Matthieu Queloz

Cognitive terms such as “understanding" or “reasoning" are increasingly applied to large language models, even by technically informed researchers. This paper argues that such applications are best understood neither as literal attributions nor as mere loose talk, but as instances of metaphor as conceptual adaptation: the creative extension of concepts beyond their original domain under pressure of what I call conceptual needs. Drawing on the Strawson–Kant tradition on imagination and concept-application, inferentialist semantics, Bermúdez's theory of rational framing, and Gentner's structure-mapping account of analogy, I develop a framework with four interlocking components. First, I argue that metaphorical extension selectively transfers inferential connections – not reference – from a source domain to a target domain. Second, I show that this transfer is guided by conceptual needs: mismatches between our conceptual repertoire and our circumstances that possess a distinctive aptic normativity (a normativity of fit). Third, I argue that whether the transferred connections hold in the target domain is empirically evaluable, and that mechanistic interpretability now furnishes the tools for this evaluation. Fourth, drawing on Bowdle and Gentner's "career of metaphor" model and Müller's sleeping/waking framework, I propose that cognitive terms applied to AI are undergoing a career – a dynamic trajectory from novel comparison through inferential negotiation towards potential conventionalisation. This framework makes headway beyond the debate over the literal applicability of cognitive terms to AI by replacing it with graduated, empirically tractable questions about the aptness of specific inferential transfers, and it reveals why the choice of cognitive vocabulary for AI is not merely terminological, but evaluatively and practically consequential.

As AI increasingly drives discovery, the concept of inventor faces severe strain. Recent judicial decisions, such as the Swiss Federal Administrative Court’s 2025 DABUS ruling, expose a deepening tension: courts demand intellectual creation by a natural person even as human contributions to AI-assisted discovery become increasingly nominal. This paper approaches the resulting tension from the standpoint of political philosophy rather than jurisprudence: the strain AI places on the concept of inventorship is too fundamental to be resolved by interpretative methods taking existing conceptual architectures for granted. Inspired by Hume’s genealogy of property, the paper reconstructs the historical “need matrices” that forged the concept of inventorship, tracing its evolution from Venetian guild economics through Romantic genius ideology to corporate R&D. This reveals the concept to be an overburdened bundle serving four social functions: incentivising innovation, disseminating knowledge, legitimating monopolies, and resolving priority disputes. It also clarifies the mismatch between the concept and the emerging realities of AI-driven discovery. To resolve this mismatch, we must disaggregate the concept of inventorship and develop specialised conceptual resources for each of these functions. If we invented the notion of inventor to perform certain functions, we can reinvent it to perform them better.

How to appraise concepts? Against approaches focusing either on the goals of concept-users or on the functions of concepts, I advocate focusing on what concepts we now need. After diagnosing the historical ambivalence of “need” between subjective want and objective exigency, I characterise conceptual needs as possessing a distinctively aptic normativity – a normativity of fittingness. They signal a cognitive privation that marks a mismatch between our conceptual repertoire and our situation, reorienting conceptual engineering from detached amelioration to situated adaptation. To render this aptic normativity tractable, I introduce the analytic device of “need matrices.” As illustrated using the concept of privacy, need matrices and the “need vectors” they generate provide blueprints for fitting concepts by modelling needs as arising at the intersection of concerns, capacities, and circumstances. Unlike function-first approaches, the need-first approach is inherently prospective rather than retrospective; unlike goal-first approaches, it acknowledges that concepts are answerable to pressures we face unwittingly and unwillingly. Starting from potentially unprecedented predicaments rather than established goals or functions makes this approach uniquely suited to guiding conceptual adaptation in times of social and technological upheaval.

