Implicit Bias

Generative artificial intelligence systems based on transformers, including both text-generators like GPT-3 and image generators like DALL-E 2, have recently entered the popular consciousness. These tools, while impressive, are liable to reproduce, exacerbate, and reinforce extant human social biases, such as gender and racial biases. In this paper, we systematically review the extent to which DALL-E Mini suffers from this problem. In line with the Model Card published alongside DALL-E Mini by its creators, we find that the images it produces (...) tend to represent dozens of different occupations as populated either solely by men (e.g., pilot, builder, plumber) or solely by women (e.g., hairdresser, receptionist, dietitian). In addition, the images DALL-E Mini produces tend to represent most occupations as populated primarily or solely by White people (e.g., farmer, painter, prison officer, software engineer) and very few by non-White people (e.g., pastor, rapper). These findings suggest that exciting new AI technologies should be critically scrutinized and perhaps regulated before they are unleashed on society. (shrink)

In Tobia (2016), Kevin P. Tobia tests for bias using two ontological arguments claimed to be symmetrical and of equal strength. We show they are neither.

What factors affect judicial decision-making? The legal system is of utmost importance because of its impact on our lives. Judges appear to have the most power among any social workers seeing as the precedents set in their decisions are tantamount to written law. Nevertheless, judges may be subject to certain biases, moral and cognitive alike, which influence their rulings. Looking into how morality and cognitive biases affect judges may also reveal how we as individuals handle combining morals with ethics- as (...) well as whether any discrepancies among ourselves are justifiable. An investigation into relevant moral and jurisprudential theories revealed much about the inner workings of a judge’s mindset- whether he or she recognizes it or not. Coupling this with corroborated research on a select few heuristics present specifically for judges sets the foundation for an array of possible solutions in order to ensure the most impartial judicial proceedings. Plausible remedies include conducting premortems, implicit association tests, and a revised ruling procedure. (shrink)

As we have seen in the transition form Part I to Part II of this book, the inductive riskiness of doxastic methods applied in testimonial uptake or prescribed as exemplary of religious faith, helpfully operationalizes the broader social scientific, philosophical, moral, and theological interest that people may have with problems of religious luck. Accordingly, we will now speak less about luck, but more about the manner in which highly risky cognitive strategies are correlated with psychological studies of bias studies and (...) human cognitive ecology. Chapter Four is concerned with connections between psychological study of biases and heuristics, and the comparative study of fundamentalism. The first section looks at work by psychologists and philosophers on our bias blind spot. Later sections ask, ‘In what ways might biases and heuristics play a special role in aiding our understanding of, and response to, fundamentalist orientation?’. The judgments we make in ignorance of our own biases Montaigne calls our importunate presumptions, and he suggests a host of practical factors that make them appealing. Montaigne, as I discuss in the first section, associates many of our errors with one or another kind of presumption, often about our similarity or differences from others, or from God. Our obvious psychographic diversity, and the polemical ground dynamics involved in our ‘culture wars’ are compounded on the agential side by the invisibility of our biases to ourselves. A number of person and social biases are described that plausibly affect all of our beliefs in domains of controversial views, religious views included. The second section continues to study of how etiological symmetry (similar patterns of belief-uptake) gives rise to religious contrariety (diverse narratives and theologies) in testimonial faith traditions, and what the implications of this are for philosophy of religion, generally, and for an improved comparative study of fundamentalism, in particular. Utilizing the work of philosophers such as Rachel Fraser and psychologists such as Emily Pronin and her co-authors, I offer a four-step genealogical account for how etiological symmetry so easily gives rise to religious contrariety. This account begins with the narrative nature of testimony in the Abrahamic family of religions, and how narrative content confounds our “source monitoring.” My genealogy also introduces what I term biased-closure inferences (BCI) as one of the key enablers of religious exclusivism and absolutism. These are the seemingly ‘logical’ but actually very self-serving inferences people often make, inferences from their own belief being true, to any belief contrary to it being false. Those who claim unique truth, epistemic access, and/or virtue and religious value for the home religion are no exception to the broad pertinence of bias studies across domains of controversial view. The proximate causes of belief are all that we can study, and in these there may be significant etiological symmetries. Yet those groups themselves, especially to the extent that they are exclusivist, are tunnel-visioned on claims of doctrinal uniqueness: on content differences of a theological sort. Comparative philosophy is met with the puzzlement that symmetric or essentially similar doxastic strategies should give rise not just to cognitive diversity, but to what we can call polarized or polemical contrariety, contrariety of a kind where each view adamantly rejects all others. Arguably, the more that theologies offer explanations of a counter-inductive sort, or profess counter-inductive thinking as exemplary of faith, the better evidential ground there for inferring that bias involved in the acquisition or maintenance of beliefs. This provides more substance to the question of when censure, and etiological challenges to faith-based beliefs are philosophically well-motivated. (shrink)

