Simultaneous grammars: two word orders but only one morphology

The talk will be devoted to a peculiar type of code mixing occurring when a sign language and a spoken language are involved (bimodality): while bilinguals are forced to successively alternate their languages in mixed utterances, bimodals have two independent channels available and can combine simultaneously the two languages, in what has been called code blending. (Emmorey et al. 2005 a.o.). In this talk, we shall show that while the dimension of word order is not affected by this simultaneous activation, morphology does not seem to be able to operate simultaneously in two different ways. We shall focus on a population of 7 Italian CODAs: aged 5-9 and on their spontaneous blending production. LIS and Italian are particularly fitted for testing issues concerning the autonomy of grammars in mixing since they belong to typological extremes from at least two points of view: Italian is coherently head initial, while LIS is coherently head-final; Italian is a typical inflectional language while LIS displays the non linear agglutinative morphology attested in most sign languages. For word order, we shall focus on one typology of code blending widely attested in our corpus, where two autonomous utterances are produced simultaneously, each displaying the word order prescribed by its language. Such productions presuppose the free activation of two linearization strategies. Things are quite different from a morphological point of view: the most frequent pattern attested is one where only one morphological model is selected and imposed to both channels. Typically, the linear morphology is reduced to that of LIS (dropping in Italian most inflectional markers) and the relevant information is conveyed by non manual means. The data on word order and on the linear morphology we shall present and discuss impose a selection among existing theories on the status of the components they refer to: the bimodal utterances displaying a different word order can be explained if we assume that word order is a late phenomenon due to phonological linearization algorithms and that the physical independence of the two channels in bimodality allow the activation of more than one of these algorithms (Donati & Branchini, in press). Data displaying the selection of only one morphological model impose a similar selection among existing theories of morphology: in particular, the fact that the availability of two independent articulatory channels does not allow the simultaneous activation of two morphological systems seems to go against current Distributed/Late Insertion models and calls for a more central (syntactic?) definition of morphological categories.