This paper addresses two central puzzles concerning the limits of sense. First, how can we meaningfully speak about transgressions of this limit without either speaking nonsensically ourselves, or inadvertently turning nonsense into sense? And second, what meaning do sentences like tautologies and contradictions have that, arguably, constitute the limits of sense? This paper sets out to provide answers to both puzzles from a noncognitivist perspective inspired by Carnap’s (1950) theory as developed in his Empiricism, Semantics, and Ontology. The central idea of the given analysis is that the statements in question communicate the speaker’s pro- or con-attitude toward a way of speaking. These statements thus mainly fulfill an expressive, rather than a descriptive function.
When a language model invents a citation, philosophers ask what kind of bad language this is. The most popular answer says it is bullshit: speech produced without caring whether it is true. I argue the answer is wrong, and Wittgenstein shows why. On the resolute reading of the Tractatus, which the Investigations keeps, nonsense has no kinds. A string is nonsense only when nobody has given it a use. Meaning is something people do with words inside a shared practice. A language model takes part in no practice. It cannot care about truth, or fail to care, because it does not assert at all. Its output, as produced, is neither sense nor nonsense nor bullshit. Those verdicts appear only when a reader treats the string as something said. A hallucination is a claim the reader made on the machine's behalf, and it failed. The meaning, and the responsibility, are ours.
DOI: http://doi.org/10.26333/sts.xxxix1.12 Reply to commentators in Semiotic Studies special issue on Unreal Beliefs. I defend the position developed in this book, i.e., minimal non-realism, against various charges. First, I defend the way I reconstructed folk commitments to realism about beliefs. Second, I deal with the arbitrariness charge against my definition of belief realism. Third, I critically evaluate a counterproposal to my view, namely the idea that beliefs are socially constructed natural kinds. Then I deal with the challenges levelled against my arguments against realism about belief. Finally, I reconsider some of the consequences of minimal non-realism and consider potential alternatives to it.
In this text, I am going to discuss when things go wrong when using language. In particular, I would like to address a special kind of ungrammaticality: contrary to what is expected within the generativist tradition, there seem to be word-sequences that are syntactically unacceptable for logical reasons. The semantic works exploring this line of research are usually grouped under the rubric “logicality of language hypothesis”. The aim of this paper is to examine the recent literature that attempts to subsume the observed ungrammaticalities under the umbrella of semantic anomaly, and thus to offer alternative (typological) explanations with respect to language logicality. Some critical issues of these alternative explanations will be highlighted and, while also drawing attention to the weaknesses of the logicality of language hypothesis, the latter will be re-proposed as an explanation that can be sensibly supported (for the time being, at least).
The article addresses the issue of nonsense, in a detailed and comprehensive manner scrutinised in the book Nonsens i zjawiska pokrewne. Studium semantyczne [Nonsense and Related Phenomena. A Semantic Study] (Wo & lstrok;k, 2014). The aim of the present text is to juxtapose elements that make it possible to determine convergences and divergences between the category of nonsense in philosophy and logic and the expression of general (more precisely: Polish) language, which undergoes linguistic research. The study focuses on a structuralist linguistic perspective. The analysis is based on a distinction which is crucial in the description of language: a term vs. an expression of general language. In line with this distinction, the rules of defining a word functioning in two registers-in specialist language and in general language-differ. This is due, among other factors, to differences in the functioning of such a word. The undertaken linguistic research is preceded by a theoretical and methodological introduction, the latter outlining the adopted assumptions and linguistic description methods, as well as the theories employed for the proposed analysis (including the concept of the unit of language). The results of the study, with a synthesis of relationships established between the compared categories, are presented in the conclusion.
