The paper argues that much of the difficulty with making progress on the issue of the normativity of logic for thought, as discussed in the literature, stems from a misapprehension of what logic is normative for. The claim is that, rather than mono-agent mental processes, logic in fact comprises norms for quite specific situations of multi-agent dialogical interactions, in particular special forms of debates. This reconceptualization is inspired by historical developments in logic and mathematics, in particular the pervasiveness of such (...) dialogical conceptions in the early days of logic in ancient Greece. The multi-agent, dialogical perspective then allows for the formulation of compelling 'bridge principles’ between the relation of logical consequence and dialogical normative principles, something that is notoriously difficult to achieve in a mono-agent setting pertaining exclusively to thought and belief. The upshot is also that the truth-preserving rules of logic generally do not have a primary normative bearing on mono-agent mental processes, and in this sense the paper sides with Harman's critique of the idea that logic has a normative import for thought and belief. (shrink)

This article is an exercise in the virtue epistemology of the internet, an area of applied virtue epistemology that investigates how online environments impact the development of intellectual virtues, and how intellectual virtues manifest within online environments. I examine online echo chambers and epistemic bubbles (Nguyen 2020, Episteme 17(2), 141–61), exploring the conceptual relationship between these online environments and the virtue of open-mindedness (Battaly 2018b, Episteme 15(3), 261–82). The article answers two key individual-level, virtue epistemic questions: (Q1) How does immersion (...) in online echo chambers and epistemic bubbles affect the cultivation and preservation of open-mindedness, and (Q2) Is it always intellectually virtuous to exhibit open-mindedness in the context of online echo chambers and epistemic bubbles? In response to Q1, I contend that both online echo chambers and online epistemic bubbles threaten to undermine the cultivation and preservation of open-mindedness, albeit via different mechanisms and to different degrees. In response to Q2, I affirm that both a deficiency and an excess of open-mindedness can be virtuous in these online environments, depending on the epistemic orientation of the digital user. (shrink)

Most human beliefs are acquired through communication, and so are most misbeliefs. Just like the misbeliefs discussed by McKay & Dennett (M&D), culturally transmitted misbeliefs tend to result from limitations rather than malfunctions of the mechanisms that produce them, and few if any can be argued to be adaptations. However, the mechanisms involved, the contents, and the hypothetical adaptive value tend to be specific to the cultural case.

In many traditional, small-scale societies, death and other misfortunes are commonly explained as a result of others’ malign occult agency. Here, we call this family of epistemic tendencies “the agential view of misfortune.” After reviewing several ethnographic case studies that illustrate this view, we argue that its origins and stability are puzzling from an evolutionary perspective. Not only is the agential view of misfortune false; it imposes costs on individuals and social groups that seem to far outweigh whatever benefits the (...) view might provide. We thus doubt that the agential view of misfortune is explainable in terms of adaptive effects. However, neither does it seem readily explainable as a consequence of belief formation strategies that are on the whole adaptive (as is plausibly the case for certain other of our false beliefs, including some that are costly). Accordingly, we contend that the commonness of the agential view of misfortune demands a special evolutionary explanation of some kind. We provide a partial explanation of this phenomenon by highlighting the adaptive benefits that often flow to occult specialists in environments where the agential view of misfortune is entrenched. What this does not explain, however, is the general lack of resistance we observe in response to occultists’ exploitative behaviours over (cultural) evolutionary timescales. We conclude by canvassing a few possible explanations for this puzzling lack of resistance, and while we commit ourselves to none, we do find one option more promising than the others. (shrink)

This paper disputes the widespread assumption that beliefs and desires may be attributed as theoretical entities in the service of the explanation and predic- tion of human behavior. The literature contains no clear account of how beliefs and desires might generate actions, and there is good reason to deny that principles of rationality generate a choice on the basis of an agent’s beliefs and desires. An alter- native conception of beliefs and desires is here introduced, according to which an attribution (...) of belief is, in a certain sense, an assertion on the believer’s behalf and an attribution of desire is, in a certain sense, a command on the desirer’s behalf. In terms of this conception of the attribution of beliefs and desires, we can begin to understand how attributions of beliefs and desires can be explanatory, although we still cannot expect the attribution of beliefs and desires to generate predictions based on an assumption of rationality. Finally, the reality of beliefs and desires is likened to the reality of money; it is the reality of things governed by objective norms. (shrink)

