ALEF

ALEF warm-up before next season (exclusively in-person). See you there!

Title: Inferentialism and AI

Abstract: Large language models succeed at many linguistic and reasoning tasks, yet it remains unclear what, if anything, their internal workings amount to in semantic terms. This talk examines whether contemporary AI architectures can sustain a notion of meaning that is public, rule-bound, and inspectable—features typically expected by inferentialist theories of meaning.

I contrast two approaches. On the one hand, distributional semantics treats meaning as emerging from patterns of linguistic co-occurrence encoded in high-dimensional vector spaces. This view has been enormously successful in practice, but its semantic commitments are not entirely transparent: the structures that drive model performance are neither rule-governed nor readily interpretable as articulable commitments and entitlements.
On the other hand, inferentialism—especially in its proof-theoretic variants—ties content to the network of inferences a speaker is entitled to draw. This framework makes semantic norms explicit and publicly assessable, and it provides tools for distinguishing correct from defective inferential behaviour.

The core question is whether current AI systems instantiate anything like such normative, rule-structured inferential roles, or whether they merely approximate them statistically.

https://philevents.org/event/fileDownload/141681?fileId=19097

Here we go again! The CFA is open.

See you at 6! Here is the link:
Zoom: https://us02web.zoom.us/j/83691371602?pwd=Ym1tQWhIWTlTbDJ4YmNwNFRRaFBIZz09

Meeting ID: 836 9137 1602
Passcode: 097096

See you Friday! Here is the abstract for the talk:

Zerilli (2022) and Frankish (2024) have argued that one should adopt Dennett’s (1989) intentional stance when interacting with Large Language Models (LLMs). When adopting the intentional stance, one interprets the behaviour of the system, assigning beliefs, goals and desires to gain predictive and explanatory power of the system’s behaviour. The goal of this paper is to argue that the adoption of the intentional stance towards LLMs is unwarranted. First, I introduce two cases that show Zerilli’s conjunction to be neither a sufficient nor a necessary prima facie condition for the adoption of the intentional stance. Second, I argue that current LLMs fail a necessary condition of interpretability: rationality. Third, I argue that Frankish and Zerilli’s strategy to assign desires and goals to LLMs is ultimately reducible to the adoption of the design stance. Fourth, I defend the adoption of the design stance and defend it against possible objections and criticisms. Ultimately, I argue that the intentional stance is unwarranted, and that the appropriate stance to adopt towards LLMs is the design stance.

See you at 6!
https://us02web.zoom.us/j/83691371602?pwd=Ym1tQWhIWTlTbDJ4YmNwNFRRaFBIZz09

Meeting ID: 836 9137 1602
Passcode: 097096

Abstract: Suppose a causal theory of reference for utterances of proper names is true. According to such theories, there are two important parts of the reference mechanism: baptisms, in which a name is bestowed upon an object, enabling later reference, and borrowing or inheritance, where later tokens of the same name refer to that object because they are appropriately related to the baptism. Much work has been done on how the later tokens must relate to the baptism. Surprisingly, very little work has been done on the nature of baptisms themselves, or even to describe desiderata for what a good theory of baptisms would be like. I offer both. This talk examines the two received approaches to baptisms across the literature and their respective shortcomings, some more sophisticated theories one might hold, and their problems, and finally my own theory which accounts for all the difficulties of the other views. On my theory, a baptism is an act (possible an 'extended' or 'composite' act) which institutes a constitutive rule in virtue of which later utterances of the relevant name will refer to the relevant object. I demonstrate that this view can account for all the cases I discuss.

See you at 4! https://us02web.zoom.us/j/83691371602..
Meeting ID: 836 9137 1602
Passcode: 097096

We are going to start 2 hours earlier than ususal. See you tomorrow at 16!

We will meet both in person (Faculty of Philosophy - Bucharest) and online.

Abstract: I start by arguing that the Knower paradox offers underappreciated resistance to standard nonclassical approaches to the paradoxes of selfreference, because of the divergence they require between rejecting a sentence and accepting its negation. I move on to address a problem which the Knower paradox raises for my own favoured noncontractive, instability-based approach, and which consists in the apparent lack of a suitable principle of “epistemic ascent”. Shifting from logic and metaphysics to pragmatics, I defend the coherence and reasonableness of accepting sentences that express unstable states-of-affairs. Consequently, appealing to unusual conceptual resources, I close by revealing whether paradoxical sentences are true or not…

The microphone has stopped working unexpectedly. The online broadcast of the talk is cancelled. We apologize for any inconvenience. See you Friday!