proceedings of the aristotelian society susanna siegel (harvard)

2014/2015 136th session volume cxv proceedings of the aristotelian society Edited by matthew soteriou (warwick) issue no. 3 E p i s t e m i c C h a r g e susanna siegel (harvard) www.aristoteliansociety.org.uk memberships latest issues podcasts virtual issues the joint session

proceedings of the aristotelian society 136th session issue no. 3 volume cxv 2014 / 2015 susanna siegel harvard university monday, 15 june 2015 17.30-19.15 room 349 senate house university of london malet street london wc1e 7hu united kingdom This event is catered, free of charge, & open to the general public contact mail@aristoteliansociety.org.uk www.aristoteliansociety.org.uk 2015 the aristotelian society

biography Susanna Siegel is Edgar Pierce Professor of Philosophy at Harvard University. She is author of The Contents of Visual Experience (Oxford University Press, 2010), and numerous articles in the philosophy of perception. Recent papers discuss the varieties of influences on perceptual experiences from cognition, affect, and learning, their impact on the epistemic role of perception, and the nature of belief. editorial note The following paper is a draft version that can only be cited or quoted with the author s permission. The final paper will be published in Proceedings of the Aristotelian Society, Issue No. 3, Volume CXV (2015). Please visit the Society s website for subscription information: www.aristoteliansociety. org.uk.

susanna siegel How is perception distinct from belief? It is often assumed that while beliefs redound on the rational standing of a subject, perceptions do not. An irrational belief detracts from the rationality of believer, according to this assumption, but perceptions cannot do the same. I argue that perceptual experiences can have a rational standing, and that their epistemic status can be modulated by other psychological states (such as fears, wishes, prior knowledge and belief) that help produce them. Drawing on a metaphor from electricity, I call the epistemic status that perceptual experiences can have. VIVEK HAS BECOME a vain performer. In the past, even after his performances solidified, Vivek felt unsure of himself. He often came away thinking he had performed badly. In those days, no matter who was in the audience, their faces looked neutral or displeased. But remarkably, no one ever looked content. Now, no matter who Vivek performs for, the faces in the audience range in their expression from neutral to pleased. Remarkably, no one ever looks disapproving. As Vivek gained confidence, his perception changed. Let us suppose that Vivek is projecting onto the audience his conception of himself. He projects either the opinion that he is pleasing (when he is vain), or that he is displeasing (when he is diffident). His projection prevents him from seeing the wider range of expressions that others who aren t invested in his point of view could readily observe. Projection comes in at least two modes, in principle. A first mode affects only how one interprets what one sees. Vain Vivek comes away believing that the audience is pleased with his performance, and some aspect of his vanity such as his desire that others like his performance, or his antecedent confidence that they will like it makes him interpret what he sees in the faces. In this mode of projection, the sense in which the audience looks pleased to vain Vivek is that Vivek is judging, in response to his visual perception, that by and large the audience is pleased. Someone else viewing the same audience might come away with different judgments, but without any significant difference in what they see. This mode of projection can clearly have adverse effects on the rationality of the perceptual judgment. Vivek jumps to the conclusion that a particular audience member thinks a performance is top-notch, when the 4

main indication of what they think is their neutral facial expression. When Vivek jumps to conclusions, his belief about the audience member s reaction is formed epistemically badly. To introduce a piece of terminology, his belief is thereby ill-founded, because being formed epistemically badly is one way to be ill-founded (another way is to form subsequent beliefs that are at odds with the first, without adjusting any of them). Vivek s vanity leads him to form a poorly justified belief. In the second, more far-reaching mode of projection, Vivek s vanity reaches more deeply into his perception of the audience. It influences the appearances he has to interpret, rather than simply influencing his interpretation of them. By the time he interprets what he sees, Vivek s vanity has already shaped his perception. Here, too, we might describe the end result by saying that to Vivek, the audience looks pleased. But this time, the sense in which the audience looks pleased is that approval is suggested directly by what Vivek is responding to, when he judges that they approve. Vivek s vanity isn t making him jump to conclusions about how great the audience thinks he is. His vanity functions like rose-tinted sunglasses, rather than a bias toward judging that the things you see are reddish. Like sunglasses, his vanity affects the visual appearances themselves, and the appearances can then be taken at face value. The two modes of projection differ in whether Vivek s vanity leaves an aspect of conscious perception untouched. What conscious perceptual state is left untouched by vanity in the first mode, but directly affected by it in the second? I ll call it Vivek s perceptual experience. A perceptual experience is a conscious mental state that one is typically in when one perceives part of the immediate environment. The experience characterizes how the things you see look to you, at a level that is an input to judgment. By having the experience, you are in a position to interpret what you see. Does the second, deeper mode of projection lead to any epistemic shortcoming? An epistemic effect most directly analogous to the effect on belief by the first mode of projection would be that the perceptual experience itself is formed epistemically badly, and would be thereby illfounded. Here is the central question of this paper: Is this epistemic status a status that experiences can have? If so, then experiences can be irrational (or, presumably, rational). If perceptual experience were simply a form of judgment, then since judgments (as a form of belief) can be ill-founded by resulting from jumping to conclusions or other epistemically faulty forms of inference, experiences on this view could be susceptible to those faults too. 5