Psychological studies on fictional persuasion demonstrate that being engaged with fiction systematically affects our beliefs about the real world, in ways that seem insensitive to the truth. This threatens to undermine the widely accepted view that beliefs are essentially regulated in ways that tend to ensure their truth, and may tempt various non-doxastic interpretations of the belief-seeming attitudes we form as a result of engaging with fiction. I evaluate this threat, and argue that it is benign. Even if the relevant (...) attitudes are best seen as genuine beliefs, as I think they often are, their lack of appropriate sensitivity to the truth does not undermine the essential tie between belief and truth. To this end, I shall consider what I take to be the three most plausible models of the cognitive mechanisms underlying fictional persuasion, and argue that on none of these models does fictional persuasion undermine the essential truth-tie. (shrink)

The novice/expert problem is the problem of knowing which apparent expert to trust. Following Alvin Goldman’s lead, a number of philosophers have developed criteria that novices can use to distinguish more from less trustworthy experts. While the criteria the philosophers have identified are indeed useful in guiding expert choice, I argue, they can’t do the work that Goldman and his successors want from them: avoid a kind of testimonial scepticism. We can’t deploy them in the way needed to avoid such (...) scepticism, because it would take genuine expertise to do so. I argue that attempts to deploy them in this sort of deep way involve a kind of transgression akin to, and at least as unreliable as, epistemic trespassing. We should give up trying to solve the novice/expert problem and instead promote better epistemic trust. (shrink)

The human species is more reliant on cultural adaptation than any other species, but it is unclear how observational learning can give rise to the faithful transmission of cultural adaptations. One possibility is that teaching facilitates accurate social transmission by narrowing the range of inferences that learners make. However, there is wide disagreement about how to define teaching, and how to interpret the empirical evidence for teaching across cultures and species. In this article I argue that disputes about the nature (...) and prevalence of teaching across human societies and nonhuman animals are based on a number of deep-rooted theoretical differences between fields, as well as on important differences in how teaching is defined. To reconcile these disparate bodies of research, I review the three major approaches to the study of teaching – mentalistic, culture-based, and functionalist – and outline the research questions about teaching that each addresses. I then argue for a new, integrated framework that differentiates between teaching types according to the specific adaptive problems that each type solves, and apply this framework to restructure current empirical evidence on teaching in humans and nonhuman animals. This integrative framework generates novel insights, with broad implications for the study of the evolution of teaching, including the roles of cognitive constraints and cooperative dilemmas in how and when teaching evolves. Finally, I propose an explanation for why some types of teaching are uniquely human, and discuss new directions for research motivated by this framework. (shrink)

How should we undertand the role of norms—especially epistemic norms—governing assertive speech acts? Mitchell Green (2009) has argued that these norms play the role of handicaps in the technical sense from the animal signals literature. As handicaps, they then play a large role in explaining the reliability—and so the stability (the continued prevalence)—of assertive speech acts. But though norms of assertion conceived of as social norms do indeed play this stabilizing role, these norms are best understood as deterrents and not (...) as handicaps. This paper explains the stability problem for the maintenance of animal signals, and so human communication, for we are animals too, after all; the mechanics of the handicap principle; the role of deterrents and punishments as an alternative mechanism; and the role of social norms governing assertion for the case of human communication. (shrink)

This paper will consider the extent to which patients' dependence on clinical expertise when making medical decisions threatens patient autonomy. I start by discussing whether or not dependence on experts is prima facie troubling for autonomy and suggest that it is not. I then go on to consider doctors' and other healthcare professionals' status as ‘medical experts’ of the relevant sort and highlight a number of ways in which their expertise is likely to be deficient. I then consider how this (...) revised picture of medical expertise should lead us to view the potential threat to patient autonomy that results from depending on such ‘experts’. I argue that, whether or not patients are aware of the limitations of medical expertise, in practice it is difficult to do other than defer to medical advice, and this presents a threat to patient autonomy that should be addressed. I conclude by suggesting some ways in which this threat to autonomy might be mitigated. (shrink)

This Element examines what we can learn from religious disagreement, focusing on disagreement with possible selves and former selves, the epistemic significance of religious agreement, the problem of disagreements between religious experts, and the significance of philosophy of religion. Helen De Cruz shows how religious beliefs of others constitute significant higher-order evidence. At the same time, she advises that we should not necessarily become agnostic about all religious matters, because our cognitive background colors the way we evaluate evidence. This allows (...) us to maintain religious beliefs in many cases, while nevertheless taking the religious beliefs of others seriously. (shrink)

