Jonathan Jenkins Ichikawa
Jennifer Windt’s (2015) Dreaming is an enormously rich and thorough book, developing many important and illuminating connections between dreaming, the methodology of psychology, and various subfields in philosophy — especially the philosophy of mind, the philosophy of psychology, the philosophy of cognitive science, and epistemology. In this commentary, I’ll focus on two of the epistemological threads that run through the book.
The first thread I wish to discuss, which is considered most directly in chapters 3 and 4, has to do with the status of certain assumptions about the nature of dreams in the methodology of dream research. Windt argues that some such assumptions — such as the assumption that dreams involve experiences, and that dream reports are reliable — are methodologically necessary default assumptions, akin to Wittgensteinian hinge propositions. I will suggest that Windt is quietly presupposing some rather strong sceptical assumptions, and that recent literature in epistemic externalism may bear in important ways on her arguments.
The second thread I will consider involves the perennial sceptical worry that dreaming threatens ordinary knowledge: it seems quite clear to me that right now I am sitting in this hotel bar writing a commentary on Windt’s book; but doesn’t the experience of dreaming rule out my belief to that effect as knowledge? If I were having a realistic dream about writing a commentary, wouldn’t things seem exactly the same? Is there any way to rule such possibilities out? Once again, my suggestion will be that Windt is making some tacit sceptical assumptions that there might be good reason to resist. Or if we are not to resist them, I hope we may at least be convinced to lay them out clearly for critical examination.
One of the running themes in Windt’s book is a critique of dream theorists, both within philosophy and psychology, who have mistakenly assumed central elements of what Windt calls a ‘Cartesian’ characterization of dream experience. According to this conception, dreams fully replicate the phenomenal character of waking experience. Windt’s critique of this tradition seems to me wholly correct. But if I am right about the two main threads described above, Windt’s departure from Cartesian assumptions is only partial. Although she carefully and insightfully articulates and critiques Cartesian assumptions about the nature and phenomenology of dream experience, I suspect that she is persisting uncritically in some substantive Cartesian assumptions about epistemology — the idea, for example, that our evidence, and the basis for any rational beliefs, is ultimately exhausted by our phenomenal experience. I think some such assumptions play important roles in the epistemological discussions of the book. I would like to draw them out for scrutiny.
2. Methodological Assumptions of Dream Research
One of the central aims of the first several chapters of the book is to articulate some of methodological assumptions that underwrite dream research. Among these assumptions are these:
The Experiential Assumption: Dreams are conscious experiences in the sense that they are phenomenal states, or that it is like something to dream. (Windt, 2015, p. 5)
The Transparency Assumption: Dream reports are epistemically transparent in the sense that they are trustworthy sources of evidence about the occurrence and phenomenal character of experience during sleep. (ibid., p. 6)
Windt will argue that researchers proceed properly in making these assumptions, despite her contention that they aren’t themselves experimentally confirmed, or even confirmable. Instead, Windt will argue that these assumptions are methodologically indispensible as assumptions, foundational in the Wittgensteinian sense. The attempt to establish them is a misplaced one; their status as assumptions is a necessary condition for anyone who wishes to investigate dreaming at all. Even to wonder about the truth of these assumptions, Windt says, would be to engage in a kind of ‘pseudoquestion’. Windt writes:[D]oes my claim that the experiential character of dreams and the transparency of dream reports have the status of methodologically necessary default assumptions retain its plausibility in the face of actual examples from dream research? Or are there any cases in which dream research can be said to yield straightforward empirical evidence confirming the experiential and transparency assumptions, thus once more turning them into genuine hypotheses? I will argue that the latter is not the case: scientific dream research is best regarded not as confirming the experiential and transparency assumptions but as assuming them, and as a matter of methodological necessity… [Q]uestions about the experiential character of dreaming and the transparency of dream reports are pseudoquestions in the sense that they are artifacts of a misconceived philosophical problem. (ibid., p. 73)