But on the face of it, perceptual experience is distinct from belief. When you know that things aren t as they appear in your experience, you don t believe your eyes. Later on, I ll reconsider whether perceptual experiences could be a form of belief, but for now, let s assume they aren t. One way a perceptual experience could come to be irrational, if that s even a coherent possibility, is that the perceptual experiences is the conclusion of an inference, where the inference has the same epistemic impact on experience as it has on beliefs. When the conclusion of an inference is a belief, inference makes the belief epistemically depend on what it is inferred from. For instance, if a subject ends up believing P on the basis of an inference, that belief can detract from the subject s rational standing, due to either the epistemic status of the inputs to the inference, or the structural features of the inference, or both. Psychologists have long argued that perceptual experiences like perceptual states in general do result from inference or at least, from a process they have happily called inference. 1 But by and large, those interested in the epistemic role of perceptual experience have generally not concluded from the fact that perceptual experiences result from inference that experiences epistemically depend on their psychological precursors that they re inferred from. Why not? Do the unconscious inferences that lead to perception differ in kind from inferences that establish epistemic dependence of beliefs? Or is there just one kind of inference? I won t try to settle this question here. Either way, it is usually presumed that experiences are not the kind of states that can have a rational standing, the way beliefs do. Ernest Sosa voices this idea when he says: [W]hen [perceptual] experiences help explain the rational standing of some other state or action, they do not thereby problematize their own rational standing. Being so passive, they have no such standing. Think of the idea that perceptual experiences can be unjustified justifiers the main ingredient in the thesis that beliefs can be immediately justified by experience, which is in turn a cornerstone of foundationalism in epistemology. 2 When proponents of immediate justification appeal to experiences as unjustified justifiers, they don t mean that the experiences 1 Helmholtz, I. Rock, Ibn al Haytham. Hatfield (2002) Perception as Unconscious Inference in Perception and the Physical World: Psychological and Philosophical Issues in Perception. Eds. Dieter Heyer and Rainer Mausfeld. 2 A subject S s belief that p is immediately justified by an experience, just in case there need be no further propositions that S must be justified in believing from a source other than the experience, in order for experience to justify her in believing that p or if there are, being justified in believing those propositions need not play a role in S s getting justification to believe p from her experience. 6

are anti-justified. They mean that experiences are off the grid of epistemic norms. The grid supports beliefs, but not experiences. Not all foundationalists are proponents of this kind of immediate justification, and not all proponents of immediate justification are foundationalists. But defenses of both positions usually rely on the assumption Sosa articulates: that experiences have no rational standing. A powerful idea is that something in the nature of perceptual experience precludes them from having a rational standing, and hence from receiving the same kinds of epistemic effects as beliefs have, by virtue of resulting from inference. Perhaps something in the nature of inference, or of one kind of inference, precludes this as well. 3 I ll focus here on the idea that it s the nature of experience, rather than inference, that underlies the powerful presumption that inferences don t modulate the epistemic standing of experience, in the way that they can modulate the epistemic standing of belief. Why explore whether a person s experiences can reflect their rational standing? First, if experiences can have a rational standing, then we ll have to re-think the boundary between a-rational and epistemically evaluable features of the mind, since that boundary has widely been taken to lie at the interface between perception and belief. 4 Second, experiences that have a rational standing are not unjustified justifiers. On the face of it, it might seem that foundationalism that puts perceptual experience in the foundation stands or falls with the status of experiences as unjustified justifiers. Ultimately I ll argue that that this isn t so, but if one is going to paint the resulting picture of mind, one has a lot of explaining to do. I argue that nothing in the nature of experience precludes it from being irrational. My first step is to consider more closely the epistemic status that s at issue, when we ask whether experiences could be irrational. After discussing this epistemic status (section 1), I consider and reject reasons to think experiences are precluded from having such a rational standing, and float a positive reason to think they do (section 2). Finally, I turn to the most plausible accounts of what the scope and grounds of the epistemic status would be, and consider their implications for the global structure of justification (section 3). 1. When a belief is formed or maintained epistemically badly, it is ill-found- 3 I visit this question in section 3. 4 Davidson, Paradoxes of irrationality [A. Smith, R. Feldman, etc] 7

ed. What does ill-foundedness consist in? We can distinguish between static and forward-looking aspects of it. The static aspect is that it detracts from the subject s rationality. The forward-looking aspect is that it has the potential to make subsequent beliefs formed on its basis ill-founded. Wellfoundedness of belief has analogous static and forward-looking aspects. These two aspects form two dimensions along which beliefs are epistemically evaluable. Let s look more closely at the static aspect, focusing on ill-foundedness. When a belief B is ill-founded, merely having the belief detracts from your overall rationality. You are less than fully rational, by virtue of having B. We can call the aspect of the ill-founded belief that makes it the case that you are in a lesser rational state the property of being anti-justified. The first aspect of a belief s being ill-founded is that the belief is anti-justified. The second aspect of ill-foundedness is forward-looking. Subsequent beliefs formed on the basis of B may inherit B s anti-justification. 5 The belief B thereby has a power to make other beliefs ill-founded, if those beliefs are formed on the basis of B. This power is the second aspect of ill-foundedness. It has three important features. First, suppose that B is the belief that p&q, and B* is the belief that q. If belief B is ill-founded, then it has the power to make B* ill-founded. And it can have this power, whether or not B is the right kind of thing to be a reason for B*. On some views, beliefs are only rarely reasons for other beliefs. According to these views, in our example, the reason for B* is not B, but rather p&q, and belief is a relation to p&q that helps the subject have p&q as a reason. 6 The forward-looking aspect of ill-foundedness does not include any commitments about whether B is or isn t a reason for subsequent beliefs formed on its basis. 5 Does any subsequent belief formed on the basis of B have to inherit B s ill-foundedness? In some cases B s ill-foundedness could wash out. For instance, if B is part of a complex belief in a scientific theory and the complex belief is well-founded, and a subsequent belief is formed on the basis of the entire theory, then the fact that B is ill-founded may not make any significant difference to the epistemic status of the subsequent belief. Consider someone who believed the same complex theory, but had a well-founded belief that B instead of an ill-founded one, and went on to form the same subsequent belief. Now compare this epistemically superior person to the subject of B, when both subjects hold the same subsequent belief on the basis of the scientific theory. The subject of B might not be comparatively less rational than the superior subject in holding the subsequent belief, although by hypothesis, in holding B, she is less rational in holding B. Cases like these show that the forward-looking aspect of ill-foundedness is a potential for having a forward-looking effect, rather than the disposition to detract from the ill-foundedness of any subsequent belief formed on its basis. 6 Raz 1975, Stampe 1987, Dancy 2000. [Dancy s example of belief that is a reason for another belief raise your hand if you believe that P.] 8

