Declan Smithies --- The Phenomenal Basis of Epistemic Justification
New Waves in Philosophy of Mind
Mark Sprevak
Hi Declan! I really enjoyed your paper. It’s brilliantly written, and something I haven’t thought about at all before—very exciting stuff.
I’ve got a really inchoate question. I’m not sure how to state it or even if it is relevant. Please ignore this if it is not useful!
On my very poor understanding, one of apparent motors of your argument seems to be the requirement that that any epistemically-relevant difference should show up from the subject’s point of view. For example, the reason why a super-blindsighter doesn’t count as having justified visual beliefs is that ‘there is nothing from within his subjective point of view on the world to distinguish [his visual beliefs] from beliefs forms on the basis of blind guesswork.’
However, I was wondering if you are loading the dice a little by understanding a subject’s point of view in this particular way: namely, as the the contents of the subject’s phenomenal consciousness. The super-blindsighter has access consciousness of the visual data, but that data isn’t phenomenally conscious, so the data doesn’t count showing up in his ‘subjective point of view’.
But there seem other ways of understanding a subjective point of view that yield different results from your view, but are still compatible with epistemic internalism (and distinct from reliabilism).
For example, one might think that many cognitive systems have a ‘point of view’ on the world—an overall ‘take’ on the state of the world in relation to them and how they should act in it. This may involve a collection of representations of how things in the world are related to them, their own state, and how they should react and use the info to guide action. These representations may be internal to the cognitive system, they may completely determine the system’s behaviour, but only a small portion of them may be phenomenally conscious.
I guess I was wondering, what is to stop someone from saying that we should understand a subject’s point of view in this way, and that epistemically-significant differences only have to show up in this (much wider) arena, not necessarily in the arena of phenomenal consciousness?
Bence
Hi Declan, thanks for the very interesting paper – those poor Zombies with their impoverished mental lives… But let me try to challenge you a bit about the first part of the paper. I’m not sure I understand the argument why unconscious perception couldn’t justify beliefs. You go through this twice (p. 8 and p. 11), but on both occasions you seem to say that justification only makes sense from a subjective perspective. So what happens if someone denies this? It seems that there is an unconscious version of a wide variety of mental states and processes: unconscious perception, unconscious attention, unconscious emotion, unconscious biases on decision-making, unconscious beliefs, unconscious desires, unconscious imagery (see Ian’s paper on this site), unconscious ‘intentions’ (see my paper on this site). What would stop us to talk about unconscious justification? You seem to say that that’s an oxymoron – just like talking about unconscious consciousness would be an oxymoron. But I don’t see why this would be so. I’m sure you have a worked out response to this, but I didn’t find it here (or I missed it).
Also, you appeal to the Gopnik and Graf findings as something that supports your view. But one could go the other way. If three year olds can’t tell where their justification comes from, they (presumably couldn’t tell if it came from subliminal priming either. Why are these findings not threatening your conclusion (rather than support it)?
Finally, I know you write about this elsewhere, but it was not at all obvious to me why beliefs are phenomenally individuated. Or, more precisely, it is not at all obvious to me why the intentional individuation will be equivalent to the phenomenal one. Suppose that I have a belief that x is F that is so suppressed (maybe because it clashes with so many of my other beliefs) that it does not dispose me to make any judgment at all. But the (normal) belief that x is F would dispose me to make judgments (with phenomenal character). So they seem to have the same content, but they would be different if we individuate them phenomenally. I guess you would deny the possibility of the former kind of beliefs, but if so, I’d like to hear more about why we should do so. Beliefs can sometimes cause judgments (with certain phenomenal character) and they can also sometimes justify other beliefs. I don’t see any principled reason why there couldn’t be beliefs that can do the latter but not the former.
Robert O'Shaughnessy
Hi Declan
Thank you for a challenging and very rewarding paper. You’ve set out some tools that are really useful for me in thinking about these issues. There are many questions I would like to ask you but I will limit myself to seeing if I can challenge what I take to be one of your central points: that when the phenomenality of perceptual experience is lacking one also lacks epistemic justification for beliefs about what is perceived.