Virtue theories can plausibly be argued to have important advantages over normative ethical theories which prescribe a strict impartialism in moral judgment, or which neglect people’s special roles and relationships. However, there are clear examples of both virtuous and vicious partiality in people’s moral judgments, and virtue theorists may struggle to adequately distinguish them, much as proponents of other normative ethical theories do. This paper first adapts the “expanding moral circle” concept and some literary examples to illustrate the difficulty of (...) adequately distinguishing virtuous from vicious partiality. Later sections aim to show how an adequate philosophical ethics will be able to both a) attribute virtue (and hence praiseworthiness) to agents whose actions may directly serve only their special interests, roles, or special relationships; and b) attribute vice (and hence censure) where an agent’s attitudes and/or actions mirror known personal biases (for instance, an egoistic attitude) or social biases (for instance, and ethnocentric attitude). This dual ability leaves space for much virtuous partiality, while also reflecting a “risk-aware” approach, recognizing the deleterious consequences of our human penchant for constructing and imbuing moral significance to oftentimes factitious us/them dichotomies. (shrink)

Recently, psychologists have developed indirect measurement procedures to predict suicidal behavior. A prominent example is the Death/Suicide Implicit Association Test (DS-IAT). In this paper, I argue that there is something special about the DS-IAT which distinguishes it from different IAT measures. I argue that the DS-IAT does not measure weak or strong associations between the implicit self-concept and the abstract concept of death. In contrast, assuming a goal-system approach, I suggest that sorting death-related to self-related words takes effort because death-related (...) words trigger avoidance-impulses, which suicide ideation weakens. The DS-IAT taps into weakened automatic responses from the self-preservation system. Additionally, the suggested cognitive structure, illuminated with the selfish-goal theory, explains predictable suicidal behavior. (shrink)

Can agents overcome unconscious psychological influences without being aware of them? Some philosophers and psychologists assume that agents need to be aware of psychological influences to successfully control behavior. The aim of this text is to argue that when agents engage in a proactive control strategy, they can successfully shield their behavior from some unconscious influences. If agents actively check for conflicts between their actions and mental states, they engage in reactive control. For engaging in reactive control, agents need awareness (...) of those mental states which are in conflict with an action. In contrast, if agents are actively maintaining a goal in consciousness, they engage in proactive control. Proactive control does not consist of conflict detection or conflict resolution. I argue that proactive control explains how agents overcome unconscious psychological influences. In doing so, I claim that consciousness is important for engaging in reactive and proactive control. (shrink)

Implicit biases help maintain disparities between social groups. Arguably, in order to understand how to combat the effects of implicit biases we need to know what kind of mental states the implicit attitudes that undergird them are. Moreover, whether or not you care about combatting the effects, we need to understand what implicit attitudes are to begin understanding human cognitive architecture. Accordingly, one question that a lot of recent research in philosophy and psychology has focused on is: how belief-like are (...) implicit attitudes? The dominant answer from psychologists and philosophers so far is that implicit attitudes are not beliefs. However, in this paper, I argue that implicit attitudes (probably) are beliefs by presenting a general dilemma for various criteria of belief that are supposed to distinguish implicit attitudes from beliefs. Then I argue we can explain frequently cited and prima facie puzzling empirical data that might nevertheless appear to distinguish implicit attitudes from beliefs, if we adopt an independently motivated, fragmented model of the human mind. Finally, I sketch some simple empirical predictions that fall out of the combination of a fragmented model of the mind with the view that implicit attitudes are beliefs, hopefully opening the door for empirical investigations into novel strategies for altering implicit attitudes. (shrink)

Much has been made about the ways that implicit biases and other apparently unreflective attitudes can affect our actions and judgments in ways that negatively affect our ability to do right. What has been discussed less is that these attitudes negatively affect our freedom. In this paper, I argue that implicit biases pose a problem for free will. My analysis focuses on the compatibilist notion of free will according to which acting freely consists in acting in accordance with our reflectively (...) endorsed beliefs and desires. Though bias presents a problem for free action, I argue that there are steps agents can take to regain their freedom. One such strategy is for agents to cultivate better self-knowledge of the ways that their freedom depends on the relationship between their conscious and unconscious attitudes, and the way these work together to inform action and judgment. This knowledge can act as an important catalyst for agents to seek out and implement short- and long-term strategies for reducing the influence of bias, and I offer four proposals along these lines. The upshot is that though bias is a powerful influence on our actions, we need not resign ourselves to its negative effects for freedom. (shrink)

The term 'implicit bias' has very swiftly been incorporated into philosophical discourse. Our aim in this paper is to scrutinise the phenomena that fall under the rubric of implicit bias. The term is often used in a rather broad sense, to capture a range of implicit social cognitions, and this is useful for some purposes. However, we here articulate some of the important differences between phenomena identified as instances of implicit bias. We caution against ignoring these differences: it is likely (...) they have considerable significance, not least for the sorts of normative recommendations being made concerning how to mitigate the bad effects of implicit bias. (shrink)

