Seemings and the Foundations of Justification: A Defense of Phenomenal Conservatism [1 ed.] 9781003281108

Citation preview

“Blake McAllister presents a highly original and well-​argued defense of phenomenal conservatism. With such able defenders, there is a reason that phenomenal conservatism is an increasingly popular view of epistemic justification.” Logan Paul Gage, Franciscan University of Steubenville, USA

Seemings and the Foundations of Justification

All justified beliefs ultimately rest on attitudes that are immediately justified. This book illuminates the nature of immediate justification and the states that provide it. Simply put, immediate justification arises from how things appear to us—​from all and only our “seemings.” The author defends each aspect of this “seemings foundationalism,” including the assumption of foundationalism itself. Most notably, the author draws from common sense philosopher Thomas Reid to present new and improved arguments for phenomenal conservatism and gives the first systematic argument that seemings alone are capable of immediately justifying. The discussion delves deeply into the nature of seemings and how it is that their assertive phenomenal character makes them (and them alone) capable of immediately justifying. Along the way, the author makes novel contributions to perennial debates such as: internalism versus externalism, deontologism and epistemic blame, epistemic circularity, and the common sense response to skepticism. Seemings and the Foundations of Justification will appeal to scholars and advanced students interested in epistemology, Thomas Reid, or the common sense tradition. Blake McAllister (PhD, Baylor University) is Associate Professor of Philosophy at Hillsdale College. He has published extensively in epistemology with special emphasis in early modern and religious epistemology. His work has appeared in venues such as Synthese, Faith & Philosophy, and History of Philosophy Quarterly.

Routledge Studies in Epistemology Edited by Kevin McCain, University of Alabama at Birmingham, USA and Scott Stapleford, St. Thomas University, Canada

Epistemic Care Vulnerability, Inquiry, and Social Epistemology Casey Rebecca Johnson The Epistemology of Modality and Philosophical Methodology Edited by Anand Jayprakash Vaidya and Duško Prelević Rational Understanding From Explanation to Knowledge Miloud Belkoniene Illuminating Errors New Essays on Knowledge from Non-​Knowledge Edited by Rodrigo Borges and Ian Schnee Digital Knowledge A Philosophical Investigation J. Adam Carter Seemings and the Foundations of Justification A Defense of Phenomenal Conservatism Blake McAllister

For more information about this series, please visit: www.routle​dge.com/​Routle​dge-​Stud​ies-​ in-​Epist​emol​ogy/​book-​ser​ies/​RSIE

Seemings and the Foundations of Justification A Defense of Phenomenal Conservatism Blake McAllister

First published 2024 by Routledge 605 Third Avenue, New York, NY 10158 and by Routledge 4 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2024 Blake McAllister The right of Blake McAllister to be identified as author of this work has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. ISBN: 978-​1-​032-​24795-​3 (hbk) ISBN: 978-​1-​032-​24994-​0 (pbk) ISBN: 978-​1-​003-​28110-​8 (ebk) DOI: 10.4324/​9781003281108 Typeset in Sabon by Newgen Publishing UK

For my wife, Bess

Contents

Acknowledgments

x

Introduction

1

1 Classical Deontologism

13

2 Immediate Justification and Seemings Foundationalism

50

3 Some Objections to Phenomenal Conservatism

75

4 What Are Seemings?

101

5 Foundationalism Defended

142

6 The Case for Seemings Exclusivism

163

7 Against Epistemic Conservatism

190

8 A Reidian Argument for Phenomenal Conservatism, Part I: Some Seemings

208

9 A Reidian Argument for Phenomenal Conservatism, Part II: All Seemings

245

Index

270

Acknowledgments

I am thankful to so many for their support throughout the long process of writing this manuscript. Many of the ideas in this book were developed while writing my dissertation at Baylor University under the direction of Jonathan Kvanvig and with extended assistance from Trent Dougherty. I also benefited there from discussions with my graduate colleagues—​ particularly John Bishop, Logan Gage, Ross Parker, Chris Tweedt, Alli Thornton, and Ryan West. Various ideas from that time were written and rewritten many times over. I am grateful to Matt Frise, Kevin McCain, Andrew Moon, Ted Poston, Bradley Rettler, Lindsay Rettler, and Chris Tucker for commenting on early versions of various chapters. Several of my students aided me through their questions and conversation including Sarah D’Spain, Thomas Reusser, and the members of my epistemology courses in the Fall of 2018 and 2022. The support of my family has been essential to making this possible, especially that of my wife, Bess. Along with our children, you have given me the best years and deepest joys of my life. Above all, I am thankful to God, the ultimate author of all good gifts. Some of the material in this book has been reused. Sections of Chapter 1 appeared previously as “Justification without Excuses: A Defense of Classical Deontologism,” American Philosophical Quarterly 59, no. 4 (2022): 353–​366, and “Rescuing a Traditional Argument for Internalism,” Synthese, 201, 144 (2023). The first half of Chapter 4 was published as “Seemings as Sui Generis” Synthese, 195, no. 7 (2018): 3079–​3096. And Chapter 7 was first published as “From One Conservative to Another: A Critique of Epistemic Conservatism” Southwest Philosophical Review, 37, no. 2 (2021): 167–​186.

newgenprepdf

Introduction

Humans are rational agents. One aspect of our rationality is manifested in the formation of beliefs and other doxastic attitudes on the basis of reasons. The ideal, of course, is not that these beliefs be based on just any reasons but on good ones—​reasons sufficient to render those beliefs (in some sense) likely to be true. When this ideal is achieved, our beliefs are epistemically justified. Often enough, our reasons for belief consist in other beliefs: we infer one thing from another. But this cannot be true in all cases. The chain of reasoning, if it is to successfully justify anything, cannot regress infinitely but instead must begin with foundational beliefs that are not justified on the basis of any other beliefs. This is because beliefs inherit their ability to justify from whatever it is that justifies them, and a regress consisting entirely in such intermediaries would have no ability to justify at all. No more than a family consisting entirely of inheritors could be wealthy. Someone must generate the money before it can be passed along to the others. Likewise, something must justify inherently, and so serve as an ultimate source of justification, if anything is to be justified inferentially. This familiar position is known as “foundationalism,” and it is my first thesis. The perennial mystery of foundationalism is how our foundational beliefs come to be justified. Their justification cannot depend on having justification for any other beliefs (via inference or otherwise) lest they cease to be the regress stoppers we need them to be. Instead, they must be justified immediately. But what sort of thing could immediately justify our beliefs? Some foundationalists claim that our foundational beliefs are immediately justified without being immediately justified by anything. Foundational beliefs have no rational basis but are, in select conditions, immediately justified nonetheless. In contrast, I appeal to experiences as the reasons on which our foundational beliefs are based. If we identify these experiential reasons as evidence, then this amounts to the claim that

DOI: 10.4324/9781003281108-1

2  Introduction all immediate justification stems from foundational evidence—​a position I call “foundations evidentialism.” This is my second thesis. The appeal to experience does not dispel the mystery of foundationalism so much as locate it. For the experiences that justify our foundational beliefs must be of a remarkable sort. They must be the sorts of things that can justify belief even if we lack verification of their truth or reliability; otherwise, the justification they provide would not be immediate and the regress would continue. These experiences must be inherently trustworthy. This raises more questions than it answers: how can it be rational to trust something without verification? Of which experiences is this true? What is it about them that makes it so? I will explain my answers to all of these questions in the following chapters. For now, let me simply tell you what those answers are. The experiences that serve as our foundational evidence, and so constitute the ultimate sources of justification for all of our beliefs, consist in all and only our seemings. Seemings are, quite simply, states that have propositional content which feels true (or false, but we’ll get to that later). Consider, for instance, whether it is the case that: If A is bigger than B, and B is bigger than C, then A is bigger than C. Surely this is true. But why do we think so? For no other reason, we might say, than that it is obvious, or evident, or that we simply see that it is true. What we are describing, I propose, is the presence of a particular kind of mental state. One that has the preceding principle as its content, and which presents that content to us as true. Many have called states of this particular sort “intuitions.” We can also call them “rational seemings.” There are many other species of seemings (perceptual, memorial, introspective, and more) that resemble these rational seemings in essential respects—​they too have propositional content that feels true—​yet differ in the domain of their content and the manner of their production. What unites all of these under the genus of seemings, however, is their forcefulness—​that distinct phenomenal character by virtue of which their content feels true. It is also this forceful presentation of propositional content that makes seemings inherently trustworthy. When a proposition is presented as true, this counts in favor of that proposition’s truth in a way that, in the complete absence of other considerations, entitles one to take some positive doxastic attitude toward it. In other words, the seeming is foundational evidence for its content. This is a species of the view known as “phenomenal conservatism,” my third thesis. My claim here is not only that some seemings are foundational evidence, but that all are. Having forcefulness and propositional content is sufficient for something to count as foundational evidence. Forcefulness and propositional content are also necessary for foundational evidence. Without forcefulness, a proposition might be entertained but not in a way that counts in favor of its truth (as when one

Introduction  3 imagines, hopes, or desires that something is the case). Without propositional content, an experience might feel revelatory, but there isn’t any specific claim to which one can assent. However, any state that possesses both forcefulness and propositional content is by definition a seeming. Thus, foundational evidence can consist in nothing other than seemings. This is my fourth thesis, called “seemings exclusivism.” Together, these four theses constitute a comprehensive theory of the foundations of justification that I will call “seemings foundationalism.” In sum, all non-​immediate justification comes from immediate justification; all immediate justification comes from foundational evidence; and foundational evidence consists in all and only our seemings. Seemings foundationalism is strictly about the foundations of justification. I do not purport to settle how it is that foundational beliefs are weighed against one another, although I will make some claims about this in the course of our inquiry. My main concern is simply to identify the need for and sources of immediate justification, as well as to explain how it is that such immediate justification is possible. I am far from the first, of course, to identify seemings as sources of immediate justification. Phenomenal conservatism is mainstream these days, thanks mostly to the work of Michael Huemer, who ushered talk of “seemings” and “phenomenal conservatism” into the contemporary debate with his superb book, Skepticism and the Veil of Perception. One of the points that I would like to stress, however, is that same basic view was present long before Huemer coined these terms (as I’m sure Huemer himself would be eager to concede). In fact, the view enjoys a long history of proponents going all the way back to ancients such as Aristotle, the Stoics, and the Academic Skeptics. I even plan to surprise many epistemologists today by asserting that they too qualify as phenomenal conservatives of one sort or another, though several will no doubt be scandalized by the label. My aim in this book is not primarily to defend phenomenal conservatism against the many objections leveled against it. The literature is full of these negative defenses. What is sorely missing are positive defenses of the view—​ fresh arguments for the truth of phenomenal conservatism. In fact, I think it fair to say that the main arguments for phenomenal conservatism have not advanced much beyond where Huemer left them over 20 years ago in his original presentation of the view. Yet few, if any, take those arguments to be fully successful. Promising? Sure. Tantalizing? Certainly. But convincing? Eh, that’s unclear. Despite this hesitation, phenomenal conservatism continues to win droves of new adherents and as many or more dissidents willing to spend their time critiquing it. Why does phenomenal conservatism possess such allure? There is, of course, its refreshing simplicity, its promise to unite all immediate justification under one banner, its power to vanquish skepticism and restore a common sense philosophical methodology. But when it

4  Introduction comes down to it, I think many people just have the strong sense (dare I say, it seems to them) that something like Huemer’s original arguments must be right, even if these arguments haven’t received their strongest formulation. This book attempts to provide that strongest formulation. And not only with respect to phenomenal conservatism. Seemings exclusivism is the other half of the equation. While Huemer gestures at this position in his case for phenomenal conservatism, he never actually defends it. Huemer’s mentality appears to be that if you have seemings then you don’t need any other source of immediate justification, so why think there is one? As far as I can tell, then, I will be the first to give any kind of systematic argument that there can’t be another one—​that seemings are the only states capable of providing immediate justification. Doing so allows me to strengthen several of Huemer’s original arguments for phenomenal conservatism in crucial ways. This is the lacuna that prevented those arguments from fully convincing. In order to defend seemings foundationalism, or even to fully explain what it is, we must first take some controversial stances on the nature of justification. These days, disagreement about justification extends all the way down such that there is no settled framework for thinking about justification from which we can all take our start. Perhaps we can say without too much controversy that justification requires good reasons (as I have done above), but that is about all we can say. As soon as we say anything about what we mean by reasons, or what makes them good, we enter into contested territory. I especially have in mind here the debate between internalists and externalists. These camps approach justification in such fundamentally different ways that we must take a stand here before we can proceed with anything else. The disagreement between internalists and externalists hinges in large part on whether justification is evaluated from a first-​or third-​ person point of view. On the first-​person approach favored by internalists, justification depends on what is accessible to us. Considerations that are not accessible to us from the first-​person perspective are irrelevant (with respect to justification, at least). Externalists take more a third-​person approach. What justifies a belief might not be accessible to us at all. Both internalists and externalists agree that justified beliefs must be, if not true, then at least likely to be true; but they understand this likelihood in different ways. For internalists, justified beliefs are likely to be true in the sense that they are indicated to be true (on-​balance) by the evidence available within one’s first-​person perspective. For externalists, justified beliefs are likely to be true in the sense that those beliefs (or the processes generating them) are objectively reliable or modally truth-​tracking. We cannot have a conversation about what provides justification and why without first deciding which of these two approaches to adopt.

Introduction  5 In Chapter 1, I argue for an internalist conception of justification. It may be that externalists are also discussing some important epistemic status, but that status is not justification. My argument is simply that, until very recently, justification has always referred to a first-​person status centered around following the evidence. This is evident if we go back to the figures from which our debates about justification descend—​namely, Descartes, Locke, and other early modern epistemologists. Such thinkers also conceived of justification as deontological, as pertaining to what we are permitted to believe when operating in accordance with our epistemic duties. This classical deontologist conception of justification is, I propose, the original one, and as such has a kind of default status for us. That default status can be overcome if classical deontologism proves problematic, but all of its main objections can be addressed. I consider in particular the accusations that classical deontologism relies on doxastic voluntarism to account for epistemic blame and reflective transparency to motivate internalism. I show how in both cases classical deontologism can be sustained without any appeal to voluntarism or transparency. The result is a significant reorientation of the classical deontological conception of justification that, unlike externalism, avoids departing from that conception in its central commitments. The rest of my project assumes this renovated version of classical deontologism. I must stress, however, that the importance of my project as a whole does not hinge decisively or even predominantly on whether the term “justification” refers to the positive epistemic status described by classical deontologism or not. (The proper use of this term is important, just not all important.) Call it “internal justification” or “rationality” or something else. What matters most is (i) that there is in fact some positive epistemic status possessing the features which classical deontologism ascribes to justification and (ii) that this epistemic status is important and so worth investigating. If we can agree on that, then we have motivation to move forward. We can understand the discussion in Chapter 1 as providing a defense of the first point. By addressing the most formidable objections to classical deontologism, I remove the most serious reasons for doubting that there is a coherent epistemic status belonging to all and only those attitudes that we are permitted to believe by the evidence available to us from the first-​person point of view. Once we grant that there is such a status, its importance is, I think, evident. The ultimate goal, to be sure, is to secure true beliefs while avoiding false ones, but any sensible attempt we make to get at truth must coincide with what our evidence indicates to be true. To amend an example from Huemer (2001, 105), say someone commands us to point to all and only the things that are apples and then, a moment later, to all and only the things that our evidence indicates to

6  Introduction be apples. If we are doing our best to follow the former, then we will already be doing the latter. This is because doing our best to secure true beliefs and avoid false ones just is to believe in accordance with what our evidence indicates to be true. Why is this the case? Because believing truly (and not believing falsely) isn’t something we can just do. If it were, then following the evidence would be unnecessary. But the human predicament is that believing truly is something we must do, if we do it at all, by trying to do it. This trying must be something that is directly up to us to ensure, and that is intelligible to us as a means of pursuing truth and avoiding falsehood. In short, to try our best to believe truly is simply to follow our evidence. Many today will contest whether following the evidence is directly up to us to ensure, but I defuse such concerns in Chapter 1. I go on in Chapter 2 to explicate a general framework for thinking about justification on the revised classical deontological approach. I then articulate seemings foundationalism within this general framework. In particular, I say more about what qualifies a view as an instance of phenomenal conservatism and how my particular formulation of it compares to others. In Chapter 3, I address a handful of objections to phenomenal conservatism. As stated above, my primary intention is not to address objections but to build a positive case. I’m playing offense, not defense. Nevertheless, it seems incumbent on me to say something in response to objections—​if only to those that immediately present themselves—​to ensure us that the view is not considered so implausible as to be doomed from the start. Thus, I examine whether my version of phenomenal conservatism is too liberal in dispensing justification, whether it licenses viciously circular reasoning, whether it leads to a problematic “double-​counting” of our evidence, and whether it can really sustain a common sense response to skepticism as advertised. On this last point, I argue that the common sense approach, when coupled with phenomenal conservatism, is actually quite formidable as a response to skepticism. Chapter 4 lays the final groundwork for my defense of seemings foundationalism. There I explicate and defend the account of seemings sketched above: that seemings are states whose propositional content feels true. After arguing that seemings are best understood as experiences instead of beliefs or inclinations to believe, I address the complaints that such experiential seemings are weird, or underdefined, or that they simply do not exist. Along the way, I provide one of the most detailed accounts of seemings on offer. I give special attention to what we mean when we say that their content feels true, as well as to the relationship between seemings and other states like perceptual experience and direct acquaintance. An important insight as it relates to later chapters is that many of the mental states that foundationalists have traditionally appealed to as foundational

Introduction  7 evidence end up qualifying as seemings, although their proponents would not have called them that. My argument for seemings foundationalism begins in Chapter 5 with a defense of foundationalism. I employ the familiar regress argument (previewed above) to show how we can sustain the alternatives to foundationalism—​infinitism and coherentism—​only by accepting significant revisions to the way we ordinarily think about justification. Such revisions come with serious and unavoidable costs. In principle, those costs might be worth paying if the costs for foundationalism are as bad or worse; but, in actual fact, foundationalism is in a much better position. The most serious worry for foundationalism consists in what is called “Sellars’ dilemma” or sometimes “BonJour dilemma,” but seemings are at least plausibly positioned to split the horns of that dilemma. The prospects for its resolution are promising, whereas the problems raised for infinitism and coherentism are irremediable. In Chapter 6, I set about examining which kinds of states might serve as foundational evidence for us. I argue that for a state to be foundational evidence it must have propositional content—​indeed, the same content as the belief it is supposed to be evidence for—​or else the subject could not appreciate the relationship between that state and the relevant belief apart from justified background beliefs. I also argue that this propositional content must be presented as true or else that state cannot indicate the truth of that content, again, apart from a set of justified background beliefs. But if a state has propositional content that feels true, then that state is a seeming. Hence, foundational evidence consists in nothing other than seemings, just as seemings exclusivism maintains. Chapter 7 then argues that immediate justification requires foundational evidence. Epistemic conservatives are the main ones who resist this conclusion, at least within the internalist, foundationalist camp, so I can effectively argue for foundations evidentialism by arguing against epistemic conservatism. The case for epistemic conservatism is mainly one based on theoretical fruit—​it is admitted to be initially implausible, but justified nonetheless by its usefulness in epistemic theorizing. I undercut these motivations by showing how phenomenal conservatism can provide the same theoretical benefits from a much more plausible starting point. This leads us into the final chapters, which contain my defense of phenomenal conservatism. Up till this point, I have suggested that seemings can plausibly serve as foundational evidence, but it is here where I argue that they do serve in this capacity—​all of them. The argument is spread over two chapters. Chapter 8 argues for a minimal version of phenomenal conservatism (Minimal-​PC): the position that at least one seeming is foundational evidence for its content. In support of this principle, I update four arguments found in Huemer’s original defense.

8  Introduction (1) (2) (3) (4)

Denying Minimal-​PC leads to skepticism. Denying Minimal-​PC is self-​defeating. Minimal-​PC seems true. Minimal-​PC is a constitutive principle of rationality.

With Minimal-​PC in place, I argue in Chapter 9 that if some seemings are foundational evidence, then all are. To avoid this conclusion, dissenters must point to some epistemically relevant distinction between the seemings that do serve as foundational evidence and the seemings that don’t. Most point to external features which, given the internalist approach defended in Chapter 1, are not relevant with respect to justification. Others point to internal differences that are relevant only if we have background evidence allowing us to appreciate their significance. Such differences do not, however, affect a seeming’s status as foundational evidence. In the end, I argue that there are no relevant differences between seemings, at least as it pertains to their status as foundational evidence. Some seemings may be stronger foundational evidence than others, but all seemings are foundational evidence if only of a minimal sort. Before diving in, I must acknowledge my indebtedness to early modern epistemologist, and Scottish minister turned philosopher, Thomas Reid. A contemporary of Hume and Kant, Reid is scandalously ignored in most undergraduate courses in early modern philosophy. Hume is presented as the culmination of British Empiricism, and Kant as the definitive and seemingly only possible alternative. And it is true in some sense that these two chart the course for much of mainstream philosophy in the years following them. Continental philosophy largely followed Kant in his Copernican revolution, though they take it far beyond anything Kant himself would have endorsed; the early analytics wed Hume’s empiricism with symbolic logic and the linguistic turn. Although Reid was arguably more popular at the time, especially outside of narrow philosophical circles, for many years Reid fell out of favor with the philosophical elites. But on the contemporary scene, Reid’s star is ascendant. In epistemology, he is one of the earliest progenitures of fallibilism, moderate foundationalism, externalism (specifically, something like proper functionalism about warrant), testimonial non-​reductionism, direct realism, particularism, and the common sense response to skepticism. No philosopher I can think of, ancient or modern, better anticipated (and in many cases inspired) the epistemological advancements of the second half of the twentieth century. When one adds to this Reid’s growing prominence in other areas (e.g., philosophy of mind, personal identity, free will), Reid has earned himself a place at the table with Hume and Kant. He may lack some of the brilliance and polish of Hume, or the towering intellect of Kant, but he has the virtue of being right—​right more often, at least, than either Hume or Kant managed to be.

Introduction  9 Reid inspires many of the central principles and arguments in this book (a fact to which I will routinely draw the reader’s attention). For one, this book is located squarely within the common sense epistemological tradition, and Reid is an early and influential advocate for many of the foremost principles within that tradition (see, e.g., Boespflug 2019a). For another, Reid is one of the primary inspirations for my defense of phenomenal conservatism in Chapters 8 and 9. Alas, though I deploy Reid’s reasoning to defend phenomenal conservatism, I cannot claim him as a proponent. As I interpret him, Reid argues that all of our natural beliefs (roughly, those resulting from the proper functioning of our natural constitutions) are prima facie justified to a degree corresponding to the strength of that belief. This differs from phenomenal conservatism in a couple of ways. For one, Reid does not acknowledge the existence of seemings or any cognitive state prior to our beliefs. Instead, foundational evidence is identified with whatever it is that serves as the immediate occasion of those natural beliefs—​the things directly triggering the natural principles of our constitution—​when everything is functioning properly. In perception, Reid understands this to be sensations. With self-​evident proposition, it is the mere consideration of those propositions themselves that triggers a natural belief (hence, they are, in a sense, evidence for themselves). While this is certainly a significant difference, it is also relatively surface level. I actually agree with Reid that our foundational evidence must consist in those things that serve as the immediate causes of our beliefs in a properly constituted and properly functioning rational agent. I argue for this view (with some qualifications) in Chapter 1. The disagreement is over which mental states we have, not over whether and how those mental states justify. So, for instance, if Reid were convinced that there were such things as seemings, and that these seemings were the immediate causes of our natural beliefs in ideal human reasoners, then he would happily concede that those seemings were foundational evidence for their content as phenomenal conservatism maintains. Of course, even if Reid were to concede this, the fact remains that the status of perceptual seemings as foundational evidence for our perceptual beliefs would be contingent for him, as opposed to being necessary on phenomenal conservatism (Boespflug 2019b). To be clear, it is necessary for Reid that whatever grounds our natural beliefs (i.e., beliefs formed through the proper functioning of our rational faculties) is good evidence for those beliefs; however, there is no state identified by Reid as necessarily grounding our natural beliefs.1 It is reasonable to suppose, then, that were Reid to have acknowledged seemings as the grounds of our natural perceptual beliefs, their status as foundational evidence would have been contingent as well. In contrast, standard versions of phenomenal

10  Introduction conservatism say it is necessary that seemings are foundational evidence for their content. This disagreement over modal status is the symptom of a deeper and more important disagreement, which is that, for Reid, what qualifies a mental state as foundational evidence isn’t anything about the state itself. Its evidential status is determined by whether belief is the response called for by the creature’s design plan.2 The stronger the belief called for, the stronger the foundational evidence. So even if Reid did come to acknowledge seemings as foundational evidence, what would make them foundational evidence is simply that we are naturally constituted so as to respond to them with belief when functioning properly. Phenomenal conservatives have usually held, however, that there is something about seemings themselves—​something about their unique phenomenal character in particular—​that makes it permissible for the agent to believe their content. For standard phenomenal conservatives, the properly functioning human responds to seemings with belief because that is the fitting response, whereas for Reid belief is the fitting response to seemings because that is what is called for by the human design plan. Now, it is possible to build a kind of Reidian phenomenal conservatism around the idea that, as a contingent fact about the human design plan, seemings (or at least some of them) are foundational evidence for their contents (see Bergmann 2013), but that is not the sort of view that phenomenal conservatives have standardly endorsed nor the sort that I wish to defend here. I cannot help but wonder, however, whether Reid would have come to a different conclusion were he to have acknowledged seemings. These points of contention with phenomenal conservatism seem forced onto him by his flawed theory of perception. Might he welcome a change were that perceptual theory to be corrected? To see what I mean, we must appreciate that Reid is enmeshed in the early modern perceptual theory that dominated from Hume all the way back to Descartes. This is the theory on which what we call “perception” is really just “introspection” of various sensory impressions in our own minds. On this view, sensory impressions had no representational content, nor did they stem any kind of direct encounter with external substances themselves. They were just raw feels bearing no similarity to the states of affairs that they prompt us to think about. Taking this view for granted, Reid concluded that those sensations cannot be foundational evidence for our perceptual beliefs about the external world by virtue of anything about those sensations themselves. In principle, we could be constituted such that the same raw feels trigger, and so serve as foundational evidence for, vastly different beliefs.3 However, had Reid acknowledged the existence of perceptual seemings as a kind of intermediary between sensations and beliefs, then there is now a state prior to belief that doesn’t seem arbitrarily connected

Introduction  11 to beliefs about the external world. This removes any need to think of their evidential status as contingent. So, it’s not out of the question to suppose that Reid might have become a standard phenomenal conservative had he acknowledged the existence of perceptual seemings.4 But these are mere speculations. In any case, it doesn’t distress me that Reid isn’t a proponent of phenomenal conservatism. He is one of its central inspirations nonetheless. Without further introduction, then, let us proceed with our inquiry. Notes 1 Indeed, evidence is largely dispensable on Reid’s approach. What matters is whether the belief is produced in accordance with the proper functioning of our natural constitutions. For if so, then we are shielded from blame with respect to them. Which states prompt such “natural beliefs,” or whether there are any such states at all, is not what is truly accounts for their justification. Hence, if we abstract away from the particularities of the design plan, I think it fair to interpret Reid as an epistemic conservative, or at least as allowing for the kind of justification that epistemic conservatives envision—​namely, that our natural beliefs are given presumption simply because we have them. Though I admit that this Reidian version of epistemic conservatism is non-​standard (just as a Reidian version of phenomenal conservatism would be). 2 It may also need to be a reliable, truth aimed design plan, as contemporary proper functionalists like Bergmann (2006) stress. Reid may have had these nuances in mind, but he does not spell them out clearly. 3 Though Reid’s theory of perception is deeply flawed, there is a grain of truth here. If perception is a direct encounter with external objects, there is still a way in which it feels to encounter those objects. It could be that the way in which it feels to encounter an object is contingent such that, in other worlds, what it feels like to perceptually encounter the very same object is completely different. (This might even be possible while holding all our bodily states constant, in which case we have an argument for dualism.) But that is very different from the epistemic view at issue here, which treats these feelings as foundational evidence for the existence of objects in the external world. 4 Boespflug argues that the explanation Reid gives for why our natural beliefs are prima facie justified (what Boespflug calls a “meta-​justification”) is incompatible with phenomenal conservatism; however, I think Boespflug misunderstands the nature of Reid’s meta-​justification. On my interpretation, Reid thinks we are justified in holding onto our natural beliefs sans defeaters because we cannot be justly blamed for doing so (see McAllister 2016 and Chapter 7, §7.4 of this volume); and the ultimate defense of our blamelessness comes in the idea that trusting our natural beliefs is a constitutive principle of rationality. This sort of meta-​justification, however, can and should be welcomed by phenomenal conservatives. Indeed, following Reid, I make precisely this sort of case in Chapter 8, §8.5.

12  Introduction References Bergmann, Michael. 2006. Justification Without Awareness: A Defense of Epistemic Externalism. Oxford: Oxford University Press. —​—​—​. 2013. “Externalist Justification and the Role of Seemings.” Philosophical Studies 166 (1): 163–​184. Boespflug, Mark. 2019a. “The Legacy of Reid’s Common Sense in Analytic Epistemology.” Journal of Scottish Philosophy 17 (1): 23–​37. —​—​—​. 2019b. “Why Reid Was No Dogmatist.” Synthese 196 (11): 4511–​4525. Huemer, Michael. 2001. Skepticism and the Veil of Perception. Lanham, MD: Rowman & Littlefield. McAllister, Blake. 2016. “Re-​Evaluating Reid’s Response to Skepticism.” Journal of Scottish Philosophy 14 (3): 317–​339.

1 Classical Deontologism

Any inquiry into the foundations of epistemic justification must begin with a general understanding of what justification is. Beliefs and other doxastic attitudes enjoy many features we find commendable, but which of these is justification specifically? Here I will build the case for classical deontologism. On this view, justification is deontological. It is a status belonging to those attitudes that are permitted by our epistemic duties. It is also internalist. The conditions relevant to justification are located entirely within the mind of the subject. My basic argument will be that classic conception of justification is the correct conception because it is the original one—​this is the status that the language of “justification” was originally used to pick out. If so, then those who use the term “justification” to denote other positive epistemic statuses are changing the subject. Those other statuses may be just as important as justification, or even more so, but they are not justification. The matter is not quite so simple as this, of course. For we can alter the meaning of a term over time, and often there are good reasons for doing so. The term “justification” may have originally referred to an internal, deontological status, but if it turns out that our conception of this status is deeply problematic—​if it turns out to be incoherent or based on some crucial misunderstanding—​then it may be perfectly sensible to shift the meaning of “justification” toward some other status in the vicinity that does not share these problems. Furthermore, those who oppose the classical deontological conception of justification generally do so precisely because they take that status to be problematic in the above sorts of ways. Thus, an adequate defense of classical deontologism will require that I address such concerns. 1.1  What Is Classical Deontologism? My starting assumption is that the contemporary notion of justification emerges predominantly from the epistemological debates of the early

DOI: 10.4324/9781003281108-2

14  Classical Deontologism modern period—​ particularly from Descartes and Locke, though also Hume, Reid, and others as well. This is not to say that discussions relevant to justification do not occur before this. Everyone is familiar, for instance, with Plato’s claim in the Theaetetus that knowledge requires true judgment with an account (201d). And the philosophical schools after Plato foreshadowed many of the same debates about justification present today. Nevertheless, our conversations about justification today are continuations of early modern debates in a way that isn’t quite true of ancient or medieval debates (even if the latter contain valuable lessons for the former). It is in this sense that the contemporary notion of justification originates from early modern epistemology (see Plantinga 1993 n24). I propose that the original conception of justification should serve as the default for our own. By this I mean two things. First, as a matter of language, if the term “justification” originally denotes X but not Y, then other things being equal, “justification” will continue to denote X and not Y, and those who try to use it to denote Y are simply misusing the term. Of course, if people continually use “justification” to denote Y instead of X, then at some point they can actually shift the meaning of the term such that “justification” (at least in some contexts) actually does come to denote Y. (I fear this has happened or is dangerously close to happening in epistemology today.) So, second, I also mean that we ought to resist this semantic shift and the linguistic behavior that leads to it. After all, it is not always improper to change the meaning of a term, say, if it is currently being used to refer to X but Y is more coherent than X and better fits the role in our system that X was originally intended to play. But the onus is on those who would deviate from the original meaning. Thus, we should resist reconceiving of justification unless the original conception proves to be problematic in serious ways. When we look at early modern figures, what we see is that justification is a decidedly deontological notion. We were assumed to have epistemic duties, responsibilities, and requirements governing how to believe, and the beliefs that failed to meet these were deemed impermissible or unjustified.1 Those that did not violate such duties were permitted or justified. (I say more about the nature of these epistemic duties in Chapter 2.) We can capture this in the following principle: Deontologism—​S’s belief is unjustified if and only if it is epistemically impermissible for S. The deontologism of the early moderns seems to be why the term “justification” was chosen in the first place—​a term that “suggests duty, obligation, requirement … the whole deontological stable” (Plantinga 1993, 4).

Classical Deontologism  15 Descartes and Locke are especially clear that they are relying on deontological conceptions of justification (see Plantinga 1993 Ch. 1, §3–​4). Locke writes: He that believes, without having any Reason for believing, may be in love with his own Fancies; but neither seeks Truth as he ought, nor pays the Obedience due his Maker, who would have him use those discerning Faculties he has given him, to keep him out of Mistake and Error. He that does not this to the best of his Power, however he sometimes lights on Truth, is in the right but by chance; and I know not whether the luckiness of the Accident will excuse the irregularity of his proceeding. This at least is certain, that he must be accountable for whatever Mistakes he runs into: whereas he that makes use of the Light and Faculties GOD has given him, and seeks sincerely to discover Truth, by those Helps and Abilities he has, may have this satisfaction in doing his Duty as a rational creature, that though he should miss Truth, he will not miss the Reward of it. For he governs his assent right, and places it as he should, who in any Case or Matter whatsoever, believes or disbelieves, according as Reason directs him. He that does otherwise, transgresses against his own Light, and misuses those Faculties, which were given him … (Locke [1689] 1975, 687–​688) In this passage, Locke indicates that we have a duty to believe in a justified manner. We are accountable for our errors when we do not and stand in need of excuse if we are to escape reactive responses. Indeed, Locke’s suggestion is that to fail to believe as one should is to disobey God and may therefore be considered a kind of epistemic sin. Descartes adds to this: But if I hold off from making a judgment when I do not perceive what is true with sufficient clarity and distinctness, it is clear that I am acting properly and am not committing an error… But were I to embrace the other alternative, it will be by sheer luck that I happen upon the truth; but I will still not be without fault, for it is manifest by the light of nature that a perception of the part of the intellect must always precede a determination on the part of the will. Inherent in this incorrect use of free will is the privation that constitutes the very essence of error. (Descartes [1641] 2006, 33) Descartes says here that we are at fault, or blameworthy, for believing as we ought not. Unjustified beliefs stem, he says, from “the privation that constitutes the very essence of error,” which he elsewhere defines as “a lack

16  Classical Deontologism of some knowledge that somehow ought to be in me” (Descartes [1641] 2006, 30). They arise, in other words, from some deficiency in us or our intellects. While neither as explicit nor severe as Descartes and Locke, other early moderns like Hume and Reid also employ deontological conceptions of justification (McAllister 2016). For instance, Reid is constantly arguing against Hume—​ his paradigm skeptic—​ that the man of common sense cannot be blamed for believing as he does. Nicholas Wolterstorff writes, “Reid’s skeptic has his eye on entitlement—​on what one is permitted to believe and on what one is not permitted to believe—​on what one may believe and on what one must not believe” (Wolterstorff 2001, 188). Reid and Hume’s point of dispute is not a practical one: both men acknowledge that, practically speaking, we cannot be blamed for holding such beliefs given their indispensability for governing the affairs of common life. Rather, the charge that Hume lays against the man of common sense is a distinctively epistemic one—​it is that, unlike the wise man, the man of common sense does not apportion his beliefs to the evidence and so believes in a manner that is epistemically impermissible. Hume doesn’t seem to think of this as a form of intellectual sin, as Locke does, but he does think it unwise. And it is this epistemic accusation that Reid attempts to refute. Regardless of who is right (I will take Reid’s side in later chapters), both Hume and Reid are thinking of justification deontologically. Another aspect to the early modern’s deontological conception—​again, most prominently in Descartes and Locke—​is that we could guarantee the justification of our beliefs if only we form those beliefs with sufficient care. Plantinga writes: “As the classical deontologist sees things, … whether we are justified in our beliefs is up to us.” He continues: We need give no hostages to fortune when it comes to justification; here our destiny is entirely in our own hands. The fates may conspire to deceive me … . Even so, I can still do my epistemic duty; I can still do my best; I can still be above reproach. (Plantinga 1993, 15) The idea is that we can ensure justification not only in the actual world, which we can suppose to be friendly, but even in those worlds most hostile to the truth of our beliefs (e.g., evil demon worlds, brain-​in-​a-​vat worlds, and the like). At best, such guarantees extend only to those beliefs that we are allowed to form with sufficient care. If something interferes with our reasoning, preventing us from reaching a settled position on the matter, then all bets are off. (Extreme examples include beliefs forced on us through neurological rewiring, brainwashing, or hypnosis. More ordinary examples include beliefs formed while sleep deprived, under a serious time crunch,

Classical Deontologism  17 or intoxicated.) Thus, we should limit our focus to non-​exempt beliefs, which are those that result from the full and unobstructed operation of our rational belief-​forming dispositions. I will refer to this traditional idea—​ that it is within our power to guarantee the justification of our non-​exempt beliefs—​by saying, “it is possible for us to be impeccable with respect to justification.” We can define impeccability more precisely as: Impeccability—​S is impeccable with respect to justification if and only if S possesses a total set of rational belief-​forming dispositions such that, in all possible worlds where S has that set of dispositions, all of S’s non-​ exempt beliefs are justified. The assumption that it is possible for us to be impeccable with respect justification led Descartes and Locke to internalism about justification. Or perhaps it was their internalism that led them to toward the possibility of impeccability. In any case, these positions are traditionally thought to go hand in hand (Plantinga 1993, Goldman 1999). By internalism, we mean (as a first pass) the following: Internalism—​The conditions that would justify S’s belief that p were they to obtain are all internal to the mind of S. We will discuss how best to characterize internality later, in §1.4. The basic idea linking impeccability and internalism is that, by placing all of the justifying conditions within our own minds, it is thereby within our power to ensure that we believe if and only if those conditions obtain. On the other hand, if some of the justifying conditions are external to us, then we do not have the power to ensure this outcome—​a reality that externalists themselves are generally willing to admit. We will return to how to flesh out this basic idea below. For now, let’s quickly take stock of the discussion thus far. The original and thus default conception of justification has the following three features: i Unjustified beliefs are those that are impermissible for us to hold. ii It is possible for us to be impeccable with respect to justification—​ to guarantee that we only form justified beliefs if allowed to do so unobstructed. iii The justifying conditions on our beliefs are all internal to our minds. Together, these form the basic contours of the classical deontological conception of justification as I defend it here. To be sure, I will end up disagreeing with Descartes and Locke on how to understand or account for

18  Classical Deontologism each of these principles, sometimes in important ways. In a sense, then, even I am proposing something of a revision to the original conception of justification. Nevertheless, my revisions remain faithful to that original vision in a way that externalist alternatives do not. Externalists adopt a conception of justification which rejects both of the last two principles and in some cases the first as well. Indeed, it is easy to sympathize with Laurence BonJour when he complains, “it finally seems possible to make sense of externalism only by construing the externalist as simply abandoning the traditional notion of epistemic justification” (BonJour 1978, 8). This might lead us to wonder how externalists came to such a radically different understanding. Are they making an egregious mistake? Here’s my take. Externalists are making a mistake, but not an egregious one. The explanation of this error begins with Descartes. Justification, we can all agree, is a means to truth; but Descartes thought of this connection to truth in two distinct ways. First, he thought of justification as an intentional means to truth, meaning that justified beliefs must be on-​balance indicated to be true (to a sufficient degree) by our evidence. This ensures that justified beliefs have a kind of first-​person, subjective connection with the truth. He also thought of justification as an effective means to truth, meaning that justified beliefs are in fact guaranteed to be true. This gives justified beliefs a kind of third-​person, objective connection with the truth. Descartes saw these two truth-​connections as related. For it is by carefully following our evidence that we guarantee the truth of our beliefs. In other words, having an intentional connection to truth entailed having an effective connection as well. In time, Descartes’s infallibilism—​his claim that justification precluded any possibility of error—​proved untenable. If justification is infallible, then we are justified in believing very little, for there is very little about which we cannot be mistaken. In light of this, we may resign ourselves to a more moderate effective connection. Perhaps following our evidence guarantees only probable truth, or objective reliability. Alas, that too is untenable. Evil demon victims are case in point. The perceptual beliefs of these victims enjoy the same evidence ours do, and so bear the same intentional connections to truth, yet their beliefs are not only false but also highly unreliable. Arguably, one of the central lessons from early modern epistemology is that intentional and effective connections to truth are logically independent. There are no substantive entailments from the latter to the former. Following the evidence does not guarantee that our beliefs are true or even reliable for anything beyond the incorrigible.2 To be sure, we must still believe that following the evidence is an effective means to truth, but only as a contingent matter of fact, and not as something following from the nature of evidence itself.

Classical Deontologism  19 In the wake of this realization, which connection do we build our concept of justification around? Do we conceive of justification most fundamentally as an intentional means to truth or an effective one?3 Externalist opt for the latter. This was a mistake, I think, but it is at least an understandable mistake. Descartes originally conceived of justification as coming with an effective connection to truth, and externalist theories of justification capture this. Furthermore, we have, because of this effective connection, traditionally thought of justification as making justified beliefs good candidates for knowledge. Once again, externalist theories of justification capture this association as well as or better than internalist theories. So, many of the features we traditionally associate with justification are well accounted for by externalism. Despite these merits, externalists picked the wrong side. For Descartes, justification was always most fundamentally about an intentional connection to truth, with the effective connection being a supposed consequence of believing in accordance with our evidence. That’s why Descartes ended up thinking about justification as guaranteeing both kinds of connections: it is natural (though ultimately mistaken) to think that following the evidence ensures the truth of our beliefs. Conversely, no one is tempted to think that coming to true beliefs requires following the evidence. Hence, if Descartes’s fundamental concern had been to secure an effective connection to truth, he never would have been led to think of justification as coming with an intentional connection at all. Thus, if forced to pick a side, we should think of justification as an intentional rather than effective means to truth, as this has been its fundamental orientation since inception. The externalist has one final card to play. If it turns out that the classical deontological conception of justification is severely problematic, then perhaps we would be warranted in moving away from this original conception toward something more tenable. And this is no idle threat. Externalists have offered serious reasons for thinking that the traditional conception is problematic. We will look at these objections more closely in the next section, and I will attempt to rebut them in the sections after that. The conclusion of this section, however, is that if the classical deontological conception can sustain these objections, then it is the one we ought to go with. 1.2  Objections to Classical Deontologism One of the most well-​known objections to the classical deontological conception of justification is that it requires doxastic voluntarism. William Alston writes: By the time honored principle that “Ought implies can,” one can be obliged to do A only if one has an effective choice as to whether to do

20  Classical Deontologism A. It is equally obvious that it makes no sense to speak of S’s being permitted or forbidden to do A if S lacks an effective choice as to whether to do A. And it seems even more obvious, if possible, that S cannot be rightly blamed for doing (or not doing) A if S is incapable of effectively deciding whether or not to do A. (Alston 1988, 259) In short, a deontological conception of justification “is viable only if beliefs are sufficiently under voluntary control to render such concepts as requirement, permission, obligation, reproach, and blame applicable to them” (Alston 1988, 259). While many throughout history have endorsed some form of doxastic voluntarism—​including (arguably) Augustine, Aquinas, Descartes, Locke, Pascal, and William James—​most today reject it, and understandably so. We seem to have direct voluntary control over few if any of our beliefs. We may have indirect voluntary control over many more, but this control does not seem extensive enough to sustain a general duty to believe in a particular manner (Alston 1988). There may be ways to defend voluntarism (e.g., Steup 2000), but for the purposes of this chapter, I will grant its demise. This threatens to render the classical deontological conception of justification untenable. The second objection we will consider is directed at the possibility of impeccability and its supposed grounding in internalism. The traditional account of impeccability held by Descartes and Locke (Plantinga 1993, ­chapter 1)—​which I will call “the reflective account”—​appeals first to our prime epistemic position with respect to internalist norms.4 On this account, the mind is reflectively transparent to us, meaning:5 Necessarily, we are always in a position to know (or justifiably believe) exactly which mental states we are in. In addition, the support relations that obtain between those mental states and our belief contents are thought to be reflectively transparent as well, meaning: Necessarily, we are always in a position to know (or justifiably believe) what our mental states do or do not indicate to be true, and to what extent. Since, on internalist norms, the normatively relevant conditions are exhausted by which mental states one is in and what those states indicate to be true, it follows that internalist norms are normatively transparent to us, meaning:

Classical Deontologism  21 Necessarily, we are always in a position to know (or justifiably believe) what we should and should not believe according to internalist norms. On the reflective account, it is because internalist norms are normatively transparent, and the norms of justification are internalist in nature, that being impeccable with respect to justification is a possibility for us. The problem is that many contemporary thinkers reject normative transparency. Regarding the transparency of the mental, there are intuitive counterexamples (Huemer 2007, 35, Srinivasan 2015, 275), empirical critiques (Gopnik 1993, Carruthers 2011, Schwitzgebel 2006; 2011), and philosophical objections—​the most notable being Timothy Williamson’s anti-​luminosity argument (Williamson 2000, ­chapter 4). It is also increasingly common to assert that we do not always have unproblematic access to which belief contents our mental states support and to what extent (Kvanvig 2014, ­chapter 2, §5). There are formidable defenses in light of such concerns (Smithies 2019), but the proponents of transparency are rapidly thinning. For the sake of this discussion, I will grant such critiques (though I do not thereby endorse them). If transparency goes, then so does the reflective account. To complete the case, note that there does not seem to be any way of re-​grounding the same kind of reflective explanation on more moderate assumptions. For example, one might appeal to the fact that we are often in a position to know what internalist norms require of us; but this more moderate control over our beliefs does not guarantee that our non-​exempt beliefs will be justified across all possible worlds. Moreover, we have this same limited control over satisfying externalist norms. Insofar as we are generally aware of the cognitive processes leading to the formation of our beliefs, we are, in that respect, also generally aware of whether we are meeting whatever demands externalist norms make of us. So, this more moderate reflective account does not create sufficient separation between internalist and externalist norms to accommodate an argument for internalism on that basis. The growing doubts about normative transparency explain why many, including many internalists, have abandoned the possibility of impeccability about justification altogether. For if justification is not normatively transparent, then there will be worlds in which we are, despite our best efforts, ignorant of what we should and should not believe. But then it seems that we have no way of ensuring that in those worlds we are always believing as we should and never as we shouldn’t. Amia Srinivasan thus writes: Without privileged access to our own minds, there are no norms that can invariably guide our actions, and no norms that are immune from

22  Classical Deontologism blameless violation. This will come as bad news to those normative theorists who think that certain central normative notions—​e.g. the ethical ought or epistemic justification—​should be cashed out in terms of subjects’ mental states precisely in order to generate norms that are action-​guiding and immune from blameless violation. ... More generally, once we have accepted that our relationship to our own minds lacks the perfect intimacy promised by Cartesianism, we are, for better or worse, left with the view that the normative realm is suffused with ignorance and bad luck. (Srinivasan 2015, 273) We have, then, two reasons for thinking that the classical deontological conception of justification is hopelessly ill-​conceived. The first is that the deontological aspects require that we have pervasive voluntary control over our beliefs, which we do not have. The second is that the possibility of impeccability requires that the norms of justification be transparent to us, which they are not. How are we to respond? Assuming voluntarism is out of the picture, the only way of salvaging classical deontologism from the first objection is to show how the application of deontological concepts to beliefs remains valid despite our lack of voluntary control over them. This I will do in §1.3. I will rebut the second objection in §1.4 by showing how we can account for impeccability entirely apart from transparency.6 1.3  Making Sense of Epistemic Blame Our first task must be to get clear on what it takes for deontological concepts like impermissibility or duty to have application within the epistemic sphere. I will then show how those conditions can be satisfied sans voluntarism. For convenience, I will focus exclusively on negative evaluations, on how belief might be impermissible or blameworthy. There are different ways of characterizing epistemic deontologism, some more strenuous than others. To start, Feldman suggests that a theory may be an instance of deontologism so long as it meets the following minimal condition: “Sentences about what one ought to believe, and what is required or permissible, follow from distinctively epistemic evaluations” (Feldman 2008, 354–​355). We must then ask ourselves what the truth conditions are for sentences like, “It is impermissible for S to believe p.” Feldman himself seems to opt for a set of truth conditions that are easily satisfied, making deontologism fairly uncontroversial. For instance, what is thought to be the main alternative to deontologism for internalists conceives of justification in terms of fit with the evidence (Conee and Feldman 2004, ­chapter 4), but Feldman himself classifies this theory as meeting the minimal characterization of deontologism (Feldman 2000; 2008, §5). If we

Classical Deontologism  23 want to avoid the concern that our deontologism is insubstantial or trivially true, we can do so by requiring more for deontological sentences to come out true. One might think that sentences like “It is impermissible for S to believe p” are true only if believing p entitles us to take certain negative reactive attitudes or actions either toward S or toward S’s belief that p. Such characterizations of deontologism come in many varieties, again some more strenuous than others, depending on the kind of reactive responses licensed by impermissible beliefs. With respect to reactive attitudes, an important question is whether we must be entitled to blame the subject for impermissible beliefs (sans an excuse)? And if so, what is blame and when is it appropriate? (Blame is far more severe, and blameworthiness harder to assign, on some views than on others.) With respect to reactive actions, an important question is whether we must be entitled to sanction or punish blameworthy subjects and in what ways? Or are lesser reactive actions sufficient? Rather than trying to answer all of these questions, I will simply propose a characterization of epistemic deontologism that is plainly substantial. It will require that the subject be blameworthy for impermissible beliefs in the absence of an excuse, and that impermissible beliefs open one up, not to sanction or punishment, but at least to some minimal reactive actions. As we begin, remember that our account of epistemic impermissibility and blame cannot suppose that we have any substantial direct, or even indirect, voluntary control over what we believe. We are looking for whether deontologism can be preserved in the absence of doxastic voluntarism. But in ruling out voluntarism, Alston’s worry looms large: how can we be blamed for our beliefs if we do not have a choice about them? This worry has at least three components. The first concerns the idea that “ought implies can.” We must be able to believe as we should to be blamed for not doing so, but presumably I can’t believe other than I do unless belief is voluntary. The second worry, related to the first, is that I must have control over what I believe in order to be blamed for believing in a particular manner. But what could that control consist in except voluntarily choosing what to believe? The third and final worry is that even if there are some non-​voluntarist ways of understanding the required kind of ability and control, those won’t be sufficient to allow us to be truly blameworthy for our beliefs (where the term is uttered with great emphasis). Regarding the first worry, let us imagine that a certain belief of ours, one which breaks a norm, is causally determined—​roughly, the conditions leading up to the belief (including my set of cognitive dispositions) in conjunction with the laws of nature entail that I form that belief. The charge is that I can’t believe as I ought to (the violating belief is determined after all) and so I cannot be blamed for it. The intuition here is similar to the one

24  Classical Deontologism driving the principle of alternate possibilities with respect to moral responsibility. We might look, then, to compatibilist treatments of this principle for guidance. We could, of course, follow certain philosophers in simply denying the need for alternate possibilities (as does Ryan 2003, §4), but we needn’t. For there is a legitimate sense in which we can believe differently even if that belief is determined: that is, if we were to possess different intellectual constitutions, then we would believe differently. Thus, so long as it is possible for us to possess an intellectual constitution, that would result in the normatively appropriate attitude, then we can believe as we ought to in the required sense. Since compatibilism is the predominant position on freedom and moral responsibility, it seems these philosophers at least should concede that the determination of our beliefs does not in and of itself preclude our being blameworthy for them (Ryan 2003, Steup 2008). The second worry says that I can only be blamed for things that are under my control, but beliefs are not things under my control. They happen apart from my will and so I have no say in the matter. Beliefs do happen apart from our wills (we are granting), but that does not mean we have no say in the matter. The objection assumes that I do not really control my beliefs unless I am able to determine what to believe in some extrinsic way: by coming to reflect on what my epistemic norms require of me and then choosing which attitude to take. On such extrinsic models of control, “I exercise [agential control over my beliefs], not in believing itself, but in doing certain other things that affect my beliefs” (Boyle 2009, 126). Assuming that real control is extrinsic control, David Owens argues that we needn’t have control over our beliefs to maintain responsibility for them (Owens 2000), and he is right given that assumption. But I think we would do better to reject that assumption. We ought instead to think about control over beliefs in a more intrinsic way. Matthew Boyle explains, “On this view, we exercise our capacity for cognitive self-​determination, not primarily in doing things that affect our beliefs, but in holding whatever beliefs we hold” (Boyle 2009, 127). In defense of this view, Boyle draws a comparison to actions: If I have agential control over anything, I certainly have it over the things I do intentionally. Whether to perform these actions is up to me. But the control I exercise over my intentional actions is surely not an extrinsic form of control. I do not control them by acting on them. Rather, my intentional actions are themselves my acts: they are not extrinsically but intrinsically under my control. (Boyle 2009, 142) On this approach, when an action arises out of the rational dispositions constituting my character, that action is under my control, not because

Classical Deontologism  25 I am able to change that action through reflection or will that I will differently, but simply because it arises from those rational dispositions which are my own. I perform that action because I am the way that I am—​because certain reasons for acting that way are compelling to me. Such actions are under my control in a significant way (and this remains true even if there are other, more robust forms of control over those actions that I lack). The same is true of doxastic attitudes. If a belief is the manifestation of my own rational dispositions—​and so is to that degree not exempt—​then that belief is under my control in an important sense. I believe it because my intellect is the way that it is—​because certain reasons for believing that way are compelling to me.7 This is sufficient control to sustain blame. If this seems like cheating, might I suggest that this is really the third worry: that blame is something severe and altogether too harsh to be appropriately directed at intellectual agents for violating the norms of belief—​not unless we could pick and choose what we believe in, which we can’t. If this is the heart of the matter (and I suspect that for most people it is), then may I suggest that this is simply too strong a view of blame. Blame can be especially harsh, but it needn’t be. Here it may be helpful to introduce Angela Smith’s distinction between the depth of an appraisal, “which concerns whether the person can legitimately be asked to justify that for which he is being appraised,” and its significance, “which concerns the kind of importance that attaches to his failing to meet the normative standards that apply in the particular case” (Smith 2008, 384). To say that someone deserves blame for a performance is only to note the depth of an appraisal to which the performing agent is thereby subject; it does not entail that the failure of this performance is of grave significance. Relatedly, the severity of the reactive attitudes and actions that constitute or accompany blame can differ dramatically from case to case. Even within the moral sphere, the manner in which we blame agents for serious moral failings is much more severe than for minor peccadillos. The former might license strong reactive attitudes (condemnation; scorn; abhorrence) and actions (rebuke; social ostracization; imprisonment), but not the latter. And when we consider what blame looks like in the less severe cases, it becomes increasingly clear that deviant beliefs can make us blameworthy in this milder way. To illustrate the point, consider one of the more prominent theories of moral blame from Gary Watson (Watson 1996). On Watson’s view, there is a certain kind of blame in which “to blame (morally) is to attribute something to a (moral) fault in the agent” (Watson 1996, 230–​231). Blame is thus, for Watson, a kind of negative evaluation in which one judges there to be some deficiency within the agent. An action can make one blameworthy, then, if it fails with respect to some standard of excellence and the explanation for that failure traces back to some deficiency within one’s moral character. Now, it is possible for there to be failings for which we are not

26  Classical Deontologism blameworthy that trace back to some deficiency within us. For instance, we may fail to meet certain standards of health because of a defect within one of our organs, but we are not necessarily blameworthy for that failing. It matters, then, that our actions are based on reasons, or at least formed through rational dispositions, in which case they disclose something about our moral agency in ways that meeting or not meeting the standards of health do not. Watson says: But if what I do flows from my values and ends, there is a stronger sense in which my activities are inescapably my own: I am committed to them. As declarations of my adopted ends, they express what I’m about, my identity as an agent. They can be evaluated in distinctive ways (not just as welcome or unwelcome) because they themselves are exercises of my evaluative capacities. (Watson 1996, 233) Pamela Hieronymi makes a similar point in arguing that performances for which one is: answerable for an account of one’s reasons … [are ones that] reveal one’s take on what is true or important or worthwhile, and so can help to determine the quality of one’s will. But the quality of one’s will just is the object of moral assessment and the reactive attitudes. (Hieronymi 2008, 362) Watson calls this “the self-​disclosure view” of responsibility (see also “real self views” in Wolf 1990). Let’s apply all of this to belief. Beliefs too are based on reasons and are thus performances for which we are answerable for an account of our reasons. Accordingly, our beliefs can disclose who we are as intellectual agents. They can reveal the quality of our intellects. Thus, if a belief fails with respect to some epistemic standard of excellence (a genuine epistemic norm) and that failure is attributable to us, then that belief opens us up to a negative evaluative judgment. That is, one is thereby entitled to judge that we are intellectually deficient, or flawed intellectual agents, and that that is why we erred. But, mimicking Watson, “to blame (epistemically) is to attribute something to an (intellectual) fault in the agent.” Thus, we are in Watson’s sense blameworthy for such beliefs. This is a plausible account of epistemic blame. Indeed, Watson’s view is an even more plausible view of epistemic blame than it is of moral blame, since the ways in which we morally blame are often much harsher, and this is a comparatively mild account. Thus, even if one demands more for

Classical Deontologism  27 moral responsibility specifically, Watson’s account may nevertheless capture a significant sense in which we can be epistemically responsible for our beliefs. Nevertheless, one may still sense that this alone isn’t a robust enough account of blame. Plausibly blame also consists in, or in some other way requires, additional negative reactive attitudes or actions. Let us grant that it is so. Notice that the recognition of a deficiency within one’s intellect would certainly license disapproval of that agent or his or her intellectual constitution. This is a negative reactive attitude called for by beliefs that have failed in the relevant sense. As for reactive actions, failed beliefs at the very least open one up to the demand that one “explain or justify her rational activity” and “re-​assess it if an adequate justification cannot be provided” (Smith 2008, 381). This alone may be enough of a reactive action to ground blame (as on Smith 2005; 2008). Failed beliefs may also permit one to encourage or pressure the offending agent to seek further evidence or intellectual development. We therefore have a plausible account of epistemic blame that does not rely on doxastic voluntarism. Neither does it rely on normative transparency. A belief is blameworthy so long as it reveals a deficiency within my intellect, thereby subjecting me to a negative evaluation on that basis. This can be true even if I am completely unaware that my belief results from such a deficiency. What of those who say that we cannot be fairly blamed for a violation we do not know we’re committing? Once again, unless one is thinking about blame as something more severe, the answer to this question has already been made clear: we are exposed to blame by that belief because it testifies to some defect within us that others can legitimately disapprove of and encourage us to correct. Awareness of our failure is unnecessary. Putting all of this together results in a substantial account of epistemic blame and, subsequently, epistemic impermissibility. To review, impermissible beliefs are those that are blameworthy absent excuses, and we are blameworthy for those beliefs that reveal a defect within our intellects. This is because the revelation of this defect permits others to (1) form negative evaluative judgments about us as defective intellectual agents; (2) form negative emotional attitudes such as disapproval toward us or our intellectual constitutions on the basis of that defect; and (3) take reactive actions toward us such as enjoining us to defend our beliefs, to reassess them, to pursue further evidence, or to set about improving the quality of our intellects. If one cannot shake the sense that more severe attitudes or actions are required for genuine blame, then we must respectfully part ways. Though I’m hopeful we can agree that using “epistemic blame” to describe this negative posture is not a gross misuse of the term.

28  Classical Deontologism 1.4  Internalism and Impeccability We now turn to the second objection, which denies that impeccability is possible apart from normative transparency. In the first two subsections, I show how a proper model of belief formation can account for impeccability without reliance on transparency so long as we adopt a specific form of internalism. In the last subsection, I explain why this account only works given internalism, thereby restoring the traditional link between internalism and impeccability. 1.4.1  The Dispositional Account of Impeccability

The reflective account of impeccability deploys what John Pollock calls “the intellectualist model” of norm guidance (Pollock 1987, 64; 1986, ­chapter 5). It envisions that we follow norms by consulting explicitly articulated rules dictating how we ought to believe. But what if there was another way to follow norms—​one that did not require transparency? Here I will outline Pollock’s non-​intellectualist model of norm guidance and show how it might be worked into a non-​reflective account of impeccability. What is this alternative model of norm guidance? Let us begin with an example from the practical realm: You may make reference to a driving manual when you are first learning to drive a car, but once you learn how to drive a car you do not look things up in the manual anymore. You do not usually give any explicit thought to what to do—​you just do it. This does not mean, however, that your behavior is no longer being guided by those norms you learned when you first learned to drive. … The point here is that norms can govern your behavior without your having to think about them. (Pollock 1987, 66) In driving, we see a form of non-​intellectualist norm guidance at work. For example, while driving we see brake lights ahead: we immediately brake. We do not consult the rule, “Brake when you see tail lights turn red,” or form any belief about what we ought to do. Nevertheless, that rule guides our driving. What has happened is that, in learning to drive, we have internalized certain patterns of action. More precisely, we developed a set of rational cognitive dispositions that automatically unfold in behavior that adheres to the norms of driving—​dispositions that are triggered not by reflecting on the fact that we have certain mental states but by the mere presence of those mental states themselves. This set of driving dispositions are not mere reflexes, mind you. With reflexes or other non-​rational dispositions, the trigger serves only as the

Classical Deontologism  29 cause of the resulting action. With a rational disposition, however, the trigger is both a cause and a reason for the resulting action. In the above example, when seeing tail lights ahead causes us to brake, that experience is the reason or rational basis for our action as well as its cause. This ensures that our driving behavior is not merely in conformance with the norms of driving—​that is, we happen to perform the right action in the right conditions—​but in compliance with it—​that is, we perform the right action precisely because we are in the right conditions. When we comply with a norm, we are following or being guided by that norm, even if we do not reflectively consult it. And that is what is happening when we drive. The conditions that dictate we should brake are also the reasons why we do brake, even though we do not reflect on the fact that we have such reasons. A moment’s reflection on our cognitive processing reveals that we usually form beliefs in the same non-​intellectualist way we drive. We are responsive to various mental states as they come into existence, adjusting our beliefs and levels of confidence in accordance with epistemic norms but rarely do we explicitly consult those norms or what they require of us in light of those mental states. We usually do not reflect on the fact that we have such states at all. For example, while driving we have a perceptual appearance of tail lights: we immediately believe that the car in front of us is braking. We do not consult any epistemic principle to tell us whether that appearance provides adequate support for our belief. We don’t form any beliefs about our perceptual appearance either—​any beliefs we have are about the car not our experience of the car. What’s going on is that the appearance itself directly guides our cognition, apart from any reflective beliefs about its existence or normative significance.8 After all, the existence of a mental state or property is a different state of affairs than reflectively believing in the existence of a mental state or property, and the rational dispositions resulting in belief formation can be directly triggered by the former as well as the latter. This is a feature of human cognition that Williamson calls “unreflective causal sensitivity to the evidence” (Williamson 2000, 180; see also Owens 2000, 22). Importantly, what we are describing here are rational belief-​forming dispositions. This means that our experiences don’t merely cause our beliefs but also serve as the reasons for them. What makes the difference between a reason and a mere cause? It seems to have something to do with the way in which these appearances cause our beliefs. I will say more about what this way is later on, but for now we needn’t fixate on why these appearances qualify as reasons, only that they do—​and that this remains the case even when we are not reflectively aware that we are having such appearances.9 On this non-​intellectualist model of belief formation, we can ensure impeccability with respect to our epistemic norms simply by virtue of

30  Classical Deontologism possessing the right “internalized pattern of behavior that we automatically follow in reasoning” (Pollock 1987, 68). As an example, consider an epistemic norm N, which says of a certain state of affairs X: believe p if and only if X obtains. For S to be impeccable with respect to N, S must meet the following condition: Conforming to N—​ S possesses a total set of rational belief-​ forming dispositions such that, in all possible worlds where S has that set of dispositions, S forms a non-​exempt belief that p if and only if X obtains. The conformance condition ensures that S forms the non-​exempt belief that p if and only if belief that p is called for by N, but this is still compatible with S basing his attitudes on all the wrong reasons. Thus, impeccability also requires that S’s beliefs are rationally based on the conditions that permit them:10 Complying with N—​S conforms to N and does so precisely because X does or does not obtain. If both the conformance and compliance conditions are satisfied, then S is impeccable with respect to N. The crucial point is that both of these conditions can be satisfied without S’s being reflectively aware of whether X obtains or not. S just needs to possess the right set of rational dispositions, which will automatically unfold in the right ways in response all and only to the right states of affairs. We will call this “the dispositional account of impeccability.” The dispositional account only works if the states of affairs identified as normatively relevant are, in Pollock’s terminology, “directly accessible” to our rational belief-​forming systems (Pollock 1987, 69). Intuitively, directly accessible states are the ones capable of immediately or directly bringing about belief without requiring that we reflect on our having them. In the driving example, for instance, the appearance of the tail lights brings about our belief without our having to reflect on the fact that this appearance exists. Crucially, then, the “direct accessibility” of these states does not refer to our ability to become immediately aware of them on reflection; it refers instead to the ability of these states to immediately cause us to believe. Thus, the first and most important feature of directly accessible states is that they can trigger our rational belief-​ forming dispositions merely by existing. Another way to put this is: Direct Accessibility (first pass) —​X is directly accessible to S’s rational belief-​forming systems only if X (rather than reflective awareness of X) can be the proximate cause of one of S’s rationally formed beliefs.