Consider playing one of those text based adventure games which used to be quite common back in the day. All that appeared on the screen was text, there was no graphics. Each time you entered a room or a dungeon you would be presented with a list of propositions describing the location, its contents, and what was going on, and a list of actions that the environment afforded you (e.g. exit south; pick up book etc.). It seems to me that the player would be entirely justified in believing the propositions encountered (provided they turned out to be mostly true) even though he had no perceptual phenomenality of what was the source of these judgements.
The predicament of the player in this kind of game seems to be exactly parallel to that of the kind of superblindsigher you describe (instead of ‘text’ appearing on the screen, the equivalent appears directly in her mind). She would soon learn that she was justified in believing the propositions she was presented with provided they were accompanied by the high level of confidence that you describe on page 11 (the propositions could also be presented with lesser levels of confidence (maybe it’s a bird, maybe it’s a plane, no, definitely it’s Superman) and those with lesser levels she would not be justified in believing).
You then say that “mere feeling of confidence is not sufficient to justify belief –justification is not so easy to come by.” What you seem to have in mind here is that the superblindsighter does not know the source of the evidence for her belief. But later you say that one would be justified in believing that Canberra is the capital of Australia even if the evidential source of this belief was not known. You say the justification would be the mere presence of the (first order) belief. But I would say that the feeling of confidence must be doing all the work. If you are responding to a quiz question you might feel the answer Canberra is a wild guess, or something you have an inkling of, or something you are sure of; what distinguishes these is the feeling of confidence (or perhaps just some functional equivalent of such a feeling).
It seems to me that the superblindsighter is as equally justified in believing the beliefs presented to her by their mere presence (when accompanied by the high level of confidence) as the person retrieving from memory that Canberra is the capital of Australia (with a high level of confidence).
The upshot, I would suggest, is that knowing the evidential source of your belief is not necessary for epistemic justification. Epistemic justification can be as easy to come by as having a high level of confidence. Knowing the evidential source (which is very often perceptual/visual information) is not necessary so it can be entirely unconscious (or even no longer existent as in the case of Canberra being the capital of Australia).
Anyway, thanks again for a great paper.
-Robert
John Gregg
First off, let me say that I think the link between epistemology and phenomenal consciousness is under-explored, and your paper helps to address that.
I am not sure I understood all of your paper - a few more concrete examples would have helped dispel the unmoored, abstract sense I had reading it. I could not shake a creeping sense of circularity: we have justification to hold an attitude if we have justification to believe that we have justification for that attitude. (This is paraphrased, only slightly, from the definition of access internalism on page 19.) Similarly, I don't quite see the distinction between justification and ideal critical reflection, which also gives me a circular feeling when one is used to define the other. I also had trouble with the distinction between judging and "adopting an attitude toward a proposition", which seemed to be an important distinction on page 16.
On page 17, you say that "beliefs are individuated wholly by their dispositions to cause the phenomenal character of phenomenally conscious judgments, whereas subdoxastic mental representations are individuated at least in part by their role in unconscious computational processes." It seems to me that the (presumably causal) role in unconscious computational processes could certainly percolate up to having an influence on phenomenally conscious judgments, so this distinction is not crisp to me. A role in unconscious processing very well might be a disposition to cause a phenomenal effect.
More generally, I question this whole idea of justification that epistemologists use. It seems kind of Old-World to me. Do our intuitions about it indicate some natural kind, or is epistemic justification a sort of ad hoc day-to-day, functional concept?
Also, I tend to interpret phenomenal consciousness broadly. Are you really sure that beliefs are not phenomenal? I'm not so sure. Is there nothing at all it is like to believe that Canberra is the capital of Australia?