My main concern here will be with Windt’s discussion of the epistemological status of the experiential and transparency assumptions, but first I’ll permit myself a paragraph on the suggestion that they are methodologically necessary assumptions. Windt takes this expression seriously — she is clear throughout the book that she takes these assumptions literally to be indispensible for dream research; there could be no science of dreaming without these assumptions, Windt says. This strikes me as too strong. Perhaps dream research would be impoverished without these assumptions, but I don’t see the case that it couldn’t exist at all. Indeed, some of the research Windt cites — such as research relating to non-human animals’ dreams — does not seem to depend in any obvious way on the assumption that dream reports are trustworthy. On p. 45 Windt notes a particular tension in some of Aristotle’s remarks on dreaming: on the one hand, Aristotle held that dreaming only occurs during sleep, and that sleep is marked by a lack of behaviour. This is why Aristotle’s view (in this respect like Windt’s, though not in others) privileged the importance of dream reports. But on the other hand, Aristotle held that non-human animal sleep behaviour — dogs barking in their sleep, for example — indicated dreaming. As Windt observes, indications of dreams in animals allow for the study of dreams without dream reports. But Windt does not extend this thought to the status of the allegedly methodologically necessary assumptions. Chapters 3 and 4, where the discussion of the status of these assumptions is foregrounded, does not discuss dreams in non-human animals.
What of Windt’s suggestion that questions about the assumptions are ‘pseudoquestions’? Windt is not explicit about what exactly she means by the term. In the quote above they are glossed as ‘artifacts of a misconceived philosophical problem’; elsewhere she sometimes leaves off the ‘misconceived’ qualification, or characterizes pseudoquestions as ones in which ‘no straightforwardly empirical answer is forthcoming’ (ibid., p. 40). There doesn’t seem to be any attempt in the book to deny that the experiential and transparency assumptions express propositions with a straightforward metaphysical status — Windt is not claiming, for example, that there’s no clean distinction between possible worlds where people have experiences when they dream and those where they don’t, or that talk about whether dreams involve experiences is disguised discourse about linguistic conventions. The point seems to be epistemological, arguing that in some sense the assumptions cannot be empirically established, or at least, they can’t be established without begging central questions against one who doubted them. Her conclusion is that
the truth of these assumptions cannot itself be the object of scientific investigation: attempts to corroborate or verify the trustworthiness of dream reports (or the reportability of dreams) by independent means are either circular or demonstrate their own insufficiency by inviting inescapable skepticism about dreaming and dream reporting. (ibid., p. 139)
But Windt’s stance isn’t wholly sceptical. Although she holds that these assumptions cannot be established scientifically, Windt does think that they are ‘theoretically justified’. In chapter 4 of the book, Windt defends the experiential and transparency assumptions via inference to the best explanation. These assumptions, she says, best cohere with the observed data. At its simplest, the abductive argument is used to account for the very existence of dream reports:[T]he master argument… goes as follows:
(1) The surprising fact, C, is observed: upon awakening, people often report having had experiences during sleep!
(2) But if A (the transparency assumption) were true, C would be a matter of course.
(3) No available competing hypothesis can explain C as well as A does.
(4) Hence A is true: dreams are experiences during sleep, and dream reports transparently show this to be the case. (ibid., p. 145)
Windt proceeds with a detailed examination of much of the empirical literature to make the case for the key premise (3), comparing her approach to those, such as Norman Malcolm’s and Daniel Dennett’s, which demur. I found her case here compelling. Windt observes, for example, that lucid dreamers can be trained to communicate with dream scientists during sleep about their dream experience, and that the apparent experiences described by this method cohere with waking reports. While it isn’t impossible to provide a story that predicts this coherence without positing dream experience, it is rather difficult to do so without positing complex mechanisms for which there is little evidence, and which seem rather ad hoc. By contrast, positing dream experience and transparent reports of it is has much more ‘explanatory loveliness’.