Classical Deontologism  31 Since we are dealing with rational belief-​forming dispositions—​ones whose causes are also reasons for belief—​we are operating here on what Wedgwood calls “the folk-​psychological level of explanation.” Wedgwood explains that the folk-​psychological order of explanation operates at “the personal, mental level,” and so features mental states “of the sort that are referred to in everyday folk-​psychological discourse,” such as beliefs, perceptions, appearances, and the like (Wedgwood 2002, 357). These kinds of explanations do not appeal to “subpersonal mechanisms or modules in the brain or anything of that sort” (ibid.). There is, of course, a different explanatory order within which the proximate causes of our beliefs will be brain states; but the states that interest us here are the proximate causes of belief within folk-​psychological explanations since this is the level at which beliefs are evaluated with respect to reasons and rationality. Hence, we must add to our understanding of direct accessibility as follows: Direct Accessibility (second pass)—​X is directly accessible to S’s rational belief-​forming systems only if X can be the proximate cause of one of S’s rationally formed beliefs within the folk-​psychological explanation of that belief. Thirdly, some explanations within the folk-​ psychological order are articulated in more detail than others. Some states of affairs will be described as the proximate causes of our beliefs in less-​than-​fully-​articulated folk-​ psychological explanations but will not show up as proximate causes within fully articulated explanations. For example, we might say truly that, “The man believed his laundry was done because the dryer went off.” However, we can add more detail to this explanation by further specifying, “The man believed his laundry was done because he heard the dryer go off.” Thus, while the dryer going off might be a proximate cause of the man’s belief in some sense, it is not the most proximate cause of his belief—​there are intermediary causal states by virtue of which the dryer’s going off caused the belief. We can confirm this by noting that when hearing the dryer go off comes apart from the dryer in fact going off, belief tracks with the former. For instance, if the dryer goes off but the man doesn’t hear it, then he does not form the relevant belief. When it comes to directly accessible states, the ones we are looking for are the most proximate causes of our beliefs. These will be the ones that feature as proximate causes within fully articulated folk-​psychological explanations, meaning that they include “all the intervening steps that can be captured at the folk-​psychological level” (Wedgwood 2002, 357). Thus, we reach the following: Direct Accessibility (third pass)—​X is directly accessible to S’s rational belief-​forming systems if and only if X can be the proximate cause of

32  Classical Deontologism one of S’s rationally formed beliefs within the fully articulated folk-​ psychological explanation of that belief. As the dryer example illustrates, the most proximate causes of our beliefs on the folk-​psychological level of explanation are always going to be mental states. In other words, “X” will always make reference to the presence or absence of some mental state M. Finally, for the dispositional account to work, all of the defining characteristics of the normatively relevant mental states need to be ones which can directly affect belief. Otherwise, a mental state M* with the same causal properties as the normatively relevant state M could bring about belief even when M is absent, inhibiting perfect conformance. Furthermore, belief would not be brought about when S has M precisely because S has M (as opposed to M*), preventing compliance. Hence, the mental states that are directly accessible in the relevant sense will be ones characterized exclusively by those properties by virtue of which they are capable of serving as the most proximate causes of our beliefs. Returning to our earlier example, say the man heard the dryer go off on a Wednesday. We can truly say then that, “The man believed his laundry was done because he heard the dryer go off on a Wednesday.” Nevertheless, the mental state hearing the dryer go off on a Wednesday is not directly accessible in the relevant sense because that state is characterized by features that make no difference to whether belief is immediately formed. A sure sign of this is that hearing the dryer on a Wednesday makes no difference to whether hearing the dryer immediately brings about belief or not. If the man heard the same sound on a day other than Wednesday, he would have formed the same belief. So, the state hearing the dryer go off on a Wednesday is not directly accessible to our rational belief-​forming systems.11 Wedgwood would go so far as to say that this mental state is not part of the correct explanation of the belief at all. He writes: [I]‌n any correct explanation there must be a certain sort of proportionality between the explanandum and the explanans. The explanans must be sufficient in the circumstances to produce the explanandum; but it also must not contain any irrelevant elements that could be stripped away without making it any less sufficient to produce the explanandum. (Wedgwood 2002, 363) If Wedgwood is correct, then all the better for my argument. In that case, our third pass at direct accessibility would be equivalent to our final characterization below since none of the mental states featured in our fully articulated folk-​psychological explanations would be characterized in ways that are irrelevant to their ability to bring about belief. Some, however,

Classical Deontologism  33 may argue that states such as hearing the dryer go off on a Wednesday do in fact explain why the man believed his laundry was done. These states are simply partially superfluous explanations in that they contain more than what is strictly required to explain why the belief came about. To accommodate this sort of position, we can give our full characterization of directly accessible states as follows: Direct Accessibility—​X is directly accessible to S’s rational belief-​forming systems if and only if X can be the proximate cause of one of S’s rationally formed beliefs within the fully articulated, entirely non-​superfluous folk-​psychological explanation of that belief. Not all mental states are directly accessible to us in the defined sense. This fact makes it impossible for us to be impeccable with respect to all norms of belief, at least if we are relying on the dispositional account. Return to the norm: believe p if and only if S has M. Humans can guarantee conformance to that norm if and only if (i) M is directly accessible to us, or (ii) M* is directly accessible to us and, necessarily, M* obtains if and only if M obtains. Otherwise, it will be possible that M obtain and for our belief-​forming dispositions not to be triggered, or for those dispositions to be triggered when M does not obtain. In short, the dispositional account can satisfy the conformance condition on impeccability vis-​à-​vis a norm if and only if the normatively relevant conditions are either themselves directly accessible to us, or else necessarily correlate with conditions that are. It is even harder to satisfy the compliance condition on impeccability. For compliance requires that our beliefs be rationally based on the states which make them permissible. In this case, that requires believing that p if and only if S has M and precisely because S does or does not have M. But this can be guaranteed only if M itself is directly accessible to us. Thus, the dispositional account can preserve impeccability with respect to a norm if and only if the normatively relevant conditions are themselves directly accessible. 1.4.2  Why This Works for Internalism

To determine whether we can be impeccable with respect to internalist norms, we must first ask what exactly makes for an internalist norm? As a first pass, we said that internalist norms are those whose normatively relevant conditions are all internal to the mind of the subject. That is, the normatively relevant conditions consist exclusively in the presence or absence of certain mental states. This much internalists agree on. But exactly which mental states are normatively relevant and why? Access internalism is the original form of internalism as endorsed by traditional figures like Locke and Descartes. It says that the normatively

34  Classical Deontologism relevant mental states are those to which the subject enjoys a special kind of reflective access—​namely, one is always in a position to know (or justifiably believe) whether they exist and perhaps also what they do or do not indicate to be true. I propose that, most fundamentally, what makes something normatively relevant for access internalists is that the subject is able to take it into account in belief formation, and the specific manner in which the subject takes it into account is by being reflectively aware of it. Given the challenges to mental transparency cited above, many internalists have moved away from access internalism. Since we are granting those challenges here, we will do the same. The main alternative to access internalism is mentalism. Mentalism says that justification, or how one ought to believe, supervenes on one’s mental states, but denies that we need any special access to those mental states.12 Complicating matters are the existence of factive mental states like knowing or perceiving—​ones characterized in part by their relation to obviously external states of affairs. If these kinds of mental states are included in the supervenience base, then the resulting mentalism would not be internalist in any recognizable sense. Thus, to be an instance of internalism, we must limit the sort of mental states that are included in the supervenience base. Usually, it is restricted to all and only non-​factive mental states. Such non-​ factive states are internal to the subject in a natural enough sense to qualify this as a form of internalism, at least on first blush. Let’s call this form of mentalism “non-​factive mentalism.” One worry about this form of mentalism is that the subject may have no way of registering reflectively or otherwise whether certain non-​factive states exist or not. In allowing such states to count as normatively relevant, however, we contravene what I proposed as the fundamental idea behind internalism as originally conceived: that the normatively relevant conditions are those that the subject can take into account in belief formation. Thus, Michael Bergmann understandably questions whether we ought to consider this a form of internalism at all.13 Worse still, by allowing some normatively relevant conditions to be completely undetectable (even to our automatic processing systems), non-​factive mentalism cannot use the dispositional account to sustain impeccability with respect to justification. That account requires that all normatively relevant conditions be directly accessible to our automatic processing systems. The consequence is that non-​factive mentalism cannot claim for itself the motivations for internalism under discussion in this paper. There is, however, a different form of mentalism inspired by John Pollock that can take advantage of the dispositional account. On this approach, the mental states included in the supervenience base are all and only those that are directly accessible to the subject’s rational belief-​forming systems. I will argue momentarily that directly accessible states are a subset of non-​factive

Classical Deontologism  35 mental states; thus, if non-​factive states are “internal” in a natural sense, directly accessible states are even more so. For while there are non-​factive mental states whose presence or absence can go completely undetected by us, and which have no potential for affecting belief formation, this is not true of directly accessible mental states. We invariably detect their presence or absence and do so in a way that allows us to take them into account in forming beliefs. To be clear, the sense in which we “detect” directly accessible states is not by having actual or potential second-​order awareness of them (though we can become aware of them in this way); rather, we detect them in the sense that our rational belief-​forming dispositions are sensitive to their presence or absence. This allows us to take them into account in the sense that they wield causal influence over which beliefs we form. Let us define “causal internalism” as a form of mentalism which says that justification (or how we ought to believe) supervenes on all and only one’s directly accessible mental states. Using the dispositional account, it is easy to see how impeccability can be sustained with respect to justification if causal internalism is true. What we need for impeccability is that all of the normatively relevant conditions be directly accessible to our automatic processing systems. This is obviously satisfied on causal internalism since the supervenience base is specifically restricted to directly accessible mental states. 1.4.3  Why This Doesn’t Work for Externalism

On first blush, it may seem that the dispositional account also provides externalists with an avenue toward preserving impeccability. Why can’t I be disposed to form only reliable beliefs, for instance, across all possible worlds in which I am similarly disposed? The problem is that, on externalism, the normatively relevant conditions are neither themselves directly accessible (as required for compliance) nor do they necessarily correlate with things that are (as required for conformance). But both of these are required for impeccability. Let’s look more closely at each of these two conditions.14 Let’s start with conformance. Consider an externalist norm that says of some external, non-​directly accessible state of affairs X: believe p if and only if X obtains. To maintain the possibility of impeccability, it would need to be the case that there is some kind of directly accessible state M such that, necessarily for S, states of kind M exists if and only if X obtains. However, the plain fact is that things on the inside do not perfectly track things on the outside. It seems there will always be possible scenarios in which an instance of M is brought about in the absence of X. As an example, a simple version of process reliabilism says that S’s belief is justified if and only if that belief results from a reliable process (e.g., Goldman

36  Classical Deontologism 1979). Clearly, there are no directly accessible mental states that always and only lead to reliably formed beliefs. The same directly accessible state that causes a reliable belief in normal worlds will cause an unreliable belief in evil demon worlds. The problem generalizes to all similar version of externalism because directly accessible states can always be caused in different ways. Skeptics take advantage of this fact in designing skeptical scenarios: they preserve the same kind of directly accessible states that one would have in a normal environment while swapping in abnormal causes. It seems, then, that there are no directly accessible states that necessarily correlate with whatever external conditions the externalist identifies as normatively relevant.15 Thus, we cannot guarantee even conformance to externalist norms. The only kind of externalism that has any chance of satisfying the conformance condition is the sort that Pollock calls “norm externalism” (Pollock 1987, 74). This sort of view adopts internalist rules but then legitimizes those rules by appealing to the external benefits of following them. For example, a version of normal worlds reliabilism says that a belief is justified for S if and only if it results from a belief-​forming process that is reliable in normal worlds (e.g., Goldman 1986). Such a theory does not require that the process producing S’s belief actually be reliable, but only that this process would be reliable in worlds that are “normal” for S. Accordingly, S can ensure justification across all of the relevant worlds (including evil demon worlds) by being disposed so as to believe reliably in normal worlds. This only works, however, if there is a set of rational belief-​ forming dispositions that prove reliable in normal worlds. The problem is that this needn’t always be true. If S’s belief-​forming processes are not reliable even in worlds that are normal for S (say S’s home environment is an evil demon world or, more mundanely, S’s faculties simply evolved to be unreliable), then S will not be able to guarantee the justification of S’s non-​exempt beliefs no matter how carefully S reasons. Regardless, even if externalism could somehow secure conformance, the second and more decisive problem is that we could not comply with externalist norms—​that is, our beliefs would not always be brought about precisely because the relevant external conditions obtained. This is because the conditions identified by externalist theories as normatively relevant are not directly accessible to us. Wedgwood effectively argues for this principle by showing that the proximate causes of belief in fully articulated folk-​psychological explanations are always internal facts about a thinker’s mental states (Wedgwood 2002, §4), but this argument should be unnecessary for the majority of externalists. On most versions of externalism, it is evident that the conditions to which they are appealing are not directly accessible. Obviously, proponents of those theories will dispute whether it is important that normatively relevant conditions be directly accessible,

Classical Deontologism  37 but they will not pretend that facts about, for instance, the reliability of one’s perceptual processes somehow show up as the proximate causes of our beliefs in fully articulated folk-​psychological explanations.16 Such externalists do not dispute, for instance, that evil demon victims would believe the same things as their ordinary counterparts, even if the external conditions to which they appeal were absent.17 Hence, most externalists will concede that their favored external conditions are not directly accessible to our rational belief-​forming systems. The only externalists who might contest our having no direct access to the external are disjunctivists and some in the knowledge-​first camp, although even these should, I think, eventually concede the point. The sort of theorists I have in mind are those who endorse norms such as: believe p (perceptually) if and only if one perceives p or believe p if and only if one knows that p. Perception and knowledge are factive mental states, meaning that to perceive or to know p is to stand in a kind of relation that is possible only if p is true. Intuitively, it might seem that states like perception or knowledge are directly accessible to us. After all, we often say things like, “I believe p because I know q” or, to return to our earlier example, “The man believes his laundry his done because he hears the dryer go off” (hearing being a species of perception). If we look more closely, however, we can see that perception, knowing, and other factive mental states do not meet the conditions on direct accessibility. This is because those states are characterized by features that make no difference to whether belief comes about, and so at best constitute partially superfluous explanations of belief. For example, in the dryer case, it makes no difference whether the man hears the dryer go off or whether it merely appears to him that the dryer is going off (a state phenomenally indistinguishable from hearing the dryer go off)—​he will believe or not believe the same way in both cases. Michael Huemer makes the point using a different example: [T]‌he appearance probabilistically screens off my belief from the external fact. That is, given that I experience exactly the sort of appearance I am now experiencing, the probability of my forming the belief that there is a cat is unaffected by the actual existence or non-​existence of the cat. If I have this sort of appearance caused by a cat, I will believe that there is a cat; but equally, if I have the same appearances when no relevant cat exists, I will believe that there is a cat. (Huemer 2007, 40) Even disjunctivists and knowledge-​firsters concede this, as shown in their admission that evil demon victims will believe in exactly the same ways that their ordinary counterparts do despite lacking the same factive mental

38  Classical Deontologism states. It follows that an entirely non-​ superfluous folk-​ psychological explanation of belief will not include mention of factive states like perception or knowledge, since some of the defining features of these states play no causal role in directly bringing about belief. To be clear, I am not saying, along with Wedgwood, that perception or knowledge do not feature in fully articulated folk-​psychological explanations of belief at all. I am not even denying that perception and knowledge are sometimes the most proximate causes of our beliefs. I am making the more modest point that if and when such states do serve as the most proximate causes of our beliefs, the properties by virtue of which they do so are not those that qualify them as factive states. Hence, such states are not directly accessible in the manner required to take advantage of the dispositional account. To drive the point home, consider the following example from Miriam Schoenfield (2015, 257–​258).18 The Colonial American lookout Paul Revere makes the following plan: The Revere Plan—​Believe that the British are coming by land if he perceives the British coming by land; believe they are coming by sea if he perceives them coming by sea. Let us imagine, however, that the British had taken to using decoys—​ objects that could make it appear as though they were coming by land or sea when they were not. The Colonialists are fretting about the reliability of Revere’s plan in light of such decoys when someone interjects: I don’t see the problem. Revere’s plan is to believe that the British are coming by land if he perceives them to be coming by land, not if it only appears to him that they are coming by land. And the same for the sea. So, where’s the issue? The issue, of course, is that Revere’s plan will in fact result in Revere believing that the British are coming by land or sea whenever it appears to him that they are coming that way (and not only when he genuinely perceives this). Indeed, the issue with the interjector’s question is so obvious to us that those listening would likely take it for an ill-​timed joke. However, if perception could be the proximate cause of belief within a fully articulated, entirely non-​superfluous folk-​psychological explanation, then the interjector’s question wouldn’t be silly at all. There would be no principled reason why Revere couldn’t have a rational disposition that was directly triggered by perception but not by a phenomenally indistinguishable appearance. What we observe, however, is that anyone who is disposed to form the belief that p upon perceiving p will also form the belief that p upon apparently perceiving p. Unless we concede that the same kinds of directly

Classical Deontologism  39 accessible states are prompting belief in both cases (i.e., appearances), then we are utterly at a loss to explain this observation. Even proponents of disjunctivism or knowledge-​first can and should accept that we have no direct access to factive states. Timothy Williamson, for example, concedes that we are “causally insensitive to any difference in evidence” between ordinary cases of knowledge and skeptical scenarios (Williamson 2000, 180). Likewise, Harold Langsam writes of John McDowell: McDowell, although a disjunctivist about experience, is not only committed to saying that perceptual experiences share a categorical property with illusory experiences (the property of being appearances), but he is also committed to saying that it is this shared property that grounds the power of perceptual experiences to produce beliefs. (Langsam 2021, 9792) Even on McDowell’s view, then, factive states are at best partially superfluous explanations of belief. How are these concessions consistent with knowledge-​first and disjunctivism? Because Williamson goes on to insist that “Our causal insensitivity to any difference in evidence between the two cases does not show that there is no difference in evidence between them” (Williamson 2000, 180). In like manner, Langsam defends McDowell’s epistemology by claiming that the inaccessibility of factive states does not prove their normative irrelevance. He writes: I deny … that only properties that ground the experience’s causal power to produce beliefs can be justification-​conferring properties. … [T]‌here is simply no reason why every justification-​conferring property of an experience must also be a property that grounds its causal power to produce beliefs. (Langsam 2021, 9799) Now, a central thesis of this chapter is that the factors relevant for justification are limited to the directly accessible (because only then can we be impeccable with respect to justification). Nevertheless, the lesson to take from Langsam and Williamson is that externalists’ real objection is to the normative claim that only directly accessible states are epistemically relevant vis-​à-​vis justification, not to the causal claim that factive mental states are not directly accessible.19 Because external states of affairs, including the existence of factive mental states, are not directly accessible to us, it follows that we cannot be disposed so as to comply necessarily with externalist norms in the manner required for impeccability. So, if justification is the sort of thing that we

40  Classical Deontologism can guarantee for our non-​exempt beliefs so long as we reason correctly, then internalism about justification must be true. For impeccability is possible only if we limit the normatively relevant conditions to those that are directly accessible to us—​namely, to the internal. 1.5  Causal Internalism In closing, I want to add some final remarks on causal internalism, mostly to address the worry that it is not a genuine form of internalism at all. My position is that while causal internalism is somewhat revisionary, it remains faithful enough to the core features of internalism as traditionally conceived that the label is not misapplied. The importance of our conclusion does not hinge on this debate, of course. We could easily reframe the discussion as a defense of mentalism about justification with no less interest. Whatever we call it, the requirement that justification supervene on directly accessible mental states would rule out every major theory of justification that marches under the banner of externalism (and some that march under the banner of internalism as well)—​and that conclusion is plenty significant on its own. My position, then, is that causal internalism is a form of internalism as that term is generally used. It’s just one which reorients the focus away from reflection and toward causation (specifically, the particular sort of causation involved in rational basing). As Schoenfield aptly puts it: Internalist norms deserve our attention, not because of any special epistemic access that we have to internalist conditions (though perhaps we have that too), but rather because of a special causal role that we should expect internalist conditions to play. (Schoenfield 2015, 254) Through this reorientation, I propose that casual internalism elegantly splits the difference between other forms of internalism, preserving their advantages while escaping their disadvantages. Let’s examine how. Access internalists correctly recognized that for something to be relevant to justification, the subject must be able to take it into account. Where they erred is assuming that this must take the form of being reflectively aware of it. However, in retreating from this assumption, non-​factive mentalism lost the original insight altogether. Causal internalism, on the other hand, preserves it in a suitably moderated way. On causal internalism, mental states are taken into account by having direct causal influence over rational belief formation. This causal influence, keep in mind, is not merely causal. When it comes to rational dispositions, the triggering state serves as a reason as well as a cause.20 And, clearly, if a mental state serves as the

Classical Deontologism  41 rational basis for a belief, then that state is taken into account by the subject in an especially robust sense. Thus, by reconceiving of access in terms of (a special kind of) casual sensitivity rather than reflective awareness, casual internalism captures some of the central motivations for access internalism while avoiding its liabilities. Is there any awareness requirement on causal internalism? Bergmann argues that without an awareness requirement, it’s hard to see casual internalism as genuinely internalist. He suggests that on causal internalism: [It’s possible that] most of our justified beliefs are justified in virtue of internal states we aren’t aware of (or even potentially aware of). Thus, most of our justified beliefs are justified despite the fact that we aren’t aware of even potentially aware of anything at all contributing to their justification. To classify such a view as an internalist view does not seem to fit at all with the way the term “internalism” has been used in the literature. (Bergmann 2006, 70) It’s not entirely clear that conscious awareness should be required for internalism. Even without any awareness requirement, causal internalism rules out all common externalist theories of justification and limits normatively relevant conditions to the states within the agent’s own mind that, on the personal level, bring about the agent’s beliefs. So, if forced to pick a label, “internalism” seems more appropriate than “externalism.”21 Nevertheless, let us assume for the sake of argument that an awareness requirement is crucial to internalism. Plausibly, casual internalism can accommodate this. Here’s how. There are reasons to think that rational belief formation (when a belief is formed on the basis of reasons rather than brutely caused) is always triggered most immediately by what Smithies calls “phenomenally individuated” mental states.22 Smithies elaborates: These “phenomenally individuated” mental states include not only your experiences, which are individuated by their phenomenal character, but also your standing beliefs, desires, and intentions, which are individuated by their dispositions to cause certain kinds of experiences under experiential conditions. At the same time, these exclude your “subdoxastic” mental states, which are individuated by their role in unconscious computational processes, and all your mental states that are externally individuated by their relations to the external world. (Smithies 2019, 25)23 Phenomenally individuated states such as these are ones that we are or could be aware of (on the first-​order level) simply by virtue of having them. We

42  Classical Deontologism can sometimes also become reflectively aware of them (on the second-​order level), but that is something different. BonJour describes this difference by contrasting the “built-​in,” “constitutive awareness” that we have of conscious phenomenal states with “higher-​order awareness” of them (BonJour and Sosa 2003, 4.2–​4.3). Here’s the crux: if phenomenally individuated states are always the proximate causes of rational belief formation within the fully articulated folk-​psychological explanation of our beliefs, then, given casual internalism, we always are or always could be consciously aware of our reasons for belief—​reasons which, good or bad, determine whether our belief is justified. We aren’t always reflectively aware of those reasons, but that isn’t a requirement for genuine internalism, or at least it shouldn’t be. So, are phenomenally individuated states always the most proximate causes of our rationally formed beliefs within folk-​psychological explanations? It seems so. Ordinarily, non-​ basic beliefs are directly caused by other beliefs, and basic beliefs are directly caused by conscious experiences (such as those present in perception, memory, reason, and introspection); and both beliefs and experiences are phenomenally individuated states. What else might serve as the most proximate cause of our beliefs? Neuroscientists might identify certain unconscious brain processes as immediately causing belief, but we are looking for the most proximate causes within folk-​psychological explanations. Such explanations operate on the personal level and so will not include subpersonal brain states. Similarly, Bergmann proposes that something like radio signals might serve as the most proximate cause of our beliefs (Bergmann 2006, 67). The counterexample fails, however, in one of two ways. If the radio signals cause belief by directly rewiring our brains, then there won’t be a folk-​ psychological explanation of the belief at all. There is no folk-​psychological explanation because the belief is not rationally formed. The radio signals bypass our rational belief-​forming dispositions rather than trigger them. Like subpersonal brain states, such radio signals will still count as causing our beliefs, but only within an altogether different order of explanation. On the other hand, if the radio signals do manage to show up in a folk-​ psychological explanation of the belief, it will only be by causing us to have a conscious mental state (like an experience) that then serves as the most proximate cause of the belief. Simply saying “S was struck by radio signals” does nothing to make intelligible or unsurprising on the personal level why S came to believe p unless something like the following is also true: “S was struck by radio signals which made it appear to S that p” (Wedgwood 2002, 359–​360). We can generalize the lesson here: nothing can serve as the most proximate cause of belief within a folk-​psychological explanation unless it is operating on the personal level, and nothing

Classical Deontologism  43 operates on the personal level unless it is a conscious mental state with phenomenal properties. What’s more, it seems that nothing other than these phenomenal properties is relevant to belief formation. Evil demon cases prove the point. It goes unchallenged that evil demon victims and their ordinary counterparts will share all and only the same rationally formed beliefs.24 This is because the victims and their counterparts are stipulated to be phenomenal duplicates—​ what it feels like to be one is exactly the same as what it feels like to be the other—​and the unstated assumption is that what we believe supervenes on how things feel to us. We are on solid ground, then, in assuming that the most proximate causes of our beliefs on the folk-​psychological level will always be phenomenally individuated mental states. That is, only phenomenally individuated mental states are directly accessible to us. There may also be another, in-​ principle reason why phenomenally individuated states are always the most proximate causes of our rationally formed beliefs. When it comes to rational belief-​forming dispositions, the triggering states must cause those beliefs in a special sort of way—​one that allows the triggering states to serve as reasons for the resulting beliefs as well as causes. What could this be? A plausible proposal is that a reason causes belief by indicating to the subject that some propositional content is true. There are seemingly only two ways in which this can occur. Either one is apparently aware of a support relation between some other proposition, which is taken to be true, and this new content. Or else one is apparently aware of the truth of the proposition itself. In both cases, the most proximate causes of the belief are phenomenal states. If our beliefs were not caused by phenomenal states, then they would not be rationally formed beliefs at all. They would result not from rational belief-​forming dispositions but something more like belief reflexes. Furthermore, the phenomenal properties of such states are sufficient to immediately cause belief in us. Any other properties of these states are superfluous to explaining why belief was rationally formed. Thus, there is a strong case that the proximate causes of our rationally formed beliefs within a fully articulated, entirely non-​superfluous folk-​psychological explanations are always phenomenally individuated states. This presentation of this in-​principle argument is perhaps too quick to be fully convincing. However, the arguments made throughout the rest of this inquiry will shore up its essential points. In particular, I will argue that the only states capable of serving as the rational bases for our beliefs are seemings (or other beliefs that are themselves based on seemings), and that what makes them and them alone fit to serve in such a role is precisely their unique phenomenal character—​namely, that their propositional content feels true. In other words, it is only and solely by virtue of something

44  Classical Deontologism feeling true that anything can be indicated to be true. If I succeed in this, then the in-​principle argument above will be proven correct. What all this suggests is that, on causal internalism, justification ends up supervening on our phenomenally individuated mental states, or some subset of them. If this is correct, then causal internalism implies some species of what Declan Smithies calls “phenomenalism” (Smithies 2019). What makes phenomenally individuated states so important, however, is not that they are reflectively transparent to us (as Smithies would have it), but that they are the only ones directly accessible to our rational belief-​forming systems.25 It is not their transparency that distinguishes them, but their causal role. More precisely, it is the fact that phenomenally individuated states can play the unique causal role assigned to reasons: causing belief, yes, but doing so by indicating the truth of its content. Plausibly, such indication is only possible because of our first-​order awareness (or potential awareness) of those states. The upshot is that, on causal internalism, it is plausible that all of the contributors to justification are ones we are either aware of or potentially aware of on the first-​order level. If so, then causal internalism manages once again to capture a central element of access internalism without taking on its excesses. I’ve now defused two most serious problems for classical deontologism. We can be genuinely blameworthy for all of our non-​exempt, unjustified beliefs without relying on the assumptions of doxastic voluntarism or reflective transparency if only we limit the normatively relevant conditions to the directly accessible. To be fair, the resulting view does depart from Locke and Descartes’s original vision of justification on several points. If we wish to call my position “neo-​classical deontologism,” I would not object. I would insist, however, that the view I have argued for here represents a kind if minimal departure from that original vision—​one that remains as faithful to its most central features while jettisoning its most problematic assumptions—​and so deserves the mantle of “justification” more so than its externalist rivals (or even alternative forms of mentalism). In any case, the rest of our inquiry will move forward on the assumption of (neo-​)classical deontologism. The question now becomes how we are to attain justification so understood. From whence does it arise? Notes 1 These same duties govern all of our doxastic attitudes, be they beliefs or disbeliefs or any other doxastic state. For convenience, however, I will focus on beliefs throughout this chapter.

Classical Deontologism  45 2 DePoe (2021) defends a classical foundationalist approach to justification as preserving both an intentional and effective connection to truth. I address this claim in Chapter 9, §9.3.1. 3 A third option is to treat these two aspects of justification as equally fundamental though logically independent. On this approach, we should simply stipulate that justified beliefs have both sorts of connections. Alston’s “internalist externalism” is an example of this (Alston 1988). Reid, I believe, takes this sort of stance with respect to warrant. 4 Some contemporary proponents of this account (or key elements of it) include BonJour (1985), Chisholm (1977), Ginet (1975), Steup (1988), and Smithies (2019). See also Srinivasan (2015), §3, for discussion. 5 On the definition of reflective transparency, see Williamson (2000) and Berker (2008). 6 See Schoenfield (2015) for a related attempt to motivate internalism despite threats to luminosity. There are significant differences in how we go about formulating and defending our ideas, nevertheless there is, I think, a common insight motivating both of our responses. 7 What is essential here is that the beliefs result from my own rational dispositions, not that they be based on reasons per se. For instance, if one’s rational dispositions were to produce a belief that is not based on reasons, then such a belief would still be under one’s intrinsic control in the relevant sense. Though I tend to think this is impossible. Plausibly what makes a disposition rational (as opposed to a mere reflex) just is that it is responsive to reasons (as opposed to having mere causes), in which case no beliefs formed by our rational dispositions are entirely baseless. I say more about this in §1.5. 8 If you think that perceptual beliefs are inferred from how things appear to us, then consider how it is we came to believe that things appear thus-​and-​so. At some point, our beliefs must be caused apart from higher-​order beliefs, or else we would be faced with forming an infinite regress of beliefs prior to forming any belief and belief formation would be impossible. 9 This does not yet commit us to the position that those appearances are good reasons for belief, or ones adequate to justify belief in the absence of defeaters. 10 With respect to the norms of justification, a belief that conforms to those norms is said to have propositional justification for its content, whereas a belief that complies with those norms is said to be doxastically justified. 11 The fact that the man heard the dryer go off on a Wednesday could make a difference to what he believes if, for instance, he knows that the dryer goes off sporadically (even when the laundry isn’t finished) on days other than Wednesday. However, the man would have to be reflectively aware of the fact that he heard the dryer on a Wednesday in order for it factor into belief-​formation, in which case its being Wednesday doesn’t make any difference to whether the sound of the dryer immediately brings about belief or not. Another way to put this is that the state hearing the dryer go off on a Wednesday is, in that particular case, not the most proximate cause of belief, but only causes belief by virtue of intermediary states such as believing that it is Wednesday. 12 Conee and Feldman (2004), Chapter 3.

46  Classical Deontologism 3 Bergmann (2006), 55–​56. 1 14 If we define “externalism” as “not-​causal internalism,” then it follows by definition that externalist norms appeal to conditions that are not directly accessible to our automatic processing systems. That being said, I should showcase that the sorts of theories usually thought of as exemplars of externalism are still “externalist” in this particular sense. 15 It is of course possible to gerrymander an external state of affairs that necessarily correlates with some internal state (e.g., the state of affairs S has internal state M and God exists is an external state of affairs that, if God necessarily exists, will perfectly track the internal state M), but these are trivial. The crucial thing is that none of the external conditions proposed by externalist theories of justification, nor anything in their vicinity, are of this sort. 16 One of these might feature in a folk-​psychological explanation of belief if, say, S1 believes that trusting in the testimony of S2 is a reliable process, and so infers the belief that p from the fact that S2 testified that p. But clearly the reliability of that process is not the most proximate cause of belief in that situation. 17 If the content of their beliefs is externally determined, then perhaps the subjects do not form the exact same beliefs. Nevertheless, it is clear that one subject will form the same kind of belief if and only if the other does. The only difference is that the content of the one belief will be “filled in” differently than the content of the other. Thus, I will set this complication aside. 18 I’ve amended the example to better fit our purposes. 19 Furthermore, nothing said above about the causes of our beliefs implies that we can have the same token state in a case of perception as we do in a case of illusion or hallucination. Say that, in the case of perception, S has a perceptual state P and, in the case of hallucination, S has an entirely different token state H. There is no token state common to both of them. The argument made above is simply that P would be the most proximate cause of a belief in S if and only if H would be as well (were S to have that state instead), and this proves that P only brings about belief by virtue of properties that H also possesses. Now, if we wish, we could define “appearances” as a type of mental state characterized exclusively by how it feels to be in a state like P or H, in which case we could say that S has an appearance in both the perceptual case and the hallucinatory case; and those appearances would be the proximate causes of S’s belief in both. But it would not be the same appearance in both cases, only the same type of appearance. And it is only the former claim that disjunctivists should be concerned to deny. Thus, once again, we see that even disjunctivists should accept the causal claims at issue here. 20 Being the proximate cause of a belief is not sufficient for being the base of that belief, but the additional conditions (whatever they are) are ones that will be ordinarily satisfied when a directly accessible mental state triggers our rational belief-​forming dispositions. 21 Something like this is perhaps by Bergmann does not see internalism and externalism as exhaustive, with mentalism (including what we are calling “causal internalism”) residing somewhere in between the two.

Classical Deontologism  47 22 The claim under consideration here adds to Wedgwood (2002) that the internal facts about our mental states that determine belief formation always have to do with their phenomenal properties. 23 On phenomenally individuated states, see also Smithies (2014). 24 Once again, issues of content externalism aside. 25 This cuts against Smithies’ claim that phenomenalism and a phenomenal conception of evidence is best motivated by accessibilism—​the position that justification is reflectively transparent to us. This claim features centrally in the argument of Smithies (2019), and especially in his critique of phenomenal conservatism in ­chapter 12. Essentially, Smithies proposes that internalists double-​ down on reflective transparency, whereas I am trying to show how internalism can be motivated apart from reflective transparency.

References Alston, William P. 1988. “The Deontological Conception of Epistemic Justification.” Philosophical Perspectives 2: 257–​299. Bergmann, Michael. 2006. Justification Without Awareness: A Defense of Epistemic Externalism. Oxford: Oxford University Press. Berker, Selim. 2008. “Luminosity Regained.” Philosophers’ Imprint 8: 1–​22. BonJour, Laurence. 1978. “Can Empirical Knowledge Have a Foundation?” American Philosophical Quarterly 15(1): 1–​14. —​—​—​. 1985. The Structure of Empirical Knowledge. Cambridge, MA: Harvard University Press. BonJour, Laurence, and Ernest Sosa. 2003. Epistemic Justification: Internalism Vs. Externalism, Foundations Vs. Virtues. Hoboken, NJ: Wiley-​Blackwell. Boyle, Matthew. 2009. “Active Belief.” Canadian Journal of Philosophy 39 (S1): 119–​147. Carruthers, Peter. 2011. The Opacity of Mind: An Integrative Theory of Self-​ Knowledge. Oxford: Oxford University Press. Chisholm, Roderick M. 1977. Theory of Knowledge. 2nd edition. Englewood Cliffs, NJ: Prentice-​Hall. Conee, Earl, and Richard Feldman. 2004. Evidentialism: Essays in Epistemology. Vol. 48. Oxford: Oxford University Press. DePoe, John. 2021. “Justification by Acquaintance.” Synthese 199: 7555–​7573. Descartes, René. [1641] 2006. Meditations, Objections, and Replies. Translated by Roger Ariew and Donald A. Cress. Indianapolis, IN: Hackett Publishing. Feldman, Richard. 2000. “The Ethics of Belief.” Philosophy and Phenomenological Research 60 (3): 667–​695. —​—​—​. 2008. “Modest Deontologism in Epistemology.” Synthese 161 (3): 339–​355. Ginet, Carl. 1975. Knowledge, Perception and Memory. Philosophical Studies Series. Heidelberg, DE: Springer Netherlands. Goldman, Alvin I. 1979. “What Is Justified Belief?” In Justification and Knowledge, edited by George Pappas, 1–​23. Dordrecht, NL: D. Reidel. —​—​—​. 1986. Epistemology and Cognition. Cambridge, MA: Harvard University Press.

48  Classical Deontologism —​—​—​. 1999. “Internalism Exposed.” Journal of Philosophy 96 (6): 271–​293. Gopnik, Alison. 1993. “How We Know Our Minds: The Illusion of First-​Person Knowledge of Intentionality.” Behavioral and Brain Sciences 16 (1): 1–​14. Hieronymi, Pamela. 2008. “Responsibility for Believing.” Synthese 161 (3): 357–​373. Huemer, Michael. 2007. “Compassionate Phenomenal Conservatism.” Philosophy and Phenomenological Research 74 (1): 30–​55. Kvanvig, Jonathan L. 2014. Rationality and Reflection: How to Think About What to Think. Oxford: Oxford University Press. Langsam, Harold. 2021. “Mcdowell’s Infallibilism and the Nature of Knowledge.” Synthese 198 (10): 9787–​9801. Locke, John. [1689] 1975. An Essay Concerning Human Understanding. Edited by P. H. Nidditch. Oxford: Clarendon Press. McAllister, Blake. 2016. “Re-​Evaluating Reid’s Response to Skepticism.” Journal of Scottish Philosophy 14 (3): 317–​339. Owens, David. 2000. Reason Without Freedom: The Problem of Epistemic Normativity. New York: Routledge. Plantinga, Alvin. 1993. Warrant: The Current Debate. Oxford, UK: Oxford University Press. Pollock, John. 1986. Contemporary Theories of Knowledge. 1st Edition. London: Hutchinson. —​—​—​. 1987. “Epistemic Norms.” Synthese 71 (1): 61–​95. Ryan, Sharon. 2003. “Doxastic Compatibilism and the Ethics of Belief.” Philosophical Studies 114 (1–​2): 47–​79. Schoenfield, Miriam. 2015. “Internalism Without Luminosity.” Philosophical Issues 25 (1): 252–​272. Schwitzgebel, Eric. 2006. “The Unreliability of Naive Introspection.” Philosophical Review 117 (2): 245–​273. —​—​—​. 2011. Perplexities of Consciousness. Denver, CO: Bradford. Smith, Angela M. 2005. “Responsibility for Attitudes: Activity and Passivity in Mental Life.” Ethics 115 (2): 236–​271. —​—​—​. 2008. “Control, Responsibility, and Moral Assessment.” Philosophical Studies 138 (3): 367–​392. Smithies, Declan. 2014. “The Phenomenal Basis of Epistemic Justification.” In New Waves in Philosophy of Mind, edited by Jesper Kallestrup and Mark Sprevak, 98–​124. London: Palgrave Macmillan. —​—​—​. 2019. The Epistemic Role of Consciousness. New York: Oxford University Press. Srinivasan, Amia. 2015. “Normativity Without Cartesian Privilege.” Philosophical Issues 25 (1): 273–​299. Steup, Matthias. 1988. “The Deontic Conception of Epistemic Justification.” Philosophical Studies 53 (1): 65–​84. —​—​—​. 2000. “Doxastic Voluntarism and Epistemic Deontology.” Acta Analytica 15 (1): 25–​56. —​—​—​. 2008. “Doxastic Freedom.” Synthese 161 (3): 375–​392.

Classical Deontologism  49 Watson, Gary. 1996. “Two Faces of Responsibility.” Philosophical Topics 24 (2): 227–​248. Wedgwood, Ralph. 2002. “Internalism Explained.” Philosophy and Phenomenological Research 65 (2): 349–​369. Williamson, Timothy. 2000. Knowledge and Its Limits. Oxford: Oxford University Press. Wolf, Susan. 1990. Freedom Within Reason. Oxford: Oxford University Press. Wolterstorff, Nicholas. 2001. Thomas Reid and the Story of Epistemology. Cambridge: Cambridge University Press.

2 Immediate Justification and Seemings Foundationalism

In this chapter I introduce the overarching thesis of this book—​a position I call “seemings foundationalism.” True to its name, seemings foundationalism says that the foundations of justification consist in a kind of mental state I have presented under the name “seemings” or “appearances.” These states have propositional content which they present as being true or false. Seemings foundationalism says that justification in general is made possible by the immediate justification of foundational doxastic attitudes; these foundational attitudes are immediately justified by foundational evidence; and all and only seemings count as foundational evidence. Hence, justification is ultimately about forming beliefs and other doxastic attitudes in accordance with what seems to be the case. To properly understand this position, we need to have an adequate understanding not only of seemings, but also of foundational evidence, immediate justification, and their interrelations. And to understand that, we better first have a sense of what evidence and justification are more generally. I have already drawn some substantial conclusions about justification in Chapter 1. In §2.1 of this chapter, I build on these conclusions to establish a basic framework for thinking about justification and, in turn, immediate justification and foundational evidence. With this groundwork in place, I can then explain seemings foundationalism in §2.2 and §2.3. 2.1  Justification and Evidence Here we say a bit more about what it means for a doxastic attitude such as belief to be justified. Note the words “a bit” in the previous sentence. The goal is not to give a complete theory of justification but simply to provide a basic framework for thinking about justification to be used throughout the rest of our inquiry. Obviously, the conclusions about the nature of justification reached in Chapter 1 are already quite contentious, but those who are thinking of justification in way described in that first chapter

DOI: 10.4324/9781003281108-3

Immediate Justification and Seemings Foundationalism  51 (particularly, as an internal status pertaining to how one should believe in the first-​person sense) will, I think, generally also end up thinking of justification in roughly the way I describe here. Up front I should note that beliefs are not the only kind of doxastic attitude that can be justified or not. The attitudes of disbelief and withholding assent can also be evaluated in this way. Venturing beyond the three traditional “square-​states,” there are also more fine-​grained doxastic attitudes like levels of confidence or (what I will treat as equivalent) credences. These too can be justified or unjustified. I do not wish to get into controversies over whether credences can be reduced to beliefs (e.g., Moon and Jackson 2020) or beliefs to credences (e.g., Dodd 2017); following Elizabeth Jackson, I will simply assume that there are both kinds of attitudes that feature into our epistemic lives in different ways (Jackson 2022). Despite this plurality of doxastic attitudes, one will often find me talking solely about the justification of beliefs. This is simply a matter of convenience, and the reader is welcome to read those statements as applying mutatis mutandis to other doxastic attitudes. In this initial chapter, I will try to make some of those applications explicit before focusing almost exclusively on beliefs in later chapters. I argued in Chapter 1 that justification is a deontological notion. That is, justified attitudes are those that do not run afoul of our epistemic duties. Notice that justified attitudes are not required by our epistemic duties but allowed by them. Plausibly, there are no attitudes one is obligated to take toward a proposition (Nelson 2010). Simply not forming an attitude at all (which is to be distinguished from adopting the attitude of “withholding assent”) is seemingly always permissible. Notwithstanding, there may be some attitudes that one is obligated to take if one forms an attitude toward that proposition. Regardless, the justified attitudes are just the permissible ones. This deontologism is one of the more controversial aspects of my framework. Though I argued for deontologism in the previous chapter—​and I stand by that argument—​it is important to note that deontologism is to some degree dispensable within the overall inquiry. That is to say, many (perhaps even all) of my central arguments can be reframed in non-​ deontological terms with only marginally less plausibility. For instance, instead of asking whether belief is objectively permitted in certain conditions, one could ask whether belief is an objectively fitting response to those conditions (Conee and Feldman 2004); and most of the arguments will retain their force under this alternative framing. Nevertheless, I have given reason to think that deontologism is true, and the strongest case for seemings foundationalism can be made under that assumption. Thus, I will continue to proceed under the deontological framework. Let’s take a closer look at these epistemic duties. An initial clarification is that these duties are distinctively epistemic, not practical. Practical duties

52  Immediate Justification and Seemings Foundationalism govern the use of the will and concern the relation of that will toward goodness. Whereas epistemic duties govern the use of the intellect and concern its relation toward truth. Thus, whether some doxastic attitude satisfies one’s epistemic duties will be a purely epistemic evaluation rather than a practical or moral one. It’s not about whether one has taken the right practical actions leading up to those attitudes (even if those actions have some bearing down the line on what one believes), but whether the attitudes produced by the intellect meet some purely epistemic standard. Relatedly, the particular duties I am interested in concern which doxastic attitudes to form at a particular time, not with how to order our investigations over time. That is to say that my focus is on synchronic justification rather than diachronic justification. I will be making the assumption that we have these sorts of synchronic, distinctively epistemic duties. (I have already addressed the most pressing reasons for doubting them in Chapter 1.) This means that, for every belief or other doxastic attitude, there is some objective fact as to whether the formation or maintenance of that attitude flouts our epistemic duties or not. Let’s explore the content of these duties and try to get a better sense of what kinds of doxastic attitudes do and do not violate them. The standard view is that our epistemic duties pertain to the proper pursuit of the epistemic goal. The epistemic goal (or at least the one that concerns us vis-​à-​vis justification) has two dimensions (see, e.g., Foley 1993). The first is to say of what is that it is and what is not that it is not (see Aristotle’s Metaphysics 1011b25). With respect to beliefs, this amounts to believing p if p is true. With respect to disbelief, this amounts to disbelieving p if p is false. With respect to credences, this amounts to placing a positive credence in p if p is true (the higher the better), and placing a negative credence in p if p is false (the lower the better). For convenience, I will refer to this first half of the epistemic goal as “attaining truth.” The second half of the epistemic goal is to not say of what is not that it is and what is that it is not. With respect to beliefs and disbeliefs, this amounts to believing p only if p is true and disbelieving p only if p is false. With respect to credences, this amounts to placing a positive credence in p only if p is true, and placing a negative credence in p only if p is false (where mistaken credences are worse the higher or lower they are). I will call this second half of the epistemic goal “avoiding falsehood.” These two goals taper each other. Using beliefs as an example, if attaining truth was our only aim, then then we could reach that goal simply by believing all propositions. Fulfilling our epistemic duties would be clear-​cut. Alas, believing indiscriminately would score very poorly with respect to avoiding falsehood. Coming from the other direction, if avoiding falsehood was our only aim, then would could achieve that goal simply by not believing any propositions—​just withhold assent toward everything.

Immediate Justification and Seemings Foundationalism  53 But again, one would make no progress toward securing truth. Thus, our epistemic duty is to form a combination of doxastic attitudes that properly balances the pursuit of both these competing goals.1 I find it helpful to think of this as something like a betting scenario. In believing p, one is placing a bet on the truth of p. In disbelieving p, one is placing a bet on the falsehood of p. In withholding assent toward p, one is not making a bet all, thereby removing any risk of loss or opportunity for reward. The ultimate goal is to make the most amount of money possible by placing the correct bets on every proposition or, at least, every proposition we form an attitude toward. Given our limited epistemic positions, certain bets are reckless for us. For instance, it is generally conceded that believing p when one has no indication that p is true is not a permissible betting strategy and violates one’s epistemic duty by improperly privileging the securing of truth over the avoidance of falsehood. In that scenario, withholding assent is the only sensible way of pursuing both aspects of the dual-​natured epistemic goal. This all translates to credences as well, with the addition that the higher or lower the credences (essentially, its absolute value from .5) the larger the bet. In a moment, I’ll say more about the conditions that must be in place for one to permissibly form some doxastic attitude toward p, but I don’t want to be too specific on this point. A substantive answer to this question—​ specifically with respect to basic attitudes—​is precisely what I intended to argue for throughout the rest of this inquiry. I don’t want to build my proposed answer into the framework itself. With that in mind, I think it neutral enough to say that permissibly believing or disbelieving p requires that one have sufficient indication of p’s truth or falsity, respectively; and likewise for positive or negative credences. This is typically described as having adequate evidence for or against the proposition in question, evidence being that which indicates the truth or falsity of something. Epistemic conservatism might make an exception for certain basic doxastic attitudes, but this is the exception to the rule even if so (and I argue against making such an exception in Chapter 7). Whether one has, on-​balance, adequate evidence is a function of one’s total evidence, of all the evidence one possesses. A particular mental state is, for subject S, evidence for p if and only if it aims toward making permissible belief that p or a positive credence in p. A state is evidence against p for S if and only if it aims toward making permissible disbelief that p or a negative credence in p. Whether a state is so aimed is an objective matter determined by necessary, logical relations existing between that state and the state it justifies, or so I will assume here. Note that particular mental states may aim toward making a certain doxastic attitude permissible even if they do not achieve that aim. Gravity may pull down on a rocket (and so aim toward grounding it) even while its thrusters overcome this force,

54  Immediate Justification and Seemings Foundationalism propelling the rocket upward. Just so, particular mental states may “push or pull” toward making permissible belief even while, on the whole, the evidence permits only disbelief.2 This occurs, for instance, when the evidence for p is opposed by other, stronger evidence against it. When one has adequate evidence for believing p, this amounts to propositional justification for believing p (and likewise for other attitudes). We refer to propositional justification when we say that “S has justification for believing,” as contrasted with “S has a justified belief.” Forming a justified belief requires more than having adequate evidence; one’s belief must also be properly based on that evidence. I won’t be taking any stand on the nature of the basing relation except to say that it is a causal one. That is, a belief is not based on a mental state unless that state causes the belief to come about. Not all causes are bases, to be sure, but all bases are causes. Thus, a justified belief that p needs to be caused by adequate evidence that p. What’s more, the adequate evidence must cause the belief in the right sort of way. For instance, a belief would not be properly based if it were caused by the epistemically irrelevant properties of a token mental state that, by virtue of other properties, counts as adequate evidence. A belief formed in this manner would conform to the norms of justification, but it would not comply with it. We must add, then, that the evidence needs to cause the belief by virtue of the properties that make it adequate evidence. In sum, a justified belief is based on adequate evidence only if that belief is caused by adequate evidence precisely because it is adequate evidence. When these conditions obtain, one has doxastic justification for that belief. So far so good. As soon as we get into the details, however, we find ourselves embroiled in controversy. What sort of mental states provide evidence for p and by virtue of which properties? What kind of evidence is adequate for belief or disbelief? These are the questions I will propose partial answers to in §2.2 and argue for in Chapters 6–​9. One controversial conclusion already established in Chapter 1 is that the states which serve as evidence for our beliefs are all internal to us in the sense that they are directly accessible to our automatic belief-​forming systems. Directly accessible states are characterized exclusively by properties to which we possess unreflective causal sensitivity—​ones that can affect whether belief is immediately brought about. Accordingly, whether a directly accessible state is evidence for something (and how strong) is entirely a matter of these “directly accessible properties.” It won’t be the case, for instance, that some mental state counts as evidence for believing p because it constitutes an objectively reliable basis for believing p, since being an objective reliable basis for belief is not a property that the agent’s belief-​ forming systems can directly detect and respond to. Rather, the properties that matter must be ones that the mental states “wear on their faces,” so to speak, such as their phenomenal properties.

Immediate Justification and Seemings Foundationalism  55 Another controversial aspect of my position is that, on the present framework, our evidence at least sometimes consists in mental states such as experiences or beliefs. Kevin McCain calls this “psychologism” about evidence (McCain 2014, 10). This is opposed to the idea that our evidence consists in something non-​psychological like publicly accessible states of affairs (like fingerprints at a crime scene) or propositions. The former is ruled out by causal internalism, since publicly accessible states of affairs are not directly accessible in the required sense. The idea of propositions as evidence, however, is ruled out by the requirement that adequate evidence causes our doxastically justified beliefs. Propositions don’t cause our beliefs at all. Mental states with propositional content do, but that’s not quite the same thing. You might think there’s an easy workaround: can’t propositionalism just require instead that justified beliefs be caused by mental states whose propositional content is adequate evidence for the content of the belief? This might work in the case of inference, but it runs into problems with respect to immediate justification. As I will propose below, certain experiences with propositional content p might serve as evidence for belief that p, but it is not the case that the content of the former—​namely, p—​is evidence for the content of the latter—​also p.3 As it pertains to the foundations of justification, at least, our foundational evidence must consist in mental states. Of course, the debate surrounding whether to identify mental states as evidence is a large one that I cannot settle here. I will simply note that the most prominent arguments against psychologism are, by my lights, inadequate.4 Now that we have a basic understanding of justification, let us turn our attention to immediate justification specifically. Immediate justification is a special case in which the justification of some doxastic attitude (be it a belief, disbelief, or some positive or negative credence) does not depend on having evidence or justification for any other doxastic attitudes. An initial worry with this characterization is that all propositional attitudes may prove dependent on a framework of background beliefs in at least the following sense: one is only able to possess and deploy the concepts by which one entertains propositional content insofar as one has a broader conceptual framework already in place—​a framework that presumably requires an extended set of doxastic attitudes such as belief. If this is true, does it imply right off the bat that immediate justification is impossible? Not as I intend it. Compare this to the situation of a priori justification. A belief is justified a priori so long as the evidence that justifies it does not derive from sensory experience, even if some of the concepts that feature in it are acquired through sensory experience. We might say that the justifying conditions are independent of sensory experience, even though some of the broader enabling conditions are not. Similarly, immediate justification requires only that the justifying conditions of belief do not consist

56  Immediate Justification and Seemings Foundationalism in having evidence or justification for other beliefs or attitudes, even if entertaining that propositional content in the first place requires a having framework of background beliefs more broadly. Foundational evidence can in turn be characterized as evidence whose ability to justify is likewise independent of evidence or justification for anything else. As such, foundational evidence aims toward immediate justification. We must remember, of course, that evidence can aim toward justification but not achieve it. For example, it can be true that one has foundational evidence for believing p, yet one is not immediately justified in believing p, because there is other evidence which diminishes the on-​ balance level of evidential support for p to below the threshold adequate for belief. There are at least three different ways in which such diminishment can happen. First, if we gain evidence against p, then that can counter the evidential support provided by the foundational evidence for p. This counter evidence is called a “rebutting defeater.” Second, we can gain evidence that disproves or introduces serious doubts about the reliability of the foundational evidence for p. This is called an “undermining defeater.” The third situation occurs when new evidence swamps or supersedes the foundational evidence, and subsequently reduces the total level of evidential support for p. To give an example, imagine we have foundational evidence that is, in and of itself, adequate for justifying belief that p; however, we later learn with certainty that the statistical probability of p is .65 (not high enough to justify belief). The foundational evidence isn’t countered or undermined so much as supplanted. If we know that the statistical probability of p is .65, then that’s all we need to know—​the foundational evidence is boxed out of the discussion. I will call this kind of defeater a “supplanting defeater.” Are defeaters the only way of blocking immediate justification for p when one has foundational evidence for p? That depends on how we characterize defeaters. Take a case in which the prior probability of p (the probability of p given our background evidence) is low.5 For instance, if p is the roll of this fair die will come up 1, then (assuming we roll the die) the prior probability of p will be 1/​6. (Indeed, this is the pure prior probability of p, meaning that it is the probability of p given only logical truths as background evidence.) In this case, having foundational evidence for p does not guarantee that we are justified in believing p or even placing a credence in p of over .5. Our background evidence—​in this case, evidence about the nature of a fair die—​prevents the foundational evidence from justifying these positive doxastic attitudes toward p. We can construe this as a form of defeat if we like. After all, our level of support for p is diminished from where it would otherwise be by the presence of this background evidence. (And this remains true even if that background evidence consists

Immediate Justification and Seemings Foundationalism  57 merely in logical truths.) Thus, it seems fair to call this diminishment a form of defeat, broadly construed. Some, however, wish to draw a distinction between the way in which low priors prevent justification and the way in which defeaters do. For convenience, I will use the term “defeat” broadly to include situations in which having a (comparatively) low prior probability for p reduces our level of support for p from where it would have been given the foundational evidence alone. Nothing hinges on this usage, however. The central point is just that having low priors is another situation in which having foundational evidence for p can fail to justify believing or placing a positive credence in p. What does foundational evidence guarantee? About all, we can say for sure is that it will immediately justify some positive doxastic attitude in the absence of all other evidence, including any and all background evidence. Following Ted Poston, when we have absolutely no evidence either for or against a proposition, we can call this being in a state of “empty symmetrical evidence.” To be clear, the state of empty symmetrical evidence excludes all evidence for or against a proposition, including any evidence provided by logical truths. Poston writes: The state of empty symmetrical evidence is special. It is one in which there is not evidence relevant for determining a unique prior probability. Take the proposition that Johnson has won the lottery. One has no direct evidence that this is true and no direct evidence that this is false. One hasn’t been told that Johnson won or that someone else did. Yet, one does possess enough evidence to determine that it is unlikely that Johnson won the lottery. This isn’t a case of empty symmetrical evidence. (Poston 2014, 24) In a true case of empty symmetrical evidence, we have no rational basis at all for taking any sort of attitude toward the proposition at hand. If we introduce foundational evidence for p into this special situation, such that this foundational evidence is truly the only thing we have to go on, then that foundational evidence will justify us in forming some positive doxastic attitude toward p. Foundational evidence immediately justifies in the state of empty symmetrical evidence. The specific kind of doxastic attitude it justifies will depend on the foundational evidence at hand. Different kinds of foundational evidence come with different justificatory potentials. For instance, it may be that a particular instance of foundational evidence for p is insufficient to justify full belief in p even in the state of empty symmetrical evidence. It may be sufficient only to justify some positive credence toward p (of a certain strength but no more). About all, we can say of foundational evidence in general that

58  Immediate Justification and Seemings Foundationalism it is sufficient to justify some positive doxastic attitude (broadly construed) toward the proposition in question, though we have to look at the specific case to see which attitude this is. As soon as we step outside of the state of empty symmetrical evidence, things get more complicated. When other relevant evidential considerations are brought into the picture (including reflections on one’s foundational evidence itself), we now need to find the on-​balance level of evidential support for the proposition at hand. To predict how things will turn out here, one needs a more detailed and comprehensive theory of justification—​one that specifies how evidential considerations are weighed against one another to reach an on-​balance level of evidential support. These theories are out there, but they will not be the topic of this inquiry. My focus is specifically on the foundations of justification. The questions I’m asking are ones such as: which mental states serve as foundational evidence? Which properties allow them to fulfill the role? Does all justification ultimately arise from foundational evidence or are there other sources? I will inevitably have to take some stances on how my theory affects justification more broadly (see especially Chapter 3), but these will not be my primary focus. We now have a sufficient framework in place for thinking about justification and evidence—​as well as the special cases of immediate justification and foundational evidence—​to introduce the theory I wish to defend. 2.2  Seemings Foundationalism Seemings foundationalism is about the ultimate origins of justification. It consists in four theses. The first is a version of foundationalism: Foundationalism (FN)—​S has non-​immediate justification for taking some doxastic attitude toward p only if (and because) S has immediate justification for taking some doxastic attitude toward q. At the root of FN is the idea that non-​immediate justification derives from immediate justification. Having the former depends on having the latter. Hence, nothing is justified unless something is immediately justified. I defend FN in Chapter 5. The second thesis of seemings foundationalism is the application of evidentialism to the foundations of justification. It says that nothing is immediately justified unless it is supported by foundational evidence. Foundations Evidentialism (EV)—​S has immediate justification for taking some doxastic attitude toward p only if (and because) that attitude is permitted by S’s total evidence for and against p.

Immediate Justification and Seemings Foundationalism  59 While evidentialism is widely accepted within internalism, there are some views—​particularly epistemic conservatism—​which exempt immediately justified beliefs from the general need for evidence. EV rejects these views, extending the need for evidence all the way down to the foundations. I defend EV in Chapter 7. Together, these first two theses imply that all justification for our beliefs ultimately derives from foundational evidence. The last two theses amount to a theory of foundational evidence. The first says that all foundational evidence consists in seemings—​ that is, thoughts whose propositional content feels true or false: Seemings Exclusivism (SE)—​If S has foundational evidence for p, then (it is because) it seems to S that p is true; if S has foundational evidence against p, then (it is because) it seems to S that p is false. Seemings exclusivism does not imply that we are justified in believing something only if it seems true. We may be justified in believing a proposition that doesn’t seem true, or even one that seems false, so long as it is supported by other propositions that are themselves justified.6 Of course, if all non-​immediate justification traces back to immediate justification, and all immediate justification derives from foundational evidence, and all foundational evidence consists in seemings, then a proposition that does not itself seem true can only be justified if it is supported by other things that do seem true. In short, FN, EV, and SE together imply that any justification we attain ultimately originates from seemings. That being said, none of the above theses give us any assurance that we can, in fact, attain justification. SE says that if we have foundational evidence, it consists in seemings; but that is not to say that foundational evidence is a genuine possibility. The last thesis assures us that having foundational evidence is indeed possible. It is a version of phenomenal conservatism, which says: Phenomenal Conservatism (PC)—​ If it seems to S that p is true, then (because of that and to that extent) S has foundational evidence for p; and if it seems to S that p is false, then (because of that and to that extent) S has foundational evidence against p. Essentially, all seemings are foundational evidence. How strong that evidence is depends on the phenomenology of the seeming and how strongly it feels true. The more forceful the seeming, the stronger the foundational evidence. The picture of justification that emerges from seemings foundationalism is one in which justification is ultimately a matter of forming attitudes in accordance with how things seem. Our seemings, along with the things

60  Immediate Justification and Seemings Foundationalism justified by them, constitute the total set of evidence on which our justification depends.7 This is an instance of what Declan Smithies calls “global phenomenal conservatism,” which says, “Justification is wholly grounded in seemings in the sense that, necessarily, you have justification to believe that p just when and because it seems to you strongly enough on balance that p” (Smithies 2019, 383). When we begin forming beliefs about the world, the attitudes justified by our seemings are the basic stances from which the rest of our theorizing must subsequently emerge. These starting points are not infallible or beyond question as they would be on classical foundationalism. Seemings foundationalism is an instance of moderate foundationalism. Like icebergs floating in the water, our foundational attitudes can bear weight—​ even tremendous weight—​ but put enough pressure on them, or rock them in the right way, and they’ll sink below the surface. Hence, the justification provided by seemings is susceptible to defeat, though that defeat can only come at the hands of other seemings somewhere down the line. Only from the vantage of one iceberg can you sink another. I’ll spend the rest of this chapter (and the next) saying a bit more about PC. The other three theses—​FN, EV, and SE—​will be discussed at greater length when I defend them in Chapters 5–​7. 2.3  Phenomenal Conservatism Although the name “phenomenal conservatism” was coined by Michael Huemer (in Huemer 2001), philosophers have endorsed phenomenal conservatism under different guises as far back as Aristotle. Indeed, principles largely equivalent to phenomenal conservatism are prevalent throughout the tradition. This practice continues today, where many (e.g., perceptual dogmatists) endorse versions of phenomenal conservatism without using that particular label.8 What qualifies a principle as an instance of phenomenal conservatism then? And how does my specific version of phenomenal conservatism compare to other such principles? 2.3.1  The Essence of Phenomenal Conservatism

We can organize the options using Jonathan Kvanvig’s schematics for principles of epistemic normativity (Kvanvig 2014). Such principles will include a state or condition that confers the relevant epistemic status (the conferrer), the enabling conditions that allow the status to be conferred (the enabler), specification of the status conferred (the operator), and the object on which the status is conferred (the target). Thus, a standard formulation of phenomenal conservatism can be broken down as follows:

Immediate Justification and Seemings Foundationalism  61 PCHuemer

If it seems to S that p, (conferrer) then, in the absence of defeaters, (enabler) S thereby has at least some degree of justification for believing (operator) that p. (target) (Huemer 2007, 30)

For comparison, consider some other principles that I take to be versions of phenomenal conservatism: PCTooley

If it seems to S that p and, in addition, its seeming to S that p is a basic seeming, (conferrer) then, in the absence of defeaters, (enabler) S thereby has at least some degree of justification for believing (operator) that p. (target) (Tooley 2013, 323)

PCTucker

If it seems to S that p, (conferrer) then [none] (enabler) S thereby has evidence which supports (operator) p. (target) (Tucker 2011, 55)

PCDougherty If it seems to S that p, (conferrer) then [none] (enabler) S thereby has a pro tanto reason to believe (operator) p. (target) (Dougherty 2011, 333) PCAlston

Whatever [proposition] (target) appears to one as so-​and-​so (conferrer) is thereby likely to be so-​and-​so, (operator)

62  Immediate Justification and Seemings Foundationalism in the absence of sufficient indications to the contrary. (enabler) (Alston 1999, 198) PCPryor

S has immediate prima facie justification for believing (operator) those propositions (target) that S’s perceptual experiences basically represent to S. (conferrer) [none] (enabler) (Pryor 2000, 539)

PCAudi

If S has a spontaneous perceptual experience in which S has the impression that x is F, (conferrer) and on this basis attentively believes that x is F, (enabler) then this belief (target) is prima facie justified. (operator) (Audi 1987, 9)

In my judgment, what qualifies a principle as a version of phenomenal conservatism is that it meets three conditions. First, it must appeal to seemings or appearances as conferrers. There are different ways of characterizing seemings, as we will see in Chapter 4, but a necessary feature seems to be that they possess a distinct phenomenal character. This is the “phenomenal” part of phenomenal conservatism. Note that principles such as Pryor’s or Audi’s that do not explicitly mention seemings still meet this condition based on how they understand the nature of perceptual experience (see, e.g., Pryor 2000, ft 37). The second condition on phenomenal conservatism is that seemings must confer some positive epistemic status on a proposition or propositional attitude. Usually this is (some degree of) justification or prima facie justification for believing, but some of the above versions characterize seemings as conferring evidential or rational support or as making a proposition epistemically likely. Though it usually goes unstated in official formulations, proponents almost always add that the strength of this positive epistemic status (the degree of justification; the strength of the evidence; the level of rational support; the likelihood of the proposition, etc.) is proportional to the strength of the seeming (see, e.g., Huemer 2007, ft 1). Whatever the status, its target is normally a proposition, although principles can also, like Audi’s, target attitudes such as belief.

Immediate Justification and Seemings Foundationalism  63 The third and final condition on phenomenal conservatism is that the enabling conditions are such that seemings confer the aforementioned positive epistemic status by default. Often, there are no special enabling conditions at all.9 If there are special enabling conditions listed, they are simply the lack of things (like defeaters) that would block seemings from conferring that status or, when the status is doxastic justification, the additional requirement of proper basing. As Pryor puts it, the “mere having” of a seeming confers the positive epistemic status (Pryor 2000, 519). In particular, one needn’t have any prior verification of that seeming’s reliability.10 Of course, there’s nothing stopping one from reflecting on whether a seeming is reliable or not, but its initial justificatory power is not derived from such reflections. This is the “conservatism” in phenomenal conservatism. Huemer describes the lack of enabling conditions as granting seemings the “presumption of innocence” (Huemer 2001, 100). The metaphor, of course, is that there is a kind of tribunal of reason before which seemings are judged. As Reid writes: As a judge, after taking the proper evidence, passes sentence in a cause, and that sentence is called his judgment; so the mind, with regard to whatever is true or false, passes sentence, or determines according to the evidence that appears. … The analogy between a tribunal of justice and this inward tribunal of the mind, is too obvious to escape the notice of any man who ever appeared before a judge. And it is probable, that the word judgment, as well as many other words we use in speaking of this operation of the mind, are grounded on this analogy. (Reid [1785] 2002, 407) Huemer’s claim is that, on phenomenal conservatism, seemings are to be declared “innocent” (its content true) unless proven “guilty” (its content false). The defense can choose to offer evidence for the innocence of its client—​and doing so can shore up the defendant’s position—​but it does not ordinarily need to do so. Even if no evidence is offered for the defendant’s innocence, the defendant should still be found innocent absent sufficient evidence against it. This analogy, while helpful, isn’t perfect since on some versions the positive epistemic status that seemings confer is something less than full justification for believing; and overcoming this presumption does not always require presenting evidence of the seeming’s “guilt” (or falsity), or proving that guilt beyond a reasonable doubt. A more exact analogy is to say that phenomenal conservatism licenses us to place implicit trust in the way things seem. When we trust a person’s testimony with respect to p, we lend credence to that testimony—​the greater the trust, the greater the credence. If the person testifies that p,

64  Immediate Justification and Seemings Foundationalism we may not always end up believing p (if there are other things we trust more that lead us toward not-​p), but we at least give their testimony weight in our theorizing. To trust a person’s testimony implicitly, then, is to lend their testimony credence prior to having evidence of its reliability. Notice that this implicit trust, as I have described it, is manifested in our response to the person’s testimony, not what we believe about that person’s testimony. A toddler implicitly trusts his mother’s testimony by automatically believing what she says, not by believing that what she says is likely to be true. The second-​order belief about reliability need not be present, and almost certainly isn’t present in the case of the toddler. To implicitly trust a seeming then is to lend credence to it in theorizing, even absent verification of its reliability. This implicit trust is manifested not by believing anything about the seeming, but by believing in accordance with it—​that is, by the forcefulness of the seeming moving one’s intellect directly. The testimonial analogy is an especially apt one. Reid and many others have maintained that there is a close analogy between experience and testimony. Speaking specifically of perceptual experience, Reid writes: There is a much greater similitude than is commonly imagined between the testimony of nature given by our sense, and the testimony of men given by language. (Reid [1764] 1997, 171) Our senses testify immediately the existence, and many of the attributes and relations of external material beings; and, by our constitution, we rely with assurance upon their testimony, without seeking a reason for doing so. (Reid [1785] 2002, 448–​449) Elsewhere he talks about the testimony of consciousness (or introspection), of memory, and of reason (Reid [1785] 2002, 463). The idea is that each of our faculties produce experiences which can be treated as the “testimony” of those faculties.11 Reid, relying on the mistaken assumption that our experiences consist in beliefs, concluded that our foundational beliefs are themselves the testimony of experience; however, that testimony is better conceived of as the seemings on which our foundational beliefs are based (McAllister 2019, 38–​39). Seemings, we have said, present their content to the subject as true via their unique phenomenal character. Fittingly, Chris Tucker describes this phenomenal character as making it “feel as though the seeming is ‘recommending’ its propositional content as true or ‘assuring’ us of the content’s truth” (Tucker 2010, 530). Hence, we can think of seemings as analogous to the testimony of our various rational

Immediate Justification and Seemings Foundationalism  65 faculties offered up to our first-​person consciousness for consideration.12 Granting this, phenomenal conservatism effectively says that we are justified in implicitly trusting in the testimony of our own minds. In this, phenomenal conservatism mirrors Reid’s own views concerning testimony more generally. Reid thinks that humans are naturally credulous, meaning that we are naturally disposed to form a belief that p when someone testifies that p, even if we do not have verification of the testifier’s reliability. As Alexander Campbell, a follower of Reid, puts it: Man is, therefore, so created and circumstanced now, as to be naturally and necessarily credulous. Credulity—​for I know no term more expressive of the native bias to receive truth upon testimony—​I say, credulity is as natural to man as breathing. (Campbell 1839, 107) Reid goes on to maintain that believing in accordance with this natural credulity is justified. That is, if the testimony that p occasions belief that p in accordance with the proper functioning of our naturally credulous disposition, then that belief is justified. This is Reid’s famous “principle of credulity” (Reid [1785] 2002, 487–​488). According to this principle, testimonial justification does not require evidence in the reliability of the testimony (or testifier) but is instead immediate. Reid’s view can be contrasted with Hume’s, which requires independent verification of the testimony’s trustworthiness before that testimony can justify belief (Hume [1748] 1999). Reid goes on to apply this same trusting approach to the testimony of our own rational faculties. Consider what he says about foundational introspective beliefs: But why did [Descartes] not prove the existence of his thought? Consciousness, it may be said, vouches for that. But who is voucher for consciousness? Can any man prove that his consciousness may not deceive him? No man can: nor can we give a better reason for trusting it, than that every man, while his mind is sound, is determined, by the constitution of his nature, to give implicit belief to it, and to laugh at, or pity, the man who doubts its testimony. (Reid [1764] 1997, 17) Just as we are naturally credulous toward the testimony of other people, so we are natural credulous toward what consciousness (and our other rational faculties) vouch to be true. And just as implicit trust is rational in the former case, so it is the latter. Thus, Reid concludes, the attitudes formed in accordance with this implicit trust are justified. Phenomenal conservatism says that seemings provide justification in a similar way to testimony for Reid.13 The comparison is not

66  Immediate Justification and Seemings Foundationalism exact—​particularly, Reid places an emphasis on naturalness and proper function that is not obviously present in phenomenal conservatism,14 and his focus is on beliefs rather than seemings. Nevertheless, the basic idea is the same in both cases: seemings are to be implicitly trusted, their testimony automatically given weight in theorizing, just as Reid says that we are to implicitly trust the testimony of both other people and our own minds. We will return to the testimonial comparison when considering objections to phenomenal conservatism below. A possible fourth condition on phenomenal conservatism is that the initial trustworthiness of the relevant seemings is necessary and stems from something about those seemings themselves. This would exclude purported externalist versions of phenomenal conservatism (e.g., Bergmann 2013), which maintain that seemings are merely contingent sources of immediate justification which justify only because the right external conditions are in place.15 I will, for the present moment, refrain from endorsing this condition. Let externalist versions of phenomenal conservatism bear the name. It remains the case, of course, that the contemporary proponents of phenomenal conservatism most responsible for its current rise to prominence all endorse versions of the principle which meet this condition, since they are operating within internalist frameworks. Hence, I think it fair to call principles meeting this fourth condition (along with the others) “standard phenomenal conservatism.” Before moving on, we must dispel a persistent misunderstanding of phenomenal conservatism. Phenomenal conservatism claims that the justifying condition for p is p’s seeming true, not having justification for believing that p seems true. Critics sometimes misconstrue phenomenal conservatism on this point. For instance, some interpret phenomenal conservatism as saying that when we update on the proposition p seems true (call this proposition “q”), it raises the probability that p (e.g., White 2006). But this is a critical misinterpretation (Miller 2016, Moretti 2020). On this way of modeling things, the justification for p is not immediate at all but is rather based on our prior justification for q. Essentially, we are inferring p on the basis of q. But the whole point of phenomenal conservatism is to reject our reliance on inferential justification of this sort! To point out a misinterpretation this grave is not mere “quibbling,” as some have suggested (Lasonen-​Aarnio and Hawthorne 2021, ft. 13), but simply to insist that critics actually engage with the view rather than a straw man of their own making. If the critics claim not to understand the sort of position I describe, then I must ask them how we are supposed to gain justification for believing that the proposition p seems true to us? Surely, we are not justified in believing that we have this seeming via updating on some higher-​ order proposition such as it seems to me that it seems to me that p. For then we’d have to ask about the justification of that higher-​order proposition,

Immediate Justification and Seemings Foundationalism  67 and so on and so forth. The correct answer is that they take our justification for believing that p seems true to be immediate. When we introspectively attend to how things seem, that reflective state immediately justifies us in believing that p seems true. No updating on other propositions is needed. According to phenomenal conservatism, this is what happens with all seemings, including those whose content extends beyond the confines of our own minds. Critics are under no obligation to agree, but they should at least characterize the position accurately. 2.3.2  PC Among the Conservatives

How does my specific version of phenomenal conservatism, PC, stack up against other versions? Here is my principle once more, broken down as above: PC

If it seems to S that p is true, (conferrer) then (because of that and to that extent) S has foundational evidence for (operator) p. (target) [none] (enabler) If it seems to S that p is false, (conferrer) then (because of that and to that extent) S has foundational evidence against (operator) p. (target) [none] (enabler)

Let us walk through each of these four categories. 2.3.2.1 Conferrer

The most noticeable difference between PC and existing versions of phenomenal conservatism is that PC mentions two different kinds of seemings. There is the kind of seeming one has when something feels true, which I call a “seeming-​true”; and there is the kind of seeming one has when something feels false, which I call a “seeming-​false.” I suspect that others have avoided drawing this distinction under the impression that its seeming false that p is the same state of affairs as its seemings true that not-​p. This would be convenient if true, but I don’t think it is. For one, the phenomenology of seemings-​true and seemings-​false is notably different. The one recommends its content to you, urging you to assent; whereas the other

68  Immediate Justification and Seemings Foundationalism rejects its content, urging you toward disbelief. These states feel as different as an aversion for p and a desire for not-​p. For another, the propositional content of the states is different. When 1+​1=​3 seems false to me, it is that precise proposition—​1+​1=​3—​I am entertaining, and not its negation. For convenience, I will almost exclusively talk about seemings-​true, especially once I get into the positive case for seemings foundationalism; however, everything I say will apply mutatis mutandis to seemings-​false. Next, PC is an unrestricted version of phenomenal conservatism, meaning that all seemings are conferrers. On restricted versions of phenomenal conservatism only some seemings are conferrers. For example, Tooley’s version of phenomenal conservatism is restricted to “basic seemings.”16 Not so for PC. A worry for unrestricted principles like PC is that it opens it up to the charge of over-​liberality. I will briefly respond to this charge in Chapter 3. The worry for restricted principles, on the other hand, is whether they can draw a principled distinction between the seemings that qualify as foundational evidence and those that do not. I argue that this distinction cannot be drawn in Chapter 9. 2.3.2.2 Operator

PC differs from most if not all versions of phenomenal conservatism in the positive epistemic status that seemings confer. On PC, seemings provide foundational evidential support for or against their content (depending on whether it’s a seeming-​true or seeming-​false). Indeed, one understands from the discussion of evidence in §2.1 that I take the seemings themselves to be the foundational evidence in question. Furthermore, when I say that we have foundational evidence for or against p “to the extent” that p seems true or false, I mean that the strength of the foundational evidence for or against p is proportional to how strongly p feels true or false, which is determined by the strength of the seeming’s phenomenal character. We also understand from our discussion in §2.1 that having foundational evidence for or against p does not guarantee that one has justification for believing or disbelieving p. For one, there is always the possibility that other considerations diminish the on-​ balance level of evidential support for p via defeaters (broadly construed so as to include low pure prior probabilities per our discussion in §2.2).17 However, even if there are no defeaters, it may still be the case that a weak seeming is incapable of providing justification for believing or disbelieving p. Perhaps the strength of that foundational evidence is only sufficient to justify a slight positive or negative credence in p. PC even leaves open the possibility that certain seemings only provide some slight “presumption” in favor of their content (McCain and Moretti 2021, 80). That being said, I will argue later on (as a contingent matter of fact, not as a part of PC) that many of our seemings

Immediate Justification and Seemings Foundationalism  69 are in fact strong enough to justify full belief or disbelief in their contents, absent defeaters. 2.3.2.3 Enabler

PC does not appeal to any special enabling conditions. One has foundational evidence for or against something simply by virtue of its seeming true or false. Some of the other versions of phenomenal conservatism include a “no defeaters” condition as an enabler. For instance, justification for believing is conferred if there are no defeaters present. The lack of such a condition suggests that seemings continue to count as foundational evidence even after their evidential support has been defeated. They are evidence that has been countered rather than no longer counting as evidence. Is this correct, or should PC include a no-​defeaters clause, saying “If it seems to S that p is true, then S has foundational evidence for p in the absence of defeaters”? I do not think a no-​defeaters clause needs to be added. A seeming that p will always be foundational evidence for p in the sense that it permits belief or some positive credence in p in the state of empty symmetrical evidence. This reflects the fact that the seeming is intrinsically aimed at making permissible belief or positive credence in its content, even if that’s not its ultimate effect in present conditions (or even if its ultimate effect is the opposite). In this sense, then, we should continue to identify it as foundational evidence for its content. There is, however, another sense in which people use the word “evidence” in which evidence for p must increase one’s level of justification for p in present conditions—​that is, given one’s current background evidence—​rather than in a state of empty symmetrical evidence. If this is how we insist on using the term “evidence” then a no-​defeaters clause does need to be added, as well as other qualifications. Notwithstanding, I will use “evidence” in the former sense as it applies to PC. Thus, what PC claims is that seemings intrinsically indicate the truth of their content such that, in the state of empty symmetrical evidence, the subject is justified in forming some positive doxastic attitude toward that content. However, nota bene, PC does not claim that seemings will always increase one’s level of justification for their content regardless of one’s background evidence. 2.3.2.4 Target

The target of PC, like most, is a proposition—​specifically, the content of the conferring seeming. Thus, the evidential support provided by seemings aims at propositional justification. Such support can come to result in a doxastically justified attitude, but that will require proper basing as well.

70  Immediate Justification and Seemings Foundationalism For that, the seemings not only need to be adequate evidence for one’s attitude but also need to cause that attitude in the right sort of way. In particular, this will include the belief being caused by the features of the seeming that make permissible the attitude in question. Finally, I intend PC to be taken as a necessary principle. There is something about seemings themselves, independent of the external conditions in which they are situated, that makes them sources of immediate justification. Indeed, given casual internalism, their inherent trustworthiness must stem from their directly accessible properties, such as what it feels like to be in them. Thus, PC is an instance of standard phenomenal conservatism. By now, we should have a much firmer grasp on the nature of PC and its place within seemings foundationalism. The next two chapters will deepen our understanding of these views in important ways. In Chapter 3, I will examine some of the implications of PC for justification more generally. In particular: Does PC substantiate a common sense reply to skepticism? Does PC allow for just anything to be justified? Does it permit circular reasoning? Does it mistakenly “double-​count” one’s evidence? In Chapter 4, I explicate the nature of seemings and address those who find them mysterious or under-​specified, as well as those who deny their existence altogether. Both discussions not only increase our understanding of seemings foundationalism, but also ward off some of the more immediate concerns one might have about the view. Notes 1 I will assume that there is an objective fact about which attitudes our epistemic duties permit. (I take a more subjective approach in McAllister 2019, following Richard Foley, which I depart from here.) On an objective framework, however, it can still be asked whether there is exactly one set of beliefs and credences that satisfy our epistemic duties. Permissivists say no. (For more on permissivism, see Jackson and Turnbull forthcoming). Proponents of uniqueness disagree (Feldman 2007). With respect to these goals, they will insist that there is only one proper weighting (likely an equal one), and only one set of attitudes that permissibly pursues those goals. The position I wish to defend is neutral with respect to this question, and so I leave it aside. 2 Indeed, a mental state may aim toward permitting belief even when its total effect is the opposite. By analogy, gravity still aims downward even if applying downward force to the rocket triggers some other, stronger force that moves the rocket upward. This can happen, for instance, when higher-​order considerations suggest that having evidence for p is, in present conditions, stronger evidence for not-​p. 3 This might be a reason to disagree with McCain’s judgment that “the disagreement [between propositionalism and psychologism] does not lead to significant epistemic differences” (McCain 2014, 27). On the other hand, if McCain is

Immediate Justification and Seemings Foundationalism  71 correct that little hangs on the disagreement, then all the better for my framework, since this removes a possible point of contention. 4 The arguments in Williamson (2000, ­chapter 9, section 5) are perhaps the most pressing. See Conee and Feldman (2008; 2011) for replies. Kelly (2016) also raises the concern that mental states—​experiences, in particular—​may not be able to play all of the roles that evidence is paradigmatically thought to play. See Dougherty and Rysiew (2014) and Gage (2018) for responses. 5 How low? Low enough that, based on this background evidence alone, one would not be justified in taking the positive doxastic attitude toward p that the foundational evidence would otherwise allow. (Technically, then, a “low” pure prior could be quite high. Say, when foundational evidence would otherwise justify a credence in p of .95, but when the background evidence only justifies a credence of .8.) 6 Smithies (2019, ­chapter 11) gives the name “strong phenomenal conservatism” to the position that every justified belief that p must be justified by a seeming that p, but no phenomenal conservative I know of endorses such a view. 7 I will leave open which seemings count as being in our possession at any particular time. This is a difficult question, and it’s not one for which I take myself to have a fully satisfying answer. Certainly, the seemings we possess at t extend beyond those that are occurrent at t (contra Feldman 1988). If it seems to us that it is impossible for a figure to be both a circle and a square, we do not lose that bit of evidence a moment later when we cease to think of it. So, there must be some way in which we “store” seemings that were once occurrent but are no longer. On the other hand, not all seemings are stored or remain stored forever. Those seemings, especially in the distant past, that leave no lasting mark on our minds might as well not have occurred at all. One idea is that a non-​occurrent seeming still counts as being in our possession so long as it was once occurrent and continues to have some lingering causal influence on how we believe. Another idea, following McCain (2014, 51), is to say something like this: S has the seeming that p available as evidence relevant to the belief that q at t if and only if, at t, S has an occurrent seeming that p or S is disposed to bring the seeming that p to mind when reflecting on the question of q’s truth. Fortunately, none of my arguments will hinge on which theory of possession we adopt. I also take comfort in the fact that such questions are not unique to seemings foundationalism. All forms of mentalism that do not limit themselves to occurrent states must wrestle with which non-​occurrent mental states count as a part of our total evidence. I see no reason to think that seemings foundationalism can’t help itself to whatever answer wins out here. 8 For more on existing support for phenomenal conservatism (or principles very much like it), see Chapter 8, §8.1. 9 Obviously, there will still be broader enabling conditions in the background such as the existence of the subject and his cognitive capacities, but those go without saying. 10 A seeming is “reliable” if and only if forming the doxastic attitude toward which that seeming inclines you is a reliable process.

72  Immediate Justification and Seemings Foundationalism 11 Nowadays, people prefer to talk of “cognitive tools” or “modules” rather than “faculties.” I will go ahead and continue to talk of “faculties” although I am happy that this should be understood in the way most consonant with our best cognitive science. 12 Our faculties may be innately hard-​wired to produce certain seemings. For instance, Reid thought that we were innately disposed to conceive of a material substance causing our sensations upon feeling those sensations. However, other seemings appear to be produced by the subconscious processing of our rational faculties. My suggestion is that our faculties can tacitly detect logical relationships between states of affairs in our environments, and/​or other things we already believe, and new propositional contents (McAllister and Dougherty 2019). This new propositional content is then be encapsulated in a seeming and presented to the conscious mind as something true. 13 Though note that one can be a phenomenal conservative without being a Reidian about interpersonal testimony. One needs only draw a relevant distinction between the “testimony” of one’s own mind and the testimony of other people that permits implicit trust in the one but not the other. Personally, I think such a distinction can be drawn (particularly because implicit trust in the former is constitutive of rationality in a way that implicit trust in the latter is not; see Chapter 8, §8.5). That being said, it is plausible that intellectual self-​trust commits us to at least some kind of trust in others (Foley 2001, Zagzebski 2012), though it is beyond the scope of this inquiry to say in what way. 14 On the other hand, if we understand “natural” cognitive dispositions in a normative sense—​as those that belong to an ideally constituted human agent—​and proper function as merely the unfolding of these cognitive dispositions without obstruction, then phenomenal conservatives would agree with Reid that implicitly trusting seemings is natural and that those attitudes formed by the properly functioning agent in accordance with this natural disposition are justified. 15 See the discussion of Reid and phenomenal conservatism in the introduction to this book. 16 Basic seemings are non-​derived seemings, where a seeming is derived “if and only if there is some proposition q such that it seems to S that q, where that causes it to seem to S that p, by a process internal to S, and where, otherwise, it would not seem to S that p” (Tooley 2013, 322). 17 Note that, by formulating PC in terms of foundational evidence, and acknowledging that foundational evidence can be prevented from immediately justifying positive doxastic attitudes by low pure priors, PC escapes some of the most serious worries raised for phenomenal conservatism by Lasonen-​Aarnio and Hawthorne (2021).

References Alston, William P. 1999. “Back to the Theory of Appearing.” Philosophical Perspectives 13: 181–​203. Audi, Robert. 1987. Belief, Justification, and Knowledge: An Introduction to Epistemology. 1st edition. Belmont, CA: Wadsworth Publishing.