Second, B can have the power to make subsequent beliefs formed on its basis ill-founded, whether or not B harbors epistemic resources that bear on B*. On some views, it is not the belief B that harbors epistemic support for B*, but rather the epistemic support the subject has for B that does that (if anything does). Third, the fact B has the power to make subsequent beliefs formed on its basis ill-founded does not entail that the subject s merely having B is sufficient to give B this power. This point may seem obvious, since someone could have belief B when B is well-founded. But even aside from B s own rational standing, the minimal unit of epistmic power may include more than the belief B. Depending on what B* is, B may need help from other mental states, or from external conditions, for B to have forwardlooking epistemic power regarding B*. Even then, B still has forwardlooking power. These three features of the forward-looking aspect of ill-foundedness are important because they can apply to perceptual experiences, even if experiences have no rational standing, as Sosa describes. Perceptual experiences can have forward-looking power to make beliefs formed on their basis ill-founded, even if they cannot be justified or unjustified. And they can have that power, whether or not they are reasons for belief, or harbor epistemic support for them, or are minimal units to which forward-looking power belongs. Perhaps the minimal unit of epistemic power includes other background beliefs, or constraints on the etiology of the experience. Now that we ve distinguished between static and forward-looking aspects of epistemic evaluation, we can ask more refined questions about Vivek. When we ask whether his experience could be ill-founded, we can be asking about the static aspect of ill-foundedness, or the forward-looking aspect, or both. I m going to ask about whether Vivek s experience can have a status that combined both aspects. Using electricity as a metaphor, let s say that an experience is negatively epistemically charged just in case it has both the static and the forward-looking aspect of ill-foundedness, and can transmit the negative epistemic status (the static feature) to subsequent beliefs formed on its basis. 7 So if Vivek s experience has a negative epistemic charge, it harbors a negative epistemic status. Just as the beliefs can be justified or anti-justified depending on whether the static aspect of epistemic evaluation is negative or positive, any experiences with negative or positive have an analogous epistemic status that reflects the subject s rational standing. And just as one belief can be more anti- 7 [transmission adds something so not just asking about both aspects.] 9

justified than another, one experience can be more negatively epistemically charged than another. Like justification, comes in increments. The electricity metaphor extends to these features. Just as electric charge both has a valence (positive or negative) and comes in increments, has a valence and a negative or positive can be larger or smaller. Let s assume that Vivek remains unaware that his vanity has any influence on the content and phenomenal character of his perceptual experience. Can Vivek s experience have any at all, and if so, is it modulated (for instance, is its increment reduced, or its valence made negative) by the influence of his vanity? Here we are asking whether Vivek s vanity gives his experience negative, and so reflects poorly on Vivek s rational standing. Charged: Vivek s vanity gives his experience negative epistemic charge. Let us distinguish this effect on Vivek s experience from a weaker effect that targets only the forward-looking aspect of ill-foundedness. I mention this other possible effect to highlight the contrast with the Charged option, and to set it aside. Downgraded: Vivek s vanity reduces the forward-looking power of his experience to provide justification to subsequent beliefs. If Vivek s experience is downgraded by vanity, then the forward-looking power of his experience is reduced, relative to the power it could otherwise have if his experience were not influenced by vanity. If Vivek s vanity negatively charges his experience, it downgrades it, but the converse does not hold. The Downgrade option is compatible with Vivek s experience lacking altogether. Because experiences could have forward-looking power to provide justification to subsequent beliefs formed on its basis, without reflecting the subject s rational standing, experiences could lose forward-looking power without that reduction reflecting on the subject s rational standing. As a first step toward putting the Charged option in a favorable light, I focus on whether any experiences could even be epistemically charged. I start by revisiting whether experience is a form of belief. I argue that if any experiences are epistemically charged, then this won t be due to their status as beliefs, because there are good reasons to think that experiences are not a form of belief. 10

