Figure 2 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Source publication
Contrary to the prevailing assumption that language is “primarily a tool for communication rather than thought”, I argue that language is, to invoke Oscar Wilde, “quite useless”. Arguing from aesthetic philosophy and the minimalist program for linguistic theory, I conject that language, like art, is not “for” anything—it simply is , conforming to a...
Citations
... As Watumull (2024) vividly illustrates, language as an internal system "exists for its own sake" though crucially it can be used by humans to "[unify] the worlds of abstractions, minds, and bodies" (Watumull, 2024, 2). The language-for-communication thesis is deficient, attempting to reduce language use to serve as a tool for mere expressiveness (Watumull, 2024, 15-16). ...
Some generative linguists have long maintained that human beings exhibit a species-specific form of intellectual freedom expressible through natural language. With roots in Descartes' effort to distinguish humans from machines, it is the human use of their linguistic capacity that demonstrates a unique intellectual and creative potential. This "creative aspect of language use," or CALU, is typically conceived as neither determined nor random yet appropriate to the circumstances of its use. CALU is therefore taken to be beyond the scope of computation. This paper argues that there is substance to this claim and that CALU is likely an unreachable threshold for artificial intelligence (AI) developed via computational means. Background is provided by distinguishing the Cartesians' original "language test" from Turing's later "Imitation Game" for machine intelligence. Moreover, it is shown how Turing's work on computability theory allowed generative linguists to reformulate the original Cartesian problem of language use in the twentieth century. The paper proceeds to show why generative Large Language Models do not reproduce CALU. Then, it argues that characteristics relevant to human intelligence such agency and embodiedness-while conceivable in future AI systems via computation-are insufficient for attaining CALU. Language use's non-teleological nature, in contrast to these characteristics, poses a formidable challenge to AI via computation. Finally, clarifying remarks are offered on the problems posed by testing for CALU, with contrasts to Turing's Imitation Game and tests designed in its lineage.
... Compared to the previous volume, book reviews have made a comeback with reviews of Merge and the Strong Minimalist Thesis (Chomsky et al., 2023;reviewed in van Gelderen, 2024) and The Philosophy of Theoretical Linguistics: A Contemporary Outlook (Nefdt, 2024;reviewed in Voudouris & Roe, 2024). Lastly, our Forum section has again seen lively discussion with contributions questioning how to evaluate the language abilities of LLMs (Leivada, Dentella, & Günther, 2024), arguing that so-called "laryngeal descent theory" for the origin of speech was actually never a popular line of thinking in speech-centric sciences (Ekström, 2024), and providing a minimalist perspective on ongoing discussions about the purpose of language claiming that language evolved primarily as a tool for communication (Watumull, 2024). ...