Immediate Justification and Seemings Foundationalism  73 Bergmann, Michael. 2013. “Externalist Justification and the Role of Seemings.” Philosophical Studies 166 (1): 163–​184. Campbell, Alexander. 1839. Debate on the Evidences of Christianity. London: R. Groombridge. Conee, Earl, and Richard Feldman. 2004. Evidentialism: Essays in Epistemology. Vol. 48. Oxford: Oxford University Press. —​—​—​. 2008. “Evidence.” In Epistemology: New Essays, edited by Quentin Smith, 83–​104. Oxford: Oxford University Press. —​—​—​. 2011. “Replies.” In Evidentialism and Its Discontents, edited by Trent Dougherty, 283–​323. Oxford: Oxford University Press. Dodd, Dylan. 2017. “Belief and Certainty.” Synthese 194 (11): 4597–​4621. Dougherty, Trent. 2011. “Further Epistemological Considerations Concerning Skeptical Theism.” Faith and Philosophy 28 (3): 332–​340. Dougherty, Trent, and Patrick Rysiew. 2014. “Experience First.” In Contemporary Debates in Epistemology, edited by Matthias Steup, John Turri, and Ernest Sosa, 2nd edition, 17–​21. Hoboken, NJ: Wiley-​Blackwell. Feldman, Richard. 1988. “Having Evidence.” In Philosophical Analysis, edited by D. F. Austin, 83–​104. Amsterdam, NL: Kluwer Academic Publishers. —​—​—​. 2007. “Reasonable Religious Disagreements.” In Philosophers Without Gods: Meditations on Atheism and the Secular Life, edited by Louise Antony, 194–​214. Oxford: Oxford University Press. Foley, Richard. 1993. Working Without a Net: A Study of Egocentric Epistemology. Oxford: Oxford University Press. —​—​—​2001. Intellectual Trust in Oneself and Others. Cambridge: Cambridge University Press. Gage, Logan Paul. 2018. “Can Experience Fulfill the Many Roles of Evidence?” Quaestiones Disputatae 8 (2): 87–​111. Hawthorne, John, and Maria Lasonen-​ Aarnio. 2021. “Not so Phenomenal!” Philosophical Review 130 (1): 1–​43. Huemer, Michael. 2001. Skepticism and the Veil of Perception. Lanham, MD: Rowman & Littlefield. —​ —​ —​ . 2007. “Compassionate Phenomenal Conservatism.” Philosophy and Phenomenological Research 74 (1): 30–​55. Hume, David. [1748] 1999. An Enquiry Concerning Human Understanding, edited by Tom L. Beauchamp. Oxford Philosophical Texts. Oxford: Oxford University Press. Jackson, Elizabeth. 2022. “On the Independence of Belief and Credence.” Philosophical Issues 32 (1): 9–​31. Jackson, Elizabeth, and Margaret Greta Turnbull. forthcoming. “Permissivism, Underdetermination, and Evidence.” In The Routledge Handbook of the Philosophy of Evidence, edited by Clayton Littlejohn and Maria Lasonen-​ Aarnio. New York: Routledge. Kelly, Thomas. 2016. “Evidence.” In The Stanford Encyclopedia of Philosophy, edited by Edward N. Zalta, Winter 2016. https://​plato.stanf​ord.edu/​archi​ves/​ win2​016/​entr​ies/​evide​nce/​. Kvanvig, Jonathan L. 2014. Rationality and Reflection: How to Think about What to Think. Oxford: Oxford University Press.

74  Immediate Justification and Seemings Foundationalism McAllister, Blake. 2019. “A Return to Common Sense: Restorationism and Common Sense Epistemology.” In Restoration & Philosophy, edited by J. Caleb Clanton, 35–​78. Knoxville, TN: University of Tennessee Press. McAllister, Blake, and Trent Dougherty. 2019. “Reforming Reformed Epistemology: A New Take on the Sensus Divinitatis.” Religious Studies 55 (4): 537–​557. McCain, Kevin. 2014. Evidentialism and Epistemic Justification. New York: Routledge. McCain, Kevin, and Luca Moretti. 2021. Appearance and Explanation: Phenomenal Explanationism in Epistemology. Oxford: Oxford University Press. Miller, Brian T. 2016. “How to Be a Bayesian Dogmatist.” Australasian Journal of Philosophy 94 (4): 766–​780. Moon, Andrew, and Elizabeth Jackson. 2020. “Credence: A Belief-​First Approach.” Canadian Journal of Philosophy 50 (5): 652–​669. Moretti, Luca. 2020. Seemings and Epistemic Justification: How Appearances Justify Beliefs. New York: Springer. Nelson, Mark T. 2010. “We Have No Positive Epistemic Duties.” Mind 119 (473): 83–​102. Poston, Ted. 2014. Reason and Explanation. New York: Palgrave Macmillan. Pryor, James. 2000. “The Skeptic and the Dogmatist.” Noûs 34 (4): 517–​549. Reid, Thomas. [1764] 1997. An Inquiry into the Human Mind on the Principles of Common Sense. Edited by Derek R. Brooks. University Park, PA: Pennsylvania State University Press. —​—​—​. [1785] 2002. Essays on the Intellectual Powers of Man, edited by Derek R. Brooks. University Park, PA: Pennsylvania State University Press. Smithies, Declan. 2019. The Epistemic Role of Consciousness. New York: Oxford University Press. Tooley, Michael. 2013. “Michael Huemer and the Principle of Phenomenal Conservatism.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 306–​327. Oxford: Oxford University Press. Tucker, Chris. 2010. “Why Open-​Minded People Should Endorse Dogmatism.” Philosophical Perspectives 24 (1): 529–​545. —​ —​ —​ . 2011. “Phenomenal Conservatism and Evidentialism in Religious Epistemology.” In Evidence and Religious Belief, edited by Kelly James Clark and Raymond J. VanArragon, 52–​73. Oxford: Oxford University Press. White, Roger. 2006. “Problems for Dogmatism.” Philosophical Studies 131 (3): 525–​557. Williamson, Timothy. 2000. Knowledge and Its Limits. Oxford: Oxford University Press. Zagzebski, Linda Trinkaus. 2012. Epistemic Authority: A Theory of Trust, Authority, and Autonomy in Belief. Oxford: Oxford University Press.

3 Some Objections to Phenomenal Conservatism

This book is an inquiry into the foundations of justification. I have proposed that seemings are the foundational evidence on which all justification supervenes. I do not aim to provide any systematic account of how such foundational evidence combines to yield an on-​balance level of support (c.f., McCain and Moretti 2021). That being said, many share concerns about the implications of seemings foundationalism for justification more broadly, particularly because of its commitment to phenomenal conservatism (PC). For example, if seemings are foundational evidence, then wouldn’t any belief be justified, no matter how outlandish, so long as it seems true? I address this and some other notable objections to PC in this chapter. I already addressed several other concerns in the previous chapter—​ or rather, I bypassed them by formulating PC in such a way as to avoid them—​and I will address still further concerns in Chapters 8 and 9, in the process of giving a positive argument for the view. To be clear, though, the primary goal of this inquiry is to develop a positive line of reasoning in favor of PC, not to dismantle arguments against it. (The literature is full of the latter, but surprisingly short on the former.) Nevertheless, I cannot develop a convincing case for PC without saying something about these objections, if only to assure us that PC has something plausible to say in response. To avoid duplicating efforts, however, I will stick to what I take to be the heart of these issues, rather than engaging in a review of various responses present in the literature.1 In addressing these objections, I will sometimes be stepping beyond what is strictly implied by PC and seemings foundationalism, supplementing it with additional assumptions and epistemic principles. For example, I will be defending the use of PC within a broader common sense response to skepticism, though PC in itself does not require that one take such an approach. These represent my preferred ways of working out the view, but they are not the only ways of doing so. We must be careful, then, not to confuse dissatisfaction with my particular outworking with the view itself. DOI: 10.4324/9781003281108-4

76  Some Objections to Phenomenal Conservatism I try to make clear throughout what is implied by PC versus what is my preferred addition. 3.1  The Common Sense Response to Skepticism The first issue we will look at is not so much an objection to PC as an attempt to chasten its ambitions. One of the main attractions of principles like PC is that, following in the tradition of Reid and G.E. Moore, they seem to make possible a common sense reply to skepticism. The most formidable skeptical arguments gain purchase by highlighting skeptical scenarios in which things appear to us exactly as they do now, and yet virtually nothing actually is as it appears to be. The skeptic insists that, to be justified, we must independently rule out such skeptical scenarios prior to trusting in what seems to be the case. In other words, we need to verify the truth, or at least the reliability, of our seemings before we are justified in believing their content. PC rebuts this skeptical challenge not by overcoming it, but by rejecting its legitimacy. Given PC, our initial justification for believing in how things seem does not depend on our having some independent means of confirming the truth or reliability of those seemings. Seemings immediately justify belief in how things appear. Trust is the default. This rejection of the skeptic’s demand, rather than a rising to meet it, is one of the defining characteristics of common sense replies to skepticism (see, e.g., Moore 1939, 144–​148). Accordingly, those drawn to the common sense reply to skepticism have looked favorably on principles like PC as placing that sort of reply on sure epistemic footing. Luca Moretti contests whether PC really can substantiate the common sense reply to skepticism (Moretti 2018; 2020, McCain and Moretti 2021). Moretti concedes that PC does account for the justification of unreflective agents—​those who simply believe in accordance with how things seem, never questioning the origins or reliability of their seemings—​but that as soon as we reflect on our seemings, as addressing skepticism forces us to do, they can no longer justify our beliefs as they once did. Thus, in confronting skepticism, PC fails to sustain a common sense response. Moretti calls this “the problem of reflective awareness.” I will examine Moretti’s reasoning more closely in a moment, but before I do, we should consider what the impact would be if we granted Moretti’s conclusion in full. It certainly wouldn’t sound the death knoll for PC (as Moretti, a proponent of phenomenal conservatism himself, certainly recognizes).2 There are many other reasons for endorsing PC besides its ability to sustain a common sense response to skepticism—​including the arguments I will offer in the latter chapters of this book. Indeed, nothing in Moretti’s argument even prohibits principles like PC from playing an indispensable role in responding to skepticism. Moretti himself endorses an

Some Objections to Phenomenal Conservatism  77 explanationist or abductive response to skepticism (McCain and Moretti 2021, ­chapter 7), but such responses must account for the immediate justification of one’s initial data; and for that you need a principle like PC. On the other hand, the place of PC within a common sense approach to skepticism has always been a central attraction of such principles. It is no coincidence that those historical figures most closely associated with PC, such as Reid and Moore, are also the forefathers of the common sense tradition. Hence, the relationship between PC and common sense is not something to be abandoned lightly. Thus, in the first subsection, I will criticize Moretti’s argument, showing how a common sense response to skepticism can be conjoined with PC. In the second subsection, I defend the plausibility of this PC-​based common sense response. 3.1.1  The Problem of Reflective Awareness

Why does Moretti think that reflecting on our seemings has a destructive effect? In reflection, we consider the fact that we have some seeming and ask whether it is veridical. Are things as they appear to be? Perhaps they are (the veridical hypothesis), but various possibilities of error (deception hypotheses) also present themselves. Either we have independent evidence that rules out these possibilities of error—​ indicating that the veridical hypothesis is more likely than any deceptive hypothesis—​or we do not have such evidence. If we do not, then Moretti insists we have an undermining defeater for any immediate justification provided by the seeming. If we do have independent verification, then Moretti maintains that the immediate justification provided by our seeming is supplanted. Our trust in that seeming is no longer given implicitly but is grounded in our reflective assessment. In either case, belief is no longer immediately justified by our initial seeming. Moretti concludes that principles like PC do not reflectively sustain a common sense reply to skepticism. What are we to make of this argument? Again, even if we grant Moretti’s argument, it remains the case that unreflective agents are immediately justified in their beliefs (assuming they lack defeaters). This is no trivial outcome. Skeptics do not ordinarily limit their accusations to those who reflect on their beliefs; they impugn the unreflective as well or more so. Compare this outcome to a common contextualist response to skepticism. The contextualist concedes that you do not know that you are driving your car once the skeptic has raised the uneliminated possibility that you’re dreaming; nevertheless, in ordinary contexts, when the possibility of dreaming is properly ignored, you can and do know that you are driving (see, e.g., Lewis 1996). Knowledge is elusive upon reflection, but still most people know most ordinary things most of the time. If Moretti is correct,

78  Some Objections to Phenomenal Conservatism then we may face a similar situation with PC. It is only those who reflect who can no longer rely on common sense justification. While perhaps not entirely satisfying—​we don’t get everything that we originally wanted from the common sense reply to skepticism—​this concessive version would still stand as a major bulwark against skepticism. This much Moretti admits. Something that Moretti does not recognize, however, is that PC also preserves the justification of those who are reflective in an ordinary way. It seems to Tom, a college student, that a bear is walking across campus. In surprise, he asks himself whether his eyes deceive him. He quickly runs through the ordinary possibilities of error—​ Did I mistake a big dog for a bear? Could it be a statue? Did I imbibe any hallucinogens?—​and dismisses them all just as quickly on the basis of additional seemings and ordinary sorts of background beliefs—​That’s way too big to be a dog; it just moved, so it’s not a statue; I didn’t take any drugs either, and hallucinogens are really uncommon anyways; seems it must really be a bear! These ordinary background beliefs, moreover, are not formed in ways designed to meet the skeptic’s challenge (he is not retreating all the way back to incorrigible beliefs about how things appear, for instance). Instead, they reflect an implicit trust that the world largely is as it seems to be. As such, they are justified in accordance with PC. If Tom uses these common sense background beliefs to confirm his original seeming, then, sure, there may be some sense in which the justification of Tom’s bear belief is no longer immediate (or not entirely immediate),3 but this hardly disqualifies it as an instance of commonsensism. Tom shores up confidence in his original seeming by relying on other seemings (via his common sense background beliefs) that the skeptic would be just as eager to contest as the first. This is akin to G.E. Moore proving that the external world exists by relying on the premises “Here’s a hand” and “Here’s another!” Therefore, PC preserves the justification of not only those who fail to reflect on their seemings, but also those who reflect on them in a commonsensical way.4 But what of those who consider skeptical scenarios directly? Does PC license a common sense response for them? It would seem so. Imagine Tom’s classmate Dave mentions to Tom the possibility that he’s a brain in a vat with scientists artificially producing in him the appearance of the bear. Tom (being in no mood to play games) dismisses this as absurd. More precisely, he dismisses it because it is incongruent with his ordinary background beliefs, which, given PC, are justified by seemings. Since Tom is relying on well-​justified beliefs to dismiss this possibility of error, it seems his swift dismissal is perfectly appropriate. After all, if we agreed that Tom was permitted to dismiss the ordinary possibilities of error listed above—​ possibilities that are far more plausible to him than his being a brain in a vat—​then we surely cannot blame him for dismissing possibilities that

Some Objections to Phenomenal Conservatism  79 are far less plausible in the same way. If we are tempted to criticize Tom in the latter case but not the former, I suspect this is because we implicitly take on the skeptic’s own standards when confronting skepticism: we are looking for Tom to rule out that scenario in a way that would satisfy the skeptic. But a hallmark of the common sense tradition is a refusal to accept the skeptic’s standards as legitimate and binding. If we couple PC with this common sense refusal, then Tom is justified in dismissing the skeptical scenario simply on the basis of its incongruence with a wealth of other things that seem true. Again, Moretti is right to note that, to the extent that Tom relies on his ordinary background beliefs, the justification of his belief is supplanted (or supplemented) by inferential support. But Tom is still offering a common sense reply to skepticism so long as, like Moore, the background beliefs to which he appeals are immediately justified in ways that the skeptic would just as soon contest. Thus, I take it that PC can be integrated into a common sense reply to skepticism, even for those who reflectively engage with the skeptic, so long as we are allowed to dismiss skeptical challenges to one’s seemings-​ based belief on the basis of other seemings-​based beliefs that are just as susceptible to the skeptic’s challenge as the first. This would reduce the degree to which we are relying on exclusively our original seeming, but that seems proper and right and not at all inconsistent with the common sense approach of Reid or Moore. What exactly is the problem with Moretti’s argument then? I see at least two errors. The first is the one I have just argued for: that it is possible for us to inferentially support the veridicality of our seemings without thereby departing from the common sense reply to skepticism, so long as that support is itself commonsensical. The second is one that I will now argue for: that, in at least some cases, one can continue to permissibly believe in accordance with how things seem even if no supplementary argument can be provided. Say it seems to George that triangles have exactly three sides. He reflects on this rational seeming, or intuition, which is particularly strong, and admits that, possibly, it is mistaken. It is not the least bit plausible to him that he is mistaken; he does not think there is any serious possibility that this intuition is false. The intuition is so compelling that he cannot find it within himself to lend any significant credence to this possibility of error. Nevertheless, he recognizes in some abstract sense that human intuitions can be mistaken, even with respect to things that we feel supremely confident about, and that, alas, he is human.5 Must George appeal to other beliefs about the reliability of mathematical intuition to shore up his original seeming; or is the original intuition strong enough to continue to permit belief all on its own? It strikes me that George’s intuition might be strong enough to continue justifying belief, despite the possibility of error

80  Some Objections to Phenomenal Conservatism being raised. The key is that this mere possibility is not a serious possibility, and only serious possibilities or error defeat justification.6 If it was plausible that he made an error, then, certainly, George would have to use other considerations to rule that error out before continuing to justifiably believe. But the mere mention of a possible error, with nothing at all behind it, seems insufficient to undermine George’s original justification for belief.7 I have said that PC calls for us to place implicit trust in our seemings. It would be a fragile trust indeed were it undermined simply by acknowledging that, possibly, our seemings are wrong. Return to the testimonial analogy. Say, someone testifies to us that triangles have exactly three sides, and we implicitly trust them as the Reidian account of testimony permits. Someone else then mentions to us, “You know, maybe he’s wrong.” This skeptic is not asserting that the person is wrong, nor is he insinuating that his being wrong is at all plausible: he just notes it as a logical possibility. Are we no longer permitted to trust the person in light of this comment? My sense is that the Reidian about testimony should say no. If belief is originally permitted, then the mention of a merely possible error doesn’t change this. Or if it does, it only partially undermines the immediate justification provided by that testimony rather than removing it entirely. That’s the position I take toward seemings and the implicit trust we place in them. I should be clear that this is my preferred interpretation of PC and the implicit trust it authorizes. It results from wedding PC to a common sense principle about which possibilities of error are serious—​serious possibilities being, in this context, ones that do not allow us to continue trusting in how things seem without verification: The Common Sense Standard—​A possibility of error is serious only if it is plausible. Moretti, on the other hand, sees our implicit trust as being more fragile. The mere absence of verification for a seeming makes the possibility of error a serious one. Thus, the dispute between Moretti and myself is not about PC per se but about the common sense standard. We agree that implicit trust in our seemings is permitted; we simply disagree over how resilient that trust is allowed to be when faced with mere possibilities of error. Accordingly, those who find themselves more sympathetic to Moretti’s approach should not see that as a reason to reject PC, just my particular wielding of it. That being said, I would like to argue that these common sense applications of PC are in fact quite plausible—​that it is reasonable for one to proceed in accordance with this common sense approach to skepticism, as proponents of principles like PC have traditionally maintained.

Some Objections to Phenomenal Conservatism  81 3.1.2  Defending Common Sense

Both the skeptic and the man of common sense can agree that a belief that p cannot be justified if there is a serious possibility that this belief is false. What the skeptic and the common sense fellow disagree over is which possibilities of error count as serious. The skeptic holds to the following standard: The Skeptical Standard—​Any possibility of error is a serious possibility of error. By this standard, if any possibility of error exists, it must be ruled out before one’s belief can be justified. The person of common sense denies this. Some possibilities of error are non-​serious and can be properly ignored. To become serious, those possibilities of error must have something going for them—​they must be plausibilities, not just possibilities. When it comes to extreme skeptical scenarios, then, the man of common sense claims that those are non-​serious. Why? Because there is nothing going for them and a lot going against them—​namely, they conflict with all sorts of other things that seem true to us as enshrined in our ordinary background beliefs (not to mention many of them seem evidently false in their own right). Relying on his skeptical standard, the skeptic deems this approach unreasonable. He points out that these ordinary background beliefs are themselves susceptible to skeptical scenarios, and so cannot be justified without first eliminating those possibilities of error. Thus, the original belief still faces its own possibility of error, which must be treated as serious, and so cannot be justifiably believed either. The proponent of common sense replies that skeptical scenarios do not raise serious possibilities of error to his background beliefs for, as in the first instance, such possibilities have nothing going for them and a lot going against them. And so the debate goes, round and round. Sooner or later, our common sense fellow will break off the conversation and go about his day. The question for us is: has he proceeded reasonably? Can we fairly criticize or blame him for the commonsensical manner in which he dismisses the challenge posed by skeptical scenarios? In answering these questions, we should first ask ourselves whether there is any plausible alternative. Return to the example of Tom, who is trying to determine whether to trust his perceptual seeming when his friend suggests that he might be a brain in a vat. Let us imagine that Tom treats this skeptical suggestion as a serious possibility. Panic sets in as he realizes that his appearances might be illusions concocted by evil neuroscientists. If there is an argument for the superiority of the real-​world hypothesis over the brain-​in-​a-​vat hypothesis—​one that does not appeal to common

82  Some Objections to Phenomenal Conservatism sense perceptual beliefs (these also being susceptible to the brain-​in-​a-​vat hypothesis)—​Tom can’t think of it. Accordingly, Tom dutifully lets go of his perceptual belief. And not only that. He relinquishes all beliefs whose justification likewise depends on perception. As a result, Tom is paralyzed, unable to move voluntarily for lack of belief that he has a body. Obviously, this is an extreme example. In normal cases, Hume is correct in saying, “such a fatal event is very little to be dreaded. Nature is always too strong for principle” (Hume [1748] 1999, 207). But what matters is not whether this commonly happens, or even whether it could happen, but whether it should happen. Is this the rational and proper response to reflecting on the brain-​in-​a-​vat scenario, assuming that Tom cannot rule it out except by relying on other common sense perceptual beliefs? It seems evident (to me at least) that the answer is no. If Tom was my student, I wouldn’t beam with pride at his fearless rationality. I would rush his slumped body to the hospital! What he needs at that point is a psychiatrist, not a doctor in philosophy. Point is, the alternative upheld by skeptics as “reasonable” seems an awful lot like lunacy (see Reid [1764] 1997, 215–​216). This is a ridiculous example, to be sure, but that’s precisely the point. Reid offers similar examples by which he means to poke fun at those who take skepticism too seriously. Consider the following yarn: Pyrrho the Elean, the father of this philosophy, seems to have carried it to greater perfection than any of his successors: for if we may believe Antigonus the Carystian, quoted by Diogenes Laertius, his life corresponded to his doctrine. And therefore, if a cart run against him, or a dog attacked him, or if he came upon a precipice, he would not stir a foot to avoid the danger, giving no credit to his senses. But his attendants, who, happily for him, were not so great sceptics, took care to keep him out of harm’s way; so that he lived till he was ninety years of age. (Reid [1764] 1997, 20) The point of this wry vignette is to provoke ridicule. This isn’t done in a mean spirit—​Reid’s goal isn’t to take pleasure in making fun (though that may be a welcome byproduct)—​rather, he is strategically prompting us to recognize that those who actually follow through on skeptical principles are being, in the truest sense of the word, ridiculous, worthy of ridicule, and so cannot be our rational exemplars. Reid believes this sense of ridicule is a gift of nature to help keep us from departing too far from matters of common sense: We may observe, that opinion which contradict first principles are distinguished from other errors by this; that they are not only false, but absurd: And, to discountenance absurdity, Nature hath given us a

Some Objections to Phenomenal Conservatism  83 particular emotion, to wit, that of ridicule, which seems intended for this very purpose of putting out of countenance what is absurd, either in opinion or practice. (Reid [1785] 2002, 462) Certainly, ridicule is not an infallible guide, but, in this case, I think we would be wise to let the ridiculousness of Tom sitting there, wide-​eyed, stupefied by skepticism guide our judgment. If the alternative to the common sense approach is lunacy, then by all means, “let my soul dwell with Common Sense” (Reid [1764] 1997, 18). Still, there might seem something less than fully satisfying about the common sense way of replying to skepticism. It seems nonchalant. It’s too easy. Baron Reed (not to be confused with Thomas Reid) offers the following criticism: At this point, no doubt, many anti-​skeptics will be protesting loudly that it just is not true that we should have skeptical doubts. The skeptic’s reply is simple: why not? If there is a reason why we should not have those doubts, let’s hear it. If such a reason could be provided, of course, the common sense approach wouldn’t have been needed in the first place. If, on the other hand, there is no reason why we shouldn’t have skeptical doubts, then the refusal to come to grips with skepticism is entirely ungrounded. Common sense, if this is what it amounts to, is not a way of doing philosophy but a way of not doing it. (Reed 2006, 192) Reed’s central worry seems to be that proponents of common sense cannot justify their claim that the possibilities of error raised by skeptical scenarios are not serious possibilities of error. Why aren’t they to be taken seriously? Without such a defense, Reed says that this common sense approach to skepticism is “completely ungrounded.” As indicated above, the man of common sense can argue that skeptical scenarios are not serious in at least two ways. First, he can point to the implausibility of such scenarios given his ordinary background evidence. For instance, channeling Moore, Tom might point out that he has hands, and if he has hands, then he isn’t a brain in a vat. Second, the man of common sense can also point to his anti-​skeptical epistemic standards, which say that possibilities of error must be treated as serious only if they have something positive going for them (and not as a default). To further substantiate these anti-​skeptical standards, our common sense fellow might point to his large collection of justified beliefs about the world—​justified beliefs that he would not have were skeptical scenarios the sort of things that must always be taken seriously.8

84  Some Objections to Phenomenal Conservatism Of course, Reed would likely consider such arguments are illegitimate. The arguments cited in favor of the common sense approach are ones we are justified in believing only if the proponent of common sense is correct. But why would that make those arguments illegitimate? You might accuse the arguments of begging the question, but that isn’t quite right. To beg the question, one must only be justified in believing the premises if one is already justified in believing the conclusion. But given PC, you don’t have to have justification for believing anything about the non-​seriousness of skeptical scenarios (the conclusion) in order to justifiably believe that you have hands, or that you have widespread knowledge about the world (the premises). Perhaps the objection is that we are only justified in believing the arguments if the common sense standard is correct. But if the common sense standard is in fact correct, then this is no objection at all. Hence, the objection presumes the falsity of the common sense position. Now, plausibly some presumption of standards is unavoidable here—​that is, there is no neutral vantage point from which to argue that does not presume the standards of one side or the other—​but then you can hardly blame the man of common sense for remaining unpersuaded by objections that operate according to the very standards he thinks are evidently false (especially if you refuse to blame the skeptic for doing exactly the same thing). Thus, if the accusation that these arguments are illegitimate is to gain any traction with the proponent of common sense, they must be framed as an objection to using those arguments in a certain dialectical context. For the skeptic does not accept that standard and so would not accept the argument. In other words, the charges laid against these defenses of the common sense approach are not that they do not justify their conclusions. Substantiating those charges would require reasons for thinking that the common sense standards underlying such arguments are false, which Reed does not offer (and which skeptics cannot consistently offer). Rather, the real complaint is that these defenses of common sense do not justify their conclusions in a way that would satisfy the skeptic. But if that is the real complaint, then we should be asking ourselves why convincing the skeptics should be our barometer for success. Bergmann rightly points out that the skeptic does not treat persuading the man of common sense as the barometer of his success, since he does not attempt to give arguments that will persuade the fellow committed to common sense standards (Bergmann 2021, 148). Why should the man of common sense be obliged to extend a courtesy that is not provided to him? Often, our goal is to give an argument that would persuade those who disagree with us, but not always. Sometimes we seek reflective-​assurance rather than persuasion. And there is reason to think that the former is all we can hope for here. After all, Reid maintains that it is impossible to convince those skeptics who faithfully adhere to their own skeptical standards:

Some Objections to Phenomenal Conservatism  85 If a Sceptic should build his scepticism upon this foundation, that all our reasoning and judging powers are fallacious in their nature, or should resolve at least to with-​hold assent until it be proved that they are not; it would be impossible by argument to beat him out of this strong hold, and he must even be left to enjoy his scepticism. (Reid [1785] 2002, 480) The reason this is impossible is because we cannot dismiss the possibilities of error facing one belief without relying on other beliefs; but those other beliefs will inevitably face their own possibilities of error, and so cannot be relied on until those possibilities are dismissed. This further dismissal will require still further beliefs, which face their own possibilities of error, and so the regress continues—​justification always deferred and never attained. But if the skeptics have adopted standards that make them impossible to persuade, then it is a fool’s errand to uphold their persuasion as our standard of success. If there were strong arguments in favor of the skeptical standard, then perhaps we would be forced into the vain attempt. But, to the man of common sense, the intuitive appeal of this skeptical standard pales in comparison to what accepting that standard would force him to give up. Indeed, let us flip Reed’s question around on the skeptic: what reason can be given for his skeptical standard? Why should mere possibilities of error, with nothing to back them, raise doubts about our beliefs? The skeptic cannot appeal to intuition for the reason just mentioned: there are things far more intuitive (such as our having many justified beliefs about the world) that conflict with this standard. Indeed, the skeptic has, by his own lights, absolutely no hope of giving a satisfactory answer to this question. For we have already seen that his skeptical standard makes it impossible for him to justify belief in anything at all, including that very standard. Thus, if any position is, as Reed says, “entirely ungrounded,” it is that of the skeptic. What about Reed’s final charge that abiding by the common sense standard is “not a way of doing philosophy but of not doing it”? To the man of common sense, this isn’t an objection. It’s his position! It isn’t the job of philosophy to grant authority to common sense judgments. Common sense doesn’t need philosophy. To the contrary, it is philosophy that needs common sense. To insist that common sense judgments cannot be justifiably believed until philosophy has independently proven them reliable gets things backward.9 Reid writes: [T]‌he votaries of this Philosophy, from a natural prejudice in her favour, have endeavoured to extend her jurisdiction beyond its just limits, and to call to her bar the dictates of Common Sense. But these decline this

86  Some Objections to Phenomenal Conservatism jurisdiction; they disdain the trial of reasoning, and disown its authority; they neither claim its aid, nor dread its attacks. In this unequal contest betwixt Common Sense and Philosophy, the latter will always come off both with dishonour and loss; nor can she ever thrive till this rivalship is dropt, these incroachments given up, and a cordial friendship restored: for, in reality, Common Sense holds nothing of Philosophy, nor needs her aid. But, on the other hand, Philosophy (if I may be permitted to change the metaphor) has no other root but the principles of Common Sense; it grows out of them, and draws its nourishment from them: severed from this root, its honours wither, its sap is dried up, it dies and rots. (Reid [1764] 1997, 19) Reid’s understanding of the relationship between philosophy and common sense is not beyond dispute. Nevertheless, its opponents must do more than simply rephrase that methodology in the form of an accusation. Something similar is true of many other complaints made against the common sense approach. For instance, some accuse the man of common sense of being dogmatic. Apart from some substantive critique, this charge amounts to little more than name-​calling. And the proponent of common sense can play that game as well. It is the skeptic who is being dogmatic. Indeed, these charges stick much better against the skeptic than they do against the man of common sense. After all, the skeptic is the one clinging to his demanding epistemic standards, even when those standards cut against things far more evident (such as the fact that we have a lot of justified beliefs about the world). Another complaint is that the man of common sense fails to take skepticism seriously.10 Well, skepticism should be treated with the seriousness that it deserves, and the skeptic could be accused of taking skepticism far too seriously just as readily as the proponent of common sense could be accused of not taking it seriously enough. Which objection proves the better cannot be decided apart from adjudicating their more fundamental dispute about which standard to adopt. And at least the man of common sense is, by his own lights, on the right side of this dispute. The consistent skeptic cannot claim even this. From the perspective of a common sense fellow, then, there is really no convincing reason to abandon his common sense approach. The objections to his approach are only good if the skeptic’s standard is correct, but he can see no convincing reasons for those standards—​indeed, he thinks he can see that, by those very standards, there can’t be good reasons for them—​and, conversely, he takes there to be many good reasons against those skeptical standards. What is this fellow to do but continue believing in accordance with what is most evident to him? I, for one, cannot blame him for it.

Some Objections to Phenomenal Conservatism  87 3.2  Bootstrapping and the Problem of Easy Knowledge In defending this common sense reply to skepticism, I have allowed us to verify the trustworthiness of one seeming on the basis of others formed in much the same way. For instance, Tom dismissed a skeptical challenge to his perceptual seeming by appealing to background beliefs that were themselves based on perceptual seemings. In this way, Tom could verify the reliability of his perceptual belief-​forming methods by relying on those very same methods. Is this not an instance of circular reasoning? And isn’t it a problem if PC allows for it? Indeed, some have objected to PC on precisely this basis.11 A prominent version of this objection is the bootstrapping problem (Fumerton 1995, Vogel 2000), which is an instance of what Stewart Cohen calls “the problem of easy knowledge” (Cohen 2002; 2010). As applied to PC, the basic idea is that we can too easily verify the trustworthiness of our seemings. For instance, Cohen says we could confirm the trustworthiness of our color seemings simply by looking at a series of color cards and reasoning, “This card seems red; and wouldn’t you know it is red (a belief justified immediately by my perceptual seeming); and likewise for all the other cards; thus, my color vision is accurate.” We could also strengthen this confirmation simply by performing the experiment over and over again. However, Cohen and others consider it evident that such bootstrapping cannot justify belief in the accuracy of our color vision (Cohen 2002, 318). The examples used to press this objection are sometimes flawed. For instance, ordinary adults in Cohen’s color card example will not receive much if any support for the reliability of their color vision, even given PC. This because we have special reasons to question the helpfulness of that experiment. We know that color blindness is common enough to be a serious possibility, and that this procedure does not protect against it. Our knowledge of this possibility places limits on how much support for our color vision, if any, we can gain through such an experiment. Likewise, repeating the same experiment will not result in greater and greater evidence of reliability because we know that the tests are not independent of one another. So, we must be careful to judge PC only by those forms of circular reasoning that it actually condones, such as Tom’s reliance on perceptual seemings to verify the reliability of our perceptual faculties. My response could be categorized as a form of biting the bullet: concede that PC allows for bootstrapping but argue that such bootstrapping is legitimate and unavoidable. Alternatively, if we define bootstrapping more narrowly, as a specific form of impermissible reasoning, then my position is that PC licenses a certain kind of circular reasoning, but not one that counts as bootstrapping per se. Without an exact definition of bootstrapping, it’s hard to say which is more accurate (compare with Cohen 2002,

88  Some Objections to Phenomenal Conservatism 321). In either case, I’ll be defending the permissibility of at least one kind of circular reasoning. Obviously, it does not follow that just any form of circular reasoning is permissible. To use Bergmann’s terminology, some forms of circularity are malignant, others benign (Bergmann 2004; 2006). So, what kinds of circularity are there? There is, first of all, a division between psychological and logical circularity. Psychological circularity is when the mere fact that we already believe something leads us to entrench, strengthen, or otherwise continue to hold belief in that thing. An explicit instance of this would be to conclude that p must be true on the basis of an argument in which p features among the premises. Surely this is an altogether rare occurrence. It is not uncommon, however, for the belief that p to affect our attitudes in more implicit ways, such that our belief in p ends up bolstered merely because we came into the situation already believing it. Logical circularity, on the other hand, is when our justification for believing p is supposed to be derived in a way that relies on our already having justification for believing p.12 Again, an explicit instance of this would be for us to claim justification for believing p on the basis of an argument in which p is a premise. Other arguments also beg the question—​ meaning that their premises can be justifiably believed only if and precisely to the extent that one already has justification for the conclusion—​but do so in more subtle, implicit ways. These are the ones we are more liable to fall for. PC does not commit us to either psychological or logical circularity in either its explicit or implicit forms. Psychologically, the seeming that p is capable of triggering belief that p without any belief that p lurking in the background. We do not have to believe that a seeming is reliable, for instance, before we are inclined to trust it. As a psychological matter, the phenomenal character of the seeming itself is sufficient to secure our trust.13 Logically, the situation is parallel. According to PC, the justificatory potential of seemings does not depend on having prior justification for believing that those seemings are reliable. That’s the whole conservative aspect of phenomenal conservatism. These are the two most troubling kinds of circularity, but PC is guilty of neither. What PC does allow us to do is to implicitly trust our seemings in the process of supporting their trustworthiness. This is not psychological circularity in that it does not require us to implicitly believe that our seemings are trustworthy. (The toddler trusts his mother by believing what she says, not by believing that she is reliable.) Nor is this logical circularity in that, given PC, trust is appropriate even without verification of reliability. Nevertheless, we are treating a certain category of things as trustworthy in proving their trustworthiness. We are, as Reid would say, “taking their own testimony for their veracity” (Reid [1785] 2002, 481). Let’s call this “source circularity.”14

Some Objections to Phenomenal Conservatism  89 Now that we have hit upon the kind of circularity allowed by PC we must ask: is source circularity vicious? My answer is the same as Reid’s: “this we must do implicitly, until God give us new faculties to sit in judgment upon the old.” Such circularity cannot be entirely avoided, unless we adopt seemingly ad hoc limits on the justification of reflective beliefs (see, e.g., Cohen 2002, 326–​327).15 For if we are allowed to go about justifying reflective beliefs as we do any other—​and I cannot see why this area should be subject to special limitations—​then such beliefs will inevitably display source circularity. And this is true whatever our foundational sources of justification turn out to be. Thus, there can be no objection to PC simply because it allows for source circularity. But is source circularity truly unavoidable? One could try to substantiate trust in one source on the basis of another source. For instance, one could potentially justify belief in the trustworthiness of perception through a priori reasoning. But what are we to say about the trustworthiness of a priori reasoning?16 When we get down to our foundational sources of justification, whatever they may be, then how are we to justify belief in the trustworthiness of those very sources? Any justified belief we form must ultimately derive its justification from them. Thus, any belief we form as to whether those foundational sources are trustworthy will inevitably display source circularity.17 This shouldn’t be surprising. Any attempt to prove that our foundational sources are trustworthy must proceed from those same sources. How could it be otherwise?18 Foley concurs: After all, the search for assurances will itself require us to marshal our cognitive resources. It will itself involve the use of methods about which we can sensibly have doubts, doubts that cannot be addressed without begging the question. Any attempt to address them will employ methods either that are themselves already at issue or that can be made so. … There is no intellectually neutral position from which to defend our most fundamental intellectual positions. (Foley 1993, 61) Accordingly, source circularity must be permitted, so long as we are to allow beliefs about the trustworthiness of foundational sources to be justified at all. Not just any potential source of justification can be verified through source circular reasoning, however. For one, not every purported source of justification is genuine. That is, only some purported sources of justification are actually worthy of implicitly trusting, and only those worthy of implicitly trusting can successfully generate justification for believing that they are trustworthy. And even then, there is no guarantee that a genuine source of justification will justify belief in its own trustworthiness. It is not

90  Some Objections to Phenomenal Conservatism trivially true, for instance, that our perceptions align with one another so as to bolster confidence in the reliability of perception. Indeed, we sometimes learn via perception that perception is unreliable in various circumstances. So, to grant the legitimacy of source circularity in certain instances is not to rubber stamp it in any form. Moreover, even when source circularity is legitimate, it still comes with significant limitations. Reid, for instance, recognizes the pointlessness of using source circular reasoning to dismiss doubts about the trustworthiness of that source: If a man’s honesty were called into question, it would be ridiculous to refer it to the man’s own word, whether he be honest or not. The same absurdity there is in attempting to prove, by any kind of reasoning, probable or demonstrative, that our reasoning is not fallacious, since the very point in question is, whether reasoning may be trusted. (Reid [1785] 2002, 480) He then points to Descartes as having perpetrated this sort of error: Descartes certainly made a false step in this matter; for having suggested this doubt among others, that whatever evidence he might have from his consciousness, his sense, his memory, or his reason; yet possibly some malignant being had given him those faculties on purpose to impose upon him; and therefore, they are not to be trusted without a proper voucher: to remove this doubt, he endeavours to prove the being of a Diety who is no deceiver; whence he concludes, that the faculties he had given him are true and worthy to be trusted. It is strange that so acute a reasoner did not perceive, that in this reasoning there is evidently a begging of the question. For even if our faculties be fallacious, why may they not deceive us in this reasoning as well as in others? And if they are to be trusted in this instance without a voucher, why not in others? (Reid [1785] 2002, 480–​481) Reid’s point here is not to criticize source circular reasoning in and of itself (contra Lemos 2004, 76–​84). His point is that it is ridiculous to use source circular reasoning to convince someone who has or should have doubts about that source or who refuses to trust it without “a proper voucher.” Source circular reasoning only works for those who already trust that source implicitly and have no reason not to continue doing so. Thus, Reid criticizes Descartes not for giving an argument for the veracity of his faculties per se, but for doing so under the guise of incredulity; for feigning doubts about the veracity of clear and distinct perceptions and then using

Some Objections to Phenomenal Conservatism  91 clear and distinct perceptions to assuage those doubts; for claiming he trusts nothing without first verifying it and then reasoning his way to the veracity of reason.19 Source circular reasoning can only verify the veracity of reason if there were never any doubts to begin with.20 Recognizing these limitations on source circular reasoning can help us address some of the seemingly problematic cases that PC and other principles are accused of condoning. Frequently, these are situations in which there are or should be doubts about the trustworthiness of a particular source—​a serious possibility of error remains uneliminated. Accordingly, it is not permitted to dismiss those doubts while relying on that very source. On the other hand, raising a question about the trustworthiness of a source, even if that question goes unanswered, does not automatically give one good reason to doubt it. That is because the possibilities of error raised by that question may not be serious enough to sustain doubt. According to the common sense reply to skepticism, this describes the situation when extreme skeptical scenarios are brought up. As defended in §3.1.2, the common sense standard allows that such skeptical scenarios do not have sufficient plausibility to qualify as serious possibilities of error, and so do not justify doubts in the trustworthiness of our faculties. Thus, such people can justly rely on source circular reasoning to dismiss those scenarios. In sum, PC doesn’t permit just any kind of circular reasoning. The most problematic kinds of circularity receive no endorsement from PC. PC does allow for source circularity, but so do all theories on which there are foundational sources of immediate justification and on which reflective beliefs are not ad hoc prohibited. Moreover, this source circular reasoning is only allowed in specific instances and with limited effect. In particular, a source cannot be relied on to dismiss legitimate doubts about the reliability of that very source. Therefore, the problems of bootstrapping and easy knowledge, while instructive, do not provide strong reasons to reject PC. 3.3  Does PC Justify Bizarre Beliefs? There is a temptation to abuse principles like PC by wielding them in overly simplistic fashion. For example, if we are trying to explain the justification of moral or religious beliefs, just posit moral or religious seemings, then poof they’re justified! But the story cannot be so simple. Toddlers may be justified in believing things on the basis of isolated seemings, but mature adults, who have doubtlessly learned something about which sorts of seemings are and are not reliable, generally must find verification for their seemings within a more comprehensive system of beliefs if they are to continue trusting them. They must, for example, have some reason to think that these moral or religious seemings are genuine revelations rather than

92  Some Objections to Phenomenal Conservatism mere whims. The failure to recognize this fact gives rise to one of the most common objections to PC: that it is too liberal in doling out justification. Seemings are a dime a dozen, after all. If all it took to be justified was to have a seeming, then practically anybody could be justified in believing practically anything, or so the worry goes. There are already answers to this objection in the literature (e.g., Gage and McAllister 2020, Huemer 2023, Tucker 2010), so I will be succinct in my response. The advantage of briefly retracing some of this territory is to draw out a couple of important points, easily missed, about the application of PC to ordinary contexts. Objectors usually illustrate their concerns by proposing an example of some bizarre belief that they claim is justified on the basis of PC. For instance, a paranormal enthusiast might be justified in believing that she is encountering a ghost simply because it seems that there is one in front of her. The first response to such examples is to point out that there are unacknowledged defeaters which prevent the bizarre belief from being justified. Indeed, part of the reason we consider them bizarre is precisely because they run contrary to the kinds of beliefs that we ordinarily take ourselves to be capable of justifying. For instance, ordinary adults have multiple defeaters for ghost seemings. This will include not only rebutting defeaters—​there are widespread reasons for believing that ghosts don’t exist or, if they do, rarely show themselves—​but also undermining defeaters. Ordinary adults have copious evidence about which fundamental sources of information are available to humans as well as when those sources are reliable. So much in fact that they are justified in taking themselves to have a fairly complete view of the situation. The result is that if a seeming arises from something other than the widespread and thoroughly vetted sources familiar to us all (things like perception, memory, reason, introspection, and perhaps a few others), then that will right away constitute a red flag. This red flag alone will often be enough to greatly diminish if not entirely defeat the justificatory support provided by that seeming.21 Note that we do not need to consciously register the presence of this undermining defeater. Merely possessing it within our background evidence is enough to undermine whatever justification would otherwise come from that seeming. The upshot is that unverified seemings—​seemings that lack a set of background evidence vouching for the veracity of their source—​will rarely provide any ultimate justification for ordinary adults,22 and all of the bizarre seemings put forward by objectors are unverified in this sense. (That’s part of what makes them bizarre.) Hence, bizarre seemings will ordinarily contribute little if any to the justification of belief. On the other hand, if the objector wishes to consider extraordinary agents whose evidence differs significantly from that of ordinary adults—​ in particular, agents who lack any reasons for doubting their bizarre seemings—​then PC would allow those bizarre seemings to justify belief.

Some Objections to Phenomenal Conservatism  93 However, this is only possible because such beliefs are not bizarre to those agents at all! They do not have reasons for doubting them as we do. Returning to our ghost example, for the subject to lack any serious reasons for doubting her ghost seemings she would either have to have very little evidence about which sources of information are available to humans, like an infant, or else her background evidence would have to differ radically from that of the ordinary adult—​say she was raised in an insular cult in which ghost seemings were always treated as normal and reliable for those who had honed their abilities. In the latter case, the agent would actually have evidence for the veracity of his ghost seemings. How could such seemings not strengthen the agent’s epistemic position? Any hesitation to admit this comes from assessing the agent’s belief from a perspective other than her own. But given internalism about justification (defended in Chapter 1), it is the latter that counts. Alternatively, if the extraordinary agent we’re considering is something more like an infant, and is somehow ignorant of what sources of information humans possess, then how are paranormal seemings about the presence of ghosts any different from perceptual seemings about the presence of other humans? We were not aware of the reliability of perceptual seemings until after we had already trusted those seemings. In fact, we almost certainly relied on perceptual seemings in order to build up evidence for the reliability of perceptual seemings. (For more on this sort of circularity, see §3.2.) If we revert to that early stage in which we have no evidence about the reliability of either kind of seeming—​perceptual or paranormal—​then, from the subject’s perspective, the two really are on equal footing. But plausibly the perceptual seemings immediately justify their content in those conditions, so plausibly the paranormal seemings do as well. I defend this last bit of reasoning at length in the latter parts of this inquiry. For now, the essential point is just that, for extraordinary agents, it is not at all evident that their bizarre seemings are incapable of justifying belief, in which case the objector has failed to establish a problem for PC. To be sure, it remains evident that something is wrong with bizarre beliefs. Part of why we consider them bizarre is because we know them to result from processes that are unreliable, or to involve malfunction, or to possess certain other epistemic defects. Accordingly, there are important positive epistemic statuses that these beliefs lack—​people often cite warrant as an example (see Tucker 2011, §5). We must remind ourselves, however, that PC concerns justification, and that justification does not require reliability or proper function or the like. To be sure, PC need not deny a close relationship between justification and statuses like warrant. It is plausible that having justification for a belief is necessary for that belief to be warranted. But the former should not be conflated with, or understood as sufficient for, the latter. So, even in the extraordinary cases where PC

94  Some Objections to Phenomenal Conservatism allows bizarre beliefs to be justified, PC does not prevent us from acknowledging the many other epistemic deficiencies of those beliefs—​deficiencies that plausibly explain why we might continue to feel unsettled about them. 3.4  Double-​Counting A final concern for PC is Michael Tooley’s “double counting objection” (Tooley 2013, 322–​323). Tucker illustrates Tooley’s objection with the following example (Tucker 2013, 13–​14). Say you have a good argument for p that provides some level of justification for believing p. Then, because of that argument and for no other reason, it begins to seem to you that p. If PC is correct, your seeming provides foundational evidence for p which, when combined with the inferential support of the original argument, ostensibly results in a “higher level of justification” for p (Tooley 2013, 323). However, this gain seems illegitimate. PC mistakenly allows the force of the argument to be double counted. To explain my response, let us return to a testimonial analogy. Imagine that a meteorologist named Mary examines the raw data coming from the National Weather Service and concludes that it will rain tomorrow (call this “p”). She asks her friend, another meteorologist, what he thinks, and he testifies that it will rain tomorrow. It turns out, however, that Mary’s friend did not look at the data himself. Rather, he based his conclusion solely on Mary’s testimony. He was, in effect, feeding Mary’s own testimony right back to her. This is analogous to the original seemings case, in which the seeming that p (like the friend’s testimony) is not independent of the subject’s own discursive reasoning. Looking at this testimonial case will, I believe, provide important lessons for seemings case as well. If we assume a Reidian view to testimony, does the friend’s testimony provide Mary with a higher level of justification for p? The answer is not straightforward. For one, we need to consider what it means for Mary to have a higher level of justification for p. This could designate that Mary is permitted to place a higher credence in p than she could before receiving the testimony, or perhaps that she is now justified in believing p when she was not before. Alternatively, it could refer to a general strengthening of Mary’s epistemic position with respect to p. Such a general strengthening does not necessarily justify Mary in forming new and stronger attitudes toward p. For instance, say that her own calculations, as well as her friend’s testimony, if taken individually, each permit Mary to place a credence of .8 in p. There is no guarantee that those two sources of justification, if taken together, permit a higher credence in p. Justification is not additive in that sort of way. Instead, her friend’s testimony may strengthen her epistemic position with respect to p simply by, say, making her justification for p more resilient to defeaters. So, adding her friend’s testimony

Some Objections to Phenomenal Conservatism  95 necessarily increases Mary’s level of justification for p in the second sense (“strengthening it”), but not in the first sense (“elevating it”). Focusing on this second sense, the Reidian view says that Mary’s justification is initially strengthened by her friend’s testimony, but it may not be ultimately strengthened by it. This is because the justificatory support lent by that testimony can be defeated by further considerations. Learning that her friend’s testimony is not independent of her own is one such defeating consideration. The dependence of her friend’s testimony is not a rebutting or undermining defeater but a supplanting one. Her friend’s testimony, being wholly derivative, becomes irrelevant. On the other hand, if Mary is unaware that her friend’s testimony depends on her own, then Mary’s justification is ultimately strengthened by that testimony. But that strengthening does not seem illegitimate at all given Mary’s ignorance of the situation. From her point of view, there is no double counting going on; rather, she has received confirmation of her judgment from another expert, and her position is strengthened accordingly. Apply these lessons to a case with seemings. Say Mary looks at the weather data and, through careful reasoning, infers that it’s going to rain tomorrow (p). As a result of these calculations, it begins to seem to her that it will rain tomorrow. Does Mary’s seeming increase her justification for believing this proposition? According to PC, Mary’s seeming is foundational evidence for p. Thus, Mary’s seeming will initially strengthen Mary’s position with respect to p. As discussed above, this strengthening does not guarantee that Mary is permitted to take a stronger attitude toward p than before—​her epistemic position may or may not be elevated. The seeming will, however, strengthen Mary’s epistemic position with respect to p in some regard, at least initially. The ultimate effect of that seeming will depend on what defeaters Mary has. In particular, if Mary knows that her seeming is dependent on her discursive reasoning, then she will to that extent possess a supplanting defeater for her seeming (Huemer 2013, 339–​ 341). PC thereby avoids any double counting. Of course, in ordinary situations, there will usually be some question as to how independent a seeming is from one’s explicit calculations. Cognitive science tells us that the processes producing intuitions are often separate from the processes behind discursive reasoning. For instance, Mary’s intuitions about the weather generally won’t always be mere regurgitations of conclusions reached through rational calculation. Often they will arise from separate, unconscious cognitive processes that could well be picking up on unique aspects of the data. Many intuitive processes are less reliable than their discursive counterparts; others are more so. For instance, standardized test takers are generally advised to “go with their gut” over the conclusions they reach through explicit reasoning; and chess grandmasters often intuit the best move more reliably than they can explicitly calculate

96  Some Objections to Phenomenal Conservatism it. In Mary’s specific case, it is plausible that her weather intuitions, honed through long experience as a meteorologist, are just as or more reliable than her rational calculations. Reliability aside, the crucial point is just that we could reasonably treat Mary’s intuition and her calculations as (partially) independent sources.23 In which case, Mary will receive a boost when the conclusion of her argument begins to seem true. But again, this isn’t double counting because as far as Mary can tell those sources are at least somewhat independent. Thus, PC doesn’t have a problem with double counting. To the extent that we are aware of the dependence of our seemings on our explicit rational calculations, the justificatory support provided by those seemings is supplanted. On the other hand, PC correctly predicts that our seemings do provide additional support when there is no indication of such dependence. None of the objections surveyed constitutes a major problem for PC. That is to say that the proponent of PC has plausible things to say in response to each of these concerns. This paves the way for my positive arguments in later chapters. Addressing these objections should also have left us with a more thorough understanding of PC and seemings foundationalism, and what some of the implications of these views might be for epistemology more generally. Really, there is only one aspect of seemings foundationalism that is still in dire need of clarification. As of yet, I have said little about the mental states that feature so prominently within this framework. What are seemings exactly? In the next chapter, I present a detailed explication of seemings—​of what they are and what they are not. Afterward, our focus will pivot from understanding seemings foundationalism to whether seemings foundationalism is true. Notes 1 Huemer (2023), Moretti (2015), Tucker (2013), and others have already done an excellent job surveying the literature on phenomenal conservatism, including charting the various back-​and-​forths between its objectors and defenders. 2 Indeed, Moretti’s criticism might be better understood as an objection to the common sense reply to skepticism in general than to PC proper. That is because his considerations apply just as much to any internalist version of that common sense reply, be it supported by phenomenal conservatism, epistemic conservatism, or some other principle of immediate justification. For example, consider a version of epistemic conservatism, which says that some basic beliefs are immediately justified simply because we possess them. What happens when we reflect on one of those basic beliefs and inquire into its reliability? Either we do not have any independent verification of its truth, which results in the loss of our immediate justification, or else we do, in which case that verification supplants whatever immediate justification belonged to that basic belief.

Some Objections to Phenomenal Conservatism  97 3 I see no reason why Tom’s belief cannot be supported by both the immediate justification provided by his perceptual seeming and the inferential justification provided by his ordinary background beliefs. To be clear, there are some cases in which, upon reflection, the inferential support fully supplants whatever immediate justification comes from the seeming. However, in other cases, it seems that the immediate justification is not supplanted but supplemented (or perhaps both occur in different measures). We will consider this in more detail below and in our discussion of double-​counting in §3.4. 4 Moretti maintains that one needs to appeal to “independent evidence” to rule out deception hypotheses. Do these ordinary background beliefs qualify? They are clearly independent in the sense that they are not based on the specific seeming under consideration. However, they do (we are assuming) rely on other seemings that could be subjected to the same kinds of doubts as the first. If this disqualifies them from being “independent” in the relevant sense, then I would simply deny that independent evidence is required. As I will argue in Chapters 6–​8, it isn’t possible to make a rationally compelling argument that isn’t ultimately based on how things seem. Thus, to accept the demand for independent evidence, so understood, is to give way to complete and total skepticism. 5 The situation described should bear more than a passing resemblance to Descartes’s doubts about clear and distinct principles of reason as raised, for instance, in Meditations 1 and 3. 6 It would be nice to have a clear way of differentiating those possibilities of error that are serious from those that are not. Following Bergmann, we could define serious possibilities of error as those that should make us withhold or disbelieve the proposition at issue (Bergmann 2006, 198), but that seems more to label rather than locate what we we’re looking for. I won’t give a theory here except to say that, to be serious, the error must at least rise to the level of being plausible for the subject (i.e., the subject is justified in placing some non-​negligible level of confidence in the belief being erroneous in that way). 7 This is why I think, in a situation like Tom’s, the immediate justification provided by the original seeming might be supplemented rather supplanted by inferential verification. The original seeming can continue to lend immediate support to the belief despite the possibility of error being raised. This immediate support would then be joined, not replaced, by the inferential verification provided by surrounding considerations. This does not always occur—​sometimes the immediate justification is entirely supplanted—​but it can. 8 This is to adopt a particularist methodology. See Chisholm (1973) on the general idea of particularism and Bergmann (2021) on its application to precisely these issues. 9 It is, we might say, to put Descartes before the horse. 10 See also the discussion in Bergmann (2021, 147ff). 11 It should be noted, however, that the problem is not with anything specific to PC but arises for any principle that allows for “justified beliefs that are obtained without prior justification for believing that the processes that produces [sic] them are reliable” (Cohen 2010, 141).

98  Some Objections to Phenomenal Conservatism 12 I am following the convention in calling this “logical circularity,” but I think it better called “epistemic circularity” since it pertains to circularity in the order of justification rather than circularity within a logical system per se. Alas, the term “epistemic circularity” has already been co-​opted for what I will call “source circularity” below. 13 See the discussion of belief-​formation in Chapter 1, §1.4.1. 14 This kind of circularity is typically called “epistemic circularity” (following Bergmann 2004; 2006; 2021), but I think this descriptor is less than ideal. The label “epistemic circularity” suggests that the justification of one’s belief is derived in a circular manner—​that A derives its justification from B, which derives its justification from A somewhere down the line. But that is decidedly not what’s going on here. (That’s logical circularity, as described above.) The beliefs produced by our faculties are immediately justified and require no reflective beliefs in the reliability of the faculties that produced them or justification for the same. Circularity only enters the picture later on, after our ordinary beliefs have already been justified, when we use those beliefs to form a judgment on whether the faculties that produced them are reliable. The term “source circularity” could also be misunderstood, I suppose, but I find it at least somewhat less susceptible to confusion. 15 Furthermore, if we do place ad hoc limits on the justification of reflective beliefs, then PC can work within these limitations just as well as any other theory. 16 See Hales (2000) on justifying the use of a priori rational intuition through a priori rational intuition. 17 It might be possible, I suppose, for one foundational source FS1 to verify another foundational source FS2, and vice versa, and so to avoid source circularity in the most direct sense. But for humans, there are serious reasons to doubt that this will work: we might be able to verify the trustworthiness of perception without relying on perception itself (though see Alston 1991, ­chapter 3), but how could we ever hope to verify reason without relying on reason? And even if we did manage this, it would be cold comfort. We would still be relying on a set of sources that, as a whole, are unverified except by the members within that set. Hence, circularity is still present. Another logical possibility is for there to be an infinite chain of foundational sources, each verified by another. But, in addition to the many theoretical problems such a view would face, I think it is a safe assumption that the number of human sources of justification is finite. 18 Even coherentism, which denies foundational sources of justification as traditionally conceived, will face a problem here. Coherentism says that beliefs are justified by cohering within a sufficiently large and interconnected system of beliefs. Call beliefs that satisfy this condition “coherent beliefs.” How can we justifiably believe that coherent beliefs are reliably formed? We can only do so if the belief that coherent beliefs are reliably formed is itself a coherent belief. But this is just to verify the trustworthiness of coherence through reliance on coherence—​it’s a kind of source circularity once more. 19 I don’t agree with Reid that Descartes is (completely) guilty of these things, but this is undoubtably a common interpretation of Descartes. And it is Descartes as commonly interpreted that Reid is criticizing.

Some Objections to Phenomenal Conservatism  99 20 See Bergmann’s distinction between QD-​situations—​that is, “Situations where, prior to the [source circular] belief’s formation, the subject is or should be seriously questioning or doubting the trustworthiness of [the relevant source]”—​ and Non-​QD-​situations (Bergmann 2006, 198). Following Reid, Bergmann proposes that source circularity (or what he calls “epistemic circularity”) is vicious in QD-​situations but not in Non-​QD-​situations. 21 If that seeming is extraordinarily strong—​so strong that it is nearly impossible not to lend some credence to it—​then that may constitute an exception. But such seemings are a rare occurrence outside of the familiar foundational sources listed above. 22 This will not be true for infants, however, who are just starting to build up their belief systems. Such infants are not justified in taking themselves to have a good grasp on which sources of information are available to them. Thus, when a seeming arises and cannot be quickly traced back to some verified source, this does not raise a red flag for infants as it does for ordinary adults. 23 They are not entirely independent, of course, in that they are both operating on a similar data set. Compare this to the case where Mary’s colleague looks at the same data, reaches his own judgment, and then gives his testimony.

References Alston, William P. 1991. Perceiving God: The Epistemology of Religious Experience. Ithaca, NY: Cornell University Press. Bergmann, Michael. 2004. “Epistemic Circularity: Malignant and Benign.” Philosophy and Phenomenological Research 69 (3): 709–​727. —​—​—​. 2006. Justification without Awareness: A Defense of Epistemic Externalism. Oxford: Oxford University Press. —​—​—​. 2021. Radical Skepticism and Epistemic Intuition. Oxford: Oxford University Press. Chisholm, Roderick M. 1973. The Problem of the Criterion. Milwaukee: Marquette University Press. Cohen, Stewart. 2002. “Basic Knowledge and the Problem of Easy Knowledge.” Philosophy and Phenomenological Research 65 (2): 309–​329. —​—​—​. 2010. “Bootstrapping, Defeasible Reasoning, and a Priori Justification.” Philosophical Perspectives 24 (1): 141–​159. Foley, Richard. 1993. Working Without a Net: A Study of Egocentric Epistemology. Oxford: Oxford University Press. Fumerton, Richard. 1995. Metaepistemology and Skepticism. Lanham, MD: Rowman & Littlefield. Gage, Logan Paul, and Blake McAllister. 2020. “The Phenomenal Conservative Approach to Religious Epistemology.” In Debating Christian Religious Epistemology: An Introduction to Five Views on the Knowledge of God, edited by John M. DePoe and Tyler McNabb, 61–​81. London: Bloomsbury Academic. Hales, Steven D. 2000. “The Problem of Intuition.” American Philosophical Quarterly 37 (2): 135–​147.

100  Some Objections to Phenomenal Conservatism Huemer, Michael. 2013. “Phenomenal Conservatism Über Alles.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 328–​350. Oxford: Oxford University Press. —​ —​ —​ . 2023. “Phenomenal Conservatism.” In Internet Encyclopedia of Philosophy. https://​iep.utm.edu/​phen-​con/​. Hume, David. [1748] 1999. An Enquiry Concerning Human Understanding, edited by Tom L. Beauchamp. Oxford Philosophical Texts. Oxford: Oxford University Press. Lemos, Noah. 2004. Common Sense: A Contemporary Defense. Cambridge: Cambridge University Press. Lewis, David K. 1996. “Elusive Knowledge.” Australasian Journal of Philosophy 74 (4): 549–​567. McCain, Kevin, and Luca Moretti. 2021. Appearance and Explanation: Phenomenal Explanationism in Epistemology. Oxford: Oxford University Press. Moore, G.E. 1939. “Proof of an External World.” Proceedings of the British Academy 25 (5): 273–​300. Moretti, Luca. 2015. “Phenomenal Conservatism.” Analysis 75 (2): 296–​309. —​ —​ —​ . 2018. “Phenomenal Conservatism and the Problem of Reflective Awareness.” American Philosophical Quarterly 55 (3): 267–​280. —​—​—​. 2020. Seemings and Epistemic Justification: How Appearances Justify Beliefs. New York: Springer. Reed, Baron. 2006. “Epistemic Circularity Squared? Skepticism about Common Sense.” Philosophy and Phenomenological Research 73 (1): 186–​197. Reid, Thomas. [1764] 1997. An Inquiry into the Human Mind on the Principles of Common Sense, edited by Derek R. Brooks. University Park, PA: Pennsylvania State University Press. —​—​—​. [1785] 2002. Essays on the Intellectual Powers of Man, edited by Derek R. Brooks. University Park, PA: Pennsylvania State University Press. Tooley, Michael. 2013. “Michael Huemer and the Principle of Phenomenal Conservatism.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 306–​327. Oxford: Oxford University Press. Tucker, Chris. 2010. “Why Open-​Minded People Should Endorse Dogmatism.” Philosophical Perspectives 24 (1): 529–​545. —​ —​ —​ . 2011. “Phenomenal Conservatism and Evidentialism in Religious Epistemology.” In Evidence and Religious Belief, edited by Kelly James Clark and Raymond J. VanArragon, 52–​73. Oxford: Oxford University Press. —​—​—​. 2013. “Seemings and Justification: An Introduction.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 1–​29. Oxford: Oxford University Press. Vogel, Jonathan. 2000. “Reliabilism Leveled.” Journal of Philosophy 97 (11): 602–​623.

4 What Are Seemings?

Seemings foundationalism says that all justification rests on foundational evidence, and that our foundational evidence consists in all and only our seemings. What are these seemings? I stated before that seemings are representational mental states with propositional content and a forceful phenomenal character. As such they are sui generis states distinct from other states with propositional content like beliefs. This characterization of seemings is an instance of what Tucker calls the Experience View of seemings (Tucker 2013, 3–​7).1 Others construe seemings differently. Tucker identifies three main alternatives to the Experience View. On the Belief View, a seeming that p is a belief that p.2 On the Inclination View, a seeming that p is a conscious inclination to believe p.3 On the Taking-​Evidence View, a seeming that p is a belief or inclination to believe that one either has evidence for p or a mental state displaying the truth of p.4 The purposes of this chapter are two-​fold. First, I will argue against these alternative views and for the Experience View. Each of the alternatives fails to accommodate central aspects of our ordinary, pre-​ theoretical understanding of seemings, whereas the Experience View accommodates them perfectly well. I also address the most serious complaints with the Experience View, the main one being that adherents have failed to give an adequately thorough and systematic description of seeming states so understood. This leads into my second goal, which is to provide a thorough and systematic description, or at least to move us a significant ways in this direction. 4.1  Alternatives to the Experience View To develop my argument, I will assume that we have a fairly accurate pre-​theoretical understanding of when we have seemings and how they function in our mental economy. This is not to say that most people give any significant thought to the nature of seemings. They don’t. (And on the rare occasions when they do, they almost certainly use terms like DOI: 10.4324/9781003281108-5

102  What Are Seemings? “appearances” or “intuitions” rather than “seemings.”) But the same is true of even the most familiar mental states. Few reflect deeply on the nature of beliefs, perceptions, desires, emotions, pleasures and pains, etc., yet they can easily identify whether or not they are in these mental states and can speak informatively about the role that such states play in thought and action. I will assume that seemings are familiar to us in much the same way. We can gain access to this pre-​theoretical understanding of seemings by gauging our intuitions about certain cases, by considering whether certain principles about seemings ring true, or by examining the conversational propriety of seemings-​statements. Reflection on these things may lead us to correct minor deficiencies in our pre-​theoretical conception, but, given my starting assumption, our final theory should remain largely faithful to this pre-​theoretical understanding and the intuitions and principles that arise from it.5 A word of caution as we proceed. The terms “appears” and “seems” can be used in different ways. For instance, sometimes when we say “it seems to me that p” we are simply hedging our assertion. The purpose of this hedge—​like the qualifier “in my humble opinion”—​is not so much to report anything about how things appear as to assert p while signaling intellectual modesty or a respect for dissenting opinions. Other times we use the language of “it seems that p” or “it seems to me that p” to indicate that a body of evidence lends some sufficient level of support to p (McGrath forthcoming). For instance, we might say, “It seems like it will rain,” to report that our evidence provides on-​balance support for this forecast.6 When we talk in this way, we are not reporting the presence of a mental state that we would call “an appearance” or “a seeming.” For instance, if someone were to follow up with the question, “So you have an appearance that it will rain?”, we would answer in the negative. In contrast, there are times when we say “it seems that p” in which we are reporting the presence of a seeming. Situations where, if someone followed up by asking, “So you have an appearance that p?”, we would agree. It is to these last instances—​ the ones in which we use seems-​language to report the presence of a mental state properly called “an appearance” or “a seeming”—​that we should look in order to fix the nature of seemings. Thus, we must be careful to rely only on examples where the usage is clear. In the rest of this section, I will criticize the alternatives to the Experience View, showing how each of them conflicts with our pre-​ theoretical understanding of seemings. The Experience View, in contrast, appears to have no such difficulties. 4.1.1  The Belief View

Let us start with the Belief View, which says:

What Are Seemings?  103 Belief View—​A seeming that p just is a belief that p. The standard objection to this view is that there are cases in which it seems to S that p but S does not believe p.7 The Naïve Comprehension Axiom is a common example.8 Some have reported that this principle seems true to them, and indeed continues to seem true to them even after learning that it is false. These people lack any belief in the axiom, though the axiom still seems true. This possibility is also supported by the felicity of saying, “It certainly seems like the Naïve Comprehension Axiom is true, but I don’t in any way believe it.” Other examples are not hard to find. The conscientious student may withhold assent toward an intuitive principle until all of the evidence is considered: “What you’re saying seems right, but I’m not ready to believe it quite yet.” The skeptic may acknowledge the presence of an appearance without assenting to it: “It certainly seems that way, but things are not always as they seem.” If it can seem that p when one does not believe that p, then a seeming that p cannot be the same as a belief that p.9 I should quickly note that the Experience View has no trouble handling these cases. There is nothing impossible or even odd about having a particular kind of experience with p as its content while not believing p. It is difficult to see what a proponent of the Belief View could say in response to the standard objection. One might try to argue that in these cases the subject does believe, for example, that the Naïve Comprehension Axiom is true; they just believe this and that it is false. In other words, they hold at the same time obviously inconsistent beliefs. This move does not help the proponent of the Belief View, for it commits them to the untenable conclusion that anyone who find the Naïve Comprehension Axiom intuitive yet knows that it is false is being irrational.10 The proponent of the Belief View might turn to Lyons for help, who argues that “the standard argument for distinguishing percepts [i.e., perceptual seemings] from beliefs is actually quite unconvincing upon closer scrutiny” (Lyons 2009, 70; see also Lyons 2013). Given functionalism, Lyons points out that “believing that p is a matter of standing in a certain functional relation to a representation, R, which has the content that p. ... When R has this causal role, it is a belief” (Lyons 2009, 71). Similarly, a representation is a seeming when it plays a certain functional role in the mind. Even if the functional roles of a seeming and a belief are different, there is no reason to think that a single representation, R, cannot play both roles simultaneously, in which case R is both a seeming and a belief.11 In certain cases, R can stand in the functional role of a seeming but not that of a belief. However, Lyons retorts, “the fact that what I believe and what percept I’m having are not necessarily the same does not show that individual percepts are not beliefs, in cases where the two agree.” In other words, the

104  What Are Seemings? standard objection fails to show that seemings are always distinct from beliefs. I grant that the standard objection fails to show this. What I must emphasize, however, is that this does nothing to undermine the standard objection to the Belief View. The Belief View maintains that what it is to be a seeming just is to be a belief—​that is, that all token representations that play the functional role of a seeming are, by virtue of playing that role, also beliefs. The standard objection does show the Belief View, so understood, to be false. Lyons’s insight that some token representations may play the functional role of both seeming and belief casts no doubt on this refutation. Lyons seems well aware of this (though some of his statements can be misread to suggest otherwise). He agrees that the standard objection shows, “The kinds belief and percept are distinct; being a belief is a different thing from being a percept” (Lyons 2009, 72), and he has no problem granting that a seeming can be distinct from any belief (ibid., Chapter 3, footnote 40). So, Lyons does not appear to have any dispute with the standard objection’s takedown of the Belief View, but this is a subtle point that is easily missed. We would do well to clearly distinguish two positions: i The Belief View, which says that all seemings are beliefs (and that those token states are beliefs by virtue of that which makes them seemings). ii The view that no token seeming is a belief. The standard objection, at least as it is featured here and in the related literature, is an objection to (i). It is no surprise then that, when considered as an objection to (ii), the standard objection is “woefully inadequate” (Lyons 2013, 23). It is, however, perfectly adequate to refute the Belief View. Here is another problem for the Belief View. Believing p comes with certain norms and duties that do not come with its seeming to one that p. If Peter believes p, then he should have sufficient evidence for p and properly base his belief on that evidence, and he is (at least sometimes) blameworthy if he does not (Cullison 2010, McCain 2012, 48–​49). But we do not require evidence for p before p seems true. Consider the difference between these two statements: “The argument seems valid, but I don’t have anything (other than my seeming) to support this.” “I believe the argument is valid, but I don’t have anything (other than my belief) to support this.” No one would criticize someone because of the first statement, but many would criticize someone because of the second. To give another example, imagine you are passing the roulette table as the ball starts to roll. For

What Are Seemings?  105 some unknown reason, it seems that it is going to come up red. Obviously, belief that the ball is going to come up red is unjustified here; you would be violating epistemic norms if you were to believe it. On the other hand, you are not violating any epistemic norms simply by its appearing to you that the ball is going to come up red. The best explanation is that something’s seeming to be true does not come with the same norms or carry the same responsibilities as believing it to be true. The difference between seemings and beliefs that we are highlighting here is not that one is governed by different epistemic norms than the other, but rather that seemings don’t appear to be subject to epistemic norms at all. Seemings aren’t the sort of thing that can be justified or unjustified (Huemer 2001, 97–​98, Farley 2017). This is why it doesn’t make sense to talk about having evidence for or against seemings as it does for beliefs (McCain 2012, 48). We can have evidence about seemings and about whether their content is true, but this is different from having evidence for seemings (as one has evidence for a belief). The latter requires that the state be one that is subject to normative assessment—​that it be capable of being justified or unjustified—​and seemings simply don’t fit into that category. This is not to say that people cannot be criticized for having seemings. We might criticize a racist, for instance, for having racist seemings. But here the criticism is not that the seemings are unjustified, strictly speaking, but that the intellectual character that produces such seemings is in some way deficient. The worst we could say about the seemings themselves are that they are delusory or non-​veridical. All of this makes sense if seemings are experiences. Undergoing an experience is not subject to the same norms as believing. The content of your experiences does not need to be supported by evidence, for instance. In fact, it is plausible that experiences are not even the sort of thing that can be justified or unjustified. Of course, a person can be epistemically or morally vicious for putting herself into a position to have certain experiences, but those experiences are not unjustified, just delusive. A third problem with the Belief View is that, intuitively, its seeming to us that p plays an important role in our coming to believe that p.12 A seeming that p paradigmatically motivates belief in p. This is why, when asked to explain why we believe something, we often respond by saying that it seems that way. The following statements are all felicitous: “I believe p because it seems that way.” “When it seems that p it inclines me to believe p.” “My belief in p is based on the fact that is seems that way.” The Experience View has no difficulties here. If seemings are experiences of the right sort—​ say, with propositional content and a distinctive

106  What Are Seemings? phenomenal character that inclines us to believe that content—​then it makes perfect sense that they would cause (or otherwise explain) our beliefs.13 The Belief View, on the other hand, seems unable to accommodate the data. Seemings obviously cannot cause or explain our beliefs if to be a seeming just is to be a belief. (I will critique a possible response to this style of argument in §4.1.2 when the same line of reasoning is leveled against the Inclination View.) The cumulative weight of these problems is too much for the Belief View to bear. It fails to line up with our pre-​theoretical understanding of seemings. 4.1.2  The Inclination View

The inclination view says that seemings are not beliefs but rather conscious inclinations to believe. Inclination View—​ A seeming that p just is a conscious inclination to believe that p. Ernest Sosa, a proponent of the Inclination View, characterizes a conscious inclination to believe p as the feeling of being pushed or compelled to believe p. He writes: What are these seemings? It is helpful to compare deliberation on a choice or the pondering of a question, where we “weigh” reasons pro or con. Switching metaphors, we feel the “pull” of conflicting considerations. No matter the metaphor, the phenomenon itself is familiar to us all. There is something it is like to feel the pull of contrary attractions as we deliberate or ponder. Such intellectual seemings, such pulls, are distinct from sensory experiences. (Sosa 2007, 47–​48) The way that Sosa describes these felt inclinations makes his view seem very similar to the Experience View. The difference is that, on the Experience View, the seeming is not the inclination itself but the thing that causes the inclination. Why do you feel your mind beginning to assent to p? Because p is presented to you as true. It is not clear to me that we can cleanly separate the presentation of p as true from the feeling of being drawn toward believing p that Sosa describes above (and if we cannot, then Sosa’s version of the Inclination View would not be a rival to the Experience View at all); but insofar as we can make a distinction, the Inclination View identifies the seeming with the inclination rather than the experience that initiates that inclination.