The public is largely innumerate, making systematic mistakes in estimating some politically relevant facts, such as the share of foreign-born citizens. In two-step or multistep flow models, such mistakes could be corrected if better-informed citizens were able to convince their peers, in particular by using good arguments citing reliable sources. In six experiments, we find two issues that dampen the potential power of this two-step flow process. First, even though participants were more convinced by good than by poor arguments, many (...) did not change their minds, even when confronted with good arguments. Second, participants are not inclined to spontaneously generate arguments that cite reliable sources, even when they have just been influenced by such arguments. Both issues should put a significant brake in the spread of political numeracy through the two-step flow process, in particular in non-dialogic contexts. (shrink)

According to relevance theory, irony comprehension invariably entails the identification of some opinion or thought and the identification of the speaker’s dissociative attitude. In this paper, it is argued that it is also essential for hearers to identify not only that propositional attitude, but also the affective attitude that the speaker holds towards the source of this echo so that an optimallyrelevant interpretive outcomeis achieved. This notion comprises feelings and emotions of a non-propositional quality which affect the propositional effects obtained (...) in ironical communication. The paper further argues for the need to incorporate non-propositional effects into the traditional propositional object of pragmatic research. (shrink)

Mercier and Sperber (2011a, 2011b; Mercier, 2011a, 2011b, 2011c, and 2011d) have presented a stimulating and provocative new theory of reasoning: the argumentative theory of reasoning. They maintain that argumentation is a meta-representational module. In their evolutionary view of argumentation, the function of this module would be to regulate the flow of information between interlocutors through persuasiveness on the side of the communicator and epistemic vigilance on the side of the audience. The aim of this paper is to discuss the (...) perspective of the authors in which they conceive this competence as the natural scenario of reflective reasoning. (shrink)

Mercier & Sperber (M&S) argue for their argumentative theory in terms of communicative abilities. Insights can be gained by extending the discussion beyond human reasoning to rodent and robot navigation. The selection of arguments and conclusions that are mutually reinforcing can be cast as a form of abductive reasoning that I argue underlies the construction of cognitive maps in navigation tasks.

Neil Levy argues that the importance of acquiring cultural knowledge in our evolutionary past selected for conformist and deferential social learning, and that contemporary bad beliefs – roughly, popular beliefs at odds with expert consensus – result primarily from the rational deployment of such conformity and deference in epistemically polluted modern environments. I raise several objections to this perspective. First, against the cultural evolutionary theory from which Levy draws, I argue that humans evolved to be highly sophisticated and vigilant social (...) learners. Given this, the ubiquity of bad beliefs in the modern world is puzzling: if humans are so smart and suspicious, why do these characteristics seem so rare in domains such as politics? I argue that the answer rests on the incentives that underlie bad beliefs, and I favorably contrast this explanation with Levy’s appeal to epistemic pollution. (shrink)

Deirdre Wilson provides a reflective overview of a volume devoted to the historic application of relevance-theoretic ideas to literary studies. She maintains a view argued elsewhere that the putative non-propositional nature of literary effects are an illusion, a view which dates to Sperber and Wilson : “If you look at [non-propositional] affective effects through the microscope of relevance theory, you see a wide array of minute cognitive [i.e., propositional] effects.” This paper suggests an alternative, that modern-day humans have two apparently (...) different modes of expressing and interpreting information: one of these is a system in which propositional, cognitive effects dominate; the other involves direct, non-propositional effects. The paper concludes by describing two ways such affects might be assimilated into relevance theory. The first, to accept that humans are much more than merely cognitive organisms; the second, to rethink quite radically what we mean by cognition. (shrink)

How is human social intelligence engaged in the course of ordinary conversation? Standard models of conversation hold that language production and comprehension are guided by constant, rapid inferences about what other agents have in mind. However, the idea that mindreading is a pervasive feature of conversation is challenged by a large body of evidence suggesting that mental state attribution is slow and taxing, at least when it deals with propositional attitudes such as beliefs. Belief attributions involve contents that are decoupled (...) from our own primary representation of reality; handling these contents has come to be seen as the signature of full-blown human mindreading. However, mindreading in cooperative communication does not necessarily demand decoupling. We argue for a theoretical and empirical turn towards “factive” forms of mentalizing here. In factive mentalizing, we monitor what others do or do not know, without generating decoupled representations. We propose a model of the representational, cognitive, and interactive components of factive mentalizing, a model that aims to explain efficient real-time monitoring of epistemic states in conversation. After laying out this account, we articulate a more limited set of conversational functions for nonfactive forms of mentalizing, including contexts of meta-linguistic repair, deception, and argumentation. We conclude with suggestions for further research into the roles played by factive versus nonfactive forms of mentalizing in conversation. (shrink)

