Sociolinguistic Variation in American Sign Language

Chapter 6

142 Chapter 6 Grammatical and Social Conditioning of Phonological Variation In chapters 4 and 5 we looked at the patterns of variation exhibited by three phonological variables: 1 handshape signs, deaf, and the location of signs represented by know. We saw that all three can be considered classic sociolinguistic variables, in that the variation that they exhibit correlates with both linguistic and social factors. That all three variables exhibit significant correlations with both linguistic and social factors is not at all surprising. As we have seen, these kinds of correlations have been described extensively for variables in spoken languages, and we are not at all surprised to find them in sign languages as well. What is striking and unexpected, however, is the consistent and strong role of grammatical factors in accounting for the behavior of all three phonological variables. Also of interest is the role of social and historical factors unique to the American Deaf community across all three variables. In chapters 4 and 5 we examined each of the three phonological variables separately; in this chapter we consider them together, first in terms of grammatical conditioning of the variables and second in terms of the sociohistorical context in which they occur. GRAMMATICAL CONDITIONING OF PHONOLOGICAL VARIATION Sociolinguistic research on spoken languages has shown that linguistic variables may be systematically conditioned by factors operating at different levels of the linguistic system. For example, as we saw in chapter 1, numerous studies have shown that -t,d deletion in English is systematically conditioned by the preceding and following phonological environments , stress, and the grammatical category of the word containing the cluster (e.g., Baugh 1983; Guy 1980; Wolfram 1969). Other sociolin2714 GUP SVA Chapter 06 6/14/01 9:22 AM Page 142 Conditioning of Phonological Variation : 143 guistic variables, such as verbal -s in English, are also constrained by both phonological and grammatical factors (Godfrey and Tagliamonte 1999; Poplack and Tagliamonte 1989), as is final -s aspiration and deletion in Puerto Rican Spanish (Poplack 1979). Although the fact that many sociolinguistic variables are constrained by factors operating at different linguistic levels may be a commonplace for students of spoken languages, phonological variation in ASL and other sign languages has heretofore been accounted for by positing phonological constraints alone, particularly the features of the preceding and/or following segments, without reference to structures other than the sequence of phonological segments. The program of research on ASL until very recently has been to demonstrate that ASL, and by analogy other sign languages, are true languages. This work has proceeded by demonstrating that the structure of ASL parallels that of spoken languages and that its phonology and syntax are subject to the same kinds of processes that operate in spoken languages. In the process, this work has not considered the possibility that factors at different linguistic levels may constrain phonological variation. For example, as we saw in chapters 4 and 5, Liddell and Johnson (1989) explain variation in all three of the variables examined here—1 handshape, deaf, location signs—exclusively by reference to features of the preceding and/or following segments. The results of our analysis do not support Liddell and Johnson’s claims. Recall that the core of our analysis of each variable involved identifying the linguistic factors that govern the observed variation. Following the model provided by studies of spoken language variation and heeding earlier claims about variation in sign languages, we hypothesized that features of the immediately preceding and following phonological environment would play key roles. For example, we assumed that the location of preceding and following signs would be important for understanding the variation in deaf and in the location of signs such as know; we assumed that the handshape of the preceding and following signs would play a role in the variation of 1 handshape signs. We therefore included factor groups consisting of the features of the preceding and following segments. However, Lucas’s earlier analysis of deaf (1995), albeit with a small number of tokens, had alerted us to the possible role played by grammatical function in explaining the variation. That analysis , based on 486 tokens, found the syntactic category of deaf to be the only significant linguistic factor, with adjectives favoring noncitation forms, predicate adjectives slightly disfavoring them, and nouns strongly 2714 GUP SVA Chapter 06 6/14/01 9:22 AM Page 143 144 : c o n d i t i o...