Saturday, May 21, 2005

Sensitivity, Dreams, and the Cogito

We want our beliefs to be true, but in a reliable sort of way, rather than by some fluke or chance. This prompts the sorts of counterfactual requirements discussed here. We want our belief to be 'sensitive' to the truth, so that if P had been false, we would not have believed that P. But this requirement is too strong, as skeptical scenarios make clear. Instead, we should be satisfied with a belief that is 'safe', i.e. one we would not easily be mistaken about. It is enough that our beliefs track the truth through close possible worlds - they need not be secure all the way out to the first 'not-P' world, if it is very distant (as is the BIV scenario).

Now, Jonathan at Fake Barn Country argues that the skeptic's dream scenario threatens the safety of our beliefs. There is a close possible world in which I am dreaming all of this, such that many of my present beliefs would be false. So even though my beliefs are all true (let us suppose), they fail to count as knowledge because I easily could have been dreaming and thus mistaken.

As I responded in comments:
the phenomenology of dreaming is quite different from that of waking life. Even though I might not realise when I'm dreaming, I certainly do realise when I'm awake. This asymmetry is important. There is no chance at all that I'm dreaming right now, because I know that I never have this quality of phenomenological experience while asleep. So I don't see dreams as skeptically dangerous.

My understanding of dreaming is influenced by Dennett's metaphor of anosognosia: sometimes when it seems we are aware of something (e.g. filling in the blind spot), it's really just that we are not aware of our deficit. I think dreams are like this: it's not that we positively think we're awake, it's more just that we aren't aware that we are asleep. Most dreams aren't convincing hallucinations, it's just that our judgment is so impaired at the time that we don't realise how different and unconvincing they are. The potential for confusion is negative, not positive; it's due to a lack, not a presence.

What I want to emphasise here is the crucial difference between my waking knowledge that I'm awake (which is in no danger of being mistaken), and my dreaming inability to tell whether I'm awake. The latter in no way threatens the former. From the fact that if I were dreaming right now I would not be able to tell the difference, it does not follow that I actually cannot tell the difference! That's the "asymmetry" idea I was trying to get at before. (I worry that I'm not expressing this very clearly, but I hope you at least get the general idea.)

As such, I consider my beliefs quite unthreatened by dream-based skepticism. If we include the appropriate background, i.e. the rich 'positive' phenomenology that grounds my beliefs, then I think they even remain 'safe'. Given my phenomenology, I could not easily be mistaken. I could have a different phenomenology and fail to realise it, e.g. if dreaming, but that is of no relevance here. It's no problem, because as it happens I'm not just overlooking a negative, I'm actually noticing a positive. I am fully aware of the richness of my present (waking) experiences. If dreaming, I would not have that awareness. Instead, I would be unaware of my deficit. Two negatives do not make a positive. We should conclude that my beliefs are safe after all.

Alternatively, if you insist that the dream scenario implies that my beliefs are technically not 'safe', then that simply shows that safety is not what matters. All that justification really requires is the sort of "asymmetrical safety" that I have described above. 'Symmetrical safety' is overly stringent in much the same way as 'sensitivity' is. So we should not be too worried if our beliefs fail to meet these tall requirements in light of various skeptical scenarios.

To demonstrate: suppose that you've been slipped a drug that screws up your logical capabilities. So you engage in wacky trains of illogical reasoning, all the while unaware of your irrationality. At some point, you reason to yourself, "I think, therefore I am", and it strikes you as utterly certain - you think you couldn't possibly be wrong. But suppose it turns out you are wrong. We can't comprehend this of course, because we're all on the wacky drug ourselves, so our logical capacities are screwed. But that, I hereby stipulate, is really how it is (in the scenario we're imagining).

Does that scenario in any way threaten our present certainty in Descartes' cogito? I should think not. This then helps reinforce my central point: The fact that if we were wrong we wouldn't know it, does not establish that we currently do not know that we're right. To think otherwise is to neglect the asymmetries I've pointed to, and Dennett's insight that neglecting absence does not create presence - it merely seems like it.


  1. This is a good argument, Richard; and quite right, I think. Dreaming may avoid the problem of remote possible worlds, but dreams seem a weak basis for a global rather than a local skepticism, for precisely the reason you note: the problem it sets up is due entirely to a lapse in the ability to judge.

    It's difficult to put the matter clearly. Would this work? I'm currently paying close, careful, critical attention to what is going on (let's say); but there is a close possible world in which I am being lazy (thus giving rise to false beliefs). We can't really build a global skepticism out of this possibility; the fact that there's a close possible world where I'm careless, uncritical, and unattentive doesn't give a reason to be generally skeptical about my mental abilities, precisely because in such cases my mental abilities are in a state of lapse. But dreaming is (effectively) just such a case, just like the anti-logic drug.

    The anti-logic drug, by the way, reminds me of William James's experiments with laughing gas. While on it, he had the intense conviction of having discovered the secret of the universe, the ultimate discovery that would revolutionize the way we look at the world. So he wrote it down. When he came out of the state induced by the gas, he found that all he had written down were a few puns.

  2. Yes, you've pinned down the idea I was trying to get at. Thanks for that (and the great William James anecdote!).

  3. I just now managed to find an article online in which James briefly analyzes the experience:

    Subjective Effects of Nitrous Oxide

    Interestingly, it's also explicitly an essay on Hegelianism!

  4. Interestingly - for the most part - I DO know I am dreaming when I am asleep 9the ones that i remember at least), although this may take a moment to realise. (this alows for interesting experiments)

    Still while I "know I am awake" now all I know is that I am more awake than in some dreams. But that in no way implies there is not a higher level. For example I could have a dream about trying to wake up - and then I do wake up but actually it is still a dream - it may also be clear that that "waking up" has a higher level of reality than the dream. for example it may obey the basic laws of physics. there is somthing in this new state not in the old one (and it could be a fundimental difference if my brain had moved into a slightly more awake stage of sleep lets say) and yet it is still "asleep".

    BTW good to see some William James !

  5. Good post, Richard -- I agree with quite a lot of it. My post was about why the dream scenario, traditionally construed, was especially important. You're arguing that the traditional construal is mistaken, which is cool. I think something similar.

  6. I just can't resist adding this nonsense.
    Suppose that as well as lesser states, there was a waking state that was "higher" and more inclusive than the waking state you are sure of now. The 'you' that you had become, would know this without doubt ... would 'you' not?
    (And mr Descartes' saying might appear somewhat differnt)

    But how would you inform another, less awake, mind what it was like? And indeed, how would your normal mind then interprate the information that was briefly perhaps glimpsed.

    Might it not sound like nonsense? The middle state of mind and the normal expression channels simply could not formulate the 'higher ' information.

    Sadly, it would be 'unprovable' that a higher state had occured unless. .unless ...

  7. Hmm .. as genius observes, it is possible to know, in a dreamlike way, that 'I am dreaming.'

    I would suggest that similar knowledge could arise in a normal waking state mind which had experienced over a period what I called a "higher" state. ( Once it recurred sufficiently for the logical mind to begin to incorporate the information)

    And that intriguing quote from William James, though it is a result of artificail modification of the brain, would appear more than just nonsense.
    Oh well, nice to express one's bias. cheers .

  8. Jonathan outlines a similar argument from Ernest Sosa, here. Sosa illustrates the asymmetry idea through an analogy with death:

    "Suppose I could now about as easily be dead, having barely escaped a potentially fatal accident. Obviously, we cannot distinguish being alive from being dead by believing oneself alive when alive, and dead when dead. But that is no obstacle to our knowing ourselves alive when alive."


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)