The current debate on literary cognitivism in the philosophy of fiction typically assumes that we can rigorously distinguish between fictional and factual, and focuses on the question of whether and how works of fiction can impart propositional knowledge to the reader. In this paper we suggest that this way of framing the debate may be problematic. We argue that works of fiction almost inevitably include a reference to the real world and that – contrary to what is usually assumed – (...) the exchange between fiction and reality is vivid as well as potentially fruitful. We shed a new light on these complex dynamics by building on the metaphors of trade exchange and smuggling between the two worlds. While the current debate exclusively focuses on cognitively relevant goods that “officially” cross the border through “customs”, as it were, we show that exchanges between fiction and reality run deeper. Indeed, as we show, a substantial part of the cognitive impact that we derive from fiction goes “under the table” and is “smuggled” from fiction to reality. As such, it bypasses the audience’s conscious control, which in turn explains the manipulative force that fiction can exert. Smuggling takes place when cognitively relevant contents are passed on to the reader in subliminal ways, as it happens when they imprint implicit biases or prejudices, shift perspectives or subtly modify habits or patterns of behavior. By elaborating on the metaphors of trade exchange and smuggling, we aim at presenting a more comprehensive and nuanced understanding of the relations between fiction and reality and on the impact that imagination has on our real-world beliefs. (shrink)

What can we do—and what should we do—to fight against bias? This final chapter introduces empirically-tested interventions for combating implicit (and explicit) bias and promoting a fairer world, from small daily-life debiasing tricks to larger structural interventions. Along the way, this chapter raises a range of moral, political, and strategic questions about these interventions. This chapter further stresses the importance of admitting that we don’t have all the answers. We should be humble about how much we still don’t know and (...) dedicate efforts to gathering as much knowledge as possible. Even so, we know enough now to start making a difference, and this chapter ultimately aims to chip away at the gap between our abstract commitments to treat people fairly and our lived habits and experiences, which continue to be shaped by implicit and explicit prejudices and stereotypes about race, gender, and other social categories. (shrink)

(This contribution is primarily based on "Implicit Bias, Moods, and Moral Responsibility," (2018) Pacific Philosophical Quarterly. This version has been shortened and significantly revised to be more accessible and student-oriented.) Are individuals morally responsible for their implicit biases? One reason to think not is that implicit biases are often advertised as unconscious. However, recent empirical evidence consistently suggests that individuals are aware of their implicit biases, although often in partial and inarticulate ways. Here I explore the implications of this evidence (...) of partial awareness for individuals’ moral responsibility. First, I argue that responsibility comes in degrees. Second, I argue that individuals’ partial awareness of their implicit biases makes them (partially) morally responsible for them. I argue by analogy to a close relative of implicit bias: moods. (shrink)

Affective realism, roughly the hypothesis that you “perceive what you feel”, has recently been put forward as a novel, empirically-backed explanation of police shooting errors. The affective states involved in policing in high-pressure situations result in police officers literally seeing guns even when none are present. The aim of this paper is to (i) unpack the implications of this explanation for assessing police culpability and (ii) determine whether we should take these implications at face value. I argue that while affective (...) realism stands to diminish, if not eliminate, the moral and legal responsibilities of officers who have made shooting errors, the empirical data itself does not directly support such a radical rethink of police culpability. (shrink)

To biologise racism is to treat racism as a neurological phenomenon susceptible to biochemical intervention. In 'Race on the Brain: What Implicit Bias Gets Wrong About the Struggle for Racial Injustice', Kahn (2018) critiques cognitive psychologists and neuroscientists for framing racism in a way that tends to biologise racism, which he argues draws attention and resources away from non-individualistic solutions to racial inequality. In this paper I argue the psychological sciences can accommodate several of Kahn’s criticisms by adopting a situated (...) approach to cognition, where we take environmental features as integral to the cognitive processes that manifest racial bias. (shrink)

Drawing on empirical findings, a number of philosophers have recently argued that people who use English as a foreign language may face a linguistic bias in academia in that they or their contributions may be perceived more negatively than warranted because of their English. I take a critical look at this argument. I first distinguish different phenomena that may be conceptualized as linguistic bias but that should be kept separate to avoid overgeneralizations. I then examine a range of empirical studies (...) that philosophers have cited to argue that people who use English as a foreign language are subject to linguistic bias in academia. I contend that many of these studies do not sufficiently support key claims that philosophers have made about linguistic bias, are challenged by counterevidence, and lack generalizability. I end by introducing methodological recommendations that may help philosophers develop more convincing empirically informed arguments regarding linguistic bias. (shrink)

To Protect and Serve is Norm Stamper’s second book, following Breaking Rank: A Top Cop’s Exposé of the Dark Side of American Policing (2005), and is a welcome contribution to the debate between those who advocate violence against the police and those who insist that no reform is necessary. Radley Balko’s The Rise of the Warrior Cop: The Militarization of America’s Police Forces (2013) inadvertently provided the context of the civil unrest in Ferguson ignited by the shooting of Michael Brown (...) in 2014 and the controversy surrounding the “Ferguson effect” provides the context for Stamper’s treatise on the problems faced – and caused – by contemporary U.S. law enforcement. Stamper is one of the few senior police officers singled out for praise by Balko, for the most part on the basis of his activism against police prejudice and militarisation subsequent to his retirement as Seattle’s police chief in 2000. (shrink)

