Table 1.3 summarizes the internal constraints on variable units.

Table 1.3. Internal Constraints on Variable Units

Earlier studies of variation in ASL focused on compositional constraints. That is, the variation was seen to be conditioned by some feature of the variable sign itself, as described earlier in Battison, Markowicz, and Woodward (1975). Sequential constraints are those that have to do with the immediate linguistic environment surrounding the variable, such as the handshape or palm orientation of the sign immediately preceding or following the variable sign, as we see with 1 handshape signs. Grammatical category constraints have to do with the role that the sign’s grammatical category plays in the variation, as seen in Lucas (1995). The constraint of structural incorporation has to do with the preceding or following syntactic environment surrounding the variable. We will be considering structural incorporation as a constraint as we try to understand what conditions the variable subjects in plain verbs (e.g., pro.1 like vs. [pro.1] like). Finally, pragmatic features may act as constraints. Hoopes (1998), for example, found that the lengthening of a sign for emphasis played a role in the occurrence of pinky extension. Emphasis is a pragmatic factor, a feature chosen by the signer in a particular context to convey a particular meaning. It is not an inherent feature of the sign.

The results of Lucas’s (1995) and Hoopes’s (1998) studies of deaf and of pinky extension show us that the analysis of internal constraints on variation in ASL needs to proceed with caution because the identification of such constraints is not always completely straightforward. Although casual observation might suggest the presence of phonological constraints, further examination might well reveal functional constraints (as in the case of deaf) or pragmatic ones (as in the case of pinky extension). Furthermore, a possible fundamental difference between sign language variation and spoken language variation is emerging from the analysis of internal constraints. This difference relates to the fact that variation in spoken languages is for the most part a boundary phenomenon—that is, a phenomenon that affects linguistic segments that occur in sequence, segments occurring at the boundaries of larger units (i.e., words). And as Wolfram (1974) and Guy (1991) have found, one constraint on -t,d deletion in English, for example, was whether the -t or -d in question was a past-tense morpheme, that is, an affix.