But Windt is careful not to claim on these grounds that this amounts to a verification of these assumptions:
In defending the explanatory loveliness of the received view, it might seem that I am claiming that the transparency and experiential assumptions can, after all, be verified. But as we have seen in chapter 3, this is not the case: no independent evidence for the trustworthiness of dream reports is forthcoming. (ibid., p. 155)
The assumption implicit in the last sentence of this quotation is that verification requires independent evidence. But this is a substantive epistemological assumption. There is room in epistemological space between scepticism and non-circular refutations of sceptics. One needn’t hold that questioning the epistemic status of a given assumption is a confused product of mistaken philosophical theorizing, just because one wants to deny scepticism even though certain sceptics can’t be refuted on their own terms. Indeed, the preponderance of contemporary epistemologists hold that this situation holds quite ubiquitously.
Consider perception generally, even setting aside any worries particular to dreaming. It is overwhelmingly natural to think that many facts about one’s external environment may be verified, learned, and known, via perception. At the moment I know via perception, for example, that it is raining outside and that there is a woman in a beige coat sitting beside me. But as everyone who has taken an introductory epistemology course knows, it is not at all obvious that there is any non-question-begging way to defend this claim against a certain sort of sceptic — one fixated on a Cartesian evil demon scenario, for instance. This is the kind of reasoning that leads epistemologists like James Van Cleve (2003) to conclude that epistemic externalism — roughly the idea that knowledge or justification requires some kind of first-personal awareness of one’s epistemic position — is the only alternative to scepticism.
The point extends into the a priori realm as well. It is a familiar thought in the epistemology of the a priori that the validity of basic logical inferential patterns can only be established by invoking those patterns themselves. One can give a soundness proof for modus ponens, for instance, but one can’t do so without engaging in conditional reasoning.
Or consider the absurdly sceptical position that holds that there are no reasons to believe anything whatsoever. Trivially, one cannot give a non-question-begging argument against such a scepticism. But that doesn’t mean the scepticism isn’t mistaken, or even that the anti-sceptical stance can’t be established by invoking reasons.
So it’s not very plausible in general that one can only establish that which one can independently verify. Whether a demonstration is independent depends on the (perhaps idiosyncratic) commitments of one’s interlocutor. Consequently, I think Windt might have been less concessive to the sceptic about the experiential and transparency assumptions. Perhaps they can be established, albeit not independently. There are various ways this sort of suggestion might be developed; here is one. Many contemporary epistemologists will accept these two principles:
Introspective Principle: People generally know what kinds of experiences they’ve recently had.
Testimonial Principle: When people know things, and tell them to others, those others can come to know them too.
Whether principles like these constitute epistemological bedrock, or whether they themselves are established by further epistemological principles, is a subject of controversy. The important thing is that they’re general non-sceptical principles that are commitments of epistemological orthodoxy. Notably, this is so on grounds that are general — not specific to dreaming or any other subject matter.
The introspective principle and the testimonial principle imply Windt’s ‘transparency assumption’. Dream reports are trustworthy guides to what experiences, if any, subjects have while dreaming. So I do not see that Windt needs to consider that assumption to be an unestablished bedrock of dream research — it may simply be a consequence of epistemological principles that are plausible on more general grounds.
The experiential assumption is not implied by the introspective and testimonial principles alone, but it does follow from them combined with the obvious empirical fact that people often report dream experiences. This seems to me to be the right result. Contrary to the letter of Windt’s suggestion, the experiential assumption is not an unestablished axiom for dream research — its status depends importantly on the observed fact that people report dream experiences. Suppose counterfactually that everybody reported that they’d had no experiences while asleep. This, I’m sure Windt would agree, would amount to substantial disconfirmation of the experiential assumption. So the experiential assumption is confirmed by the dream reports we do observe.