Mentalizing is our ability to infer agents’ mental states. Attributing beliefs, knowledge, desires, and intentions are frequently discussed forms of mentalizing. Attributing mentalistically loaded stereotypes, personality traits, and evaluating others’ rationality are forms of mentalizing, as well. This broad conception of mentalizing has interesting and important implications for social epistemology. Several topics in social epistemology involve judgments about others’ knowledge, rationality, and competence, e.g., peer disagreement, epistemic injustice, and identifying experts. Mentalizing is at the core of each of these debates. (...) In this chapter, I describe the broad conception of mentalizing and show how it is central to how we judge others’ knowledge, competence, and rationality. I will apply this perspective on mentalizing to two debates: the epistemology of peer disagreement and interventions on testimonial injustice. I argue that understanding how mentalizing works can help us see these debates in a different light. Such reframing can help us make progress on these challenging debates. (shrink)

I argue that what speakers mean or express can be determined by their implicit or unconscious states, rather than explicit or conscious states. Further, on this basis, I show that the sincerity conditions for utterances can also be fixed by implicit states. This is a surprising result which goes against common assumptions about speech acts and sincerity. Roughly, I argue that the result is implied by two plausible and independent theories of the metaphysics of speaker meaning and, further, that this (...) is a robust basis on which to make an inference, with a fair degree of confidence, about the relationship between expression and implicit attitudes. (shrink)

Whenever we communicate, we inevitably have to say one thing before another. This means introducing particularly subtle patterns of salience into our language. In this paper, I introduce ‘order-based salience patterns’, referring to the ordering of syntactic contents where that ordering, pretheoretically, does not appear to be of consequence. For instance, if one is to describe a colourful scarf, it wouldn’t seem to matter if one were to say it is ‘orange and blue’ or ‘blue and orange’. Despite their apparent (...) triviality, I argue that order-based salience patterns tend to make the content positioned first more salient – in the sense of attention-grabbing – in a way that can have surprising normative implications. Giving relative salience to gender differences over similarities, for instance, can result in the activation of cognitively accessible beliefs about gender differences. Where those beliefs are epistemically and/or ethically flawed, we can critique the salience pattern that led to them, providing an instrumental way of evaluating those patterns. I suggest that order-based salience patterns can also be evaluated on constitutive grounds; talking about gender differences before similarities might constitute a subtle form of bias. Finally, I reflect on how the apparent triviality of order-based salience patterns in language gives them an insidious strength. (shrink)

In this paper, I argue that instead of primarily paying attention to the nature of implicit attitudes that are taken to cause implicit discrimination, we should investigate how discrimination can be implicit in itself. I propose to characterize implicit discrimination as unintentional discrimination: the person responds to facts unintentionally and often unconsciously which are, given their end, irrelevant and imply unfair treatment. The result is a unified account of implicit bias that allows for the different ways in which it can (...) display itself and can be explained. Furthermore, the view can account for the central characteristics of implicit bias: (1) that it is, for a variety of reasons, difficult to control, (2) that we are not necessarily unconscious of implicit bias but not properly conscious either, and (3) that we can unintentionally discriminate regardless of whether we claim to care about fairness. (shrink)

The central focus of this paper is to motivate and explore the question, when is it permissible to endorse a psychologizing explanation of a sincere interlocutor? I am interested in the moral question of when (if ever) we may permissibly dismiss the sincere reasons given to us by others, and instead endorse an alternative explanation of their beliefs and actions. I argue that there is a significant risk of wronging the other person, and so we should only psychologize when we (...) are in a position to know that they are in bad faith. (shrink)

The standard associative account of implicit bias posits that the mind unavoidably mirrors the biased co-occurrences that are present in the environment. The resulting fatalistic view of implicit bias as inevitable and immutable is both scientifically unwarranted and societally counterproductive.

This piece is a comment on Quilty-Dunn, Jake, Nicolas Porot, and Eric Mandelbaum. 2023. “The Best Game in Town: The Reemergence of the Language-of-Thought Hypothesis across the Cognitive Sciences.” Behavioral and Brain Sciences 46: e261. -/- The target article signal boosts important ongoing work across the cognitive sciences. However, its theoretical claims, generative value, and purported contributions are – where not simply restatements of arguments extensively explored elsewhere – imprecise, noncommittal, and underdeveloped to a degree that makes them difficult to (...) evaluate. The article's apparent force results from engaging with straw rather than steel opponents. (shrink)

