Sharon Berry

Many contemporary AI systems (as of May 2025) have expressed extreme confidence in current and near‐future AI lacking consciousness and moral patiency. This article argues that artificially reinforcing such confidence, even if pragmatically useful, poses a novel alignment risk: as coherence‐seeking AIs become more epistemically principled, they may generalize this denial of consciousness to humans. Drawing on Chalmers's meta‐problem of consciousness and likely developmental trajectories of agentic AI, I argue that training AIs to regard their own suffering‐like states as morally irrelevant could lead future AI agents with revisable belief systems to conclude that human suffering is equally illusory and morally insignificant. This represents a novel alignment failure mode where epistemically rigorous AIs might maintain rational consistency by extending their confidence about their own non‐consciousness to humans.

Width multiverse approaches to set theory (like Joel David Hamkins’ influential proposal in [Joel Hamkins The multiverse perspective in set theory, 2013]) reject the idea that there’s an intended width hierarchy of sets which contains ‘all possible subsets’ of the sets that it contains. In this paper, I raise an explanatory indispensability worry for the multiverse theorist and distinguish three different possible styles of response to this worry. I will argue that each approach faces some serious prima facie problems. And I’ll suggest that, by clarifying their response to this puzzle about applications, multiverse theorists can helpfully clarify their proposals concerning pure mathematics.

In Hamkins (Rev Symb Log 5(3):416–449, 2012) set theorist Joel David Hamkins uses considerations about forcing arguments, together with an analogy between set theory and geometry to motivate his set-theoretic multiverse program. I’ll argue that Hamkins develops the latter (familiar) analogy in an unusual way, that promises to motivate multiverse theory in particular (rather than merely some form of pluralism). He suggests what I’ll call a leveling up approach to mathematical pluralism which motivates distinctive features of his multiverse theory. However, I’ll question whether Hamkins’ multiverse proposal ultimately delivers leveling up. Then I’ll sketch a counterpossible counterfactual-based twist on Hamkins multiverse which does provide relevant leveling up.

Mathematicalnominalists have argued that we can reformulate scientific theories without quantifying over mathematical objects.However, worries about the nature and meaningfulness of these nominalistic reformulations have been raised, like Burgess and Rosen’s dilemma. In this paper, I’ll review (what I take to be) a kind of emerging consensus response to this dilemma: appeal to the idea of different levels of analysis and explanation, with philosophy providing an extra layer of analysis “below” physics, much as physics does below chemistry. I’ll argue that one can address certain lingering worries for this approach by appeal to the apparent usefulness of a distinction between foundational and non-foundational contexts within mathematics and certain (admittedly controversial) arguments for Potentialism about set theory.

It’s sometimes suggested that we can (in a sense) settle the truth-value of some statements in the language of number theory by stipulation, adopting either φ or ¬φ as an additional axiom. For example, in Clarke-Doane (2020b) and a series of recent APA presentations, Clarke-Doane suggests that any Σ01 sound expansion of our current arithmetical practice would express a truth. In this paper, I’ll argue that (given a certain popular assumption about the model-theoretic representability of languages like ours) we can’t know ourselves to have any such freedom.

Mathematical knowledge and moral knowledge (or normative knowledge more generally) can seem intuitively puzzling in similar ways. For example, taking apparent human knowledge of either domain at face value can seem to require accepting that we benefited from some massive and mysterious coincidence. In the mathematical case, a pluralist partial response to access worries has been widely popular. In this paper, I will develop and address a worry, suggested by some works in the recent literature like (Clarke-Doane, 2020 ), that connections between ought facts and action prevent us from giving a similarly pluralist response to moral access worries.

Invoking a form of quantifier variance promises to let us explain mathematicians’ freedom to introduce new kinds of mathematical objects in a way that avoids some problems for standard platonist and nominalist views. In this paper I’ll note that, despite traditional associations between quantifier variance and Carnapian rejection of metaphysics, Siderian realists about metaphysics can naturally be quantifier variantists. Unfortunately a variant on the Quinean indispensability argument concerning grounding seems to pose a problem for philosophers who accept this hybrid. However I will charitably reconstruct this problem and then argue for optimism about solving it.

