Go back to the Descartes page for more texts and other resources.

Evaluating Descartes' Dreaming Hypothesis

A critical evaluation of a slightly modified version of Descartes' argument.

Descartes? dreaming hypothesis appears, at first glance, to be a very compelling argument. The argument in question, which is a modification of Descartes? hypothesis, initially appears to be almost irrefutable. However, I will argue that although it is deductively valid, the argument is not cogent. The first premiss is not acceptable, and equivocation makes the conclusion of the argument ambiguous in its meaning and deceptive in its implications.

The claim that the argument is deductively valid requires some justification. It can be argued that, assuming the premisses are true, there are circumstances in which the conclusion would be false. For example, it may be suggested that the statement ?I know that I am thinking? could be true regardless of whether I am awake or asleep. However, relies on the assumption that thought in dreams is essentially the same as thought in everyday life. I will argue later that this is not true, and that it is impossible for someone who is dreaming to be conscious of the fact that she is thinking. Other claims like ?I know that I exist?, are similarly problematic. Given that there are no other imaginable circumstances in which the conclusion would not follow from the premisses, the main argument must be valid.

I have established that the move from ?I do not know that I am now awake? (hereafter, (b)) and ?There is nothing I know with greater certainty than that I am now awake? (hereafter, (c)) to the conclusion, ?I do not know any statement to be true? (hereafter, (d)) is deductively valid. It is also necessary to establish that the move from the first premiss (hereafter, (a)) to the intermediate conclusion (b) is valid. It clearly is. If we cannot exclude the possibility that I am dreaming, it follows by virtue of the definition of dreaming that I cannot be certain that I am now awake. Hence, the entire argument is valid.

Consequently, any attack on the cogency of the argument must be on the grounds that either of the premisses (a) or (c) is unacceptable. Premiss (a) attempts to establish the fact that we can?t prove that we are not dreaming. The skeptic here does not have a considerable burden of proof - it is not necessary for her to establish that we are dreaming, but merely that we can offer no proof to the contrary. Even so, this premiss is not acceptable. We may object to this premiss by simply arguing ?I can tell that I?m not dreaming?, but this argument begs the question ? we may be dreaming that we?re not dreaming. In order to justify our objection to premiss (a), we need to establish that there is a way for us to distinguish dreaming from waking.

It is true that in many dreams we may have the sensation that we are awake. However, this doesn?t mean we can rule out the possibility that when we are awake we can tell with absolute certainty that we are not dreaming. An analogy will help illustrate how this can be the case. I am awake, absorbed in reading my favourite novel. If I were to connect my brain to a device which filtered all thoughts about the real world, so that I would no longer experience the sensations of hunger, of sitting on a couch or of noise in the background, I would effectively feel as though I was part of the world in the novel, and I would be happy to say ?I?m in the novel?. However, without the filtering device, I am certain that I am only reading, and that I am not in the novel. A dream is much the same, although it is much more absorbing (like a good novel, really). If my sleep wasn?t filtering my thoughts about reality, it would be possible for me to assert that I?m only dreaming. This illustrates the fact that what I believe in my dream (?I?m in the dream?) is not the same as what I know to be true outside the dream (?I?m only dreaming?). Although I may be deceived when I?m dreaming, when I?m not dreaming I know with certainty that I am awake. This is because I can engage in certain thought processes and perform certain mental activities which I do only when I am awake, and which are not possible in the ?filtered? realities of dreams and novels. I can question, deduce, analyse and hypothesise; and I can use these skills and my self-awareness to establish that I am awake. If I am awake, I am not dreaming; hence, I can offer proof that I am not dreaming, and challenge the acceptability of (a).

The dreaming hypothesis rests on the concept of knowledge, so it is important to determine whether the word ?know? has a consistent meaning in the argument. In the intermediate conclusion (b), ?know? is used to imply certainty, suggesting that, as a result of (a), I cannot be certain that I am now awake. In (c), however, the word is used in a more familiar way, in line with our usual understanding of knowledge: to ?be aware through observation, inquiry or information?.[1] Statement (c) appeals to our beliefs and our awareness; statement (c) is concerned instead with what we are justified in believing. It is unclear which of these definitions of knowledge is used in the conclusion of the argument, (d). If it is the first, then the argument contends that I am aware and I believe that no statement is true. If it is the second, then the argument concludes that it is a justified fact that no statement is true. There is considerable difference between these alternatives. Either the skeptic has argued that it is true that I cannot know any statement to be true, or she has argued that my understanding of the world is such that I acknowledge that I cannot know any statement to be true. If it is the later, then the conclusion seems a contradiction with (c), and if it is the former, then we wonder how the skeptic could have arrived at that conclusion by using in (c) an entirely incompatible concept of knowledge. In either case, the fallacy of equivocation is involved, and the conclusion is ambiguous.

In ?Four Forms of Scepticism?, G.E. Moore argued that any premiss of a skeptical hypothesis is less certain than the belief that the skeptic is challenging.[2] In relation to our dreaming argument, Moore would contend that we should sooner accept the fact that we?re not dreaming than accept any of the skeptic?s premisses. Moore based his argument on the idea that our bodies exist in a definite external world which requires no justification. This relates closely to our discussion of knowledge. We all believe that we are not dreaming, because we see evidence ourselves in the external world. Whether or not that belief is justified is another matter entirely; a matter which does not affect our everyday sensibilities.

Accordingly, the dreaming argument can never demonstrate that there is doubt about what we actively believe ? that is, knowledge as it is used in premiss (c). It is thus considerably less useful than we may have originally thought, and suffers from a lack of clarity which makes it difficult to determine the implications of its conclusion. Furthermore, there are several grounds on which the acceptability of the first premiss can be challenged. For these reasons, the dreaming hypothesis is not a cogent argument.

[!1] ?Knowledge? in J. Pearsall, ed., The Concise Oxford Dictionary, 10th edition, 2001
[!2] G.E. Moore, Selected Writings, Routledge, London, 1993.


R. Descartes, Meditations, in The Philosophical Writings of Descartes, tr. J Cottingham et al, Cambridge, Cambridge University Press, 1985.
T. Honderich, ed., The Oxford Compaion to Philosophy, Oxford, Oxford University Press, 1995
J. Pearsall, ed., The Concise Oxford Dictionary, 10th edition, Oxford, Oxford University Press, 2001
B. Stroud, The Significance of Philosophical Scepticism, Oxford, Oxford University Press, 1989

Authors | Quotes | Digests | Submit | Interact | Store

Copyright © Classics Network. Contact Us