Finally, I have a Dennett-type comment, in that you seem to assume there is a monolithic "I" that is or is not justified; that believes or does not believe certain things; that holds or does not hold certain doxastic attitudes. Might the self be a bit messier than that? Might belief, like Dennett's consciousness, be more like fame in the brain or multiple drafts? Maybe we aren't as well integrated as we think, such that there might not be a single coherent set of our beliefs.
-John Gregg
http://www.jrg3.net/mind/
Philip Goff
Declan
I am very sympathetic with your paper.
In order to bring in beliefs, your position seems to require a commitment to cognitive phenomenology. I am well up for this, but I think such a commitment is more controversial than you seem to suggest in the paper (might be good to more explicitly acknowledge the need for cog phem, and link it to recent debates, notably Bayne and Montague volume). You favour a dispositional account of beliefs, but there are other ways in which fans of cog phem make sense of beliefs, e.g. Uriah Kriegal has an interpretationalist account of unconscious beliefs (which ultimately grounds non-conscious beliefs in conscious interpretation). I wonder whether your account could be open to a wider variety of ways of linking non-conscious beliefs to cognitive phenomenology?
Declan Smithies
Thanks, Mark. For precisely the reasons you give, I don’t want to rest too much weight on the platitude that justification depends on “the subject’s perspective” or “the subject’s point of view”. Sometimes I do make use of these locutions, but the question remains: which mental states figure within the subject’s perspective or point of view in such a way as to play an epistemic role in determining which doxastic attitudes one has epistemic justification to hold? That is, in effect, a version of the generalization question that I’m concerned to address in part one. So, your proposal that we can understand the subject’s perspective in broadly functional terms, without appealing to phenomenal consciousness, is in much the same spirit as access mentalism, metacognitive mentalism, and the other functionalist answers to the generalization question that I consider in part one. Still, your point remains that I can’t reject these proposals simply by appealing to an unreconstructed notion of the subject’s point of view. Instead, my strategy would be to rest the argumentative weight on the intuitive responses to examples in section one and the argument from access internalism developed in sections two and three. Let me think about how best to clarify this issue in the final version.
Declan Smithies
Thanks, Bence. This is a good question: why not talk about “unconscious justification”? One example of someone who talks this way is John Campbell in Reference and Consciousness: he talks about justification for various information-processing routines conducted by unconscious computational systems. I’m happy to talk this way too if someone insists on it. But I’d want to maintain that unconscious justification in this sense is a different kind of thing from what I’m talking about. There may be some sense in which we can evaluate the performance of unconscious computational systems, but it seems quite different from the kind of evaluation that seems appropriate for beliefs and actions. We need some kind of terminology to pick out this distinctive kind of evaluation, so if ‘justification’ is used in the broader sense, then we need some other terminology, e.g. ‘epistemic justification’ to pick out the more narrow sense that I have in mind.
On three year-old children: I don’t think they’re in the same boat as the super-blindsighter because although they’re unable to access their phenomenally individuated mental states, these mental states are nevertheless accessible to them. What I mean is that they have introspective justification to believe that they’re in those states just by virtue of being in them, although they lack the psychological capacities required to take advantage of this introspective justification in forming introspectively justified beliefs. By contrast, the subdoxastic mental representations of the blindsighter, the super-blindsighter, and the hyper-blindsighter are not accessible in this sense. (NB: accessibility here is an epistemic notion, rather than a merely psychological one.)
On suppressed beliefs: I want to recognize such cases, but treat them as cases in which the disposition to cause phenomenally conscious judgments is blocked, i.e. the conditions for the manifestation of the disposition do not obtain. Even so, these states are individuated by their dispositions to cause phenomenally conscious judgments under the appropriate manifestation conditions.
Declan Smithies
Thanks, Robert. I’m not sure I understood the comparison between the super-blindsighter and the video-gamer. The key difference, as far as I can see, is that the video-gamer is forming beliefs about the fictional game-world on the basis of her conscious experience of what she’s reading on the screen, whereas the super-blindsighter has no such conscious experience to go on. Am I missing something crucial about the example?
