Speaking logically
By Salvatore Pistoia Reda (Beatriu de Pinós fellow at Universitat Pompeu Fabra)
Speaking a language does not involve logical reasoning. In linguistic theory, this widely accepted assumption entails that expressions are syntactically formed without reference to their logical properties. As a consequence, a common observation in the generative tradition in linguistics is that the set of linguistic constructions that are possible (or accepted or “grammatical”) in a given language can include expressions that amount to contradictions and tautologies. Familiar evidence involves Chomsky’s celebrated example «Colorless green ideas sleep furiously» (cf. N. Chomsky, Syntactic Structures, 1957) but the number of attested precedents reveals early awareness of the problem, e.g. «It did and did not rain at the same time and place» (cf. J. S. Mill, An examination of Sir William Hamilton's philosophy, 1865), «Diese runde Tafel ist viereckig» (cf. H. Steinthal, Grammatik, Logik, und Psychologie, 1855), «Questa tavola rotonda è quadrata» (cf. B. Croce, Questa tavola rotonda è quadrata, 1905). Interestingly, this observation is also widely accepted in the analytic tradition in philosophy where its implications for a general theory of meaning and of knowledge are thoroughly discussed. According to a Carnapian approach, for instance, while being part of the same formal theory syntax (i.e. the set of formation rules) and logic (i.e. the set of transformation or inference rules) work independently from each other, whence the adoption of what resembles a modular division of labor concerning linguistic components (cf. R. Carnap, Logiche Syntax der Sprache, 1934; Y. Bar-Hillel, Logical Syntax and Semantics, 1954).
More recent discussion in formal semantics however questions the assumption that syntactic formation is independent from logical inference. It is observed, in particular, that there is a subset of analytic (i.e. contradictory or tautological) expressions that seem to be excluded from the language (i.e. from the set of possible expressions in a given language) precisely on account of their logical status. Evidence includes the interpretation of quantifiers when interacting with special linguistic environments, e.g. «Some students except John smoke» (cf. K. von Fintel, Exceptive constructions, 1993), «There is every curious student» (cf. J. Barwise, R. Cooper, Generalized quantifiers and natural language, 1981), «There are any cookies left» (cf. G. Chierchia, Logic in Grammar, 2013). In other words, this alternative perspective submits an analyticity puzzle, i.e. the fact that only some contradictions and tautologies are excluded from the language while other contradictions and tautologies are judged as perfectly acceptable by speakers. In order to respond to the puzzle, efforts are made on the formal side of the debate to formulate and justify theoretically a deductive system that only accesses unacceptable linguistic constructions thus explaining the asymmetric acceptability of analyticities. The proposal is then made that what motivates the exclusion of certain analytic construction from the language is an inferential system which is inherently and essentially linguistic and crucially differs from classical reconstructions in formal logic. As a consequence, this new perspective, while breaking with the generative tradition in linguistics with respect to the relationship between syntax and logic, also conflicts with standard conventionalist approaches in logic and seems to be bound to a substantive interpretation of the inferential system. Current discussion investigates whether this interpretation is indeed obligatory for proponents of the new perspective. In conclusion, the new perspective understandably maintains that speaking a language is not the same as logical reasoning but emphatically submits that nonetheless you cannot really speak illogically.
Contacts: salvatore.pistoiareda@upf.edu | https://salvatorepistoiareda.me/
Social media: Facebook | ResearchGate | Twitter