An Accessory Study of "phonetic Symbolism"

In 1929 Sapir published a brief study from which he concluded that certain vocalic and consonantal sounds have a definite symbolic significance unrelated to the associative and linguistic value commonly attached to words.' His method consisted in presenting two 'unfamiliar' words (e.g. mal and mil), to which a meaning (e.g. 'table') was arbitrarily attached. His subjects reported whether mal symbolized a larger or a smaller table when compared with the word mil. The results indicated that there was a tendency for vowels near the a-end of the vocalic scale to imply 'largeness' and for those near the i-end to imply 'smallness,' quite apart from linguistic uses of these sounds. A like reference was also discovered for certain consonants. In the immediately preceding article in this JOURNAL S. S. Newman has, at the instance of Professor Sapir, again attacked the problem of phonetic symbolism, applying to his research a method of statistical treatment not used in the earlier study, and using a much larger body of material than the earlier study commanded.2 Newman's results generally confirm Sapir's earlier conclusions, making the interpretation much more specific. Newman again makes phonetic symbolism apply differentially to 'large' and 'small' and adds also the contrasting pair 'bright' and 'dark.' The distribution of his vowel-sounds and consonant-sounds upon the scale of magnitude (large-small) and the scale of brilliance (dark-light) leads him to believe that it is determined by such mechanical factors as position of the tongue in articulation, resonance within the mouth, the size of the oral cavity, and length of the vowel sound. Some differences in these factors with respect to their influence upon the two scales are thought to be significant. Upon the appearance of the first article, we were led to undertake a bit of experimentation upon the subject, partly to adapt the method to the psychological laboratory and partly to discover how far "phonetic