3. Dreams and Waking Knowledge
I now turn from third-personal questions about how scientists can study dreams to first-personal questions about how one should think about the possibility that one is currently dreaming. Such thoughts have figured centrally in many important epistemological projects, especially as motivations for scepticism, as in this passage from al-Ghazāli, c. 1100 CE:
For a brief space my soul hesitated about the answer to that objection, and sense-data reinforced their difficulty by an appeal to dreaming, saying: ‘Don’t you see that when you are asleep you believe certain things and imagine certain circumstances and believe they are fixed and lasting and entertain no doubts about that being their status? Then you wake up and know that all your imaginings and beliefs were groundless and unsubstantial. (al-Ghazāli, 2000, p. 22)
(Descartes’ more famous version of the argument, given some 500 years later, is in Meditations I; Descartes, 1985, p. 13.) The general form of this kind of sceptical argument runs thus:
- The experiences I am currently having, apparently the result of my perceptual engagement with the external world, are phenomenally exactly like the kinds of experiences I sometimes undergo while dreaming.
- Consequently, I have no way to know that dreams aren’t responsible for my current experiences.
- So I have no perceptual knowledge of the external world.
Windt denies the first premise in this argument — she holds that dreams do not typically involve the same kind of experience as waking experience. In this respect — though not in all respects — she is in agreement with philosophers such as Ernest Sosa and myself. See e.g. Sosa (2007, chapter 1) and Ichikawa (2009), which defend an imagination model of dreaming, holding that dreams involve imaginative experiences rather than perceptual ones. (Windt will disagree on these specifics, going on to develop a view on which dreaming ought to be considered in important ways sui generis, not to be understood in terms of waking belief or imagination.)
Sosa uses the differential experience claim to rebut dream scepticism. He holds that when we dream, we do not typically believe the contents of our dreams; instead, we undergo a different kind of mental state. Call this a ‘dream belief’. (Whether it is a propositional imagining, or sui generis, or something else, we can remain neutral on, so long as we deny that it’s a genuine belief.) Sosa concludes that dreams are not, after all, a possible source of error, since they don’t involve false beliefs. Believing that one is awake is a method that guarantees avoiding error via dreaming.
Windt’s response to Sosa is that even if one’s dream beliefs aren’t beliefs they still feel like beliefs — they share belief’s phenomenology, as demonstrated by the fact that dreamers are typically unable to recognize the difference between dreaming and being awake. And this, Windt says, is where the epistemological action is. If it feels like a belief, then, for all epistemological intents and purposes, it’s a belief. And so likewise for perceptual experience.
Here are two illustrative passages — one about perceptual experience, and one about belief.
Perception is commonly used as a success word. Visual experience, for instance, only counts as perceptual if it is sufficiently accurate with respect to the mind-independent objects it represents and appropriately sensitive to changes in the subject’s actual environment. Because quasi-perceptual experience, as I use the term, only refers to the phenomenal character of an experience, it can be applied to illusions or hallucinations that from a third-person, epistemological perspective do not count as genuine instances of perception. To say that dreaming involves quasi-perceptual experience, then, is to say that dreaming is phenomenally indistinguishable from genuine perception. In this view, dreams and waking perception, but also illusions and hallucinations, form a common kind of mental state (cf. Macpherson, 2013): they are all quasi-perceptual experiences in a phenomenological sense. (Windt, 2015, p. 199, my emphasis)
For present purposes, the main point is that thinking at least on occasion occurs in one’s head, metaphorically speaking, and it is like something to think in this manner. This subgroup of cognitions in the phenomenological sense is what Descartes is interested in, and for present purposes, it is enough to note that theorists of drifferent stripes can accept their existence. The upshot is that to deny that dreaming involves belief in the relevant sense, one will have to deny that dreaming involves the phenomenology of believing, but also more generally of engaging in cogitations. As long as dreams involve the phenomenology of believing and of cogitating more generally, they induce epistemic discomfort. (ibid., p. 407)
Note Windt’s emphasis on the epistemological significance of the phenomenological. Windt’s idea here is that even if mental states differ in kind, if they feel the same then they are alike with respect to their epistemic significance. This is a strong internalist assumption. Notice, in the second italicized passage above, that Windt seems to be treating sameness of phenomenological character as tantamount to indistinguishability. This kind of inference occurs frequently in the book (more on this below). If it is possible to know things on the basis of something beyond phenomenology, then sameness of phenomenology does not imply sameness of epistemic access.