As conceptual engineering fractures into explication pursuing exactness and amelioration pursuing justice, the field risks losing its focus. I argue that unifying these projects requires retrieving a crucial insight from Rudolf Carnap: that attempts to improve concepts must start with the preliminary stage of practical clarification. However, Carnap’s account of clarification in terms of predictive proficiency remains normatively inert and biased towards exactness. I expand it into a normative diagnosis of the needs underpinning a concept’s inferential structure. This reveals whether properties like vagueness are flaws that need fixing or features worth preserving. Ultimately, practical clarification can orient a division of conceptual labour, determining when concepts call for scientific exactness or social utility and accommodating both aspirations within one methodology.

En soustrayant le concept de vérité à l’historicisme foucaldien, Pascal Engel finit par exposer davantage les « vertus de la vérité » à la généalogie négative de Foucault. Cet article propose une lecture plus ambitieuse de la généalogie positive de ces vertus, montrant que cultiver l’exactitude et la sincérité comme valeurs intrinsèques est une nécessité fonctionnelle, et non un accident historique. Établir le statut de vertu de ces dispositions offre une défense plus robuste contre le cynisme foucaldien et l’indifférence contemporaine à l’égard de la vérité.

What makes understanding an important cognitive state? And what does having the concept of understanding do for us? This paper offers a unifying account of understanding by jointly reverse-engineering the function of both the state and the concept. We argue that we care about understanding because it grounds and predicts robust competence: the stable ability to succeed across novel scenarios. Our concept of understanding evolved as an efficient proxy to track this elusive property, allowing us to identify who to trust and learn from. This highlights the sociality of understanding and how it shapes what kinds of understanding we are apt to form. We then argue that understanding is the result of convergent pressures on social agents to predict the world using models that are not only accurate, but also compressed enough to be stored, demonstrated, and transmitted. This allows us to integrate a number of ostensibly competing accounts of understanding. Finally, we show how the forces at the root of human understanding elucidate debates over AI understanding.

This article critically appraises Benjamin Lang’s defence of “co-reasoning” with personalized AI advisors through the lens of the liberalism of fear. Lang’s co-reasoning model promises to steer between “dropping anchor” in a user’s past values and “chasing the horizon” of the aspirational values they come to avow. I argue that, while morally attractive at the level of individual psychology, this picture risks exacerbating precisely the patterns of dependency and vulnerability that the liberalism of fear urges us to treat as primary. Once the epistemic, structural, and temporal asymmetries of power involved are recognized, the envisaged “co-reasoning” threatens to look more like an information-asymmetric and structurally constrained csonsultation with a system whose ultimate loyalties lie elsewhere. Moreover, personalization risks translating structural injustices into individualized aspirational challenges, thereby stabilizing domination. I therefore propose three political constraints on personalized AI advisors: the priority of non-domination over aspirational optimization, the public contestability of operative norms, and the recognition that some civic burdens are not fit objects for personalization at all.

Large language models (LLMs) are often portrayed as merely imitating linguistic patterns without genuine understanding. We argue that recent findings in mechanistic interpretability (MI), the emerging field probing the inner workings of LLMs, render this picture increasingly untenable—but only once those findings are integrated within a theoretical account of understanding. We propose a tiered framework for thinking about understanding in LLMs and use it to synthesize the most relevant findings to date. The framework distinguishes three hierarchical varieties of understanding, each tied to a corresponding level of computational organization: conceptual understanding emerges when a model forms “features” as directions in latent space, learning connections between diverse manifestations of a single entity or property; state-of-the-world understanding emerges when a model learns contingent factual connections between features and dynamically tracks changes in the world; principled understanding emerges when a model ceases to rely on memorized facts and discovers a compact “circuit” connecting these facts. Across these tiers, MI uncovers internal organizations that can underwrite understanding-like unification. However, these also diverge from human cognition in their parallel exploitation of heterogeneous mechanisms. Fusing philosophical theory with mechanistic evidence thus allows us to transcend binary debates over whether AI understands, paving the way for a comparative, mechanistically grounded epistemology that explores how AI understanding aligns with—and diverges from—our own.