I think your comparison between super-blindsight and memory is a very interesting one. Ultimately, though, I want to argue that these cases are quite different, since super-blindsight is a process of belief-formation, whereas memory is a process of belief-retention. On memory, I’m inclined towards a form of epistemic conservatism on which merely believing a proposition provides defeasible justification to retain the belief in the absence of defeaters. On this view, I’m justified in retaining beliefs in memory without retaining any memory of the evidence on which those beliefs were originally formed. But it doesn’t follow that the super-blindsighter is justified in forming beliefs on the basis of a feeling of confidence alone. Here’s another way to put the point: when the super-blindsighter forms beliefs about the blind field, she has defeasible justification to retain those beliefs, but she also a defeater – namely, that these beliefs have just been formed in an irrational way. The same does not apply in the case of ordinary memory beliefs, e.g. my belief that Canberra is the capital of Australia.
Declan Smithies
Thanks, John. You’re right to point out that the paper is pitched at a fairly abstract level. Unfortunately, I’m afraid that’s in the nature of the beast, although I’ll do my best to try and make things a bit more user-friendly.
On circularity: there is a circularity in the JJ principle (i.e. if one has justification to believe p, then one has justification to believe that one has justification to believe p), but I don’t think it’s a vicious circularity. At any rate, that would have to be argued.
On subdoxastic mental representations: you’re right to point out that these unconscious mental representations might figure in processes that influence phenomenal consciousness, but I’d want to insist that, unlike beliefs, they’re not individuated wholly by their dispositions to cause phenomenal consciousness, since they’re individuated at least in part by their role in unconscious computational processes. If you’re interested, you can find a slightly more detailed discussion of this point in section 3 of my paper, “The Mental Lives of Zombies”, which is linked below:
http://philpapers.org/archive/SMITML.1.pdf
On justification: I don’t know if justification is a natural kind, and by the same token, I don’t know whether morality is a natural kind, but I do think these are among our most central dimensions of evaluation and that’s why I’m interested in understanding them, whether or not they play an important role in the natural sciences.
On beliefs and phenomenal consciousness: I don’t think beliefs are phenomenally conscious because one’s beliefs can persist even when one has no phenomenally conscious states at all, e.g. someone dreamlessly sleeping might have certain beliefs about where they live, etc. But I do think beliefs have phenomenally conscious manifestations – I call these ‘judgments’ – so, like you, I’m working with quite a broad conception of phenomenal consciousness.
Finally, on the self: I am indeed assuming that there is a unique self that is the subject of all my thoughts and conscious experiences. My mind boggles at the thought that there might be multiple selves: in that case, which of them is thinking my thoughts – all of them, or just some of them, and if so, then which ones? But I can’t pretend to have grappled with these issues properly, so I’ll have to leave them for another day!
Declan Smithies
Thanks, Philip. You’re absolutely right that my theory of belief commits me to controversial claims about the nature of cognitive phenomenology. In particular, I’m committed to the claim that judgments are individuated by their phenomenal character and that beliefs are individuated by their dispositions to cause the phenomenal character of judgments. I don’t think I’m committed to the further claim that judgments have non-sensory phenomenal character, since that depends on the further claim that judgments and beliefs cannot be individuated by sensory phenomenal character alone. My own view is that the distinction between sensory and non-sensory phenomenal character is not clear enough to make the issue tractable, so I don’t take a stand on this. But I do recognize that my claims about individuation are controversial and perhaps I should acknowledge this more explicitly in the chapter. In order to keep things manageable, I’m leaving a more extended discussion of cognitive phenomenology for another time, but I am currently in the process of writing two articles on this topic for Philosophy Compass.