As Windt points out (pp. 481–2), there is a parallel here between her own sceptical remarks about dreams and some classical objections to Descartes. Descartes held that clear and distinct perception was an epistemically secure guide to truth — that which one can clearly and distinctly perceive, one cannot be wrong about. But a sceptic might worry at the higher order: what if one mistakenly thinks that one clearly and distinctly perceives something? Arnauld famously pressed this worry in his Fourth Objections to Descartes. If this is possible, says Arnauld, then relying on clear and distinct perception isn’t infallible after all (Descartes, 1985, p. 142).
Sosa holds that judging that one is not dreaming, on the basis of one’s experience, is a foolproof way to judge truly; paralleling Arnaud, Windt worries that it is possible to be in a dream state one is unable to recognize, and merely dream-believe that one is awake. Since being in such a state is obviously no conclusive indication of wakefulness, Windt rejects Sosa’s stance in much the same way that Arnaud objected to Descartes’.
I have two concerns with this argument, each having to do with potentially sceptical and internalist epistemological assumptions. First, since in general there is a difference between X and it seeming as if X, so too is there a difference between the epistemological thesis that X justifies some belief, and one that says a seeming as if X justifies that belief. When Sosa says he knows he’s not dreaming because he believes that p, or because of his perceptual experiences, and he wouldn’t believe that p or have those experiences if he were dreaming, he’s citing the belief itself, or the experiences themselves — not its feeling phenomenologically to him as if he has the belief or the experiences, nor his being in a state he’s unable to distinguish from that of having the belief or the experiences. When Descartes cites clear and distinct perception, that’s a different argument from one that cites the appearance as of clear and distinct perception.
There is a natural internalist tendency to restrict our evidence to that which we can be sure about, or that which is immune from certain kinds of sceptical doubts. But whether we should succumb to that tendency is a contentious epistemological question; following mainstream externalist epistemological orthodoxy, I am among those who think we should not. If this is right, then the move from the evidence claimed (the belief, the clear and distinct perception) to an appearance of the evidence amounts to an unnecessary epistemological restriction.
It’s also worth considering G.E. Moore’s (1939) famous ‘proof of an external world’ in this context. Moore cited the fact that he had hands as a way of ruling out the hypothesis that there is no external world. An insistence that he cite his appearance as if he had hands would beg the question against Moore. Just this distinction is crucial to the currently widespread neo-Moorean externalist epistemological stance.
So once again, I think that Windt may be assuming a more internalist, and a more sceptic-friendly, approach to evidence and epistemology than is required. Indeed there is a sense in which it is even more Cartesian than Descartes was — Descartes wouldn’t accept external world perceptual contents as given, but he at least accepted that he had the thoughts he did — Windt wants to deny even this: all Descartes gets to go on is that it seems as if he has the thoughts he does.
A brief passage later in the book foregrounds Windt’s internalist commitments on this score. She responds to Descartes’ own clarification that he intends clarity and distinctness themselves, rather than the appearance as if one has them, as the markers of truth, thus:
The impatience manifest in [Descartes’] reaction obscures the point that ‘clarity’ and ‘distinctness’ are success words: essentially, Descartes is admitting that a thought only counts as clear and distinct if it tracks the truth. This is important, because if true, it makes the criterion of clarity and distinctness unusable because from the first-person perspective, we can never distinguish between correctly and incorrectly applying it. (Windt, 2015, p. 481)
But it simply doesn’t follow from the fact that clear and distinct perception of p requires the truth of p that we can never distinguish first-personally whether we are clearly and distinctly perceiving something. This would follow only given the rather sceptical assumption that we can never know first-personally whether p. Notice that ordinary perception, for example visual perception, is also veridical — one cannot see that the plane has landed unless the plane has landed. (Windt herself pointed this out in the p. 199 passage quoted above.) But none but a radical sceptic would deny on these grounds that there is no first-personal way to know whether one sees that the plane has landed.