What Are Seemings?  107 Many of the problems facing the Inclination View are the same as or similar to those confronting the Belief View. As before, the first objection to the Inclination View is that it can seem to one that p without one being inclined to believe p. Let us return to the Naïve Comprehension Axiom. To someone who is familiar with axiom and its history, it may seem that it is true while lacking any inclination to believe it. Samuel Taylor (2015, 10) claims that it is infelicitous to state, “It seems (or appears) to me that P is true but I don’t feel at all compelled to actually believe P,” but that statement sounds perfectly felicitous to my ears. Consider also the statement, “I admit that the Naïve Comprehension Axiom seems true but, since I know better, I’m not at all inclined to believe it.” Earlenbaugh and Molyneux (2009, 106ff) offer what I consider to be the strongest response to this objection.14 Their first move is to distinguish between competing inclinations and net inclinations.15 As Chudnoff points out, this distinction is analogous to the distinction between component and resultant vectors (Chudnoff 2011b, 633). One may have multiple inclinations that counter or supplement one another resulting in a net inclination that is stronger or weaker than any particular inclination. In the case of competing inclinations, two inclinations oppose one another (i.e., incline in opposite directions), negating to some degree the effects of the other. The example Earlenbaugh and Molyneux (E&M) give is of having “a competitive inclination to eat the cake and a simultaneous competitive inclination to stay on one’s diet” (Earlenbaugh and Molyneux 2009, 106). E&M might argue that when considering the Naïve Comprehension Axiom one has a competing inclination to believe that it is true but no net inclination to believe this, and this is compatible with the Inclination View. I don’t think this explanation works. When I say, “I have no inclination to believe that the Naïve Comprehension Axioms is true,” I am saying that I have no inclination whatsoever to believe this. I have neither a competing inclination nor a net inclination to believe it. I think this is what many others are intending to report as well.16 Regardless of what people in general mean, as long as there is one case where something seems true without the subject having any inclination to believe it, the Inclination View is false. E&M could change their strategy. Perhaps they should maintain that those who report having no competing inclination to believe what seems true are simply mistaken.17 When one of these people introspects, one sees that one has no net inclination to believe and mistakenly concludes that one has no competing inclination to believe. To help us analyze this new response, let us label two scenarios as follows (p is the proposition that for every well-​defined predicate with a free variable, P(x), there is a set whose members consist in all and only those who satisfy P(x)):

108  What Are Seemings? No Inclination—S has no competing inclination to believe p and, hence, no net inclination to believe p. Countered Inclination—S has a competing inclination to believe p and a greater or equal competing inclination to resist believing p (or to believe ~p) and, hence, no net inclination to believe p. With respect to the Naïve Comprehension Axiom, people like myself report that p seems true and that No Inclination obtains. This is impossible if seemings are inclinations to believe. The explanation I am proposing for E&M (which I will call “E&M’s explanation” for convenience) is that these people mistake No Inclination for Countered Inclination. For E&M’s explanation to work, it must be the case that the phenomenology of No Inclination and Countered Inclination are the same or at least similar enough that one can easily mistake being in Countered Inclination for being in No Inclination. Otherwise, it is implausible that those who introspect are mistaking one for the other. The problem (for E&M) is that the phenomenology of these two scenarios is very different. Let us take E&M’s own example: that of having a competing inclination to eat cake and a competing inclination to stay on one’s diet. Let’s say that one’s competing inclination to maintain one’s diet is slightly stronger than one’s inclination to eat the cake, resulting in a small net inclination not to eat the cake. If you have been in this situation (or one similar), recall how it felt. In my own experience, I was well aware of my competing inclination to eat the cake, to say the least. It felt like a battle was raging between my various inclinations—​each one tugging me in opposite directions. One inclination did not mute the other; I just had both inclinations at full strength. Compare this to the situation in which you see some cake and you have no inclination to eat it and a slight inclination not to do so. There is no battle raging in this case. You do not feel pulled in opposite directions. One scenario feels completely different from the other. Furthermore, if one is strongly inclined to eat the cake but barely manages to resist thanks to a stronger inclination to stay on one’s diet, this is often a difficult, draining event. In contrast, there is no distress in having no inclination to eat the cake and a slight inclination to not eat it. The upshot is that No Inclination and Competing Inclination feel completely different from one another. This renders E&M’s potential response untenable. It is implausible to think that people like myself are mistaking No Inclination for Competing Inclination. Thus, we should take these people’s reports at face value and, accordingly, as evidence against the Inclination View. A second popular argument against the Inclination View arises out of reflection on the functional role of seemings (Huemer 2007, Cullison 2010). As we saw with beliefs, when asked why we are inclined to believe

What Are Seemings?  109 something it is common to cite the fact that it seems true. It is natural to say, for instance, “I am inclined to believe p because it seems that p.” In these cases, we appear to offer our seemings as explanations for why we are inclined to believe. On reflection, we find that seemings are standardly thought of as the causes of our inclinations—​as those things that incline us to believe. But if a seeming that p explains one’s inclination to believe p, then what it is to be a seeming cannot be the same as what it is to be an inclination to believe. The upshot is that the Inclination View is at odds with our ordinary understanding of seemings and their functional role in bringing about inclinations to believe. Taylor (2015) provides one of the most promising responses to this charge.18 He suggests that when we say, “I am inclined to believe p because it seems that p,” we have slipped into using what Chisholm called the comparative use of “seems.”19 What we are reporting is that our current sensations are of the sort that is typically present when p and that it is these sensations which incline us to believe. To use Taylor’s example, we come to associate certain sensations with there being a white cat on a couch. In this way, it comparatively seems that there is a white cat on a couch. When these sensations incline us to believe, we are prone to state, “I am inclined to believe that there is a white cat on a couch because it (comparatively) seems that there is.” Thus, Taylor finds that such statements do not threaten the view that seemings are inclinations to believe. The most immediate problem with Taylor’s strategy is that it does not cover all the cases. In particular, it fails to account for rational seemings, memorial seemings (those associated with semantic memory), or other kinds of seemings that do not have any accompanying sensory phenomenology. For instance, say you find yourself inclined to believe that everything is identical to itself. When asked about the origin of this inclination, you report, “I’m inclined to believe that everything is identical to itself because it seems that this is true.” Taylor cannot say that you are using “seems” comparatively since there are no sensations associated with everything’s being identical to itself; much less can he appeal to these missing sensations as the cause of your inclination to believe. Thus, Taylor’s explanation fails to square the Inclination View with the all the data. A third argument, closely related to the second, can be gleaned from Earl Conee. In differentiating inclinations to believe from “seeing the truth,” Conee writes: When seeing the truth of an intrinsically obvious proposition, we do feel strongly inclined to believe it. But this feeling of impulsion cannot be the phenomenal quality by which the truth is seen, nor can it be the evidence that is provided by the seeing for the truth of the proposition. Seeing a proposition to be true is a reason for believing the proposition—​perhaps

110  What Are Seemings? the best sort of reason there could be. In contrast to this, feeling driven to believe a proposition is no reason at all on its own. (Conee 1998, 850) A similar argument might be made with respect to something’s seeming true. We do not need to assume as Conee does that something’s seeming to be true is in fact a good reason for believing that proposition; rather, it is enough that seemings are, in our pre-theoretical understanding, plausibly good reasons for believing their content. The fact that seemings are plausible candidates for such a role, but inclinations to believe are not, suggests that seemings aren’t mere inclinations. A fourth argument against the Inclination View is that seemings are representational—​they have accuracy conditions.20 The seeming that my dog is sleeping can be accurate or inaccurate. Mere inclinations, however, are not accurate or inaccurate. At most, they can incline you to form some other mental state that is accurate or inaccurate. Nor are inclinations derivatively accurate or inaccurate by virtue of inclining us to form mental states with accurate or inaccurate content. An inclination toward something inaccurate is misleading, not inaccurate. The problem seems to be that inclinations do not have propositional content. Inclinations incline us toward mental states with propositional content, but this content is not a component of the inclination itself. In contrast, on the Experience View seemings have propositional content and, thereby, accuracy conditions. Thus, insofar as we think that seemings can be accurate or inaccurate, we have reason to endorse the Experience View over the Inclination View. I will offer a final line of reasoning that displays the superiority of the Experience View over the Inclination View. A number of philosophers note that not all inclinations to believe are seemings.21 For instance, I may desire p and so be inclined to believe p, but this inclination cannot plausibly be considered a seeming. Say, I am an avid Chiefs fan and so strongly desire to see the Chiefs win the Super Bowl that I am inclined to believe that this will happen. It does not thereby seem true to me that the Chiefs will win the Super Bowl. It may even seem that the Chiefs will not win the Super Bowl, though I am still inclined to believe it by my desire to see it happen. The same reasoning will disqualify inclinations to believe caused by hopes, wishful thinking, emotions, or appetites. In light of this, proponents of the Inclination View need to limit seemings to some particular kind of inclination to believe. The natural amendment (considered by Cullison 2010, 266–​269) is to restrict seemings to truth-​ directed inclinations to believe. Inclinations caused by desires are not aimed at the truth of the proposition but at the desirability of the proposition. Similarly, inclinations caused by positive emotions are not aimed at the truth of the proposition but at the goodness of the proposition. It is

What Are Seemings?  111 because that proposition feels good that we are inclined to believe it. Thus, by limiting seemings to truth-​directed inclinations to believe we can weed out those inclinations caused by desires, emotions, or the like. Obviously if one is to make this amendment one must be able to spell out what it is for an inclination to be truth-​directed or aimed at truth. Cullison considers the following proper functionalist analysis: It seems to S that P if, and only if, i S feels an inclination to believe P, and ii that inclination was caused by cognitive faculties designed to secure true beliefs. (Cullison 2010, 267) This analysis, however, won’t work for multiple reasons. First, Cullison points out that in a brain-​in-​a-​vat scenario, one’s faculties may not be designed to secure true beliefs. According to this analysis, then, nothing could seem true to the brain in a vat. But, of course, things would seem true to the brain in a vat. The reason this is a common skeptical scenario is that, were you a brain in a vat, things could seem precisely the same to you as they do now. Second, say there is a benevolent demon watching over you. This demon implants a faculty in you that inclines you to believe p whenever you desire p, and then ensures that whatever you desire comes about. This faculty is designed to secure true beliefs in the sense that it is highly reliable and was intended to produce true beliefs by its designer. But as we saw before, inclinations to believe produced by desires are not seemings. Thus, this analysis will not do. Thankfully there is a more natural way to characterize truth-​directed inclinations to believe. I said earlier that inclinations to believe caused by positive emotions are aimed at the goodness of the object rather than its truth. Perhaps, then, we should be looking at the causes of our inclinations. We can then characterize truth-​directed inclinations as those inclinations that have the right sort of cause. At first, you might be tempted to characterize truth-​directed inclinations as those that are caused by the truth of the proposition one is inclined to believe, but there can be truth-​directed inclinations to believe things that are false—​for example, when one is inclined to believe by misleading evidence. You might try to characterize truth-​directed inclinations as those that are caused by the truth-​maker of the proposition one is inclined to believe, but this fails for the same reason as the first proposal. There can be truth-​directed inclinations to believe false propositions, though these obviously lack truth-​makers. Furthermore, it may be that some true propositions lack truth-​makers (e.g., that there are no white ravens), and one can have truth-​directed inclinations to believe these. The best characterization of truth-​directed inclinations is that they are those caused by the apparent

112  What Are Seemings? truth of the proposition one is inclined to believe.22 That is, if an inclination is caused by a mental state in which the proposition is presented as true or feels true to the subject, then that inclination is aimed at truth. Taking this position, however, requires one to posit the existence of forceful mental states with propositional content. Once you admit that there are mental states of this sort, it becomes clear that they are the better candidates for seemings than the inclinations they cause. For example, all of the problems mentioned previously go away if one understands seemings as the experiences that cause truth-​directed inclinations rather than the truth-​directed inclinations themselves. Furthermore, one of the primary motivations for the Inclination View is to avoid positing (supposedly) weird experiences like forceful mental states (see §4.2). But if the Inclination View is forced to posit these kinds of experiences as well, then there is little reason left to endorse the view. 4.1.3  The Taking-​Evidence View

The last view we will discuss, the Taking-​Evidence View, has come onto the scene relatively recently thanks to Conee and Tooley.23 We noted at the start of §4.1 that the language of “it seems that p” can be used to report that we have evidence which supports p to some sufficient degree (or at least that there is such evidence). The Taking-​Evidence View proposes that seemings should be understood as the mental states we are in when we recognize this fact—​the state of taking ourselves to have sufficient evidence for p. We face multiple decision points in fleshing out this view. First, we can understand the act of taking in multiple ways. For example, as: i believing that we have sufficient evidence for p, ii being inclined to believe that we have sufficient evidence for p, iii being directly aware of having sufficient evidence for p, iv being in a state that feels like one is directly aware of having sufficient evidence for p. Nothing in our discussion will hinge on how we understand this taking relation, so I will continue to talk about us “taking” ourselves to have sufficient evidence in a way that is neutral between various options.24 A second decision point lies in how strong we must take the evidence in favor of p to be. Consider the following options: 1 S’s seeming that p just is taking herself to have some evidence for p. 2 S’s seeming that p just is taking herself to have evidence for p sufficient to justify belief that p (or some credence in p of over .5).

What Are Seemings?  113 3 S’s seeming that p just is taking herself to have evidence for p sufficient to justify belief that p (or some credence in p of over .5) absent defeaters. Option 1 is too weak. Learning that a certain dog was at the scene of a shooting is some evidence (extremely weak though it may be) that the dog was the shooter, but it doesn’t thereby seem to me that the dog is the shooter when I recognize this fact. Option 2 seems both too weak and too strong. I can have a seeming that p even when I don’t take myself to have a level of evidential support for p sufficient to justify belief. For instance, the Naïve Comprehension Axiom can seem true to me even though, having heard of Russell’s paradox, I don’t take myself to have justification for believing it.25 On the flip side, I can also take myself to have evidential support sufficient to justify belief in p and yet fail to have a seeming that p. For instance, if you fill a balloon with helium, suspend it in the middle of a non-​moving vehicle, and then step on the gas, the balloon will move toward the front of the vehicle. This doesn’t seem right to many people; they would describe it as “counterintuitive.” And it can remain counterintuitive even if they are justified in believing that the balloon will move forward based on their understanding of physics or, even better, the testimony of a trusted physicist. Here, they have evidence that justifies them in believing that the balloon will move forward, even though it doesn’t seem to them that it will. Option 3 helps with the first of these problems: I would have evidence justifying belief in the Naïve Comprehension Axiom if I didn’t know about Russell’s paradox. But it exacerbates the second: just because I would have justification-​level support for p were I to lack defeaters doesn’t make p seem true. There are lots of counterintuitive conclusions—​ones which do not themselves seem true—​that are supported by evidence absent defeaters. It might help at this point to specify which kind of evidence we are supposed to take ourselves to have for p. Perhaps we need to take ourselves to have evidence of a specific sort rather than just evidence in general. Teasing out a suggestion from Conee, perhaps we need to take ourselves to have a mental state displaying the truth of p in order for p to seem true—​ this mental state constituting a special kind of evidence for p (Conee 2013, 54–​55). This would avoid the problems with counterintuitive conclusions. For although I can take myself to have general evidence for p without p seeming true (because p remains counterintuitive), it is less plausible that I can take myself to have a mental state displaying the truth of p without p seeming true. A conclusion whose truth is directly displayed to you would not ordinarily be described as “counterintuitive.” For this reason, I will treat this sort of “mental state formulation” to be the strongest version of the Taking-​Evidence View.

114  What Are Seemings? Taking-​Evidence View—​A seeming that p just is the state of taking oneself to have a mental state displaying the truth of p. Though I will direct my objections toward this version of the view, those criticisms will apply mutatis mutandis to the more general evidential formulations discussed above as well. The first objection is that the Taking-​Evidence View has the implausible implication that nothing seems true unless one is engaging in second-​order reflection about one’s own mental states.26 Clearly, though, something can seem true to S without S taking anything to be true about her mental states. This happens constantly. The wood seems rough to the carpenter as she runs her hand along the board though she neither engages nor is inclined to engage in any higher-​order reflection. Examples are endless. This alone may be enough to dismiss the Taking-​Evidence View. To voice my second critique, we need to look at the origins of the position. Conee comes up with the Taking-​Evidence View by examining the conditions under which it is correct for S to report, “it seems to S that p.” He writes, The topic of interpretation here is some correct uses of the terms seems and appears and their variants. This “correct” means nothing technical. It means that the terms are employed with semantic and conversational propriety to communicate something. The interpretation offers a description of the conditions that occasion this proper use. No more specific interpretive claim is intended. (Conee 2013, 53) Conee eventually concludes that it is appropriate for S to report, “it seems to S that p,” when S takes oneself to have a mental state displaying the truth of p (or, as we discussed above, evidence for p more generally). Indeed, we acknowledged above that the language of “it seems that p” often is used in this way. This leads Conee to suggest (but not endorse) that for p to seem true to S is just for S to take herself to have a mental state displaying the truth of p. Ironically, the Taking-​Evidence View is refuted by the same considerations that led Conee to suggest it. Let us grant that S can appropriately report, “it seems to S that p,” when S takes herself to have a mental state displaying the truth of p. Now take a step back and notice that being in a mental state M does not, by itself, make it appropriate to report, “I am in M.” I also need to take myself to be in M before I can properly assert this. If I didn’t take myself to be in M, then I would be asserting something that I didn’t believe or wasn’t even inclined to believe. Now the Taking-​ Evidence View says that a seeming is the second-​order mental state of

What Are Seemings?  115 taking some first-​order mental state to display the truth of p. Let’s call this second-​order mental state “A.” Given what we said moments ago, it is only appropriate for S to assert, “I am in A” when S takes herself to be in A. So, on the Taking-​Evidence View, S can appropriately assert “it seems to S that p,” only if S takes herself to take herself to have some first-​order mental state displaying the truth of p. That is, S needs a third-​order mental state—​she needs to believe that she believes something about her first-​order mental states—​just in order to appropriately assert, “it seems to me that p.” This severely overcomplicates matters. As Conee himself argues, it is appropriate for S to assert, “it seems to me that p,” when S takes herself to have a mental state displaying the truth of p; she needn’t take herself to take herself to have such a mental state. Thus, the Taking-​Evidence View is falsified rather than confirmed by Conee’s conclusions about seeming reports. In fact, what Conee’s analysis suggests is that the mental state displaying the truth of p is the seeming. For it is appropriate for S to assert, “it seems to S that p,” precisely when she becomes aware that she in a mental state displaying p’s truth. The best explanation is that the mental state displaying p’s truth is the seeming and it becomes appropriate for S to report this fact when she becomes aware of it. Ironically, Conee’s analysis of seeming reports ends up confirming the Experience View rather than challenging it. Our discussion reveals severe problems for the Taking-​Evidence View. The same is true, we have seen, for all of the other alternatives to the Experience View as well. In contrast, the Experience View has at every turn accounted for our ordinary understanding of seemings without difficulty. I take this to be strong evidence for the Experience View. 4.2  Concerns for the Experience View Before we crown the Experience View the winner, we must consider its own problems. If they are as bad or worse than those facing the alternatives, then perhaps one of those alternatives will still be preferrable (or else we should keep looking for better options). The main complaints against the Experience View are as follows. First, seemings as described by the Experience View, or “experiential seemings,” are novel, or unfamiliar, or weird (the exact nature of the complaint is hard to pin down), and so should be approached with a healthy dose of suspicion. Lyons writes: If the appearance is not a belief, it’s not a property of the belief, and it’s not the familiar (so to speak spatial, colorful) sensory experience whose connection to belief is merely contingent, then it is something quite mysterious. (Lyons 2013, 25)

116  What Are Seemings? Second, proponents of the Experience View have not done enough to overcome this suspicion by carefully and systematically articulating the nature of experiential seemings. Though Feldman is writing prior to several seminal discussions, he still speaks for many in saying: It must be admitted that the idea of what it is for one to have an experience as of its being the case that p is far from clear. Some better account of how this all works is needed. One has the sense that defenders of this view are simply saying whatever they must to achieve the desired results. A more general theoretical basis for their view is desirable. (Feldman 2003, 147) Third and finally, some prominent figures in epistemology are “seemings skeptics” and claim not to have seemings or anything like seemings. For instance, here is Williamson talking about his intuitions in the Gettier case: Although mathematical intuition can have a rich phenomenology, even a quasi-​perceptual one, for instance in geometry, the intellectual appearance of the Gettier proposition is not like that. Any accompanying imagery is irrelevant. For myself, I am aware of no intellectual seeming beyond my conscious inclination to believe the Gettier proposition. Similarly, I am aware of no intellectual seeming beyond my conscious inclination to believe Naïve Comprehension, which I resist because I know better. (Williamson 2007, 217) Conee furnishes another example. He writes: The suggested characterization of the phenomenology of intuition is that when a proposition is seen to be true, we are aware of some sort of brightness or shine that apparently radiates from the proposition. This would not be a literal shine, since that is a visual quality. It would be some cognitive analogue to a visual radiance. The plain fact is that no such quality exists. … Further reflection shows that there is no special sort of attractiveness displayed by the propositions that are seen to be true. In fact, intrinsically obvious propositions are largely rather dull to consider, even while we see that they are true. (Conee 1998, 849–​850) Some take these failures to find any seeming or to locate their forcefulness to be evidence that experiential seemings do not exist. I will address each of these worries, beginning with the last.

What Are Seemings?  117 In response to the seeming skeptics, I offer several responses. The first is that the appeal to experiential seemings is not justified solely through introspection but also by virtue of the sorts of considerations outlined earlier in this chapter. Something is going on when we consider the Naïve Comprehension Axiom and, for the reasons already given, other states like beliefs or inclinations to believe cannot adequately account for it. Experiential seemings make the most sense of how we ordinarily think and talk about seemings, and that gives us reason to posit them. My second response to seeming skeptics is that the failure to find a seeming in one particular instance provides little evidence that seemings do not exist. Even for the most obvious propositions—​the one’s that most strongly seem the true to us—​we do not have occurrent seemings in their truth every time we consider them.27 Furthermore, the failure to find the experiential seeming or its distinct phenomenal character may simply be a mistake. To this end, Elijah Chudnoff offers the following insight: In order to find—​ not just have—​ an intuition experience, your understanding of what intuition experiences are shouldn’t be wildly mistaken. If your understanding of what intuition experiences are is wildly mistaken, then you might very well have many intuition experiences, but fail to recognize them as such. (Chudnoff 2011b, 644; 2013, 53) Both Williamson and Conee seem to be looking for a state with quasi-​ perceptual phenomenology—​that is, the sort of imagery you have when you form of mental picture of a right angle or an equilateral triangle. But that’s not at all the sort of phenomenology one should be looking for with respect to intellectual seemings. I would also add that even if your understanding is not “wildly mistaken,” it can still be easy to conflate your seeming state with something else that feels similar to it. Note that Williamson acknowledges a felt “conscious inclination” to believe the Gettier proposition and the Naïve Comprehension Axiom. When first introducing the Inclination View above, I noted that it can be difficult to distinguish the presentation of something as true from the feeling of one’s own mind being drawn toward believing it. Plausibly, Williamson is aware of an experiential seeming, but is mistakenly categorizing it as a feature of his conscious inclination instead of the cause of that inclination. Finally, even if we decide to trust Williamson or Conee when they claim not to have any experiential seemings, should we not also trust the many more who claim that they do have them? The contemporary literature is full of epistemologists who claim to recognize experiential seemings upon introspection.28 And many more can be claimed if we include those who describe states very much like experiential

118  What Are Seemings? seemings, though they may not call them by that name.29 What adds even more credibility is that seemings of this sort have been widely recognized throughout the history of philosophy. Again, the word “seemings” won’t show up much if any in the tradition, but what you will see are mental states that share the basic contours of experiential seemings: that is, mental states with propositional content that feel as though they are revealing reality to us and which thereby incline us to assent to their content. While there is no time for a full historical survey, I will point to a few choice examples. We can begin with the ancients. Stoics such as Zeno of Citium held that foundational beliefs (the term he uses is katalēpsis, or “apprehensions”) were based on phantasia, which might be translated as “appearances,” “presentations,” or commonly “impressions.” Usually thought of as sensory, these impressions have several important features for our purposes. First, impressions have “content that can be expressed in language,” and so are typically called “rational” or “cognitive impressions” (Sedley 1998, sect. 12). Today, we would say that they have propositional content.30 Second, cognitive impressions are distinct from beliefs: “Whether or not you take the impression to be true depends on a further cognitive act, assent (synkatathesis), which you may give or withhold at will” (Sedley 1998, sec. 12). Third, “To have an impression is simply for things to strike you as being a certain way” (Sedley 1998, sect. 12) They are described as having a kind of “clarity and distinctness” that assures you of their truth (Allen 2020 sect. 2). Fourth and finally, though not irresistible, cognitive impressions “induce us to give assent” to the content represented therein (Løkke 2015, 72). This understanding of impressions was mirrored by Academic Skeptics like Carneades. The Academics disagreed with the Stoics over the importance of a special sort of cognitive impression—​one that was “impressed and sealed and moulded from a real object, in conformity with its reality,” such that it could not be false (Cicero 1933, 2.77).31 The Academics instead held that “no presentation proceeding from a true object is such that a presentation proceeding from a false one might not also be of the same form” and so emphasized merely probable impressions (Cicero 1933, 2.77). Regardless, the Stoics and Academics are in step on the general nature of the impressions themselves. It is evident that despite differences in terminology, the impressions of the Stoics and the Academics are seemings as I have defined them (or something close enough). They are mental states, distinct from beliefs, with propositional content and a forceful phenomenal character that inclines one to believe their content. Even if there are important differences in the ways we characterize these states, the fact that these ancients identified something with the same general features lends credence to the Experience View. Much like two witnesses who both claim to have seen a tall white

What Are Seemings?  119 male commit the crime lend support to each other, even if they disagree about the color of his shirt. Something similar is plausibly true of Aristotle. As with the Stoics, Aristotle posited a middle state falling between perception and perceptual belief called appearance (phantasia) (Irwin 1988, 326). Though closely related to perception (Aristotle calls them a “trace or echo of perceptual activity” (Caston 2009, 324)), appearances are seen to be distinct in that, unlike perception, they can be false (Irwin 1988, 326). This means that appearances must be representational states possessing propositional content that may or may not be true. At least in rational perceivers, this content seems to be conceptual given that the same object might appear differently to different observers depending on which concepts they possess (Irwin 1988, 325–​326). Despite having conceptual content, appearances are differentiated from beliefs in much the same way we argued against the Belief View: The sun appears to be a foot wide, he argues, even though we believe, and even know, that it is not, but larger than the whole earth (An III.3 428a24–​b9), just as the two lines in a Müller-​Lyer diagram (to use a modern example) continue to look unequal even after we have convinced ourselves that they are in fact equal. (Caston 2009, 324) These appearances are eventually taken up in thought and believed. Of course, it is unclear why such appearances would lead to belief unless there was something about their phenomenal character in that instant that motivated assent—​something like the feel of a state that is representing things as they really are. So, once again, there is a mental state in Aristotle’s system that plausibly fits the description of an experiential seeming. Departing the ancient period, we can see parallels to experiential seemings in the works of many early modern epistemologists. The most well-​known are Descartes’s “clear and distinct perceptions.” Throughout the Meditations, Descartes says that he clearly and distinctly perceives that various principles are true, indicating that these perceptions take propositions as their objects and so have propositional content. The term “perception” is, of course, metaphorical. Indeed, one of the central goals of Descartes’s project is to establish that perception is not a foundational source of justification, but must be re-​grounded on the clear and distinct “perceptions” of reason. Why, then, does Descartes use the term “perceptions” to describe the states that are to replace actual perceptions as the ultimate justifiers of our beliefs? Presumably because, following a long tradition in ancient and medieval philosophy, Descartes found an analogy between the way in which propositions are presented to the subject in

120  What Are Seemings? “clear and distinct perceptions” and the way in which objects are generally thought to be presented in perceptual experience. He adds elsewhere that these propositions are “shown me by the light of nature” (Descartes [1641] 2006, 21–​22). Such metaphors are also found in other early moderns like Locke, who describes certain propositions as displaying an “evident lustre” and “perfect Clearness and Distinctness” (Locke [1689] 1975, 533).32 Plausibly, what Descartes and Locke are referring to is the way in which certain mental states present their content as true. Just as light reveals the presence and features of various physical objects, so our “clear and distinct perceptions” feel as though they are revealing to us the truth of their content. Finally, Descartes notes that “whenever I turn my attention to those very things that I perceive with such great clarity, I am so completely persuaded by them…” (Descartes [1641] 2006, 20). Locke says they grant “full assurance” (Locke [1689] 1975, 533). In other words, something about the presentation of these propositions, in these instances at least, not only encourages assent, but compels certitude. Thus, we see in Descartes and Locke states very much like those I have described as seemings. Examples could be multiplied, but my point is already clear. Whatever evidence seeming skeptics provide against the Experience View is far outweighed by the counterevidence provided by the great cloud of witnesses who do acknowledge experiential seemings. There may even be scientific evidence for experiential seemings. Lu Teng (2022, §4) outlines findings from the psychological sciences which point to the existence of source-​monitoring mechanisms. True to their name, these mechanisms monitor the origins of our mental states—​for example, is this a random thought or was it retrieved from memory?—​and render a verdict as to whether those states are likely to be veridical. Schizophrenics, for instance, seem to have defects in their source-​monitoring mechanisms which make it difficult to distinguish their own imaginings from veridical experience. Such mechanisms, Teng says, produce a distinct kind of epistemic feeling in us, one that assures us that a particular experience is veridical. Teng suggests we identify this metacognitive feeling with the phenomenal force characteristic of experiential seemings. She calls this the metacognitive account of phenomenal force. If Teng is right, then we have confirmation of experiential seemings not only from the philosophical tradition but from our best science as well. In light of all this, it is hard to see the small handful of seeming skeptics as posing an especially strong challenge to the Experience View. Another important lesson to be learned from our brief historical survey is that experiential seemings are not some newfangled invention or contemporary fad. They have deep roots in the tradition, showing up repeatedly throughout the Western canon.33 This goes a considerable way toward addressing another concern with the Experiential View: that seemings

What Are Seemings?  121 so understood are novel, or unfamiliar, or weird, and should thereby be approached with great suspicion. They aren’t novel but have been discussed as plausible sources of immediate justification from ancient times. They aren’t unfamiliar. We can easily become introspectively aware of them within ourselves and prompt this awareness within others. As to whether they are weird, I would point out that what counts as weird often has far more to do with the intellectual fashions of the time than with the intrinsic plausibility of the position being considered. The word “weird” is like the early modern pejorative “occult.” What counts as occult depends on the paradigm operative in one’s intellectual community and can change rapidly. For instance, the idea of gravity acting at a distance was occult until Newton said it wasn’t. The point is that some may find the idea of experiential seemings weird from their particular perspectives; but unless we have reason to think their perspectives are especially reliable in this regard, that fact has little weight in the grand scheme of things. The most plausible accusation that can be made against the Experience View is that its proponents have not given a sufficiently detailed analysis of experiential seemings. Of course, even if this is granted (though I do not think it should be), I do not see how it undermines the Experience View. What proponents have done, and done well, is to guide us in introspectively identifying experiential seemings within our own minds, usually by giving examples or by comparing and contrasting them with other kinds of mental states. This provides sufficient evidence that there are experiential seemings, even if we do not fully understand what these mental states are or how they do what they do. What’s more, if the distinctive features of experiential seemings are primitive, then little more can be done than to gesture at them. No deeper analysis can be given. Nevertheless, I will attempt to say a least a bit more than what proponents have said so far. In particular, I will try to give us a better understanding of what it means for a seeming to present its content as true. The phenomenology associated with this presentation—​what has been called “felt veridicality” or “the feel of truth”—​has proven difficult to describe. Some don’t even try! Plantinga, for instance, writes: So what is it, then, to see that a proposition p is true? All I can say is this: it is … to form this belief with that peculiar sort of phenomenology with which we are well acquainted, but which I can’t describe in any way other than as the phenomenology that goes with seeing that such a proposition is true. (Plantinga 1993, 106) Those who do describe the feel of truth further sometimes use metaphors, such as saying that seemings feel as though they are “recommending” their

122  What Are Seemings? content or “asserting” it as true (Tucker 2010, 530)—​metaphors that, according to Tooley, are never cashed out: Since mental states are not agents, they cannot literally assert anything, or recommend anything, or assure us that anything is the case. So such metaphorical talk needs to be jettisoned and replaced with characterizations that are literally true. Since Huemer and Tucker have not done this, they have not really provided us with a satisfactory account of what they take seemings to be. (Tooley 2013, 310) To complicate matters, it is not always made clear how experiential seemings are related to other kinds of mental states such as perceptions. Thus, it would also be helpful to situate experiential seemings within a more general taxonomy of mental states. In the rest of this chapter, I will try to address all of these concerns. Though the Experience View can survive without such an analysis, adding it will place the Experience View on even surer footing. From here on, as in earlier chapters, “seemings” will refer to experiential seemings. 4.3  Fleshing Out the Experience View Seemings are a species of appearance. As a genus, appearances are mental states in which states of affairs are presented to a subject as obtaining or not obtaining. The specific manner in which seemings present states of affairs is via propositional content that feels true or false.34 There are, however, other ways in which states of affairs might be presented to the subject that do not involve propositional content. We can better understand seemings by comparing and contrasting them with another kinds of appearance—​namely, sensory appearances. This will also provide a valuable opportunity to clarify the relationship between perceptual seemings and perception more broadly. 4.3.1  Perceptual Seemings versus Sensory Appearances

Sensory appearances present physical states of affairs to us via the rich sensory phenomenology we all know and love. Beyond this, there is little we can say about sensory appearances without wading into controversy. A prominent dispute is whether sensory appearances represent physical states of affairs or not. Representationalism says yes. Those who say no usually endorse relationalism, which takes sensory appearances to involve some kind of direct relation between mind and object. I do not need to take

What Are Seemings?  123 a stand on this issue. Accordingly, I will talk of sensory appearances “(re) presenting” states of affairs, where this remains neutral on whether those objects are presented via representations or direct relations. I will also refer to the “content” of sensory appearances, although this should not be taken to imply representational content specifically. What is important for our purposes is that the kind of presentation that occurs in sensory appearances—​whatever exactly it turns out to be—​is not conceptual. Thus, the one view I do reject here is conceptualism, a species of representationalism which says that sensory appearances represent their objects via conceptual (i.e., propositional) content.35 There are many reasons for thinking that sensory appearances are non-​ conceptual. Two compelling ones are that the content of one’s sensory appearances far exceeds in both richness and determinate detail whatever is conceptually articulated by us in thought, in actuality if not in principle (see Martin 1992 and Tye 1995, respectively). Another is that infants and animals are capable of having sensory appearances even if they are not capable of deploying concepts (Bermúdez 1998, Evans 1982, Peacocke 2001). I would add that the sensory capacities of infants do not grow in proportion with their growing conceptual abilities—​the thought of an adult far outstrips the thought of an infant, but their sensory appearances are not much different. The use of demonstrative concepts to refer to the objects of sensory appearances also points to the need for some kind of prior, non-​ conceptual presentation of those objects (Campbell 2002). These and other arguments cut against conceptualism about sensory appearances. It follows that no seemings (not even perceptual seemings) are sensory appearances. Seemings have propositional content; sensory appearances do not. In this, I am siding with Tucker and others in distinguishing seemings from “sensations,” which is how they refer to states with sensory phenomenology (Tucker 2010, Bergmann 2013b).36 Drawing a distinction between perceptual seemings and sensory states does not, however, prevent perceptual seemings from depending on sensory states in important ways. What do I have in mind here? Something like the following. Say you have a sensory appearance in which a dog is non-​conceptually (re)presented as running toward you. This triggers a second mental state, downstream from the first, which conceptually represents that same state of affairs (or some aspect of it) and does so in a way that feels true to you. This second state is a perceptual seeming. While distinct from the sensory appearance, the perceptual seeming still depends on it in several ways. One, the perceptual seeming would not exist were it not triggered by the sensory appearance. Two, plausibly, the perceptual seeming would not be classified as a perceptual seeming were it not triggered by a sensory appearance. Three, the perceptual seeming could deploy a demonstrative concept that refers to something (re)presented in the sensory appearance. For instance, it may

124  What Are Seemings? seem to you that that dog, or even just that thing, is running toward you. In which case, the perceptual seeming is partially constituted by the sensory appearance—​the seeming not only would not, but also could not exist without the prior appearance. Fourth and finally, the content of the perceptual seeming may only feel true because you can “see” a correspondence between the seeming’s propositional content and the states of affairs presented in the sensory appearance (more on this in §4.3.3). In this way, the sensory appearance can explain why it seems that a dog is running toward you as opposed to merely thinking about a dog running toward you. We can, if we like, go on to identify perceptual experience in its fullness with the complex mental state involving both a sensory appearance and the perceptual seeming to which it gives rise. If we take this option, our account would qualify as a “dual-​component” theory of perception (Smith 2002). This is, in its broad outlines, the sort of theory that Reid endorsed although Reid understood both the conceptual and non-​conceptual components differently than we have here (see Reid [1785] 2002, 210). Perception began for Reid with the body being affected so as to generate non-​conceptual sensations, understood as raw feels. These sensations then triggered a conception—​a belief—​that referred to the cause of those sensations in accordance with both natural and acquired principles of the human constitution. In contrast, the theory we are considering understands sensations to be part of the non-​conceptual (re)presentation of an object, and it takes the conceptual component to be a seeming rather than a belief. An advantage of this sort of dual-​component theory is that it can account for both the motivations that push people toward conceptualist theories of perception and those that push people toward non-​conceptualist theories (Bengson, Grube, and Korman 2011, Cullison 2013, 48). The main argument for conceptualism is that perception must play a certain role in justifying beliefs, and that is a role it cannot play without propositional content (McDowell 1994).37 Our dual-​component theory can acknowledge the strength of this argument: perceptions can justify only by virtue of their conceptual components.38 On the other hand, I have cited numerous motivations for thinking that perceptual states are non-​conceptual: the richness and detail of their content; the fact that infants and animals can share them; the need for non-​conceptual states to fill in the content of demonstrative concepts. Once again, our dual-​component theory can accommodate these arguments. They all point to the need for a non-​conceptual component to perceptual experience—​one that is prior to the conceptual component—​ and our dual-​component theory offers that. Finally, consider that when two people look at the same painting, their perceptual experience of that object can be both the same and different. It can be the same with respect to sensory phenomenology—​the colors and patterns manifested in conscious

What Are Seemings?  125 experience are exactly alike—​ but different with respect to conceptual content—​to one, it only seems that this is a beautiful painting; to the other, it seems that this is a Botticelli. Our dual-​component theory explains this brilliantly. The alternative is to identify perceptual experience exclusively with sensory appearances. Our perceptions would then cause perceptual seemings rather than partially consist in them. This view has the undesirable implication that perceptions do not serve as foundational evidence for our beliefs (assuming, as I will go on to argue, that seemings do all the justificatory work). Notwithstanding, I am satisfied with the conclusion that humans have both non-​ conceptual perceptual states (sensory appearances) and conceptual perceptual states (perceptual seemings), whatever their exact relation to “perceptual experience” turns out to be. 4.3.2  Assertive Force and the Feel of Truth

Though distinct, sensory appearances share something in common with perceptual seemings: namely, their force or mode of presentation. The force of a mental state can be compared to the force of a linguistic utterance. The same content can be uttered in assertive, interrogative, or imperative modes. Similarly, the same content can, within a mental state, be entertained by the subject in some or all of these ways. Importantly, the force of a mental state is not to be understood as part of the content of that state. As Smithies writes, “Just as the name suggests, forcefulness is a distinctive kind of representational force, rather than a distinctive kind of representational content” (Smithies 2019, 386). Accordingly, two mental states can have the exact same content while differing in force. Such states will have different phenomenal characters—​what it feels like to be in one will be different from what it feels like to be in the other—​but this is due not to a difference in content, but to the way in which that content is presented. So, for instance, if it seems to us that p and if we imagine that p, these states will have the same propositional content. Nevertheless, their phenomenal characters are not at all the same. Likewise, it is logically possible (even if infeasible) to visualize something as vividly as if we had a sensory appearance of it. This visualization might have the same content as a sensory appearance but would nevertheless feel quite different. This difference is accounted for by the different representational force of those mental states (Siegel and Silins 2015, §4, Smithies 2019, 92–​93). Both sensory appearances and perceptual seemings share a similar kind of representational force: specifically, assertive force. Much like an assertive utterance, mental states with assertive force present their content to the subject as being the case. Such states feel as though they are revealing how things really are. We can contrast this with other states that do not

126  What Are Seemings? feel this way. The classic illustration is to juxtapose imagining that p with its seeming that p, but we could also compare wondering whether p to its seeming that p. If wondering has a distinctive kind of force at all, it is something like interrogative force: it is put forward as a question whether p. Contrast this to a seeming, in which p is put forward as representing how things really are.39 Because sensory appearances and seemings have different kinds of content, their assertive force is not exactly the same. The assertive force of a seeming makes its content feel true, whereas the assertive force of a sensory appearance makes its content feel accurate (assuming representationalism) or real (assuming relationalism). Thus, seemings have a unique kind of assertive force that we can call “felt veridicality” or “the feel of truth” (Tolhurst 1998, 298–​299). It is this specific sort of assertive force—​the feel of truth—​that we have been calling “forcefulness” in our discussions thus far (a practice that will continue throughout this inquiry). The term “assertiveness” is sometimes used to characterize beliefs. No doubt beliefs do involve some kind of assertive orientation, but the kind of assertiveness characteristic of beliefs is not the same kind of assertive force characteristic of seemings (c.f., Smithies 2019, 93 and 387–​388). In a belief, we take an assertive stance toward the proposition in question. Notice our active role here. It is we who are doing the asserting. In a seeming, however, it is the seeming that presents its content with assertive force. We—​the subjects—​are passive (Skene 2013, 542). We are receiving the assertion. This is why beliefs need justification in a way that seemings do not. We can only be assessed on stances that we take toward a proposition; but we are not taking any stance toward p simply by virtue of p’s seeming true. Of course, having p presented to us with assertive force can and usually will incline us to take an assertive stance toward p, but the former (the seeming) and the latter (the belief) are not to be conflated.40 What about Tooley’s complaint that we are relying too heavily on metaphorical language? Such a complaint has little going for it. The description I have just given of seemings is not predominantly metaphorical at all. I have pointed out that mental states, like illocutionary acts, have force—​a mode in which they present their content—​which is reflected in their phenomenal character. We can easily become aware of this aspect of our mental states through introspection. I then pointed out that seemings have a kind of force that is shared with sensory appearances. Again, we can fix the referent here by highlighting various states that do and do not share this force. The name “assertive force” does draw an analogy between what’s going on in these mental states and what’s going on in assertive utterances, but that’s simply a label for a feature of our experience that we are attending to directly. Thus, our understanding of this feature does not critically depend on that metaphor.

What Are Seemings?  127 But isn’t this idea of assertive force weird? Setting aside the general feebleness of such complaints (see §4.2), I don’t see how it is particularly weird. Is it any weirder for a state to present its content under a particular mode than for it to have content in the first place? Alternatively, if by “weird” the objector really just means “wonderous,” then I admit that assertiveness is wonderous. But so are all aspects of consciousness! Indeed, a surprising number of thinkers have found consciousness so weird that they have tried to do away with it altogether. Behaviorists of a certain stripe are a prime example. The error in privileging their own sense of weirdness over the introspectively evident fact that conscious does exist is, I take it, painfully obvious (pun intended). Though the matter is less stark when it comes to assertiveness, a similar principle applies. We ought not privilege our sense of weirdness over the fact that there is a clear introspective difference between, say, entertaining p and intuiting p, and that this difference is plausibly described as a difference in representational force rather than representational content: the intuition presents its content assertively, but the state of entertaining does not. Another consideration that bolsters the legitimacy of assertiveness is that we see parallels to it in other areas of our mental lives. For instance, there is a strong case that emotions should not be thought of as mere bodily perturbations or raw feels but as a kind of quasi-​perceptual state in which states of affairs are presented as good or bad in a distinct sort of way. For instance, when we are angry about something, that state of affairs is presented to us as being worthy of indignation. To feel anger is not to judge that something is worthy of indignation but to feel that it is. The evaluation is wrapped up in the very mode of presentation. This is presumably what Jonathan Mitchell refers to as “feeling-​towards-​value” (Mitchell 2021). Bennett Helm calls emotions so understood “felt evaluations” (Helm 2001; 2002; 2009). Robert Roberts calls them “concern-​based construals” (Roberts 1988; 2003). On this understanding of emotions, emotions have a kind of evaluative force. Plausibly other kinds of affective states display other species of evaluative force. For instance, desires may present their targets as good to bring about; feelings of shame may present the subject’s own self as worthy of rejection; the pangs of conscience may present actions as being morally obligatory or forbidden. The existence of such felt evaluations lends plausibility to our characterization of seemings. If something can feel good or bad, then why can’t it feel true or false? Finally, as mentioned in a previous section, Teng (2022) finds evidence for something very much like assertive force within the psychological sciences. She describes an epistemic feeling originating from our source-​monitoring mechanisms that, when conjoined with another mental state, makes that state feel veridical. If the mental state in question is a thought with propositional content, then presumably we would describe this epistemic feeling as

128  What Are Seemings? making that propositional content feel true. This sounds exactly like what we have been saying about assertive force. Plausibly, Teng is simply identifying the psychological mechanisms behind what we have thus far been describing from an interior, phenomenological point of view.41 This adds further legitimacy to our appeal to assertive force. 4.3.3  Direct Acquaintance and Seemings

A final and especially illuminating characterization of the feel of truth comes from classical foundationalism. Classical foundationalists explain immediate justification by appeal to direct acquaintance.42 In direct acquaintance, an object is presented to the subject in a way that isn’t mediated by representations. The object is given directly to the subject—​ laid bare before the subject’s conscious mind. Indeed, being aware of an object via representations is possible only because of this more direct form of awareness. How else do we become aware of the mental representation itself? If we can become aware of representations only via more representations, then we quickly set off an infinite regress. This is prevented only by admitting a more direct kind of awareness that theorists call “acquaintance.” Acquaintance is the kind of awareness one has of conscious things simply because one is conscious of them (and not because one is reflecting on one’s conscious experience). Say, for instance, that you feel a dull ache in your toe. You are directly acquainted with this painful feeling, even if you do not take the time to reflect on the fact that you have it. If we do choose to reflect on our conscious states, however, then classical foundationalism looks to acquaintance to help immediately justify our beliefs about them. For reasons I discuss in Chapter 6, §6.2, mere acquaintance with a fact is not enough to justify belief in it (a “fact” being a state of affairs that obtains). Instead, the subject also needs to be acquainted with both a thought (a mental state with propositional content) and the correspondence between that thought and the fact it describes (Fumerton 1995, BonJour and Sosa 2003). Returning to the ache in your toe, to justifiably believe that you are in pain requires (i) acquaintance with that pain; (ii) acquaintance with the thought that you are in pain; and (iii) acquaintance with the correspondence or “match” between your thought and the pain it describes. You will have acquaintance with your pain simply by consciously feeling that pain. Similarly, you will be acquainted with the thought that you are in pain simply by entertaining that thought. But what are we to make of this last requirement? Certainly there is something familiar to us in this description. We can, for instance, draw our attention back and forth between our thought and our pain and, in doing so, we can “just see” that the one corresponds to the other. There is a way it feels to be acquainted

What Are Seemings?  129 with this correspondence between thought and fact. That feeling, I propose, is the feel of truth. Indeed, the name, “the feel of truth,” could hardly be more fitting. If truth is a kind of correspondence between a proposition and the world,43 then feeling that one is aware of the correspondence between a proposition and the world is, almost by definition, what it feels like to be aware of that proposition’s truth. My suggestion is that when a thought has assertive force, it amounts to this thought being accompanied by the feeling that we are directly aware of its content’s correspondence with reality.44 To be clear, we may not actually be aware of its correspondence with reality; we may only feel like we are aware of it. Imagine a case where you feel an itch. The sensation begins to swell, becoming sharp and poignant before dying down again. Let’s stipulate that even at the height of its intensity, this feeling remains an itch rather than a pain. However, when the sensation is at its climax, you form the thought, “I am in pain,” and it feels as though you are acquainted with the correspondence between your thought and the fact it describes. In actual fact, however, you are not acquainted with any such thing. You cannot be acquainted with something that does not exist, and there is no pain to be acquainted with. Neither, then, can there be any acquaintance with the correspondence between your thought and the world, since no such correspondence obtains. Nevertheless, it can still feel as though you are acquainted with this correspondence. In order for our thoughts to feel true, and so display the kind of forcefulness characteristic of seemings, do we also have to be acquainted with the fact to which those thoughts supposedly correspond—​that is, their truth-​maker? Smithies thinks so. He says that the kind of force characteristic of seemings “is the kind of phenomenal character you have when it seems to you that you’re presented with the very things that make your experience true” (Smithies 2019, 93). I disagree. Some seemings come with an awareness of their own truth-​makers. For instance, when it seems to us that we are in pain, we are directly acquainted with the fact that our thought is about. In other cases, however, this is not the case. For instance, it seems to me that no squirrel is as big as my car. Here I am directly acquainted with a thought, and it feels as though I am acquainted with its correspondence to reality; however, I do not feel as though I am directly acquainted with the state of affairs there being no squirrel as big as my car. It’s not clear that there even is a truth-​maker for that thought. To give another example, say we intuit that if A is bigger than B, and B is bigger than C, then A is bigger than C. It’s controversial whether we are aware of any truth-​maker for this proposition. What isn’t controversial, however, is that this principle seems true. All a seeming requires, then, is the thought and the feeling that this thought corresponds to reality. Awareness of a truth-​maker is optional.

130  What Are Seemings? Of course, there could still be something special about seemings where we feel directly acquainted with a truth-​ maker for its content. Some epistemologists will argue that only these kinds of seemings are capable of immediately justifying belief. Though I argue against such a position in Chapter 9, it will be helpful to give a special name to such seemings. Following McCain and Moretti, we will call them “presentational appearances” or, better yet, “presentational seemings” (McCain and Moretti 2021, 78–​80). They seem to derive this name from Eli Chudnoff, who describes these kinds of seemings as having a special sort of “presentational phenomenology” (Chudnoff 2011a, 2013). Smithies too calls the special phenomenal character belonging to such seemings “presentational force” (Smithies 2019, 93–​94). I am happy to go along these labels, so long as we acknowledge that not all seemings are presentational seemings. On the direct acquaintance view of seemings articulated here, seemings end up being complex mental states composed of at least two more basic states: the thought that p and felt acquaintance with its correspondence to reality. The former provides the propositional content of the seeming; the latter makes that content feel true. This separation explains our observation in §4.3.2 that assertive force is not a kind of representational content but a mode in which representational content is entertained. It also fits rather well with the metacognitive account of phenomenal force (Teng 2022), since that account attributes the feeling of truth to metacognitive states coming alongside our first-​order thoughts. 4.3.4  Varieties of Seemings

In drawing to a close, I want to highlight the many different kinds of seemings we possess. We have already touched upon several. We have, for instance, discussed perceptual seemings at length. We have also discussed introspective seemings, such as the seeming that we are in pain. It is important to emphasize that the conceptual awareness we have of our conscious states through introspective seemings is distinct from the direct awareness we have of our conscious states merely by being in them. This is analogous to perceptual seemings, through which we become conceptually aware of physical states of affairs that we were already non-​conceptually aware of through sensory appearances. Extending the parallel, some introspective seemings will deploy demonstrative concepts that point to the conscious states with which we are directly acquainted—​for example, it seems to us that that is pain. These introspective seemings will then be partially constituted by, or at least inseparable from, the existence of those conscious states. Another kind of seeming already mentioned is rational or intellectual seemings. Rational seemings originate from a set of cognitive abilities we

What Are Seemings?  131 call “reason” and are commonly called “intuitions.”45 The phenomenology that accompanies rational seemings is generally much more meager than that surrounding perceptual or introspective seemings. With certain geometrical intuitions, my rational seeming might show up alongside some quasi-​sensory phenomenology—​intimations of figures and the like—​but most rational seemings aren’t accompanied by any sensory phenomenology at all (quasi or otherwise). This is not to say that they have no phenomenal character. They do. That character is simply exhausted by what it feels like to entertain the seeming’s propositional content and the feeling that such content corresponds to the way things really are. As McCain and Moretti explain: Seemings are thought of as experiences because they have a specific cognitive phenomenology quite independently of their also having, or being accompanied by, a rich phenomenology, such as that typical of sensory experience. (McCain and Moretti 2021, 61) Memorial seemings are another important type. Some memorial seemings may be more like perceptual seemings (those associated with episodic memory), whereas others may be more like rational seemings (those associated with semantic memory). All of them, however, will have propositional content that concerns states of affairs in the past.46 There may be many other species of seemings as well: such as inferential seemings (Huemer 2016), moral seemings (Huemer 2005), and religious seemings (Tucker 2011), each set apart by the kinds of propositions they take as their content or the manner in which they are produced. There will be healthy debate over whether such species are reducible to the four fundamental kinds listed above.47 There are still other seemings that don’t fit naturally into any of the categories named thus far: e.g., a random hunch or “lucky feeling” in which it seems to you that you will win the lottery. Different kinds of seemings can feel quite different from one another. Nevertheless, they all share the same basic core: they have propositional content which feels true (or false). We’ve covered a lot of groundwork in this chapter. To quickly review, I have argued that seemings are best understood as sui generis experiences with propositional content and assertive force. The main alternatives to this understanding of seemings all proved inadequate to our pre-​theoretical understanding of seemings and their role in our cognitive lives. Despite this, some remain suspicious of the Experience View because they find seemings so understood to be absent, weird, or under-​developed. I addressed all of these worries, first, by pointing to the great cloud of witnesses, both now

132  What Are Seemings? and throughout the tradition, who acknowledge something like experiential seemings; and second, by giving a detailed explication of experiential seemings and their relationship to other mental states. In particular, I described the assertive force of seemings and suggested that this force be understood as feeling like one is directly aware of the correspondence between propositional content and the world it describes. With this understanding of seemings in place, I am finally in a position to develop my argument for seemings foundationalism. The argument stretches out over the remainder of the book. I defend the thesis that all justification stems from immediate justification in the next chapter, and that all immediate justification comes from all and only seemings in the chapters after that. These arguments will depend crucially on adopting the view of seemings defended here. Not every aspect of my account is essential, but the plausibility of the forthcoming arguments requires at least that seemings (i) have propositional content and (ii) that this content feels true. Notes 1 Tucker tags Bealer (2000), Chudnoff (2011a), Cullison (2010), Huemer (2001; 2007), Lycan (2013), Pryor (2000), Skene (2013), and Tucker (2010) as proponents of this view. I would add Bedke (2008), Beillard (2016), Bergmann (2013a; 2013b; 2021), Brogaard (2013), Dougherty (2014), Farley  (2017), Gage  (2018), Gage and McAllister (2020), Koksvick (2011), Markie (2013), McCain (2012), McGrath (2013), Moretti (2020), Pust (2000), Rosen (2001), and Smithies (2019). 2 Tucker cites early Lycan (1988) and Swinburne (2001) as proponents of the Belief View. It is also worth including in this camp those who think of intuitions as beliefs since intuitions are often taken to be one kind of seeming. These include Devitt (2006), Goldman and Pust (1998), Gopnik and Schwitzgebel (1998), Kornblith (1998), Lewis (1983), and Ludwig (2007) among others. Chudnoff notes that conceiving of intuitions as beliefs (or inclinations to believe) is especially prominent among those who focus on the experimental data concerning intuitions. See, for example, some of the essays in French and Wettstein (2007) and Knobe and Nichols (2008). 3 Tucker points to Rogers and Matheson (2011), Sosa (1998; 2007), and once again Swinburne (2001) as proponents of this view. I would add Dogramaci (2013), Machery (2017), and Yablo (1993). Taylor (2015) defends this view though may not endorse it. We should also include those like Earlenbaugh and Molyneux (2009) and (perhaps) Williamson (2004; 2007) who think that intuitions are inclinations to believe. 4 See Conee (2013), Tooley (2013), and Wagner (forthcoming). 5 This appears to be the same strategy used in many of the most prominent discussions of seemings, such as Huemer (2007), Chudnoff (2011b), and Cullison (2010).

What Are Seemings?  133 6 If seemings foundationalism is true, then the fact that we sometimes use “it seems that p” to report having sufficient evidence for p is not coincidental. For if all foundational evidence consists in seemings, then any proposition with on-​ balance evidential support must be indicated to be true by other propositions for which we do have seemings. Thus, p will still be supported by things that seem true. And it is natural to refer to this situation loosely as “p’s seeming true.” 7 Among those who raise this objection are Bealer (2000), Bergmann (2013b), Cullison (2010), Chudnoff (2011b), Huemer (2001; 2007), Koksvick (2011), and Tucker (2013). 8 An even more common example is the Müller–​Lyer illusion: the lines continue to seem different lengths even though I do not believe this. In the past, I have used that example myself. I have now moved away from it, however, since I find it unclear how we are using seems-​language in this case. Plausibly, when we say, “it seems that one line is longer than the other,” we are operating with what Chisholm calls the comparative use of “seems” (Chisholm 1957, 44ff)—​that is, we may just be saying that we find ourselves with the kind of sensory phenomenology that is typically displayed when lines of unequal length are present. But this is to report ourselves as having a certain kind of sensory appearance, which I go on to differentiate from perceptual seemings in §4.3. The Naïve Comprehension Axiom provides a clearer example since there is no sensory phenomenology involved. 9 It can also be the case that one believes p when p does not seem true (contra Lyons 2013, 24). One may believe that some counterintuitive theorem in mathematics is true—​say, that the sum of the natural numbers is –​1/​12—​even though this certainly doesn’t seem true, and even seems false. (Notice the claim is not that this proposition didn’t seem true at an earlier time before one believed it. The claim is that it still doesn’t seem true even though one believes it. C.f. Lyons 2013, 24–​25.) However, this in itself is not a strong objection to the Belief View because that view may say that seemings are identical only to a certain kind of belief. 10 See Koksvick (2011, 44ff) for more discussion on these matters. 11 Lyons (2009, ­chapter 3, section 2) offers a more detailed account of how this is possible. 12 This argument is usually given against identifying seemings with inclinations to believe, but it applies mutatis mutandis to beliefs. See §4.1.2. 13 Above, Lyons (2009; 2013) argues that a token representation, R, can simultaneously possess the properties of being a seeming and of being a belief, in which case the seeming and the belief are the same state. Would this prevent that seeming from explaining why we have the belief in those cases? I don’t think so. To explain why we have this belief, we must explain how R came to possess the property of being a belief, and R’s property of being a seeming could play an explanatory role in how R came to possess the new property of being a belief. A quarterback’s pass can gain the property of being a touchdown pass partially by virtue of possessing other properties such as having a certain velocity, initial direction, or flight pattern. Similarly, R can gain the property of being a belief partially in virtue of possessing the property of being a seeming. Furthermore,

134  What Are Seemings? I do not think there is any problem with a seeming justifying or serving as evidence for belief in cases where the same token representation is both a seeming and a belief, contra Lyons (2013, 24). Just as an action’s having the property of, say, bringing about a lot of pleasure can, in conjunction with other things, bring it about that this action is morally justified, so R’s having the property of being a seeming can bring it about both that R has the property of being epistemically justified. 14 I would classify Earlenbaugh and Molyneux as proponents of the Inclination View, though they only claim that intuitions (rather than seemings more generally) are inclinations to believe. 15 See also Sosa (2007, 49–​50), on prima facie versus resultant attractions. 16 Chudnoff (2011b) is explicit about this with respect to other standard examples like the Müller–​Lyer illusion. 17 Chudnoff (2011b, 632–​634) also stresses that Earlenbaugh and Molyneux need an error theory for their response to be plausible. 18 Notice that Taylor’s response applies mutatis mutandis to the parallel argument against the Belief View. The same is true of my response to Taylor. 19 Taylor (2015, 17–​20). See Chisholm (1957, 44ff) on the comparative use of “seems.” See also Huemer (2013, section 1.3) for an argument that there is no need to distinguish multiple usages of “seems” or “appears” as Chisholm does. 20 Even proponents of the inclination view admit this (see Taylor 2015, 363–​364). 21 See Bergmann (2013b, 156), Cullison (2010, 264–​265), Huemer (2007, 31), and Tolhurst (1998, 297–​298). 22 Or, perhaps also, the apparent truth of propositions that apparently support the proposition one is inclined to believe. 23 Conee does not endorse the view but merely offers it as a suggestion (Conee 2013, 54; see also 2004, 15). Others, however, have taken Conee up on his suggestion (e.g., Wagner forthcoming), and so the view deserves our attention here. 24 There is one exception. We better not understand the taking relation as its seeming to S that she has sufficient evidence for p. If we do, then the Taking-​ Evidence View would define seemings in terms of other seemings, and so cannot serve as a general account of those states. At best, the Taking-​Evidence View would characterize a certain kind of higher-​order seeming which is different from some more fundamental kind. And it is this more fundamental kind of seeming that interest us here. 25 See also the example of Hume in Huemer (2013, 335). 26 See Huemer (2013, 335–​336) for more discussion on this point. 27 Sometimes when we say that a proposition “seems true” we are reporting that we have an occurrent seeming in its truth. Other times we are saying only that we possess a seeming in its truth, where that includes non-​occurrent seemings that have been stored in our minds. In this latter sense, obvious propositions can “seem true” to us perpetually, even when we are not attending to them at all. 28 See the proponents of the Experience View cited above. 29 Some who describe propositions feeling true (or something comparable to this) include, just to name a few, Audi (2013), Heck (2000), Kriegel (2021), Plantinga (1979; 1993), Pollock (1974), Roberts (1988), and Tolhurst (1998).

What Are Seemings?  135 Moreover, I will argue in §4.3.3 that classical foundationalists who appeal to acquaintance—​specifically, to acquaintance with a thought and its correspondence to reality—​are de facto acknowledging seemings in doing so. Thus, we can add people like BonJour (1998), DePoe (2021), Fumerton (1995), and Hasan (2013) to the list as well. 30 Moreover, the precise propositional content of a cognitive impression will depend on the skill and knowledge of the perceiver. An art expert will have a different sensory cognitive impression than a novice will. See Løkke (2015, 74). 31 The Stoics are plausibly interpreted as holding a kind of disjunctivist theory of perception (see Reed 2002). 32 I take it that Locke is not (contra Plantinga 1993, 105) describing any actual sensory phenomenology that is supposed to accompany consideration of the proposition, but is speaking metaphorically. 33 And in the Eastern canon as well, though I am not able to speak with any authority about its prevalence there. 34 As I have been doing throughout, I will focus primarily on seemings-​true—​that is, the states wherein it seems true to us that p. However, our discussion will also apply to seemings-​false—​that is, the states wherein it seems false to us that p—​with suitable changes. 35 Having conceptual content does not entail having propositional content. Propositional content is conceptual content that can be true or false. It is possible for a state to have conceptual content that does not amount to propositional content because it cannot be true or false (simple ideas are prime examples). However, for a state to conceptually represent some state of affairs, that state must have propositional content specifically. 36 I prefer to reserve the term “sensations” to refer to the sensory phenomenology itself—​the sensory qualia, if you will—​that belongs to sensory appearances. 37 For more on the importance of conceptual content for justification, see Chapter 6. 38 Another argument for conceptualism about perception—​one that has not yet been developed in the literature—​could appeal to the role of perception in concept acquisition. Plausibly, we can only learn how to deploy concepts, and so acquire them, by deploying them (in something like the manner discussed in McAllister 2021). Thus, if we are supposed to acquire our concepts through perceptual experience, our perceptual experience must involve the deployment of concepts. On the other hand, certain concepts seemingly couldn’t be deployed in the required manner unless there was some prior, non-​conceptual form of awareness for them to piggie-​back on (Campbell 2002, Roskies 2008; 2010). Dual-​component theory thus allows us to accept the conceptualist argument—​ perception can play the required role in conceptual acquisition by virtue of perceptual seemings—​while also acknowledging the inadequacy of pure conceptualism. 39 Seemings-​false have their own kind of force that complements assertive force. Instead of asserting their content, they deny it—​ presenting that content as failing to describe how things really are. We can call this “dismissive force.” When a state with dismissive force has propositional content, then that content feels false.

136  What Are Seemings? 40 In the past, I (following others) called experiential seemings “propositional attitudes” because they have propositional content. However, if propositional attitudes also require that the subject take some sort of active stance toward the proposition in question, then seemings are not propositional attitudes after all. That being said, I am not confident that this is a requirement of propositional attitudes. 41 Because Teng’s account is operating on a different level of explanation, I do not see it as directly competing with the descriptions offered above or in the following section. This is not to say that conflict is impossible. Certainly there could be interior explanations of assertive force that do not harmonize with Teng’s account (or with whatever accounts emerge from our best science), but I do not see that as being the case here. To the contrary, I think the account of assertiveness offered in this chapter meshes rather well with Teng’s. 42 See Fumerton (1995) on direct acquaintance. 43 I am not hereby committing myself to the correspondence theory of truth narrowly construed. I am referring simply to the general and relatively uncontroversial idea that, as Merricks says, “truth depends on the world” (Merricks 2009). 44 Likewise, when a thought has dismissive force, it feels as though we are directly aware of a mismatch between the content of that state and what is actually the case. 45 Some wish to reserve the term “intuitions” for seemings whose content is presented as necessarily true or necessarily false. It’s unclear whether we are to take necessity to be a part of the representational content itself (e.g., it seems that p where p =​necessarily, squares are not circles) or whether this is supposed to be included in the representational force with which that content is presented (e.g., it seems that p where p =​squares are not circles, and p is presented as being necessarily true). On the latter option, there is a special kind of assertive force—​modal assertive force—​in which the proposition not only feels true but necessarily true. In any case, it strikes me as a purely semantic debate whether we call all rational seemings “intuitions” or only those whose content is purportedly necessary. 46 In the past with respect to one’s personal time. See Lewis (1976). 47 Some might argue that all of the suggested kinds of seemings can be reducible to subspecies of rational seemings; others will resist. As an example, see the discussion over whether religious seemings are a sui generis kind of seeming (as seems presumed in Tucker 2011) or whether they are reducible to rational seemings (as in McAllister and Dougherty 2019).

References Allen, James. 2020. “Carneades.” The Stanford Encyclopedia of Philosophy. 2020. https://​plato.stanf​ord.edu/​archi​ves/​fall2​020/​entr​ies/​carnea​des/​. Audi, Robert. 2013. “Doxastic Innocence: Phenomenal Conservatism and Grounds of Justification.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 181–​201. New York: Oxford University Press.

What Are Seemings?  137 Bealer, George. 2000. “A Priori Knowledge.” The Proceedings of the Twentieth World Congress of Philosophy 5: 1–​12. Bedke, Matthew S. 2008. “Ethical Intuitions: What They Are, What They Are Not, and How They Justify.” American Philosophical Quarterly 45 (3): 253–​270. Beillard, Julien. 2016. “Phenomenal Conservatism, Reflection and Self-​Defeat.” Logos and Episteme 7 (2): 187–​199. Bengson, John, Enrico Grube, and Daniel Z. Korman. 2011. “A New Framework for Conceptualism.” Noûs 45 (1): 167–​189. Bergmann, Michael. 2013a. “Externalist Justification and the Role of Seemings.” Philosophical Studies 166 (1): 163–​184. —​—​—​. 2013b. “Phenomenal Conservatism and the Dilemma for Internalism.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 154–​178. Oxford: Oxford University Press. —​—​—​. 2021. Radical Skepticism and Epistemic Intuition. Oxford: Oxford University Press. Bermúdez, José Luis. 1998. The Paradox of Self-​Consciousness: Representation and Mind. Cambridge, MA: MIT Press. BonJour, Laurence. 1998. In Defense of Pure Reason. Cambridge: Cambridge University Press. BonJour, Laurence, and Ernest Sosa. 2003. Epistemic Justification: Internalism Vs. Externalism, Foundations Vs. Virtues. Hoboken, NJ: Wiley-​Blackwell. Brogaard, Berit. 2013. “Phenomenal Seemings and Sensible Dogmatism.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 270–​289. Oxford: Oxford University Press. Campbell, John. 2002. Reference and Consciousness. Oxford: Oxford University Press. Caston, Victor. 2009. “Phantasia and Thought.” In A Companion to Aristotle, edited by Georgios Anagnostopoulos, 322–​334. Hoboken, NJ: Wiley-​Blackwell. Chisholm, Roderick M. 1957. Perceiving: A Philosophical Study. Ithaca, NY: Cornell University Press. Chudnoff, Elijah. 2011a. “The Nature of Intuitive Justification.” Philosophical Studies 153 (2): 313–​333. —​—​—​. 2011b. “What Intuitions Are Like.” Philosophy and Phenomenological Research 82 (3): 625–​654. —​—​—​. 2013. Intuition. Oxford: Oxford University Press. Cicero. 1933. “Academica, Book 2.” Translated by H Rackham. Attalus. 1933. www.atta​lus.org/​transl​ate/​acad​emic​a2b.html. Conee, Earl. 1998. “Seeing the Truth.” Philosophical and Phenomenological Research 58 (4): 847–​857. —​ —​ —​ . 2004. “First Things First.” In Evidentialism: Essays in Epistemology. Oxford: Oxford University Press. —​—​—​. 2013. “Seeming Evidence.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 52–​68. Oxford: Oxford University Press. Cullison, Andrew. 2010. “What Are Seemings?” Ratio 23 (3): 260–​274.

138  What Are Seemings? —​ —​ —​ . 2013. “Seemings and Semantics.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 33–​51. Oxford: Oxford University Press. DePoe, John. 2021. “Justification by Acquaintance.” Synthese 199: 7555–​7573. Descartes, René. [1641] 2006. Meditations, Objections, and Replies. Translated by Roger Ariew and Donald A. Cress. Indianapolis, IN: Hackett Publishing. Devitt, Michael. 2006. “Intuitions in Linguistics.” British Journal for the Philosophy of Science 57 (3): 481–​513. Dogramaci, Sinan. 2013. “Intuitions for Inferences.” Philosophical Studies 165 (2): 371–​399. Dougherty, Trent. 2014. “Faith, Trust, and Testimony: An Evidentialist Account.” In Religious Faith and Intellectual Virtue, edited by Laura Frances Callahan and Timothy O’Connor, 97–​123. Oxford: Oxford University Press. Earlenbaugh, Joshua, and Bernard Molyneux. 2009. “Intuitions Are Inclinations to Believe.” Philosophical Studies 145 (1): 89–​109. Evans, Gareth. 1982. The Varieties of Reference. Oxford: Oxford University Press. Farley, R. M. 2017. “The Sellarsian Dilemma: Not What It Seems.” Southwest Philosophy Review 33 (1): 115–​123. Feldman, Richard. 2003. Epistemology. Hoboken, NJ: Prentice Hall. French, Peter A., and Howard K. Wettstein. 2007. Philosophy and the Empirical. Hoboken, NJ: Wiley-​Blackwell. Fumerton, Richard. 1995. Metaepistemology and Skepticism. Lanham, MD: Rowman & Littlefield. Gage, Logan Paul. 2018. “Can Experience Fulfill the Many Roles of Evidence?” Quaestiones Disputatae 8 (2): 87–​111. Gage, Logan Paul, and Blake McAllister. 2020. “The Phenomenal Conservative Approach to Religious Epistemology.” In Debating Christian Religious Epistemology: An Introduction to Five Views on the Knowledge of God, edited by John M. DePoe and Tyler McNabb, 61–​81. London: Bloomsbury Academic. Goldman, Alvin I., and Joel Pust. 1998. “Philosophical Theory and Intuitional Evidence.” In Rethinking Intuition: The Psychology of Intuition and Its Role in Philosophical Inquiry, edited by Michael Depaul and William Ramsey, 179–​200. Lanham, MD: Rowman & Littlefield. Gopnik, Alison, and Eric Schwitzgebel. 1998. “Whose Concepts Are They, Anyway? The Role of Philosophical Intuition in Empirical Psychology.” In Rethinking Intuition, edited by M. R. DePaul and William Ramsey, 75–​91. Lanham, MD: Rowman & Littlefield. Hasan, Ali. 2013. “Phenomenal Conservatism, Classical Foundationalism, and Internalist Justification.” Philosophical Studies 162 (2): 119–​141. Heck, Richard. 2000. “Nonconceptual Content and the ‘Space of Reasons.’” Philosophical Review 109 (4): 483–​523. Helm, Bennett, W. 2001. Emotional Reason: Deliberation, Motivation, and the Nature of Value. Cambridge, MA: Cambridge University Press. —​ —​ —​ . 2002. “Felt Evaluations: A Theory of Pleasure and Pain.” American Philosophical Quarterly 39 (1): 13–​30. —​—​—​. 2009. “Emotions as Evaluative Feelings.” Emotion Review 1 (3): 248–​255.

What Are Seemings?  139 Huemer, Michael. 2001. Skepticism and the Veil of Perception. Lanham, MD: Rowman & Littlefield. —​—​—​. 2005. Ethical Intuitionism. London: Palgrave Macmillan. —​ —​ —​ . 2007. “Compassionate Phenomenal Conservatism.” Philosophy and Phenomenological Research 74 (1): 30–​55. —​ —​ —​ . 2013. “Phenomenal Conservatism Über Alles.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 328–​350. Oxford: Oxford University Press. —​ —​ —​ . 2016. “Inferential Appearances.” In Intellectual Assurance: Essays on Traditional Epistemic Internalism, edited by Brett Coppenger and Michael Bergmann, 144–​160. Oxford: Oxford University Press. Irwin, Terence. 1988. Aristotle’s First Principles. Oxford: Oxford University Press. Knobe, Joshua, and Shaun Nichols. 2008. Experimental Philosophy. Oxford: Oxford University Press. Koksvick, Ole. 2011. “Intuition.” Dissertation, Canberra: The Australian National University. Kornblith, Hilary. 1998. “The Role of Intuition in Philosophical Inquiry: An Account with No Unnatural Ingredients.” In Rethinking Intuition: The Psychology of Intuition and Its Role in Philosophical Inquiry, edited by M. DePaul and W. Ramsey, 129–​141. Lanham, MD: Rowman & Littlefield. Kriegel, Uriah. 2021. “The Structure of Phenomenal Justification.” Australasian Journal of Philosophy. Lewis, David K. 1976. “The Paradoxes of Time Travel.” American Philosophical Quarterly 13 (2): 145–​152. —​—​—​. 1983. “New Work for a Theory of Universals.” Australasian Journal of Philosophy 61 (4): 343–​377. Locke, John. [1689] 1975. An Essay Concerning Human Understanding, edited by P. H Nidditch. Oxford: Clarendon Press. Løkke, Håvard. 2015. Knowledge and Virtue in Early Stoicism. Vol. 10. Studies in the History of Philosophy of Mind. New York: Springer. Ludwig, Kirk. 2007. “The Epistemology of Thought Experiments: First Person Versus Third Person Approaches.” Midwest Studies in Philosophy 31 (1): 128–​159. Lycan, William G. 1988. Judgement and Justification. Cambridge: Cambridge University Press. —​ —​ —​ . 2013. “Phenomenal Conservatism and the Principle of Credulity.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 293–​305. Oxford: Oxford University Press. Lyons, Jack C. 2009. Perception and Basic Beliefs: Zombies, Modules and the Problem of the External World. Oxford: Oxford University Press. —​—​—​. 2013. “Should Reliabilists Be Worried About Demon Worlds?” Philosophy and Phenomenological Research 86 (1): 1–​40. Machery, Edouard. 2017. Philosophy Within Its Proper Bounds. Oxford: Oxford University Press.

140  What Are Seemings? Markie, Peter. 2013. “Searching for True Dogmatism.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 248–​269. Oxford: Oxford University Press. Martin, Michael G. F. 1992. “Perception, Concepts, and Memory.” Philosophical Review 101 (4): 745–​763. McAllister, Blake. 2021. “Conceptualism and Concept Acquisition.” Theoria 87 (1): 69–​86. McAllister, Blake, and Trent Dougherty. 2019. “Reforming Reformed Epistemology: A New Take on the Sensus Divinitatis.” Religious Studies 55 (4): 537–​557. McCain, Kevin. 2012. “Against Hanna on Phenomenal Conservatism.” Acta Analytica 27 (1): 45–​54. McCain, Kevin, and Luca Moretti. 2021. Appearance and Explanation: Phenomenal Explanationism in Epistemology. Oxford: Oxford University Press. McDowell, John. 1994. Mind and World. Cambridge, MA: Harvard University Press. McGrath, Matthew. 2013. “Phenomenal Conservatism and Cognitive Penetration: The Bad Basis Counterexamples.” In Seemings and Justification, edited by Chris Tucker, 225–​247. Oxford: Oxford University Press. Merricks, Trenton. 2009. Truth and Ontology. Oxford: Oxford University Press. Mitchell, Jonathan. 2021. Emotion as Feeling Towards Value: A Theory of Emotional Experience. Oxford: Oxford University Press. Moretti, Luca. 2020. Seemings and Epistemic Justification: How Appearances Justify Beliefs. New York: Springer. Peacocke, Christopher. 2001. “Does Perception Have a Nonconceptual Content?” Journal of Philosophy 98 (5): 239–​264. Plantinga, Alvin. 1979. “The Probabilistic Argument from Evil.” Philosophical Studies 35 (1): 1–​53. —​—​—​. 1993. Warrant and Proper Function. Oxford: Oxford University Press. Pollock, John. 1974. Knowledge and Justification. Princeton, NJ: Princeton University Press. Pryor, James. 2000. “The Skeptic and the Dogmatist.” Noûs 34 (4): 517–​549. Pust, Joel. 2000. Intuitions as Evidence. New York: Garland. Reed, Baron. 2002. “The Stoics’ Account of the Cognitive Impression.” Oxford Studies in Ancient Philosophy 23: 147–​180. Reid, Thomas. [1785] 2002. Essays on the Intellectual Powers of Man, edited by Derek R. Brooks. University Park, PA: Pennsylvania State University Press. Roberts, Robert C. 1988. “What an Emotion Is: A Sketch.” Philosophical Review 97 (April): 183–​209. —​—​—​. 2003. Emotions: An Essay in Aid of Moral Psychology. Cambridge, MA: Cambridge University Press. Rogers, Jason, and Jonathan Matheson. 2011. “Bergmann’s Dilemma: Exit Strategies for Internalists.” Philosophical Studies 152 (1): 55–​80. Rosen, Gideon. 2001. “Nominalism, Naturalism, Epistemic Relativism.” Noûs 35 (s15): 69–​91. Roskies, Adina L. 2008. “A New Argument for Nonconceptual Content.” Philosophy and Phenomenological Research 76 (3): 633–​659.