Kevin Ryan
Hi Declan,
Thanks for your paper. I have several quick questions about the affective character of phenomenal individuation, although I'm not sure exactly how it fits with your current project. I also have a feeling my questions may be missing the meat of what matters in phenomenal individuation, at least in relation to epistemic justification. In short, what is the relationship between affect and epistemic justification? Is affect distinct or inseparable from the aspects of phenomenal individuation relevant to the issue at hand? Moreover, does one need to be in a dispassionate, neutral, or "rational" affective/emotional state in order to have epistemic justification? While I think many cases, such as "Canberra is the capital of Australia", can be considered without regard to affect, the group of issues I have in mind are generally instances where two intuitions seem to be at an impasse, and pointing to a purported fact of the matter "in the world" is not enough to justify one or the other.
Robert O'Shaughnessy
Hi Declan
Thanks for the reply. Sorry this is so long. Feel free to ignore or skip to the last paragraph.
The analogy between the superblindsighter and the videogamer I now see I could have made much more simply with the notion of testimony. The way I see the superblindsighter as you describe her is that she is receiving spontaneous testimony about what is going on. This is something the ordinary blindsighter never gets. He has to be cued as to what his options are by an outsider (is it a X or a O). And he never experiences getting an answer in inner speech in the form of, say, ‘it is a X’. Instead he just experiences being forced to select X or O at random (or at least that’s how they report it). The remarkable thing is that he gets it right way above chance.
However the superblindsighter is in the position of spontaneously being presented with (perhaps in inner speech) testimony to the effect that ‘there is a X on the screen’. This testimony does not have to cued by an outside questioner, nor does it have to be cued by her own questions, it just as it were pops into her mind in the form of testimony.
So rather than the videogamer, imagine you are blindfolded and there is someone telling you ‘there is a X on the screen’. This is very different from seeing the X yourself and spontaneously forming a belief to that effect. In the visual case you could go on to form many other beliefs that that identical visual information is evidence for (poised to provide) e.g. ‘one of the two black lines is longer than the other’, ‘there is a dot at the fixation point’ and so on.
In the blindfold testimony case you only gain access to the single proposition expressed in the testimony, and you do not know what the visual information that grounded it looks like. But I think from what you say you agree that one would be (defeasibly) justified in believing that there is a X on the screen provided the testifier is generally found to be reliable (compare first forming the belief ‘Canberra is the capital of Australia’ on the basis of testimony from a reliable source).
The analogy with the superblindsighter is that she is in effect receiving spontaneous testimony (in inner speech) and this source she has discovered from experience is highly reliable. The testimony in the case of the superblindsighter, I would say, is effectively “I am sure there is a X on the screen” (the ‘I am sure’ bit may be carried by a feeling of confidence).
Your intuition seems to be that the superblindsighter will always feel her own testimony is a shot in the dark in the absence of the visual evidence. My intuition is that she will treat the testimony (even though she does not know what grounds it) that ‘pops’ into her mind just like she would treat an outsider’s testimony if she was blindfolded. She will feel she is justified in believing it. To put it another way she would surely be irrational not to take it as justified given how reliable it is.
Of course one thing to note is that the title of your paper
still holds: what the superblindsighter has that the ordinary blindsighter does
not is the phenomenality of the testimony (plus perhaps the phenomenality of
the feeling of confidence). All I am challenging is your claim on page 11 that the
superblindsighter is in the same epistemic predicament as the blindsighter because of the lack of visual evidence!
-Robert
Adam Pautz
Very cool paper. I have some questions about your simple theory of introspection. I am attracted to it, but I think a more complex theory might be needed. This is stuff we've talked about in emails lately; I am bringing it up here in case others are interested, and to take some points further.
Here's a quick argument against the simple theory as you now formulate it in your paper (or at least an argument that it is incomplete).
(1) By virtue of having colors experiences C1, C2, C3, John has an introspective (propositional) justification for believing that color experience C1 more like C2 than C3
(2) The simple view says "for some [phenomenally individuated] mental states M, necessarily, one is in M if and only if one has introspective justification to believe that one is in M". So, given that C1, C2, and C3 are the relevant kinds of mental states, the simple view, as now formulated, only entails that John has a justification to believe that he is in C1, that he has a justification to believe that he is in C2, and that he has a justification to believe that he is in C3 -- and maybe that he has a justification to believe that he is in T - where T is his total experience, of which C1-C3 are 'parts'. In general, the simple view only ever entails that agents have a justification to believe simple claims like 'I'm in mental state M'.