In “Against Personalized AI Moral Advisors” (2025), Muriel Leuenberger has argued that the personal nature of practical deliberation, which I stressed in my “Can AI Rely on the Systematicity of Truth? The Challenge of Modelling Normative Domains” (2025a), counterintuitively militates against the development of personalized AI moral advisors and in favour of generalist AI moral advisors. Here, I take up and develop this line of thought, drawing out how the asystematicity of normative domains reveals the fundamental limitations of both personalized and generalist AI moral advisors.

Why is the concept of truth so important to us? After all, it is not at all obvious why human intelligence would have evolved to do anything other than to dissimulate, deceive, cheat, and trick. Pragmatic genealogies like the genealogies of the value of truth told by Nietzsche and Williams can help us grasp why we think as we do. But instead of explaining concepts by tracing them to antecedent objects in reality, they trace them to practical needs and reverse-engineer the functions performed by the concepts.

This paper develops Bernard Williams’s suggestion that for philosophy to ignore its history is for it to assume that its history is vindicatory. The paper aims to offer a fruitful line of inquiry into the question whether philosophy has a vindicatory history by providing a map of possible answers to it. It first distinguishes three types of history: the history of discovery, the history of progress, and the history of change. It then suggests that much of philosophy lacks a vindicatory history, for reasons that reflect philosophy’s character as a humanistic discipline. On this basis, the paper reconstructs Williams’s conception of what it means for philos- ophy to engage with its own history. The paper concludes that it is a mistake to think that a vindicatory history is what we would really like to have, and that in fact, the resulting picture gives philosophy several reasons to engage with its own history.

In Epiphanies: An Ethics of Experience, Sophie Grace Chappell offers a phenomenology of epiphanies—those high points in experience when values most vividly reveal themselves to us. Yet Chappell’s method of using phenomenological descriptions to show that we live by our epiphanies leaves open the question of their authority. Why should the epiphanic carry more authority than more sober experiences? The answer, I argue, had better be sensitive to our explanatory understanding of epiphanies. Moreover, it should be sensitive to how the very power of epiphanies threatens to distort our experience of other values. We must beware of what Lichtenberg called transcendent ventriloquism, whereby subjective experience is made to sound like something more than what it is—an eternal truth to be enacted at any cost. I then turn to the politics of epiphanic experience: how to live together given that we live by different epiphanies. I raise several problems for Chappell’s proposal that the normativity inherent in conversation both imposes transcendental constraints on epiphanic ethics and is determinate enough to prescribe specific political arrangements. These constraints determine either too little or too much; they come too late not to beg crucial questions; they cannot offer an independent check on epiphanic ethics; and they introduce a tension at the heart of the book between the romantic openness to experience, emotion, and enthusiasm and the rationalist aspiration to exclude non-rational influences. This threatens to cast as an illicit intrusion precisely the experiential dimension that epiphanies were to introduce into the conversation.

Bernard Williams’ books demand an unusual amount of work from readers. This is particularly true of his 1985 magnum opus, Ethics and the Limits of Philosophy (ELP)—a work so charged with ideas that there seems to be nothing more to say, and yet at the same time so pared-down and tersely argued that there seems to be nothing left to take away. Reflecting on the book five years after its publication, Williams writes that it is centrally concerned with a Nietzschean question: the question of philosophy’s authority, in particular when it comes to telling us how to live (1990, p. XIX). Some ethical theories seem implicitly committed to the idea that philosophy has everything to tell us about how to live. This Williams rejects. But the question then is how much philosophy has to tell us, and as critical as Williams may be of philosophy’s ambitions in this regard, his answer in ELP is certainly not nothing. The book even suggests some things that philosophy might say. But what Williams emphatically insists on, both in the book and in his later reflection on it, is that the question needs to be taken more seriously than it has been.

