Tracking the origins of transformational generative grammar 1

Tracking the main influences of 19thand 20th-century mathematics, logic and philosophy on pre-1958 American linguistics and especially on early Transformational Generative Grammar (TGG) is an ambitious cross-disciplinary endeavour. Ideally it would call for expertise in the methods of intellectual historiography, the history and content of 20th-century American linguistics, the history and philosophy of science (including logic and mathematics), the tools and results of mathematical logic, and the theory of computable functions. Scholars fully versed in all of these fields are rare indeed. If Marcus Tomalin makes some mistakes in his book (henceforth LFS), that should not be surprising. What is surprising is how much progress he makes in furthering intellectually serious work on the history of modern linguistics, and how wide his reading in the relevant technical literature has been. LFS locates the intellectual roots of TGG in the methods developed by 19thand 20th-century mathematics and logic for exhibiting the conceptual structure of theories and constructing rigorous proofs of theorems. Tomalin discusses the methods developed by Augustin-Louis Cauchy for the rigorisation of the calculus in the 1820s; Whitehead & Russell’s use of the axiomatic method in Principia Mathematica (1910–1913); the Hilbert program (in the 1920s) to prove all of mathematics consistent; Bloomfield’s early axiomatisation of a general linguistic theory (1926); Carnap’s logical empiricist proposals for the logical reconstruction of science on an experiential basis in the 1920s and 1930s; and Goodman’s (1951) adaptation and revision of Carnap (1928). Earlier histories of TGG have not investigated the linguistic relevance of this literature. Tomalin argues persuasively that the two-level approach to theorising now associated with early TGG (grammars as theories of languages, plus a metatheory about the form of grammars) was a specialised adaptation to linguistics of techniques developed for doing metatheoretical work on mathematics and logic, which were also adopted and applied in the philosophy of science. And he points out that the approach can be found in linguistics from Bloomfield (1926) on. In this article we concentrate mainly on places where we think there are further questions that should be asked, or where we disagree with Tomalin, or where we think he made mistakes. However, we want to emphasise from the start that Tomalin deserves much credit: the scope of his reading of the primary literature is broad and deep, and his book provides a valuable scaffolding for future work in the area, even if it goes wrong in some of the details.