Undershoot in two modalities: evidence from fast speech and fast signing

This study seeks to extend methodologies previously used to analyze target undershoot in vowels to the analysis of English stop consonants. The same methods are used to elicit variation in the location and handshape parameters of a signed language, specifically American Sign Language (ASL). Research has shown that variation in the F2 values of vowels can be related to speaking rate such that shorter vowels are less likely to achieve the expected target value for that vowel. Further, this kind of target undershoot is dependent on the phonetic environment in which the vowel is placed. Previous research on consonant articulations is mixed. While variation has been shown in most cases, some results conflict. The previous research on variation in signing is less extensive. Studies indicate that variation in signed languages can be linked to aspects of the phonetic environment. A few studies also indicate that some variation is related to signing rate or style. Three experiments were conducted to confirm the existence of rate dependence on English F2 onsets and on the location and handshape parameters of ASL. Speakers and signers were asked to produce short utterances at differing rates of production ranging from normal, relaxed utterances to articulations that were as fast as possible. Rate and context dependence were also shown to fit expectations about phonetic undershoot based on the earlier research on vowel undershoot. Multiple linear regression analyses were used to quantify the effects of speaking or signing rate and phonetic environment on English F2 onsets, and on ASL locations and handshapes. The articulators used for the production of speech sounds and those used for the production of ASL signs are radically different in physical form. Despite the differences between the two systems, there may be important similarities in the motor control of these systems. As a result, constraints on language production in one modality may be quite similar to constraints on production in the other modality. This dissertation shows that undershoot phenomena are similar across the two modalities providing support for the validity of transferring methodologies for the analysis of speech to the study of signed languages.