Mirco Sambrotta

Can Large Language Models (LLMs), such as ChatGPT, be considered genuine language users? Can they truly understand the meanings of natural language? This paper adopts an inferentialist perspective, arguing that grasping the meaning of an expression is nothing but grasping the inferential role the expression plays. But roles are conferred by rules. An expression’s contentfulness consists of its use being governed by inferential rules. Meaningful items incorporate norms of inference, which they are subject to. Thus, grasping meanings is mastering this bundle of rules. The paper questions whether LLMs, despite their ability to generate coherent text and perform inference-like tasks, follow such rules. However, unlike humans, LLMs are trained to detect statistical patterns. While it might be suggested that LLMs ‘know’ which words typically follow others based on statistical patterns, the norms they learn to master differ fundamentally from the norms of inference that govern language use. The paper concludes that the current version of LLMs, despite their advanced language processing capabilities, do not genuinely grasp or understand conceptual content and should, therefore, be viewed as simulations of language users rather than true participants in the logical space of reasons.

In most of her publications, Amie Thomasson develops and defends what she calls a ‘easy’ approach to ontology, which leads to both a simple ontological realism and a form of metaontological deflationism. In her view, very roughly, the ontological judgment that some first-order claim properly describes worldly facts, or depicts what really exists, adds nothing to the first-order claim itself. With the former, one merely reaffirms the latter, without taking on any burden of more substantial explanations. For the latter is not supposed to avert to some deep or substantive criteria that might fail to be met. As a result, ontological questions are easy to answer, and ontological disputes are misguided. The aim of this paper is to offer an account that does justice to ontological discourse. This, though, without bearing the burden of substantial (i.e., represantionalist) standpoints. According to the proposed account, neither questions of that sort are trivial nor are ontological disputes misguided. Instead, ontological judgments serve a significant normative role in shaping and arranging our discursive practice. In particular, they play a distinctive kind of expressive function: that of expressing commitments to the adoption of conceptual structures within which alone descriptions may occur. Ontological discourse thus does not aim to describe, but rather, by showing what should and should not count as reasons or evidence, it articulates the conditions and common ground that make description possible.

This book offers an insight into some of the central topics in the current panorama of metaphysics. Thorny issues, such as the ontological state of fictional entities, abstract objects, and possible and impossible worlds, are addressed through an open conversation with Amie Thomasson, one of the leading figures in the current scene and landscape of metaphysics. This collection of heterogeneous and critical essays on her work, and her replies to them, stem from the SIUCC (Seminario Interuniversitario de Ciencia Cognitiva/Inter-University 3-day Workshop on Philosophy and Cognitive Science). By means of this interaction between a group of top scholars and Amie Thomasson, the book introduces the philosophically curious to the way the "first philosophy" is carried out today. This volume is of great interest to academics, philosophy majors who are taking an upper-division metaphysics class, and members of the general public who are curious about the work of Amie Thomasson.

The aim of this paper is to explore the possibility that at least some ontological dispute are better understood as what David Plunkett and Timothy Sundell have called ‘metalinguistic negotiations’. I will take the debate between the dominant approaches of realism and anti-realism (especially fictionalism) about the ontological status of scientific models as a case-study. I will argue that such a debate is best seen as normatively motivated, insofar as a normative and non-factual question may be involved in it: how ought the relevant piece of language to be used? Even though I will generally assess the prospects for a broadly deflationist approach, I shall outline a sense in which such a dispute can be recast as ‘minimally substantive’.

This paper aims to shed some light on the metaphysical inquiry about the nature of intentional (and fictional) objects by putting forward a semantic analysis of the expressions we use to refer to them. I argue that such terms are formed by nominalizing expressions that belong to other logical categories than names and other parts of speech than those apt to refer. They then map the presentation of non-full-fledged referential terms into objects. However, in playing this function, their use involves a distinctive kind of metalinguistic expressive role: that of being elaborated from and explicative of pragmatic features of the underlying nominalized expressions. As a result, they turn out to be merely ostensible singular terms, which refer to objects in the world only in the broad sense in which the world comprises linguistic norms and roles. Therefore, one does not need to countenance intentional (and fictional) objects to understand and grasp the use of the expressions that purport to refer to them.