What Are Seemings?  141 —​—​—​. 2010. “ ‘That’ Response Doesn’t Work: Against a Demonstrative Defense of Conceptualism.” Noûs 44 (1): 112–​134. Sedley, David. 1998. “Stoicism.” In Routledge Encyclopedia of Philosophy. Abington, UK: Routledge. Siegel, Susanna, and Nicholas Silins. 2015. “The Epistemology of Perception.” In Oxford Handbook of Philosophy of Perception, edited by Mohan Matthen, 781–​811. Oxford: Oxford University Press. Skene, Matthew. 2013. “Seemings and the Possibility of Epistemic Justification.” Philosophical Studies 163 (2): 539–​559. Smith, A.D. 2002. “The Problem of Perception.” Philosophical Quarterly 54 (217): 640–​642. Smithies, Declan. 2019. The Epistemic Role of Consciousness. New York: Oxford University Press. Sosa, Ernest. 1998. “Minimal Intuition.” In Rethinking Intuition, edited by Michael DePaul and William Ramsey, 257–​269. Lanham, MD: Rowman & Littlefield. —​—​—​. 2007. A Virtue Epistemology: Apt Belief and Reflective Knowledge. Oxford: Oxford University Press. Swinburne, Richard. 2001. Epistemic Justification. Oxford: Oxford University Press. Taylor, Samuel A. 2015. “What Seemings Seem to Be.” Episteme 12 (3): 363–​384. Teng, Lu. 2022. “A Metacognitive Account of Phenomenal Force.” Mind and Language. Tolhurst, William E. 1998. “Seemings.” American Philosophical Quarterly 35 (3): 293–​302. Tooley, Michael. 2013. “Michael Huemer and the Principle of Phenomenal Conservatism.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 306–​327. Oxford: Oxford University Press. Tucker, Chris. 2010. “Why Open-​Minded People Should Endorse Dogmatism.” Philosophical Perspectives 24 (1): 529–​545. —​ —​ —​ . 2011. “Phenomenal Conservatism and Evidentialism in Religious Epistemology.” In Evidence and Religious Belief, edited by Kelly James Clark and Raymond J. VanArragon, 52–​73. Oxford: Oxford University Press. —​—​—​. 2013. “Seemings and Justification: An Introduction.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 1–​29. Oxford: Oxford University Press. Tye, Michael. 1995. Ten Problems of Consciousness: A Representational Theory of the Phenomenal Mind. Cambridge, MA: MIT Press. Wagner, Verena. forthcoming. “Epistemic Seemings in Inquiry.” In Seemings: New Angles, New Arguments, edited by Kevin McCain and Scott Stapleford. Abington, UK: Routledge. Williamson, Timothy. 2004. “Philosophical ‘Intuitions’ and Scepticism About Judgement.” Dialectica 58 (1): 109–​153. —​—​—​. 2007. The Philosophy of Philosophy. Hoboken, NJ: Wiley-​Blackwell. Yablo, Stephen. 1993. “Is Conceivability a Guide to Possibility?” Philosophy and Phenomenological Research 53 (1): 1–​42.

5 Foundationalism Defended

Here begins my positive case for seemings foundationalism. My focus in this chapter is on foundationalism. Specifically, I will argue that there must be some source of immediate justification if anything is to be justified at all. I do not yet say what that foundational source of justification is or is not (that is the task of Chapters 8 and 9 and 6 and 7, respectively), or even whether there is one at all. My claim in this chapter is just that there needs to be such a foundation if there is to be justification. My argument will be largely indirect. That is, I will examine the possible alternatives to foundationalism—​ infinitism and coherentism—​ and argue that both are decidedly less adequate than foundationalism. Foundationalism faces its own challenge, to be sure—​in particular, how can anything justify without itself requiring justification? But there are plausible responses to that challenge, whereas there is reason to believe that the problems facing infinitism and coherentism are irresolvable. Thus, foundationalism is our best option. 5.1  The Regress Argument for Foundationalism Foundationalism is the view that all justification ultimately stems from immediate justification, but there are different ways of specifying what exactly this means. As a first pass, we might say that all justified doxastic attitudes are either immediately justified or else justified by other beliefs that are themselves immediately justified, but that’s not quite right. Foundationalism does not require that our noetic structures rest on a foundation of immediately justified beliefs specifically. Thus, as a second attempt, we might characterize foundationalism as the view that all justified doxastic attitudes are either immediately justified or else justified by other doxastic attitudes that are themselves immediately justified. When framed in terms of propositional justification rather than doxastic, we get the following:

DOI: 10.4324/9781003281108-6

Foundationalism Defended  143 FN—​S has non-​immediate justification for taking some doxastic attitude toward p only if (and because) S has immediate justification for taking some doxastic attitude toward q. FN is best understood as being about the structure and origins of justificatory support, that is, that quantity enough of which suffices to justify us in taking a particular doxastic attitude toward a proposition. FN says that all non-​immediate justificatory support—​the kind of support that one attitude provides for another—​is derived from immediate justificatory support somewhere down the line. On this view, non-​immediate justification might be thought of as a passing along of the justificatory support received from prior sources, and this is possible only if there is some foundational source of justificatory support which does not depend on any prior sources. Whatever this foundational source is, it is one of the true founts from which justificatory support flows, arguments and inferences being mere conduits through which the reach of such support is extended. FN is neutral with respect to many different kinds of foundationalism. One way in which versions of foundationalism can differ is with respect to how strong immediate justification must be. On one end of the spectrum, classical foundationalism requires that immediate justification for p be strong enough to permit certainty in p, or to render belief in p infallible, or to place p beyond all doubt.1 On the other end, one might allow only for a very weak kind of immediate justification—​one sufficient to justify only some minimally positive level of confidence in p in the state of empty symmetrical evidence.2 FN is compatible with both of these extremes. In later chapters, I will argue for a version of foundationalism that falls between them, but, at the moment, my focus is on FN alone.3 Another way in which versions of foundationalism can differ is with respect to how justificatory support is passed along to other propositions. Some maintain that immediately justified attitudes can, apart from any other justified beliefs, be deployed in inferences to justify new conclusions. Others maintain that immediately justified attitudes cannot be used to justify other propositions unless situated within a larger framework.4 As before, FN makes no claim here. In fact, seemings foundationalism as a whole makes no claim here. The focus of our inquiry is on the foundational sources of justification, not how that justificatory support gets transferred to other propositions. FN is thus a fairly expansive understanding of foundationalism, but I take this to be an advantage. The essence of foundationalism, as I see it, is the idea that the kind of justificatory support that one attitude provides for another is derived from the more immediate justificatory support that certain attitudes enjoy apart from other attitudes. This can be developed

144  Foundationalism Defended in many different ways, all of them rightfully flying under the banner of “foundationalism.” The traditional argument for foundationalism, going back at least to Aristotle, is the regress argument. I develop what I take to be the strongest formulation of this argument here.5 The first step in the argument says that if one belief is to justify another, then the justifying belief must itself be justified. After all, beliefs are the sort of thing that need justification—​they are either held permissibly or not—​and how could a belief held without permission grant us permission to believe something else? Such a belief might commit us to believing something as a matter of consistency. For instance, if we believe that p and q without permission, and we are forced to take a stance toward p, there might be some sense in which we are committed to believing p. But being committed to believe p in this way is not the same thing as being permitted to believe p. It does not guarantee that believing p is in accordance with our epistemic duties as directed toward attaining truth and avoiding falsehood. What it does guarantee is that if believing p and q is in accordance with our epistemic duties, then so is believing p. In other words, holding a belief, in and of itself, can only conditionally justify us in believing something else. But conditional justification is not actual justification, and if our concern is with the latter, then a belief can bestow that status only if it is itself justified. Applying this rule, we see that if belief B1 is justified by B2, then B2 must itself be justified; and if B2 is justified by B3, then, applying our principle once more, B3 must also be justified. This sets off a regress of beliefs. Foundationalism says that all such regresses must come to an end in beliefs that are immediately justified—​that is, justified in a way that does not depend on the justification of any other beliefs. If a regress come to an end in a belief Bn that is not justified, then, according to our rule, Bn–​1 (the belief that Bn conditionally justifies) would not be actually justified, and so neither would Bn–​2 (the belief that Bn–​1 conditionally justifies), and so on back to B1. Is there any alternative to the foundationalist account here? That is, is there some account of how B1 can be justified that does not appeal to an immediately justified belief somewhere down the line? 5.1.1 Infinitism

Infinitism suggests that B1 can be justified so long as the regress is infinite. More specifically, if there is an infinite regress of beliefs available to the subject in which every belief in the chain, Bn, is justified by some further belief, Bn+​1, that has not been featured in the chain before, then and only then are the beliefs in that chain justified (Klein 1999, §I). In mandating that these beliefs be available to the subject, infinitism does not require that the subject actually possess those beliefs, much less that they be occurrent,

Foundationalism Defended  145 but only that the subject be disposed to form such beliefs when called upon (ibid, 308). This qualification is necessary to avoid skepticism since it seems impossible for humans to possess an infinite number of beliefs. There are many problems with infinitism. One problem is that, despite the aforementioned qualification, infinitism seems to usher in widespread skepticism about the justification of our beliefs; and I take that to be a decisive reason to reject the position.6 There are several ways of pressing this skeptical worry. One is to offer reasons for thinking that an infinite series of beliefs isn’t available to us. For instance, when the child (or overly eager philosopher) incessantly asks “why?” over and over, it does not take long to reach a point where we can seemingly go no further. This may be evidence that in fact none of us are disposed to indefinitely form higher-​ order beliefs in the manner required by infinitism. What’s more, it may not even be possible for us to be so disposed (c.f., Klein 1999, 307ff). Such dispositions must be materially realized within the human brain, and there’s only so much room to go around. The problem intensifies when we add that at least some of the beliefs in the regress may (in many cases) need to be increasingly complex as we proceed further back along the regress. Do we have the concepts and mental capacities to handle such beliefs? Finally, even if such an infinite series is feasible with respect to certain kinds of beliefs—​for instance, mathematical ones—​that is no guarantee that the same sort of strategies which apply to those beliefs will work for all of the many different kinds of beliefs for which we take ourselves to have justification. Thus, there are very real concerns that the sort of infinite series of beliefs that infinitism requires for justification are not available to us. And even if such infinite series are available, we clearly don’t use them in the formation of our beliefs. Frequently, our chains of beliefs end in basic beliefs based on nothing more than the fact that their content seems true (see Chapter 8, §8.2.2). This gives us a second way of pressing the skeptical problem for infinitism. For our beliefs will only be doxastically justified if they are properly based on that which propositionally justifies them. But the fact of the matter is that no infinite chain of reasoning is ever the base of our beliefs. Indeed, we can’t base our beliefs on any infinite chain of beliefs because, as mentioned above, the human mind cannot actually possess an infinite number of beliefs, and a belief must be actual in order to base another belief.7 The basing relation is causal (though not merely causal), and a belief cannot cause another belief if it does not exist.8 Infinitists might respond to these objections by attempting to mitigate the skeptical implications of their position. What the above arguments show, they might say, is that, on infinitism, none of our beliefs are fully (i.e., maximally) justified, either propositionally or doxastically; however, infinitism still allows them to attain levels of justification falling short of the max. The longer the regress available to us, and the further it is realized within

146  Foundationalism Defended our minds, the higher the level of propositional or doxastic justification we enjoy, respectively. The problem with this response is that adding beliefs to a regress of finite length does nothing to increase our level of justification in any meaningful sense. At least, this is true if we stick with the rule that beliefs can only lend justificatory support (of any quantity) if they have received it from something prior. For so long as our chain of beliefs remains finite (and immediate justification is not admitted), then the chain will terminate in a final belief Bn that enjoys no justificatory support. It follows that the belief which depends on this final belief (Bn–​1) enjoys no justificatory support, and so on and so forth back to B1. In what sense, then, is a belief supported by a chain of n number of beliefs any better off than one supported by a chain of n–​1 beliefs? Both enjoy no level of justification at all. To illustrate the point, consider a deductive argument. The following argument is invalid. Argument 1 2. p ⊃ q 1. ∴ q

Nor does it become valid if we add another conditional premise: Argument 2 3. o ⊃ p 2. p ⊃ q 1. ∴ q

But is there any sense in which the second argument is closer to being valid than the first? I don’t see how. Likewise, how is a chain of n conditional justifiers any closer to achieving full justification than a chain of n –​1? I suppose infinitists could simply stipulate that chains of greater length provide higher levels of “justification” in some specialized sense, but then it is entirely unclear why we should care about this specialized sense of “justification.” It would not, for instance, address the skeptical worry that, on infinitism, none of our beliefs enjoy any level of justification in the ordinary sense. Our discussion here has foreshadowed a second and even greater problem for infinitism, which is that even if we had an infinite chain of beliefs available to us, and could use these to base our beliefs, that still wouldn’t be enough to justify our beliefs.9 The reasons why have already been hinted at above. By themselves, beliefs are merely conditional justifiers. They pass along justificatory support that they have received from prior sources. In this way, beliefs are, at best, intermediate sources of justification (and unjustified beliefs are not sources of justification at all). This can be contrasted with ultimate sources of justification, which are capable of

Foundationalism Defended  147 bestowing justificatory support without themselves receiving that support from anything prior. Ultimate sources are true originators of justificatory support, whereas intermediate sources merely redistribute it. It should be apparent, then, that a regress consisting only in beliefs, even if infinite, cannot lend justificatory support to anything. For a regress of beliefs is a series without an ultimate source of justificatory support; and without an ultimate source of justificatory support, there are no intermediate sources of justificatory support, for there is nothing for those intermediaries to pass from one to another.10 At best, what a regress of beliefs does is establish relations of conditional justification between the beliefs in the regress—​B1 is justified if B2 is justified, and so on. But relations of conditional justification can never amount to actual justification.11 To see this better, return to our deductive argument from above. We agreed that Argument 1 is invalid. And it does not become valid, nor get any closer to becoming valid, by adding another conditional premise, as we saw in Argument 2. But what if we add an infinite regress conditional premises? Argument 3 … 3. o ⊃ p 2. p ⊃ q 1. ∴ q

As before, not even this would make the argument valid. No set of conditional premises, however large, entails an unconditional conclusion. What would change things, in fact the only thing that could change things, is an unconditional premise—​something that can trigger the antecedent of these conditionals and set the dominos falling. The same is true of conditional justification. No set of conditional justifiers, even an infinite set, can unconditionally justify anything. An unconditional justifier is needed. To be clear, the objection here is not that infinitism leads to skepticism but that it’s contradictory. It says that a belief can be actually justified by a set consisting only in conditional justifiers, which is impossible. Returning to a previous discussion, we see now the deeper reason why talk of less than maximal levels of justification doesn’t make sense on infinitism. For if the infinitist ideal fails to provide any justification, then approximating that ideal does not bring us any closer to doing so either. The last hope for infinitism is to drastically reorient our understanding of justification. On this reorientation, justification is a property not of beliefs proper but of belief systems—​ namely, infinite chains of non-​ repeating beliefs standing in the right kinds of logical relations to one another. Beliefs inherit justification by their participation in such an infinite chain or, at least, to the extent that the system in which they participate approximates

148  Foundationalism Defended such an infinite chain. Reorienting justification in this way prevents us from thinking about beliefs as intermediate sources of justification which pass along justificatory support that they have received from prior sources, and so blocks the theoretical argument developed above. I have two responses to this maneuver. The first is to issue the challenge of explaining why we should care about our beliefs being justified in this revisionary sense. Return once more to Argument 3 above. Someone could stipulate that an argument with an infinite number of conditional premises, each one leading to the next, is “valid,” and that other arguments are “partially valid” to the extent that they approximate this ideal. It’s a free country. But we would then need an explanation for why we should concern ourselves with this revisionary notion of validity. To the contrary, Argument 3 doesn’t look any better to me than Argument 2, nor does Argument 2 look any better than Argument 1. To make matters worse, even if there were some great-​making property that belonged to Argument 3 but not (to the same extent) to Arguments 2 or 1, the revisionist would need to convince us that this was a genuine revision of the concept of validity rather than a blatant replacement of it. Are we still talking about validity or about something else entirely? Applying this to infinitism’s revisionary notion of justification, we must ask ourselves why we should care about our beliefs being justified in this sense. Does extending the chain one notch further before coming to a baseless belief improve our situation in any significant way? Both are ultimately baseless. One just takes a little bit longer to get there. Even if we do find something better about longer chains, we must also ask whether that great-​making property can fairly be identified with the property of justification. Does infinitism’s revisionary notion of justification make such a departure that it unduly co-​opts the language of “justification” to refer to something altogether different? It is hard not to think so. Second, this revisionary move is very similar to the one made by coherentists in response to the regress argument, as we will see momentarily. Both end up with a conception on which justification emerges solely from a large set of beliefs whose members bear appropriate logical relations to one another. Because of this, both will end up sharing the same crucial problems. This is convenient as the objections I level against coherentism throughout the rest of this chapter will also constitute objections to this revisionary version of infinitism. 5.1.2 Coherentism

The other alternative to foundationalism is coherentism. Coherentism says the regress of beliefs ultimately circles back on itself—​that is, some beliefs will end up being cited multiple times in a singular chain. When those beliefs show up again, we must defend them in a similar way as before,

Foundationalism Defended  149 and so the argument repeats itself. What justification requires, then, is a coherent web of beliefs in which each belief is adequately supported by other beliefs in the web. With respect to the regress argument, coherentism escapes the feasibility worries facing infinitism, since the subject need only have a finite number of beliefs, but it remains susceptible to the same deeper problem: namely, it seeks to derive actual justification from a set of exclusively conditional justifiers. If there is no ultimate source of justificatory support for the chain, then there is nothing for the beliefs in that chain to pass along to one another. At best, forming a coherent web of beliefs ensures that if some of the beliefs in that chain were justified, then all the others would be as well; but there isn’t any way to satisfy that antecedent. This is especially apparent given the circularity that features in that reasoning. B1 is supposed to receive its justification from B2 and B2 is supposed to receive it from B1. In such a case, there is no explanation for why either of these beliefs would have any justification to pass along in the first place. The only way for coherentism to address these concerns is to challenge the linear conception of justification at play in the regress argument.12 Instead of justification passing from one belief to another, justification is a property of belief systems that emerges when the constituents of those systems bear the appropriate kind of logical relations to one another. Beliefs inherit their justification through participation in this kind of coherent system. On this more holistic conception of justification, the regress argument fails to convince. This does not mean coherentism is in the clear, however, since this revisionary notion of justification comes with problems of its own. The question simply becomes how these problems stack up against those facing foundationalism. 5.2  Perennial Objections In this section, I lay out the most serious objections to foundationalism on the one hand and coherentism and infinitism on the other. I then argue that foundationalism has the resources to address its primary difficulty while coherentism and infinitism do not. 5.2.1  Two Problems for Coherentism and Infinitism

I will press two objections in this section, first against coherentism and then against infinitism. Both of these objections are illustrated by Plantinga’s famous example of the epistemically inflexible climber: Consider the Case of the Epistemically Inflexible Climber. Ric is climbing Guide’s Wall, on Storm Point in the Grand Tetons; having just

150  Foundationalism Defended led the difficult next to last pitch, he is seated on a comfortable ledge, bringing his partner up. He believes that Cascade Canyon is down to his left, that the cliffs of Mount Owen are directly in front of him, that there is a hawk gliding in lazy circles 200 feet below him, that he is wearing his new Fire rock shoes, and so on. His beliefs, we may stipulate, are coherent. Now add that Ric is struck by a wayward burst of high-​energy cosmic radiation. This induces a cognitive malfunction; his beliefs become fixed, no longer responsive to changes in experience. No matter what his experience, his beliefs remain the same. At the cost of considerable effort his partner gets him down and, in a desperate last-​ditch attempt at therapy, takes him to the opera in nearby Jackson, where the New York Metropolitan Opera on tour is performing La Traviata. Ric is appeared to in the same way as everyone else there; he is inundated by wave after wave of golden sound. Sadly enough, the effort at therapy fails; Ric’s beliefs remain fixed and wholly unresponsive to his experience; he still believes that he is on the belay ledge at the top of the next to last pitch of Guide’s Wall, that Cascade Canyon is down to the left, that there is a hawk sailing in lazy circles 200 feet below him, that he is wearing his new Fire rock shoes, and so on. Furthermore, since he believes the very same things he believed when seated on the ledge, his beliefs are coherent. But surely they have little or no warrant for him. … Clearly, then, coherence is not sufficient for positive epistemic status. (Plantinga 1993, 82) I think it fair to say that most epistemologists, coherentists included, have intuited that the climber’s beliefs are unjustified. Why is this exactly? When we unpack the case, there are actually two distinct intuitions at play here. The first and less acknowledged intuition is that the climber’s ongoing perceptual experiences just do lend credibility to that which they present as being true. In other words, part of what’s going wrong epistemically is that the climber isn’t moved at all by his experiences when it really seems he should be. His duties with respect to the epistemic goal—​that is, attaining true beliefs and avoiding false ones—​seem to require that his attitudes change in accordance with his changing experiences. After all, if the goal is to believe p if and only if p is true, how can the felt awareness of p’s truth not matter even one iota?13 And note that all this remains true even though the hiker does not believe himself to have those experiences (since even his beliefs about his own beliefs remain locked in place). It is something about the experiences themselves, not the hiker’s beliefs about those experiences, that gives them epistemic weight. Thus, Plantinga’s inflexible climber suggests that experiences can and do provide immediate justification and, in turn, that coherence is not necessary for justification.

Foundationalism Defended  151 Kvanvig and Riggs have proposed that coherentists can define coherence relations over certain experiential states with propositional content as well as beliefs—​that is, the system that must be made coherent includes not only our beliefs but our appearances as well (Kvanvig and Riggs 1992). If adopted, this proposal can help the coherentist accommodate the first intuition elicited by Plantinga’s inflexible climber (Poston 2014, ­chapter 1). Unfortunately, this treatment of the case does not avoid conceding the dispensability of coherence. For why is it that we must take experiences into account in theorizing? Why must we give their content credence but not the propositional content of other states (imagining p, hoping that p, being angry that p)? The answer seems to be that experiences lend some kind of credibility to their content that these other states do not; but that is to concede that those experiences provide some level of immediate justificatory support for their content, and we are once again left with the conclusion that coherence is not essential for justification. To be clear, defining coherence relations over experiential states does not on its own commit the coherentist to the possibility of immediate justificatory support (Kvanvig and Riggs 1992, 210–​212). For experiences to immediately justify, those experiences must be, in and of themselves, sufficient to lend justificatory support to their contents. However, the coherentist may understand experiences to be necessary for justification while also insufficient on their own.14 For example, the coherentist may require for justification not only that you have an appearance, but that it also be accompanied by a suitable cadre of beliefs about that appearance—​for example, the belief that we have such an appearance and that such appearances are generally reliable. While this version of coherentism avoids any admission of immediate justification, it subsequently becomes useless in responding to Plantinga’s inflexible climber. For the inflexible climber lacks some of the supplementary beliefs the coherentist will require for justification: in particular, he does not believe himself to be having experiences of an opera since his beliefs are fixed in place. Accordingly, the climber’s experiences of the opera should not be epistemically relevant. But the climber’s experiences of the opera evidently are epistemically relevant—​they demand to be taken into account in theorizing—​despite lacking whatever supplementary beliefs the coherentist deems requisite. The upshot is that we can accommodate our starting intuition only by acknowledging that there is something about the experiences in and of themselves that lends credibility to their content. And this is to concede that such experiences provide some level of immediate justificatory support. This is not to say that experience can immediately justify belief in their content full stop. Our experiences may only be able to immediately justify some minimal positive credence in their content; perhaps they can only grant some slight “presumption” to their content (McCain and

152  Foundationalism Defended Moretti 2021, 80). But even this still admits that experiences justify us in taking some positive doxastic attitude in the state of empty symmetrical evidence, however minimal that attitude may be. Notwithstanding, the concession that there is immediate justification does not get us all the way to foundationalism as I’ve defined it in FN. We still need to show that all non-​immediate justification depends on our having this immediate justification. Nevertheless, by removing the main objection to foundationalism, this concession clears the way for foundationalism’s ultimate victory. The case for coherentism is built in large part on the failure of foundationalism—​in particular, on the supposed impossibility of immediate justification. As BonJour writes, “the main motivation for coherentism is purely dialectical: simply the avoidance of foundationalism, with no very substantial initial plausibility attaching to coherentism itself” (BonJour and Sosa 2003, 43). This is to overstate the matter, but the central point rings true. Thus, if the coherentist concedes that the myth of the given is in fact no myth at all, then he has undermined the primary motivation for his own theory and its revisionary conception of justification. Notice that we can direct this same objection at infinitism as well. Amend Plantinga’s case so that instead of having a coherent belief system, the hiker has an infinite chain of non-​repeating beliefs supporting the conviction that he is climbing the Grand Tetons. Once again, freeze these beliefs in place and take him to the opera. As before, we intuit that the hiker’s beliefs are not justified, and part of the explanation seems to be that his ongoing experiences must be taken into account. This can’t be attributed to the fact that the hiker believes himself to have those experiences, since he lacks any such belief. But if the infinitist admits that the experiences themselves provide immediate justificatory support for their content, then he has undermined the central motivation against foundationalism by conceding that a non-​arbitrary end to the regress is possible. Let us turn to the second and more widely discussed intuition underlying Plantinga’s inflexible climber: that coherence is not sufficient for justification. The climber’s beliefs are perfectly coherent, after all, but this alone seems inadequate for them to be justified. The belief system needs to have something going for it, something like experience backing it up, for the beliefs in that system to align with our epistemic duties. The reason is that our epistemic duties require a level of caution—​that we take great care in the pursuit of truth to avoid falsehood as well. Unless there is something outside of the belief system vouching for its credibility, however, it just seems reckless to hold the beliefs in this system. After all, there are an infinite number of coherent belief systems, most of which are mostly false. If we were to arbitrarily pick one of these and adopt it as our own, said adoption would not be permissible. But that is essentially what is going on if we have a coherent belief system that is completely disconnected from

Foundationalism Defended  153 experience. This is perhaps a unique way of pressing the well-​known alternative systems objection. I think this reasoning provides the strongest basis for BonJour’s famous input objection as well: Coherence is purely a matter of the internal relations between the components of the belief system; it depends in no way on any sort of relation between the system of beliefs and anything external to that system. Hence, if, as a coherence theory claims, coherence is the sole basis for empirical justification, it follows that a system of empirical beliefs might be adequately justified, indeed might constitute empirical knowledge, in spite of being utterly out of contact with the world that it purports to describe. Nothing about any requirement of coherence dictates that a coherent system of beliefs need receive any sort of input from the world or be in any way causally influenced by the world. (BonJour 1985, 108) Why is input from the world so essential? Because without the (apparent) awareness of the world in experience backing up our belief systems, those beliefs are simply too risky to be allowed. In response to doubts about the sufficiency of mere coherence, most coherentists have abandoned “pure coherentism,” which is the view that justification can arise through coherence relations alone. BonJour overgeneralizes, but only slightly, when he says, “there is probably no one who has ever seriously advocated a pure coherence theory of empirical justification, one in which the coherence of a set of beliefs is claimed to be by itself sufficient for justification” (BonJour and Sosa 2003, 43). Instead, most advocates endorse “impure coherentism,” which concedes that a coherent system of beliefs cannot be justified unless key beliefs enjoy immediate epistemic support that does not itself arise from coherence relations.15 For example, Poston, a self-​acclaimed explanatory coherentist, writes: I do not see how coherence by itself can provide justification unless the data used for coherence reasoning has some presumption in its favor. Unless the data used for coherence reasoning has some initial plausibility, mere coherence cannot raise the probability of belief. (Poston 2014, 43) But this is simply to give the game away to foundationalism, for it concedes that non-​ immediate justification (via coherence relations) depends on immediate justification just as FN maintains. This brings us to the dirty little secret of coherentism, which is that almost all of those who style themselves “coherentists” are actually what BonJour calls “weak foundationalists” (BonJour 1978, 4). That is, they concede

154  Foundationalism Defended that there are sources of immediate justification, and that a coherent belief system apart from these sources would not be justified; but they insist that these sources are not able to immediately justify belief specifically and/​or that this immediate justificatory support cannot be extended through inference until enough of these sources are integrated into a coherent framework. If such theorists wish to call themselves “coherentists” to distance themselves from stronger forms of foundationalism which do allow for such things, then I won’t complain (at least not too loudly); but we must recognize that this impure coherentism is really a form of foundationalism as I have defined it, and so is not a competitor to the position I am defending here.16 Infinitism is also susceptible to much the same charges. The sense of inadequacy does not depart if we envision the hiker having an infinite regress of non-​repeating beliefs instead of a coherent belief system. Indeed, all of the things noted about the inadequacy of coherent beliefs systems in and of themselves apply to the non-​repeating infinite regresses envisioned by infinitism. For example, I noted above that there are many coherent belief systems, most of which are false, and that without some backing from experience, the adoption of one over the others seems reckless. Similarly, there are many possible infinite chains of beliefs, as many for false conclusions as for true, and unless one of these chains has something going for it, then it is hard to see how the beliefs in that chain honor our duty to be cautious in avoiding falsehood.17 For another example, consider again BonJour’s input objection, slightly revised: [A non-​repeating infinite regress of beliefs] is purely a matter of the internal relations between the components of the belief system; it depends in no way on any sort of relation between the system of beliefs and anything external to that system. Hence, if, as [infinitism] claims, [this sort of infinite regress] is the sole basis for empirical justification, it follows that a system of empirical beliefs might be adequately justified, indeed might constitute empirical knowledge, in spite of being utterly out of contact with the world that it purports to describe. Nothing about any requirement of [non-​repeating infinite regression] dictates that [an infinitely regressing] system of beliefs need receive any sort of input from the world or be in any way causally influenced by the world. (BonJour 1985, 108) The same objections apply to both coherentism and infinitism because they make the same problematic claim: that justification requires only that our beliefs stand in the appropriate logical relations to other beliefs.18 But without something else supporting that system of beliefs—​ something like experience—​all we have are the subject’s assertions and a

Foundationalism Defended  155 mutually-​supporting collection of assertions (even if infinite in number) simply isn’t adequate. There could be a form of “impure infinitism” that grants some level of immediate justification to key premises but insists that full endorsement of those premises must always be justified via additional argumentation (see Aikin 2005; 2008, 178–​179). However, this would simply be another form of weak foundationalism and no competitor to foundationalism as characterized in FN. In summary, only pure coherentism and pure infinitism stand in opposition to foundationalism as characterized by FN. But those positions have no good way of responding to either of the objections pressed in this section. The most plausible responses give the game away to foundationalism by conceding both the possibility of and the need for immediate justification to ground our belief systems. The only hope for the pure coherentist or pure infinitist is to bite the bullet and argue that, nevertheless, the foundationalists is in a worse spot than he is. 5.2.2  BonJour’s Dilemma for Foundationalism

The most formidable challenge to foundationalism is BonJour’s dilemma (BonJour 1978; 1985, ­chapter 4). The argument is often called “Sellars’ dilemma,” but it is BonJour’s version of it that I will treat here.19 The conclusion of the argument is that immediate justification is impossible because it requires that there be a state that is simultaneously capable of justifying without itself requiring justification. This is impossible, BonJour maintains, because the very same features that enable a state to justify also render that state needful of justification. Let’s unpack this reasoning a bit further. We first need to draw a distinction between beliefs and the mental states that are supposed to immediately justify those beliefs.20 BonJour calls these mental states “intuitions” or “immediate apprehensions,” but I will stick with the term foundational evidence. What must these states of foundational evidence be like in order to immediately justify belief? For one, BonJour argues that they will need to be “cognitive states,” by which he means that they possess propositional content. If they do not possess propositional content, then it is “difficult to see how the intuition is supposed to justify the belief” (BonJour 1978, 10). BonJour’s defense of this premise is rather thin, but I will argue in the next chapter that it nevertheless rests on sure footing. While it may be possible for a non-​propositional state to serve as evidence, this is not possible apart from a larger framework of background evidence; hence, such states cannot serve as foundational evidence since they cannot immediately justify anything. It follows that whatever states are to serve as foundational evidence must be propositional. It is not enough, however, that these states merely possess propositional content, for there are all sorts of ways in which that content might be hosted.

156  Foundationalism Defended If I imagine that p, or hope that p, or get angry that p, I will have a mental state with propositional content; but none of these are plausible candidates for foundational evidence. This is perhaps why BonJour focuses specifically on cognitive states that involve “the thesis or assertion that-​p” or “assertive content”—​beliefs being his prime examples (BonJour 1978, 10). States with assertive content fare better than non-​propositional states in that “it is easy enough to understand in a rough sort of way how an intuition can serve to justify a belief with this same assertive content.” However, “the problem is to understand why the intuition, involving as it does the cognitive thesis that-​p, does not itself require justification” (BonJour 1978, 10). BonJour concludes that the idea of foundational evidence is faced with “a fundamental dilemma”: [I]‌f his intuitions or immediate apprehensions are construed as cognitive, then they will be both capable of giving justification and in need of it themselves; if they are non-​cognitive, then they do not need justification but are also apparently incapable of providing it. This, at bottom, is why epistemological givenness is a myth. (BonJour 1978, 11) To resolve the dilemma, the foundationalist needs a state that can justify without itself requiring justification. Can there be such a state? BonJour argues that there can’t be, because the very same properties that render a state capable of justifying are also the properties that render it in need of justification. He writes: It is one and the same feature of a cognitive state, viz. its assertive content, which both enables it to confer justification on other states and also requires that it be justified itself. (BonJour 1978, 12) For good measure, BonJour adds that any attempt to posit such a state is, if not incoherent, bound to be “hopelessly contrived and ad hoc” (BonJour 1978, 11). I think it fair to say that BonJour’s dilemma, or something in the neighborhood, is the problem for foundationalism. How can we trust what a state presents as true without first establishing (via other states) whether it is trustworthy? 5.3  Resolving BonJour’s Dilemma I will argue here that seemings are well-​ situated to split the horns of BonJour’s dilemma. I will not be arguing that seemings do resolve the

Foundationalism Defended  157 dilemma, for that would require an argument that seemings immediately justify their contents, and that is the argument I plan to make throughout the rest of this inquiry. What I will argue here is just that seemings show how the dilemma might be resolved. That is, they provide a plausible account of how a state can lend justificatory support to its content without itself receiving justificatory support from anything else. The upshot is that there is a plausible path forward for foundationalism, but not for (pure) coherentism or infinitism. The first thing to note is that there is a problematic conflation in BonJour’s presentation of the dilemma. In discussing states with “assertive content,” BonJour fails to distinguish between the way in which beliefs are assertive and the way in which experiences are. In the last chapter, we discussed a mode of representation known as assertive force. States with assertive force are those that present their content as representative of the way things really are, seemings being our prime examples. Call this experiential assertiveness. I also noted there that beliefs are not assertive in this sense. A belief does not assert its content to us as true. Rather, in believing, we are asserting that proposition to be true. The subject is making the assertion rather than the experience. Call this doxastic assertiveness. Doxastic assertiveness is what makes something capable of being justified or unjustified. Only when we take a stand on what is true does our cognitive activity become assessable with respect to the norms of justification. Accordingly, states that are assertive only in the sense of having assertive force, like seemings, are not subject to normative evaluation. This is why seemings cannot be permissible or impermissible, only veridical or illusory. Once we recognize the distinction between doxastic and experiential assertiveness, it becomes plain that BonJour’s argument trades on an ambiguity in the meaning of “assertive content.” He claims first that states which do not involve “the assertion that-​p” cannot justify. Fair enough. If the state has neither doxastic nor experiential assertiveness, then it could not justify belief. He then claims that any state with assertive content must itself be justified. But this is only true of states that are assertive in the way beliefs are. States with experiential assertiveness are not subject to the norms of justification. Thus, this second horn of the dilemma seems plausible only if we are conflating these two different kinds of assertiveness. This spells bad news for BonJour. For seemings have assertive content (by virtue of their experiential assertiveness) but do not require justification (by lacking doxastic assertiveness); and by BonJour’s own admission, “it is easy enough to understand in a rough sort of way how an intuition can serve to justify a belief with this same assertive content” (BonJour 1978, 10). Furthermore, the appeal to seemings cannot be dismissed as a “contrived” or “ad hoc” posit since numerous epistemologists not only today but throughout the course of history have recognized the existence of

158  Foundationalism Defended seeming states (see Chapter 4, §4.2) and considered them good candidates for the role of foundational evidence (see Chapter 8, §8.1). Thus, seemings stand poised to pass through the horns of BonJour’s dilemma. We’re not yet in the clear, however. For we can reframe much the same problem in terms of verification rather than justification. That is, if we have a seeming that presents its content to us as true, we can still ask whether we should trust this seeming’s “assertion” (BonJour 2004, 358; 1985, 78). Moreover, it is not initially implausible to say that we need some verification of its reliability before we are permitted to trust it (e.g., Hasan 2013, §2). In which case seemings wouldn’t qualify as foundational evidence. Though BonJour’s dilemma is not fully defeated, the foundationalist has nevertheless gained ground in this exchange. For the second horn of the dilemma might have initially seemed quite obvious. If a state is assertive in the same way beliefs are, then of course it cannot justify without receiving justificatory support from something prior. This was a crucial principle in the regress argument for foundationalism, and so, in addition to being evident on its own, cannot not be denied without abandoning one of the central motivations for foundationalism. However, it is not at all obvious that an experience with assertive force is unable to justify without receiving verification from other justified states. This position has some initial plausibility, but it is far from evident. Indeed, on initial inspection, its contrary enjoys equal if not greater plausibility. When a proposition seems true to us, it feels as though we are directly aware of that proposition’s truth. With presentational seemings, we also feel directly aware of a state of affairs which makes that proposition true. Even without verification, it is quite plausible that these states bring belief (or some lesser doxastic attitude) into alignment with our epistemic duties vis-​à-​vis the pursuit of truth and the avoidance of falsehood.21 In other words, seemings are plausible candidates for foundational evidence. This moderate conclusion is all I need to complete the case for foundationalism, for it shows that there is a plausible means of escaping the central problem for foundationalism. In contrast, we saw that the most resilient versions of coherentism and infinitism do not have any way of responding to their central problems (without giving the game away to foundationalism) except to bite the bullet and hope that foundationalism is even worse off. Recall also that these versions of coherentism and infinitism survive the regress argument only by abandoning our natural conception of justification as linear. This is a revision we might be willing to make if forced, but these revisionary conceptions of justification have turned out to be more problematic than the one we are being asked to abandon them for. Thus, if we must choose between these three views (and we must), foundationalism is the way to go. This verdict will be vindicated in later

Foundationalism Defended  159 chapters when I make the case that seemings not only might, but in fact do, serve as foundational evidence for their content. Having defended FN—​ that all justification traces back to immediate justification—​we can now turn our attention to where this immediate justification is supposed to come from. The next two chapters argue that it cannot come from anywhere but seemings. The chapters after that will argue that it can and in fact does come from seemings. Notes 1 Classical foundationalism is an especially strong version of what BonJour calls “strong foundationalism,” which requires the immediate justification of our beliefs be knowledge level (BonJour 1978, 4). 2 This is close to what BonJour calls “weak foundationalism” (BonJour 1978, 4), although it is even a bit weaker than that. For BonJour still construes weak foundationalism as requiring that we have immediate justification for believing (although the degree of justification required may fall short of being sufficient for knowledge) as opposed to justification for a mere positive level of confidence. 3 I refer to my own outworking of seemings foundationalism here rather than seemings foundationalism proper. In and of itself, seemings foundationalism says only that the strength of immediate justification will differ in accordance with the strength of the seeming (with how forceful it is), but this does not entail anything about what that range might be. I’d like to leave seemings foundationalism neutral in this respect so as to accommodate a wider variety of adherents. My preferred outworking of seemings foundationalism, however, which I will defend briefly in Chapters 8 and 9, is that the strength of immediate justification provided by seemings runs the gamut. In some cases, it may be sufficient only to justify a minimally positive level of confidence. In others, it may be sufficient to justify certainty (or something close to it). It depends on the seeming. 4 BonJour builds these positions on inferential justification into his characterization of strong and weak foundationalism, respectively, although I think this is unfortunate. For it is possible to mix and match positions on the strength of immediate justification with what is required for such justification to be extended through inference. For example, it may be that all inferential justification necessarily requires a larger framework of beliefs, no matter how strong our immediate justification for believing p may be. 5 While the argument applies to all doxastic attitudes, it will be convenient to focus specifically on beliefs. 6 See my defense of this assumption in Chapter 8, §8.2. 7 Beliefs do not need to be occurrent in order to serve as the base of other beliefs, but we do need to actually have them. 8 See Bergmann (2007) for further discussion. 9 The same basic argument is made by many. See, for example, Ginet (2005, 149) and Dancy (1985, 55).

160  Foundationalism Defended 0 The similarity to Aquinas’s second way is intentional. 1 11 This is why we should not follow Klein in calling the regress a regress of reasons, for one belief serves as a reason for another only when the former belief is a source of justification for the other. But that is precisely what is being denied here. 12 See, for example, the discussion in Pollock (1986). 13 See Chapter 8, §8.4, for more on this as it pertains to seemings specifically. 14 Kvanvig and Riggs suggest that having an appearance might be an “Insufficient but Non-​redundant condition of e which is part of a larger condition that is itself Unnecessary but Sufficient for e” (Kvanvig and Riggs 1992, 211)—​an idea they take from J.L. Mackie’s work on causation (Mackie 1974). 15 Examples of coherentists (or those sympathetic to coherentism) who plausibly fit into this camp include Harman (1973), Lycan (1988), Poston (2014), Quine (1960), Sellars (1963), and Thagard (2000). To be clear, these thinkers do not necessarily admit experience specifically as a source of immediate justification. Indeed, most in this camp endorse some sort of epistemic conservatism, which grants immediate support to the contents of our beliefs simply because they are believed. See Chapter 7 for my critique of epistemic conservatism. 16 Later, in Chapter 8, I will argue that seemings can sometimes immediately justify full belief in their content, but this is not a part of seemings foundationalism proper. Neither seemings foundationalism nor I take any stance on what is required to transmit justificatory support through inference. 17 What Klein and Turri (2023, §4.c) call the “AC/​DC objection” seems related to (but still different from) my objection here. 18 Klein writes: I might note in passing that Davidson’s characterization of coherence theories—​namely that “what distinguishes a coherence theory is simply the claim that nothing can count as a reason for holding a belief except another belief” might distinguish it from foundationalist theories, but it does not distinguish it from infinitism. (Klein 1999, 317) 19 Sellars’ presentation of the argument (in Sellars 1956) is infamously difficult to suss out. As BonJour notes, “I suspect that something like the argument to be given here is lurking somewhere in Sellars’ ‘Empiricism and the Philosophy of Mind,’ … but it is difficult to be sure” (BonJour 1978, 9). 20 Epistemic conservatives deny that there is any need for a mental state (distinct from the belief itself) to justify belief, but I will set that aside here. I argue against epistemic conservatism in Chapter 7. 21 Note that classical foundationalists like Hasan (who I previously cited as requiring verification of seemings) should be at least somewhat favorable toward this more credulous position once the nature of seemings is properly understood. For in presentational seemings, it feels as though we are directly acquainted with a thought, the state of affairs that thought describes, and the correspondence between them. In other words, it feels as though we are triple

Foundationalism Defended  161 acquainted with p. But classical foundationalists like Hasan maintain that the state of triple acquaintance is a state that can justify without itself requiring verification (Hasan 2013). Plausibly, if triple acquaintance states can justify, then so can states that are phenomenally indistinguishable from those triple acquaintance states like presentational seemings. (See Chapter 6 for an argument that it is the feeling of triple acquaintance, rather than triple acquaintance itself, that is doing the justifying if anything is.)

References Aikin, Scott F. 2005. “Who Is Afraid of Epistemology’s Regress Problem?” Philosophical Studies 126 (2): 191–​217. —​ —​ —​ . 2008. “Meta-​ Epistemology and the Varieties of Epistemic Infinitism.” Synthese 163 (2): 175–​185. Bergmann, Michael. 2007. “Is Klein an Infinitist about Doxastic Justification?” Philosophical Studies: An International Journal for Philosophy in the Analytic Tradition 134 (1): 19–​24. BonJour, Laurence. 1978. “Can Empirical Knowledge Have a Foundation?” American Philosophical Quarterly 15 (1): 1–​14. —​—​—​. 1985. The Structure of Empirical Knowledge. Cambridge, MA: Harvard University Press. —​—​—​. 2004. “In Search of Direct Realism.” Philosophy and Phenomenological Research 69 (2): 349–​367. BonJour, Laurence, and Ernest Sosa. 2003. Epistemic Justification: Internalism Vs. Externalism, Foundations Vs. Virtues. Hoboken, NJ: Wiley-​Blackwell. Dancy, Jonathan. 1985. An Introduction to Contemporary Epistemology. Hoboken, NJ: Blackwell. Ginet, Carl. 2005. “Infinitism Is Not the Solution to the Regress Problem.” In Contemporary Debates in Epistemology, edited by Steup Matthias and Sosa Ernest, 140–​149. Hoboken, NJ: Blackwell. Harman, Gilbert. 1973. Thought. Princeton, NJ: Princeton University Press. Hasan, Ali. 2013. “Phenomenal Conservatism, Classical Foundationalism, and Internalist Justification.” Philosophical Studies 162 (2): 119–​141. Klein, Peter D. 1999. “Human Knowledge and the Infinite Regress of Reasons.” Philosophical Perspectives 13: 297–​325. Klein, Peter, and John Turri. 2023. “Infinitism in Epistemology.” In Internet Encyclopedia of Philosophy. https://​iep.utm.edu/​inf-​epis/​. Kvanvig, Jonathan L., and Wayne D. Riggs. 1992. “Can a Coherence Theory Appeal to Appearance States?” Philosophical Studies 67 (3): 197–​217. Lycan, William G. 1988. Judgement and Justification. Cambridge: Cambridge University Press. Mackie, J.L. 1974. The Cement of the Universe: A Study of Causation. Oxford: Clarendon Press. McCain, Kevin, and Luca Moretti. 2021. Appearance and Explanation: Phenomenal Explanationism in Epistemology. Oxford: Oxford University Press.

162  Foundationalism Defended Plantinga, Alvin. 1993. Warrant: The Current Debate. Oxford: Oxford University Press. Pollock, John. 1986. Contemporary Theories of Knowledge. First Edition. London: Hutchinson. Poston, Ted. 2014. Reason and Explanation. New York: Palgrave Macmillan. Quine, W.V.O. 1960. Word and Object. Cambridge, MA: MIT Press. Sellars, Wilfrid. 1956. “Empiricism and the Philosophy of Mind.” Minnesota Studies in the Philosophy of Science 1: 253–​329. —​—​—​. 1963. Science, Perception and Reality. New York: Humanities Press. Thagard, Paul. 2000. Coherence in Thought and Action. Cambridge, MA: MIT Press.

6 The Case for Seemings Exclusivism

In the last chapter, I made the case that all justification stems from immediate justification (FN)—​immediately justified attitudes being those whose justification does not depend on having evidence or justification for any other doxastic attitudes. Now we ask where immediate justification comes from. Seemings foundationalism says that immediate justification requires foundational evidence (EV), and that our foundational evidence consists in all seemings (PC) and only seemings (SE). The defense of these last three theses will spread over the next four chapters. I begin this chapter with a defense of seemings exclusivism: SE—​If S has foundational evidence for p, then (it is because) it seems to S that p is true; if S has foundational evidence against p, then (it is because) it seems to S that p is false. Essentially, no mental states other than seemings qualify as foundational evidence. In the following chapter, I dismiss the possibility of immediate justification absent foundational evidence. The final two chapters argue that some and in fact all seemings qualify as foundational evidence, at which point our defense of seemings foundationalism will be complete. The plan of this chapter is to survey our various mental states and consider which of them, if any, are capable of fulfilling the unique role of foundational evidence: that role being to make permissible belief in a manner that does not depend on background evidence or justified background beliefs. Throughout our survey, a consistent pattern will emerge. Either the candidates end up being seemings, or else they lack something essential—​a lack that is fixed only by making those states more seeming-​ like. In drawing out this pattern, we will uncover a principled reason why no states other than seemings can serve as foundational evidence. The goal here is not to prove that seemings can serve as foundational evidence (that’s the purview of Chapters 8 and 9), but that all other mental states can’t.

DOI: 10.4324/9781003281108-7

164  The Case for Seemings Exclusivism You may wonder why I have chosen to defend SE (and, in the next chapter, EV) before PC. This is because the plausibility of the lattermost is greatly boosted by the truth of the former. For if seemings are the only potential sources of immediate justification, then given FN, skepticism can be avoided only by allowing seemings to be actual sources of immediate justification. In this way, the arguments of these next two chapters (Chapters 6 and 7) will provide significant, though not indispensable, support to the arguments in the two that follow (Chapters 8 and 9). 6.1  The Foundationalist Landscape A great many foundationalists today are externalists about justification. On these views, whether a mental state counts as foundational evidence will depend on features of that state that are external to the subject. For instance, a state may be foundational evidence if basing belief on that state in the absence of background evidence is, for example, a reliable process, or a safe process, or results from the proper functioning of one’s reliably truth-​aimed faculties, or manifests intellectual virtue. I argued in Chapter 1 that externalism about justification is false. Accordingly, we will limit our focus to internalist forms of foundationalism. Among internalists, the candidates for foundational evidence consist in various kinds of non-​doxastic mental states.1 Internalists generally agree that doxastic attitudes like belief cannot serve as foundational evidence for the reasoning underlying BonJour’s dilemma: the same features of belief that allow it to justify also render it in need of justification (BonJour 1978, 12). This is not to say that beliefs cannot be evidence for other beliefs; they can be. They simply cannot be foundational evidence since they are only capable of justifying when they themselves have already been justified by something prior. Among non-​doxastic states, the first major divide is between those candidates that have conceptual content (conceptual states) and those that do not (non-​conceptual states). Allow me to characterize each of these in more detail, starting with conceptual states. An important class of conceptual states are those with conceptual content that is capable of being true or false, also known as propositional content. Propositions, I will assume, are sentence-​like mental representations. They are constituted by concepts in much the way that words constitute a sentence.2 Those states with propositional content can be called thoughts. When one thinks, one is related to a proposition in such a way that one’s mind is directed toward whatever states of affairs that proposition represents—​ that is to say those states of affairs become intentional objects for the subject. There are many kinds of thoughts. There’s entertaining p, imagining p, believing p, hoping that p, fearing that p, and more. Some even think that sensory states are species of thought with distinctive phenomenology—​a

The Case for Seemings Exclusivism  165 position known as “conceptualism” (e.g., Brewer 1999, McDowell 1994, Peacocke 1983). Foundationalists who appeal to conceptual states must go on to specify which species of thought are supposed to serve as foundational evidence. Here we can further divide candidates into those that possess assertive force and those that do not. We will consider both kinds of conceptual states in §6.3 and §6.4. Non-​conceptual states are simply those that lack conceptual content. Among non-​ conceptual states, some candidates for foundational evidence are representations while others are relations. Non-​conceptual representations, like thoughts, represent states of affairs but do so differently from thoughts. To get a rough idea, we might compare this to the way in which maps or pictures represent a landscape differently than does a written description. Many think that sensory appearances are non-​ conceptual representations—​a position known as “representationalism” about perception (e.g., Evans 1982, Peacocke 1992, Tye 1995). Some foundationalists hold that non-​conceptual representations can be foundational evidence. Others in the non-​conceptualist camp will appeal to some kind of direct relation between the subject and a state of affairs to fill the role of foundational evidence. Direct acquaintance is the classic example (Russell 1912), though it has gone by a variety of names. In direct acquaintance, a state of affairs is directly presented to the conscious mind of the subject. As Fumerton says, “the fact is there before consciousness” (Fumerton 1995, 76). Laurence BonJour explains: The underlying idea here is that of confrontation. In intuition, mind or consciousness is directly confronted with its object, without the intervention of any sort of intermediary. It is in this sense that the object is given to the mind. (BonJour 1978, 11) Though originally suspicious of direct acquaintance, BonJour later becomes a proponent, describing it as the kind of “built-​in” awareness that one has of one’s own occurrent mental states simply by virtue of their being conscious, as opposed to the kind of apperceptive awareness one can gain of those states upon reflection (BonJour and Sosa 2003, sect. 4.1–​4.2). Acquaintance is not conceptual, since it is possible to be acquainted with a fact “without even possessing the conceptual resources to represent that fact in thought” (Fumerton 1995, 74). Nor does it involve any kind of non-​conceptual representation. Thus, Ali Hasan can say, “To be directly aware of or acquainted with a fact is to be aware of that fact without the mediation of any thought, representation, or other state of awareness” (Hasan 2013b, 122). The relation of direct acquaintance is usually limited

166  The Case for Seemings Exclusivism to states of affairs such as being in pain or two things being more than one. However, relationalists about perception extend a similar kind of direct relation to states of affairs in one’s immediate physical environment (e.g., Brewer 2011, Campbell 2002, Fish 2009). On such views, sensations do not have any representational function but are simply what it feels like for states of affairs in the physical world to be directly presented to one’s conscious awareness. This kind of direct acquaintance, or direct awareness, is another candidate for foundational evidence. So, we have two different kinds of non-​ conceptual states to consider: non-​conceptual representations and non-​conceptual relations. I treat both of these together, in §6.2, since the lack of conceptual content in and of itself will ensure their undoing. Treating them together introduces some terminological challenges, however. For convenience, I will continue to talk of non-​conceptual states “(re)presenting” states of affairs, where this refers either to their non-​conceptual representation or direct presentation. For the sake of thoroughness, I should mention one final non-​ conceptualist view of perceptual states, if only to dismiss it—​that is the Reidian view of sensations as raw feels (Reid [1764] 1997; [1785] 2002). On this view, sensory states do not intrinsically represent anything at all, nor are they the product of an object’s direct presentation. I bracket off this view because no internalists propose that such states serve as foundational evidence. As McCain and Moretti say, “After all, if sensations do not represent the world as being a particular way, it is unclear how they can give us (on their own) reason to think that the world is a certain way” (McCain and Moretti 2021, 67). Perhaps they could be evidence when situated within a larger framework of evidence which we could use to discern their significance, but then they would not be foundational. 6.2  Non-​Conceptual States This section argues that non-​conceptual states (whether representations or relations) cannot be foundational evidence. There are arguments for this conclusion already in existence, though I do not find any of them fully convincing. Rehearsing these arguments will, however, lay the groundwork for introducing my own. To foreshadow our discussion, the basic problem is that lacking conceptual content makes it difficult for non-​conceptual states to satisfy some of the basic conditions on evidence. My argument, in contrast to existing arguments, is not that it is impossible for non-​conceptual states to meet these conditions, but that they can only do so when certain background evidence is in place. Hence, they cannot be foundational evidence. To understand the difficulties facing non-​conceptual states, we must first delve into the nature of belief and its content. It is to this that we now turn.

The Case for Seemings Exclusivism  167 6.2.1  The Nature of Belief

At root, beliefs involve taking an affirmative stance toward what is the case. There must therefore be at least two components to beliefs: i That by virtue of which the mind is directed toward a state of affairs as an intentional object, and ii That by virtue of which one affirms that this state of affairs obtains. The latter dimension is sometimes called the belief’s “assertiveness” (though I must reiterate the point that this doxastic assertiveness is different from the assertive force characteristic of seemings). I won’t say more about it here other than that it is by virtue of this assertiveness that beliefs have a mind-​to-​world direction of fit (Searle 1983), and so are in some manner defective if the state of affairs toward which the mind is directed does not actually obtain. It is also by virtue of their doxastic assertiveness that beliefs require justification (BonJour 1978, 12). But how do beliefs direct the mind toward some state of affairs in the first place? The predominant answer, which I will assume here, is that beliefs direct our minds toward states of affairs via conceptual content that is capable of being true or false. There are three things to note here. One, beliefs have representational content. Two, that content is conceptual. Three, that conceptual content is true or false. All this amounts to the claim that the content of a belief is a proposition. Let’s run through a few reasons for thinking each of these is true. First, if belief is not representational, then it must place one’s mind in a direct, unmediated relation to the state of affairs being affirmed. But if this were so, then any belief in that same state of affairs would be the same as any other, possessing the same cognitive significance, and this is clearly false. Consider Lois Lane, the love interest of Superman. Lois is also the coworker of Clark Kent, who (unbeknownst to her) is Superman’s alter ego. When Lois sees Superman fly by the window, she believes that Superman is flying. She does not, however, believe that Clark Kent is flying despite the fact that both of these beliefs direct her mind toward the same concrete state of affairs. Thus, the belief that Superman is flying and the belief that Clark Kent is flying are not the same (Lois possesses one without possessing the other), nor do these beliefs possess the same cognitive significance for her. Situations like Lois’s exemplify what is known as Frege’s puzzle (Frege [1892] 1952). This puzzle compels us to introduce a bearer of cognitive significance to serve as a kind of mediator between subject and object, allowing for the same state of affairs to be cognitively approached in different ways.3 This is a role filled by representations. In believing, the subject’s mind is directed toward a state of affairs by entering into a relation with some

168  The Case for Seemings Exclusivism mental representation, the bearer of cognitive significance, that represents that state of affairs. Second, this mental representation, the content of the belief, is conceptual. Beliefs only arise once conceptual capacities are developed, whereas non-​ conceptual representation is something that infants and animals can readily do (as in sensory appearances). This is not to say that infants and animals cannot form beliefs, but only that, if they do so, it is by virtue of whatever limited conceptual abilities they possess. Moreover, when beliefs do arise, their specific contents are proscribed by one’s conceptual repertoire. The expert and the amateur will form very different beliefs when looking at the same item. Third, the conceptual content of beliefs is also capable of being true or false. This means that it qualifies as propositional content since propositions are the bearers of truth and falsity. For another, belief contents can stand in certain rational and logical relations to one another that can seemingly only exist between propositions. For these and other reasons, beliefs are standardly classified as propositional attitudes. These observations about belief introduce several challenges to the idea of non-​conceptual states as foundational evidence. For a mental state is evidence only if it indicates the truth of some proposition in a way that permits one to believe it. But belief content is conceptual. This means that non-​conceptual states cannot share the same content as beliefs. Worse, they do not share even the same kind of content as beliefs. How is it, then, that they can make belief permissible? The challenges here are twofold. One condition on evidential states is the following: The Relation Condition—​The content of the evidential state must bear the right kind of logical relations to the proposition it supports. For now, we can leave unspecified what the right kind of logical relations are. The point is that there must be some pertinent relations between those contents, and the general worry is that, whatever they are, they cannot obtain between non-​conceptual and conceptual contents. Even if those relations can obtain, the second and greater worry is that the subject will not be able to recognize that they obtain. After all, if the subject cannot appreciate the relevance of some non-​conceptual content to a proposition, then how is it supposed to make it permissible for the subject to believe that proposition? In other words, a second condition on evidential states is this: The Recognition Condition—​The subject must be in a position to recognize that the content of the evidential state bears the right kinds of logical relations to the proposition it supports.

The Case for Seemings Exclusivism  169 Satisfying the recognition condition is a special challenge for non-​ conceptual states. As Reiland puts it, “non-​conceptual states do not talk to us in our language: they do not articulate their contents in terms of our conceptual categories” (Reiland 2015, 524). So even if a non-​conceptual state is “saying” the very same thing as a proposition, it does not follow that the subject will be in a position to recognize this fact. The question for us becomes, is it possible to translate between non-​conceptual and conceptual contents so as to appreciate the relations between one and the other? And if so, can this translation occur apart from background evidence? We will consider all of these worries more closely in the following section. 6.2.2  Bridging the Conceptual and the Non-​Conceptual

Some philosophers have argued that non-​conceptual states cannot contribute to the justification of belief in any capacity. Non-​ conceptual states can cause beliefs, to be sure, but they cannot be evidence for them. A classic argument to this effect comes from Davidson who maintains that the content of non-​conceptual states cannot stand in the requisite kinds of logical relations to the contents of our beliefs (Davidson 1986). That is, non-​conceptual states cannot satisfy the relation condition. There is some plausibility to this objection. Notice that relations of entailment, consistency, probability, and coherence only exist between propositions, but surely for a mental state to be evidence for a proposition, its content must entail or make probable or cohere with or at least be consistent with that proposition. If Davidson’s argument is sound, that’s good news for SE. For then non-​ conceptual states are not any kind of evidence, much less foundational evidence. If, on the other hand, we assume that Davidson is wrong, there are other problems for non-​conceptual states waiting in the wings. To see these additional problems, let us consider how one might address Davidson’s objection. BonJour gives an example of the most plausible response. He writes: The relation between this non-​conceptual content and such a conceptual description thereof may not be strictly logical, as Davidson uses the term, i.e., may involve nothing like the relations of inference, consistency or inconsistency, etc., that may exist between two propositions, but it is also obviously not merely a causal relation. Rather it is a descriptive relation, having to do with the accuracy or inaccuracy of fit between a conceptual description and a non-​conceptual object that the description purports to describe. (BonJour and Sosa 2003, 72)

170  The Case for Seemings Exclusivism In other words, though a non-​conceptual experience may not be able to entail or probabilify belief, there can still be a kind of match between that which is (re)presented in experience and that which a belief describes via its propositional content. A common way of parsing this out is in terms of accuracy conditions. That is to say that non-​conceptual states still have accuracy conditions (Siegel 2010), and those accuracy conditions can be the same as (or otherwise suitably related to) those belonging to a proposition. However exactly we understand this “matching,” BonJour proposes that when a belief’s content matches the content of one’s experience, the direct awareness of this latter content “can seemingly constitute a kind of reason for thinking that the description is true or correct… thus, apparently, providing a basis for the justification of the conceptual claim” (BonJour and Sosa 2003, 72; see also Heck 2000 and Peacocke 2001). In short, the relation condition can be satisfied despite the lack of propositional content. This cannot be the whole story, however, for non-​conceptual states must also satisfy the recognition condition. McDowell directs his criticism of non-​conceptual states at this condition. In order to recognize the relation between contents, one must be able to subject that relation to reflective scrutiny, and the only sorts of relations that submit themselves to scrutinization in the required sense are those existing between conceptual contents (McDowell 1994, 52–​53). Again, if this is correct, then we are one step closer to SE. Nevertheless, there are other problems for non-​conceptual states even if McDowell’s specific worries can be addressed. The most plausible reply to McDowell is exemplified by Christopher Peacocke. Peacocke thinks that the relation between non-​conceptual states and beliefs can be scrutinized. According to Peacocke, it is sufficient for the subject to be able to meaningfully ask, “Is something’s looking (or feeling) that way a reason for believing that it’s F?” And this the subject can do simply by demonstrating toward the thought That’s F and then toward the content of one’s non-​conceptual state (Peacocke 2001, 254–​256). If we grant Peacocke’s reply, an additional problem remains—​namely, that such scrutinizations seemingly require a framework of background evidence with which to make them. Specifically, being in a position to recognize the match between a belief content and a non-​conceptual content requires that the subject have evidence about what it feels like to non-​ conceptually (re)present various states of affairs.4 Notice that the issue is not whether the match between non-​conceptual content and belief content is discernable a priori (though I am inclined to think with Reid that such connections are not so discernable in perception), for it remains the case that this match cannot be recognized without background evidence already in place.

The Case for Seemings Exclusivism  171 Bill Brewer argues for a similar point. Reasons, Brewer says, “if they are genuinely to be the subject’s reasons,” must be “recognized by him as such” (Brewer 1999, 169). He then argues: To characterize a state as non-​conceptual just is to deny that it is given to the subject in such a way as to provide for the possibility of this recognition directly. Instead, he must argue about the state as follows: this state is F; anything which is F is a reason for believing that p (or for ϕ-​ing); therefore I have reason to believe that p (or to ϕ). Only then can he recognize his position as one of having a reason for believing that p (or for ϕ-​ing); only then, therefore, does he have a genuine reason for doing so. (Brewer 1999, 168) His conclusion is that: a non-​conceptual mental state can only possibly figure in an account of a person’s own reason for believing or doing something if it is conjoined with his second-​order knowledge of the appropriate rational relations between the two—​between mental states of that type, which are F, and beliefs or actions of the relevant kind. (Brewer 1999, 168–​169) The need for background evidence to satisfy the recognition condition is apparently conceded by many who identify non-​conceptual states as evidence. Peacocke admits that it is “a philosophical task to show that the holding of the correctness condition for the non-​conceptual content in question ensures the holding of the correctness condition for the conceptual content” (Peacocke 2001, 254). BonJour acknowledges that appreciating the match vis-​à-​vis direct acquaintance requires that one “understand what an experience would have to be like in order to satisfy the conceptual description” (BonJour and Sosa 2003, 73). Gareth Evans, as interpreted by McDowell, says that a non-​conceptual perception “is intelligible to its subject as purporting to be awareness of a feature of objective reality … only against the background of an understanding of how perception and reality are related” (McDowell 1994, 54). And what could such understanding consist in except evidence about which propositions match which non-​ conceptual contents? Brewer uses the need for background evidence to argue that non-​ conceptual states cannot serve as evidence at all. His reasoning is that if second-​order beliefs are needed to link one’s experience and belief, then “[the subject’s] reason is really provided by his second-​order belief that he is in a mental state which is F. This is clearly a conceptual state. Hence it

172  The Case for Seemings Exclusivism follows that all reasons are conceptual” (Brewer 1999, 169). Once again, if Brewer is right, then that’s a boon for SE; however, I don’t need Brewer’s argument to reach the intended conclusion. The fact that non-​conceptual states cannot serve as evidence apart from other background evidence proves on its own that non-​conceptual states are not foundational. It is important to recognize that Davidson, McDowell, and Brewer are all arguing for the stronger conclusion that non-​ conceptual states cannot serve as evidence at all. My conclusion is the more moderate one that non-​ conceptual states cannot serve as foundational evidence, even if they can play some role in justifying belief more generally. To summarize my argument (where S is a subject, M is a mental state, and p is a proposition): 1. If M is evidence for p for S, then S must be in a position to recognize the match between the content of M and p. (The recognition condition) 2. If M does not have conceptual content, then S is in a position to recognize the match between the content of M and p only if S has certain background evidence. 3. If M is foundational evidence for p for S, then it is not the case that M is evidence for p for S only if S has certain background evidence. 4. Thus, if M does not have conceptual content, then M cannot be foundational evidence for p for S.5 Is there any way to avoid my conclusion? Only two objections present themselves—​one to premise 1 and the other to premise 2. (Premise 3 is simply the definition foundational evidence.) I’ll treat each of these in turn. 6.2.3  The Recognition Condition

One way to avoid my argument against non-​conceptual states would be to reject premise 1 and the recognition condition. Peacocke apparently opts for this response. He admits up front that, “It is then a philosophical task to show that the holding of the correctness condition for the non-​ conceptual content in question ensures the holding of the correctness condition for the conceptual content”; yet he insists that “the ordinary thinker does not need to know what the explanation of the validity of a particular transition is” (Peacocke 2001, 254). Rather, the subject is entitled to move from her non-​conceptual state to a matching belief directly, without any philosophical account being in place (ibid, 257–​ 258). Indeed, it is by “drawing on one’s ordinary, ground-​level abilities to react rationally to one’s perceptual states” that a philosophical case might be built (ibid, 258). To be clear, Peacocke is not making the more moderate point that ordinary subjects needn’t be actively entertaining beliefs about the match between

The Case for Seemings Exclusivism  173 non-​conceptual and conceptual contents in order to justifiably transition from one to the other. He’s taking the more radical position that subjects do not even need to be in a position to form such beliefs. We can see this because, according to Peacocke, the subject is only in a position to form such beliefs after (and because) one has already repeatedly made this rational transition. Peacocke’s denial of the recognition condition leads to the untenable conclusion that a non-​conceptual experience is evidence for any matching conceptual content, even if the subject can’t recognize the connection between them.6 To illuminate the problem, let us return to Frege’s puzzle and our example of Lois Lane. Lois sees the man who is both Superman and Clark Kent fly by her office window, clad in his traditional red, yellow, and blue garb. A concrete state of affairs (consisting in a person and the property of flying) is non-​conceptually (re)presented to Lois in experience. Does this experience constitute foundational evidence for Lois? Well, evidence for what exactly? This same state of affairs is accurately described by many different propositions, including Superman is flying but also Clark Kent is flying or A Kryptonian is flying or even The future father of your child is flying. Which proposition is Lois supposed to have evidence for? One thing we can say for sure is that Lois does not have evidence for any of the latter propositions, even though these propositions accurately describe what she’s experiencing. Lois has no idea that Clark Kent is Superman, or that he’s Krypotnian, or that he will father her child. The blur of blue, yellow, and red streaking by her window simply doesn’t indicate to her, for instance, that her affable but awkward coworker is flying around town, and so does not permit belief in this proposition. Even if, for some odd reason, Lois found herself automatically believing that Clark Kent was flying while having that experience, that belief would be, from Lois’s point of view, random and without rational basis. The same is true of the propositions a Kryptonian is flying or the future father of your child is flying, since Lois (at this point in the story) knows virtually nothing about Superman’s past or future. The issue in all of these examples is that Lois can’t tell that the content of her experience matches those propositions. She must be in a position to recognize the connection between her experience and those conceptual contents before the former can be evidence for the latter. Lois’s experience may be evidence that Superman is flying, but if so, this is only because Lois is in a position to recognize that the proposition Superman is flying matches the content of her non-​conceptual experience. She justifiably believes that Superman has dark hair, wears a distinctive blue, yellow, and red outfit, and is often flying about town. She is also aware of how it feels when these things are (re)presented in perceptual

174  The Case for Seemings Exclusivism experience. So, when she experiences that familiar rush of color and shape and sound, she has evidence that Superman is flying. So far, my examples have been perceptual, but the same points apply to all non-​conceptual states. Take, for example, the non-​conceptual state of direct acquaintance. If Lois becomes acquainted with a red sensation (perhaps the red of Superman’s cape), there will once again be many propositions matching her experiential content—​one’s such as I’m being appeared to redly or I’m being appeared to scarletly or even I’m being appeared to in the way that properly functioning humans are appeared to in normal conditions when light having a wavelength of 680 nm enters the eye. Which propositions Lois has evidence for will depend on her ability to recognize that such propositions accurately describe the state of affairs with which she is acquainted. If she lacks any knowledge of wavelengths, for instance, then her acquaintance is not evidence for the lattermost proposition. Chisholm’s classic example of the speckled hen bolsters the point (Chisholm 1942). Say, one perceives a 39-​speckled hen and becomes acquainted with the resulting phenomenal array. It is obvious (for one not gifted with extraordinary subitizing abilities) that such acquaintance is not evidence for the proposition that one is being appeared to 39-​speckledly. If it were evidence, then it would permit belief in that proposition (or at least a positive credence in it), but it doesn’t. And this is precisely because the subject can’t recognize the match between that proposition and that with which she is directly acquainted. For reasons like these, those appealing to direct acquaintance (mainly classical foundationalists) usually concede the recognition condition. BonJour, for instance, responding to the problem of the speckled hen, adds the following condition on justification by acquaintance: “the person must apprehend or recognize the agreement or fit between the aspect of experience being attended to and the conceptual description given by the belief” (BonJour and Sosa 2003, 193).7 There is, however, at least one proponent of acquaintance that denies this. Declan Smithies maintains that one’s acquaintance with the 39-​speckled phenomenal array does in fact constitute evidence for the proposition I am being appeared to 39-​speckledly (call this “p”)—​that is, the acquaintance state propositionally justifies p for the subject. The problem, Smithies says, is that the subject cannot properly base belief in p on that evidence—​ that is, the subject cannot form a doxastically justified belief in p (Smithies 2019, ­chapter 10, §2). And, presumably, part of the reason why is that the recognition condition has not been met.8 Essentially, Smithies would identify recognition as a condition on doxastic justification rather than propositional justification. While ingenious, such a view remains counterintuitive. It says that when Lois sees Superman fly by her window, she thereby has evidence for the

The Case for Seemings Exclusivism  175 proposition Clark Kent is flying, and the future father of my child is flying, and in fact for every proposition that accurately describes the state of affairs with which she is directly acquainted. Smithies mitigates this by telling us that Lois can’t justifiably believe those propositions, despite having adequate evidence for them, but the oddity is not altogether dispelled. Fortunately, I think we can bypass this dispute without resolving it. The non-​conceptual acquaintance states that Smithies identifies as foundational evidence are, on their own, incapable of immediately justifying belief (or any other doxastic attitude). They can provide immediate propositional justification for p, but they cannot sustain a doxastically justified belief in p apart from other evidence. The focus of this inquiry, however, is on the sort of foundational evidence that is capable of immediately justifying doxastic attitudes like belief (compare to Conee and Feldman 2008, 85). We must have foundational evidence of this sort if we are to be justified in believing anything at all. Why? Because the kind of non-​conceptual evidence that Smithies wishes to admit cannot be put to work in justifying beliefs or other doxastic attitudes unless the recognition condition can be met. But satisfying the recognition condition will require that we justifiably affirm the “match” between the content of that non-​conceptual state and the content of our belief, and that affirmative attitude cannot be justified—​ doxastically justified—​unless we have foundational evidence that is capable of immediately justifying. So, while I will continue to operate under the assumption that all foundational evidence is capable of immediately justifying some doxastic attitude or other, those who side with Smithies can simply reframe our present discussion as a search for the special kind of foundational evidence that can immediately justify our doxastic attitudes. And even Smithies agrees that mere acquaintance states do not qualify. If the need for recognition (premise 1) is conceded, the next question becomes whether one might gain the requisite recognition without relying on background evidence (premise 2). Let’s turn to this question now. 6.2.4  Direct Recognition

Instead of relying on background evidence, is it possible for someone to directly see the match between a proposition and the content of one’s non-​ conceptual experience? Possibly so. And if so, this might furnish one with a way of rejecting premise 2. How are we to understand this act of direct recognition? It cannot consist in a belief (or other doxastic attitude) that the proposition matches one’s experiential content, for then that belief must be justified, and how is this to happen without appealing to background evidence? The classical foundationalist Richard Fumerton suggests that this direct recognition be understood as simply another instance of direct acquaintance. He thus

176  The Case for Seemings Exclusivism proposes the following three conditions on immediate justification, where p is a proposition and X is a fact—​that is, a state of affairs that obtains (Fumerton 1995, 75): i S is directly acquainted with X. ii S is directly acquainted with the thought that p (where p represents X). iii S is directly acquainted with the correspondence between the thought that p and X.9 Classical foundationalists like Fumerton and BonJour generally limit direct acquaintance to necessary facts as well as facts about one’s own conscious mind—​that is, to introspection and a priori rational intuition. However, on relationalism, perception is the direct awareness of facts in one’s immediate physical environment. Such direct awareness is seemingly sufficient to satisfy Fumerton’s first condition. Thus, the view can be stretched to allow for immediate perceptual justification as well. In any case, it should be clear by now that we have moved beyond a simple acquaintance view of immediate justification. It is no longer being suggested that mere acquaintance with a fact is foundational evidence; we must also be acquainted with conceptual content and the correspondence between these two. Thus, the foundational evidence is actually a complex mental state consisting in three separate acquaintance states. Following Huemer (2011), we can call this view (and those like it) “triple acquaintance views.” Do triple acquaintance views give a plausible, non-​conceptual candidate for foundational evidence? They do not. The problem lies not in the accounts themselves but in the fact that their chosen candidate is not non-​conceptual. What serves as foundational evidence is a complex of three different mental states, including a state with conceptual content. BonJour thus writes, “the cognitive act in question certainly cannot be entirely non-​conceptual, since it essentially involves the conceptual content of the propositional claim.” He adds, “it is also extremely natural to formulate it by saying that I see that the content of the experience fits the content of the proposition. Thus, something like a propositional element is somehow involved” (BonJour 2006, 747). Accordingly, these triple acquaintance views do not give us any support for denying premise 2. That premise says only that non-​conceptual states require background evidence in order to satisfy the recognition condition; and that principle remains as sure as ever. For if the recognition is direct, then the candidate for evidence becomes conceptual; but if the recognition is not direct, then it will require background evidence. On the other hand, what are we to make of these conceptual states of triple acquaintance? Could they serve as foundational evidence? We will take a closer look at these in the following section, but as far as non-​ conceptual states go, the last avenue of escape has been closed off.