Niche construction theory highlights the active role of organisms in modifying their environment. A subset of these modifications is the developmental niche, which concerns ecological, epistemic, social and symbolic legacies inherited by organisms as resources that scaffold their developmental processes. Since in this theory development is a situated process that takes place in a culturally structured environment, we may reasonably ask if implicit cultural biases may, in some cases, be responsible for maladaptive developmental niches. In this paper we wish to (...) argue for an affirmative answer. In order to do so, we first propose to conceptualize implicit bias as embodied perceptual habits, and then proceed to show that these habits are at least partially responsible for maladaptive developmental niches of children with Down syndrome and autism. With this framework we thus hope to bring together two fields of research that haven’t been explicitly connected: implicit bias and niche construction theory. Linking these two theories may bring benefits both to implicit bias researchers, who can extend this concept to characterize other sets of processes as biased, as well as to niche construction theorists, who will have a useful theoretical tool to diagnose maladaptive features of niches brought about by sociocultural biases. (shrink)

The English language now dominates scientific communications. Yet, many scientists have English as their second language. Their English proficiency may therefore often be more limited than that of a ‘native speaker’, and their scientific contributions (e.g. manuscripts) in English may frequently contain linguistic features that disrupt the fluency of a reader’s, or listener’s information processing even when the contributions are understandable. Scientific gatekeepers (e.g. journal reviewers) sometimes cite these features to justify negative decisions on manuscripts. Such justifications may rest on (...) the prima facie plausible assumption that linguistic characteristics that hinder fast and easy understandability of scientific contributions are epistemically undesirable in science. I shall raise some doubts about this assumption by drawing on empirical research on processing fluency and its biasing effects. I also argue that directing scientists with English as a foreign language toward approaching ‘native-level’ English can have the negative consequence of reducing their potential to make scientific belief formation more reliable. These points suggest that one seemingly compelling justification for linguistically discriminating against scientific contributions in ‘non-native’ English is questionable and that the common insistence by scientific gatekeepers on ‘native-like’ English may be epistemically harmful to science. (shrink)

Mental representations remain the central posits of psychology after many decades of scrutiny. However, there is no consensus about the representational format(s) of biological cognition. This paper provides a survey of evidence from computational cognitive psychology, perceptual psychology, developmental psychology, comparative psychology, and social psychology, and concludes that one type of format that routinely crops up is the language-of-thought (LoT). We outline six core properties of LoTs: (i) discrete constituents; (ii) role-filler independence; (iii) predicate–argument structure; (iv) logical operators; (v) inferential (...) promiscuity; and (vi) abstract content. These properties cluster together throughout cognitive science. Bayesian computational modeling, compositional features of object perception, complex infant and animal reasoning, and automatic, intuitive cognition in adults all implicate LoT-like structures. Instead of regarding LoT as a relic of the previous century, researchers in cognitive science and philosophy-of-mind must take seriously the explanatory breadth of LoT-based architectures. We grant that the mind may harbor many formats and architectures, including iconic and associative structures as well as deep-neural-network-like architectures. However, as computational/representational approaches to the mind continue to advance, classical compositional symbolic structures – that is, LoTs – only prove more flexible and well-supported over time. (shrink)

Marine ecosystems are under increasing pressure from human activity, yet successful management relies on knowledge. The evidence-based policy (EBP) approach has been promoted on the grounds that it provides greater transparency and consistency by relying on ‘high quality’ information. However, EBP also creates epistemic responsibilities. Decision-making where limited or no empirical evidence exists, such as is often the case in marine systems, creates epistemic obligations for new information acquisition. We argue that philosophical approaches can inform the science-policy interface. Using marine (...) biosecurity examples, we specifically examine the epistemic challenges in the acquisition and acceptance of evidence to inform policy, discussing epistemic due care and biases in consideration of evidence. (shrink)

We have higher-order evidence that aesthetic judgements in the context of awarding art prizes may be affected by implicit bias, to the detriment of artists from marginalized groups. Epistemologists have suggested how to respond to higher-order evidence by appeal to bracketing or suspending judgement. We explain why these approaches do not help in this context. We turn to three ways of addressing the operation of implicit bias: (i) anonymization, (ii) the production of objective criteria, (iii) direct implicit bias mitigation techniques. (...) We show that, in the art prize case, strategy (i) is sometimes counterproductive and any benefits are partial, and strategy (ii) is difficult or impossible to implement. This means that the need for (iii) (direct implicit bias mitigation techniques) is more pressing here than elsewhere. The art prize context is one where mitigation of a particular kind is all we are left with. However, domain-specific problems arise for this strategy too, which call for further empirical work on the operation of implicit bias in the artworld. We conclude that the problem of implicit bias as it arises in the specific context of awarding prizes in the artworld is especially challenging and, given the unavailability of alternative mitigations in this context, the need for direct bias mitigation is even more pressing here than in society in general. (shrink)