This paper offers an analytic perspective on epistemic dependence that is grounded in theoretical discussion and field observation at the same time. When in the course of knowledge creation epistemic labor is divided, collaborating scientists come to depend upon one another epistemically. Since instances of epistemic dependence are multifarious in scientific practice, I propose to distinguish between two different forms of epistemic dependence, opaque and translucent epistemic dependence. A scientist is opaquely dependent upon a colleague if she does not possess (...) the expertise necessary to independently carry out, and to profoundly assess, the piece of scientific labor which her colleague is contributing. If the scientist does possess the necessary expertise, I argue, her dependence is translucent. However, the distinction between opaque and translucent epistemic dependence does not exhaust dependence relations in scientific practice, because many dependence relations are neither entirely opaque nor translucent. I will discuss why this is the case, and show how we can make sense of the gray zone between opaque and translucent epistemic dependence. (shrink)

Based on an empirical study of a research team in natural science, the author argues that collaborating scientists do not trust each other completely. Due to the inherent incompleteness of trust, epistemic trust among scientists is not sufficient to manage epistemic dependency in research teams. To mitigate the limitations of epistemic trust, scientists resort to specific strategies of indirect assessment such as dialoguing practices and the probing of explanatory responsiveness. Furthermore, they rely upon impersonal trust and deploy practices of hierarchical (...) authorship. (shrink)

La notion de témoignage est au cœur d’une branche de la théorie de la connaissance, appelée « épistémologie du témoignage », qui s’interroge sur le bien-fondé des croyances acquises par le biais d’autrui – par opposition aux seules capacités sensorielles et intellectuelles de l’individu. Le témoignage est aussi un des modes de preuve les plus importants dans le cadre judiciaire, aussi bien au cours de l’enquête et de l’instruction que du procès, et a fait l’objet de nombreuses réflexions de la (...) part de juristes, en particulier dans les pays de common law. Par-delà les différences disciplinaires, les distinctions des juristes sont à même d’apporter un éclairage nouveau à la réflexion épistémologique, laquelle invite à reconsidérer certaines définitions juridiques. (shrink)

Human attention is limited in its capacity and duration. In language, this is manifested in many ways, but more conspicuously in the strategies by which information is distributed in utterances, that is, their information structures. We contend that the pragmatic categories of Topic and Presupposition precisely meet the necessity to modulate attentional resources on sentence contents, and they do this by “directing” certain contents to automatic and others to controlled processing mechanisms. We discuss experimental findings suggesting that presupposed or topicalized (...) information correlates with automatic processing, and we suggest that this association grounded for the emergence of topic and presupposition units in human communication. We also put forth the processing automaticity induced by these units as the rationale behind their persuasiveness in some specific contexts of language use. (shrink)

I argue for a theory of the optimal function of the speech act of referring, called the edenic theory. First, the act of singular reference is defined directly in terms of Gricean communicative intentions. Second, I propose a doxastic constraint on the optimal performance of such acts, stating, roughly, that the speaker must not have any relevant false beliefs about the identity or distinctness of the intended object. In uttering a singular term on an occasion, on this theory, one represents (...) oneself as not having any confused beliefs about the object to which one intends to refer. This paves the way for an intentionalist theory of reference that circumvents well-known problems, which have not been adequately addressed before in the literature. (shrink)

I argue that introspection recruits the same mental mechanism as that which is required for the production of ordinary speech acts. In introspection, in effect, we intentionally tell ourselves that we are in some mental state, aiming thereby to produce belief about that state in ourselves. On one popular view of speech acts, however, this is precisely what speakers do when speaking to others. On this basis, I argue that every bias discovered by social epistemology applies to introspection and other (...) forms of self-directed representation. If so, it becomes unclear in what sense social epistemology is social. (shrink)