The Case for Seemings Exclusivism  177 6.3  Conceptual States We now turn our attention to conceptual states and whether any of these might be foundational evidence. My goal in this section is not, as with non-​conceptual states, to rule out the possibility of foundational evidence entirely but rather to argue that no conceptual states other than seemings can serve in that role. We’ll start by looking at the various candidates foundationalists have put forward. The upshot is that all of them are either inadequate or else species of seemings. Before getting into it, I should address the worry that my argument against non-​conceptual states unintentionally rules out foundational evidence altogether, thereby undermining my overall defense of seeming foundationalism. This worry is understandable, though ultimately unfounded. Conceptual states do not face the same problems with the recognition condition that thwarted non-​conceptual states. As suggested above, it may be possible to directly recognize the match between a belief content and the content of a conceptual experience without relying on background evidence. This is especially plausible if those contents are one and the same—​as when, for instance, a seeming that p is thought to be evidence for p. To take this even further, there may not be any need of recognition in cases where the token content of one’s belief and the token content of the evidential state are numerically identical representations. To give an example, imagine that S forms a token thought M with propositional content p and a forceful phenomenal character. By virtue of its propositional content and forceful phenomenal character, M falls into the kind Seeming States—​that is, M is a seeming. Subsequently, S forms the belief that p. The way in which this happens, however, is not by forming a second token thought M*. Rather, M itself comes to take on the doxastic assertiveness, which is definitive of belief. Thus, by virtue of its propositional content and doxastic assertiveness, M also falls into the kind Belief States—​that is, M is now both a seeming and a belief.10 In such a case, the subject apparently doesn’t need to recognize the match between the content of her seeming and the content of her belief because they are not only of the same type, but are in fact the same token content. And recognition only seems necessary when the subject moves from one token content to another.11 At the very least, no framework of background beliefs is needed, which is the essential point. 6.3.1  Triple Acquaintance States

Let us begin our survey of conceptualist candidates with the states of triple acquaintance introduced above. Such things can be difficult to envision in the abstract, so consider an example. Say that Lois directs her attention

178  The Case for Seemings Exclusivism toward her red sensation and thinks I am being appeared to thusly. This final concept is demonstrative, its content being filled in by the state of affairs toward which Lois is attending. In securing the reference of this demonstrative concept, it is plausible that Lois is directly acquainted not only with her own thought but also with the correspondence between that thought and the target sensation. Something similar might be possible in a perceptual case. Assume, as on relationalism, that Lois is directly aware of the concrete state of affairs consisting in the person that is Superman and his property of flying. She then forms the thought that is such-​and-​such (where both of these are demonstrative concepts) while attending to that state of affairs. Plausibly, Lois is directly aware with her own thought, of a fact, and of the correspondence between them. I find it plausible that conceptual states of the sort just described can serve as foundational evidence. Setting aside worries about whether we are actually ever triple acquainted with things, if we were that would plausibly be a genuine source of immediate justification. After all, the state of affairs about which you are thinking is laid bare before your conscious mind, and you can directly see (without any need for background evidence) that your thought’s content is accurately capturing it.12 If belief is not permitted in these conditions, then when is it? Do we therefore have a counterexample to SE? We do not, due in part to the fact that triple acquaintance states are also seeming states. I argued for this identification in Chapter 4. There I proposed that forcefulness, the distinguishing feature of seemings, should be understood as what it feels like to be directly aware of the correspondence between a proposition and the fact it represents. A triple acquaintance state is thus the very paragon of a forceful mental state. Moreover, triple acquaintance states have propositional content because they involve direct awareness of the content of one’s thought. It follows that when one is in a state of triple acquaintance, necessarily, one will also be in a mental state with propositional content and a forceful phenomenal character—​that is, one has a seeming. This vindicates Huemer’s own assessment when he says, “I thus find it impossible to imagine a case in which a person has all the acquaintance states required by the Triple Acquaintance Theory without having the corresponding appearance” (Huemer 2011, 7). This, of course, does not mean that the kinds Triple Acquaintance States and Seeming States are identical. For one, not all seeming states are triple acquaintance states. One can feel like one is triple acquainted with something, and thereby have it seem true, when one is not actually triple acquainted with it. Furthermore, the features by virtue of which a mental state qualifies as a seeming are different from the features by virtue of which it qualifies as a triple acquaintance state. Specifically, factiveness is a defining feature of triple acquaintance states but not of seeming states—​that

The Case for Seemings Exclusivism  179 is to say that triple acquaintance with p entails that p is true, whereas it can seem that p but p be false. These differences notwithstanding, all token mental states that qualify as triple acquaintance states will also qualify as seemings. SE isn’t out of the hotseat quite yet, however. SE not only says that one has a seeming that p whenever one has foundational evidence for p but also that one has the foundational evidence because of the seeming. Thus, I need to show that if these token mental states are serving as foundational evidence, then they are doing so because they are seemings and not because they are triple acquaintance states. In order words, what gives them the ability to immediately justify are not those features unique to triple acquaintance states—​namely, their factiveness—​but rather the features shared by seemings—​namely, their forcefulness and propositional content. To delve into this issue further, we need to return to some earlier conclusions on the nature of justification. We concluded in Chapter 1 that the only normatively relevant properties of our mental states are those that are directly causally accessible to our automatic processing systems. This rules out the position that it is triple acquaintance states qua triple acquaintance states that justify. The problem is that none of the features that set triple acquaintance states apart from seemings are directly causally accessible to us, meaning that those features cannot be normatively relevant to whether a mental state is capable of justifying belief or not. Consider: when we are triple acquainted with something, and believe accordingly, what is the most proximate cause of our belief within a fully articulated, entirely non-​superfluous folk-​ psychological explanation? Is it that we are actually triple acquainted with that state of affairs? Or is it that we feel like we are triple acquainted with that state of affairs. Clearly, it’s the latter. On a purely psychological level, the same belief will be formed when one is apparently triple acquainted (i.e., when something seems true) as when one is actually triple acquainted. If it feels the same from the inside, it doesn’t make any psychological difference whether the acquaintance is genuine or not. In this way, the state of seeming true screens off the state of triple acquaintance—​were one to be in the former state but not the latter, one would still form belief. What we are seeing here, as we saw in our discussion of disjunctivism in Chapter 1, §1.4.3, is that factiveness, the distinguishing feature of triple acquaintance states, is not directly accessible to our automatic processing systems. It follows that token triple acquaintance states do not directly cause belief because they are triple acquaintance states but because they feel like triple acquaintance states—​that is, because they possess the kind of forceful phenomenal character definitive of seemings. And so it is by virtue of being seemings, if anything, that such token states are capable of immediately justifying belief.

180  The Case for Seemings Exclusivism Tellingly, one of the main proponents of triple acquaintance, Fumerton, effectively concedes that factiveness is not playing an essential role in the process of immediate justification (see also Hasan 2013b, 130). Fumerton allows the belief that p to be immediately justified in the following conditions (Fumerton 1995, 77 and 186): i S is directly acquainted with X. ii S is directly acquainted with the thought that p, where p represents state of affairs Y and Y is very similar but not identical to X. iii S is directly acquainted with something very similar to the correspondence between the thought that p and X. In these conditions, there is no genuine acquaintance with the fact represented in one’s thought, nor with the correspondence between one’s thought and that fact. It feels like one is directly acquainted with these things, to some extent, but those feelings are not genuine. Yet, Fumerton concedes, the belief that p is still to some degree immediately justified. This suggests that even when the triple acquaintance is genuine, it is not triple acquaintance per se that is doing the justifying; instead, it is feeling like one is triple acquainted that is doing the justificatory work. In other words, it is because of the state’s forceful phenomenal character and propositional content that it is capable of immediately justifying belief (assuming it is). And these are the defining characteristics of seemings. 6.3.2  Sensory States

What other mental states have been put forward as potential candidates for foundational evidence? A natural proposal is that sensory appearances might be foundational evidence for our perceptual beliefs. The most pressing problem with this proposal is that sensory states—​those with the distinctive sensory phenomenology we all know and love—​do not have conceptual content. I cited several reasons for thinking this back in Chapter 4, §4.3, where I used this to distinguish sensory appearances from perceptual seemings (which have propositional content). Despite the arguments against it, there are still some who defend conceptualism about sensory appearances. Let us imagine that they are correct. If so, then it is no longer clear that sensory appearances, so understood, are distinct from perceptual seemings. Indeed, the main reason I gave for distinguishing perceptual seemings from sensory appearances was that the former had propositional content while the latter did not. If we are granting propositional content to sensory appearances, however, then there is no longer any reason to draw that distinction. Sensory appearances would simply be a species of seeming—​one with a uniquely sensory way of

The Case for Seemings Exclusivism  181 conceptually representing states of affairs. In that case, however, sensory appearances would pose no threat to SE. What if we tried to imagine sensory states with propositional content but lacking any assertive force? Earl Conee suggests that certain sensory states might justify perceptual beliefs even if they do not have any distinctive “feel of truth” (Conee 2013, §6). In assessing this proposal, our first task is to get clear on what these sensory states are supposed to be like. They should display the full richness of sensation displayed in ordinary perception but lack whatever phenomenal character makes it feel as though they are descriptive of the way things really are. This is difficult to imagine. Perhaps it is even impossible. Part of the difference between being in a sensory state and simply bringing sensations to mind (as one might do in imagination) consists in the fact that the former sensations feel as though they are authentic presentations of reality. To the extent that we can imagine sensory states without forcefulness, however, it would have to feel something like immersing oneself in an extremely detailed virtual reality gaming system. All sorts of things are represented in rich sensory detail, but since you know it’s a game, it doesn’t feel real to you in the way that ordinary perceptual experience does.13 Could states like this be foundational evidence? At first blush, it may seem so. Sure, they don’t feel true in and of themselves, but a moment’s reflection highlights some striking features. For one, the sensations displayed in them are especially vibrant or, as Hume might say, “vivacious”—​certainly more so than in ordinary cases of imagining something. Another interesting feature is that these states are passively received, whereas in imagination one has an obvious role in their generation. These sensory states are also richly detailed (far more than in dreams, for instance) and display a kind of coherence with each other across time. What do we make of all this? The best explanation, we may assume, is that such states are reliable indications of the way things really are, in which case those sensory states serve as evidence for our perceptual beliefs. This seems to be something like what Conee has in mind (Conee 2013, 62–​63). Alas, none of this shows that non-​forceful sensory states are foundational evidence. They are only capable of justifying after reflection on their vividness, passive reception, richness, coherence, and the like—​features whose significance can be ascertained only when situated within a larger framework of background evidence. Now is a good time to address an example from Kevin McCain, which might be construed as issuing a challenge to seemings exclusivism on the basis of sensory states (McCain 2017). McCain re-​envisions Plantinga’s example of the inflexible climber, which we discussed in Chapter 5 as an objection to coherentism. The example involves a rock climber whose beliefs are “locked in place” thanks to an unlucky burst of cosmic radiation.

182  The Case for Seemings Exclusivism Though his experiences of the world continue to change, his beliefs remain fixed. These beliefs remain as coherent with one another as they ever were and, thus, on coherentism, should remain justified. However, it seems evident that the climber’s beliefs are not justified. This is because those beliefs no longer match his experiences, which implies that those experiences are evidence that the climber must take into account, contra coherentism. McCain amends the example by adding that the climber’s seemings are fixed in addition to his beliefs—​only his sensory states change. In this revised case, it still seems that the climber’s beliefs are not justified. This plausibly implies, as before, that those sensory states are evidence that the climber must take into account, contra seemings exclusivism. In treating this example, we must first ask how to understand these sensory states. If they are Reidian sensations (raw feels) or non-​conceptual states (either representations or relations), then I have already shown above why they cannot be foundational evidence. That being said, there is nothing stopping them from being non-​foundational evidence if one has the right kind of background evidence. This explains why the climber’s beliefs are no longer justified. As an ordinary person, the climber will have evidence about the significance of his sensory states operating in the background; hence, those sensory states serve as non-​foundational evidence for the climber—​evidence that is not being taken into account by his unchanging beliefs.14 The same thing will be true if we understand the climber’s sensory states as non-​assertive states with conceptual content. They can and will serve as evidence, just not foundational evidence. Finally, if that climber’s sensory states are understood as assertive states with conceptual content (propositional content specifically), then they are simply seemings and the example is incoherent as described, for holding fixed his seemings would also require holding fixed his sensory states. In any case, McCain’s example is not a threat to seemings exclusivism. 6.3.3  Affective States

Some maintain that emotions serve as foundational evidence for evaluative judgments (e.g., Döring 2007). This is a position known as emotional dogmatism. Emotional dogmatism looks very much like phenomenal conservatism except that emotions are identified as foundational evidence rather than seemings. I will argue here that emotions are, sadly, unsuited to the role. Though our discussion will focus exclusively on emotions, the same considerations will also disqualify other affective states such as desires, hopes, or even aesthetic experiences. To begin, if emotions turn out to be non-​conceptual states, then they will flounder for reasons already articulated. But what if emotions are conceptual? There are plausible theories on which some (if not all) emotions

The Case for Seemings Exclusivism  183 have propositional content. Take Robert Roberts theory of emotions as concern-​based construals (Roberts 1988; 2003). On this view, emotions present some state of affairs to us as good or bad in certain way—​they have evaluative force (see Chapter 4, §4.3.2). And at least sometimes that state of affairs is represented via propositional content. To be clear, the evaluation of this state of affairs is not a part of the emotion’s propositional content per se but is instead contained in the distinctive force with which that propositional content is presented. For instance, if we are angry at our friend’s mistreatment, then the propositional content of our anger is simply our friend is being mistreated, and not our friend’s mistreatment is worthy of indignation. We saw something similar with seemings—​for example, if it seems that 1+​1=​2, then the propositional content of our seeming is 1+​1=​2, not 1+​1=​2 is true. Could emotions so construed be foundational evidence? Evidence for what, exactly? We could go two ways here. We could say that an emotion with propositional content p is foundational evidence for p or that it’s foundational evidence for q, where q is it is good or bad that p in some particular way. For instance, we could say that feeling angry about our friend’s mistreatment is supposed to be evidence that our friend is being mistreated or that our friend’s mistreatment is worthy of indignation. There are problems with either option. Starting with the former, the main issue is that feeling good or bad about something does not permit us to believe the corresponding propositions to be true (or false, for that matter). The fact that we feel indignation while thinking about the mistreatment of a friend does not in and of itself give us any evidence that our friend actually is being mistreated. If I simply imagined the mistreatment, and found myself getting angry just thinking about it, I wouldn’t thereby gain evidence that mistreatment is occurring. Of course, we usually only get worked up about states of affairs that we think actually obtain, but that makes our anger a response to evidence of mistreatment, not evidence of mistreatment itself. To take a positive emotion, consider a brand-​new father whose infant son is in the intensive care unit. He feels comfort and joy at the thought of his son fully recovering and coming home with him and his wife. These positive emotions toward his son’s recovery are not evidence that his son will recover. They are not the sorts of things that permit belief from a purely epistemic point of view, at least not by themselves. This is because they do not in themselves indicate the truth of that proposition, only its goodness. Now, we may come to learn that feeling joyful or angry about a proposition is, in select contexts, an indicator of its truth; but even then, the emotions would not be foundational evidence since they gain their justificatory power from our background evidence. What if the content of the emotion feels not only good or bad, but true? Say, for instance, that the newborn’s father has a strong feeling that his

184  The Case for Seemings Exclusivism son really will recover in addition to the comfort and joy he feels at the prospect. Or say that we feel our friend really is being mistreated, as well as feeling indignant about this. Such states might serve as foundational evidence for their content, but this is no threat to seemings exclusivism since those states are also seemings. They are states with propositional content that feels true. What seems to be going on here is the sort of situation discussed at the beginning of §6.3, where the same token mental state qualifies as both a seeming and an emotion. Seemings and emotions continue to be different types of mental states—​the former characterized by assertive force and the latter by evaluative force—​but token representations can simultaneously meet the requirements for both. What’s more, these emotional mental states only stand a chance at immediately justifying belief by virtue of their felt truth (the properties that qualify them as seemings), not their felt goodness (the properties that qualify them as emotions). Thus, it is qua seemings, not qua emotions, that the token states are doing any justificatory work. On the other hand, if an emotion is purported to support not its own propositional content, p, but some other proposition, q, then it will run into the same obstacles that tripped up non-​conceptual states. We must be able to recognize that q “matches” that emotion. For example, a negative emotion may phenomenally present its object as offensive, disgusting, or harmful. We must therefore be in a position to recognize that the propositional content of our belief, and the evaluative concepts deployed in it, accurately captures the particular way in which the target state of affairs is being presented as bad. This recognition is not always easy to attain. Miranda Fricker rightly says that “the language of emotion is not reducible to the language of words,” and that “rational, linguistic processes” are needed to articulate “those nebulous or ‘subterranean’ meanings which emotions already possess” (Fricker 1991, 15). The essential point, of course, is not that such recognition is difficult, but that it cannot occur apart from background evidence about evaluative properties and how those properties manifest themselves in the phenomenology of an emotion. This means that emotions cannot be foundational evidence. We could try to avoid the need for background evidence by positing some kind of direct awareness of the relationship between q and the emotion, but this is just another appeal to something like triple acquaintance states. (We are acquainted with our emotion, with a thought, and with the correspondence between them.) We know how this plays out: it’s not the emotion per se that immediately justifies but a complex conceptual state that turns out to be a species of seeming. None of this, by the way, suggests that emotions cannot be strong evidence for believing something to be good or bad; only that they are not foundational evidence for these judgments. The same is true of beliefs and

The Case for Seemings Exclusivism  185 sensory states—​they can be evidence, just not foundational evidence. This is in no way a minimizing position to take toward emotions or their role in rational cognition. 6.4  Seemings or Bust At this point, one might ask how many times we need to beat our heads against the wall before we admit that there is no way through. Perhaps we can cut off any remaining hope (and save ourselves the headache) with the following reasoning. Consider a mental state M that is supposed to be foundational evidence for p for S. Does M have propositional content? If not, then it cannot satisfy the recognition condition apart from background evidence, and so cannot be foundational. If M does have propositional content, however, then does that content feel true or not? If it does not feel true, then M is incapable of permitting belief apart from background evidence, in which case whatever evidential support M provides is not foundational. But if the content of M does feel true, then M is a seeming. Moreover, it is by virtue of its propositional content and forceful character—​its seemingness—​that it possesses whatever justificatory power it does. For it is precisely when such things are present or absent that immediate justification plausibly can or cannot occur. Thus, any mental state capable of immediately justifying belief is a seeming, and it is only capable of justifying because it is a seeming. This is an in-​principle argument for SE. The main step of this reasoning that needs further elaboration is the claim that forcefulness, the feel of truth, is necessary for permitting belief apart from background evidence. We saw this illustrated with non-​forceful sensory states, but I want to generalize the point here. The content of non-​ forceful states does not feel true. These states may have propositional content, but they do not present that content to us as being representative of the way things actually are. Examples include imagining p, considering whether p, or merely entertaining p. Envision having such a mental state in the state of empty symmetrical evidence. That is, we entertain the thought that p but have zero evidence concerning the reliability of that mental state or its truth. Does this state by itself make permissible belief in its content? The answer is plainly no. On a purely psychological level, such a mental state wouldn’t incline us toward belief in the first place. And even if it somehow did, the resulting belief would not be permitted by our duties to pursue truth and avoid falsehood. Belief would be completely arbitrary from our first-​person point of view, making it a reckless bet. The problem is that there is nothing indicating the truth of that content. Nothing about those non-​forceful states in and of themselves connects that content with truth, and there is no background evidence with which to draw such a

186  The Case for Seemings Exclusivism connection either. As before, this is not to say that non-​forceful mental states can’t be evidence at all, but only that they cannot serve as evidence apart from other background evidence. In a nutshell, for a mental state to be a plausible candidate for foundational evidence, it’s got to have propositional content and that propositional content has got to feel true, but if it has both those things then it’s a seeming. Thus, no mental states other than seemings are plausible candidates for foundational evidence. Notes 1 One could construe epistemic conservatism as saying that beliefs are foundational evidence for themselves (McCain 2021, §4), but it is usually understood as the view that immediate justification can be present in the absence of foundational evidence altogether (McCain 2008, 187). Regardless, I argue against epistemic conservatism in the following chapter. 2 I am hereby using the notion of concepts in what Byrne (2005, 231) calls the “Fregean sense” (pun intended). 3 Salmon (1986) says we can approach the same state of affairs under different guises; Braun (2002) that there are different ways of believing. These guises, or ways of believing, are the bearers of cognitive significance mediating between the subject and the state of affairs grasped. The major difference between these approaches and my own is that Salmon and Braun assume a Russellian view of propositions. That is, they identify the concrete state of affairs as the proposition. In contrast, I assume a Fregean view of propositions, on which the bearer of cognitive significance—​the mental representation mediating between the subject and the concrete state of affairs—​is the proposition. Such differences do not, for our purposes, affect the essential lesson of Frege’s puzzle, which is that beliefs require a representational mediator between subject and object. 4 Sellars makes a different but related argument in Sellars (1956, 297–​298). 5 A tip of the cap to my logic student, Anna Jackson, for finding a direct proof of this argument in symbolic logic. 6 Consistent with my approach throughout this book, I am bracketing out externalist options that privilege certain conceptual contents by virtue of, say, the fact that the process of forming those sorts of beliefs in response to those sorts of non-​conceptual experiences is a reliable one (which is perhaps something like what Peacocke is thinking here). 7 BonJour seems initially reluctant to admit that this direct recognition is a cognitive act distinct from one’s original, built-​in awareness of one’s experiential and belief contents (BonJour and Sosa 2003, 193), but he eventually confirms it as such (BonJour 2006, 746–​747). 8 Smithies focuses more on the fact that one’s belief is not safe from lacking evidential support (Smithies 2019, ­chapter 10, §2), but difficulties with recognition seem to be responsible for that lack of safety.

The Case for Seemings Exclusivism  187 9 I believe this to be a largely faithful representation of BonJour’s ultimate position as well, though his terminology is different. See also, for example, Hasan (2013a) and DePoe (2011; 2021). 10 I discuss the possibility of a token state being both a seeming and a belief in Chapter 4, §4.1.1. 11 If we go in this direction, we may need to amend the recognition condition to say that recognition is needed except in these specific cases. 12 I take it that acquaintance with the correspondence between thought and fact resolves the dilemma raised in Harrison (2021, §3.3). 13 You’d have to keep reminding yourself it’s not real, of course, and even then there’s no guarantee that things wouldn’t suddenly start feeling real. After all, the popularity of virtual reality is that it is easy to “lose yourself” in the virtual environment and start reacting to the things presented therein as though they are genuinely happening. 14 It is hard to envision anyone lacking any and all background evidence about what sorts of sensory phenomenology are indicative of which sorts of states of affairs; however, if the person genuinely had no idea what to make of his sensory phenomenology—​if it was a completely indecipherable flux of color and shape with no articulable significance—​then I think the climber’s beliefs would remain justified, despite his changing sensory states.

References BonJour, Laurence. 1978. “Can Empirical Knowledge Have a Foundation?” American Philosophical Quarterly 15 (1): 1–​14. —​—​—​. 2006. “Replies.” Philosophical Studies 131 (3): 743–​759. BonJour, Laurence, and Ernest Sosa. 2003. Epistemic Justification: Internalism Vs. Externalism, Foundations Vs. Virtues. Hoboken, NJ: Wiley-​Blackwell. Braun, David. 2002. “Cognitive Significance, Attitude Ascriptions, and Ways of Believing Propositions.” Philosophical Studies 108 (1–​2): 65–​81. Brewer, Bill. 1999. Perception and Reason. Oxford: Oxford University Press. —​—​—​. 2011. Perception and Its Objects. Oxford: Oxford University Press. Byrne, Alex. 2005. “Perception and Conceptual Content.” In Contemporary Debates in Epistemology, edited by Ernest Sosa and Matthias Steup, First edition, 231–​250. Oxford: Blackwell. Campbell, John. 2002. Reference and Consciousness. Oxford: Oxford University Press. Chisholm, Roderick M. 1942. “The Problem of the Speckled Hen.” Mind 51 (204): 368–​373. Conee, Earl. 2013. “Seeming Evidence.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 52–​68. Oxford: Oxford University Press. Conee, Earl, and Richard Feldman. 2008. “Evidence.” In Epistemology: New Essays, edited by Quentin Smith, 83–​104. Oxford: Oxford University Press. Davidson, Donald. 1986. “A Coherence Theory of Truth and Knowledge.” In Truth and Interpretation: Perspectives on the Philosophy of Donald Davidson, edited by Ernest LePore, 307–​319. Oxford: Blackwell.

188  The Case for Seemings Exclusivism DePoe, John. 2011. “Defeating the Self-​ Defeat Argument for Phenomenal Conservativism.” Philosophical Studies 152 (3): 347–​59. —​—​—​. 2021. “Justification by Acquaintance.” Synthese 199, 7555–​7573. Döring, Sabine A. 2007. “Seeing What to Do: Affective Perception and Rational Motivation.” Dialectica 61 (3): 363–​394. Evans, Gareth. 1982. The Varieties of Reference. Oxford: Oxford University Press. Fish, William. 2009. Perception, Hallucination, and Illusion. Oxford: Oxford University Press. Frege, Gottlob. 1892. Translations from the Philosophical Writings of Gottlob Frege, edited by Peter Geach and Max Black. Oxford: Blackwell. Fricker, Miranda. 1991. “Reason and Emotion.” Radical Philosophy 57 (Spring): 14–​19. Fumerton, Richard. 1995. Metaepistemology and Skepticism. Lanham, MD: Rowman & Littlefield. Harrison, Eilidh. 2021 “Seemings, Truth-​Makers, and Epistemic Justification.” Synthese 198 (6): 5689–​5708. Hasan, Ali. 2013a. “Internalist Foundationalism and the Sellarsian Dilemma.” Res Philosophica 90 (2): 171–​184. —​ —​ —​ . 2013b. “Phenomenal Conservatism, Classical Foundationalism, and Internalist Justification.” Philosophical Studies 162 (2): 119–​141. Heck, Richard. 2000. “Nonconceptual Content and the ‘Space of Reasons.’” Philosophical Review 109 (4): 483–​523. Huemer, Michael. 2011. “Phenomenal Conservatism and Self-​Defeat: A Reply to DePoe.” Philosophical Studies 156 (1): 1–​13. McCain, Kevin. 2008. “The Virtues of Epistemic Conservatism.” Synthese 164 (2): 185–​200. —​—​—​. 2017. “An Isolation Objection to Phenomenal Conservatism.” Erkenntnis 82 (6): 1381–​1390. —​—​—​. 2021. “Epistemic Conservatism: A Non-​Evidentialist Epistemology?” In Non-​Evidentialist Epistemology, edited by Luca Moretti and Nikolaj Pedersen, 3: 152–​169. Brill Studies in Skepticism. Leiden, NL: Brill. McCain, Kevin, and Luca Moretti. 2021. Appearance and Explanation: Phenomenal Explanationism in Epistemology. Oxford: Oxford University Press. McDowell, John. 1994. Mind and World. Cambridge, MA: Harvard University Press. Peacocke, Christopher. 1983. Sense and Content: Experience, Thought and Their Relations. Oxford: Oxford University Press. —​—​—​. 1992. A Study of Concepts. Cambridge, MA: MIT Press. —​—​—​. 2001. “Does Perception Have a Nonconceptual Content?” Journal of Philosophy 98 (5): 239–​264. Reid, Thomas. [1764] 1997. An Inquiry into the Human Mind on the Principles of Common Sense. Edited by Derek R. Brooks. University Park, PA: Pennsylvania State University Press. —​—​—​. [1785] 2002. Essays on the Intellectual Powers of Man, edited by Derek R. Brooks. University Park, PA: Pennsylvania State University Press. Reiland, Indrek. 2015. “Experience, Seemings, and Evidence.” Pacific Philosophical Quarterly 96 (4): 510–​534.

The Case for Seemings Exclusivism  189 Roberts, Robert C. 1988. “What an Emotion Is: A Sketch.” Philosophical Review 97 (April): 183–​209. —​—​—​. 2003. Emotions: An Essay in Aid of Moral Psychology. Cambridge: Cambridge University Press. Russell, Bertrand. 1912. The Problems of Philosophy. London: Williams and Norgate. Salmon, Nathan. 1986. Frege’s Puzzle. Cambridge, MA: MIT Press. Searle, John. 1983. Intentionality: An Essay in the Philosophy of Mind. Cambridge: Cambridge University Press. Sellars, Wilfrid. 1956. “Empiricism and the Philosophy of Mind.” Minnesota Studies in the Philosophy of Science 1: 253–​329. Siegel, Susanna. 2010. The Contents of Visual Experience. Oxford: Oxford University Press. Smithies, Declan. 2019. The Epistemic Role of Consciousness. Oxford: Oxford University Press. Tye, Michael. 1995. Ten Problems of Consciousness: A Representational Theory of the Phenomenal Mind. Cambridge, MA: MIT Press.

7 Against Epistemic Conservatism

Having established seemings as the only plausible candidates for foundational evidence, we now consider whether immediate justification might be attainable apart from foundational evidence. Assuming internalism, the only position I know of that allows for this (with any proponents) is epistemic conservatism. Epistemic conservatism says that immediate justification for our basic beliefs arises simply from the possession of those beliefs in and of themselves. The bare state of having a belief gives us some presumption in its favor. In light of phenomenal conservatism, epistemic conservatism is perhaps better titled “doxastic conservatism,” although I will continue to use the traditional moniker for continuity. Epistemic conservatism, in one form or other, boasts a surprisingly impressive list of adherents, including (arguably) Thomas Reid, Roderick Chisholm, William Lycan, and Richard Swinburne. Though its popularity in the contemporary landscape has waned in light of well-​known criticisms (e.g., Foley 1983, Christensen 1994), recent work shows epistemic conservatism to be more plausible and resilient than it is usually given credit for (McCain 2008; 2019; 2021, Poston 2012; 2014). In particular, epistemic conservatism promises to do serious theoretical work if admitted. This chapter offers a critique of epistemic conservatism. The objections are not directed at the conservatism but at the claim that the conservative presumption derives simply from the possession of a belief in and of itself. Eventually, epistemic conservatism must admit of justification even in cases where the subject has no evidence, nor any indication at all that his belief is true—​a claim that’s implausible on its face. This implausibility might be overcome by theoretical benefits, except that phenomenal conservatism provides all the same theoretical benefits while avoiding the initial implausibility.

DOI: 10.4324/9781003281108-8

Against Epistemic Conservatism  191 7.1  Epistemic Conservatism Before voicing our critique, it is first necessary that we understand precisely what we are critiquing. There are many different principles that march under the banner of epistemic conservatism, some of which do not concern immediate justification at all. These include forms of what Hamid Vahid calls “differential” and “perseverance conservatism” (Vahid 2004). For instance, Lawrence Sklar’s differential conservatism maintains only that, when faced with an equally good alternative hypothesis, one is justified in continuing to believe a previously held hypothesis if that belief was initially formed “on the basis of whatever positive warrant may accrue to it from the evidence, a priori plausibility, and so forth” (Sklar 1975, 378). These forms of conservatism “are concerned with the justificatory status of a belief after its conception and acquiring positive epistemic value” (Vahid 2004, 113). Prominent proponents of differential and perseverance conservatism, respectively, include W.V.O. Quine (Quine and Ullian 1970) and Gilbert Harman (1986). Despite their pedigree, such principles do not bear on the immediate justification of basic beliefs and so do not fall within our purview. Conservative principles that do pertain to immediate justification are forms of what Vahid calls “generation conservatism” (Vahid 2004). Unrestricted versions extend the same initial presumption to all beliefs, whereas restricted versions grant it only to some. For instance, Chisholm endorses a fairly unrestricted form of epistemic conservatism: “Anything we find ourselves believing may be said to have some presumption in its favor—​provided it is not explicitly contradicted by the set of other things we believe” (Chisholm 1980, 551–​552). Whereas Swinburne restricts his principle of credulity to basic beliefs: “every proposition that a subject believes or is inclined to believe has (in so far as it is basic) in his noetic structure a probability corresponding to the strength of the belief or semi-​ belief or inclination to believe” (Swinburne 2001, 141). Finally, Reid grants immediate justification to all and only natural beliefs—​that is, those formed by the proper functioning of our natural constitutions (McAllister 2016). Formulations can also differ with respect to the exact status bestowed on beliefs, as these examples illustrate, as well as the conditions for its maintenance and defeat. On this final point, some versions of generation conservatism, such as Chisholm’s, claim only that belief grants the agent some “presumption in its favor” (see also Lycan 1988). Chisholm explicitly defines this to mean, “Accepting h is more reasonable for S than accepting not-​h” (Chisholm 1977, 8). As McCain points out, presumption thus defined does not imply that the belief is justified, even in the absence of defeaters (McCain 2019, 203–​ 204). These weaker versions of epistemic conservatism are often

192  Against Epistemic Conservatism favored by weak foundationalists, or impure coherentists, who maintain that full justification requires background beliefs. The idea is that the presumption in favor of our beliefs is leveraged into justification only by being incorporated into a coherent and mutually supporting framework of such beliefs. These weaker versions of epistemic conservatism can bypass some of the objections faced by their stronger brethren (McCain 2019), but they do so at the cost of theoretical potential. Most of the theoretical benefits claimed by epistemic conservatism require that it explain the immediate justification of our beliefs. For convenience, I will focus the discussion on the stronger versions of epistemic conservatism—​those pertaining to immediate justification—​although the general criticism I offer will apply to these weaker principles as well. The point is that whatever work can be done by epistemic conservatism can be done by phenomenal conservatism with less cost. Among the stronger versions of epistemic conservatism, the most plausible formulation comes from Kevin McCain (2008, 186): EC—​If S believes that p and p is not incoherent, then S is justified in retaining the belief that p and S remains justified in believing that p so long as p is not defeated for S. Notice that EC does not say that beliefs are evidence for themselves (c.f. Adler 1996). McCain clarifies that, on EC, believing p bolsters one’s justification for p, but not by providing evidence for p (McCain 2008, 187). The idea behind EC, and these versions of epistemic conservatism more generally, is that our beliefs can sometimes be justified without any evidence for them at all.1 (McCain talks about “reasons” rather than “evidence,” but, for the purposes of this discussion, I will treat these as equivalent.) Because it admits of justification absent evidence (or any rational grounds at all), it is generally conceded that epistemic conservatism is not especially intuitive. Indeed, I think it is fair to say that it is positively counterintuitive by most lights. The main complaint (expressed as a question) is: Why should the mere fact that I believe something even if I have no evidence for doing so justify me in continuing to believing it? As Sklar puts it, “surely not only is believing p not sufficient grounds for believing p…; in general, believing p is no grounds at all for believing p” (Sklar 1975, 377). This is made especially salient by considering that, “A person may happen to hold a belief because of whim, prejudice, emotional manipulation, specious arguments, self-​interested rationalization, subliminal advertising, drugs, or post-​hypnotic suggestion” (Goldstick 1971, 187). I will unpack these intuitions in a moment, but first it is worth asking why anyone would endorse such a principle.

Against Epistemic Conservatism  193 7.2  Theoretical Motivations for Epistemic Conservatism Usually, people tolerate EC and principles like it because it is theoretically fruitful. Some of the uses ascribed to it are suspect (Vahid 2004, 99–​101), but others appear genuine. Among these, the most significant advantages claimed for epistemic conservatism are that it helps us make sense of (i) common sense philosophical methodologies; (ii) the justification of belief in the external world despite underdetermination by the evidence; (iii) the justification of memory beliefs; and (iv) cases of forgotten evidence. Another, oft-​cited advantage of epistemic conservatism is that it helps account for our conservatism in belief-​revision, but this seems more a matter of differential or perseverance conservatism rather than generational. The introduction of phenomenal conservatism severely undercuts these motivations. On phenomenal conservatism, immediate justification arises not from the beliefs themselves but from the apparently revelatory experiences on which they are based. Most of the things epistemic conservatism promises to explain can be accommodated as well or better by its phenomenal cognate. As Vahid says, “With such a view in place, there would be no need to appeal to doxastic conservatism” (Vahid 2023). Add to this that phenomenal conservatism is significantly more intuitive. Thus, if there is need to posit one of these two principles to secure the aforementioned advantages, then phenomenal conservatism seems the better choice. As a result, the motivation for epistemic conservatism is lacking from the get-​go. To shore up my critique, let us quickly run through some of the areas where epistemic conservatism is supposed to help and see how phenomenal conservatism handles them. Remember, given the greater intrinsic plausibility of phenomenal conservatism, I only need to show that it explains things as well as epistemic conservatism (not better than it). 7.2.1  Common Sense Methodology

A common sense philosophical methodology is sometimes summarized: Start with the obvious and never give up the more obvious for the less. This approach has two components to it. First, “The metaphilosophy of common sense,” Stephen Boulter writes, “insists that common sense beliefs are to be treated as default positions. … one has shifted the burden of proof onto the shoulders of those who would reject common sense beliefs” (Boulter 2007, 24). Second, this presumption is taken to be quite formidable, generally warranting the rejection of philosophical theories or arguments when they come into conflict with common sense beliefs. This is sometimes true even if their exact error remains unknown. Reid writes:

194  Against Epistemic Conservatism Zeno endeavoured to demonstrate the impossibility of motion; Hobbes, that there was no difference between right and wrong; and [Hume], that no credit is to be given to our senses, to memory, or even to demonstration. Such philosophy is justly ridiculous, even to those who cannot detect the fallacy of it. It can have no other tendency, than to shew the acuteness of the sophist, at the expense of disgracing reason and human nature, and making mankind Yahoos. (Reid [1764] 1997, 21) The rejection of philosophical theories in this way has been called “the G.E. Moore Shift” (Rowe 1979) because of its use by G.E. Moore (Moore [1925] 1959, 41), although Reid almost certainly served as Moore’s inspiration. While it is controversial how much weight is to be given to our common sense beliefs, the fact that they are given some presumption—​that they serve as our starting points in philosophical theorizing—​is ubiquitous. Going back to ancient times, Aristotle recommends that we begin with the endoxa—​often translated as “reputable,” “received,” or “credible” beliefs or opinions—​and seek to preserve as many of them as possible. He writes: As in our other discussions, we must first set out the way things appear to people, and then, having gone through the puzzles, proceed to prove the received opinions about these ways of being affected—​at best, all of them, or, failing that, most, and the most authoritative. For if the problems are resolved, and received opinions remain, we shall have offered sufficient proof. (Aristotle 2004, 1145b2–​7) This same mentality pervades contemporary philosophy as well. Boulter attributes it to philosophers such as Ryle, Austin, Grice, and Searle (Boulter 2007, 24–​25). It arguably underlies Rawls’s method of reaching reflective equilibrium (Rawls 1971). David Lewis directs us to “collect all the platitudes you can think of” and use these as our starting point (Lewis 1972, 256). James Pryor recommends, “[W]‌e start with what it seems intuitively natural to say about perception, and we retain that natural view until we find objections that require us to abandon it” (Pryor 2000, 538). Epistemic conservatism accounts for such methodologies by granting presumption to one’s beliefs simply because they exist; but clearly, this is not our only option. The reason common sense beliefs are so common is because they seem true—​ often so strongly that we would consider them evident or obvious. Either that or they are the natural conclusions of things that are obvious. Notice that for Aristotle, the endoxa are reputable attempts to account for the phainomena, or appearances, which

Against Epistemic Conservatism  195 are experiences similar to seemings (see Chapter 4, §4.2). Thus, we can account for the presumption given to such beliefs by appealing to a principle like phenomenal conservatism on which seemings serve as foundational evidence for their content. Indeed, phenomenal conservatism does a better job than epistemic conservatism in fitting with these methodologies. For it is not all beliefs that are used as starting points but only the ones that seem obvious (“the platitudes”), or the ones that are “intuitively natural to say,” or the ones credibly formed in response to the appearances. 7.2.2  Response to Skepticism

Another way in which epistemic conservatism is supposed to help is by furnishing a response to skepticism, specifically skepticism stemming from underdetermination by the evidence. This form of skepticism is most prominently connected to Descartes but goes back at least as far as the epistemological debates between the Stoics and the Academic Skeptics. The worry is that all of one’s evidence for, say, the existence of some tree is logically compatible with the falsity of that hypothesis. For instance, one’s experience of the tree might be produced by an evil deceiver. But if that evidence does not eliminate all of the alternatives, then, it is argued, one cannot justifiably believe in the tree on that basis. Epistemic conservatism provides a way out by granting belief in the tree immediate justification, despite uneliminated skeptical possibilities, simply because this belief is held. Phenomenal conservatism does the same by granting immediate justification to what seems to be the case (Huemer 2001, ­chapter 8). It is apparent to us that there is tree (which is why we believe it in the first place), and so we are immediately justified in believing in that tree on the basis of that experience. The crucial point is that phenomenal conservatism’s response to skepticism is just as good or better than that of epistemic conservatism. Luca Morretti has recently argued that the phenomenal conservative response to skepticism isn’t all it’s made out to be (Moretti 2018; 2020, McCain and Moretti 2021). The problem of reflective awareness, as Moretti calls it, is that: seeming-​based justification is elusive, in the sense that S’s becoming reflectively aware of her seeming that P and wondering about its possible causes results in either just destroying S’s seeming-​based justification for believing P or replacing this non-​inferential justification with inferential justification. (Moretti 2018, 267) For example, if I reflect on my perceptual seeming of a tree and inquire as to its origins, there will be multiple explanations available—​some on which

196  Against Epistemic Conservatism the content of my seeming is true (veridical explanations) and others on which it is false (deceptive explanations). If I cannot independently verify that some veridical explanation is correct, then I thereby acquire a defeater for the initial justification provided by my tree-​seeming. On the other hand, if I can independently verify the veridical explanation, then I have inferential justification for the tree’s existence which supplants any immediate justification provided by the tree-​seeming. I address Moretti’s problem of reflective awareness at length in Chapter 3, §3.1.1. In this context, suffice it to say that, while Moretti is correct in a certain respect—​the phenomenal conservative response to skepticism is more complicated than it has often been made out to be—​it is not defanged by Moretti’s considerations. Even more importantly, if phenomenal conservatism fails to provide a compelling response to skepticism for these reasons, then so does epistemic conservatism. The problem of reflective awareness applies equally to both. That is, if one reflects on the origins of one’s beliefs, one will recognize veridical and deceptive explanations. Either one cannot independently verify the veridical explanation, in which case one’s immediate justification is defeated, or one can verify it, in which case one’s immediate justification is replaced. Thus, even if the full force of the problem of reflective awareness is granted, it provides no point of separation between the two forms of conservatism. 7.2.3  Memorial Beliefs and Forgotten Evidence

The justification of memorial beliefs and cases of forgotten evidence are other areas where epistemic conservatism is supposed to help (McCain 2008, 188). On epistemic conservatism, one’s belief that, say, one did or did not eat breakfast this morning is justified not on the basis of any evidence for that claim but simply because one holds that belief (and it remains undefeated). Furthermore, in cases where we no longer remember the original justification for our beliefs, those beliefs remain justified simply by virtue of the fact that one continues to hold them (so long as they remain undefeated). Now, I am not convinced that this approach is the best way of handling such cases. It is not clear to me why, in both cases, one’s beliefs cannot continue to be justified on the basis of the original evidence one had for forming those beliefs, despite the fact that those evidential states are no longer occurrent or have been forgotten. After all, their causal effects linger in the form of one’s current beliefs, and that may be sufficient to say that those beliefs are still based on, and thus justified by, that original evidence. Assessing this further would take us too far afield. All we need to see for our purposes is that if this approach is correct, phenomenal conservatism can account for it just as well as epistemic conservatism can.

Against Epistemic Conservatism  197 One’s memorial beliefs can be justified on the basis of memorial seemings: for instance, it seems to you that you did or did not eat breakfast this morning. McCain denies that memorial beliefs are “justified by any kind of sense perception or distinctive memory experience” (McCain 2008, 118), but one needn’t go along with McCain’s denial. In fact, McCain himself has come to acknowledge a role for memorial seemings in later work (McCain and Moretti 2021). What is true, and what I figure McCain is really getting at, is that many memories (semantic memories, in particular) aren’t based on any state with anything like sensory phenomenology; but that is compatible with those memories being based on seemings in which the remembered content is presented as accurately representing what occurred in the past. Memorial seemings, like rational seemings, needn’t have any sensory phenomenology. And such seemings do in fact seem to serve as the bases for many or most of our memorial beliefs (see Chapter 8, §8.2.2.4). Likewise, a belief can remain justified even when one has forgotten one’s original evidence for it so long as the content of that belief still seems true, even if it only does so because of the evidence one previously had for it. In short, phenomenal conservatism can account for such cases in a way that mirrors epistemic conservatism. None of this should be particularly surprising. Both principles are forms of conservatism, and so share similar approaches and, in turn, advantages. I obviously haven’t surveyed every advantage claimed for epistemic conservatism, but it is fair to expect that most of them will be accommodated just as well by phenomenal conservatism. Even if a few are not, the net effect is still to greatly dilute the overall motivations for epistemic conservatism. To make matters worse (or better, depending on which camp you’re in), phenomenal conservatism also avoids the most serious problem facing epistemic conservatism. We will dissect this problem in the next section and see how phenomenal conservatism does better. 7.3  Epistemic Conservatism’s Fatal Flaw Let us return to the intuitive complaint above. Many have generated counterexamples attempting to capture this complaint. Richard Feldman makes the following contribution: Detective Jones has definitively narrowed down the suspects in a crime to two individuals, Lefty and Righty. There are good reasons to think that Lefty did it, but there are equally good reasons to think that Righty did it. There is conclusive reason to think that no one other than Lefty or Righty did it. (Feldman 2003, 144)

198  Against Epistemic Conservatism If Jones randomly believes Lefty did it, then it seems that principles like EC would grant justification to Jones’s belief, simply because that is the belief that he holds. But clearly Jones’s belief would not be justified in such circumstances; rather, he should withhold assent. The most plausible versions of epistemic conservatism try to escape such examples by narrowing their scope. In the case of EC, this is accomplished by McCain’s explication of the conditions of defeat. On McCain’s view, defeat can occur in either of two cases (McCain 2008, 186): DC1 If S has better reasons for believing that ~p than S’s reasons for believing that p, then S is no longer justified in believing that p. DC2 If S has reasons for believing that ~p which are as good as S’s reasons for believing that p and the belief that ~p coheres equally well or better than the belief that p does with S’s other beliefs, then S is no longer justified in believing that p. In the Lefty–​Righty case, Jones’s reasons for believing that Lefty committed the crime are equal to his reasons for believing that not-​Lefty (or Righty) did. Moreover, the belief that Righty is guilty presumably fits just as well with Jones’s background beliefs as his belief in Lefty’s guilt does. Thus, DC2 obtains and the justification for Jones’s belief is defeated, reconciling our intuition with EC. McCain avoids the counterintuitive implications by constricting the circumstances to which EC applies; however, the problem cannot be avoided forever. Eventually, one must bite the bullet—​accepting that belief in p can be justified even when the balance of reasons doesn’t support it—​ or else restrict the scope of one’s principle so aggressively that it doesn’t apply to anything. To prove this, I will first show that EC does not escape all counterintuitive implications. Consider a scenario designed to avoid triggering either of McCain’s conditions of defeat. In this scenario, S’s reasons for p and ~p are of equal strength, but belief that p coheres better with S’s other beliefs than belief that ~p. For instance, envision a situation in which S has strong experiential reasons for ~p but S’s background beliefs cohere better with p, balancing S’s reasons for p and ~p against one another. To put a face on it, imagine one intuits that murdering a drifter and harvesting his organs is always wrong (~p); however, one’s background beliefs cohere slightly better with consequentialism and the conclusion that, in some cases, murdering the drifter is morally permissible (p). Here EC maintains that, if S happens to believe that murdering the drifter is sometimes permissible, then that belief is justified in accordance with EC. However, the intuition of Feldman and others seems to apply in this case just as strongly as in the Lefty–​Righty case. If the reasons

Against Epistemic Conservatism  199 really are equally strong on both sides, then the only justifiable stance is to withhold assent. To add to the problem, imagine a second person S* phenomenally indistinguishable from S except that, following his or her intuition, S* believes that murdering the drifter is always wrong. Here DC2 would obtain—​the belief that the drifter’s murder is sometimes permissible coheres better with S’s other beliefs than does belief that his murder is always wrong—​and so S*’s immediate justification for believing that this is always wrong would be defeated. But why the asymmetry? In both cases, S and S* believe something when there are equal reasons on both sides. In the one case, these reasons come from one’s background beliefs and, in the other, they come immediately from experience, but why should that make any difference? As a last resort, one could argue that the proposed scenario is impossible—​ that S’s reasons for p will always be superior to the reasons for ~p when p coheres better with S’s background beliefs. If this is true, it is presumably because coherence with background beliefs is one’s exclusive source of reasons. Thus, this strategy requires doing away with experiential evidence altogether (a daunting prospect).2 Regardless, it would have the effect of eliminating the problematic cases. If S’s reasons for p and ~p are equal, this is because both p and ~p cohere equally well with S’s background evidence, in which case DC2 obtains and S’s justification for p is defeated. As before, this avoids the counterintuitive implications only by narrowing the conditions in which EC provides immediate justification. Indeed, it narrows them so much that they become almost non-​existent. Consider four possible scenarios in which S believes p, which are together exhaustive: i S believes p and S’s reasons for p are better than S’s reasons for ~p. ii S believes p and S’s reasons for p are equal to S’s reasons for ~p. iii S believes p and S’s reasons for p are worse than S’s reasons for ~p. iv S believes p and S has no reasons for p or ~p at all. If (iii) obtains, then so does DC1 and S’s immediate justification for p is defeated. If (ii) obtains, then so does DC2 and S’s immediate justification for p is once again defeated. If (i) obtains, then S retains justification for believing p; but S’s belief is also mediately justified by the stronger reasons for p present in S’s larger belief system. This inferential justification supplants the immediate justification granted by EC, rendering it irrelevant. Thus, the only scenario in which the immediate justification from EC shows itself is the extreme one in which S has no reasons for or against p at all, including no framework of background evidence or beliefs that bears on the issue. This is what we have called the state of empty symmetrical evidence (Poston 2014, 21), and it seems clear that no human ever actually

200  Against Epistemic Conservatism finds herself in it (Coren 2021). It is fair to say that this would be a much more limited role for epistemic conservatism than its proponents originally envisioned. Yet the most plausible form of epistemic conservatism, defended by Ted Poston (2014, ­chapter 2), limits itself to beliefs of exactly this sort. Poston’s version of epistemic conservatism is restricted to “mere beliefs,” which are defined as those accepted in a state empty symmetrical evidence. It states: EC*—​If S merely believes p, then S has some justification for maintaining her belief that p. Poston describes EC* as a principle of last resort, since it is only relevant in those situations in which reasons tell us nothing. This does not, however, make it inconsequential. Poston follows Wittgenstein (1969) in arguing that hinge propositions are accepted in the state of empty symmetrical evidence (Poston 2014, ­chapter 2). Hinge propositions are fundamental assumptions “upon which the door of inquiry turns,” and so must already be in place before the provision of considerations for or against any of them can even begin to occur. They include things like the reliability of our cognitive faculties, the uniformity of nature, and the stability of meaning over time. The presumption in favor of such beliefs provides just enough leverage to get inquiry off the ground, at which point Poston favors an explanatory coherentist account of justification. Though extremely restrictive, Poston’s EC* is plausible. It escapes counterexamples of the sort given above, and it is less counterintuitive to grant presumption to such bedrock assumptions if only because they are believed. Indeed, Poston asks us what alternative we really have? (Poston 2014, 39–​41). If we find ourselves in the state of mere belief, we cannot wait for various arguments to weigh in since the proffering of considerations for or against such assumptions will itself presuppose them. Whatever path we take—​believing, disbelieving, or withholding assent—​will be selected in the dark. Since we already find ourselves believing, it makes the most sense just to continue in that direction. There is another option here. It may be true that hinge propositions cannot be supported by certain kinds of evidence (such as that provided by deductive, inductive, or explanatory inferences) without relying on those very propositions, but what about foundational evidence, which does not depend on any framework of background beliefs? It is possible that the justification of hinge propositions comes not from the fact that they are believed but from the fact that they are utterly evident—​the sort of things that seem so obviously true that it is ridiculous to deny them. In which case those seemings serve as foundational evidence for those hinge propositions.

Against Epistemic Conservatism  201 We thus return to my central theme: anything epistemic conservatism can do, phenomenal conservatism can do better—​or, really, just as well but from a more plausible starting point. What’s more, reflecting on the evident character of hinge propositions should lead us to reconsider the plausibility of EC*. It turns out that EC* must grant justification to hinge propositions even if they do not seem the least bit true. Why? Because if one concedes that believing in hinge propositions is justified only when they appear true, then one must also concede that those seemings are acting as foundational evidence, in which case one is not in the state of empty symmetrical evidence. Thus, to properly evaluate EC*, we must try to envision a situation in which those hinge propositions are not evident—​a difficult prospect, as such propositions are so naturally apparent to us that it is hard to imagine feeling indifferent toward them. Nevertheless, we shall try. Imagine an epistemological Adam of sorts—​an adult human brought into the world fully formed. Our Adam has the conceptual framework necessary to form thoughts but has no indication whatsoever as to how the world works. He hears a sound. The question comes to mind whether there must be some cause of this sound. Again, our Adam has no background beliefs that bear on this claim. Moreover, the proposition does not in any way strike him as being true. Or even plausible. As far as its apparent truth is concerned, he feels completely indifferent toward it. It feels no different than if he were considering any random proposition—​ say, that in 2073 the Royals will win the World Series. The one difference is that he finds himself assenting to the former claim. He, for no reason at all, believes that the sound has a cause. Is Adam’s belief justified? I submit that it is not.3 If the proposition is not believed because of its apparent truth, then why is it affirmed? The answer seems to be that the belief is, from the agent’s perspective, brutely caused, having no basis whatsoever.4 From the subject’s perspective, how could belief in such conditions be anything other than completely arbitrary? One way of gauging justification is asking whether it makes sense for someone solely concerned with securing true beliefs and avoiding false ones to risk assent. Well, if there is no evidence counting in favor of a belief, and it doesn’t even seem true in the slightest, then that belief doesn’t appear a risk worth taking. Perhaps if the pursuit of truth were all that mattered, then belief might be wagered, but the goal of avoiding falsehood prohibits reckless doxastic ventures, which is precisely what belief in these conditions would be. In such an instance, Locke seems more or less right in saying, He that believes without having any Reason for believing, may be in love with his own Fancies; but neither seeks Truth as he ought, no pays the

202  Against Epistemic Conservatism Obedience due to his Maker, who would have him use those discerning Faculties he has given him, to keep him out of Mistake and Errour. (Locke [1689] 1975, 687–​688) This objection is at the core of the intuitive complaint with which we began. There has to be some indication that a proposition is true before it can be affirmed in a way that isn’t reckless with respect to the avoidance of falsehood. There has to be some on-​balance evidential support for it. And epistemic conservatism says that belief can be justified even if there is no evidence at all. This is the unavoidable problem plaguing all forms of epistemic conservatism and, in my judgment, a decisive one. Granted, the epistemic conservative might have an out if there was no alternative; but there is. Phenomenal conservatism provides a decidedly more plausible option. For on this view, the claim is that Adam is justified in believing that the sound has a cause because that proposition seems obviously true to him. Belief seems a sensible response to such obviousness for those concerned with finding the truth and avoiding falsehood. That’s a wager worth making. The case for this must wait till the next chapter, but, for now, we can at least say that this is a much more plausible starting point than on epistemic conservatism. Thus, phenomenal conservatism not only serves the same purposes as epistemic conservatism, it does so from a more intuitive foundation. 7.4  Reid’s Defense In my judgment, the strongest defense that can be offered for epistemic conservatism comes from early modern philosopher Thomas Reid. These arguments are largely neglected in contemporary debates surrounding epistemic conservatism, to their loss. Alas, these too are undermined by the introduction of phenomenal conservatism. That is, they work just as well or better as arguments for phenomenal conservatism than for epistemic conservatism, as we shall see. Reid’s first argument appeals to the irresistibility of certain beliefs (belief in hinge propositions being prime candidates). These are beliefs that cannot be changed by any human effort, either direct or indirect. Regarding such, Reid writes “An invincible Error of the Understanding, of Memory, of Judgment or of Reasoning is not imputable for this very Reason that it is invincible” (Reid [1772] 2002, 66). The idea is that we cannot be at fault for something we do not control (ought implies can), and so if we do not have power to resist forming a belief, that belief is ipso facto permissible for us (McAllister 2016). Reid’s reasoning makes sense if our notion of epistemic responsibility required voluntary control over our beliefs, for it is true that we have no control over many of our beliefs in this sense. But,

Against Epistemic Conservatism  203 as seen in Chapter 1, contemporary accounts of epistemic responsibility do not require voluntary control, doxastic voluntarism being largely maligned. We also saw in that chapter that we retain a more moderate form of control over the relevant beliefs in that they result from cognitive dispositions that are our own, and it is possible for us to be otherwise disposed. In this sense we can resist even “irresistible” beliefs, and we should. If the irresistible belief is not occasioned by some evidence of its truth, then our previous complaint returns: in the complete absence of evidence, believing isn’t justified. Reid’s second argument appeals to the naturalness of certain beliefs (again, belief in hinge propositions being prime candidates). Natural beliefs are those formed in accordance with innate principles of our constitutions. The fact that there is nothing we did to bring about these beliefs is thought to shield us from any responsibility pertaining to them (McAllister 2016). Reid offers the plea, “This belief, Sir, is none of my manufacture; it came from the mint of nature; it bears her image and superscription; and, if it is not right, the fault is not mine” (Reid [1764] 1997, 169). But the fact that a belief results from our natural constitution is no guarantee that it’s justified, even initially, for we might be naturally constituted in a deficient way. Cognitive scientists regularly expose how “natural” it is for us (in this sense) to engage in motivated reasoning and all other sorts of problematic cognition. My critique here treats naturalness as a descriptive concept—​natural beliefs being something like those that all humans will form in the statistically normal course of human development, barring some developmental obstacle. It is obvious that belief in hinge propositions is natural in this sense, but that counts little in its favor. What if Reid is using naturalness as a normative concept—​natural beliefs being those that result from a properly constituted human intellect? That counts a lot in their favor, but there is no non-​question begging reason to think that brute belief in hinge propositions is natural in this sense. To the contrary, I have made the case that if the belief is formed without any indication of its truth, then that belief is not one that the properly constituted human intellect would form. Despite these deficiencies, there remains something extraordinarily insightful about Reid’s defenses. Consider the possibility of cognitive dispositions so fundamental to the human intellect that, if removed, one would cease to operate as an intellectual agent at all. Let us call these “constitutive principles of rationality.” The beliefs formed in accordance with such constitutive principles would be irresistible in the most formidable sense. They could be avoided only by divesting oneself of one’s intellectual agency altogether. The fact that one cannot resist beliefs in this sense does seem to shield us from criticism on their basis. Moreover, such beliefs are guaranteed to be natural not only in the descriptive but also the normative

204  Against Epistemic Conservatism sense. We know that the dispositions producing them are parts of the properly constituted human intellect because they are parts of any functioning human intellect. If a belief resulted from this kind of principle, it would surely be justified. This is, I think, what Reid is really trying to get at in his arguments—​they are attempts to establish trust in our natural beliefs as a constitutive principle of rationality (Rysiew 2002). This is, I think, the strongest sort of case that could be made for EC*. If the very process of having and giving reasons requires belief in hinge propositions, then the disposition to believe them is a constitutive principle of rationality and we cannot be faulted for believing them. Furthermore, there is a plausible case to be made that belief in hinge propositions is in fact required in order to make inferences or engage in other forms of discursive reasoning. This does not, however, make it a constitutive principle of rationality to believe in hinge propositions without any indication of their truth. That is not required for rational activity. For it may be that such hinge propositions can be justifiably believed on the basis of seemings, as already discussed. Indeed, if we are looking for constitutive principles of rationality, the following seems to be our most plausible candidate: not that beliefs must be given presumption whenever we hold them, but that beliefs must be given presumption whenever they seem true. Starting from this more plausible point, the case might still be made that trusting in how things seem is constitutive of a certain variety of rational activity in that no rational activity of that sort would be possible without it. For what else could constitute evidence for hinge propositions but seemings? How could we even get started in reasoning, much less make any progress, if we did not place a basic trust in how things seemed to us? A full defense of phenomenal conservatism (a restricted version of it) as a constitutive principle of rationality must wait till the next chapter. The point at hand is that phenomenal conservatism once again undercuts motivation for its doxastic counterpart. The long and the short of it is this: epistemic conservatism starts from an implausible claim—​that certain basic beliefs are justified simply because they are believed. This places the burden of proof against it. Some epistemic conservatives have sought to avoid this burden by limiting its scope, but I have argued that they cannot escape it entirely. When it comes down to it, epistemic conservatism posits basic beliefs justified without any evidence for accepting them, and that’s a significant cost to the position. Epistemic conservatives have proposed that we pay this cost in exchange for the theoretical fruit yielded by their position. However, such a defense is undermined by the introduction of phenomenal conservatism, which accomplishes much the same work from a more plausible starting point. After all is said and done, then, epistemic conservatism should be rejected.

Against Epistemic Conservatism  205 This also completes my defense of foundations evidentialism (EV): EV—​S has immediate justification for taking some doxastic attitude toward p only if (and because) that attitude is permitted by S’s total evidence for and against p. Without epistemic conservatism, the only way of explaining immediately justified beliefs is by appealing to foundational evidence. We also saw in Chapter 6 that nothing can serve as foundational evidence expect perhaps seemings. It follows that if there is immediate justification for our beliefs at all, it arises from seemings. In the next two chapters, I make the case that seemings are, in fact, sources of immediate justification. We already surveyed some of the theoretical motivations for this position in the present chapter (see Moretti 2015 for more). In what follows, I provide a more direct argument for its truth. To do so, I return once again to Thomas Reid, whose argument for first principles serves as the inspiration of my own. We will see first (in Chapter 8) that at least some seemings must be capable of immediately justifying belief, then (in Chapter 9) that all of them must be. Notes 1 The weaker versions of epistemic conservatism, since they do not pertain to justification, are not committed to justification apart from evidence (McCain 2021), but the stronger versions are. 2 For this reason, McCain himself presumably would not opt for this solution, as he allows for experiential evidence (McCain and Moretti 2021). 3 Coren (2021) argues that it is practically impossible to evaluate mere beliefs (or what Coren calls “bare beliefs”) since we do not ever encounter them in our actual circumstances. While I am sympathetic to the concern, this reasoning does not, in my judgment, establish that evaluating mere beliefs is impossible; only that it is exceedingly difficult. If we are careful, we might still elicit lessons from imagining a subject in these extreme theoretical conditions. 4 It’s not clear that this is possible. How could a belief be sustained if it neither seemed true nor had any other conscious psychological motivation? Even McCain says, “I am dubious about whether we do in fact form beliefs for which we have no reasons (even bad ones)” (McCain 2008, 198, ft. 33).

References Adler, J.E. 1996. “An Overlooked Argument for Epistemic Conservatism.” Analysis 56 (2): 80–​84. Aristotle. 2004. Nicomachean Ethics. Translated by Roger Crisp. Cambridge Texts in the History of Philosophy. Cambridge: Cambridge University Press.

206  Against Epistemic Conservatism Boulter, Stephen. 2007. The Rediscovery of Common Sense. London: Palgrave Macmillan. Chisholm, Roderick M. 1977. Theory of Knowledge. 2nd edition. Englewood Cliffs, NJ: Prentice-​Hall. —​—​—​. 1980. “A Version of Foundationalism.” Midwest Studies in Philosophy 5 (1): 543–​564. Christensen, David. 1994. “Conservatism in Epistemology.” Noûs 28(1): 69–​89. Coren, Daniel. 2021. “Epistemic Conservatism and Bare Beliefs.” Synthese 198: 743–​756. Feldman, Richard. 2003. Epistemology. Hoboken, NJ: Prentice Hall. Foley, Richard. 1983. “Epistemic Conservatism.” Philosophical Studies 43(2): 165–​182. Goldstick, D. 1971. “Methodological Conservatism.” American Philosophical Quarterly 8 (2): 186–​191. Harman, Gilbert. 1986. Change in View: Principles of Reasoning. Cambridge, MA: MIT Press. Huemer, Michael. 2001. Skepticism and the Veil of Perception. Lanham, MD: Rowman & Littlefield. Lewis, David K. 1972. “Psychophysical and Theoretical Identifications.” Australasian Journal of Philosophy 50 (3): 249–​258. Locke, John. [1689] 1975. An Essay Concerning Human Understanding, edited by P. H. Nidditch. Oxford: Clarendon Press. Lycan, William G. 1988. Judgement and Justification. Cambridge: Cambridge University Press. McAllister, Blake. 2016. “Re-​Evaluating Reid’s Response to Skepticism.” Journal of Scottish Philosophy 14 (3): 317–​339. McCain, Kevin. 2008. “The Virtues of Epistemic Conservatism.” Synthese 164 (2): 185–​200. —​—​—​. 2019. “Epistemic Conservatism and the Basing Relation.” In Well-​Founded Belief: New Essays on the Epistemic Basing Relation, edited by J. Adam Carter and Patrick Bondy, 201–​214. Abington, UK: Routledge. McCain, Kevin, and Luca Moretti. 2021. Appearance and Explanation: Phenomenal Explanationism in Epistemology. Oxford: Oxford University Press. Moore, G.E. [1925] 1959. “A Defence of Common Sense.” In Philosophical Papers. New York: Macmillan. Moretti, Luca. 2015. “Phenomenal Conservatism.” Analysis 75 (2): 296–​309. —​ —​ —​ . 2018. “Phenomenal Conservatism and the Problem of Reflective Awareness.” American Philosophical Quarterly 55 (3): 267–​280. —​—​—​. 2020. Seemings and Epistemic Justification: How Appearances Justify Beliefs. New York: Springer. Poston, Ted. 2012. “Is There an ‘I’ in Epistemology?” Dialectica 66 (4): 517–​541. —​—​—​. 2014. Reason and Explanation. New York: Palgrave Macmillan. Pryor, James. 2000. “The Skeptic and the Dogmatist.” Noûs 34 (4): 517–​549. Quine, W.V.O., and J.S. Ullian. 1970. The Web of Belief. New York: Random House. Rawls, John. 1971. A Theory of Justice: Original Edition. Cambridge, MA: Belknap Press.

Against Epistemic Conservatism  207 Reid, Thomas. [1764] 1997. An Inquiry into the Human Mind on the Principles of Common Sense, edited by Derek R. Brooks. University Park, PA: Pennsylvania State University Press. —​ —​ —​ . [1772] 2002. “To Lord Kames.” In The Correspondence of Thomas Reid, edited by Paul Wood, 64–​68. University Park, PA: Pennsylvania State University Press. Rowe, William L. 1979. “The Problem of Evil and Some Varieties of Atheism.” American Philosophical Quarterly 16 (4): 335–​341. Rysiew, Patrick. 2002. “Reid and Epistemic Naturalism.” Philosophical Quarterly 52 (209): 437–​456. Sklar, Lawrence. 1975. “Methodological Conservatism.” Philosophical Review 84 (3): 374–​400. Swinburne, Richard. 2001. Epistemic Justification. Oxford: Oxford University Press. Vahid, Hamid. 2004. “Varieties of Epistemic Conservatism.” Synthese 141 (1): 97–​122. —​—​—​. 2023. “Doxastic Conservatism.” In Internet Encyclopedia of Philosophy. https://​iep.utm.edu/​doxas​tic-​conse​rvat​ism/​. Wittgenstein, Ludwig. 1969. On Certainty, edited by G. E. M. Anscombe and G. H. Von Wright. New York: Harper Torchbooks.

8 A Reidian Argument for Phenomenal Conservatism, Part I Some Seemings

At the core of Thomas Reid’s common sense response to skepticism lies a simple yet compelling argument. It begins with a relatively modest claim: at least some of the things we naturally believe are immediately justified, absent defeaters. This is to say that the testimony of our faculties can, in at least one instance, be trusted sans any independent verification of its truth or reliability. Seemingly all of Reid’s interlocutors conceded this premise—​even Hume. The cost of denying it appeared to be complete and total skepticism, an implication that prevented one from consistently offering reasons against it. Indeed, Reid argued that to refuse all trust apart from independent verification amounted to an abandonment of rationality itself; for all rational belief formation must proceed from such trust. But, Reid continues, “if they are to be trusted in this instance without a voucher, why not in others?” (Reid [1785] 2002, 481). In other words: if some natural beliefs are immediately justified, absent defeaters, then all of them are. According to Reid, the only basis we have for any of our natural beliefs is that, “every man, while his mind is sound, is determined, by the constitution of his nature, to give implicit belief to it, and to laugh at, or pity the man who doubts its testimony” (Reid [1764] 1997, 17). But the same can be said for any natural belief. Thus, if such grounds sustain justified belief in one case, they do so in the others as well. He concludes: all of our natural beliefs are immediately justified in the absence of defeaters. In drawing this conclusion, Reid is effectively reorienting our conception of rationality. The attempt of our cognitive faculties to grasp truth is not fundamentally a skeptical enterprise but a credulous one: the default is to grant trust until it is lost, not to withhold it until earned. That is to say, rationality operates on a principle of credulity. In Reid’s writings, the term “principle of credulity” refers specifically to a principle of testimony (Reid [1764] 1997, 194)—​to automatically placing trust in the witness of others—​but it is an apt descriptor of Reid’s broader epistemological approach as well. Indeed, as we saw in Chapter 2, Reid considers DOI: 10.4324/9781003281108-9

A Reidian Argument for Phenomenal Conservatism, Part I  209 our natural beliefs to be akin to testimony: “There is a much greater similitude than is commonly imagined between the testimony of nature given by our sense, and the testimony of men given by language” (Reid [1764] 1997, 171). And his stance is that the testimony of our own minds is to be approached with as much credulity as the testimony of others. In the following years, this core insight migrated beyond Reid’s particular epistemological system. The same basic principle of credulity is carried on in internalist versions of moderate foundationalism as propounded by G.E. Moore, Roderick Chisholm, Richard Swinburne, Michael Huemer, and others. Each version of foundationalism, in its own way, urges a basic trust in what our minds testify to be true. I believe this insight receives its most plausible articulation in the version of phenomenal conservatism constituting seemings foundationalism: PC—​If it seems to S that p is true, then (because of that and to that extent) S has foundational evidence for p; and if it seems to S that p is false, then (because of that and to that extent) S has foundational evidence against p. This position extends the principle of credulity to whatever seems to be the case—​we are justified in lending credence to how things appear until there is reason to be suspicious. While this applies both when things seem true and when things seem false, we can simplify the discussion by focusing primarily on seemings-​true and the first half of PC. The considerations offered will apply mutatis mutandis to seemings-​false. My case for PC follows the broad strokes of Reid’s original argument. Of course, the devil is in the details, and my argument will diverge significantly from Reid’s in those details. Nevertheless, enough of the main argument and sub-​arguments find parallels in Reid that it is no misnomer to label it “Reidian.” True to its origins, the Reidian argument for PC is refreshingly simple: some seemings should be given the presumption of credibility; and if some should, all should. More formally: The Reidian Argument for PC Step One

For some seemings, if it seems to S that p is true, then (because of that and to that extent) S has foundational evidence for p. (Minimal-​PC) Step Two If this is true for some seemings, then it is true for all seemings. Conclusion Thus, for all seemings, if it seems to S that p is true, then (because of that and to that extent) S has foundational evidence for p. (PC)

210  A Reidian Argument for Phenomenal Conservatism, Part I Step One is essentially a minimal version of phenomenal conservatism. It says only that some appearances (perhaps even just one) can be approached with credulity. This is significantly more moderate than PC, as there are many who would welcome Minimal-​PC but not its extended cousin. In the present chapter, I offer four lines of support for Minimal-​PC. The first two merely provide evidence that Minimal-​PC is true, but the last two try to illuminate why it is true. These later arguments promise insight into one of the deepest and most fascinating areas of epistemology: the foundations of justification. I defend Step Two in the following chapter. There I consider, and reject, possible reasons why some seemings might be foundational evidence and others not. The very same features that permit trust in one case are present in all of them, leading us to an unrestricted version of phenomenal conservatism. 8.1  Preexisting Support for Minimal-​PC Before examining the arguments, we should first acknowledge just how many foundationalists are already committed to Minimal-​PC, or something very close to it. There are, of course, many epistemologists who explicitly acknowledge that some seemings or appearances can immediately justify belief absent defeaters (e.g., Audi 2013, Bealer 2000, Bedke 2008, Beillard 2016, Bergmann 2013, Berghofer 2020, Brogaard 2013, Chudnoff 2018a; 2018b, Dougherty 2018, Farley 2017, Gage and McAllister 2020, Huemer 2001, Koksvick 2011, Kriegel 2021, Lycan 2013, Markie 2013, McCain and Moretti 2021, McGrath 2013, Moretti 2020, Plantinga 1993, Plug 2011, Pryor 2000, Rogers and Matheson 2011, Reiland 2015, Rosen 2001, Skene 2013, Smithies 2019, Swinburne 2018, Tucker 2010).1 More than this, however, we saw in Chapter 4, §4.2, that many other philosophers today and throughout the Western canon acknowledge the existence of mental states very similar to seemings, and take such states to underlie our basic beliefs. As foundationalists, these same thinkers acknowledge the ability of those states to immediately justify those basic beliefs.2 For example, Aristotle held that beliefs could be immediately justified on the basis of appearances (phantasia), which we saw fit the description of seemings. Indeed, Christopher Shields writes of Aristotle: Aristotle adheres to an evidentiary principle which we may call the Principle of Phainomenological Conservativism (PPC): If it appears (phainetai) to a subject S as if p, then, in the absence of evidence to the contrary, S has grounds for accepting p. (Shields 2013, 8)3

A Reidian Argument for Phenomenal Conservatism, Part I  211 Shields elaborates that “phainomena serve as evidence” for Aristotle (Shields 2013, 9), just as seemings serve as foundational evidence in phenomenal conservatism. Shield’s principle is unrestricted (all appearances serve as evidence), but that is not essential to our purposes here. Others interpret Aristotle as saying that perceptual appearances only impart immediate justification when those appearances accurately represent the object of perception. Even still, Aristotle’s position would entail Minimal-​PC. Much the same is true of the Stoics and Academics, who also acknowledged the existence of phantasia, or impressions. According to the Stoics, the possession of a special kind of cognitive impression—​namely, those “stamped and impressed in exact accordance with what is, and such as could not be from what is not” (Allen 2020, sec. 2)—​entitled one to the certain apprehension of its content. The Academic Skeptic Carneades rejected the need for infallible cognitive impressions, appealing instead to merely probable impressions. These impressions did not guarantee the truth of their content, and so did not license certainty, but nonetheless permitted one to place measured confidence in what appeared to be the case. Though the Stoics and Academics disagree over which seemings justify, both agree that at least some do. Hence, both concede Minimal-​PC. Descartes can be added to the list as well. We previously observed that to clearly and distinctly perceive something—​for its truth to be “evident by the light of nature”—​is plausibly just for that proposition to seem true, albeit in a particularly strong way that is supposed to guarantee its truth. And Descartes seems to have held that when something seems true in this infallible way, one is immediately justified in believing it. He writes, “For whatever is shown to me by this light of nature, for example, that from the fact that I doubt, it follows that I am, and the like, cannot in any way be doubtful” (Descartes [1641] 2006, 21). He adds elsewhere: I am certain I am a thinking thing. But do I not therefore also know what is required for me to be certain of anything? Surely in this first instance of knowledge, there is nothing but a certain clear and distinct perception of what I affirm. (Descartes [1641] 2006, 19) In short, a strong enough seeming permits not only belief but certainty.4 Of course, Descartes does not extend justification to everything that seems true—​some seemings provide only a fallible basis for belief, which he deems insufficient—​but he does grant that immediate justification arises from at least some kinds of seemings. And this is all that Minimal-​PC claims. Consider also contemporary defenders of classical foundationalism, who argue that when one is acquainted with a thought, a fact, and the correspondence between them (a state we called “triple acquaintance”), then

212  A Reidian Argument for Phenomenal Conservatism, Part I one has immediate justification for believing the content of that thought. I argued previously (in Chapter 4, §4.3) that triple acquaintance states are also seemings, and (in Chapter 6, §6.3.1) that it is from those properties that qualify them as seeming that they derive their justificatory powers. In this way, classical foundationalists of this sort should accept Minimal-​ PC. Not all seemings will qualify as foundational evidence on classical foundationalism, but some will. BonJour, for instance, writes: Making a case for the positive part, viz., the claim that certain specific sorts of seemings do yield genuine reasons, is harder and cannot be attempted here; I have argued elsewhere that this is true for at least two specific sorts of seemings, those that involve apprehensions of the contents of one’s own conscious states of mind and those that involve apprehensions of self-​evident necessary truths. (BonJour 2004, 359) Presumably, BonJour focuses on these two kinds of seemings because they are in some sense infallible. Very well—​if only infallible seemings can immediately justify, Minimal-​PC still remains true. Many moderate foundationalists who maintain that perceptual experiences, or memories, or rational intuitions, or even emotions can immediately justify our beliefs will also find themselves already committed to Minimal-​PC. For example, William Alston endorses the following principle: Whatever appears to one as so-​and-​so is thereby likely, in the absence of sufficient indications to the contrary, to be so-​and-​so. (Alston 1999, 198) If Alston’s appearances are characterized as having propositional content that feels true, then they will be seemings and their status as foundational evidence entails Minimal-​PC.5 In sum, a great many foundationalists, both historically and in contemporary circles, are already committed to the position that some seemings are foundational evidence for their content, although they may not use that terminology in formulating their views. Indeed, I would wager that Minimal-​ PC (or something close to it) is a plurality position among internalist foundationalists. Unrestricted versions of phenomenal conservatism like PC are held by only a small minority, but Minimal-​PC enjoys wide acceptance. 8.2  The Threat of Skepticism One of the strongest motivations for Minimal-​PC is the threat of skepticism stemming from its denial. Such arguments take the following general form:

A Reidian Argument for Phenomenal Conservatism, Part I  213 The General Anti-​Skeptical Argument for Minimal-​PC 1 If Minimal-​PC is not true, then skepticism of type X ensues. 2 Skepticism of type X is false. 3 So, Minimal-​PC is true. By “skepticism,” here I simply mean a position on which we do not have justification for certain beliefs that we otherwise firmly believe to be justified. There are many different forms of skepticism so defined, and it is incumbent on the defender of Minimal-​PC to identify a particular type that is both (i) an implication of denying Minimal-​PC and (ii) justifiably believed to be false. In the spirit of Reid, and those like G.E. Moore who follow him, I will take the falsity of skepticism, at least in its most egregious forms, to be so evident that it needn’t be argued for. I have already defended this approach in Chapter 3 as a part of the common sense response to skepticism, but I will add a few thoughts here. As suggested above, the specific kinds of skepticism I have in mind are those which challenge our justification for believing things rather than our knowledge of those things. Excepting propositions such as things appear to me thusly or other “Cartesian propositions” (Unger 1975, ­chapter 3), it is not obvious to me that we do know things. This is because it is not obvious to me that infallibilism is false, and many things we think we know are held fallibly. To dismiss challenges to justification in this way is not to deny the allure of skeptical arguments, nor is it to say that skeptical arguments do not deserve our attention. There is much to be learned by investigating where these arguments go wrong, as well as why they prove so alluring in the first place. What I am saying is that when we approach such arguments, it is fair to do so in confidence that they do go wrong somewhere. What’s more, our justification for the falsity of such skepticisms is strong enough that the positions implying them can be dismissed precisely on the basis of that implication, even if one has not identified precisely where the reasoning leading to such views falls into error: “Such philosophy is justly ridiculous, even to those who cannot detect the fallacy of it” (Reid [1764] 1997, 21). Elsewhere Reid writes: But it is possible, that, by setting out from false principles, or by an error in reasoning, a man may be led to a conclusion that contradicts common sense. … A man of common sense may fairly reject the conclusion, without being able to shew the error of the reasoning that led to it. (Reid [1785] 2002, 434) As noted in the previous chapter, this kind of dismissal has become known as the G.E. Moore Shift. In essence, the falsity of the conclusion is more

214  A Reidian Argument for Phenomenal Conservatism, Part I evident than the truth of the view that led to it, and the view is dismissed on that basis. This maneuver may seem unacceptably dogmatic, and in some situations it is. But the fact that it can be misused makes it no different from any other cogent form of reasoning, the perversion of which is what we call a fallacy. The question to ask is whether it is necessarily dogmatic, and the answer to that is no. To the contrary, the G.E. Moore shift is eminently reasonable in certain conditions. Consider an example from Reid: A traveller who has good judgment may mistake his way, and be led unawares onto a wrong track; and while the road is fair before him, he may go on without suspicion, and be followed by others; but when it ends at a coal-​pit, it requires no great judgment to know that he has gone wrong, nor perhaps to find out what misled him. (Reid [1764] 1997, 23) If anything is dogmatic here, it is to stand by the accuracy of one’s directions in the face of their evident failure. In the same way, if the falsity of the conclusion really is evident, it is not dogmatic to dismiss the position leading to that conclusion; it would in fact be dogmatic not to. Think of it in the following way (Pollock 1974, 4–​5). A valid argument is one in which the premises logically entail the conclusion. All this means, really, is that at least one of the premises must be false or else the conclusion must be true. Which of these one is justified in believing (the truth of the conclusion or the falsity of one of the premises) will depend on how one’s evidence for the premises compares to one’s evidence for the conclusion. If one’s evidence against the conclusion is far superior to one’s evidence for the truth of the premises, then so much the worse for the argument! That is the situation that we find ourselves in with respect to many forms of skepticism. If denying Minimal-​PC leads to such skepticism, then the most reasonable response is to deny that denial. The trick, then, will be showing that the falsity of Minimal-​PC leads to one of these evidently false forms of skepticism. In building such a case, one must first identify the set of ostensibly justified beliefs being threatened. The more cherished the beliefs, the greater the potential of the argument. For instance, the strongest possible skeptical argument would be one that undermined the justification of all of our beliefs: The Total Anti-​Skeptical Argument for Minimal-​PC 1T. If Minimal-​PC is not true, then total skepticism ensues. 2T. Total skepticism is false. 3T. So, Minimal-​PC is true.