Experience and belief Are experiences beliefs? They aren t, if they differ from beliefs in their basic structure. For instance, if beliefs are relations to propositions but experiences are not, then experiences will differ from beliefs. What might experiences consist in, if they differ from belief in this way? A first proposal is that they are not directed toward the world at all: they don t even seem to present the subject with environment distinct from them. Think of seeing stars from being hit on the head, or the pink glow that one experiences with eyes closed in sunlight. According to the the raw-feel view, all experiences are undirected toward the external things. A second proposal is that experiences, when they aren t hallucinations or any kind of illusion, are relations of perceiving external things and their properties. This structure for experiences is distinct from belief, since beliefs can be false, whereas perception as construed here can only relate perceivers to objects that exist and properties that that those objects actually have. Call this proposal Naïve Realism. The raw-feel view and Naïve Realism each entail that experiences are distinct from belief, on the assumption that beliefs are relations to propositions. What s more challenging is to argue that experiences are distinct from belief, even if they are belief-like by being the kind of state that can be accurate or inaccurate about the subject s environment. Just as beliefs can be true or false, on a third view about experiences, all experiences have accuracy conditions: conditions under which they would be accurate about the environment. 8 Call this proposal the Content View. Are experiences, even as the Content View construes them, distinct from beliefs? 9 Consider the temporal profile that standardly attaches to experiences and beliefs. Beliefs have inertia that experiences lack. If you acquire a belief, then it tends to stay in the mind, with no need to reestablish it. Of course it is possible to forget what one once believed. But the inertia of beliefs facilitates their role in planning and guiding behavior, and to that extent forgetting is not typical. Suppose you learn that your friends plane will land at 10pm. If your belief didn t last as long as your plan to meet them at the airport shortly after 10pm, then you d have to revisit the plan. And if the world changes in a way that makes the belief 8 Arguably experiences as Naïve realist disjunctivists construe them have accuracy conditions derived from basic structure. Siegel (2010) chapter 2 contains an argument to this effect. 9 For defenses of experiences as belief, see Glüer (2009) and Byrne (2009), and before them Pitcher and Armstrong s accounts, which are criticized by Jackson (1971), Johnston (2006), and Pautz (xxxx). 11

false for instance, if the plane is delayed - that is no guarantee that your belief will change. In contrast, experiences lack the psychological inertia that characterizes beliefs. Whether your experience persists typically depends on whether you remain in contact with the relevant part of the environment. If you see a skyscraper ahead, and then turn so that it is no longer in view, your skyscraper-experience comes to an end, but your belief about where the skyscraper is located will typically persist. Fading into memory extinguishes experience, but not belief. These considerations suggest that if experience were a form of belief, then experiential beliefs would be short-lived, and the reason for their short lives would differ from the reasons for losing other beliefs. Experience don t die from forgetting, or from that subject s responses to counterevidence. A proponent of assimilating experiences to beliefs could respond that beliefs simply vary in the range of temporal profiles they can have experiential beliefs have one profile, and non-experiential beliefs have another. But the substantive point would remain that experiences have a distinctive temporal profile. And the more differences one finds between experiences and non experiential belief, the less dialectical power an assimilation of experience to belief would have. For instance, if beliefs admit of a range of temporal profiles, perhaps they admit of a range of epistemic profiles as well. Given the diversity of beliefs that the proponent of the experienceas-belief thesis has to accommodate, they seem poorly positioned to claim that experiences share the epistemic profile of belief, on the grounds that experiences are beliefs. From here, I ll assume that experiences are not beliefs, when considering whether any experiences are epistemically charged. 2. does anything preclude experiences from being epistemically charged? The thesis I d like to promote is: Epistemic Charge thesis: Some experiences are epistemically charged. A much less challenging idea, already widely accepted, is that experiences merely have forward-looking power to provide justification to subsequent beliefs formed on their basis, without having any additional epistemic sta- 12

tus comparable to justification or anti-justification. 10 Since such forwardlooking power is part of, the less challenging idea is a deflated version of the Epistemic Charge thesis. Deflated Charge thesis: Some experiences have forward-looking power to provide justification to subsequent beliefs formed on their basis, but no. This thesis does not speak to the minimal unit that has forward-looking power. The most minimal a unit could plausibly be is an experience, regardless of its etiology, but with the constraint that defeaters are absent. This idea is captured in a position often called Dogmatism, because of the response to skepticism that it recommends. Dogmatism: Absent defeaters, having an experience with content P suffices to provide immediate justification for believing P, in virtue of its phenomenal character and regardless of its etiology. 11 As a theory of the minimal unit of forward-looking epistemic power of experiences, dogmatism is only one theory among many that is compatible with deflated charge thesis. If the Epistemic Charge thesis is true, then epistemically charged experiences reflect the rational standing of the subject. If no experiences are epistemically charged, then experiences as a class do not reflect the rational standing of the subject. Earlier we met a reason to think that experiences cannot be epistemically charged: because they are passive. What kind of passivity might underwrite the epistemic unevaluability of experience? We can distinguish between three kinds of passivity. I argue that none of them preclude experiences from being epistemically charged. A first kind of passivity is phenomenological. It is not part of the phenomenology of perception that our experiences seem to result in mental activity of any sort. But the same is true of many of our beliefs. They do not seem to result from active reasoning either we simply find ourselves believing that it is time for lunch, that our neighbors are kind, or that the music is too loud. Phenomenological passivity is a poor diagnostic for. 10 For overview of this idea, see Siegel and Silins 2015 11 Goldman. Pryor s terminology 13