The notion of epistemic responsibility applied to memory has been in general examined in the framework of the responsibilities that a collective holds for past injustices, but it has never been the object of an analysis of its own. In this article, I propose to isolate and explore it in detail. For this purpose, I start by conceptualizing the epistemic responsibility applied to individual memories. I conclude that an epistemic responsible individual rememberer is a vigilant agent who knows when to (...) engage in different kinds of mental and non-mental actions in order to monitor and update her memories, and who develops and nurture different kinds of virtuous attitudes that guide those actions. These (epistemic) virtuous attitudes are oriented not only towards herself but also towards others. Whereas this conception of epistemic responsibility does not pose a problem to understand shared memories of family members and friends, it may seem suspicious when applied to large-scale collective memories. These memories, which I name historical memories, are memories of events that have a traumatic impact for the community, are permeated by unequal relations of power, keep a complex relationship with historical science, and present other characteristics that distinguish them from individual memories. But despite these differences, the analysis undertaken in this work shows that the general principles that govern the epistemic responsibility of individual and (large-scale) collective rememberers are similar, and are based on similar grounds: pragmatic considerations about the consequences of misremembering or forgetting and a feeling of care. The similarities at the individual and collective scale of the epistemic vigilant attitude that is and should be taken toward our significant past may partially justify the use of the same epithet—“memory”—to refer to these different kinds of representations. (shrink)

In this paper, I approach epistemic norms from an ontogenetic point of view. I argue and present evidence that to understand epistemic norms – e.g., scientific norms of methodology and the evaluation of evidence – children must first develop through their social interactions with others three key concepts. First is the concept of belief, which provides the most basic distinction on which scientific investigations rest: the distinction between individual subjective perspectives and an objective reality. Second is the concept of reason, (...) which in the context of science obligates practitioners to justify their claims to others with reasons by grounding them in beliefs that are universally shared within the community. Third is the concept of social norm, which is not primarily epistemic, but provides children with an understanding of norms as collective agreements. The theoretical argument is that all three of these concepts emerge not from just any kind of social interaction, but specifically from social interactions structured by the human species’ unique capacities for shared intentionality. (shrink)

This paper reports the first empirical investigation of the hypothesis that epistemic appraisals form part of the structure of concepts. To date, studies of concepts have focused on the way concepts encode properties of objects and the way those features are used in categorization and in other cognitive tasks. Philosophical considerations show the importance of also considering how a thinker assesses the epistemic value of beliefs and other cognitive resources and, in particular, concepts. We demonstrate that there are multiple, reliably (...) judged, dimensions of epistemic appraisal of concepts. Four of these dimensions are accounted for by a common underlying factor capturing how well people believe they understand a concept. Further studies show how dimensions of concept appraisal relate to other aspects of concepts. First, they relate directly to the hierarchical organization of concepts, reflecting the increase in specificity from superordinate to basic and subordinate levels. Second, they predict inductive choices in category-based induction. Our results suggest that epistemic appraisals of concepts form a psychologically important yet previously overlooked aspect of the structure of concepts. These findings will be important in understanding why individuals sometimes abandon and replace certain concepts; why social groups do so, for example, during a “scientific revolution”; and how we can facilitate such changes when we engage in deliberate “conceptual engineering” for epistemic, social, and political purposes. (shrink)

Most explanations of the rational authority of testimony provide little guidance when evaluating individual pieces of testimony. In practice, however, we are remarkably sensitive to the varying epistemic credentials of testimony: extending trust when it is deserved, and withholding it when it is not. A complete account of the epistemology of testimony should, then, have something to say about when it is that testimony is trustworthy. In the typical case, to judge someone trustworthy requires judging them to be competent and (...) sincere. In this essay we develop an exchange-based account of testimony that shows how those who receive testimony are in a position to evaluate the sincerity of speakers. (shrink)

Both science and religion are human endeavours that recruit and modify pre-existing human capacity to engage in epistemic vigilance. However, while science relies upon a focus on content vigilance, religion focusses on source vigilance. This difference is due, in turn, to the function of religious claims not being connected to their accuracy – unlike the function of scientific claims. Understanding this difference helps to understand many aspects of scientific and religious institutions.