The final topic I wish to consider involves getting clearer on the notion of indistinguishability. For two things to be indistinguishable to a subject is for the subject to be unable to tell them apart — to know which is which, as it were. I have already remarked on the close tie Windt seems to be assuming between indistinguishability and phenomenology. Here is another passage illustrating that assumed tie:
Already it would seem that the deception hypothesis involves two separate claims, where the first concerns the phenomenal character of dreams, and the second the involvement of false beliefs.
Deception hypothesis: Dreams are deceptive experiences because they are subjectively indistinguishable from wakefulness and because they involve beliefs, some of which might be false. (Windt, 2015, p. 460, my emphasis added)
The second italicized passage here cites indistinguishability, but is treated as closely connected to — as a way of spelling out — the first italicized passage, which is about phenomenal character. Again, a Cartesian picture of evidence seems to be assumed.
Here is a different way of expressing the critique. The logic of distinguishability admits of additional interesting possibilities. Notice that this argument form, though tempting, is invalid:
- Subjects in state X can be mistaken about whether they are in state Y.
- So X and Y are indistinguishable.
- So people in state Y can’t know that they’re in state X.
The substitution instance of this form where X stands for dreaming and Y stands for wakeful perception is in effect Windt’s argument against Sosa. But remember that distinguishability amounts to a capacity for a certain subject to know a certain content (viz. which state one is in). As such, it is necessarily relative to the subject’s position and abilities. Some people, in some circumstances, are capable of knowing things that other people, in other circumstances, aren’t. So it is entirely possible, consistent with someone in X’s inability to know they’re not in Y, that someone in Y is perfectly capable of knowing they’re not in X. To take one of Sosa’s starker examples, let Y be your current condition and let X be the possibility that you ceased to exist this morning — say, because you were hit by a truck and died. Your evidence rules out X, but someone in position X would be unable to rule out Y.
Late in the book Windt does mention something like this possibility, describing it as a view according to which dreamers suffer delusions. She writes for example that[i]f our discriminative abilities are, while dreaming, crippled by the instability of dream imagery, then we might fail to realize that we are dreaming even if dreams have different contents and a different phenomenological profile from waking experiences (cf. Hill, 2004a). This type of view immediately invites a nosological analysis, placing dreams in the vicinity of madness and delusions. Call this the delusion hypothesis. (Windt, 2015, p. 462)
But delusions aren’t the only examples of possible cases where one lacks epistemic abilities that one otherwise has — deceptive scenarios generally have this form. (Sometimes, when I have had too much to drink, I’m unable to tell that I’ve had too much to drink. But it doesn’t follow that I’m never able to tell whether I’ve had too much to drink. For example, when I am completely sober it is quite easy for me to know that I haven’t had too much to drink.)
But this seems to be just the kind of argument form Windt is running, e.g. here:
If there are indeed dreams in which I erroneously take myself to be awake because of the realistic phenomenal character of dreaming, then I can never be sure that I am now awake. Even if such dreams are rare, I can never exclude that I am not subject to this type of deception at this very moment. Likewise, if sometimes while dreaming I come to believe that I am awake based on a faulty line of reasoning that I nonetheless take to be rational and hence reliable, then I can never tell whether my beliefs are founded on rational reflection rather than on confabulatory reasoning. (ibid., p. 504)
To reiterate: my suggestion is that the fact that I might be wrong while dreaming doesn’t imply that I might be wrong while waking; moreover, the implicit restriction of the evidence to the phenomenal character reflects an internalist, sceptical epistemological assumption.
I have focused on epistemological points of disagreement because they are the points for which I felt I was best placed to engage in a way likely to be interesting and useful. One of the virtues of Windt’s book, in my view, is that she is sensitive to the many points at which her work on dreaming and cognitive science touches on other philosophical and psychological domains. As the other contributions to this symposium will no doubt make obvious, there is much, much more to this wonderful book than the epistemology, and I learned a great deal by reading it. But for this particular epistemologist author, I was pleased to find that there was plenty of epistemology to chew on.
al-Ghazāli, M. (2000) Al-Ghazali’s Path to Sufism, his Deliverance from Error al-Munqidh min al-Dalal, McCarthy, R.J. (trans.), London: Fons Vitae.