LLMs are, in the first instance, models of the statistical distribution of tokens in the vast linguistic corpus they have been trained on. But their often surprising emergent capabilities raise the question of how much understanding of the extralinguistic world LLMs can glean from this statistical distribution of words alone. Here, I explore and evaluate the idea that the probability distribution of words in the public corpus offers a window onto the conditional structure of the world. To become a good next-token predictor, an LLM has to become a good pattern completer, and the patterns laid up in language mirror the patterns embodied in the world to a considerable extent. When this is the case—and this is no empty condition—it allows LLMs to compress into their weights two distinct, but complementary forms of understanding of the world. The first form of understanding stored is an understanding of which properties tend to cluster together; I spell out this idea by drawing on Millikan’s account of a multidimensional clumpy world. The second is an understanding of the inferential structure that propositions about these property clusters tend to be enmeshed in; I spell out this idea by drawing on Brandom’s account of the isomorphism between deontic normative conceptual relations of incompatibility and consequence among commitments and alethic modal relations of incompatibility and consequence among states of affairs. On the resulting picture, word models can be world models insofar as linguistic patterns track real patterns between facts.

What is it that one fundamentally rejects when one criticizes a way of thinking as moralistic? Taking my cue from the principal leveller of this charge in philosophy, I argue that the root problem of moralism is the dualism that underlies it. I begin by distinguishing the rejection of moralism from the rejection of the moral/nonmoral distinction: far from being something one should jettison along with moralism, that distinction is something that any human society is bound to develop. But this valuable distinction is transformed into a problematic dualism when it casts the two sides of the distinction as contrasting sharply in nature, value, and structure. In ethics, the resulting dichotomy takes the form of a dualism of morality and prudence. In politics, it takes the form of a dualism of principle and interest. I explain the enduring appeal of such dualisms before laying out the costs of moralism thus conceived: moralism erodes our sensibility to the moral and political costs of value conflicts; it projects an unrealistic conception of agency that sets up scepticism about responsibility; and it limits our ability to appreciate and realize the wider variety of nonmoral values that sustain us, our achievements, and morality itself.

This paper argues that explainability is only one facet of a broader ideal that shapes our expectations towards artificial intelligence (AI). Fundamentally, the issue is to what extent AI exhibits systematicity—not merely in being sensitive to how thoughts are composed of recombinable constituents, but in striving towards an integrated body of thought that is consistent, coherent, comprehensive, and parsimoniously principled. This richer conception of systematicity has been obscured by the long shadow of the “systematicity challenge” to connectionism, according to which network architectures are fundamentally at odds with what Fodor and colleagues termed “the systematicity of thought.” I offer a conceptual framework for thinking about “the systematicity of thought” that distinguishes four senses of the phrase. I use these distinctions to defuse the perceived tension between systematicity and connectionism and show that the conception of systematicity that historically shaped our sense of what makes thought rational, authoritative, and scientific is more demanding than the Fodorian notion. To determine whether we have reason to hold AI models to this ideal of systematicity, I then argue, we must look to the rationales for systematization and explore to what extent they transfer to AI models. I identify five such rationales and apply them to AI. This brings into view the “hard systematicity challenge.” However, the demand for systematization itself needs to be regulated by the rationales for systematization. This yields a dynamic understanding of the need to systematize thought, which tells us how systematic we need AI models to be and when.

A key assumption fuelling optimism about the progress of large language models (LLMs) in accurately and comprehensively modelling the world is that the truth is systematic: true statements about the world form a whole that is not just consistent, in that it contains no contradictions, but coherent, in that the truths are inferentially interlinked. This holds out the prospect that LLMs might in principle rely on that systematicity to fill in gaps and correct inaccuracies in the training data: consistency and coherence promise to facilitate progress towards comprehensiveness in an LLM’s representation of the world. However, philosophers have identified compelling reasons to doubt that the truth is systematic across all domains of thought, arguing that in normative domains, in particular, the truth is largely asystematic. I argue that insofar as the truth in normative domains is asystematic, this renders it correspondingly harder for LLMs to make progress, because they cannot then leverage the systematicity of truth. And the less LLMs can rely on the systematicity of truth, the less we can rely on them to do our practical deliberation for us, because the very asystematicity of normative domains requires human agency to play a greater role in practical thought.