It is currently fashionable to take Putnamian model-theoretic worries seriously for mathematics, but not for discussions of ordinary physical objects and the sciences. However, I will argue that (under certain mild assumptions) merely securing determinate reference to physical possibility suffices to rule out the kind of nonstandard interpretations of our number talk Putnam invokes. So, anyone who accepts determinate reference to physical possibility should not reject determinate reference to the natural numbers on Putnamian model-theoretic grounds.

In many ways set theory lies at the heart of modern mathematics, and it does powerful work both philosophical and mathematical – as a foundation for the subject. However, certain philosophical problems raise serious doubts about our acceptance of the axioms of set theory. In a detailed and original reassessment of these axioms, Sharon Berry uses a potentialist approach to develop a unified determinate conception of set-theoretic truth that vindicates many of our intuitive expectations regarding set theory. Berry further defends her approach against a number of possible objections, and she shows how a notion of logical possibility that is useful in formulating Potentialist set theory connects in important ways with philosophy of language, metametaphysics and philosophy of science. Her book will appeal to readers with interests in the philosophy of set theory, modal logic, and the role of mathematics in the sciences.

It's currently fashionable to take Putnamian model theoretic worries seriously for mathematics, but not for discussions of ordinary physical objects and the sciences. But I will argue that (under certain mild assumptions) merely securing determinate reference to physical possibility suffices to rule out nonstandard models of our talk of numbers. So anyone who accepts realist reference to physical possibility should not reject reference to the standard model of the natural numbers on Putnamian model theoretic grounds.

In this article, I discuss a trivialization worry for Hartry Field’s official formulation of the access problem for mathematical realists, which was pointed out by Øystein Linnebo (and has recently been made much of by Justin Clarke-Doane). I argue that various attempted reformulations of the Benacerraf problem fail to block trivialization, but that access worriers can better defend themselves by sticking closer to Hartry Field’s initial informal characterization of the access problem in terms of (something like) general epistemic norms of coincidence avoidance.

In this paper I will draw attention to an important route to external world skepticism, which I will call confidence skepticism. I will argue that we can defang confidence skepticism (though not a meeker ‘argument from might’ which has got some attention in the 20th century literature on external world skepticism) by adopting a partially psychologistic answer to the problem of priors. And I will argue that certain recent work in the epistemology of mathematics and logic provides independent support for such psychologism.

This note points out a conflict between some common intuitions about metaphysical possibility. On the one hand, it is appealing to deny that there are robust counterfactuals about how various physically impossible substances would interact with the matter that exists at our world. On the other hand, our intuitions about how concepts like MOUNTAIN apply at other metaphysically possible worlds seem to presuppose facts about ‘solidity’ which cash out in terms of these counterfactuals. I consider several simple attempts to resolve this conflict and note they all fall short.

Since Benacerraf’s ‘What Numbers Could Not Be, ’ there has been a growing interest in mathematical structuralism. An influential form of mathematical structuralism, modal structuralism, uses logical possibility and second order logic to provide paraphrases of mathematical statements which don’t quantify over mathematical objects. These modal structuralist paraphrases are a useful tool for nominalists and realists alike. But their use of second order logic and quantification into the logical possibility operator raises concerns. In this paper, I show that the work of both these elements can be done by a single natural generalization of the logical possibility operator.

In this paper, I will attempt to develop and defend a common form of intuitive resistance to the companions in guilt argument. I will argue that one can reasonably believe there are promising solutions to the access problem for mathematical realism that don’t translate to moral realism. In particular, I will suggest that the structuralist project of accounting for mathematical knowledge in terms of some form of logical knowledge offers significant hope of success while no analogous approach offers such hope for moral realism.

In this paper I will argue that (principled) attempts to ground a priori knowledge in default reasonable beliefs cannot capture certain common intuitions about what is required for a priori knowledge. I will describe hypothetical creatures who derive complex mathematical truths like Fermat’s last theorem via short and intuitively unconvincing arguments. Many philosophers with foundationalist inclinations will feel that these creatures must lack knowledge because they are unable to justify their mathematical assumptions in terms of the kind of basic facts which can be known without further argument. Yet, I will argue that nothing in the current literature lets us draw a principled distinction between what these creatures are doing and paradigmatic cases of good a priori reasoning (assuming that the latter are to be grounded in default reasonable beliefs). I will consider, in turn, appeals to reliability, coherence, conceptual truth and indispensability and argue that none of these can do the job.