(3) So, the simple view, as currently formulated, does not entail, and does not explain, (1) - that John has an introspective justification to believe 'C1 is more like C2 than C3'.
(4) There are some truths about introspective justification that the simple view doesn't explain.
Of course, the problem is general. In psychophysical studies, if a subject consecutive auditory experience A1, A2 of a tone that increases by 10db, the subject has an introspective justification to believe that A2 is twice more intense than A1. But, by the same kind of argument, the simple view, as you formulated it in your paper, does not explain that kind of case.
A response is to offer a more complex formulation of the simple view to get over this problem - a revised version. Something like: for some mental states M1-M2...., you have a justification to believe p iff you have (and *by virtue of having*) M1, M2.....and p is a proposition wholly about M1, M2.... Call this 'the Revised simple view'. It does indeed answer the problem: it entails that John has an introspective justification to believe 'C1 is more like C2 than C3', etc. because that's a prop wholly about his experiences.
In discussion, you said you liked something in the vicinity of the revised simple view. But now a follow up problem arises. There are lots of counterexamples to the revised simple view as I just just formulated it. There are just tons of propositions wholly about your experiences that you might have absolutely no propositional justification for believing.
For example: if your experience represents red, you don't have any justification to believe red=reflectance r, or whatever, but that (or something like is) is (on some views) a true proposition wholly about your experience. similarly, you don't have immediate justification to believe your experience is identical with physical-functional state S. again i wonder how to formulated revised simple view so as not to have these consequences. (of course primitivism would be another response.) Maybe you could add a further proviso: the props can only feature special phenomenal concepts??
Another counter-example: if my experience represents a triangle, I don't thereby have immediate introspective justification to believe that represented lengths are such that the (represented) length of the hypotenuse squared roughly equals the sum of the (represented) lengths of the bases. But that's a truth wholly about my experience (at least on any intentionalist view).
Another issue I'm really interested in is the graded character of immediate introspective justification (something that hasn't to my knowledge been discussed in the literature). It seems to me that your view also entail, and so doesn't explain, truths like
[6] John has more (propositional) justification for believing that experience E1 and E2 are different than he has for believing that E3 and E4 are (where E1 and E2 are totally different whereas E3 and E4 are barely different).
Examples of graded, apparantly immediate introspective justification are everywhere in psycophysics - you just have more introspective justification for believing some things than others. For instance, you have more immediate introspective justification for believing that loudness sensation 2 is more intense than loudness sensation 1 (ordinal judgment), than you have for believing that 2 is "twice more intense" than 1 (ratio judgment). Or again take highly theoretical issues like the fact that some colors you experience are Herring primaries while others are binary. You have some propositional introspective justification for believing that, but much less than you have for simpler stuff, like "I'm in severe pain now".
In the helpful discussion of these issues we've been having by email, you responded by denying " that John has more justification to believe that E1 and E2 are distinct than that E3 and E4 are distinct". You suggested you have equal (and ideal?) propositional justification for both. (Likewise I suppose for highly theoretical issues about e. g. the Herring primaries - you'd say we have just as much propositional justification for this stuff, as we have for "I'm in severe pain" and the like.) This strikes me as counterintuitive, but suggested that what I have called intuitions about graded character "propositional justification" are really intuitions about graded character of doxastic justificaiton. but then i wonder how to spell out or analyze claims about graded character of doxastic justification. a worry for the view is that the only way to do that is to invoke the graded character of propsotional justification (belief B more doxastically justified than belief D iff the basis of B provides a higher degree of prop justification for B than the basis of D does for D). and in that case we are back with the issue of how to explaing the graded character of propositional justification in these cases.
Take care, Adam