By integrating Bernard Williams’ internalism about reasons with his later thought, this article casts fresh light on internalism and reveals what wider concerns it speaks to. To be consistent with Williams’ later work, I argue, internalism must align with his deference to the phenomenology of moral deliberation and with his critique of ‘moral self-indulgence’. Key to this alignment is the idea that deliberation can express the agent's motivations without referring to them; and that internalism is not a normative claim, but an example of sense-making from ‘the ethnographic stance’. This leaves a worry over whether moral conviction can coexist with an internalist understanding of reasons. Here too, however, Williams’ later thought provides an answer. Differentiating corrosive from corroborative sense-making, it elucidates how internalism, though not normative, can nonetheless affect our confidence in reason statements, thereby informing how we deliberate and how we address those whose motivations differ from our own.

Are there virtues that constitutively involve using certain concepts? Does it make sense to speak of rights or duties to use certain concepts? And do consequentialist approaches to concepts necessarily have to reproduce the difficulties that plague utilitarianism? These are fundamental orientating questions for the emerging field of conceptual ethics, which invites us to reflect critically about which concepts to use. In this article, I map out and explore the ways in which conceptual ethics might take its cue from virtue-ethical, deontological, and consequentialist traditions of ethical thought, flagging the main difficulties facing each approach. I end by sketching how the various dimensions of evaluation singled out by these three traditions might be combined in a single approach.

Bernard Williams argued that historical and philosophical inquiry were importantly linked in a number of ways. This introductory chapter distinguishes four different connections he identified between philosophy and history. (1) He believed that philosophy could not ignore its own history in the way that science can. (2) He thought that when engaging with philosophy’s history primarily to produce history, one still had to draw on philosophy. (3) Even doing history of philosophy philosophically, i.e. primarily to produce philosophy, required a keen sense of how historically distant from us past philosophers were, on his view, because the point of reading them was to confront something different from the present. (4) He held that systematic philosophy itself needed to be done historically, engaging not necessarily with its own history, but with that of the concepts it sought to understand. The chapter closes with an overview of the volume’s structure and content.

Philosophy strives to give us a firmer hold on our concepts. But what about their hold on us? Why place ourselves under the sway of a concept and grant it the authority to shape our thought and conduct? Another conceptualization would carry different implications. What makes one way of thinking better than another? This book develops a framework for concept appraisal. Its guiding idea is that to question the authority of concepts is to ask for reasons of a special kind: reasons for concept use, which tell us which concepts to adopt, adhere to, or abandon, thereby shoring up—or undercutting—the reasons for action and belief that guide our deliberations. Traditionally, reasons for concept use have been sought either in timeless rational foundations or in concepts’ inherent virtues, such as precision and consistency. Against this, the book advances two main claims: that we find reasons for concept use in the conceptual needs we discover when we critically distance ourselves from a concept by viewing it from the autoethnographic stance; and that sometimes, concepts that conflict, or exhibit other vices such as vagueness or superficiality, are just what we need. By considering not what concepts are absolutely best, but what concepts we now need, we can reconcile ourselves to the contingency of our concepts, determine the proper place of efforts to tidy up thought, and adjudicate between competing conceptions of things—even things as contested as liberty or free will. A needs-based approach separates helpful clarification from hobbling tidy-mindedness, and authoritative definition from conceptual gerrymandering.

Virtue ethics is frequently billed as a remedy to the problems of deontological and consequentialist ethics that Bernard Williams identified in his critique of “the morality system.” But how far can virtue ethics be relied upon to avoid these problems? What does Williams’s critique of the morality system mean for virtue ethics? To answer this question, we offer a more principled characterisation of the defining features of the morality system in terms of its organising ambition—to shelter life against luck. This reveals the system to be multiply realisable: the same function can be served by substantively different but functionally equivalent ideas. After identifying four requirements that ethical thought must meet to function as a morality system, we show that they can also be met by certain constellations of virtue-ethical ideas. We thereby demonstrate the possibility of virtue-ethical morality systems raising problems analogous to those besetting their deontological and consequentialist counterparts. This not only widens the scope of Williams’s critique and brings out the cautionary aspect of his legacy for virtue ethics. It also offers contemporary virtue ethicists a more principled understanding of the functional features that mark out morality systems and lie at the root of their problems, thereby helping them to avoid or overcome these problems.