Can people efficiently navigate the modern communication environment, and if yes, how? We hypothesize that in addition to psychological capacities of epistemic vigilance, which evaluate the epistemic value of communicated information, some social institutions have evolved for the same function. Certain newspapers for instance, implement processes, distributed among several experts and tools, whose function is to curate information. We analyze how information curation is done at the institutional level and what challenges it meets. We also investigate what factors favor the (...) cultural evolution of institutions of epistemic vigilance: these include people’s preference for accurate and reliable information and their ability to assess communicated information in view of the source’s epistemic authority; but also contingent historical factors that make it worth—or not—to contribute to the maintenance of institutions of epistemic vigilance. We conclude the paper by considering the challenges and vulnerabilities of these institutions in the Digital Age. (shrink)

I propose a new model of implicit bias, according to which implicit biases are constituted by unconscious imaginings. I begin by endorsing a principle of parsimony when confronted with unfamiliar phenomena. I introduce implicit bias in terms congenial to what most philosophers and psychologists have said about their nature in the literature so far, before moving to a discussion of the doxastic model of implicit bias and objections to it. I then introduce unconscious imagination and argue that appeal to it (...) does not represent a departure from a standard view of imagination, before outlining my model and showing how it accommodates characteristic features of implicit bias. I argue for its advantages over the doxastic model: it does not violate the parsimony principle, it does not face any of the objections so far raised to doxasticism, and it can accommodate the heterogeneity in the category of implicit bias. Finally, I address whether my view limits our ability to hold people accountable for their biases (it does not), and whether it is consistent with what we know about intervention strategies (it is). I conclude that implicit biases are constituted by unconscious imaginings. (shrink)

Mercier & Sperber (M&S) claim that the phenomenon of belief bias provides fundamental support for their argumentative theory and its basis in intuitive judgement. We propose that chronometric evidence necessitates a more nuanced account of belief bias that is not readily captured by argumentative theory.

Mayr’s distinction between proximate and ultimate explanation is justly famous, marking out a division of explanatory labor in biology. But while it is a useful heuristic in many cases, there are others in which proximate factors play an important role in shaping evolutionary trajectories, and in such cases, each project is sensitive to, and relevant to, the other. This general methodological claim is developed in the context of a discussion of human cooperation, and in particular, in a discussion on the (...) puzzling stability of the social contract over the Pleistocene–Holocene transition. For I argue that while we have a plausible account of the stability of Pleistocene cooperation, the stabilizing factors of the Pleistocene disappear in the Holocene, but cooperation does not. Holocene humans solved many collective action problems; cooperation did not collapse despite the apparent growth of free-riding elites. So the article combines a methodological claim about the interaction of proximate and evolutionary biology with a substantive one about the ecology of human cooperation. (shrink)

Short abstract (98 words). Reasoning is generally seen as a means to improve knowledge and make better decisions. However, much evidence shows that reasoning often leads to epistemic distortions and poor decisions. This suggests that the function of reasoning should be rethought. Our hypothesis is that the function of reasoning is argumentative. It is to devise and evaluate arguments intended to persuade. Reasoning so conceived is adaptive given humans’ exceptional dependence on communication and vulnerability to misinformation. A wide range of (...) evidence in the psychology of reasoning and decision making can be reinterpreted and better explained in the light of this hypothesis. (shrink)

In Kourken Michaelian questions the basic tenets of our article (Sperber et al. 2010). Here I defend against Michaelian's criticisms the view that epistemic vigilance plays a major role in explaining the evolutionary stability of communication and that the honesty of speakers and the reliability of their testimony are, to a large extent, an effect of hearers' vigilance.

This article contributes to the debate around the abilities of large language models such as GPT-3, dealing with: firstly, evaluating how well GPT does in the Turing Test, secondly the limits of such models, especially their tendency to generate falsehoods, and thirdly the social consequences of the problems these models have with truth-telling. We start by formalising the recently proposed notion of reversible questions, which Floridi & Chiriatti propose allow one to ‘identify the nature of the source of their answers’, (...) as a probabilistic measure based on Item Response Theory from psychometrics. Following a critical assessment of the methodology which led previous scholars to dismiss GPT’s abilities, we argue against claims that GPT-3 completely lacks semantic ability. Using ideas of compression, priming, distributional semantics and semantic webs we offer our own theory of the limits of large language models like GPT-3, and argue that GPT can competently engage in various semantic tasks. The real reason GPT’s answers seem senseless being that truth-telling is not amongst them. We claim that these kinds of models cannot be forced into producing only true continuation, but rather to maximise their objective function they strategize to be plausible instead of truthful. This, we moreover claim, can hijack our intuitive capacity to evaluate the accuracy of its outputs. Finally, we show how this analysis predicts that a widespread adoption of language generators as tools for writing could result in permanent pollution of our informational ecosystem with massive amounts of very plausible but often untrue texts. (shrink)

