Phonological blending or code mixing? Why mouthing is not a core component of sign language grammar
The level of integration of mouthing into the sign language system was investigated using a novel experimental procedure. We constructed a word/sign matching task in which the signer has to indicate whether a LIS (Italian Sign Language) sign matches the written Italian word that follows the video presentation of the sign. In the congruent condition, the word matches the sign, while in the incongruent condition the word matches a sign which forms a minimal pair with the sign that has been presented in the video. To form minimal pairs, all four traditional formational parameters for signs plus mouthing were considered. Lip movements were present only in mouthing minimal pairs. In the incongruent condition we compared mouthing minimal pairs separately to handshape minimal pairs, location minimal pairs, movement minimal pairs, and palm orientation minimal pairs. Accuracy was markedly lower for minimal pairs distinguished by mouthing than for minimal pairs distinguished by one of the four parameters. In the congruent condition we compared mouthing minimal pairs to all the other minimal pairs, in which lips movements were absent. Reaction times were shorter in the presence of mouthing as a consequence of the strong mapping between orthography and mouthing, confirming that mouthing is highly connected to the Italian lexicon. Participants seem to consider mouthing external to the sign to be matched with the word. We propose that cases of disambiguation by mouthing should be interpreted as cases of simultaneous code mixing. Therefore, our experimental results suggest that mouthing is not a core component of sign languages.