A second kind of passivity is passivity with respect to any kind of reasoning. Could this kind of passivity underwrite the epistemic unevaluability of experience? We apply epistemic norms to all beliefs, even when they don t result from reasoning. For instance, self-ascriptions of experiences are sometimes not the result of reasoning, and cases where we simply believe our eyes also do not result from reasoning. And in some social contexts, allegiances can pull the strings of belief in ways that seem to involve no reasoning at all. 12 And yet none of these routes to belief preclude us from evaluating them as justified or anti-justified. Perhaps the most powerful version of the idea that experiences are precluded from being epistemically charged draws on a third kind of passivity: passivity with respect to deliberation, or explicit reasoning. One might think that experiences cannot be epistemically evaluable, because one cannot form an experience as the result of deliberation, or of explicit reasoning. In contrast, any belief could in principle be formed in those ways even if it not every belief is actually so formed. And that is why all beliefs are epistemically charged, according to this line of thought. A first reply is that the generalization about beliefs seems false. On one disambiguation, the generalization is that for any believer, all of that subject s beliefs could have been formed by deliberation. But if believers need to have starting assumptions, the generalization won t be true. There will be initial prior beliefs that are needed to get a system of believing off the ground. Some examples for human belief might be the built-in assumption that light comes from above, or assumptions about spatiotemporal continuity of ordinary objects. 13 On a different disambiguation, the generalization is that for any belief, it could be formed by deliberation even if it isn t the case that all of a subject s belief could be. Beliefs with the content I believe that p could in principle by formed by deliberation even if they are typically formed by introspection and without deliberation. The same point seems to hold for endorsements of perceptual experience, in which a belief with content P is formed on the basis of an experience whose content includes P. Even if one could believe P by endorsing an experience with content P, that same belief could be reached by deliberation. In reply, making deliberation the main diagnostic of 12 Tamir and Mitchell (2012. April 16). Anchoring and Adjustment During Social Inferences. Journal of Experimental Psychology: General. Advance online publication. doi: 10.1037/a0028232 13 For a suggestion that our system of belief needs starting assumptions, see Railton (2013) and Hohwy (2013). For potential examples of starting assumptions, see Carey (2009) case for core cognition. 14

treats beliefs formed in those ways as ideal forms of epistemically charged beliefs leaving other beliefs as pale approximations. But belief in general may not have any ideal form. Even if it does, it may have multiple ideal forms. A different paradigm of epistemically charged belief is the toddler s knowing, and hence believing, that her socks are on (after putting them on herself with much effort). The route to belief wasn t deliberation, but rather a mix of observation and action. A more mature subject could reach the same first-person belief ( I just put my socks on ) by deliberation, but the toddler couldn t. Yet the toddler s belief seems just as epistemically charged as the older subject s. And its role in the subject s mind seems just as beliefy as beliefs formed by deliberation: its felt strength comes in increments, and those increments belong to the toddler s outlook on the world. They shape her sense of possible futures: if her belief is firm, she might put on her socks again tomorrow, where if her feeling that she can put on her socks is unstable, the time she put them on might feel to her like a fluke. They shape her memory of putting on her socks, and her sense of herself and others: she can put her socks on unlike a baby, who can t. They make her disposed to respond to challenges as to whether she can or can t put on her socks. With this cognitive profile, the toddler s belief is not less beliefy than a belief formed by deliberation. 14 These considerations suggest that the of belief is not grounded in each belief arising from a process that is either deliberation or a pale approximation of it. Various kinds of passivity of experience seem poor grounds for epistemic unevaluability. A different idea is that experiences cannot be epistemically charged, because they cannot be rationally adjusted in response to criticism. According to this idea, for a mental state to be epistemically charged, it must be possible for a subject to adjust it, in order to make it conform to any epistemic norms that can be used to evaluate it. But now picture being told (and believing) that your experience came about because of your vanity. It doesn t seem possible to adjust your experience, other than by looking away, covering your ears, or otherwise closing off perceptual input (assuming the experience is not an internally generated hallucination). If the experience is over by the time you come to criticize it, there seems to be no way to adjust it at all, rationally or otherwise. So if being able to adjust perceptual experience without managing the intake of perceptual information is necessary for, then by this measure, it seems that experiences can t be epistemically charged. In reply, we can distinguish between two kinds of adjustment of a mental state in response to rational criticism: adjustment by deliberation, 14 I owe this insight to Peter Railton s discussion of belief and epistemic authority in Railton 2015. 15

and adjustment by disowning the mental state. If experiences were adjustable by deliberation in response to rational criticism, then in response to criticism, such as the information that the contents of one s experience is heavily influenced by vanity, or that the experience is irrational, one would have to be able to explicitly reason to a new experiential conclusion that rationally addresses the critical information. Adjustability by deliberation is a poor diagnostic for, for reasons we saw earlier. Many beliefs are not formed by deliberation, and can t be adjusted in response to it. But that doesn t stop the beliefs from having. Many beliefs are formed and adjusted without deliberation. The issue raised by Vivek s vain projection onto experience is whether his experience is like those beliefs in that respect. The second kind of adjustment is disowning a mental state. If this kind of adjustability is a good diagnostic of, then experiences satisfy it. Even if you couldn t make yourself stop having the experience, you can cease to rely on it in your reasoning and action. Ceasing to rely on an experience can even be done to a past experience. So there is such a thing as disowning an experience. When we respond to rational criticism of beliefs by giving up the belief, or by weakening it, this is what we do: we cease to rely on what we believed in reasoning and action (or we cease to rely on it so heavily). So what happens when you cease to rely on a belief happens as well when you cease to rely on an experience. There is also a difference. In the case of belief, ceasing to rely on a belief can t come apart from having the belief. But experience can persist, even if you don t use it in reasoning or action. (Here we find another difference in temporal profiles between experience and belief). If it persists, then there s residual irrationality, analogous to someone obtuse who disowns an attitude (e.g., disrespect for someone they treat badly), but lacks the understanding needed to correct all the perspectives that go with it. So far, I have rejected what strike me as the two most powerful kinds of reasons to think that experiences are epistemically unevaluable: experiences are passive, and experiences are not rationally adjustable. Several varieties of passivity and unadjustability pertain to beliefs as well. This continuity could naturally lead someone to wonder: what might ground the epistemic evaluability of belief, and could that factor also ground the epistemic evaluability of experience? A natural idea is that what grounds the epistemic evaluability of both states is their role in the mind. Perhaps there is no further feature of belief, or routes to belief, that explains why beliefs can be evaluated as epistemically better or worse. Instead, it is their role as states that constitute our 16