The chief purpose of this paper is to advance a defense of the old-fashioned view that empty names are neither proper names nor any other kind of interpretable expressions. A view of this sort usually makes it easy to account for the meaning of first-order sentences in which they occur in subject position: taken literally, they express no fully-fledged particular propositions, are not truth-evaluable, cannot be used to make assertions, and so on. Yet, semantic issues arise when those very sentences are embedded in the scope of propositional attitude verbs. Such (intensional) constructions, indeed, turn out to be literally meaningful, truth-evaluable, and eligible for making assertions. The novel solution put forward here is to combine a version of sententialism with the idea that de dicto reports play a distinctive kind of metalinguistic expressive function. Roughly, that of enabling the ascriber to make explicit a mismatch between the way the embedded sentences are used by the ascribee and the way they are ordinarily used ̶ and, in turn, a mismatch between the way the (empty) names occurring in them are used by the ascribee and the way they are ordinarily used. Fictional names are then regarded as a mere subset of empty names. Accordingly, the above strategy is applied to account for the meaning and use of parafictional (and fictional) sentences and fictional vocabulary in general.

The purpose of this paper is to better understand what ontologists are doing when they ask questions about the categories of the world. I will take Cumpa’s attempts to find out the fundamental structure of the world as a case-study. In one of his latest paper (Cumpa 2014), he conceives the classical ontological question about the existence of the fundamental categories of the world (what are the fundamental categories of the world?) as a question about the category able to unify the two Sellarsian images of the world: the manifest and scientific images, considered as two different languages. According to him, the only category with such an explanatory power is the category of ‘facts’ (or ‘state of affairs’): the fundamental category of what he calls ‘the metaphysical language.’ I will argue that if Cumpa takes the latter to be a broader language or framework, in Carnap’s terms, common to both the ordinary and the scientific ones, then his proposal turns out to be rather problematic (as they are ultimately ‘incommensurable’). On the other hand, if he understands it as external to both of them, then his solution ends up being meaningless and devoid of any cognitive content, with at best a practical character and/or an expressive function.

Carnap famously argued that there are two kinds of questions and claims concerning the existence or reality of entities: internal and external ones. We focus on Carnapian external ontological claims of the form: ‘Xs really exist’, where ‘X’ stands for some traditional metaphysical category, such as ‘substance’, ‘fact’, or ‘structure’. While Carnap considered them meaningless, we consider them faultlessly meaningful. However, in line with an expressivist guise, we do not claim that they have the meaning they have in virtue of representing a certain range of entities in the world or describing pieces of reality. Quite the contrary, we maintain that they are meaningful because they perform a function that is fundamentally expressive, that is, a non-descriptive function. In particular, they may imply a distinctive type of metalinguistic expressive function. Moreover, we avoid accuses of mentalism insofar as we explain the meaning of external ontological claims in terms of the practical commitments to the adoption of certain linguistic forms they express, rather than as expressing some kind of mental state.

A possible way out to Kripke’s Puzzle About Belief could start from the rejection of the notion of epistemic transparency. Epistemic transparency seems, indeed, irremediably incompatible with an externalist conception of mental content. However, Brandom’s inferentialism could be considered a version of externalism that allows, at least in some cases, to save the principle of transparency. Appealing to a normative account of the content of our beliefs, from the inferentialist’s standpoint, it is possible to state that a content is transparent when name-components of that content are a priori associated with some application conditions and, at the same time, reflection alone provides an a priori access to those application conditions, with no need of any empirical investigation. Nevertheless, such requirements are only met in trivial cases. The aim of this paper is to argue that some application conditions of that sort, albeit trivial, can be ontologically ampliative. As a result, the related contents can be regarded as transparent in a substantial and rich way.

The aim of this paper is to explore the possibility that, at least, some metaphysical debates are ‘metalinguistic negotiations’. I will take the dispute between the dominant approaches of realism and the anti-realism ones about the ontological status of scientific models as a case-study. I will argue that such a debate may be better understood as a disagreement, at bottom normatively, motivated, insofar as a normative and non-factual question may be involved in it: how the relevant piece of language ought to be used. Even though I will generally assess the prospects for a broadly deflationist approach, I shall outline a sense in which the dispute can be recast as ‘minimally substantive’.