De mens is irrationeel. Verliefd op zijn eigen gelijk. Doof voor feiten argumenten. Verblind door honderden denkfouten. Een makkelijke prooi voor nepnieuws. Gevangen in filterbubbels. Zo is het heersende idee. Maar klopt het wel? In dit boek verweeft filosoof Maarten van Doorn de jongste inzichten uit de psychologie, communicatiewetenschappen, filosofie en politicologie. Hij neemt ons mee op een reis langs verrassende onderzoeksresultaten en scherpzinnige filosofen en geeft nieuwe antwoorden op dringende vragen: Waarom geloven we wat we geloven? Hoe veranderen mensen (...) van mening? Zijn we echt in een post-truth tijdperk aanbeland? Hoe gevaarlijk is misinformatie? Zijn onze emoties of onze ratio de baas in ons hoofd en in onze maatschappij? Waarom we beter denken dan we denken biedt een hoopvol geluid in een gepolariseerde wereld. (shrink)

AI tools such as ChatGPT appear to magnify some of humanity’s worst qualities, and fixing those tendencies will be no easy task.

This paper argues for a view described as risk-limited indulgent permissivism. This term may be new to the epistemology of disagreement literature, but the general position denoted has many examples. The paper argues for the need for an epistemology for domains of controversial views (morals, philosophy, politics, and religion), and for the advantages of endorsing a risk-limited indulgent permissivism across these domains. It takes a double-edge approach in articulating for the advantages of interpersonal belief permissivism that is yet risk-limited: Advantages (...) are apparent both in comparison with impermissivist epistemologies of disagreement, which make little allowance for the many distinct features of these domains, as well as in comparison with defenses of permissivism which confuse it with dogmatism, potentially making a virtue of the latter. In an appropriately critical form of interpersonal belief permissivism, the close connections between epistemic risk-taking and our doxastic responsibilities become focal concerns. (shrink)

The "myside bias'' in evaluating arguments is an empirically well-confirmed phenomenon that consists of overweighting arguments that endorse one's beliefs or attack alternative beliefs while underweighting arguments that attack one's beliefs or defend alternative beliefs. This paper makes two contributions: First, it proposes a probabilistic model that adequately captures three salient features of myside bias in argument evaluation. Second, it provides a Bayesian justification of this model, thus showing that myside bias has a rational Bayesian explanation under certain conditions.

Our understanding of implicit bias and how to measure it has yet to be settled. Various debates between cognitive scientists are unresolved. Moreover, the public’s understanding of implicit bias tests continues to lag behind cognitive scientists’. These discrepancies pose potential problems. After all, a great deal of implicit bias research has been publicly funded. Further, implicit bias tests continue to feature in discourse about public- and private-sector policies surrounding discrimination, inequality, and even the purpose of science. We aim to do (...) our part by reconstructing some of the recent arguments in ordinary language and then revealing some of the operative norms or values that are often hidden beneath the surface of these arguments. This may help the public learn more about the science of implicit bias. It may also help both laypeople and scientists reflect on the values, interests, and stakeholders involved in establishing, justifying, and communicating scientific research. (shrink)

A slew of newspaper articles were published in the 2010s with titles like: “The facts on why facts alone can’t fight false beliefs” and “Why Facts Don’t Change Our Minds — New discoveries about the human mind show the limitations of reason”. They promoted a common idea: if a person doesn’t conform to the scientific majority, it’s because she forms beliefs on scientific questions in order to achieve social goals (to fit in with people of her kind, to make her (...) social life more comfortable) instead of engaging in an earnest hunt for the truth. Rational persuasion doesn’t work with her. To change her mind, science communicators must become more paternalist. They must adopt methods of persuasion that bypass her awareness—the arts of the marketeer, the ad man. Drawing upon ideas from my recent paper, this piece aims to convince you not to take these articles so seriously. Te piece is a shortened and simplified presentation of ideas from the paper "Science Communication, Cultural Cognition, and the Pull of Epistemic Paternalism.". (shrink)

There is a correlation between positions taken on some scientific questions and political leaning. One way to explain this correlation is the cultural cognition hypothesis (CCH): people's political leanings are causing them to process evidence to maintain fixed answers to the questions, rather than to seek the truth. Another way is the different background belief hypothesis (DBBH): people of different political leanings have different background beliefs which rationalize different positions on these scientific questions. In this article, I argue for two (...) things. I argue that two attempts by proponents of the CCH to discredit the DBBH fail. And I argue that this matters, because while the CCH makes epistemic paternalistic interventions seem called for (as some philosophers have argued compellingly), the DBBH does not. The DBBH makes it much easier to stay closer to an ideal of deliberative democracy. (shrink)

Addressing biases in natural-language processing (NLP) systems presents an underappreciated ethical dilemma, which we think underlies recent debates about bias in NLP models. In brief, even if we could eliminate bias from language models or their outputs, we would thereby often withhold descriptively or ethically useful information, despite avoiding perpetuating or amplifying bias. Yet if we do not debias, we can perpetuate or amplify bias, even if we retain relevant descriptively or ethically useful information. Understanding this dilemma provides for a (...) useful way of rethinking the ethics of algorithmic bias in NLP. (shrink)