Descartes, R. (1985) The Philosophical Writings of Descartes, vol. II, Cottingham, J., Stoothoff, R. & Murdoch, D. (trans.), Cambridge: Cambridge University Press.
Ichikawa, J. (2008) Scepticism and the imagination model of dreaming, Philosophical Quarterly, 58 (232), pp. 519–527.
Ichikawa, J. (2009) Dreaming and imagination, Mind and Language, 24 (1), pp. 103–121.
Ichikawa, J. (2016) Imagination, dreaming, and hallucination, in Kind, A. (ed.) Routledge Handbook of Philosophy of Imagination, pp. 149–162, London: Taylor & Francis.
Ichikawa, J. (2017) Contextualising Knowledge, Oxford: Oxford University Press.
Kornblith, H. (2004) Does reliabilism make knowledge merely conditional?, Philosophical Issues, 14 (1), pp. 185–200.
Moore, G.E. (1939) Proof of an external world, Proceedings of the British Academy, 25 (5), pp. 273–300.
Moran, R. (2001) Authority and Estrangement: An Essay on Self-Knowledge, Princeton, NJ: Princeton University Press.
Pritchard, D. (2012) Epistemological Disjunctivism, Oxford: Oxford University Press.
Reid, T. (1785/2002) Essays on the Intellectual Powers of Man, Brookes, D. (ed.), University Park, PA: Pennsylvania State University Press.
Sosa, E. (1999) How to defeat opposition to Moore, Philosophical Perspectives, 13 (s13), pp. 137–149.
Sosa, E. (2007) A Virtue Epistemology, Oxford: Oxford University Press.
Srinivasan, A. (2015) Normativity without Cartesian privilege, Philosophical Issues, 25, pp. 273–299.
Van Cleve, J. (2003) Is knowledge easy — or impossible? Externalism as the only alternative to skepticism, in Luper, S. (ed.) The Skeptics: Contemporary Essays, Farnham: Ashgate.
Williamson, T. (1990) Identity and Discrimination, Malden, MA: Blackwell.
Williamson, T. (2000) Knowledge and Its Limits, Oxford: Oxford University Press.
Windt, J. (2015) Dreaming, Cambridge, MA: MIT Press.
 ‘…trusting dream reports is, in this view, a methodological necessity and the condition of possibility for scientific dream research’ (Windt, 2015, p. xxii).
 ‘…a creeping suspicion arises that [the problem of dream experience] might be a pseudoproblem, an artifact of philosophical debate’ (Windt, 2015, p. 38).
 See e.g. p. 137.
 This stance is very widespread; it probably deserves the title of contemporary epistemological orthodoxy. Prominent defenders include Sosa (1999) and Williamson (2000). Moore (1939) is an important earlier proponent of this kind of stance — as is Reid (1785/2002).
 As Windt uses the term, that dream reports are ‘transparent’ just means that they are trustworthy; it is not meant to connote, as e.g. Moran (2001) holds, that beliefs about one’s experience are derived from judgments about the external world.
 As Windt observes (pp. 257–60, 408–9) I offered much the same critique of Sosa in Ichikawa (2008). My own stance has shifted in the past decade; I am much more sympathetic to Sosa’s line now than I was then.
 Examples include reliabilists like Sosa (1999), disjunctivists like Pritchard (2012), and knowledge-first approaches like Williamson (2000, chapter 8) and Ichikawa (2017, chapter 3).
 I have developed this line of thought in general in Ichikawa (2017, pp. 103–4), and with respect to dreams in particular in Ichikawa (2016).
 On externalism, normativity, and higher-order knowledge of knowledge, see Kornblith (2004) and Srinivasan (2015).
 See especially Williamson (1990).