Bernard Williams articulated his later political philosophy notably in response to Ronald Dworkin, who, striving for coherence or integrity among our political concepts, sought to immunize the concepts of liberty and equality against conflict. Williams, doubtful that we either could or should eliminate the conflict, resisted the pursuit of conceptual integrity. Here, I reconstruct this Dworkin–Williams debate with an eye to drawing out ideas of ongoing philosophical and political importance. The debate not only exemplifies Williams's political realism and its connection to his critique of the morality system. It also illustrates the virtues and hazards of contemporary efforts to ameliorate or engineer our concepts; it indicates what political philosophy might look to in appraising political concepts; it adverts to the different needs these concepts have to meet if they are to sustain a politics of pluralism, deal with polarization, and secure the consent of those who end up on the losing side of political decisions; and it presents us with two starkly contrasting conceptions of politics itself, of the place of political values within it, and of our prospects of reducing the uncomfortably conflictual character of those values through philosophy.

For Bernard Williams, philosophy and history are importantly connected. His work exploits this connection in a number of directions: he believes that philosophy cannot ignore its own history the way science can; that even when engaging with philosophy’s history primarily to produce history, one needs to draw on philosophy; and that when doing the history of philosophy primarily to produce philosophy, one still needs a sense of how historically distant past philosophers are, because the point of reading them is to confront something different from the present. But Williams also holds that systematic philosophy itself needs to be done historically, engaging not just with its own history, but with that of the concepts it seeks to understand. To explore these different ways in which philosophy and history intertwine, this volume assembles specially commissioned contributions by A. W. Moore, Terence Irwin, Sophie-Grace Chappell, Catherine Rowett, Marcel van Ackeren, John Cottingham, Gerald Lang, Lorenzo Greco, Paul Russell, Carla Bagnoli, Peter Kail, David Owen, Giuseppina D’Oro, James Connelly, Matthieu Queloz, Nikhil Krishnan, John Marenbon, Ralph Wedgwood, Garrett Cullity, Hans-Johann Glock, Geraldine Ng, Ilaria Cozzaglio, Amanda R. Greene, and Miranda Fricker. They critically appraise Williams’s work in and on the history of philosophy as well as his historicist turn and his use of genealogy. The resulting collection uniquely combines substantive discussions of historical figures from Homer to Wittgenstein with methodological discussions of how and why the history of philosophy should be done, and how and why philosophy should draw on history.

This chapter argues that several aspects of Bernard Williams’s style, methodology, and metaphilosophy can be read as evolving dialectically out of Wittgenstein’s own. After considering Wittgenstein as a stylistic influence on Williams, especially as regards ideals of clarity, precision, and depth, Williams’s methodological debt to Wittgenstein is examined, in particular his anthropological interest in thick concepts and their point. The chapter then turns to Williams’s explicit association, in the 1990s, with a certain form of Wittgensteinianism, which he called ‘Left Wittgensteinianism’. It is shown how this is not a sudden conversion, but the direct product of Williams’s longstanding critical engagement with Wittgenstein’s methodology and metaphilosophy: Williams arrives at this position by envisaging a Wittgensteinianism that thinks in concrete sociohistorical terms, embraces genuine explanation, and relinquishes its insistence on the purity of philosophy. When properly understood, moreover, this critique turns out to be continuous with Williams’s advocacy of a conception of philosophy as a humanistic discipline. Finally, it is shown that Williams inherits from Wittgenstein a certain understanding of how philosophy can help us to live, in particular the therapeutic ambition to liberate us from distortions in our self-understanding by assembling reminders.

