Published online by Cambridge University Press: 22 September 2009
The main purpose of this chapter is to continue to reassess the development of TGG in the light of advances in the formal sciences. Consequently, Chomsky's writings from the years 1955–1957 will be the focus of the discussion, although, as previously, connections will be made between Chomsky's work and that of both his predecessors and his contemporaries. To this end, in section 5.2 Chomsky's rejection of stochastic grammars is considered as part of his assertion that syntax can be studied autonomously. In section 5.3 his redefinition of syntactic research is assessed, particularly the recommended shift away from discovery procedures and towards evaluation procedures. Since certain of the arguments that caused Chomsky to reconsider the role of discovery procedures in linguistic research were initially articulated by Goodman and Quine, it is necessary to explore the influence of constructional system theory upon the approach to syntactic analysis outlined in LSLT and Syntactic Structures (hereafter SS), and this is achieved in section 5.4 where it is shown that Chomsky's definition of a linguistic level is derived from constructional system theory. Given the name of the syntactic theory discussed in this book, ‘Transformational Generative Grammar’, it is necessary to discuss both the concept of ‘transformation’ and the process of ‘generation’. Accordingly, the complex evolution of grammatical transformations in syntactic theory is traced in section 5.5, while in section 5.6 the generative role of recursive definitions in TGG is considered.