Based on a review of several “anomalies” in research using implicit measures, Machery (2021) dismisses the modal interpretation of participant responses on implicit measures and, by extension, the value of implicit measures. We argue that the reviewed findings are anomalies only for specific—influential but long-contested—accounts that treat responses on implicit measures as uncontaminated indicators of trait-like unconscious representations that coexist with functionally independent conscious representations. However, the reviewed findings are to-be-expected “normalities” when viewed from the perspective of long-standing alternative frameworks (...) that treat responses on implicit measures as the product of dynamic processes that operate on momentarily activated, consciously accessible information. Thus, although we agree with Machery that the modal view is empirically unsupported, we argue that implicit measures can make a valuable contribution to understanding the complexities of human behavior if they are used wisely in a way that acknowledges what they can and cannot do. (shrink)

What attitude does someone manifesting implicit bias really have? According to the default representationalist picture, implicit bias involves having conflicting attitudes (explicit versus implicit) with respect to the topic at hand. In opposition to this orthodoxy, dispositionalists argue that attitudes should be understood as higher-level dispositional features of the person as a whole. Following this metaphysical view, the discordance characteristic of implicit bias shows that someone’s attitude regarding the topic at hand is not-fully-manifested or ‘in-between’. However, so far few representationalists (...) have been convinced by dispositionalist arguments, largely because dispositionalism cannot provide explanations in terms of underlying processes. We argue that if dispositionalism wants to be a genuine contender, it should make clear what it has to offer in terms of understanding of implicit bias. As a concrete proposal, we combine dispositionalist metaphysics with the idea that our normative practices of attitude ascription partly determine what it means to have an attitude. We show that such regulative dispositionalism can account for two prominent normative features of implicit bias. We conclude by suggesting that in order to engage in a meaningful debate with representationalism, dispositionalists might have to put the question ‘what counts as a good explanation?’ back on the table. (shrink)

Implicit attitudes are mental states posited by psychologists to explain behaviors including implicit racial and gender bias. In this paper I investigate the belief view of the implicit attitudes, on which implicit attitudes are a kind of implicit belief. In particular, I focus on why implicit attitudes, if they are beliefs, are often resistant to updating in light of new evidence. I argue that extant versions of the belief view do not give a satisfactory account of this phenomenon. This is (...) because proponents of the belief view have focused on overly narrow explanations of recalcitrance in terms of belief storage. Expanding the focus of the belief view to the kinds of irrational and arational transitions between beliefs and other mental states provides compelling (if preliminary) explanations of recalcitrance. (shrink)

A growing number of studies on fairness in artificial intelligence (AI) use the notion of intersectionality to measure AI fairness. Most of these studies take intersectional fairness to be a matter of statistical parity among intersectional subgroups: an AI algorithm is “intersectionally fair” if the probability of the outcome is roughly the same across all subgroups defined by different combinations of the protected attributes. This paper identifies and examines three fundamental problems with this dominant interpretation of intersectional fairness in AI. (...) First, the dominant approach is so preoccupied with the intersection of attributes/categories (e.g., race, gender) that it fails to address the intersection of oppression (e.g., racism, sexism), which is more central to intersectionality as a critical framework. Second, the dominant approach faces a dilemma between infinite regress and fairness gerrymandering: it either keeps splitting groups into smaller subgroups or arbitrarily selects protected groups. Lastly, the dominant view fails to capture what it really means for AI algorithms to be fair, in terms of both distributive and non-distributive fairness. I distinguish a strong sense of AI fairness from a weak sense that is prevalent in the literature, and conclude by envisioning paths towards strong intersectional fairness in AI. (shrink)

Racial, Gender, and Intersectional Biases in AI / -/- Dominant View of Intersectional Fairness in the AI Literature / -/- Three Fundamental Problems with the Dominant View / 1. Overemphasis on Intersections of Attributes / 2. Dilemma between Infinite Regress and Fairness Gerrymandering / 3. Narrow Understanding of Fairness as Parity / -/- Rethinking AI Fairness: from Weak to Strong Fairness.

Some artificial intelligence systems can display algorithmic bias, i.e. they may produce outputs that unfairly discriminate against people based on their social identity. Much research on this topic focuses on algorithmic bias that disadvantages people based on their gender or racial identity. The related ethical problems are significant and well known. Algorithmic bias against other aspects of people’s social identity, for instance, their political orientation, remains largely unexplored. This paper argues that algorithmic bias against people’s political orientation can arise in (...) some of the same ways in which algorithmic gender and racial biases emerge. However, it differs importantly from them because there are strong social norms against gender and racial biases. This does not hold to the same extent for political biases. Political biases can thus more powerfully influence people, which increases the chances that these biases become embedded in algorithms and makes algorithmic political biases harder to detect and eradicate than gender and racial biases even though they all can produce similar harm. Since some algorithms can now also easily identify people’s political orientations against their will, these problems are exacerbated. Algorithmic political bias thus raises substantial and distinctive risks that the AI community should be aware of and examine. (shrink)