Genealogies of belief have dominated recent philosophical discussions of genealogical debunking at the expense of genealogies of concepts, which has in turn focused attention on genealogical debunking in an epistemological key. As I argue in this paper, however, this double focus encourages an overly narrow understanding of genealogical debunking. First, not all genealogical debunking can be reduced to the debunking of beliefs—concepts can be debunked without debunking any particular belief, just as beliefs can be debunked without debunking the concepts in terms of which they are articulated. Second, not all genealogical debunking is epistemological debunking. Focusing on concepts rather than beliefs brings distinct forms of genealogical debunking to the fore that cannot be comprehensively captured in terms of epistemological debunking. We thus need a broader understanding of genealogical debunking, which encompasses not just epistemological debunking, but also what I shall refer to as metaphysical debunking and ethical debunking.

In ‘What Has Philosophy to Learn from Tort Law?’, Bernard Williams reaffirms J. L. Austin’s suggestion that philosophy might learn from tort law ‘the difference between practical reality and philosophical frivolity’. Yet while Austin regarded tort law as just another repository of time-tested concepts, on a par with common sense as represented by a dictionary, Williams argues that ‘the use of certain ideas in the law does more to show that those ideas have strength than is done by the mere fact that they are part of the currency of common sense’. But what does it mean to show that ideas or concepts ‘have strength’? How does conceptual strength relate to the distinction between practical reality and philosophical frivolity? And what special features of the law are supposed to make it a better test of conceptual strength than common sense? In this chapter, I reconstruct and develop Williams’s answers to these questions. I show why Williams believes that we need to test the concepts forming the currency of common sense against practical reality as embodied by legal practice; I identify seven features of tort law that make it particularly suitable to act as such a test; I distinguish three respects in which concepts can show strength, and unpack Williams’s metaphor of microwave-resistant concepts: concepts capable of holding and presenting material for intense critical scrutiny without succumbing to it themselves; lastly, I show how philosophy can learn to identify systematically weak concepts, and the limits of otherwise valuable concepts, by considering which concepts fail the test.

In this précis of The Practical Origins of Ideas: Genealogy as Conceptual Reverse-Engineering (OUP 2021), I summarize the key claims of the book. The book describes, develops, and defends an underappreciated methodological tradition: the tradition of pragmatic genealogy, which aims to identify what our loftiest and most inscrutable conceptual practices do for us by telling strongly idealized, but still historically informed stories about what might have driven people to adopt and elaborate them as they did. What marks out this methodological tradition, I argue, is that it synthesizes two genres of philosophical genealogy that are standardly set against each other: state-of-nature fictions on the one hand and patiently documentary historiography on the other. These two genres of genealogy are usually taken to be mutually exclusive and to answer to radically different philosophical interests and temperaments. But I offer a systematic account of a tradition that combines both genres into a single genealogical method, augmenting genealogy’s power and range by harnessing the strengths and possibilities of both genres.

In this paper, I respond to three critical notices of The Practical Origins of Ideas: Genealogy as Conceptual Reverse-Engineering, written by Cheryl Misak, Alexander Prescott-Couch, and Paul Roth, respectively. After contrasting genealogical conceptual reverse-engineering with conceptual reverse-engineering, I discuss pragmatic genealogy’s relation to history. I argue that it would be a mistake to understand pragmatic genealogy as a fiction (or a model, or an idealization) as opposed to a form of historical explanation. That would be to rely on precisely the stark dichotomy between idealization and history that I propose to call into question. Just as some historical explanations begin with a functional hypothesis arrived at through idealization as abstraction, some pragmatic genealogies embody an abstract form of historiography, stringing together, in a way that is loosely indexed to certain times and places, the most salient needs responsible for giving a concept the contours it now has. I then describe the naturalistic stance that I find expressed in the pragmatic genealogies I consider in the book before examining the evaluative standard at work in those genealogies, defusing the charge that they involve a commitment to a ‘stingy axiology’.