A Reidian Argument for Phenomenal Conservatism, Part I  215 I will make the case in this chapter that total skepticism is in fact an implication of denying Minimal-​ PC. That being said, more moderate anti-​skeptical arguments are also available and also perfectly effective in establishing Minimal-​PC. For instance, consider an argument based around common sense beliefs: The Common Sense Anti-​Skeptical Argument for Minimal-​PC 1C. If Minimal-​PC is not true, then skepticism about certain common sense beliefs ensues. 2C. Skepticism about those commonsense beliefs is false. 3C. So, Minimal-​PC is true. What common sense beliefs do I have in mind? Take, for example, some of the beliefs put forward by G.E. Moore in his famous paper, “A Defence of Common Sense”: “that there exists at present a living human body, which is my body”; “the earth had existed also for many years before my body was born”; “I have had dreams”; and “I have had feelings of many different kinds” (Moore [1925] 1959). If the justification of such beliefs is jeopardized by denying Minimal-​PC, then we still have an extremely formidable argument. 2C may not be quite as plausible as 2T, but it isn’t far behind. Another option is to focus on a highly specific kind of common sense belief—​for instance, introspective beliefs: The Introspective Anti-​Skeptical Argument for Minimal-​PC 1I. If Minimal-​PC is not true, then skepticism about introspective beliefs ensues. 2I. Skepticism about introspective beliefs is false. 3I. So, Minimal-​PC is true. If denying Minimal-​PC has the implication that one doesn’t have justification for believing that one is in pain (in certain familiar and paradigmatic conditions), then that is in itself an extremely strong argument in favor of Minimal-​PC. Once the relevant set of beliefs is identified, one must then show that denying Minimal-​PC entails, or is likely to entail, skepticism about such beliefs. That is, one must support premise 1 (or 1T, 1C, 1I, etc.). There are at least two ways to establish this premise. The first way, already present in the literature, is the strategy employed by Huemer in his defense of phenomenal conservatism. On this tack, one shows that the selected beliefs are, as a matter of psychological fact, ultimately based on seemings. It follows that those beliefs are not doxastically justified unless the seemings

216  A Reidian Argument for Phenomenal Conservatism, Part I on which they are based provide immediate propositional justification for their contents. We can call this the “psychological approach” to defending premise 1. The second way of defending premise 1 is by arguing that there cannot be any ultimate source of propositional justification for some set of cherished beliefs apart from seemings. Let’s call this the “principled approach” to defending premise 1. We will discuss each of these in turn, beginning with the latter. 8.2.1  The Principled Approach

The principled defense of the anti-​skeptical argument does not depend on any empirical claims about the psychological bases for our beliefs. On this approach we simply look, from a theoretical standpoint, at what kind of state could serve as an adequate basis for our justified basic beliefs. What other source of immediate justification is there besides seemings? If we conclude that nothing else fits the bill, then we have available an anti-​skeptical argument for Minimal-​PC. Indeed, we have the strongest anti-​skeptical argument possible! Given foundationalism, there must be some source of immediate justification if anything is to be justified at all. Thus, the principled approach licenses the Total Anti-​Skeptical Argument (and thus all others as well). It is at this point that I appeal to the other tenets of seemings foundationalism. In Chapter 5, I argued that foundationalism (FN) is true and, hence, that all justification comes down to immediate justification. In Chapters 6 and 7, I defended seemings exclusivism (SE)—​that only seemings can serve as foundational evidence—​and foundational evidentialism (EV)—​ that immediate justification is a function of our foundational evidence. In short, no justification without immediate justification, and no immediate justification apart from seemings. The result of denying Minimal-​PC is thus total and complete skepticism. If I am right about this, then it is conclusive evidence for Minimal-​ PC. That being said, I obviously haven’t given a conclusive case for the above three theses: SE, EV, and FN. What’s more, the plausibility of these three theses will to some degree depend on the independent plausibility of Minimal-​PC. This is most evident with respect to EV and FN. My argument against epistemic conservatism, and thus for EV, appealed to the preferability of seemings as an alternative source of justification for hinge propositions, and my case for FN rested on the premise that seemings could plausibly overcome BonJour’s dilemma. The same is also true in less direct fashion for SE. My argument for that position, in conjunction with EV, leads to total skepticism unless some seemings are in fact sources of immediate justification. If we do not take seemings to be plausible candidates for this role, then the most reasonable reaction would be to reject that

A Reidian Argument for Phenomenal Conservatism, Part I  217 argument using the same kind of G.E. Moore shift defended above. Thus, it behooves me to offer additional, independent reasons for endorsing Minimal-​PC, including independent ways of motivating the anti-​skeptical argument. 8.2.2  The Psychological Approach

The psychological approach, unlike the principled approach, does not deny a source of immediate propositional justification apart from seemings. Instead, it claims that humans fail to take advantage of any alternative, whatever it may be, with respect to some important set of beliefs—​be it all of our beliefs, or certain common sense beliefs, or introspective beliefs. Those beliefs, it is argued, are ultimately based on seemings. Thus, if the seemings on which such beliefs are based do not provide immediate propositional justification for their content, then the basic beliefs based on them are not doxastically justified, in which case all of the beliefs based on those basic beliefs are not justified, and so on. Four clarifications are helpful here. First, the claim that certain beliefs are ultimately based on seemings is a claim about the folk-​psychological causes of the belief. Which mental states bring about the belief in question on the folk-​psychological level of explanation?6 Being the folk-​psychological cause of a belief is not sufficient to qualify as the “base” of that belief, but it is necessary. Thus, to claim that seemings are the ultimate bases of a belief entails that the causal chain resulting in that belief (on this personal level of explanation) can be traced back to one or more seemings. Second, a belief can be ultimately based on seemings even if not directly based on seemings. If a belief is directly based on a seeming, then that seeming must be the most proximate cause within a fully articulated folk-​psychological explanation of the belief. However, a belief is still ultimately based on seemings if it is based on beliefs that are directly based on seemings somewhere down the line. Third, the focus here is on justified beliefs (or at least those that we take to be justified) since those are the ones relevant to the anti-​skeptical arguments. Thus, it is no objection that humans have all sorts of pernicious motivations for belief apart from seemings (ones such fear, prejudice, wishful thinking, etc.), since these are obviously not justified. Finally, the claim here is not that it is impossible for justified beliefs to be formed apart from seemings, but only that, as a matter of human psychology, this is how our justified beliefs are in fact formed (Huemer 2007, 40). As one reflects on our justified beliefs, one quickly notices that many if not most of them are based on seemings. We will survey the four main sources of justified beliefs below—​introspection, reason, perception, and memory—​and show how the beliefs from each of these sources are ultimately based on seemings. For the moment, however, let us consider more

218  A Reidian Argument for Phenomenal Conservatism, Part I general reasons for thinking that seemings underlie a great many, or perhaps even all, of our justified beliefs. Consider the act of deception. If your intent is to deceive someone who is proceeding justifiably, how do you do it? By making things seem to be one way when they really are another. And you do this by making the proposition in question seem true, or else by making other things seem true that will eventually lead to the desired belief. Notice that this is tacitly assumed by epistemologists when discussing skeptical scenarios. How does the evil demon or the brain-​in-​vatting scientist deceive his victim? By making things appear true that are not (Huemer 2011, 12). The universal allowance is that changing how things appear will change how the subject believes, at least insofar as they are proceeding in a normal and potentially justifiable manner. Hence, those beliefs must be based (at least in large part) on how things appear. From a general point of view, this makes perfect sense. If a proposition did not seem true, and was not supported by things that seemed true, then why would a rational person believe it? To be a bit more systematic, note that inferred beliefs are all ultimately based basic beliefs (or other basic attitudes). We can then ask ourselves which kinds of mental states might directly bring about basic beliefs in a manner that does not make the belief obviously unjustified? Such states cannot lack propositional content because they cannot bring about belief in a specific proposition without relying on background evidence—​at least not in a manner that stands any chance of being justified—​in which case the resulting belief will not be basic (see Chapter 6, §6.2). Merely possessing propositional content is not enough, however. When we imagine that p we have a state with propositional content, but that state has no psychological tendency to bring about belief in p in and of itself. And if it somehow did, it would be obviously unjustified. Rather, the mental state must present its propositional content to us in a manner that draws us to believe it. One way to draw out belief is for a mental state to present its content as good or desirable, but the resulting beliefs are not good candidates for justification (see Chapter 6, §6.3.3). What else could directly cause our justified basic beliefs then except states which present their propositional content as true? No good alternatives come to mind. It is possible, I suppose, that our basic beliefs have no rational basis at all, and are instead brutely caused; but reflection on our own basic belief-​ forming processes does not bear this out as the norm. Our basic beliefs are usually caused by experience—​be it a perception, a memory, an intuition, etc. The case might be made that belief in hinge propositions is simply hardwired into us, but even if that is granted (and I do not think that it should be; see Chapter 7, §7.3), it would remain true that the vast majority of beliefs formed in the normal course of human life would be ultimately

A Reidian Argument for Phenomenal Conservatism, Part I  219 based on experience. And the above reasoning suggests that seemings are the only experiences that fit the bill. We can bolster the case by looking at representative examples of how basic beliefs are formed in perception, memory, introspection, and rational intuition. If I can show that the basic beliefs resulting from even just one or two of these sources are based on seemings, that will furnish us with a formidable anti-​skeptical argument. If all of these sources rely on seemings, then we may even be able to run the Total Anti-​Skeptical Argument, since these are plausibly the only fundamental sources of basic beliefs.7 8.2.2.1  Introspective Beliefs

Let us start with basic beliefs formed via introspection: Introspection—​Jarvis has a headache while reviewing a contract at work. He finally notices the pain and attends to it. He forms the thought I am in pain and attends to the correspondence between that thought and the pain with which he is acquainted. He forms the belief that he is in pain. Here Jarvis’s belief is based on the state of triple acquaintance (acquaintance with a fact, a thought, and the correspondence between them). Mere acquaintance (acquaintance with the fact alone) does not provide an adequate psychological basis for belief in this case or in cases of introspection more generally. This should already be apparent from our discussion of mere acquaintance in Chapter 6, §6.2.3. For instance, we saw there that acquaintance with a 39-​speckled hen will not prompt the belief that there is a 39-​speckled hen unless the subject appreciates the connection between the thought there is a 39-​speckled hen and the fact with which one is acquainted. Thus, it is the more complex state of being triple acquainted that occasions belief. Another lesson from earlier chapters (Chapter 4, §4.3.3 and Chapter 6, §6.3.1) is that token triple acquaintance states are also seeming states. Hence, Jarvis’s belief, if it is based on a state of triple acquaintance, is thereby also based on the introspective seeming that he is in pain. If we go in this direction, we must then ask ourselves what is it about that token mental state that is doing the real psychological work in basing Jarvis’s belief. Once again, I have already argued in Chapter 6, §6.3.1, that it isn’t any of the factive properties definitive of triple acquaintance states, but rather those properties characteristic of seemings: namely, having propositional content and a forceful phenomenal character. This is not to say that the token triple acquaintance state doesn’t base Jarvis’s belief—​I can concede that it does—​but it does not do so because it is a triple acquaintance state. It does so because it is a seeming.

220  A Reidian Argument for Phenomenal Conservatism, Part I To drive the point home, consider a false introspective belief: Faulty Introspection—​Jarvis has a mosquito bite on his leg. The persistent itchy feeling around the bite suddenly spikes in a sharp tickling sensation that grabs Jarvis’s attention, then fades away. During the spike, Jarvis has the thought I’m in pain, and attends to (something similar to) the correspondence between this thought and his sensation. He forms the belief that he is in pain. However, the sensation was not actually painful, just poignantly itchy. In this case, Jarvis is not triple acquainted with his being in pain (since he’s not actually in pain), yet he still forms the belief. What this suggests is that so long as it feels like one is triple acquainted with something, that will be sufficient to bring about belief. Returning to the veridical case, then, it seems to make no psychological difference that Jarvis is actually triple acquainted with his being in pain. The determining factor—​ and the most fundamental basis for his belief—​is that it feels like he is triple acquainted with his pain, and those are the features definitive of seemings. At this point, we already have the grounds for the formidable Introspective Anti-​Skeptical Argument. If Minimal-​PC is not true, then our introspective beliefs would not be justified. But they are. So, Minimal-​PC is true. Let us see, however, if we cannot strengthen the overall case. 8.2.2.2  A Priori Beliefs

Consider a representative example of how basic a priori beliefs are formed via intuition: Intuition—​Norm initially holds no attitude concerning the truth of a proposition of the form ((p ⊃ q) & p) ⊃ q. Say, if it’s true both that it’s wet outside if it’s raining, and that it’s raining, then it follows that it’s wet outside. He then considers this proposition and it strongly seems true. He forms the belief that this principle is true. One may dispute whether the particular belief I have selected is genuinely basic. No matter; if it isn’t basic, then it must be inferred from other a priori beliefs that are basic, and those (if they stand any chance at being justified) will be believed in the way I have described above—​namely, because their content seems true. What else could basic a priori beliefs be based on? Acquaintance is perhaps the only alternative, but all the considerations of the last section apply here as well. To summarize, acquaintance with the fact alone does not

A Reidian Argument for Phenomenal Conservatism, Part I  221 provide an adequate psychological basis for belief, and triple acquaintance constitutes a seeming. Moreover, the properties of that token triple acquaintance state that actually trigger belief are those that qualify it as a seeming, not those which set it apart as a triple acquaintance state. In particular, the factiveness of the triple acquaintance state doesn’t seem to matter psychologically. We can observe this once again by looking at a priori beliefs based on faulty intuitions: Faulty Intuition—​Norm initially holds no attitude concerning a proposition of the form ((p ⊃ q) & ~p) ⊃ ~q. Say, if it’s true both that it’s wet outside if it’s raining, and that it’s not raining, then it follows that it’s not wet outside. He then considers this proposition and it strongly seems true. He forms the belief that this principle is true. His belief is false. The fact that the content of his intuition is false makes no psychological difference to whether Norm believes that content—​belief is formed so long as that content seems true. And if this is true in the case of Faulty Intuition, it is true in the veridical case as well. Combining this with the previous section, we now have an even more formidable anti-​skeptical case for Minimal-​PC. Without introspection and reason, a great many of our beliefs about the world would not be justified. Again, stopping here still leaves us with an exceedingly formidable argument for Minimal-​PC. Nonetheless, let’s see just how low we sink into the skeptical bog without seemings. 8.2.2.3  Perceptual Beliefs

We turn now to a perceptual example which, I take it, will represent how we form most of our perceptual beliefs more generally: Perception—​Wallace smells bacon in the kitchen. As bacon particles waft from the kitchen into his nose, triggering various olfactory responders, they produce in Wallace a familiar array of sensations. It then seems to Wallace that there is bacon in the kitchen. He believes that there is bacon in the kitchen. He gets off the couch to get some. Quite plausibly, Wallace’s belief is based directly on his perceptual seeming. Obviously, then, that belief is doxastically justified only if Minimal-​PC is true. Is there any way to resist this reading of the situation? Dissenters might argue that Wallace’s belief is based on some factive state such as perceiving the bacon, but I have already responded to this sort of position (in Chapter 1, §1.4.3). There we saw, following Huemer (2007), that factive states like perceiving bacon are screened off by appearances.

222  A Reidian Argument for Phenomenal Conservatism, Part I So long as everything feels the same on the inside, Wallace will believe that there is bacon whether he genuinely perceives this or not. For instance: Faulty Perception—​Something other than bacon particles trigger Wallace’s olfactory responders, producing in Wallace the array of sensations typically produced by the presence of bacon. (Perhaps there is a bacon candle in the kitchen or, more dramatically, Dr. Snifter, mad scientist, has found a way to stimulate Wallace’s senses through some dastardly machine.) It then seems to Wallace that there is bacon in the kitchen. Wallace believes that there is bacon in the kitchen. He gets off the couch to get some. The fact that Wallace’s perceptual states are not factive makes no psychological difference to what he believes. What this shows is that, even in the veridical case, Wallace’s belief is not brought about by his perception because it is a perception, but rather because of properties characteristic of a perceptual seeming. Another suggestion is that Wallace’s perceptual belief is not caused by a perceptual seeming but by his sensations, or what I have called sensory appearances. I have already mentioned several reasons for thinking that sensory appearances are non-​conceptual (see Chapter 4, §4.3); however, let us assume for just a moment that they have propositional content. If they present that content as true, then they will be species of seemings (which only proves my point); and if they have propositional content but don’t present their content as true, then they won’t prompt belief. We considered what this situation might be like back in Chapter 6, §6.3.2. I compared it to immersing yourself in a virtual reality gaming system that you know isn’t real. The sensations buzz and hum along, but it doesn’t feel as though they are representative of the way things really are. Such sensory states, despite their conceptual content, will not prompt belief unless (i) their content starts seeming true or (ii) one relies background beliefs (along with an introspective belief about the existence of one’s sensory appearances) to discern something about their significance. And such background beliefs will at least partially consist in a priori and introspective beliefs that are, we have already seen, based on seemings. If sensations are non-​conceptual, however, then the olfactory sensations alone will not bring about Wallace’s belief that there is bacon present. Those sensations will prompt belief only if Wallace has background evidence concerning what bacon smells like. But such background evidence will once against consist at least partially in beliefs based directly on seemings. In either case, then, seemings still feature among the ultimate bases of our perceptual beliefs. If no seemings are foundational evidence, then our perceptual beliefs won’t be justified.

A Reidian Argument for Phenomenal Conservatism, Part I  223 8.2.2.4  Memorial Beliefs

Let us finish our survey with a look at memory. In particular, we will focus on forming new beliefs through memory, as opposed to merely accessing a belief previously stored in memory. Our interest is in how beliefs are formed, not retained. For example: Memory—​Abilene initially has no attitude concerning whether she paid her rent this month. She then considers the matter and it seems to her that she did pay her rent this month. Abilene forms the belief that she paid her rent this month. This case might serve as an example of either semantic or episodic memory. On the semantic interpretation, Abilene may simply have a seeming with the content I already paid my rent this month without any accompanying quasi-​ sensory phenomenology. On the episodic interpretation, Abilene may experience something like a faint phenomenal array in which it feels like she is writing out a check and slipping it under her landlord’s door. This prompts the memorial seeming that she already paid her rent, which then leads to belief. In either case, Abilene’s belief seems to be based directly on her memorial seeming. Some will argue that the story is more complicated. As with the perceptual case, one may insist that Abilene’s memorial belief is not basic but instead based on a framework of background beliefs, including beliefs about the reliability of one’s memory. As you might predict, my response here is the same as in the perceptual case: those background beliefs themselves will be based on seemings, implying that in one way or another, Abilene’s memorial belief is ultimately based on how things seem to her. Finally, lest someone claim that some factive state of remembering is what truly explains Abilene’s belief, we can simply amend the case as follows: Faulty Memory—​Abilene initially holds no attitude concerning whether she paid her rent this month. She then considers the matter and it seems to her that she did pay her rent this month. Abilene forms the belief that she paid her rent this month. She did not pay her rent this month. The properties responsible for triggering Abilene’s belief are going to be the same in both cases, in which case she cannot believe precisely because she is in any factive state. We are now in a position to conclude that our introspective, a priori, perceptual, and memorial beliefs are all ultimately based on seemings and, thus, are not doxastically justified unless those seemings count

224  A Reidian Argument for Phenomenal Conservatism, Part I as foundational evidence. This entitles us to run at least the Common Sense Anti-​Skeptical Argument for Minimal-​PC, and perhaps even the Total Anti-​Skeptical Argument. Is the case definitive? Of course not. Philosophical arguments rarely are. The more important question is whether the above considerations make it likely that denying Minimal-​ PC comes with serious skeptical consequences, and I maintain that they do—​quite likely, in fact. When you look at the things that actually serve as the causes of our beliefs (the beliefs that we take to be justified, that is) time and time again they trace their way back to how things seem. If those seemings aren’t sources of immediate justification, then we’re is serious trouble. Returning to the broader discussion, I have presented multiple arguments that none of our beliefs would be justified were Minimal-​PC not true. For there is no source of immediate justification apart from seemings (the​ principled approach), and even if there is, we do not take advantage of it in forming our beliefs (the psychological approach). If my arguments fail to establish this totalizing conclusion, they may still establish that a great many of our justified beliefs would not be so justified were Minimal-​PC not true. Falling short of that, they may suffice to show that one important kind of justified belief (such as introspective beliefs) would not be justified were Minimal-​PC not true. In any of these cases, we have strong reason to endorse Minimal-​PC. 8.3  The Self-​Defeat Argument for Minimal-​PC Naturally, the skepticism that ensues from denying Minimal-​PC threatens our epistemological beliefs as well, including the denial of Minimal-​PC itself. This takes us straight to Huemer’s well-​known self-​defeat argument (Huemer 2001; 2007; 2011). Huemer argues that any rejection of phenomenal conservatism is self-​defeating because it will ultimately be based on how things seem to the objector. Hence, if the objector is correct and phenomenal conservatism is false, the objector’s disbelief would itself be doxastically unjustified, since it would then be based on seeming states that do not provide propositional justification for the content of that belief. In short, their disbelief only stands a chance of being justified if it is mistaken. We can structure the argument as follows: The Psychological Self-​Defeat Argument for PC 1. Any disbelief in PC (that is a plausible candidate for being justified) will ultimately be based on how things seem to the subject. 2. A disbelief that is ultimately based on how things seem to the subject is doxastically justified only if the seemings on which it is ultimately based are foundational evidence for what seems true.

A Reidian Argument for Phenomenal Conservatism, Part I  225 3. The seemings on which that disbelief is ultimately based are foundational evidence for what seems true only if PC is true. 4. Thus, any disbelief in PC is doxastically justified only if PC is true. In this way, denying phenomenal conservatism is supposedly like trying to argue that none of your faculties can be rationally trusted—​the argument only succeeds if the conclusion is false. Indeed, the very act of formulating that argument requires that one proceed as if the conclusion is false. Unsurprisingly, Huemer’s argument is heavily contested. Some object to premise 1, suggesting that other mental states provide an alternative mental basis for disbelief. In arguing that all or most of our beliefs are ultimately based on seemings, I have already responded to such objections. For instance, one prominent objection of this sort is the claim that our beliefs might be based on direct acquaintance rather than seemings (DePoe 2011, Hasan 2013, Tooley 2013). We have already seen that acquaintance with a fact will not prompt belief in any particular proposition without reliance on background beliefs—​background beliefs which will ultimately be based on seemings—​or without also being acquainted with a thought and its correspondence to that fact. But, concerning the latter, triple acquaintance states are also seemings states, and it is by virtue of their membership in the latter category that they base belief. Thus, in either case, acquaintance does not provide a way out of self-​defeat. In my judgment, the strongest objections to the self-​defeat argument are leveled at premise 3 (DePaul 2009, Markie 2013, Tooley 2013). There are, recall, restricted versions of phenomenal conservatism which grant the implicit trust to some seemings but not others. For instance, a classical foundationalist might grant introspective and rational seemings status as foundational evidence while denying it to perceptual or memorial seemings. One of these more moderate principles might be sufficient to justify disbelief in the unrestricted PC. Thus, the argument fails to establish PC in its unrestricted form. The argument does, however, constitute a formidable defense of Minimal-​PC. Remember that at this point, I am only defending the claim that some seemings are foundational evidence for their content; extending this to all seemings is the task of the final chapter. Consider then a revised argument: The Psychological Self-​Defeat Argument for Minimal-​PC 1*. Any disbelief in Minimal-​PC (that is a plausible candidate for being justified) will be ultimately based on how things seem to the subject. 2*. A disbelief that is based on how things seem to the subject is doxastically justified only if the seemings on which it is ultimately based are foundational evidence for what seems true.

226  A Reidian Argument for Phenomenal Conservatism, Part I 3*. The seemings on which that disbelief is ultimately based are foundational evidence for what seems true only if Minimal-​PC is true. 4*. Thus, any disbelief in Minimal-​ PC is doxastically justified only if Minimal-​PC is true. Premise 3* is uncontestable in this revised argument. And I have already made the case for premise 1* in the previous section—​if all or most of our beliefs are based on seemings, then this is true of our epistemological beliefs as well. In fact, epistemological beliefs seem to be based largely on introspection and a priori intuition, and there is an especially strong case that beliefs of those sorts are ultimately based on seemings (even if certain others aren’t). Thus, the famous self-​defeat argument turns out to have some oomph after all, so long as it is directed toward Minimal-​PC. Furthermore, we needn’t rely on the psychological approach embodied in Huemer’s original self-​ defeat argument. In the previous section, I introduced an alternative, in-​principle approach that applies here as well. This approach does not require us to speculate about the psychological bases of disbelief in Minimal-​PC; rather, it argues on the basis of seemings exclusivism that there is no ultimate source of justification other than seemings. Or, more moderately, there is no source of justification for epistemological beliefs other than seemings. More formally: The In-​Principle Self-​Defeat Argument for Minimal-​PC 1’. The negation of Minimal-​PC is justified only if there is foundational evidence supporting that proposition. 2’. If there is foundational evidence supporting that proposition, then (it is because) Minimal-​PC is true. 3’. Thus, the negation of Minimal-​PC is justified only if (and because) Minimal-​PC is true. Even if one or both of these self-​defeat arguments are sound, it is not entirely easy to say how they support Minimal-​PC. For one, their plausibility rises and falls with the anti-​skeptical arguments of the previous section, so it is unclear whether they provide any support independent from the case already present there. For another, the self-​defeat arguments show that a justified denial of Minimal-​PC is in some sense reflectively unsustainable, but that does not entail that Minimal-​PC is true. So, what unique support, if any, do self-​defeat arguments provide for Minimal-​PC and how? Reid once again shows the way. In addressing skepticism and its attack on common sense, Reid observes that any philosophical attempt to impugn our natural credulity must itself proceed from that same credulity. As Reid puts it, “philosophy’s only root is the principles of common sense; it grows

A Reidian Argument for Phenomenal Conservatism, Part I  227 out of them, and draws its nourishment from them” (Reid [1764] 1997, 19). The result is that arguments for total skepticism can have no merit unless their conclusions are false. But if such arguments are either baseless or misleading, then “Sensible men,” says Reid, will dismiss them out of hand, saying: “He must either be a fool, or want to make a fool of me, that would reason me out of my reason and senses” (Reid [1764] 1997, 24). Reid follows up by writing: I confess I know not what a sceptic can answer to this, nor by what good argument he can plead even for a hearing; for either his reasoning is sophistry, and so deserves contempt; or there is no truth in the human faculties, and then why should we reason? (Reid [1764] 1997, 24)8 This insight places boundaries on how far skepticism can reach. No skeptical argument, no matter how seductive, can overturn all the dictates of common sense: “In this unequal contest between common sense and philosophy the latter will always come off with both dishonour and loss” (Reid [1764] 1997, 19). Because philosophical argumentation is grounded on common sense first principles, the former simply cannot rationally undermine the latter. Reid explains it this way: It is a bold philosophy that unceremoniously rejects principles which irresistibly govern the belief and the conduct of all mankind in the common affairs of life—​principles to which the philosopher himself must surrender after he imagines he has refuted them. Such principles are older than philosophy, and have more authority than she does; she is based on them, not they on her. If she could overturn them, she would inevitably be buried in their ruins; but all the siege-​machines that philosophical subtlety can create are too weak for this purpose; and the attempt is just as ridiculous as it would be for a mechanic to construct a windlass for winching the earth out of its circuit, or for a mathematician to claim he could demonstrate that things equal to the same thing are not equal to one another. (Reid [1764] 1997, 21) This realization positions us to reject skeptical arguments from the get-​go, even if we do not detect any fallacy within them. To the extent that we are aware of what is going on, the arguments of the skeptic are, as Reid says, “ridiculous” and are more suitably greeted by laughter, or even ridicule (Reid [1785] 2002, 462), than sober concern. Applying this all to Minimal-​ PC, the self-​ defeat argument shows that the denial of Minimal-​PC ultimately derives its justification from

228  A Reidian Argument for Phenomenal Conservatism, Part I seemings. If this is correct (and I think I’ve shown that it likely is correct), then any attempt to argue against Minimal-​ PC is, like the skeptic’s attempt to overturn common sense, doomed from the get-​go. If we are aware of this, we are likewise positioned to reject those arguments out of hand, even if we do not know exactly where they go wrong. For we can see up front that they are incapable of sustaining any reasonable resistance to Minimal-​PC. To be fair, the preemptive rejection of these arguments only extends as far as one’s support for the self-​defeat arguments, which is far from absolute. A convincing enough rebuttal of Minimal-​PC may compel us to revisit our endorsement of those arguments. They make Minimal-​PC resilient, not indefeasible. Still, I have not seen any argument against Minimal-​PC that meets this standard. The most convincing arguments against PC more generally don’t impugn Minimal-​PC which is, well, minimal in comparison. Are we really to maintain that one is never justified in trusting how things seem prior to independent verification? Perhaps one day someone will provide a convincing enough case, but they have their work cut out for them. 8.4  Forcefulness and the Pursuit of Truth Phenomenal conservatism is intuitive, and Minimal-​PC especially so. It says that one is sometimes (just sometimes) permitted to trust how things appear, if only when that thing is completely and utterly obvious, like p ⊃ p. This is liable to strike many as true, or even obviously true. Intuitions like these help explain why phenomenal conservatism has retained its allure despite difficulties formulating convincing arguments for its truth. While many share these intuitions, others need a bit more prompting. Here we will dig into Minimal-​PC directly to see if we can’t make its truth even more widely apparent. The intuitiveness of phenomenal conservatism (which entails Minimal-​ PC) is sometimes supported in the following way: Suppose it seems to you that P and you have no defeaters (i.e., no good evidence for not-​P and no good evidence that this seeming is unreliable as to whether P). Which doxastic attitudes would it be reasonable for you to have toward P? Disbelieve P, without good evidence for not-​P? Withhold judgment on P? It does seem to you that P, and you lack evidence for not-​P and for the unreliability of the seeming with respect to P? The only reasonable attitude to take is belief. Even if we do not find this answer clearly correct, let alone self-​evident, it does have intuitive plausibility. Its seeming to one that P seems to be at least some evidence weighing in favor of one’s believing P. (McGrath 2013, 226)

A Reidian Argument for Phenomenal Conservatism, Part I  229 This appeal has some merit, but it is not formed as strongly as it might be. It is notably negative—​neither disbelief nor withholding assent make any sense, belief fares somewhat better, thus, if we must accept one as permissible, belief is the way to go. It is telling that Tucker (forthcoming) calls this the “Bad to Resist Belief Argument.” To make matters worse, focusing on the impermissibility of the alternatives rather than the permissibility of belief leaves it open to criticism. Tucker (2013, 9–​12; forthcoming, §5), building off Jackson (2011), raises the possibility that none of the three options is permissible. Unless we rule out the possibility of such epistemic dead-​ends, Tucker claims, we cannot be sure whether belief is genuinely permitted or else an impermissible end to which we are in some sense rationally committed. The other sections of this chapter provide me with the resources to respond to Tucker, at least when construed as undermining Minimal-​PC.9 For in order to remove support for Minimal-​PC, as opposed to unrestricted PC, one must propose that all seemings result in epistemic dead-​ends, not only that some do. But if all seemings result in dead-​ends, then that leads to skepticism (see §8.2) and self-​defeat (see §8.3). Worse yet, rationality itself is defunct (see §8.5). Thus, the possibility Tucker raises cannot apply to all cases. However, my intention in this section is to develop an additional line of support for Minimal-​PC, one that is independent of those other arguments. Presently, I will address the aforementioned vulnerabilities by making a positive case for the permissibility of belief, not just a negative case via the impermissibility of the alternatives. Before making my intuitive appeal, let’s remind ourselves precisely what Minimal-​PC is claiming. On the framework laid out in Chapter 2, foundational evidence for p is a mental state capable of permitting belief (or positive credence) in p in a state of empty symmetrical evidence. Such a mental state must fill dual psychological and normative roles. Psychologically, it must be the sort of thing that can cause belief in a state of empty symmetrical evidence. This means its ability to prompt belief cannot depend on having background evidence or beliefs; rather something about its mere presence must, in and of itself, trigger our automatic belief-​forming systems. Normatively, this mental state must also be the sort of thing that, without any background evidence, brings belief into conformance with our epistemic duties in the pursuit of attaining truth and avoiding falsehood. With this framework in mind, the question at hand is whether seemings fill the role of foundational evidence. I submit that they can and do. Seemings have a forceful phenomenal character. This character is aptly nicknamed “the feel of truth” because it is the kind one enjoys when acquainted with the correspondence between some proposition and the reality it describes. Seemings easily fill the psychological role of foundational evidence by virtue of this phenomenal

230  A Reidian Argument for Phenomenal Conservatism, Part I character. If something feels true, then that naturally inclines one to believe it (or at least to become more confident in it) even if one lacks any background evidence. What’s more, forcefulness also positions seemings to fill the normative role of foundational evidence. The feel of truth seems like precisely the sort of thing that could bring belief into line with our epistemic duties. After all, one is literally being drawn to believe the proposition by its felt truth. If our duty to pursue truth does not allow us to believe in these circumstances, then when does it? As Huemer questions, “What more are you looking for? This is as good as it gets” (Huemer 2001, 105). The more pressing question is whether there is any phenomenal character that could serve as a more respectable, or even equally respectable, basis for belief. Try to think of some other mental state that could permit belief in a state of empty symmetrical evidence. If its propositional content feels true, then it’s a species of seeming. And if its content doesn’t feel true, then are we supposed to believe that a state whose content doesn’t feel true can permit belief, while a state whose content does feel true cannot? For most of our discussion, I have been talking exclusively about the justification of belief rather than positive credences, but at this point the difference becomes especially relevant. Believing something plausibly requires more than simply placing a positive credence in it. For instance, my credence that some U.S. quarter will land heads when flipped is slightly above .5 (the additional weight on that side of the coin makes it slightly more likely), but I do not believe that the quarter will land on heads. All that is required of foundational evidence for p, however, is that it permit a credence over .5 in p in the state of empty symmetrical evidence. This makes it even harder to deny seemings that status. Surely some seemings, if only the most evident, permit one to become slightly more confident in their content than not. To deny this, one has to claim that even the strongest seemings count for literally nothing in the state of empty symmetrical evidence—​we are not permitted to be moved by them in the slightest. But that seems deeply implausible. To be clear, I think some especially forceful seemings are sufficient to permit full belief in the state of empty symmetrical evidence. But it is worth noting that Minimal-​PC is compatible with the more moderate claim that, at most, some seemings are sufficient to motivate positive credences falling short of belief. For at that point, one has already conceded that seemings are foundational evidence; we’re just debating how strong. That being said, I will continue to use the language of belief because I wish to defend the stronger claim if possible—​namely, that some seemings are capable of immediately justifying full belief. What objections might there be? A predictable worry is that belief is too risky. After all, in a state of empty symmetrical evidence one has zero verification of that seeming’s reliability. One could argue that withholding

A Reidian Argument for Phenomenal Conservatism, Part I  231 assent, seemingly the safe option, is what is called for by our epistemic duties, particularly with respect to the goal of avoiding falsehood. The error here is thinking that there is a safe option at all. Whatever stance one takes carries with it some risk. This is because the epistemic goal has two parts, which pull us in opposite directions. We want to avoid being duped, sure, but we also want to avoid missing out on the truth. Avoiding falsehood in itself is easy: just withhold assent from everything and you will never believe something false. Attaining truth is easy as well: just believe everything and you will never fail to believe what is true. It is finding a proper balance between these competing goals that is challenging. You cannot be so credulous that you become gullible, jumping to belief without sufficient indication of truth; but neither can you become so suspicious that you become skeptical, refusing to follow the signs of truth when they present themselves. There’s nothing you can do that doesn’t run the risk of falling afoul of the golden mean. Once again, a betting analogy is helpful. Your aim is to win money and avoid losing it. You don’t want to bet recklessly, but neither do you want to leave money on the table if there’s a good bet to be made. It may seem that not betting is always a safe option, but that’s a mistake. Both placing a bet and not placing a bet comes with risk. In the same way, it is a mistake to think that withholding belief is the safe play when something seems true—​ that venturing belief invites risk that can be avoided simply by withholding assent. Withholding assent carries with it risks of its own: namely, of foolishly squandering an opportunity to attain truth. When we acknowledge that there is risk either way, we can begin to see that, if the proposition’s truth is sufficiently apparent, not believing is in fact the riskier option. Withholding assent toward what strongly seems true is, as far as our subject can tell, like walking away from a good bet (one with a positive expected outcome). Nevertheless, the sense that belief is riskier can be hard to shake. We find an explanation, I think, in the well-​documented psychological bias called “loss aversion” (Kahneman and Tversky 1992). For instance, people will refuse to take favorable bets that, by the numbers, have positive expected outcomes, simply because we are more sensitive to loss than to gain. To be clear, we are not only talking about situations where the badness of losing really is greater than the goodness of winning. For instance, one might decline betting $10 on a coin flip with 1-​to-​1 odds, or even 2-​to-​1 odds, simply because losing the $10 is considerably worse than gaining an extra $10 or $20 is good. This form of loss aversion is perfectly rational. But loss aversion extends beyond this. For instance, I’ve seen people decline a series of one hundred bets on coin flips with 2-​to-​1 odds. At this point, the natural tendency toward loss aversion has spilled over into irrationality. It is understandable, then, why someone might feel resistant to believing

232  A Reidian Argument for Phenomenal Conservatism, Part I in the state of empty symmetrical evidence. The prospect of getting things wrong looms large. But reflection can help dispel this illusion. In the betting case, the math can show us that the bet is favorable and (at least in some cases) our aversion irrational. Likewise, my hope is that our present discussion may help one appreciate that “taking the bet”—​believing p solely based on its apparent truth—​is sometimes the proper strategy. A second objection stems from a closely related misconception—​one made prevalent by the influence of Descartes and his method of doubt. The idea is that being skeptical, or hard to convince, or incredulous is somehow an unqualified epistemic ideal. There are times when incredulity is called for and times when it is not. Returning to our betting analogy, there are times in which refusing to make a bet is a cautious, prudent strategy. There are other times, when the odds are astoundingly in your favor, when the refusal to make a bet is not cautious but cravenly (or at least overly conservative). It is the same in the epistemic sphere. The exaltation of incredulity represents an undue fixation on not being duped. Huemer calls it “an extreme and unreasonable bias toward the ‘avoiding error’ part of your epistemic goal” (Huemer 2001, 105). And this, really, is the crux of the matter. The categorical refusal to trust how things appear without independent verification ultimately boils down to an overprioritizing of the avoidance of falsehood.10 William James describes this mentality well, if with exaggeration: He who says, “Better go without belief forever than believe a lie!” merely shows his own preponderant private horror of becoming a dupe. He may be critical of many of his desires and fears, but this fear he slavishly obeys. He cannot imagine any one questioning its binding force. For my own part, I have also a horror of being duped; but I can believe that worse things than being duped may happen to a man in this world … Our errors are surely not such awfully solemn things. In a world where we are so certain to incur them in spite of all our caution, a certain lightness of heart seems healthier than this excessive nervousness their behalf. (James [1896] 2009, 18–​19) I would add to James that not only do we often exaggerate the horror of getting things wrong, but we fail to properly value getting things right. Where is our zeal for truth? If we do not allow ourselves to be moved by the felt truth of a proposition, then we have passed beyond virtuous suspicion into a form of intellectual sloth—​a failure to properly appreciate and zealously pursue the attainment of truth. Not even Descartes, it should be noted, remained perpetually incredulous. When the natural light of reason

A Reidian Argument for Phenomenal Conservatism, Part I  233 made something so apparent that it became impossible to doubt, even he permitted belief in how things seemed. Having dealt with these objections, our original intuition stands vindicated: some seemings, if only especially strong ones, permit belief in the state of empty symmetrical evidence. At the very least, they permit a positive level of confidence. And this entails that some seemings are foundational evidence. There is something amusing about using our intuitions about such cases (a species of seeming) to support the conclusion that seemings immediately justify. It’s circular. Not logically circular—​I’m not relying on Minimal-​PC as a premise in an argument for its truth—​but there is a kind of circularity in practice. I’m proceeding as if Minimal-​PC is true in order to argue for Minimal-​PC. (I defend this kind of source circularity in Chapter 3, §3.2.) But this only underscores the arguments of the preceding sections. How else am I supposed to argue for Minimal-​PC, or any conclusion for that matter, except by appealing to premises that seem true, or that are supported by other things that seem true? Psychologically, any belief you or I form either for or against Minimal-​PC (if it stands any chance of being justified) is going to be ultimately based on seemings. Furthermore, if FN, SE, and EV are true, then there is no other way to support Minimal-​PC except by tacitly relying on its truth. This kind of circularity, therefore, is unavoidable. It is also benign. It does not beg the question in that people can be, and in fact often are, moved by appearances before believing in Minimal-​PC (or having any epistemological beliefs at all, for that matter). If anything, this circularity simply illustrates how all reasoning, at least as we humans do it, tacitly presupposes a basic trust in how things seem—​ that is, in the truth of Minimal-​PC. This, conveniently enough, transitions us into our final and most illuminating argument. 8.5  A Constitutive Rule of Rationality Perhaps the fundamental answer concerning why it is at least sometimes permissible to believe solely on the basis of apparent truth is that doing so is a constitutive rule of rationality. A constitutive rule more generally is one that is definitive of the kind of activity we are engaging in. If we cease to act in accordance with that rule, then we cease to engage in that activity altogether. “The rules of chess,” John Searle explains, “do not just regulate, but they constitute the activity they regulate. … If you do not follow these rules, or at least a sufficiently large subset of the rules, you are not playing chess” (Searle 2018, 51–​52). These kinds of rules can be compared to other, merely regulative rules that we can break while acting within the bounds of that activity. For instance, in chess one should not give up a more valuable piece for a less valuable one without some compensation

234  A Reidian Argument for Phenomenal Conservatism, Part I in position or material later on. This is a mere regulative rule because, in violating it, we do not thereby stop playing chess; we simply play chess poorly.11 Compare this to someone who stops making allowable moves altogether. Perhaps he starts moving his King to non-​adjacent squares all over the board or, better yet, he categorically refuses to make any moves at all. This person has ceased to play chess entirely. He is no longer acting as a chess player or engaging in chess activity. Applying this to the domain of rational activity, the systematic refusal to operate in accordance with a constitutive rule results not in irrational activity per se but in the cessation of rational activity altogether. Patrick Rysiew observes that: constitutive rules have both a descriptive and a prescriptive aspect: they describe the behaviour (at least within certain limits) of one engaged in the activity in question; but for one who is so engaged, these rules also prescribe (and prohibit) certain ways of acting. (Rysiew 2002, 450) That is, the fact that (descriptively) one must proceed in such-​and-​such a manner when engaged in some type of activity has implications for how (prescriptively) one ought to proceed when engaged in that activity. In particular, it must sometimes be permissible to proceed in accordance with the constitutive rules. Otherwise, no token activity of that type would be permissible, and a type of activity with no permissible tokens is not a genuine type of activity at all. For example, if someone invites you to play “Dead-​End Chess,” where this is a game with the same pieces and aims as chess, except that the pieces are not permitted to move anywhere or do anything, then you have met a false friend who has not invited you to play any real game. You certainly can’t be a “Dead-​End Chess” player, even if you sit down at the board and claim to be playing it.12 Applying this to the matter at hand, if rational activity is to be a genuine kind of activity, and if we are to be identified as rational agents, then it must sometimes be possible for us to form attitudes in a way that is permitted by the rules governing rational activity. How do we discern the constitutive rules of an activity? One tell-​tale sign is this: if it is impossible to begin or sustain participating in that activity without abiding by a rule, then following that rule is constitutive of the activity in question. So if we find that rational activity, or some crucial kind of it, cannot get off the ground without granting implicit trust to at least some appearances, then we have good grounds for establishing that trust as a constitutive rule of rationality. And, in what follows, that is precisely what we will see to be the case. The formation of beliefs (or other doxastic states) is a central form of rational activity. (I will ignore rational activity that is not of this particular

A Reidian Argument for Phenomenal Conservatism, Part I  235 sort.) Notice, however, that forming beliefs in and of itself does not necessarily qualify as such. If a belief is brutely caused in us—​say, by the tinkering of an evil neuroscientist—​the formation of that belief is not an instance of rational activity (not by us, anyhow). It is only when we form a belief on some rational basis that it counts as our rational activity. For internalists, such rational bases will be understood in terms of evidence. Hence, we can say that the formation of a belief is an instance of rational activity only if it is formed on the basis of evidence. This principle—​the evidential requirement—​is a constitutive rule of rational belief formation. To be clear, the evidential requirement does not say that our evidence must always be adequate. Beliefs based on bad evidence remain rational in a broader sense (rational versus a-​rational), though not rational in a narrower sense (rational versus irrational). Irrational beliefs of this sort break regulative rules about the quality of our evidence, not the constitutive rule requiring evidence more generally. In other words, the evidential requirement is not that our beliefs be based on good evidence, but that those beliefs be based on mental states having the general character of evidence—​that the causes of those beliefs possess the kinds of features that might make them evidence for something in some circumstances. Now, when beliefs are formed through inference, our evidence consists in other beliefs. On the foundationalist framework, however, inferential belief formation is dependent on non-​inferential belief formation such that the former is a rational activity only if the latter is. Beliefs only have the general character of evidence if they are justified, and, per foundationalism, you can’t get any justified beliefs at all unless some beliefs enjoy non-​ inferential support. Put more intuitively, arguments can’t justify their conclusions apart from immediately justified beliefs any more than a telescope can extend the vision of a blind man (Reid [1764] 1997, 71). Thus, belief formation is a rational activity only if basic belief formation is a rational activity. Given the evidential requirement, the formation of basic beliefs is a rational activity only if it is caused by mental states having the general character of evidence (and caused by the features of those mental states that give them that general character). However, in defending seemings exclusivism, I argued that the only kinds of states that plausibly qualify as evidence for basic beliefs (i.e., foundational evidence) are seemings. States lacking either propositional content or forcefulness are entirely unfit to play the role. They are a lot like beliefs, in fact—​having the general character of evidence only if there is some existing framework of background evidence in place. But then, like beliefs, those non-​conceptual or non-​forceful states have the general character of evidence only if there are other, prior states that possess this general character in and of themselves. And what sort of mental state fits that description? Seemings are the only plausible option.

236  A Reidian Argument for Phenomenal Conservatism, Part I Hence, if we categorically refuse to trust in how things seem absent outside verification, we thereby close ourselves off to even the possibility of rational belief formation. So, an implicit trust in how things appear, at least on occasion, proves to be a constitutive rule of rationality. This conclusion, I think, makes sense on an intuitive as well as a theoretical level. If the apparent truth of a proposition moves us not one iota, not even if that proposition is utterly and altogether obvious, then we cannot engage in rational belief formation. How could we even get started? The smallest advancement in reasoning requires at least minimal trust in the way things seem to us. The refusal to grant such trust is essentially a refusal to engage in rational belief formation altogether, and thereby to divest ourselves of our identity as rational agents. We needn’t fret, of course, about someone successfully divesting themselves of rationality in this way. Even the insane grant credibility to how things seem to them. Only in the complete cessation of mental activity will one totally cease to function as a rational agent. But even if we could become entirely distrusting, the point remains the same. To function as a rational agent is to sometimes place trust in how things seem. To be absolutely clear, the claim is not that rational agents must believe that Minimal-​PC is true in order to engage in rational activity. That over-​ intellectualizes the matter. The majority of people engage in rational belief formation perfectly well without stopping to consider any epistemic principles, much less Minimal-​PC specifically. And many of those who do consider Minimal-​PC actively disbelieve it. Nevertheless, all of these people still abide by it. They place implicit trust in how things seem at least some of the time. They treat Minimal-​PC as true in practice, even though they do not explicitly believe it. If it were not so, then those people could not engage in rational activity at all. Once again, my overall argumentative strategy is derived from Reid, particularly as interpreted by Rysiew (2002). Reid argues that abiding by some of the first principles of common sense—​essentially, placing initial trust in the deliverances of at least some of our rational faculties—​is constitutive of rational activity. For instance, Reid is willing to engage with those who admit some first principles but not others, who he calls “semi-​sceptics,” for there is still a rational agent there to engage with; but to the “thorough and consistent sceptic,” Reid writes, “I have nothing to say.” For while such a skeptic remains in his skeptical frame of mind, “you can never oblige him to yield any thing else” (Reid [1764] 1997, 71). By abandoning all first principles, he has cut off even the possibility of rational belief formation. If we compare rational belief formation to sight, he has blinded himself and is no longer a seeing creature. In distinguishing the two offices of reason, Reid adds that the acceptance of first principles is what “entitles [most people] to the denomination of reasonable creatures” (Reid [1785] 2002, 433).

A Reidian Argument for Phenomenal Conservatism, Part I  237 The implication is that someone who completely abandons first principles ceases to be numbered among the population of rational agents. He has become, to borrow a different image from Aristotle, no better than a plant (Aristotle 1995, 253).13 This is why Reid says that the abandonment of first principles “is not the act of the philosopher, but the fool or the madman” (Reid [1764] 1997, 37). As Rysiew puts it, “gross failure to accept the first principles of common sense is,” not mere irrationality, but rather “just plain lunacy” (Rysiew 2002, 450). And madmen are to be pitied, not reasoned with “a man who disbelieves in his own existence, is surely as unfit to be reasoned with, as a man that believes he is made of glass” (Reid [1764] 1997, 16). What the complete skeptic is missing is trust: trust that the world actually is the way that one’s mind presents it as being. This trust needn’t be absolute—​an initial, defeasible trust is enough to get started. Furthermore, we can’t yet conclude that it must be extended to everything our minds purport to reveal to us about reality—​rational activity is possible so long as we grant trust to at least some of the deliverances of our rational faculties. But a minimal level of trust in our own minds is an unavoidable requirement for rational activity and agency. Here we see again that being rational does not mean being incredulous, or hard to convince, at least not as a default. Precisely the opposite is true. Our identification as rational creatures requires some measure of credulity—​a ready assent to the deliverances of our rational faculties even in the absence of independent verification. Such credulity, when placed in the testimony of others, might be called faith. If we are allowed to import this concept into an individual context, then we can say that placing faith in ourselves and in the ability of our minds to reveal reality is necessary a precondition of rationality. Thus, Foley writes, “we have no choice but to acknowledge that significant intellectual projects require correspondingly significant leaps of intellectual faith” (Foley 2001, 24). Reid adds, “I am persuaded, that the unjust live by faith as well as the just” (Reid [1764] 1997, 4). And, for good measure, Chesterton remarks, “It is idle to talk always of the alternative of reason and faith. Reason is itself a matter of faith. It is an act of faith to assert that our thoughts have any relation to reality at all” (Chesterton 1908). On the essentiality of self-​trust, Reid and I agree. Where we diverge is on where precisely that trust is placed. Reid thinks that we must place implicit trust in the basic beliefs produced by the proper functioning of our natural faculties. (To “place implicit trust in a belief” is to continue to hold that belief without independent verification until there is some positive reason to doubt it.) Our basic natural beliefs are rock bottom for Reid. They are the ultimate starting points for all rational activity. Reid’s conclusion here seems to be based on his understanding of how basic beliefs are formed.

238  A Reidian Argument for Phenomenal Conservatism, Part I To see this, notice that trust (of the sort that concerns us) only makes sense in a context where something is purported to be true. In the domain of mental states, this requires first that the state be capable of being true or false—​that it possess propositional content—​and also that it put forward that content as true. But Reid does not acknowledge any states matching this description different from or prior to belief. Basic perceptual beliefs, for instance, are occasioned by raw sensations devoid of propositional content. Self-​evident beliefs are occasioned by merely conceiving of the proposition (where such conceptions do not present their content as true). Thus, for Reid, our basic beliefs are the first opportunities we have to show trust in our rational faculties. These basic beliefs are their testimony concerning what the world is like. Hence, it is in those beliefs that our trust must be placed. In contrast, I have argued that there are other states that put forward propositional content as true—​namely, seemings—​and that such seemings are the bases for our basic beliefs. Thus, while Reid’s epistemological instincts are sound, he gets the psychology wrong: our basic beliefs are not rock bottom. There are prior, experiential states on which those basic beliefs are based. It is here, in experience, that we reach rock bottom. Experience is the ultimate and only way we gain insight into how things are. It is the most basic way in which our minds “testify” to us as to what the world is like.14 Not all experiences do this, of course, but only those that purport to reveal something about reality: that is, those experiential states that possess propositional content and that present that content true. I have called such states seemings. To implicitly trust a seeming is, absent other considerations, simply to believe (or place positive credence in) what seems to be true, even if one has no independent verification of the seeming’s reliability. In other words, to implicitly trust a seeming is to treat it as foundational evidence for its content; and this is a trust that we must place in seemings, at least on occasion, if we are to engage in any rational belief formation whatsoever. Our identities as rational believers require it. Accordingly, trusting in how things seem must sometimes be permissible for rational agents. If it is not sometimes permissible, then rationality itself is defunct. For the very conditions that must obtain in order for something to count as an instance rational activity would also make it impermissible. Huemer writes: If phenomenal conservatism is false, then dialectic as we practice it is fundamentally and intrinsically irrational. In calling it “intrinsically” irrational, I mean to imply that the problem would be irremediable; there is nothing that could be recognized as a form of dialectic or reasoning that would not be irrational. (Huemer 2001, 107)

A Reidian Argument for Phenomenal Conservatism, Part I  239 Contra Huemer, this is not true of PC in its unrestricted form—​someone who doesn’t place trust in all seemings can still engage in rational activity so long as he or she trusts some. But it is true of Minimal-​PC. Without it, all “rational activity” would be irrational. That is to say, there simply wouldn’t be any such thing as rational activity anymore. It would be like Dead-​End Chess. But, of course, we are rational agents (and any attempt to disagree with that is inherently self-​defeating). Thus, since a minimal trust in how things appear is constitutive of a central aspect of rational activity, it follows that granting such trust must sometimes be rationally permissible for us. That is, Minimal-​PC is true. In closing, let us draw the connection between the argument in this section and the Reidian defense of epistemic conservatism criticized in the final section of Chapter 7. Reid argued that the irresistibility and naturalness of our basic natural beliefs guaranteed that we are not at fault for, and therefore justified in, believing them. In response, I pointed out that irresistibility and naturalness were sure defenses against fault only if construed in an especially strong way: irresistible in that it is not possible to engage in rational activity without them; natural in that no rational agent could possibly lack them. If certain principles are essential to human rationality in these exceptional ways, then following them must be permissible, the beliefs formed in accordance with them justified. For I do not see how one can be rationally criticized for abiding by a principle that is required to engage in any rational activity at all. Moreover, we know that following such principles is proper to the ideal human intellect since abiding by them is necessary for any functioning human intellect. I submit that Minimal-​PC is one such principle. Notes 1 Others seem to acknowledge that some seemings can immediately justify but in fact do not. For instance, BonJour (2004) says that certain rational and introspective seemings can immediately justify, but he requires that we have a meta-​ justification vouching for the reliability of these seemings. Hence, whatever justification they provide is not immediate at all. Similarly, Steup (2013) says that seemings can justify if we have memory data verifying their reliability, but whatever justification is provided by such seemings cannot be immediate in the way we have defined it here. 2 The same could arguably be said of many in the Eastern canon as well. To give just one example, the Indian philosopher Gaṅgeśa acknowledges that perceptual cognitions are to be assumed true in the absence of defeaters (Phillips 2017). 3 In a footnote, Shields says that the name “phainomenological conservatism” was inspired by Huemer’s phenomenal conservatism, which he takes to be closely related. 4 It is a common mistake to think that, for Descartes, the justification of principles clearly and distinctly conceived (like one’s own existence) derives from some

240  A Reidian Argument for Phenomenal Conservatism, Part I kind of argument. Instead, its justification is immediate, stemming directly from its clarity and distinctness. The arguments for these clear and distinct principles (like the cogito) come downstream from that immediate justification and attempt to preserve it from encroaching doubts. 5 Even some who deny seemings foundational status de dicto may in fact endorse it de re. For instance, Bengson (2015) contends that presentational states can immediately justify, and characterizes them in a way that closely mirrors seemings; however, he specifically denies that presentational states are seemings (Bengson 2015, 729–​ 731). He says that with presentational states, unlike seemings, the subject cannot always “fully articulate” their content. If by this he means that the content of presentational states is non-​conceptual, and cannot always be articulated conceptually, then there is a genuine distinction between presentations and seemings. On the other hand, if Bengson just means that the subject may struggle to consciously rehearse the conceptual content belonging to his or her presentational state, then presentational states just are seemings (and Bengson is wrong to require that seemings be “fully articulable” in some special sense). 6 See Chapter 1, §1.4.1 on folk-​psychological explanations of belief. 7 Whether these four are the only fundamental sources of basic beliefs is controversial. Possible alternatives include testimony, moral conscience, a sensus divinitatis, etc. My general treatment of these potential sources is that they will either be reducible to one or more of the four main sources, or else the basic beliefs they produce will be based on their own kind of seeming. On testimony and seemings, see Dougherty (2014). On moral seemings, see Huemer (2005) and McAllister (2021, §2). On theistic seemings, see Tucker (2011) and McAllister and Dougherty (2019). 8 I’m reminded of the G.K. Chesterton quote, “They are all on the road to the emptiness of the asylum. For madness may be defined as using mental activity so as to reach mental helplessness; and they have nearly reached it” (Chesterton 1908). 9 I also find Moretti’s criticism of Tucker persuasive. See Moretti (2020, 14). 10 The issue of how to weigh our competing epistemic goals is rarely discussed (see Riggs 2003 for an exception). William James seemed to think that our passional natures are allowed to decide the relative importance of each goal, but I take there to be a more objective, purely epistemic answer as to how these goals ought to be balanced against one another. Though I am not inclined toward permissivism myself, my assumption here is compatible with forms of permissivism that allow flexibility with respect to how these goals are weighed against one another (e.g., Kelly 2013), so long as we acknowledge that there are limits to this flexibility. Certain weightings are simply unallowable, even if more than one weighting is. 11 I can vouch for this from much personal experience. 12 Walking into class one day, I was greeted by two epistemology students, Justin Lee and Micah Miller, sitting in front of a chess board “playing” Dead-​End Chess. Neither won.

A Reidian Argument for Phenomenal Conservatism, Part I  241 13 I must include a nod to one of my former students here, Sarah D’Spain, who would fight off her worst skeptical impulses by keeping a small plant on the windowsill by her workspace. 14 In a way, Reid agrees with the general idea that experience is what must be trusted in order to engage in any rational inquiry whatsoever. He simply holds that basic natural beliefs partially constitute our experiences. For instance, perceptual experience just is the combination of a basic perceptual belief and the sensations that triggered it. Once again, the mistake is with his account of experience, not the epistemological principles per se. See McAllister (2019, §4).

References Allen, James. 2020. “Carneades.” The Stanford Encyclopedia of Philosophy. 2020. https://​plato.stanf​ord.edu/​archi​ves/​fall2​020/​entr​ies/​carnea​des/​. Alston, William P. 1999. “Back to the Theory of Appearing.” Philosophical Perspectives 13: 181–​203. Audi, Robert. 2013. “Doxastic Innocence: Phenomenal Conservatism and Grounds of Justification.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 181–​201. New York: Oxford University Press. Bealer, George. 2000. “A Priori Knowledge.” The Proceedings of the Twentieth World Congress of Philosophy 5: 1–​12. Bedke, Matthew S. 2008. “Ethical Intuitions: What They Are, What They Are Not, and How They Justify.” American Philosophical Quarterly 45 (3): 253–​270. Beillard, Julien. 2016. “Phenomenal Conservatism, Reflection and Self-​Defeat.” Logos and Episteme 7 (2): 187–​199. Bengson, John. 2015. “The Intellectual Given.” Mind 124 (495): 707–​760. Berghofer, Philipp. 2020. “Towards a Phenomenological Conception of Experiential Justification.” Synthese 197 (1): 155–​183. Bergmann, Michael. 2013. “Externalist Justification and the Role of Seemings.” Philosophical Studies 166 (1): 163–​184. BonJour, Laurence. 2004. “In Search of Direct Realism.” Philosophy and Phenomenological Research 69 (2): 349–​367. Brogaard, Berit. 2013. “Phenomenal Seemings and Sensible Dogmatism.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 270–​289. Oxford: Oxford University Press. Chesterton, G. K. 1908. Orthodoxy. New York: John Lane Co. Chudnoff, Elijah. 2018a. “Epistemic Elitism and Other Minds.” Philosophy and Phenomenological Research 96 (2): 276–​298. —​—​—​. 2018b. “The Epistemic Significance of Perceptual Learning.” Inquiry: An Interdisciplinary Journal of Philosophy 61 (5–​6): 520–​542. DePaul, Michael. 2009. “Phenomenal Conservatism and Self-​Defeat.” Philosophy and Phenomenological Research 78 (1): 205–​212. DePoe, John. 2011. “Defeating the Self-​ Defeat Argument for Phenomenal Conservativism.” Philosophical Studies 152 (3): 347–​359.

242  A Reidian Argument for Phenomenal Conservatism, Part I Descartes, René. [1641] 2006. Meditations, Objections, and Replies. Translated by Roger Ariew and Donald A. Cress. Indianapolis, IN: Hackett Publishing. Dougherty, Trent. 2014. “Faith, Trust, and Testimony: An Evidentialist Account.” In Religious Faith and Intellectual Virtue, edited by Laura Frances Callahan and Timothy O’Connor, 97–​123. Oxford: Oxford University Press. —​—​—​. 2018. “Seemings, Reasons, and Knowledge: A Defense of Phenomenal Conservatism.” In Believing in Accordance with the Evidence: New Essays on Evidentialism, edited by Kevin McCain, 39–​51. Synthese Library. Cham, CH: Springer International. Farley, R.M. 2017. “The Sellarsian Dilemma: Not What It Seems.” Southwest Philosophy Review 33 (1): 115–​123. Foley, Richard. 2001. Intellectual Trust in Oneself and Others. Cambridge: Cambridge University Press. Gage, Logan Paul, and Blake McAllister. 2020. “The Phenomenal Conservative Approach to Religious Epistemology.” In Debating Christian Religious Epistemology: An Introduction to Five Views on the Knowledge of God, edited by John M. DePoe and Tyler McNabb, 61–​81. London: Bloomsbury Academic. Hasan, Ali. 2013. “Phenomenal Conservatism, Classical Foundationalism, and Internalist Justification.” Philosophical Studies 162 (2): 119–​141. Huemer, Michael. 2001. Skepticism and the Veil of Perception. Lanham, MD: Rowman & Littlefield. —​—​—​. 2005. Ethical Intuitionism. London: Palgrave Macmillan. —​ —​ —​ . 2007. “Compassionate Phenomenal Conservatism.” Philosophy and Phenomenological Research 74 (1): 30–​55. —​—​—​. 2011. “Phenomenal Conservatism and Self-​Defeat: A Reply to DePoe.” Philosophical Studies 156 (1): 1–​13. Jackson, Alexander. 2011. “Appearances, Rationality, and Justified Belief.” Philosophy and Phenomenological Research 82 (3): 564–​593. James, William. [1896] 2009. “The Will to Believe.” The Project Gutenberg. 2009 1896. www.gutenb​erg.org/​files/​26659/​26659-​h/​26659-​h.htm. Kahneman, Daniel, and Amos Tversky. 1992. “Advances in Prospect Theory: Cumulative Representation of Uncertainty.” Journal of Risk and Uncertainty 5 (4): 297–​323. Kelly, Thomas. 2013. “Evidence Can Be Permissive.” In Contemporary Debates in Epistemology, edited by Matthias Steup, John Turri, and Ernest Sosa, Second edition, 298–​311. Hoboken, NJ: Wiley-​Blackwell. Koksvick, Ole. 2011. “Intuition.” Dissertation, Canberra: The Australian National University. Kriegel, Uriah. 2021. “The Structure of Phenomenal Justification.” Australasian Journal of Philosophy. Lycan, William G. 2013. “Phenomenal Conservatism and the Principle of Credulity.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 293–​305. Oxford: Oxford University Press.

A Reidian Argument for Phenomenal Conservatism, Part I  243 Markie, Peter. 2013. “Searching for True Dogmatism.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 248–​269. Oxford: Oxford University Press. McAllister, Blake. 2019. “A Return to Common Sense: Restorationism and Common Sense Epistemology.” In Restoration & Philosophy, edited by J. Caleb Clanton, 35–​78. Knoxville, TN: University of Tennessee Press. —​—​—​. 2021. “Understanding Moral Disagreement: A Christian Perspectivalist Approach.” Religions 12 (5): 318. McAllister, Blake, and Trent Dougherty. 2019. “Reforming Reformed Epistemology: A New Take on the Sensus Divinitatis.” Religious Studies 55 (4): 537–​557. McCain, Kevin, and Luca Moretti. 2021. Appearance and Explanation: Phenomenal Explanationism in Epistemology. Oxford: Oxford University Press. McGrath, Matthew. 2013. “Phenomenal Conservatism and Cognitive Penetration: The Bad Basis Counterexamples.” In Seemings and Justification, edited by Chris Tucker, 225–​247. Oxford: Oxford University Press. Moore, G. E. [1925] 1959. “A Defence of Common Sense.” In Philosophical Papers. New York: Macmillan. Moretti, Luca. 2020. Seemings and Epistemic Justification: How Appearances Justify Beliefs. New York: Springer. Phillips, Stephen H. 2017. “A Defeasibility Theory of Knowledge in Gaṅgeśa.” In The Oxford Handbook of Indian Philosophy, edited by Jonardon Ganeri, 541–​ 558. Oxford: Oxford University Press. Plantinga, Alvin. 1993. Warrant and Proper Function. Oxford: Oxford University Press. Plug, Allen. 2011. “Epistemic Justification and Truth-​Conduciveness.” Dissertation, University of Rochester. Pollock, John. 1974. Knowledge and Justification. Princeton, NJ: Princeton University Press. Pryor, James. 2000. “The Skeptic and the Dogmatist.” Noûs 34 (4): 517–​549. Reid, Thomas. [1764] 1997. An Inquiry into the Human Mind on the Principles of Common Sense, edited by Derek R. Brooks. University Park, PA: Pennsylvania State University Press. —​—​—​. [1785] 2002. Essays on the Intellectual Powers of Man, edited by Derek R. Brooks. University Park, PA: Pennsylvania State University Press. Reiland, Indrek. 2015. “Experience, Seemings, and Evidence.” Pacific Philosophical Quarterly 96 (4): 510–​534. Riggs, Wayne D. 2003. “Balancing Our Epistemic Goals.” Noûs 37 (2): 342–​552. Rogers, Jason, and Jonathan Matheson. 2011. “Bergmann’s Dilemma: Exit Strategies for Internalists.” Philosophical Studies 152 (1): 55–​80. Rosen, Gideon. 2001. “Nominalism, Naturalism, Epistemic Relativism.” Noûs 35 (s15): 69–​91. Rysiew, Patrick. 2002. “Reid and Epistemic Naturalism.” Philosophical Quarterly 52 (209): 437–​456. Searle, John. 2018. “Constitutive Rules.” Argumenta 4 (1): 51–​54.

244  A Reidian Argument for Phenomenal Conservatism, Part I Shields, Christopher. 2013. “The Phainomenological Method in Aristotle’s Metaphysics.” In Aristotle on Method and Metaphysics, edited by Edward Feser, 7–​27. Philosophers in Depth. London: Palgrave Macmillan. Skene, Matthew. 2013. “Seemings and the Possibility of Epistemic Justification.” Philosophical Studies 163 (2): 539–​559. Smithies, Declan. 2019. The Epistemic Role of Consciousness. Oxford: Oxford University Press. Steup, Matthias. 2013. “Does Phenomenal Conservatism Solve Internalism’s Dilemma?” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 135–​153. Oxford: Oxford University Press. Swinburne, Richard. 2018. “Phenomenal Conservatism and Religious Experience.” In Knowledge, Belief, and God: New Insights in Religious Epistemology, edited by Matthew A. Benton, John Hawthorne, and Dani Rabinowitz, 322–​ 338. Oxford: Oxford University Press. Tooley, Michael. 2013. “Michael Huemer and the Principle of Phenomenal Conservatism.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 306–​327. Oxford: Oxford University Press. Tucker, Chris. 2010. “Why Open-​Minded People Should Endorse Dogmatism.” Philosophical Perspectives 24 (1): 529–​545. —​ —​ —​ . 2011. “Phenomenal Conservatism and Evidentialism in Religious Epistemology.” In Evidence and Religious Belief, edited by Kelly James Clark and Raymond J. VanArragon, 52–​73. Oxford: Oxford University Press. —​—​—​. 2013. “Seemings and Justification: An Introduction.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 1–​29. Oxford: Oxford University Press. Tucker, Chris. forthcoming. “Experience as Evidence.” In The Routledge Handbook of the Philosophy of Evidence, edited by Maria Lasonen-​Aarnio and Clayton M. Littlejohn. Abington, UK: Routledge. Unger, Peter K. 1975. Ignorance: A Case for Scepticism. Oxford: Oxford University Press.

9 A Reidian Argument for Phenomenal Conservatism, Part II All Seemings

In the previous chapter, we concluded that at least some seemings are foundational evidence—​a principle I called Minimal-​PC. In this chapter, I show that if Minimal-​PC is true, then so is PC in its unrestricted form. Put succinctly, if some seemings are foundational evidence, then all are. This is Step Two in the Reidian Argument for PC: The Reidian Argument for PC Step One

For some seemings, if it seems to S that p, then S thereby has some foundational evidence for p. (Minimal-​PC) Step Two If this is true for some seemings, then it is true for all seemings. Conclusion Thus, for all seemings, if it seems to S that p, then S thereby has some foundational evidence for p. (PC) After defending Step Two, the case for phenomenal conservatism, and seeming foundationalism as a whole, will be complete. 9.1  From Some to All As we begin, it is important to notice that the arguments of the previous chapter entitle us to conclude quite a bit more than just Minimal-​PC. We saw that seemings are the only potential sources of immediate justification available to us, both in principle and as a matter of how we form beliefs. To avoid problematic forms of skepticism, then, it is not enough that only a single token seeming serve as foundational evidence. There need to be enough token seemings of enough different types to justify the great panoply of beliefs that are obviously justified. To unpack this, there must be multiple kinds of seemings that are foundational evidence. This must include, at minimum, rational intuitions and introspective seemings. Moreover, if perceptual and memorial beliefs are sometimes basic rather than inferred from rational and introspective

DOI: 10.4324/9781003281108-10

246  A Reidian Argument for Phenomenal Conservatism, Part II beliefs, as I strongly suspect they are, then we must include perceptual and memorial seemings as well. Avoiding skepticism also requires that many token seemings of these various types be foundational evidence, including the sorts of token seemings operative in the ordinary contexts in which we take ourselves to be justified. Put roughly, we can conclude not only that some seemings are foundational evidence, but that a great many are. Were we to end our discussion here, it would already have far-​reaching implications. Nevertheless, I think it is possible to go further. We can conclude not only that some seemings serve as foundational evidence, or even that most seemings do, but that all seemings do. In moving from “some” to “all,” we will once take our cues from Thomas Reid, who writes the following: It is a good argument ad hominem, if it can be shewn, that a first principle which a man rejects, stands upon the same footing with others which he admits: For, when this is the case, he must be guilty of an inconsistency who holds the one and rejects the other. Thus the faculties of consciousness, of memory, of external sense, and of reason, are all equally the gifts of Nature. No good reason can be assigned for receiving the testimony of one of them, which is not of equal force with regard to the others. The greatest Sceptics admit the testimony of consciousness, and allow, that what it testifies is to be held as a first principle. If therefore they reject the immediate testimony of sense, or of memory, they are guilty of an inconsistency. (Reid [1785] 2002, 463) Though I have already voiced disagreement with Reid over some of the details (our trust is placed in seemings, not beliefs), the basic principle underlying Reid’s argument is insightful nonetheless. What is it that permits us to trust certain seemings without prior verification of their reliability? If all seemings possess those same features, then all seemings deserve the same implicit trust. Nor is this simply an ad hominem argument as Reid suggests. If there is no epistemically relevant difference between seemings, then there can be no difference between their justificatory abilities.1 The rest of this chapter will look at attempts to distinguish between those seemings that are foundational evidence (“foundational seemings”) and those that are not. I will show that none of these attempts point to any relevant difference between types of seemings that would explain why one is worthy of implicit trust but not the other. Indeed, the most popular proposals are of entirely the wrong sort. Near the end, I will argue that the very same features that make some seemings capable of serving as foundational evidence are present in all seemings.