outlook on the world. The outlook may be unstable or temporary, or it may be better characterized by incremental states (such as credences) than binary ones. Parts of the outlook will be at odds with others, and not all parts will have equal weight in the subject s considered view. But experience forms part of our outlook all the same. An outlook on the world can be appraised. They also plausibly anchor appraisals of other mental states, such as fears and desires. For instance, desires are most plausibly seen as rationally unappraisable when they are construed as devoid of any representation of what is desired as favorable or unfavorable, and fears are most plausibly seen as rationally unappraisable when they are construed as devoid of representations of anything as frightening. 15 I explore further the idea that is grounded in belonging to an outlook in section 3. Combined with the reasons to think that experiences are not precluded from being epistemically charged, these considerations go some way toward motivating the Epistemic Charge thesis. To support the thesis even more, further explanation is needed on at least three fronts. Scope: What is the scope of among experiences? Ground: What is the ground of in experiences? Modulation: What kinds of factors can modulate the epistemic charge of an epistemically charged experience? The scope question asks whether all perceptual experiences have epistemic charge. Is having exceptional or standard? The ground question asks what features of charged experiences explain why they are epistemically charged at all, as opposed to neutral. The ground question bears directly on the scope question. The scope of will depend on the distribution across experiences of the features that ground its. For instance, if the features are exclusively contingent ones for an experience to have, such as resulting from a projection like Vivek s, then only the experiences with those features will be charged, and these experiences may be limited to occasional. At the other extreme, if the features are constitutive of experience, then all experiences will be epistemically charged. The modulation question assumes that experiences can be epistemically charged, and asks what kinds of factors can increase or decrease its charge, or flip its valence from positive to negative. Taken together, the ground and the modulation questions ask for an account of which fea- 15 [I am leaving aside appraisability of desires that comes solely from coherence requirements. [outlook also underlies modes of valuation that don t directly manifest in overly evaluative language. s-norms, selection effects.] 17

tures bestow an experience with the specific that it has. In the rest of the paper, I sketch the most plausible and powerful potential answers to these questions. These possible answers let us see more clearly how the Epistemic Charge thesis could be developed, what it entails, and how it impacts the global structure of justification. 3. the scope and ground of Let us start with the scope question. If some but not all experiences have, then experiences can still be unjustified justifiers, and they can in principle relate to other experiences and beliefs in the ways that epistemologists have tried to make familiar. 16 The impact on the global structure of justification is more interesting if all perceptual experiences have. When we consider possible grounds, there s good reason to think that if any experiences are epistemically charged, then being charged is standard rather than exceptional. And some potential grounds for epistemic charge suggest that it is universal. What are the most plausible grounds for of experiences? Earlier I suggested that the fact that experiences form part of our outlook on the world motivates the Epistemic Charge thesis, once the misleading apparent reasons to reject are cleared away. We can distinguish between our considered outlook on the world, and our complete outlook on the world at a moment. Experiences that we disown or disbelieve might not belong to our considered outlook, but they are part of the complete outlook. When we consider which features of experience underlie its contribution to an outlook, a first suggestion is its phenomenal character - at least those aspects of it that are closely tied to the properties presented in experience. 17 Just as phenomenal conservatives and others say that phenomenal character bestows experiences with powers to provide justification to subsequent beliefs formed on their basis, according to this idea, phenomenal character also bestows experiences with epistemic standing of their own. The proposal that is phenomenally grounded can be put in the form of an argument for a universal thesis. The phenomenal ground argument P1. All experiences have phenomenal character. 16 Feldman 2000, etc 17 Siegel 2010. Chapter 2. 18

P2. Phenomenal character gives experiences epistemic charge. Conclusion: all experiences have an. The key premise of the phenomenal ground argument, P2, raises several questions. First, if experiences are epistemically charged just by having a phenomenal character, then what valence does the charge have, and in what increment? As a starting observation, normally it is reasonable to believe one s eyes and other senses. If you want to know whether the sunset has started, you can find out by looking. Ordinary perceptual experiences like this one can provide a baseline amount justification for believing the contents of those experiences, or closely related contents. If the phenomenal character of experience bestows experiences with, a natural suggestion is that it bestows a baseline amount of positive charge on experience, so that if experience retains its power to provide justification to subsequent beliefs formed on its basis, it is providing that power by passing on its own positive epistemic status. The thesis that phenomenal character of experience bestows it with does not entail that an experience s having the phenomenal character (and therefore a subject s having the experience) suffices for the experience to have a baseline positive a point we met earlier (section 1) in discussing minimal units of forward-looking power. 18 The phenomenal character of experience could be a factor that contributes positive, but in the presence of the other factors could modulate the charge and make it negative, or reduce its increment of positive charge. For instance, Vivek s vanity could modulate the of his experience of the faces in the audience, so that his experience is negatively charged. Say that P is a content for believing which Vivek s experience could provide a baseline amount of justification, were it not influenced by his vanity. How does the phenomenal character of Vivek s experience, with its power to provide baseline justification for P, interact with the influence of Vivek s vanity on his experience, which we re assuming leaves his experience negatively epistemically charged? Different models of this interaction are compatible with the Epistemic Charge thesis. On an aggregative model, due to its phenomenal character, the experience has a positive charge that provides some pro tanto justification for P, but the influence of Vivek s vanity leaves the experience with more anti-justification for P. Here, the pro tanto justification bestowed by the experience is simply outweighed- it does not go away. In contrast, on a prevention model, the influence of Vivek s vanity prevents the phenomenal 18 Contra Byrne 2015 19

