDIES (reviewed by Taylor, 1962) have indicated that there is consistency in the choice of meanings given to nonsense syllables by different people. This consistency, known as phonetic symbolism, has been attributed to qualities inherent in the sounds themselves. For example, Newman (1933) suggested that nonsense syllables with the vowel i are perceived as smaller than similar syllables with the vowel U, because of the higher pitch, or smaller vocal cavity associated with i. Any hypothesis of this nature leads to the prediction, to date untested, that the symbolic values of sounds must be similar in all languages. The present paper reports an experiment in which the symbolic values of nonsense syllables were determined for four unrelated languages on four different dimensions of meaning. Also reported are two preliminary experiments using English-speaking subjects, in which several factors affecting phonetic symbolism, and relevant to the design of die main experiment, were tested. PROCEDUREThe present investigation is analytical in the tradition of Sapir (1929) and Newman (1933). There were three experiments, in which the same general procedure was used, with same differences according to the specific requirements of the particular experiments. In each experiment the test material was composed of CVC (consonant-vowel-consonant) nonsense syllables mimeographed on three sheets of paper. A test booklet was made by stapling a page of instructions in front of the three randomly ordered pages of CVC syllables.In each experiment S was required to rate each syllable on some dimension of meaning, such as warmth or size, using a five-point rating scale. To rate a syllable S made a check mark on one of five dashes which followed each syllable. The S was reminded of the dimension to be rated, and the direction of his rating task with such an illustration as "small large" on the top of each test sheet There was no time limit for the task. For the purpose of statistical analysis, the five-point rating scale was used without transformation throughout the experiments. 1 The study reported here is based on part of the doctoral dissertation submitted by the senior author to the Johns Hopkins University. The authors wish to express thengratitude to Dr. James Deese for his valuable assistance.

No abstract