A Reidian Argument for Phenomenal Conservatism, Part II  247 9.2  External Differences Between Seemings We can rule out straight away any attempt to distinguish foundational seemings from others based on properties external to the agent. This follows from casual internalism about justification. No property can have any bearing on a seeming’s potential to justify unless we are in a position to take it into account in belief formation. If it is an internal property of the seeming—​one that is directly causally accessible to us—​then we are in a position to take that property into account apart from forming any further belief about whether the seeming possesses it. We can simply respond to that property of our seemings, without reflection. If, on the other hand, the property of the seeming is external to us, then it cannot have any bearing on the justificatory potential of that seeming unless we justifiably believe the seeming to possess it.2 But the matter that concerns us is whether seemings are foundational evidence—​that is, whether they can justify in a manner that does not require having justification for any other beliefs. Hence, the external properties of seemings are completely irrelevant to whether they count as foundational evidence or not. I find it helpful to think about the matter this way: no property can have any bearing on a seeming’s foundational status unless it can affect what we are permitted to believe in the state of empty symmetrical evidence. The state of empty symmetrical evidence, we may recall, is an extreme condition in which we have absolutely no evidence about the reliability of the seeming or the truth of its content. This includes anything we might learn about that seeming through a priori reflection. The only properties that could cause belief in those circumstances, and so the only properties that could bear any relevance to whether those beliefs are justified, are properties we can respond to without reflecting on the fact that we have them—​namely, internal properties. Thus, external properties cannot mark the difference between those seemings that can immediately justify and those that cannot. To make all of this less abstract, consider some externalist proposals. The common thread is that the etiology of the seemings—​their causal history—​ affects their status as foundational evidence. For instance, Susanna Siegel claims that an experience immediately justifies only if it is not cognitively penetrated in an epistemically bad way (Siegel 2012). Michael Bergmann claims that only those seemings that base belief in accordance with the proper functioning of one’s faculties operating according to a reliable, truth-​aimed design plan are capable of immediately justifying their contents (Bergmann 2013). Peter Markie suggests that the foundational status of a perceptual seeming may depend on whether it arises from knowing how to identify objects of the relevant sort (Markie 2013). Lu Teng argues that seemings fabricated by the

248  A Reidian Argument for Phenomenal Conservatism, Part II subject through the imagination are not foundational evidence, even if these feel exactly like other, non-​fabricated seemings (Teng 2018). Each of these proposals marks the difference between foundational and non-​ foundational seemings based on properties that are not directly accessible to our automatic processing systems. To put it colloquially, seemings don’t wear their causal histories on their faces. That means we cannot take those properties into account unless we have reason to believe in them; we cannot respond to them automatically in the way we can to, say, the seeming’s phenomenal properties. But then those causal histories matter not at all to whether those seemings are foundational. Put differently, if we were in the state of empty symmetrical evidence, seemings with and without those external properties would be, from our point of view, exactly the same. Obviously, learning about the external properties of a seeming will affect its justificatory potential more generally. If we have reason to believe that a seeming is unreliable, for instance, then that will defeat any immediate justification it may have initially provided. What external properties don’t affect is whether immediate justification is initially provided. In that regard, seemings must stand and fall only on the basis of directly accessible properties. This reasoning rebuts the majority of ways people try to discriminate between types of seemings. Their proposals are almost always externalist, appealing to features of the seeming that are not directly accessible, and so are irrelevant to whether that seeming is capable of immediately justifying belief. 9.2.1  Cognitive Penetration

I must specifically address one more externalist proposal given its prevalence in the literature. It has become common to propose, following Markie (2005) and Siegel (2012), that those seemings originating from a tainted source—​for example, wishful thinking—​are not foundational evidence, even if other seemings may be. This is typically supported by examples, such as the following from Peter Markie about a gold prospector named Virgil: Virgil is a novice. He has a general sense of what gold looks like, but he is not very good at its visual identification. Virgil, though, is consumed by a lust for gold. He wants very, very badly to make a discovery. … When Virgil looks at his nugget, his strong desire that it be gold comes into play and as a result, it just seems to him that it’s gold. He … believes accordingly. (Markie 2013, 257)

A Reidian Argument for Phenomenal Conservatism, Part II  249 We are supposed to intuit here not only that Virgil’s belief is unjustified all things considered, but that Virgil’s gold-​seeming does not lend even initial justificatory support to his belief, even if he is completely unaware of the seeming’s tainted origins (Markie 2013, 257–​259). This is the problem of cognitive penetration, so called because the taint has infiltrated our very experiences. Surveys of the problem of cognitive penetration already exist (Moretti 2020, ­chapter 3), so I will focus on what I take to be the crux of the dispute. The problem plainly relies on an externalist way of thinking. The tainted seemings are deemed to lack immediate justificatory power because of deficiencies in the way they were brought about, but the etiology of our seemings is not directly accessible to our automatic processing systems. Thus, given causal internalism, such features can make no difference to a seeming’s immediate justificatory potential, and the problem of cognitive penetration is no problem at all. Surprisingly, then, one sees a fair bit of hand-​wringing from internalist proponents of phenomenal conservatism about the problem of cognitive penetration. I wonder if this doesn’t simply amount to a failure of nerve. For if we genuinely have no indication that a seeming arises from a tainted source, then it is pretty clear that internalism prohibits this from having any effect on its ability to immediately justify. If we learn about those sordid origins down the line, then, sure, that will affect whether the seeming justifies belief all things considered (by providing a defeater), but that’s not what’s at issue here. Perhaps then the problem of cognitive penetration is best viewed as a challenge to internalism more generally. Nevertheless, I don’t see how it advances that debate significantly. Anyone with standard internalist intuitions will not share Markie’s verdict regarding Virgil and similar cases, at least when properly spelled out. In fact, I suspect that the only ones who will confidently share Markie’s intuition are card-​carrying externalists. To be clear, it should be evident to everyone (internalists alike) that Virgil is not justified in believing he’s found gold; but that is explained perfectly well by the fact that he has defeaters for any immediate justification that might come from his gold-​seeming. Virgil knows he’s not very good at visually identifying gold. He knows, as we all do, that desire can make us see what we want to see. He therefore has reason to be wary of his gold-​seeming. But that makes no difference to whether that gold-​seeming is foundational evidence—​to whether it can immediately justify in a state of empty symmetrical evidence. As to Markie’s claim that, intuitively, the gold-​seeming doesn’t lend even initial justificatory support to Virgil’s belief, I intuit precisely the opposite. If it seems that the substance before us is gold, and we truly have zero indication that this seeming is unreliable, then why wouldn’t that experience give us some level of justification for thinking we’ve found gold?3

250  A Reidian Argument for Phenomenal Conservatism, Part II To push this further, consider two agents who are phenomenal duplicates. One has a gold-​seeming caused in a normal, reliable manner. The second has a gold-​ seeming caused by wishful thinking. Despite their different origins, there is no detectable difference between the seemings. They feel exactly the same. I not only fail to have Markie’s intuition that the former seeming lends initial justificatory support while the latter seeming does not, I have a strong intuition in the opposite direction: these seemings stand or fall together. This is essentially the same intuition underlying the New Evil Demon Problem (Lehrer and Cohen 1983, Cohen 1984)—​an intuition that seems to be quite prevalent. Taking a slightly different case, imagine that the same agent has a reliable gold-​seeming on one occasion and a gold-​seeming caused by wishful thinking on another. However, as before, these seemings are completely indistinguishable from one another. Notice that if we do not accept the parity of such seemings, then we are forced to accept the truth of the following speech (adapted from Huemer 2006, 151): Previously, it seemed that the object before me was gold, and I believed it on that basis. That experience prima facie justified my belief. Now, it once again seems that this object before me is gold, but I do not believe it. This experience does not prima facie justify belief. What is the difference between these two experiences? There is none. As far as I can tell, they are exactly the same. Yet, for no reason at all, I contend that the one justifies but not the other. The absurdity of this speech makes it plain just how arbitrary the externalist position is from the first-​person point of view. None of this is to say that badly penetrated seemings are epistemically unproblematic. Clearly, beliefs formed on the basis of tainted seemings are deficient in important respects, but the question that interest us is whether they are unjustified, and that is a very specific charge. There can be many things wrong with a belief without its being unjustified. Alas, there is an unfortunate tendency to gloss over such distinctions when evaluating cases. Jonathan Kvanvig rightly warns us: Philosophers often have a predilection for blanket terminology to describe that which is objectionable, to the point where the term becomes so abstract and vapid as to be cognitively uninformative, serving only to express distaste or contempt. Such predilections mirror the behavior we acquired early on in having catch-​all terminology when our vocabularies fail us. (Kvanvig 2014, 159–​160)

A Reidian Argument for Phenomenal Conservatism, Part II  251 Suffice it to say that there are several alternative proposals as to what is wrong with such beliefs (or the subjects forming them) that are perfectly compatible with internalism and PC (see Moretti 2020, ­chapter 3, §5). The upshot is that the problem of cognitive penetration, and the examples motivating it, shouldn’t sway many toward externalism. Once the arbitrariness of the position is made clear, only those already inculcated into an externalist way of thinking are likely to have strong intuitions in that direction. It certainly shouldn’t distress those with more internalist intuitions (like those motivating the New Evil Demon Problem), nor do I see it winning many neutral observers. Finally, if framed as an objection to PC directly, rather than to internalism more generally, the proper response is simply to “hang tough,” insisting that tainted seemings do serve as foundational evidence, although their justificatory contributions are usually defeated in ordinary circumstances (Huemer 2013a; 2013b, Tucker 2010, §6). 9.3  Internal Differences Between Seemings Let us set externalism aside and focus on the internal properties of our seemings. The direct accessibility of these properties allows them, on a psychological level, to affect what we believe in the state of empty symmetrical evidence. If we needed to know something about them in order for them to cause belief—​for instance, that they exist or why they are significant—​then they could wield no influence in that extreme state in which all other evidence is barred. The question for us now becomes: which of these internal properties must our seemings possess in order to permissibly bring about belief in the state of empty symmetrical evidence? These are the ones relevant to whether a seeming can immediately justify. Consequently, those who wish to defend a restricted form of phenomenal conservatism must point to internal properties possessed by some seemings but not others that explain why the former seemings can permit belief in the state of empty symmetrical evidence but not the latter. We will discuss each of the most popular proposals in turn. 9.3.1 Infallibility

One of the more prominent and plausible proposals stems from classical foundationalism. A belief is infallible if it is formed in such a way that precludes its being false—​no belief formed in that same way, or on that same basis, could possibly be in error. A classical foundationalist might say that only infallible seemings can be foundational evidence, where a

252  A Reidian Argument for Phenomenal Conservatism, Part II seeming that p is infallible if and only if it guarantees the truth of any belief that p which is properly based on it. Effectively, this limits foundational evidence to certain subsets of introspective and rational seemings. The main motivation for such a view is that requiring infallibility at the foundations ostensibly guarantees that all justified beliefs are effectively connected to truth, and such a connection is often taken to be a feature of justification (DePoe 2021). Descartes went so far as to say that all justified beliefs are guaranteed to be true. Classical foundationalists today don’t go quite this far. Apart from the foundations, justified beliefs are at most guaranteed to be externally probable—​that is, probable in the way that beliefs based on objectively reliable grounds are probable. In either case, it is the guarantee of truth or reliability that motivates classical foundationalists to limit themselves to infallible foundations. The classical foundationalist proposal faces skeptical worries. It remains controversial whether a strong argument can be made starting only from infallible foundations for the truth of common sense beliefs such as the existence of an external world, the existence of other minds, that the world has existed for more than five minutes, etc., despite admirable efforts from explanationists to craft one. Even if there are strong arguments for such conclusions, however, it seems implausible that the arguments are strong enough to justify the especially high level of confidence we place in those conclusions. We do not tentatively believe that the world has existed for more than five minutes—​we are certain of it! But the aforementioned arguments, even if successful, are not conclusive in the way they would need to be to support such certitude. Finally, even if there are conclusive arguments, ordinary people don’t seem to base their beliefs on them. They do not reflect on the fact that things appear thus-​and-​so and infer from this that the world has existed for more than five minutes. In which case the common sense beliefs of almost everyone are doxastically unjustified. Another problem with the classical foundationalist proposal is that it is actually a sneaky form of externalism. Seemings do not wear their infallibility on their faces. The property of guaranteeing the truth of the beliefs based on it is not directly accessible to our automatic processing systems. I am not denying that infallibility is discernable a priori; let’s grant that it is. The point is that we only become aware that a seeming is infallible through reflection. We have to believe that the seeming is infallible to respond to it as such. However, this makes infallibility an external property. Accordingly, it has no bearing on a seeming’s foundational status. Classifying the classical foundational proposal as externalist, though ironic, is not as odd as it initially seems. Classical foundationalism is essentially akin to disjunctivism about perception—​ only genuine and direct awareness of a fact can immediately justify belief—​the only difference is that classical foundationalism limits direct awareness to necessary truths

A Reidian Argument for Phenomenal Conservatism, Part II  253 and the contents of our own minds. These restrictions make classical foundationalism appear internalist, since we are always looking inward, but it remains externalist in the deeper sense that the properties by virtue of which our mental states justify include features of our mental states that are not directly accessible to us (though we can often become aware of them on reflection). It is this external connection to reality that gives classical foundationalism its guarantee of truth, at least with respect to the foundations of our noetic structures. These considerations are sufficient to dismiss the classical foundationalist proposal. It nevertheless retains a certain allure. To break the spell, we must reflect further on the motivations underlying classical foundationalism. There is an important lesson to be learned by doing so. As I noted above, classical foundationalism restricts implicit trust to infallible seemings in an attempt to preserve an effective connection with truth. Seen rightly, this limitation of trust is in fact an attempt to do away with the need for trust altogether. I don’t need to trust that a seeming is reliable if I’m guaranteed that it is. It’s an enticing illusion, but it remains just that—​an illusion. We get no guarantees, at least not for anything beyond the most meager of foundations. As Richard Foley writes: [T]‌he reality of your intellectual life is that like the rest of us, you are working without a net. Regardless of how you marshal your cognitive resources, you are not going to have non-​question-​begging guarantees of truth or reliability. No procedure, no amount of reflection, no amount of evidence gathering can guarantee that you won’t fall into error and perhaps even great error. (Foley 1993, 135; see also 1993, ­chapter 2 and 2001, ­chapter 1) This is one of the key lessons of early modern epistemology and the radical doubt of Descartes, a lesson that Reid learned and accepted.4 He writes, “If we are deceived in [our natural beliefs], we are deceived by Him that made us, and there is no remedy” (Reid [1764] 1997, 72). The reality of our intellectual lives belies not only Descartes’s dream that all justified beliefs come with guarantees of truth, but also the more modest dream that all justified beliefs come with guarantees of likely truth. To be sure, all justified beliefs are “objectively probable” on classical foundationalism in the sense that they bear certain objective logical relations to the foundations, but this sort of objective probability carries no guarantee of external probability (i.e., reliability). The objectively probable beliefs of classical foundationalism can not only be wrong, they can be way wrong. So wrong that there are no nearby possible worlds in which they are right (Gage and McAllister 2020a, §1). So, when classical foundationalism touts itself as maintaining an “objective connection to truth”—​one that

254  A Reidian Argument for Phenomenal Conservatism, Part II moderate forms of foundationalism cannot preserve—​it falsely postures as though it can rule out being radically mistaken. In the end, the allure of classical foundationalism largely rests on the misbegotten hope that, by following its advice, we can ensure that all of our beliefs are reliable. Once we own that there are no guarantees—​at least none that are within our power to secure—​the attraction should fade. The lack of guarantees is hard to accept, however, especially for us epistemologists. We don’t want to rely on trust. There is a choice to be made here. We can rage against the lack of guarantees, or we can follow Reid and embrace it. Wolterstoff calls this humble acceptance of our limitations “Reidian piety.” He explains: Ungrounded trust, trust without reasons for trusting, that’s what is deepest in Reidian piety. Though that’s not putting it quite right. According to the Reidian, that’s what’s deep in in the piety of all humans. What’s deepest in Reidian piety, is acknowledging that fact, and acknowledging the darkness which that fact implies, and not railing against the mystery but accepting it humbly and gratefully. (Wolterstoff 2001, 261) In the spirit of such Reidian piety, Foley writes the following: Our lack of non-​question-​begging guarantees is not a failing that needs to be corrected. It is a reality that needs to be acknowledged. We must acknowledge our vulnerability to error, and acknowledge also that inquiry also involves a substantial element of trust in our own intellectual faculties and in the opinions they generate, the need for which cannot be eliminated by further inquiry. Significant inquiry requires an equally significant leap of intellectual faith. (Foley 2001, 19–​20) I said above that classical foundationalism tries to eliminate the need for self-​trust or, as Foley calls it, intellectual faith. Once we have piously embraced that such faith is ineliminable, we can see that there is no compelling reason to retain classical foundationalism’s demand for infallibility. 9.3.2  Propositional Content

Could differences in propositional content—​ the fact that the seeming is about this sort of thing rather than that sort of thing—​provide some basis for discrimination? In answering this question, we must exclude any property whose relevance depends on what we have learned about that property. For example, we have learned that our earliest memories are

A Reidian Argument for Phenomenal Conservatism, Part II  255 notoriously unreliable, often being completely fabricated (Akhtar et al. 2018). Accordingly, we have an undermining defeater for any memorial seeming that pertains to the earliest years of one’s life (particularly before the age of three). Nevertheless, such considerations are irrelevant to whether those memorial seemings permit belief in the state of empty symmetrical evidence where we have no evidence of their unreliability. Note that this applies even to things we have learned about our seemings through a priori reflection. On this basis, I would submit that differences in propositional content cannot serve as a principle for distinguishing foundational from non-​ foundational seemings. Obviously, which propositional content a seeming has will affect what it is foundational evidence for, but it cannot disqualify that seeming from being foundational evidence altogether. Again, I am not denying that we can and often do learn that seemings with certain kinds of content are less reliable than others. Such considerations do not, however, affect whether seemings have the potential to immediately justify. 9.3.3  Paired Appearances

Some seemings present alongside other mental states. A common example is that perceptual seemings are always, or nearly always, accompanied by states with sensory phenomenology (although it is controversial whether that sensory phenomenology belongs to a state distinct from the seemings themselves). McCain and Moretti call such seemings “paired appearances.” Could the fact that a seeming does or does not coincide with some paired state affect its foundational status? Paired states cannot have any effect on whether a seeming is foundational. Whether a seeming can immediately justify depends on whether it permits belief in and of itself, independently of any other mental states that might boost (or weaken) our epistemic position. Paired states can, of course, be relevant to the seeming’s ultimate justificatory potential. The presence or absence of paired states can give us indication of whether the seeming is reliable or not. But that is simply beside the point as concerns the ability of a seeming to justify considered in and of itself. 9.3.4  Felt Evidential Insensitivity

This brief discussion of paired appearances prepares us to consider a proposal from Berit Brogaard. Brogaard argues that whether a seeming displays felt evidential insensitivity is crucial to its ability to lend immediate justificatory support (Brogaard 2018). Brogaard herself thinks that seemings without this felt insensitivity are not foundational evidence at all, while others like McCain and Moretti think that lacking felt insensitivity

256  A Reidian Argument for Phenomenal Conservatism, Part II merely diminishes the seeming’s foundational evidential value (McCain and Moretti 2021, Ch. 3, §3.4). Before evaluating these proposals, we must get clear on the nature of felt evidential insensitivity. A seeming displays felt evidential insensitivity when one has a feeling, or apparent awareness, that it won’t go away with the discovery of further evidence.5 Not all seemings feel this way. For instance, it may seem to me that I told my wife about the dinner party on Friday. However, this seeming feels flimsy. I recognize that if my wife were to deny being told about the party, protesting that this is the first she is hearing about it, my seeming would likely vanish. In contrast, consider the seeming that 1 ≠ 2. This seeming feels resilient. Even if someone were to give me a plausible-​looking “proof” that 1=​2, it feels like my original seeming wouldn’t change—​the conclusion would remain as counterintuitive as ever.6 That’s felt insensitivity to the evidence. Whether a seeming has this felt insensitivity seems epistemically relevant all things considered. As McCain and Moretti point out, “felt evidence insensitivity can give S reason to think that the appearance is a genuine experience imposed upon her by the world, rather than, say, merely something that S already believes or an impression which simply comes to her mind” (McCain and Moretti 2021, 75). I do not deny that felt insensitivity can affect the ultimate evidential value of the seemings that do or do not enjoy it. Nevertheless, it does not affect the foundational status of those seemings for the reasons given in the last two sections. That is to say the feeling of evidential insensitivity either (i) belongs to a special kind of paired state or (ii) only becomes epistemically relevant in the context of background evidence. Regarding the first option, it seems that the feeling of evidential insensitivity only shows up when we reflect on our original seeming and become apparently aware of its dispositional qualities. If this is correct, then that feeling of insensitivity does not belong to the seeming itself but to our second-​ order reflective state. McCain and Moretti concur, suggesting that “felt evidential sensitivity is in fact just an additional appearance” concerning one’s initial, first-​order seeming (McCain and Moretti 2021, 78, ft. 55). Brogaard appears to resist this characterization of felt evidential insensitivity. She specifically rejects the proposal that “evidence insensitivity [is] a metacognitive feeling directed toward the seemings,” insisting instead that it is “a subjective property of the phenomenology of seemings” (Brogaard 2018, 60).7 I find this suggestion implausible. Returning to the intuition that 1 ≠ 2, I don’t find anything in the phenomenology of the intuition itself that we could fairly describe as a feeling of evidential insensitivity. The seeming is especially forceful—​it feels very strongly to be true—​but that’s not quite the same thing as being resilient in the face of counterevidence. Only when I think about that seemings and its forcefulness do I gain (in

A Reidian Argument for Phenomenal Conservatism, Part II  257 Brogaard’s words) “an apparent awareness of the epistemic robustness of one’s experience” (Brogaard 2018, 61). So, most plausibly, felt evidential insensitivity belongs to a paired state, which accompanies some seemings but not others. And we have already established that paired states cannot affect whether seemings are foundational or not. Even if we grant that felt evidential sensitivity does belong to the phenomenology of our first-​order seemings, it still seems irrelevant to their foundational status. That is because the significance of that feeling can only be appreciated with the aid of background evidence, and such things do not affect a seeming’s potential for immediate justification. To better see the point, return to McCain and Moretti’s explanation for why felt evidential insensitivity is relevant—​namely, it is the mark of a genuine experience rather than one that is made up. Those in a state of empty symmetrical evidence are not in a position to appreciate this explanation. To do that, we must have some understanding of what genuine and fabricated experiences are like and how to tell the difference. But if we cannot appreciate this explanation, then felt insensitivity does not indicate to us (in that specific context) the truth of the seeming’s content. Thus, it will not bring belief into line with our duty to attain truth and avoid falsehood in the state of empty symmetrical evidence. In sum, felt evidential sensitivity makes no difference to a seeming’s immediate justificatory potential whether we consider it as belonging to a paired state or to the seeming state itself. 9.3.5  Sensory Phenomenology

Could the presence or absence of sensory phenomenology determine which seemings count as foundational and which do not? Perceptual dogmatists, for instance, might claim that such phenomenology is an integral part of what allows seemings to justify. Seemings that do not possess such phenomenology, therefore, are not foundational. The most obvious problem with this proposal is that we need a lot more than just perceptual seemings to be foundational in order to avoid skepticism, including ones that lack sensory phenomenology. Not only this, but many rational and introspective seemings that lack sensory phenomenology are among the very strongest candidates for foundational evidence. The arguments in the previous chapter support the ability of these rational and introspective seemings to immediately justify as well or better than they do perceptual seemings. Furthermore, it is hard to see any principled reason why sensory phenomenology specifically would be necessary (see also Kriegel 2021). Why couldn’t whatever it is that sensory phenomenology accomplishes be realized in different ways in introspection or rational intuition?

258  A Reidian Argument for Phenomenal Conservatism, Part II At this point, we might retreat to the more moderate proposal that among perceptual seemings there is something about sensory phenomenology that allows them to immediately justify. Hence, perceptual seemings that lack this sort of phenomenology (assuming that’s even possible) are not foundational. Before we evaluate this proposal, we must clarify how exactly this sensory phenomenology is related to perceptual seemings. One option is that sensory phenomenology belongs to paired states—​ namely, sensations or sensory appearances—​that accompany, but remain distinct from, the perceptual seemings themselves. If this is correct, then sensory phenomenology cannot affect a seeming’s foundational status for the reasons discussed in §9.3.3. When it comes to immediate justification, the seeming must stand or fall on its own. Obviously, if a perceptual seeming lacked sensory phenomenology in normal circumstances—​as some interpret cases of blindsight (Tucker 2010, 530–​531)—​then that should give us serious doubts about the reliability of the seeming. But it won’t affect how we should respond to that seeming in the state of empty symmetrical evidence, where all evidence about what is normal is suspended. Another option, however, is that sensory phenomenology belongs to perceptual seemings themselves.8 In this case, the question becomes whether that sensory phenomenology could make a difference to those seemings’ ability to justify belief in the state of empty symmetrical evidence. The main obstacle here is that the significance of sensory phenomenology seems appreciable only with the benefit of background evidence. Take, for example, a standard green-​and-​brown phenomenal array that we enjoy while looking at a pine tree from 20 feet away in good lighting. The fact that pine trees look this way is something we must learn; and only after learning this does that sensory phenomenology indicate to us that a pine tree is present. That phenomenal array would be meaningless, for instance, to an infant who knew nothing about trees or what it felt like to look at them. Well, put us in the state of empty symmetrical evidence and we are a lot like that infant. We would not be able to appreciate the significance of the pine sensory phenomenology or any other. But then the presence or absence of that sensory phenomenology wouldn’t make any difference to what we were justified in believing in the state of empty symmetrical evidence. The only possible exception would be if we were somehow able to recognize the significance of certain phenomenology directly, apart from any background evidence. This seems implausible in most instances—​we cannot “just see” that the state of affairs presented through the pine tree array corresponds to the thought there’s a pine tree—​but perhaps there are some situations in which it may be possible. For example, we might demonstrate toward whatever state of affairs is (re)presented through sensory

A Reidian Argument for Phenomenal Conservatism, Part II  259 phenomenology and think that exists. Plausibly, we don’t need background evidence to appreciate the connection there. This last case is an example of what McCain and Moretti (2021, 78–​ 80) call “presentational appearances,” or what I will call “presentational seemings.” Presentational seemings make it feel as though we are being presented with a truth-​maker for the content of our seemings, like the very fact at issue is directly presented to our conscious awareness (Chudnoff 2011, §3). Here we hit upon a suggestion with some initial plausibility: perhaps the only perceptual seemings that can immediately justify are those with this distinctive sort of “presentational phenomenology.” This proposal also accommodates our earlier observations about rational and introspective seemings. Plausibly, such seemings have their own brands of presentational phenomenology, which is what allows at least some of them to immediately justify. We will consider the proposal that only seemings with presentational phenomenology are foundational directly. What our discussion in this section has shown us, however, is that the presence or absence of sensory phenomenology cannot have any bearing on whether a seeming is foundational except perhaps when that phenomenology is part of what qualifies it as a presentational seeming. 9.3.6  Presentational Phenomenology

Several philosophers have suggested that only presentational seemings count as foundational evidence (Berghofer 2020, Chudnoff 2018, Reiland 2015). Chudnoff, for instance, proposes the following constraint on immediate justification: Presentational Constraint—​If a perceptual experience can immediately justify believing that p, then it must both make it seem as if p and make it seem as if it makes you aware of a truth-​maker for p. (Chudnoff 2018, 288) Strictly speaking, this constraint doesn’t entail that non-​ presentational seemings aren’t foundational at all—​they could simply be foundational evidence that is too weak to immediately justify full belief—​but we will treat it as denying foundational status to non-​presentational seemings altogether. (If this extension is considered implausible by Chudnoff or others, then all the better for PC.) I will argue against that position here. Others make the more moderate claim that seemings with presentational phenomenology are always stronger foundational evidence than those without it (McCain and Moretti 2021). I will argue instead that presentational phenomenology can and often does increase the level of evidential support that a seeming lends to its content, but not necessarily. In sum, the lack of presentational

260  A Reidian Argument for Phenomenal Conservatism, Part II phenomenology will usually place an upward limit on the level of immediate justification that a seeming is capable of providing, but it does not disqualify a seeming as foundational evidence entirely. Thus, presentational phenomenology provides no means of escaping an unrestricted version of phenomenal conservatism. Presentational seemings are essentially the non-​factive versions of triple acquaintance states. We feel directly aware not only of a thought and its correspondence to reality (as in all seemings) but also of the fact that our thought describes. To give an example, imagine looking at our desks and seeing a cup of coffee. Our perception, let us assume, makes us directly aware of this state of affairs. We go on to form the thought there is a cup of coffee on my desk, and the content of that thought feels as though it accurately represents the state of affairs given through perception. This is a presentational seeming.9 To compare, imagine someone asks us whether the Earth was created 100 years ago and made to look much older than it is. It immediately seems that the answer is no. That is, we form the thought the Earth was not created 100 years ago with the appearance of age, and that content feels as though it accurately represents the way things really are. It does not, however, feel as though any of the corresponding states of affairs is being given to our conscious awareness. All we have is the general sense that whatever states of affairs are out there correspond to the content of our thought. My position is that presentational seemings will generally be stronger foundational evidence than non-​ presentational ones. This is because seemings with presentational phenomenology are usually more forceful than those without it. There are exceptions, however. Say we’re driving on a long stretch of highway, and it seems that there is a car parked along the side of the road a mile or two ahead of us, right at the edge of our visual capacities. That presentational seeming may feel only mildly true. Compare that to the seeming that the earth was not created 100 years ago with the appearance of age. It seems imminently plausible that the latter seeming (which for me is quite strong) will be more forceful than the former. Regardless, in most ordinary cases, presentational seemings are likely to be especially forceful. After all, if you are directly aware of both your thought and the state of affairs it represents, the felt correspondence between them may be particularly strong. You could even attend back and forth from one to the other, strengthening the feeling that your thought really does correspond to the way things actually are. Hence, presentational seemings are liable to be among the most forceful of seemings. Such a position is, of course, perfectly compatible with PC. PC states that a seeming is foundational evidence for p to the extent that p seems true. In other words, the more forceful the seeming, the stronger the foundational

A Reidian Argument for Phenomenal Conservatism, Part II  261 evidence. This is standard among phenomenal conservatives more generally: the degree of justification that a seeming confers correlates with its forcefulness (e.g., Huemer 2007, ft 1). What isn’t compatible with PC is the claim that non-​presentational seemings have no ability to immediately justify at all, as the presentational constraint maintains. There are at least two serious problems with this constraint. First, we said in §9.1 of this chapter (building off the arguments in Chapter 8, §8.2) that to avoid various forms of skepticism, we need to admit a great many seemings as foundational evidence, and this will almost certainly include seemings without presentational phenomenology. Consider hinge propositions, which we discussed back in Chapter 7, §7.3. Examples include the reliability of our cognitive faculties, the uniformity of nature, and the stability of meaning over time. Any attempt to inferentially justify these propositions is going to embroil us in circularity. And even if inferential justification could somehow be given, our beliefs in such propositions would not be properly based on that justification, being based instead on their apparent truth. Thus, hinge propositions, and our beliefs in them, must be immediately justified by seemings. But many of these seemings lack presentational phenomenology. When we intuit that our faculties are reliable or that nature is uniform, we’re not directly presented with any truth-​ maker for those propositions. If extensive skepticism is to be avoided then, some non-​presentational seemings must count as foundational evidence. Other skeptical problems arise in the case of perception specifically (see also Harrison 2021, §3.2). Arguably, all perceptual seemings involve some kind of presentational phenomenology; however, it is often the case that the specific state of affairs presented in perception does not match the proposition that seems true and that is subsequently believed. To give an example, imagine we see a dog standing behind a pole (Chudnoff 2018, §3). Chudnoff says the front and back half of the dog are directly presented to us, but not its middle portion (which is blocked by the pole). Say, however, that we have a perceptual seeming with the content there is an (entire) dog behind that pole, and we come to believe that proposition. Is our belief immediately justified? According to the presentational constraint, the answer is no. Immediate justification requires that we be apparently presented with a truth-​maker for our belief. But that isn’t the case here. To have that, we’d have to see the entire dog. Thus, while we have a presentational seeming, it lacks presentational phenomenology with respect to the specific proposition it’s supposed to immediately justify. Hence, given the presentational constraint, it would be unable to immediately justify our perceptual belief. This is bad news. Very often the propositional content of our perceptual seemings, and in turn our perceptual beliefs, outruns the facts directly

262  A Reidian Argument for Phenomenal Conservatism, Part II presented to us through our sensory states. Descartes provides a wonderful example: But then were I perchance to look out my window and observe men crossing the square, I would ordinarily say I see the men themselves… But what do I see aside from hat and clothes, which could conceal automata? (Descartes [1641] 2006, 17–​18) When we look into the square, only clothes are directly presented to us (the clothes cover the people just as the pole covered the middle portion of the dog). So, we would not gain any immediate justification for believing that there are people crossing the square from that experience. Examples can easily be multiplied. For instance, when we look at a car on the road ahead, we aren’t directly presented with a truth-​maker for the proposition a car is up ahead, but only the backside of a car is up ahead. Or when we hear a dog barking, we are not directly presented with a truth-​maker for the proposition there’s a dog, but only there’s a dog bark. Most disturbingly, when looking someone square in the face, we are not directly presented with a truth-​maker for the proposition there is a person (for a person has a conscious mind, and we have no direct awareness of that), but only there is a body (see Chudnoff 2018, sections 4 and 5). Arguably, not even the truth-​makers for these things are presented, and we will be forced to retreat back to indirect realism—​a position which has its own well-​known skeptical problems. In any case, none of these ordinary perceptual beliefs about people or cars or dogs are immediately justified according to the presentational constraint. Chudnoff will retort that these perceptual beliefs are nonetheless mediately justified by background beliefs. Using the example of other minds, Chudnoff says that all we are immediately justified in believing is that there is a bodily manifestation of a mental state. This premise, in conjunction with our justified background beliefs, gives us mediate justification for believing that there is a person before us. In this, Chudnoff effectively endorses Descartes’s own conclusion: But what do I see aside from hat and clothes, which could conceal automata? Yet I judge them to be men. Thus what I thought I had seen with my eyes, I actually grasped solely with the faculty of judgment, which is in my mind. (Descartes [1641] 2006, 18) There are a couple of issues with this reply. One is that ordinary perceptual beliefs like the ones above will not generally be based on that mediate

A Reidian Argument for Phenomenal Conservatism, Part II  263 justification, even if it is available. Such beliefs are often basic, being based directly on a perceptual seeming. We see a human body in front of us, it seems that there’s a person there, and we immediately believe it. At no point are we thinking about bodily manifestations of mental states. (Most people will lack that concept entirely.) In which case, many of our ordinary perceptual beliefs will be doxastically unjustified. To make matters worse, we should not be so confident that mediate justification is always available. This is especially true when we start thinking about how we came to possess these justified background beliefs that we are supposed to use in inferring other minds. How did we establish the link between bodily behaviors and conscious minds in the first place if we are never immediately justified in believing that the body before us is a conscious person? Setting that general worry aside, there are additional reasons specific to Chudnoff’s presentational constraint to doubt the availability of mediate justification. Notice that Chudnoff’s presentational constraint demands two things: (i) that it seem as if p and (ii) that one seemingly be aware of a truth-​maker for p. It is perhaps true that, when looking at someone, our sensory experience apparently presents us with a truth-​maker for the proposition there is a bodily manifestation of a mental state; nevertheless, it will not typically seem to us that this is true. That is to say the propositional content of our perceptual seeming is just there is a person. The concept of a bodily manifestation of a mental state doesn’t come to mind at all. It follows that, in these ordinary cases, we are not immediately justified in believing that there is a bodily manifestation of a mental state. What’s missing is not awareness of the truth-​maker, but the first condition of the presentational constraint: we don’t have a seeming with that specific content. The same line of reasoning rules out our having immediate justification for any proposition on the basis of that sensory experience since (we can stipulate) the content of our perceptual seeming is just there is a person and nothing else. But we need immediate justification for some proposition—​if only that things appear thus and so—​if we are to build up mediate justification for the existence of another person. So the presentational constraint comes with serious skeptical baggage. A second problem is that forcefulness can be found apart from presentational phenomenology, and forcefulness without presentation remains sufficient to permit belief (or mere positive credence) in the state of empty symmetrical evidence. Notice that the intuitive case for Minimal-​ PC (Chapter 8, §8.4) did not rely on presentational phenomenology to make its case. It was forcefulness, the felt awareness of a proposition’s truth, that brought belief into line with our epistemic duties. That feeling alone was enough to make belief, or at least some positive credence, a “good bet” with respect to attaining truth and avoiding falsehood. Presentational

264  A Reidian Argument for Phenomenal Conservatism, Part II phenomenology is not an essential feature of this explanation, and so is not required of all foundational evidence. 9.3.7 Forcefulness

Thus far, we’ve discovered only one legitimate basis for discrimination with respect to foundational status—​namely, forcefulness. Presentational phenomenology also had some effect, but only by altering the seeming’s forcefulness; thus, forcefulness remained the fundamental difference-​ maker. Forcefulness can help explain why some seemings count as stronger foundational evidence than others—​the stronger the feel of truth, the stronger the evidence—​but it provides no basis for saying that some seemings are not foundational at all. To the contrary, if the content of a seeming feels true to any degree (which is true by definition), then it seems capable of immediately justifying to precisely that extent. This final observation raises a challenge to McCain and Moretti’s treatment of “mere seemings.” Mere seemings are those that lack both presentational phenomenology and paired states. McCain and Moretti agree that mere seemings are foundational evidence, but they consider the evidence they provide to be “vanishingly close to zero” (McCain and Moretti 2021, 78). In contrast, what our discussion suggests is that the evidential value of mere seemings corresponds to their level of forcefulness; and while generally speaking, mere seemings may be on the less forceful side, there is no in-​principle reason why mere seemings cannot be especially forceful. Indeed, seemings in hinge propositions are plausibly among the strongest we have, but many of them would qualify as mere seemings. Accordingly, there is no reason why mere seemings should be unable to immediately justify full belief in their content. 9.4  Seeming Parity Our task in this chapter has been to find a principled distinction between different kinds of seemings—​one that could account for why some seemings are foundational evidence while others are not. Without such a principled distinction, the attempt to differentiate between foundational and non-​foundational seemings is inconsistent and arbitrary: why should one qualify but not the other if there is no relevant difference between them? We looked at all of the most prominent proposals. Almost all of these were of entirely the wrong sort. They either appealed to external features of our seemings or else to internal features of other states distinct from our seemings. And when they finally did appeal to internal features of our seemings, those features didn’t make any difference to what we were permitted to believe apart from background evidence, making them irrelevant

A Reidian Argument for Phenomenal Conservatism, Part II  265 to immediate justification. The only exception was forcefulness (and those features affecting forcefulness like presentational phenomenology), but differences in forcefulness only account for different levels of foundational evidence and provide no basis for denying some seemings status as foundational evidence altogether. What other basis for discrimination could be given? It’s hard to think of remaining candidates.10 And those candidates we did consider—​including all of the major ones on offer—​did not fare well. At this point, the most reasonable conclusion to draw is that there is no legitimate distinction between those seemings that are foundational and those that are not. Nevertheless, we needn’t rely entirely on process of elimination. We have already come across several positive reasons for thinking that all seemings are foundational. As discussed earlier, the argument in Chapter 8, §8.4, built the intuitive case that forcefulness, coupled with propositional content, is sufficient for motivating belief (or positive credence) in the state of empty symmetrical evidence. Adding to the case, there are clear instances of foundational seemings where their foundational status is attributable solely to the forceful presentation of propositional content. Rational seemings provide some of the strongest examples.11 Seemings in hinge propositions provide more. But if forcefulness and propositional content are sufficient in these cases to immediately justify (and it seems they must be on pain of skepticism), then why not in all of them? In closing, Reid issues the following challenge to those who trust one kind of natural belief but not others: I should beg to know, why they believe the existence of their impressions and ideas. The true reason I take to be, because they cannot help it; and the same reason will lead them to believe many other things. (Reid [1764] 1997, 71) Though I disagree with Reid on the details, the spirit of his challenge remains incisive. Why do people believe that they are in pain, or that p ⊃ p, or that the universe is more than five minutes old? The true reason I take to be, because those things seem true; and the same reason will lead them to believe many other things as well. This concludes my defense not only of phenomenal conservatism, but of seemings foundationalism as a whole: all and only seemings are foundational evidence and, thus, sources of immediate justification. All other justification originates from this foundation of seemings. Of course, seemings foundationalism has no pretensions of being a comprehensive theory of justification, as I made clear from the beginning. It is

266  A Reidian Argument for Phenomenal Conservatism, Part II nonetheless plain that our inquiry into the foundations of justification will have profound implications for epistemology more generally—​implications that I have begun to draw out elsewhere.12 Let us not, however, move too quickly past the lessons of our present inquiry. In particular, we learned that the business of rational belief formation rests on a foundation of trust—​trust in what one’s experiences present as true. This is a trust that must be given implicitly and automatically. It cannot be given only to those appearances which we have reason to believe are trustworthy, for there could be no reasons available to us were trust not already granted. This same implicit trust must, therefore, be given to everything that seems true. For in the one case trust is granted automatically and without independent verification, so why not in the other? Insofar as both seem true, there is nothing to tell between them—​not until we can rationally reflect on their differences which, again, is only possible as a result of such trust already being given. The cost of withholding trust, or placing preconditions on its bestowal, is not only skepticism but the complete abdication of our identity as rational agents. Perhaps the final lesson to be learned is that rationality just is the movement of the intellect in accordance with how things appear. Notes 1 The claim is not that we must possess a rational explanation for why one seeming differs from another (in the form of beliefs about those seemings), but only that there must be some explanation for why one seeming differs from another. One seeming must possess some epistemically relevant property that the other does not. The argument here does not, then, depend on the Rational Explanation Thesis (see Tucker forthcoming, ft. 15). 2 We could also be justifiably confident that the seeming possesses it but, as I have done throughout this inquiry, I will just talk about beliefs for convenience. 3 See Huemer (2013a) for compelling examples of badly cognitively penetrated beliefs that still justify. 4 Indeed, Reid arguably goes too far in embracing the fallibility of our beliefs. He argues that all of our faculties are “limited and imperfect,” and thus that none of their deliverances are guaranteed to be true. This is part of why Reid maintains parity between all natural beliefs, no matter which faculty produced them: Why, Sir, should I believe the faculty of reason more than that of perception; they came both out of the same shop, and were made by the same artist; and if he put one piece of false ware into my hands, what should hinder him from putting another? (Reid [1764] 1997, 169) I find it plausible, however, that there are certain incorrigible beliefs, produced by reason and introspection, that do in fact come with guarantees of truth. I just

A Reidian Argument for Phenomenal Conservatism, Part II  267 don’t think such guarantees (of truth or even reliability) can be extended significantly beyond those. 5 Brogaard formerly focused on actual evidential insensitivity (see, e.g., Brogaard 2013), which is an external property and so cannot determine foundational status (see §9.2 of this chapter); however, she now focuses on felt evidential insensitivity, which is an internal property (see Brogaard and Gatzia 2018). 6 For an excellent example of such a proof, see Spencer (1998). 7 I concede that evidential insensitivity itself is a property that belongs to our (first-​order) seemings, which is technically all that this quote claims. My position is that the feeling of evidential sensitivity does not belong to our (first-​order) seemings but to a (second-​order) state in which we reflect on those seemings. This is why I say that Brogaard “appears to” resist my characterization. She may not. 8 I argued in Chapter 4, §4.3.1 that while sensory phenomenology belongs most properly to sensory appearances, which are distinct from perceptual seemings, there can also be cases where perceptual seemings are partially constituted by sensory appearances. An example would be when our perceptual seemings demonstrate toward the states of affairs (re)presented by those sensory appearances. 9 It isn’t a presentational seeming that we could form in the state of empty symmetrical evidence though, per our discussion at the end of §9.3.5. We have to know what coffee cups look like before we could feel aware of the correspondence between our thought’s content and the state of affairs (re)presented through that phenomenal array. This might make us wonder whether this particular sort of presentational phenomenology can affect a seeming’s foundational status. I will, however, set this issue aside here. 10 This is especially so given phenomenalism—​the position touted in Chapter 1 as a complement to casual internalism. On this view, the only properties relevant vis-​à-​vis justification are phenomenal properties. For our present purposes, then, the detractor of PC would need to identify some specific phenomenal difference between types of seemings that allows the one the serve as foundational evidence but not the other. We’ve already surveyed the leading candidates here (e.g., sensory phenomenology, presentational phenomenology, felt evidential insensitivity). Additional candidates are not readily forthcoming. 11 The case might be made that certain rational seemings have presentational phenomenology (Chudnoff 2013), but I am skeptical that this can be maintained for all of them. 12 For instance, I apply seemings foundationalism to religious epistemology in McAllister (2018; 2019; 2020; 2021a; 2021b), McAllister and Dougherty (2019), and Gage and McAllister (2020b). Religious epistemology serves as a convenient microcosm for exploring broader epistemological issues such as knowledge, disagreement, naturalistic epistemology, and more.

References Akhtar, Shazia, Lucy V. Justice, Catriona M. Morrison, and Martin A. Conway. 2018. “Fictional First Memories.” Psychological Science 29 (10): 1612–​1619.

268  A Reidian Argument for Phenomenal Conservatism, Part II Berghofer, Philipp. 2020. “Towards a Phenomenological Conception of Experiential Justification.” Synthese 197 (1): 155–​183. Bergmann, Michael. 2013. “Externalist Justification and the Role of Seemings.” Philosophical Studies 166 (1): 163–​184. Brogaard, Berit. 2013. “Phenomenal Seemings and Sensible Dogmatism.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 270–​289. Oxford: Oxford University Press. —​—​—​. 2018. “Phenomenal Dogmatism, Seeming Evidentialism and Inferential Justification.” In Believing in Accordance with the Evidence: New Essays on Evidentialism, edited by McCain Kevin, 53–​ 67. Synthese Library. Cham, CH: Springer International. Brogaard, Berit, and Dimitria Gatzia. 2018. “The Real Epistemic Significance of Perceptual Learning.” Inquiry 61 (5–​6): 543–​558. Chudnoff, Elijah. 2011. “The Nature of Intuitive Justification.” Philosophical Studies 153 (2): 313–​333. —​—​—​. 2013. “Awareness of Abstract Objects.” Noûs 47 (4): 706–​726. —​ —​ —​ . 2018. “Epistemic Elitism and Other Minds.” Philosophy and Phenomenological Research 2: 276–​298. Cohen, Stewart. 1984. “Justification and Truth.” Philosophical Studies 46 (3): 279–​295. DePoe, John. 2021. “Justification by Acquaintance.” Synthese 199: 7555–​7573. Descartes, René. [1641] 2006. Meditations, Objections, and Replies. Translated by Roger Ariew and Donald A. Cress. Indianapolis, IN: Hackett Publishing. Foley, Richard. 1993. Working Without a Net: A Study of Egocentric Epistemology. Oxford: Oxford University Press. —​—​—​. 2001. Intellectual Trust in Oneself and Others. Cambridge: Cambridge University Press. Gage, Logan Paul, and Blake McAllister. 2020a. “A Phenomenal Conservative Response to Classical Evidentialism.” In Debating Christian Religious Epistemology: An Introduction to Five Views on the Knowledge of God, edited by John M. DePoe and Tyler McNabb, 34–​38. London: Bloomsbury Academic. —​ —​ —​ . 2020b. “The Phenomenal Conservative Approach to Religious Epistemology.” In Debating Christian Religious Epistemology: An Introduction to Five Views on the Knowledge of God, edited by John M. DePoe and Tyler McNabb, 61–​81. London, UK: Bloomsbury Academic. Harrison, Eilidh. 2021. “Seemings, Truth-Makers, and Epistemic Justification.” Synthese 198 (6): 5689–​5708. Huemer, Michael. 2006. “Phenomenal Conservatism and the Internalist Intuition.” American Philosophical Quarterly 43 (2): 147–​158. —​ —​ —​ . 2007. “Compassionate Phenomenal Conservatism.” Philosophy and Phenomenological Research 74 (1): 30–​55. —​—​—​. 2013a. “Epistemological Asymmetries Between Belief and Experience.” Philosophical Studies 162 (3): 741–​748. —​ —​ —​ . 2013b. “Phenomenal Conservatism Über Alles.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 328–​350. Oxford: Oxford University Press. Kriegel, Uriah. 2021. “The Structure of Phenomenal Justification.” Australasian Journal of Philosophy.

A Reidian Argument for Phenomenal Conservatism, Part II  269 Kvanvig, Jonathan L. 2014. Rationality and Reflection: How to Think about What to Think. Oxford: Oxford University Press. Lehrer, Keith, and Stewart Cohen. 1983. “Justification, Truth, and Coherence.” Synthese 55 (2): 191–​207. Markie, Peter. 2005. “The Mystery of Direct Perceptual Justification.” Philosophical Studies 126 (3): 347–​373. —​—​—​. 2013. “Searching for True Dogmatism.” In Seemings and Justification: New Essays on Dogmatism and Phenomenal Conservatism, edited by Chris Tucker, 248–​269. Oxford: Oxford University Press. McAllister, Blake. 2018. “The Perspective of Faith: It’s Nature and Epistemic Implications.” American Catholic Philosophical Quarterly 92 (3): 515–​533. —​ —​ —​ . 2019. “Evidence Is Required for Religious Belief.” In Contemporary Debates in Philosophy of Religion, edited by Michael Peterson and Ray VanArragon, Second ed., 269–​278. Hoboken, NJ: Wiley-​Blackwell. —​ —​ —​ . 2020. “The Perspectival Problem of Evil.” Faith and Philosophy 37 (4): 421–​450. —​—​—​. 2021a. “The Partiality of Faith.” Australasian Philosophical Review 5 (1): 36–​45. —​—​—​. 2021b. “Understanding Moral Disagreement: A Christian Perspectivalist Approach.” Religions 12 (5): 318. McAllister, Blake, and Trent Dougherty. 2019. “Reforming Reformed Epistemology: A New Take on the Sensus Divinitatis.” Religious Studies 55 (4): 537–​557. McCain, Kevin, and Luca Moretti. 2021. Appearance and Explanation: Phenomenal Explanationism in Epistemology. Oxford: Oxford University Press. Moretti, Luca. 2020. Seemings and Epistemic Justification: How Appearances Justify Beliefs. New York: Springer. Reid, Thomas. [1764] 1997. An Inquiry into the Human Mind on the Principles of Common Sense, edited by Derek R. Brooks. University Park, PA: Pennsylvania State University Press. —​—​—​. [1785] 2002. Essays on the Intellectual Powers of Man, edited by Derek R. Brooks. University Park, PA: Pennsylvania State University Press. Reiland, Indrek. 2015. “Experience, Seemings, and Evidence.” Pacific Philosophical Quarterly 96 (4): 510–​534. Siegel, Susanna. 2012. “Cognitive Penetrability and Perceptual Justification.” Noûs 46 (2): 201–​222. Spencer, Philip. 1998. “1=​2: A Proof Using Beginning Algebra.” University of Toronto Mathematics Network. May 26, 1998. www.math.toro​nto.edu/​math​ net/​fals​ePro​ofs/​first1​eq2.html. Teng, Lu. 2018. “Is Phenomenal Force Sufficient for Immediate Perceptual Justification?” Synthese 195 (2): 637–​656. Tucker, Chris. 2010. “Why Open-​Minded People Should Endorse Dogmatism.” Philosophical Perspectives 24 (1): 529–​545. Tucker, Chris. Forthcoming. “Experience as Evidence.” In The Routledge Handbook of the Philosophy of Evidence, edited by Maria Lasonen-​Aarnio and Clayton M. Littlejohn. Abington, UK: Routledge. Wolterstorff, Nicholas. 2001. Thomas Reid and the Story of Epistemology. Cambridge: Cambridge University Press.

Index

Note: Endnotes are indicated by the page number followed by ‘n’ and the endnote number e.g., 20n1 refers to endnote 1 on page 20. academic skeptics 118–​19, 195, 211 accuracy conditions 110, 126, 169–​70 acquaintance, direct 128, 136n42, 165–​6; mere acquaintance 128, 171, 174–​6, 219–​21, 225; triple acquaintance 128–​30, 135n29, 160n21, 175–​80, 184, 187n12, 211–​12, 219–​21, 225, 260; see also classical foundationalism Adler, J. E. 192 Akhtar, Shazia 254 Aikin, Scott 155 Allen, James 118, 211 Alston, William 19–​20, 23, 45n3, 61–​2, 98n17, 212 appearances see seemings; sensory appearances Aquinas, Thomas 20, 160n10 assertiveness: doxastic assertiveness 126, 156–​8, 167; experiential assertiveness 126, 157–​8; assertive force 125–​8, 157; see also assertive force; forcefulness Aristotle 52, 144, 194–​5, 210–​11, 237 Audi, Robert 62, 134n29, 210 Augustine 20 Austin, J. L. 194 awareness, direct see direct acquaintance Bealer, George 132n1, 133n7, 210 Bedke, Matthew 132n1, 210 Beillard, Julien 132n1, 210 beliefs 167–​8

beliefs, bases of: basing relation 40, 46n20, 54, 145, 217; folk psychological explanations of belief 31–​3, 36–​8, 42–​3, 46n16, 179, 217, 240n6; non-​intellectualist belief formation 28–​30; proximate causes of belief 30–​3, 36–​8, 42–​3, 45n11, 46n16, 46n19, 46n20, 179, 217; rational belief forming dispositions 17, 24–​5, 29–​31, 33, 35, 36, 42–​3, 45n7, 46n20; seemings as 63–​4, 88, 217–​24, 230–​1, 233, 238 beliefs, types of: a priori beliefs 89, 220–​4; incorrigible beliefs 18, 78, 266n4; introspective beliefs 65, 215, 217, 219–​20, 222–​4; irresistible beliefs 202–​3, 239; memorial beliefs 197, 223–​4; mere beliefs 200, 205n3; natural beliefs 9, 11n4, 203, 237–​9, 241n14; non-​exempt beliefs 16–​17, 21, 25, 30, 36, 39–​40, 44; perceptual beliefs 9–​11, 29, 37–​9, 45n8, 46n19, 64, 78–​9, 81–​2, 87, 93, 119, 181, 221–​4, 238, 241n14 Bengson, John 124, 240n5 Berghofer, Philipp 210 Bergmann, Michael 10, 11n2, 34, 41, 42, 46n13, 46n21, 66, 84, 88, 97n6, 97n8, 97n10, 98n14, 99n20, 123, 132n1, 133n7, 134n21, 158n8, 210, 247 Berker, Selim 45n5 Bermúdez, José Luis 123

Index  271 blame, epistemic 5, 11n1, 11n4, 15–​16, 20, 22–​7, 44, 78–​9, 81, 84, 86, 104 Boespflug, Mark 9–​10, 11n4 Bonjour, Laurence 7, 18, 42, 45n4, 128, 135n29, 152–​4, 155–​8, 159n1, 159n2, 159n4, 160n19, 164, 165, 167, 169–​71, 174, 176, 186n7, 187n9, 212, 216, 239n1 Bonjour’s dilemma 7, 155–​9, 160n19, 164, 216 Boulter, Stephen 193, 194 Boyle, Matthew 24–​5 brain in a vat 16, 78–​9, 81–​3, 111, 218 Braun, David 186n3 Brewer, Bill 164–​6, 171–​2 Brogaard, Berit 132n1, 210, 255–​7, 267n5, 267n7 Campbell, Alexander 65 Campbell, John 123, 135n38, 166 Carneades 118, 211 Carruthers, Peter 21 Caston, Victor 119 Chesterton, G. K. 237, 240n8 Chisholm, Roderick 45n4, 97n8, 109, 133n8, 174, 190, 191, 209 Christensen, David 190 Chudnoff, Elijah 107, 117, 130, 132n1, 132n2, 132n5, 133n7, 134n16, 134n17, 210, 259–​64, 267n11 Cicero 118 circularity: begging the question 84; logical circularity 88, 98n12, 233; psychological circularity 88; source or epistemic circularity 88–​91, 93, 98n12, 98n14, 98n17, 98n18, 233; see also phenomenal conservatism (circularity or bootstrapping objection to) classical deontologism 5–​6, 13–​23, 44 cognitive penetration, problem of see phenomenal conservatism (problem of cognitive penetration) Cohen, Stewart 87–​9, 97n11, 250 coherentism 7, 98n18, 148–​55, 158–​9, 191–​2 commitment, rational 144, 229 common sense methodology 3, 85–​6, 97n8, 193–​5

common sense response to skepticism 6, 76–​86, 96n2, 195–​6, 213–​14 conceptual content 55–​6, 119, 123–​5, 130, 135n38, 145, 164–​78, 180–​2, 184, 186n2, 186n6, 222, 235, 240n5 Conee, Earl 22, 45n12, 51, 71n4, 109–​10, 112–​17, 132n4, 134n23, 175, 181 compliance with a norm 30, 36–​7, 39–​40 confidence, levels of see credences conformance to a norm 30, 35–​6 conscience, moral 127, 240n7 conservatism see epistemic conservatism; phenomenal conservatism constitutive rules of rationality see epistemic conservatism (Reid’s defense of); phenomenal conservatism (constitutive rules of rationality and); Thomas Reid (on constitutive principles of rationality) content externalism 46n17, 47n24 control 20–​5, 45n7, 202–​3 Conway, Martin 254 Coren, Daniel 199–​200, 205n3 credences 29, 51–​3, 55–​7, 68, 69, 70n1, 71n5, 79, 94–​5, 97n6, 112–​13, 143, 159n2, 174, 211, 229, 230, 233, 238, 252, 263–​5, 266n2 Cullison, Andrew 104, 108–​12, 124, 132n1, 133n7, 134n21 Dancy, Jonathan 159n9 Davidson, Donald 160n18, 169–​70, 172 dead-​ends, epistemic 229, 234, 238–​9, 204n12 defeaters 56–​7, 60, 63, 68, 69, 77, 79–​80, 92, 94, 95, 113, 191–​2, 196, 198–​9, 248, 249, 251, 254–​5 deontologism, epistemic 5, 13–​16, 22–​3, 51–​3 DePaul, Michael 225 DePoe, John 225, 252 Descartes, René 5, 10, 13–​22, 33–​4, 44, 65, 90–​1, 97n5, 97n9, 98n19, 119–​20, 195, 211, 213, 232–​3, 239n4, 252–​3, 262 Devitt, Michael 132n2

272 Index direct causal accessibility 30–​44, 54, 179, 247–​8 disjunctivism 37–​9, 46n19, 135n31, 179, 221–​2, 252–​3 dispositional account of impeccability see impeccability Dodd, Dylan 51 dogmatism, emotional 182–​5 dogmatism, perceptual 60, 257; see also phenomenal conservatism dogmatic attitudes 86, 214 Dogramaci, Sinan 132n3 Döring, Sabine 182 Dougherty, Trent 61, 71n4, 72n12, 132n1, 136n47, 210, 240n7, 267n12 doxastic attitudes (non-​beliefs) 44n1, 51, 151–​2, 159n5, 230 doxastic voluntarism 5, 19–​20, 22, 23, 27, 44, 202–​3 duties, epistemic: betting analogy and 53, 231–​2; dual nature of 52–​3, 231–​2; objectivity of 52, 53, 70n1; overview of 51–​3; risky or reckless beliefs 52, 152–​4, 185–​6, 201–​2, 230–​2; seemings and 104–​5, 157, 230–​3; synchronic vs. diachronic 52; see also deontologism Earlenbaugh, Joshua 107–​8, 132n3, 134n14, 134n17 emotions 27, 82–​3, 102, 110–​11, 127, 182–​5, 192, 212 empty symmetrical evidence 57–​8, 69, 185, 199–​201, 229–​30, 247–​9, 251, 255, 257–​9, 265, 267n9 endoxa 194–​5 epistemic conservatism 190–​2, 200; adherents of 190, 191; argument against 7, 190, 193, 202, 204; common sense methodology and 193–​5; counterintuitiveness of 7, 192, 197–​202; differential conservatism 191, 193; epistemic duties and 201–​2; generation conservatism 191; memorial beliefs 196; perseverance conservatism 191, 193; problem of forgotten evidence and 196–​7; problem of reflective awareness and 96n2, 196; Reid’s defense of 11n4, 202–​4; restricted vs. unrestricted versions of 191;

skepticism and 195–​6; strong or weak versions of 191–​2; theoretical benefits of 7, 193–​7; epistemological Adam 201–​2, 205n3; evaluative force 127, 183; Evans, Gareth 123, 165, 171 evidence: overview of 1–​2, 53, 69, 70n2; psychologism about 54–​5, 70n3, 71n4; see also foundational evidence evil deceiver 16, 18, 36–​7, 43, 195, 218, 250–​1 evil demon see evil deceiver externalism about justification 4–​5, 8, 17–​19, 21, 35–​40, 46n14, 46n15, 46n16, 46n21, 164, 186n6, 247–​53 faith 237, 253–​4 Farley, R. M. 105, 132n1, 210 feel of truth see forcefulness Feldman, Richard 22, 45n12, 51, 70n1, 71n4, 71n7, 116, 175, 197–​8 felt evidential insensitivity 255–​7, 267n5, 267n7 Fish, William 166 Foley, Richard 52, 70n1, 72n13, 89, 190, 237, 253–​4 forcefulness 2, 6, 64, 111–​12, 118–​20, 125–​9, 150, 178–​81, 185–​6, 229–​30, 260, 263–​5; see also metacognitive account of phenomenal force foundationalism 1, 58, 142–​4, 216, 235; argument for 7, 142, 158–​9; classical foundationalism 45n2, 60, 128, 143, 211–​12, 225, 251–​4; moderate foundationalism 60, 209, 212; weak foundationalism 143, 153–​5, 192; see also regress argument; seemings foundationalism foundational evidence 56–​8, 69; candidates for 164–​6; dual psychological and normative role of 55, 179, 185–​6, 229–​30; non-​ conceptual states as 166–​76, 235; non-​forceful states as 181, 185–​6, 230, 235; foundations evidentialism 1–​2, 58–​9, 205, 216–​17; see also recognition condition; relation condition Frege, Gottlob 167, 173, 186n2

Index  273 Frege’s puzzle 167–​8, 173–​4 French, Peter 132n2 Fricker, Miranda 184 Fumerton, Richard 87, 128, 135n29, 136n42, 165, 175–​6, 180 Gage, Logan Paul 71n4, 92, 132n1, 210, 253, 267n12 Gaṅgeśa 239n2 Gatzia, Dimitria 267n5 G. E. Moore Shift 194, 213–​14, 216–​17 Ginet, Carl 45n4, 159n9 Goldman, Alvin 17, 35–​36, 132n2 Goldstick, D. 192 Gopnik, Alison 21, 132n2 Grice, Paul 194 Grube, Enrico 124 Hales, Steven 98n16 Harman, Gilbert 191 Harrison, Eilidh 261 Hasan, Ali 135n29, 158, 160n21, 165, 180, 187n9, 225 Hawthorne, John 66, 72n12 Heck, Richard 134n29, 170 Helm Bennett 127 Hieronymi, Pamela 26 hinge propositions 200–​1, 203–​4, 261, 264, 265 Huemer, Michael 3–​8, 21, 37, 60–​3, 92, 95, 96n1, 105, 108, 122, 131, 132n1, 133n7, 134n19, 134n21, 134n25, 134n26, 176, 178, 195, 209, 210, 217, 218, 221, 224–​6, 230, 232, 238–​9, 239n3, 240n7, 250, 251, 266n3 Hume, David 8, 10, 14, 16, 65, 82, 181, 194, 208 impeccability 16–​17, 20–​2, 28–​40 infallibility 18, 213, 251–​4; see also classical foundationalism; infallible seemings infinitism 7, 144–​50, 152, 154–​5, 158 inflexible climber 149–​55, 181–​2 internalism: access internalism 33–​4, 40–​1, 44; argument for 5, 17–​19, 39–​40; awareness requirement on 41–​4; causal internalism 34–​5, 40–​4, 54, 55, 70, 93, 164, 179,

221–​2, 247, 249, 267n10; general description of 4, 13, 17, 33; impeccability and 17, 20–​2, 28, 34–​5, 39–​40; intuitions favoring 249–​51, 266n3; non-​factive mentalism 34, 40; intuitions 2, 79–​80, 85, 95–​6, 98n16, 101–​2, 116–​17, 127, 130–​1, 132n2, 132n3, 134n14, 136n45, 155–​6, 165, 176, 199, 212, 218–​21, 226, 233, 245–​6, 251–​2, 256–​7, 265, 267n11; see also types of seemings (rational or intellectual seemings) Irwin, Terence 119 Jackson, Alexander 229 Jackson, Elizabeth 51, 70n1 James, William 20, 232, 240n10 justification, epistemic: conditional justification 144, 146–​7; doxastic justification 45n10, 54, 55, 69–​70, 145, 174–​5, 215–​17, 223–​6, 252, 262–​3; general description of 1, 4, 50–​5; immediate justification 1–​3, 55–​6, 66–​7, 76, 142–​3, 155–​6, 185–​6, 190, 204–​5, 247, 255, 263–​5; inferential justification 1, 66–​7, 79, 94–​6, 97n3, 143, 144, 146–​7, 149, 159n4, 160n16, 204, 235, 261; original conception of 5, 13–​19, 44; propositional justification 45n10, 54, 69, 142–​3, 145–​6, 174–​5, 215–​17, 224; truth and 4, 5–​6, 18–​19, 45n2, 251–​4; see also deontologism; externalism; internalism Kahneman, Daniel 231 Kelly, John 71n4, 240n10 Klein, Peter 144–​5, 160n11, 160n17, 160n18 Knobe, Joshua 132n2 Koksvick, Ole 132n1, 133n7, 210 Korman, Daniel 124 Kornblith, Hilary 132n2 Kriegel, Uriah 210 Kvanvig, Jonathan 21, 60, 151, 160n14, 250 Langsam, Harold 39 Lasonen-​Aarnio, Maria 66, 72n12

274 Index Lehrer, Keith 250 Lemos, Noah 90 Lewis, David 77, 132n2, 136n46, 194 Locke, John 5, 13–​18, 20–​1, 33–​4, 44, 120, 135n32, 201–​2 Løkke, Håvard 118, 135n30 loss aversion 231–​2 Ludwig, Kirk 132n2 luminoscity see reflective transparency Lycan, William 132n1, 132n2, 160n15, 190, 191, 210 Lyons, Jack 103–​4, 115, 133n9, 133n11, 133n13, 134n13

210, 240n9, 248, 251, 255–​7, 259, 264 Morrison, Catriona M. 254

Machery, Edouard 132n3 Mackie, J. L. 160n14 Markie, Peter 132n1, 210, 225, 247–​50 Martin, Michael 123 Matheson, Jonathan 210 McAllister, Blake 11n4, 16, 64, 72n12, 70n1, 92, 132n1, 135n38, 136n47, 191, 202–​3, 210, 240n7, 241n14, 253, 267n12 McCain, Kevin 55, 68, 70n3, 71n7, 75–​7, 104, 105, 130, 131, 132n1, 151–​2, 166, 190–​2, 196–​8, 205n1, 205n2, 205n4, 210, 255–​7, 259, 264 McDowell, John 39, 124, 164–​5, 170–​2 McGrath, Matthew 102, 132n1, 210, 228 memory 42, 64, 90, 92, 109, 120, 131, 194, 196–​7, 202, 212, 218, 223–​4, 239n1, 246; see also types of beliefs (memorial beliefs); types of seemings (memorial seemings) Merricks, Trenton 136n43 metacognitive account of phenomenal force 120, 127–​8, 130, 136n41 Miller, Brian 66 Mitchell, Jonathan 127 Molyneux, Bernard 107–​8, 132n3, 134n14, 134n17 Moon, Andrew 51 Moore, G. E. 76–​9, 83, 194, 209, 213–​15 Moretti, Luca 66, 68, 75–​80, 96n1, 96n2, 97n4, 130, 131, 132n1, 151–​2, 166, 195–​7, 205, 205n2,

particularism 8, 97n8 Pascal, Blaise 20 Peacocke, Christopher 123, 164–​5, 170–​3, 186n6 perception: analogies to 119–​20; conceptualism about 123–​5, 135n38, 164–​5, 180–​1; indirect realism about 261–​2; non-​conceptualism about 122–​5, 135n38, 165, 166, 182, 222; perceptual experience 10, 62, 124–​5, 135n38, 212; representationalism about 122–​3, 126, 165; relationalism about 122–​3, 126, 166, 176, 178; see also types of beliefs (perceptual beliefs); disjunctivism; types of seemings (perceptual seemings); sensory appearances; sensory phenomenology permissivism 70n1, 240n10 phainomena see phantasia phantasia 118–​19, 194–​5, 210–​11 phenomenalism 44, 47n25, 267n10 phenomenally individuated mental states 41–​4, 47n23 phenomenal conservatism 2, 59–​70, 209; anti-​skeptical argument for 8, 164, 212–​4, 245–​6, 261, 265, 266; Bayesian updating and 66; bizarre beliefs objection to 6, 68, 91–​4; circularity or bootstrapping objection to 6, 87–​91, 93, 98n12, 98n14, 98n17, 99n20, 233; common sense methodology and 3, 193–​5; constitutive rules of rationality and 8, 204, 233–​9, 266; double-​counting objection to 6, 94–​6; externalism and 66, 247–​51; history of 3–​4, 60,

naïve comprehension axiom 103, 107–​8, 113, 117, 133n8 Nelson, Mark 51 new evil demon problem see evil deceiver Nichols, Shuan 132n2 norm guidance 28–​30, 33–​9 Owens, David 24, 29

Index  275 208–​12; intuitive case for 8, 228–​33; Minimal-​PC 7, 210–​11; preexisting support for 71n8, 210–​12; problem of cognitive penetration and 248–​51; problem of forgotten evidence and 197; Reidian argument for 7–​8, 209–​10, 245–​6, 264–​6; self-​defeat argument for 8, 224–​8; testimonial analogy and 63–​6, 80, 94–​6, 209; see also common sense response to skepticism; problem of reflective awareness Phillips, Stephen 239n2 Plantinga, Alvin 14–​17, 20–​1, 121, 134n29, 135n32, 149–​55, 181, 210 Plug, Allen 210 presentational appearances see types of seemings (presentational seemings) presentational constraint 259–​64 presentational phenomenology see types of seemings (presentational seemings) Pollock, John 28–​30, 34–​6, 134n29, 160n12, 214 Poston, Ted 57, 151, 153–​4, 160n15, 190, 199–​200 presumption 68, 151–​2, 193–​4 proper functionalism 8–​10, 11n1, 11n2, 65–​6, 72n14, 93, 111, 164, 191, 237, 247 propositions 55, 164, 168, 169, 186n3 Pryor, James 62, 63, 132n1, 194, 210 Pust, Joel 132n1, 132n2 Pyrrho the Elean 82 Quine, W. V. O. 191 Rawls, John 194 recognition condition 168–​77, 187n11 Reed, Baron 83–​6, 135n31 reflective account of impeccability see impeccability reflective awareness, problem of 76–​80, 96n2, 97n4, 195–​6 regress argument 1, 7, 144–​9, 152, 158–​9 Reid, Thomas: on common sense methodology 85–​6, 193–​4; on constitutive principles of rationality 203–​5, 236–​8; on epistemic conservatism 11n1, 11n4, 190, 191,

202–​4; on our fallibility 253–​4, 266n4; on foundationalism 235; his legacy 8; on perception 3, 10–​11, 64, 72n12, 124, 166, 238, 241n14, 245–​6, 261–​3; on phenomenal conservatism 9–​11, 65–​6; on the principle of credulity 65, 208–​9; on proper functionalism 8, 11n2; on ridicule 82–​3, 90–​1, 194, 213, 227; on semi-​skepticism 246, 265; on skepticism 76–​7, 79, 82–​6, 193–​4, 208–​9, 213–​14, 226–​7; on source circularity 88–​91, 99n20; on testimony 64–​6, 72n13, 80, 94–​5, 209; on trust in our rational faculties 64–​6, 72n14, 208, 236–​8, 246, 253–​4; on warrant 8, 45n3 Reiland, Indrek 169, 210 relation condition 168–​70 Riggs, Wayne 240n10 Roberts, Robert 127, 134n29, 182–​3 Rogers, Jason 210 Rosen, Gideon 132n1, 210 Roskies, Adina 135n38 Rowe, William 194 Russell, Bertrand 113, 165, 186n3 Ryan, Sharon 24 Ryle, Gilbert 194 Rysiew, Patrick 71n4, 204, 234, 236–​7 Salmon, Nathan 186n3 Schoenfield, Miriam 38, 40, 45n6 Schwitzgebel, Eric 21, 132 Searle, John 194, 233 Sedley, David 118 seemings: basic description 2, 50, 101, 131–​2; basing belief on 105–​6, 217–​24, 226; belief view of 6, 101–​6, 132n2; cognitive science and 95–​6, 120, 127–​8, 130, 136n41; experience view of 2, 6, 101–​3, 105–​6, 110, 115–​32, 132n1; inclination view of 6, 101, 106–​12, 132n3; other uses of “seems” language 102, 109, 112, 114; possession of 71n7; pre-​ theoretical understanding of 101–​2; propositional attitudes and 136n40; skeptics about 6, 116–​20, 131–​2; taking-​evidence view of 101, 112–​15; triple acquaintance states

276 Index and 178–​80, 211–​12, 219–​21, 225, 260; underdeveloped objection to 6, 116, 121–​2, 126, 131–​2; weirdness objection to 6, 115, 120–​1, 126–​8, 131–​2; see also assertiveness (experiential assertiveness); forcefulness; types of seemings seemings exclusivism 2–​4, 59, 163, 185–​6, 216–​17, 233, 235–​6 seemings foundationalism 3, 50, 58–​60; argument for 3–​8, 75, 163–​4 seemings, types of: foundational seemings 246; infallible seemings 211–​12, 251–​4; inferential seemings 131; introspective seemings 2, 128–​30, 212, 239n1, 245–​6, 251–​2, 257, 259; memorial seemings 2, 109, 131, 197, 245–​6; mere seemings 264; moral seemings 91–​2, 131, 240n7; paired appearances 255, 257, 258; perceptual seemings 2, 10–​11, 78, 81–​2, 87, 93, 97n3, 122–​6, 130, 221–​2, 245–​6, 257–​8, 261–​3, 267n8, 180–​2; presentational seemings 130, 259–​64, 267n9; rational or intellectual seemings 2, 79–​80, 109, 130–​1, 136n45, 197, 212, 239n1, 245–​6, 251–​2, 256–​7, 259, 265, 267n11; religious or theistic seemings 91–​2, 131, 240n7; seemings-​false 67–​8, 127, 135n34, 135n39, 209; tainted seemings 248–​51; testimonial seemings 204n7; see also seemings Sellars, Wilfred 7, 155, 160n15, 160n19, 186n4 Sellar’s dilemma see Bonjour’s dilemma sensations see perception; sensory appearances; sensory phenomenology sensory appearances 122–​6, 180–​2, 222, 258, 267n8 sensory phenomenology 10, 11n3, 109, 122–​6, 131, 166, 180–​2, 187n14, 255, 257–​9, 261, 267n8 sensus divinitatis 240n7 Shields, Christopher 210–​11, 239n3 Siegel, Susanna 125, 170, 247, 248 Silins, Nicholas 125 skepticism: classical foundationalism and 252; contextualist response

to 77–​8; explanationist response to 76–​7, 252; infinitism and 145–​7; knowledge skepticism 213; presentational constraint and 261–​3; underdetermination by the evidence 195; see also common sense response to skepticism; phenomenal conservatism (anti-​skeptical argument for); Thomas Reid (on skepticism) Skene, Matthew 126, 132n1, 210 Sklar, Lawrence 191, 192 Smith, A. D. 124 Smith, Angela 25, 27 Smithies, Declan 21, 41, 44, 45n4, 47n23, 47n25, 60, 71n6, 125, 126, 129–​30, 132n1, 186n8, 174–​5, 210 Sosa, Ernest 106, 132n3, 134n15 speckled hen 174, 219 Spencer, Philip 267n6 Srinivasan, Amia 21–​2, 45n4 Steup, Mattias 20, 24, 45n4, 239n1 Stoics 118–​19, 135n31, 195, 211 Swinburne, Richard 132n2, 132n3, 190, 191, 209, 210 Taylor, Samuel 107, 109, 132n3, 134n18, 134n19, 134n20 Teng, Lu 120, 127–​8, 130, 136n41, 247–​8 testimony 63–​6, 80, 94–​5, 72n13, 79, 240n7 Thagard, Paul 160n15 Tolhurst, William 126, 134n21 Tooley, Michael 61, 68, 72n16, 94, 112, 121–​2, 126, 132n4, 225 transparency, reflective 5, 20–​2, 28, 45n5, 45n6, 47n25 trust 2, 63–​5, 72n13, 76–​8, 80–​1, 87–​91, 93, 208–​9, 236–​9, 253–​4, 266 truth 52, 129, 136n43 truth-​makers 111, 129–​30, 259, 261–​3 Tucker, Chris 61, 64, 92–​4, 96n1, 101, 121–​2, 123, 131, 132n1, 132n2, 132n3, 133n7, 136n47, 210, 229, 240n7, 240n9, 251, 258, 266n1 Turnbull, Greta 70n1 Turri, John 160n17 Tversky, Amos 231 Tye, Michael 123, 165

Index  277 Ullian, J. S. 191 Unger, Peter 213 Vahid, Hamid 191, 193 Vogel, Jonathan 87 Wagner, Verena 132n4, 134n23 warrant 8, 45n3, 93–​4 Watson, Gary 25–​7 Wedgwood, Ralph 31, 32, 36, 38, 42, 47n22 Wettstein, Howard 132n2

White, Roger 66 Williamson, Timothy 21, 29, 39, 45n5, 71n4, 116–​17, 132n3 Wittgenstein, Ludwig 200 Wolf, Susan 26 Wolterstorff, Nicholas 254 Yablo, Stephen 132n3 Zagzebski, Linda 72n13 Zeno of Citium 118