character from bestowing its customary in the first place. Here the contest between the two factors takes place prior to any bestowing of. Both aggregative and prevention models provide a way to explain the baseline epistemic support we often have for believing our eyes. And both models respect the conjunction of two ideas: first, that the ultimate epistemic profile of an experience depends on its relationship to other psychological states, and second, that the phenomenal character of experience bestows it with an that will stay positive, absent modulation by other factors. 19 An un-modulated experience will have positive. If a universal Epistemic Charge thesis is true, as per the phenomenal ground argument, then no experiences will be unjustified justifiers, according to the usual sense whereby unjustified justifiers are not admissible for being justified or anti-justified. If there are no unjustified justifiers in this sense, then does that entail that justification of beliefs either continues along a regress, or that a belief can only be justified by cohering with other mental states? No. If were phenomenally grounded, then the overall structure of justification could be in many respects the way that foundationalists took it to be. Phenomenal grounds could introduce self-justifying experiences. These experiences would be similar in one respect to the traditional unjustified justifiers : nothing else would justify them. But rather than being epistemically neutral, these experiences would be epistemically charged by their own features. They would be self-justifying justifiers. Could a belief be immediately justified by an epistemically charged experience? A subject S s belief that p is immediately justified by an experience, just in case there need be no further propositions that S must be justified in believing from a source other than P in order for experience to justify her in believing that p or if there are, being justified in believing those propositions need not play a role in S s getting justification to believe p from her experience. Consider first whether an experience with content P could immediately justify a belief that p. Here, there won t be any other proposition besides P that S has to be justified in believing, in order for her experience to justify her in believing that p. The situation is no different when we consider the role of epistemically 19 Either model may deserve the label phenomenal conservativism more than dogmatism does, as it is more closely analogous to epistemological conservatism, the thesis that having a belief makes it pro tanto justified. The prevention model stays closer to the metaphor of electricity, but that doesn t seem to count against it. 20

charged experience with content p (a P-experience) in justifying a subject in believing a different proposition q. Even if justification from the P- experience for believing P may play a role in providing the justification for believing Q, this justification need not come from a source distinct from experience. Even epistemically charged experiences can immediate justify beliefs, and those beliefs, together with the experiences that immediately justify them, could form a foundation for knowledge. The phenomenal ground argument develops the idea that epistemic charge is grounded in the fact that experience belongs to one s outlook on the world. However, one might ask whether highly inattentive experiences contribute to one s outlook in a way that grounds. 20 One might try to develop the outlook idea slightly differently, by excluding highly inattentive experiences, and adjusting P2. But even this adjusting would make the scope of far-reaching, so that epistemically charged experiences would be standard rather than exceptional. Besides phenomenal grounding, a different way to develop the idea that is grounded in the fact that experiences belong to one s outlook on the world is that is grounded in inference. We draw inferences from information (including misinformation) we have already, and that information is part of our overall outlook. On this proposal, too, there is reason to think that if any experiences are epistemically charged by inference, then being charged is standard rather than exceptional. To see why, suppose that Vivek s experience is not untouched by his vanity, and that inference is the vehicle by which Vivek s vanity epistemically affects his experience. We can now appreciate two different senses in which inference could give his experience a negative : the fact that Vivek s experience arises from inference either adjusts an that is grounded in something other than inference; or this inferential etiology is the of ground epistemic charge. If inference plays a grounding role, then generalizing from Vivek s experience, the scope of depends on the scope of inferential routes to experience. Vivek s sort of projection onto experience may not be widespread. But on the hypothesis we re considering, projection per se isn t the grounding factor inference is. At a minimum, some experiences would have due to resulting from inference, as per the following argument: The inferential ground argument 20 For discussion of the epistemic role of highly inattentive experiences, see Siegel and Silins. 21