We argue that the evolutionary function of reasoning is to allow us to secure more accurate beliefs and more effective intentions through collective deliberation. This sets our view apart both from traditional intellectualist accounts, which take the evolutionary function to be individual deliberation, and from interactionist accounts such as the one proposed by Mercier and Sperber, which agrees that the function of reasoning is collective but holds that it aims to disseminate, rather than come up with, accurate beliefs. We argue (...) that our collectivized intellectualism offers the best explanation of the range of biases that human reasoning is prone to, and that it does better than interactionism at offering a function of reasoning that would have been adaptive for our distant ancestors who first evolved this capacity. (shrink)

In this paper I use philosophical accounts on the relationship between trust and knowledge in science to apprehend this relationship on the Web. I argue that trust and knowledge are fundamentally entangled in our epistemic practices. Yet despite this fundamental entanglement, we do not trust blindly. Instead we make use of knowledge to rationally place or withdraw trust. We use knowledge about the sources of epistemic content as well as general background knowledge to assess epistemic claims. Hence, although we may (...) have a default to trust, we remain and should remain epistemically vigilant; we look out and need to look out for signs of insincerity and dishonesty in our attempts to know. A fundamental requirement for such vigilance is transparency: in order to critically assess epistemic agents, content and processes, we need to be able to access and address them. On the Web, this request for transparency becomes particularly pressing if (a) trust is placed in unknown human epistemic agents and (b) if it is placed in non-human agents, such as algorithms. I give examples of the entanglement between knowledge and trust on the Web and draw conclusions about the forms of transparency needed in such systems to support epistemically vigilant behaviour, which empowers users to become responsible and accountable knowers. (shrink)

According to anti‐reductionism in the epistemology of testimony, testimonial entitlement is easy to come by: all you need to do is listen to what you are being told. Say you like anti‐reductionism; one question that you will need to answer is how come testimonial entitlement comes so cheap; after all, people are free to lie.This paper has two aims: first, it looks at the main anti‐reductionist answers to this question and argues that they remain unsatisfactory. Second, it goes on a (...) rescue mission on behalf of anti‐reductionism. I put forth a novel, knowledge‐first anti‐reductionist account, which I dub ‘Testimonial Contractarianism’. According to the view defended here, in virtue of the social contract in play, compliance with the norms governing speech acts is the default position for speakers. Insofar as norm compliance is the default for speakers, I argue, all else equal, entitlement to believe is the default for hearers. (shrink)

Ruth Millikan has hypothesised that human cognition contains ‘consistency testers’. Consistency testers check whether different judgements a thinker makes about the same subject matter agree or conflict. Millikan’s suggestion is that, where the same concept has been applied to the world via two routes, and the two judgements that result are found to be inconsistent, that makes the thinker less inclined to apply those concepts in those ways in the future. If human cognition does indeed include such a capacity, its (...) operation will be an important determinant of how people use concepts. It will have a major impact on which concepts they deploy and which means of application (conceptions) they rely on. Since consistency testers are a selection mechanism at the heart of conceptual thinking, they would be crucial to understanding how concepts are selected – why some are retained and proliferate and others die out. Hence, whether consistency testers for concepts exist, and how they operate, is an important question for those seeking to understand the cultural evolution of concepts, and of the words we use to express them. (shrink)

Concepts are our tools for thinking. They enable us to engage in explicit reasoning about things in the world. Like physical tools, they can be more or less good, given the ways we use them – more or less dependable for categorisation, learning, induction, action-planning, and so on. Do concept users appreciate, explicitly or implicitly, that concepts vary in dependability? Do they feel that some concepts are in some way defective? If so, we metacognize our concepts. One example that has (...) been studied is a person’s judgement about how well they have learnt a new category. There are many other forms that concept-metacognition could take. This paper offers a preliminary taxonomy of different forms of metacognition directed at concepts. It suggests that concept-metacognition may affect the way one concept from a range of candidates is selected for use, and the way a concept is relied on in reasoning. Concept-metacognition may also play a pivotal role in the social process of constructing concepts, in replacing the old and constructing the new tools for thinking. (shrink)