Aphasia & Syntax working draft

William Matchin

The representations, operations, and principles of syntactic theories are generally held to be claims about how language is actually implemented in the human brain (Chomsky, 1965; 1995; Sprouse & Hornstein, 2016). For this reason, there is powerful potential for research on the nature of linguistic deficits due to brain damage, or aphasia, to inform syntactic theory. This is particularly so given that there exist disorders that appear to impair core aspects of language, such as agrammatism. Likewise, researchers and clinicians that seek to characterize the deficits in patients with aphasia and to develop assessment and treatment protocols can in principle greatly benefit from the insights into the nature of language provided by syntactic theory. However, there is currently little interaction between theoretical syntax and aphasiology. This is likely due to several reasons, including sociological ones such as the lack of researchers proficient in both fields and ineffective communication among researchers from these very different traditions. However, we suspect that there are deeper reasons for this disconnect. In particular, we suggest two fundamental obstacles: (i) a lack of insight into how grammatical operations apply to real-time sentence processing, and (ii) a focus by syntactic theories on grammatical operations, principles and modules that do not line up well with the currency of functional neuroimaging and neuropsychology: the cortical area. In addition, the assumption that 'agrammatism' is a syndrome caused by a single underlying cognitive source potentially related to a syntactic module is likely false, as is the assumption that damage to Broca's area is necessary and sufficient to cause agrammatism and/or Broca's aphasia. These are related to issues that have been raised by previous authors (Badecker & Caramazza, 1985; Embick & Poeppel, 2005; Embick & Poeppel, 2015; Mohr et al., 1978), and we reinforce them here. In this chapter we will first outline the methods of research in aphasia and how they have been applied to syntax. Following this, we will review the history of the interaction of these two fields, particularly with respect to the putative syndrome of " agrammatism " that is most relevant to syntactic theory. We will make key observations about the successes and failures of this research. In light of these failures, we propose splitting agrammatism into at least two separate syndromes: one that is tied to deficits in domain-general verbal working memory resources, and another that is tied to a content-addressable memory (CAM) retrieval system operating over syntactic features (McElree et al., 2003; Lewis & Vasishth, 2005). This distinction allows us to capture aspects of agrammatism that appear to be domain-general as well as those that appear to be specific to language. We then suggest some helpful steps to reconnect syntactic theory to the study of aphasia.