P1. Some experiences result from inference. P2. If an experience results from inference, then the inference gives the experience an. Conclusion: Inferred experiences have an. The wider the range of experiences that result from inference, the more far-reaching the scope of the. At an extreme, inferential grounds could support the Universal Epistemic Charge thesis. The ubiquitous-inference argument P1. All experiences result from inference. P2. If an experience results from inference, then the inference gives the experience an. Conclusion: All experiences have an. Why think that if any experiences are epistemically charged by inference, then being charged is standard rather than exceptional? Vivek s inference is unconscious, he is unaware that his experience results from his other beliefs (e.g. the beliefs that go with his vanity or his diffidence), and he may not even be aware that he has those other beliefs. We can find what seem to be similar inferences leading to perceptual experience in a wide range of cases. Those inferences are unconscious, perceivers are unaware that experiences result from them, and are often unaware that they have representations (whether they are beliefs or some other form of stored information) that the inference is drawn from. For instance, Helmholtz provided strong evidence that the visual system stores the information that light comes from above, and that this information helps explain why we end up with a perceptual experience of convexity when looking at concave surfaces. The assumption that light comes above seems to be a starting assumption, rather than one that the perceiver learns. But other evidence suggests that the character of perceptual experience can be heavily influenced by information that the subject learns. Familiarity with the typical color of things (smurfs, bananas, mailboxes, etc) seems to influence the apparent colors of things. Here, information that F s tend to have color C (bananas, are banana-shaped and textured things tend to be yellow) seems to influence what color a banana looks to have when one sees it. Training with certain conjunctions of features (such as shape and position) can make perceivers treat the conjunction as a unit, thereby influencing perceptual experiences of uniformity. 21 Here, information that features F1 and F2 tend to be found together influences how much uniformity one seems to find in a scene. 21 Witzel et al 2011, Goldstone 1998 and 2015. 22

To make the case for the continuity between these processes and the inferences that carry out Vivek s projection, the nature of inference would have to be examined more fully, so that we could better assess whether the processes leading to Vivek s experiences, or to perceptual experiences in general, are inferences of the kind that have epistemic effects. For instance, one might object that regardless of whether inference is a vehicle for projection, the kinds of inference that lead to experiences are not the kind that can have epistemic effects. Something in the nature of the inferences that generate experiences precludes them from having epistemic impact on experiences. Alternatively, one might object to generalizing from the idea that Vivek s projection makes his experiences epistemically evaluable, to the conclusion that inferences that operate over processes internal to the visual system are epistemically evaluable. 22 A fuller defense of the continuity among these kinds of inferences would meet these objections. But prima facie, inferences from perception and inferences to perception, and the inferences to perception from other perceptual states and from beliefs are similar in ways that lay a good claim to grounding epistemic evaluability all around. We have already noted some similarities between inferences to perception and from perception: they can be unconscious, sensitive to semantic and logical features, and their inputs need not be states that the subject is aware of being in. On the inference to perceptual states from other perceptual states, the mechanisms by which the inferences from perceptually learned are no different from the mechanisms by which beliefs are formed. Once a single kind of inference is pinned down, so that there is no equivocation on inference between the premises in each Inferential Ground argument, premise 1in each argument is an empirical premise. Premise 2 raises the same questions as the key premise raised in the phenomenal grounds argument. What kind of does an inferential route to experience leave an experience with? Presumably, it would depend on the epistemic status of the inferential inputs. If so, then a single factor, inference, makes it the case that an experience is charged rather than neutral, and determines the specific that the experience has. A single factor would ground and modulate. How would inferential grounds of influence the global structure of justification? The answer depends on what gives the inferential inputs forward-looking power to pass on epistemic status to experiences. Here the options are the same as the ones we find when we ask about the ultimate sources of justification of belief: (i) there is an infinite 22 [Zeimbekis papers. For replies that go further than those in the text, see Jenkin (in preparation) and Jenkin and Siegel (2015). 23

chain of inferences, or (ii) a non-linear structure by belonging to which a belief can be justified, or (iii) a structure that contains some elements that provide their own. On the face of it, it might seem that the thesis on its own would impact the global structure of justification. In particular, it might seem to stack the deck against foundationalism, by robbing that position of its unjustified justifiers. On closer examination, however, reconfiguring the domain of epistemic normativity by itself doesn t determine what the structure of justification can or cannot be. If any inferential inputs to experience have forward-looking power to influence the of the inference s conclusions, then from this assumption and the superficial similarities noted earlier, we can draw a provisional conclusion: if the inference epistemically effects Vivek s experience, then inferences that aren t vehicles for projection have analogous epistemic effects. If this epistemic effect is grounding, then the scope of is as widely distributed across experience as inferential routes to experiences. On this proposal, the more experiences arise from inference, the more extensive the range of epistemically charged experiences will be. If inference grounds the of experiences, and there are not other grounds, that seems to rule out forms of foundationalism where the foundation includes an experience. But foundationalism itself is still in the cards, as is any form of coherentism. The inferential ground theses do not determine what other factors modulate the of experiences just as the fact that one infers a belief from other mental states is potentially just one factor among many that determines the ultimate epistemic standing of the belief. I haven t argued fully or decisively for the Epistemic Charge thesis. But I ve explored what else would have to be true, if the Epistemic Charge thesis is. When we consider experiences such as Vivek s that seem to be epistemically evaluable, perhaps the correct philosophical suspicion is that all experiences are epistemically evaluable. We just didn t see it, because we weren t focusing on the cases where experiences are made irrational by their etiology, and we were presuming that experiences differed from beliefs more than they do. 24

the aristotelian society president: Adrian Moore (Oxford) president-elect: Susan James (Birkbeck) honorary director: Rory Madden (UCL) editor: Matthew Soteriou (Warwick) lines of thought series editor: Scott Sturgeon (Oxford) executive committee: Corine Besson (Sussex) / Kimberley Brownlee (Warwick) Rowan Cruft (Stirling) / Alison Hills (Oxford) / Samir Okasha (Bristol) / David Papineau (KCL) Robert Stern (Sheffield) managing editor: Lea Salje (UCL) assistant editor: David Harris web designer: Mark Cortes Favis administrator: Hannah Carnegy (UCL) www.aristoteliansociety.org.uk