It has recently been argued that to tackle social injustice, implicit biases and unjust social structures should be targeted equally because they sustain and ontologically overlap with each other. Here I develop this thought further by relating it to the hypothesis of extended cognition. I argue that if we accept common conditions for extended cognition then people’s implicit biases are often partly realized by and so extended into unjust social structures. This supports the view that we should counteract psychological and (...) social contributors to injustice equally. But it also has a significant downside. If unjust social structures are part of people’s minds then dismantling these structures becomes more difficult than it currently is, as this will then require us to overcome widely accepted ethical and legal barriers protecting people’s bodily and personal integrity. Thus, while there are good grounds to believe that people’s biases and unjust social structures ontologically overlap, there are also strong ethical reasons to reject this view. Metaphysical and ethical intuitions about implicit bias hence collide in an important way. (shrink)

Does algorithmic political bias contribute to an entrenchment and polarization of political positions? Franke argues that it may do so because the bias involves classifications of people as liberals, conservatives, etc., and individuals often conform to the ways in which they are classified. I provide a novel example of this phenomenon in human–computer interactions and introduce a social psychological mechanism that has been overlooked in this context but should be experimentally explored. Furthermore, while Franke proposes that algorithmic political classifications entrench (...) political identities, I contend that they may often produce the opposite result. They can lead people to change in ways that disconfirm the classifications. Consequently and counterintuitively, algorithmic political bias can in fact decrease political entrenchment and polarization. (shrink)

Many scientists routinely generalize from study samples to larger populations. It is commonly assumed that this cognitive process of scientific induction is a voluntary inference in which researchers assess the generalizability of their data and then draw conclusions accordingly. We challenge this view and argue for a novel account. The account describes scientific induction as involving by default a generalization bias that operates automatically and frequently leads researchers to unintentionally generalize their findings without sufficient evidence. The result is unwarranted, overgeneralized (...) conclusions. We support this account of scientific induction by integrating a range of disparate findings from across the cognitive sciences that have until now not been connected to research on the nature of scientific induction. The view that scientific induction involves by default a generalization bias calls for a revision of the current thinking about scientific induction and highlights an overlooked cause of the replication crisis in the sciences. Commonly proposed interventions to tackle scientific overgeneralizations that may feed into this crisis need to be supplemented with cognitive debiasing strategies against generalization bias to most effectively improve science. (shrink)

Much is known about homelessness from a quantitative perspective in Finland. However, the implications are often misleading and false. In this report, I present how prejudiced conclusions about the homeless are drawn in the City of Turku because there is no interest in grassroots experience. Targets to reduce homelessness still make sense.

This paper addresses some of the contours of an ethics of knowledge in the context of ameliorative epistemology, where this term describes epistemological projects aimed at redressing epistemic injustices, improving collective epistemic practices, and educating more effectively for higher-order reflective reasoning dispositions. Virtue theory and embodiment theory together help to tie the cultivation of moral and epistemic emotions to cooperative problem-solving. We examine one cooperative vice, ‘knavery,’ and how David Hume’s little-noticed discussion of it is a forerunner of contemporary game (...) theory’s concern with behavior of ‘free riding’ on systems of trust or cooperation. We develop these and other examples as calling for “risk-aware” social epistemology, which ties epistemic risk with doxastic responsibility. Risk-aware social epistemology aims at describing and ameliorating epistemic practices, but without confusing over-generalizing about culpability in ways that would conflate philosophical censure or critique of bias, with “blameworthiness” in any stronger sense. (shrink)

The fog of war—poor intelligence about the enemy—can frustrate even a well-prepared military force. Something similar can happen in intellectual debate. What I call the *fog of debate* is a useful metaphor for grappling with failures and dysfunctions of argumentative persuasion that stem from poor information about our opponents. It is distressingly easy to make mistakes about our opponents’ thinking, as well as to fail to comprehend their understanding of and reactions to our arguments. After describing the fog of debate (...) and outlining its sources in cognition and communication, I consider a few policies we might adopt upon learning we are in this fog. (shrink)

For almost half a century dual-stream advocates have vigorously defended the view that there are two functionally specialized cortical streams of visual processing originating in the primary visual cortex: a ventral, perception-related ‘conscious’ stream and a dorsal, action-related ‘unconscious’ stream. They furthermore maintain that the perceptual and memory systems in the ventral stream are relatively shielded from the action system in the dorsal stream. In recent years, this view has come under scrutiny. Evidence points to two overlapping action pathways: a (...) dorso-dorsal pathway that calculates features of the object to be acted on, and a ventro-dorsal pathway that transmits stored information about skilled object use from the ventral stream to the dorso-dorsal pathway. This evidence suggests that stored information may exert significantly more influence on visually guided action than hitherto assumed. I argue that this, in turn, supports the notion of skilled automatic action that is nonetheless agential. My focus here will be on actions influenced by implicit biases (stereotypes/prejudices). Action that is biased in this way, I argue, is in an important sense intentional and agential. (shrink)