Where to look for the Fundamentals of Language

Rigor in syntax is widely taken to require constituency-based formalisms. However, classical European structural linguistics provides other ways of meeting the methodological requirements. The advantages of these ways are becoming increasingly evident. In section 1 of this paper, we discuss how this would allow the distinction between the core of language, patterned in a relatively simple way, and its complex periphery to be expressed in terms of the notion of markedness. As Chomsky (2004) claims, the properties of the language core are to be explained as "attributable to general principles" (see section 2). It is the main objective of the present paper to discuss how the core of language, exhibiting unrestrictedly complex sentences, perhaps could be understood as based on common human mental capacities. In section 3, we consider the possibility that the earliest, and therefore the most basic form of combination is based on information structure, that is, on the distinction between topic and focus. Under this hypothesis, syntactic dependency, or valency,might be taken to be a later development, and we discuss this in section 4.1. In section 4.2, we focus these speculations more narrowly by examining two typologically different languages, English and Czech within the framework of Functional Generative Description (FGD) developed in Prague. In section 5, we show how an empirical foundation might be provided for this view of language by examining underlying syntactic annotations in the Czech Dependency Tree Bank. We show that, while the core of language then can be found to be patterned in a relatively very simple way, coming close to the elementary logic that is presumably a shared human capacity, the complex periphery of the language system is extremely large and complicated (section 6).