Cognitive Mechanisms & Syntactic Theory: Arguments against Adjuncts in the Lexicon

As a psychologist who studies sentence comprehension and holds a joint appointment in the departments of psychology and linguistics, I have frequent opportunities to observe the interaction, or lack thereof, between the two disciplines. Although cognitive psychology and formal linguistic theory share some common history in Chomsky's (1959) pivotal review of Skinner's (1957) book on language behavior, these two disciplines have not continued to influence one another to the degree one might expect. For example, theoretical developments in syntax have rarely if ever been motivated by an experimental finding about sentence comprehension. In fact, there is good reason for this. I will argue that most psycholinguistic data is irrelevant to formal linguistic theory. Nonetheless, there may be a subset of psycholinguistic data that formal linguists ought to consider. I will attempt to delineate this largely hypothetical subset, using as an example some actual psycholinguistic data on the argument/adjunct distinction. First, let us consider the domains of cognitive psychology and formal linguistic theory. Cognitive psychology encompasses the study of mental representations and the mental operations for manipulating (creating, accessing, etc.) these mental representations. The central goal is the description of a processing system. In the context of sentence comprehension, for example, a cognitive psychologist might develop a theory of syntactic parsing within sentence comprehension that specifies what types of mental representations are involved, what aspects of linguistic and non-linguistic knowledge are used to create those mental representations, and so forth. In this chapter, I will consider experimentally-oriented psycholinguists like myself to be cognitive psychologists, regardless of their academic credentials. Formal linguistic theory encompasses the study of what speakers know about their language. The central goal is the development of the simplest and most elegant description of this knowledge (i.e. grammar). By placing a premium on elegance, linguists hope to characterize the properties of the human mind that make language possible and not merely catalogue facts about languages. In addition to elegance, grammars must achieve descriptive adequacy by generating all and only those utterances accepted by native speakers. Descriptive adequacy is evaluated by analyzing human behavior, i.e., linguistic intuitions. However, descriptive adequacy receives somewhat less attention, perhaps because it is difficult to thoroughly evaluate the descriptive adequacy of a grammar for a particular language and completely impossible to fully evaluate its descriptive adequacy for all existing and possible languages—although this is understood to be the ultimate goal. Descriptive adequacy is assessed via …