Waking and Dreaming:
Illusion, Reality, and Ontology in Advaita Vedanta
Seth Miller
October 29, 1998
Phil 715: Vedanta Seminar
Prof. A. Chakrabarti

     It is generally taken for granted that our dreams are a wholly different type of experience than our everyday waking experience. In dreams seemingly anything can (and often does) happen; laws of logic are thrown to the wind, physical laws are blatantly ignored, and there is no necessary progression of events either in time or through space. Our common sense tells us that our dreams are illusory (to be sure, in colloquial language these terms are almost interchangeable), and that our waking experience is somehow more real, because of the very fact that dreams lack such things as (waking) logic. But should we trust our common sense on this matter? Beyond the fairly obvious differences that we can draw in terms of explanatory content (i.e. in a dream, blueberry muffins can turn into pink flamingos, whereas in waking experience this never seems to happen), how is it that dream experience differs from waking experience? What is the status of a dream in relation to the status of waking experience?1

      In trying to deal with these sorts of questions, it is helpful to examine a tradition that has dealt explicitly with these very issues, often in ways that are striking or unexpected. Advaita Vedanta, particularly as formulated in the Mandukya Upanishad with both Gaudapada's karikas and Shankara's commentary, takes upon itself the task of exploring the relationship between dreaming and waking, ultimately with the conclusion that they are in all important aspects equal. As this conclusion is highly non-intuitive, the question arises: "How is it possible to conclude that dream experience and waking experience are essentially the same sort of experience?" The working through of this question will give insight not only into the problem of dreaming and waking, but into the central focus of the Advaita Vedanta tradition in general.

The Mandukya Upanishad

      Chapter two of the Mandukya Upanishad, entitled "Illusion", sets about proving the "unreality (illusoriness) of duality through pure reasoning"2 in order to establish that there is only one reality--Atman (literally: self) or Brahman (literally: the all). The duality which Advaita Vedanta focuses on proving the illusoriness of is that between reality and illusion, partly because this duality is the one that is most prevalent, and partly because it is the most difficult duality to get rid of. Adult common sense feels at home with the prospect that there exists a duality between reality and illusion into which the things of the universe can be separated in a more or less systematic manner, such that those things there are just illusions, while it is these things here which are real. Advaita Vedanta, however, recognizes this as logically impossible, for to say that 'these things are illusions' is to say that 'these things are really illusions', which destroys the initial duality between reality and illusion in the first place--we are always only left with reality and not illusion. The question then becomes "what is the reality that we are left with?". The Advaitin answer to this question is precisely that we are not left with our 'everyday' kind of reality (our common-sense reality), but rather with an ultimately transcendent, non-dual Brahman.

In order to make this point, Advaita Vedanta takes an extreme view--that any duality, even the duality between reality and illusion, is itself illusory. It is not that there is reality and illusion, but that there is only reality and no illusion to begin with. The Advaitin therefore takes the strongest possible view, proclaiming that illusions are themselves illusory!3 When we speak of illusions, our common-sense automatically wishes to consider them as real illusions--which is why focusing on the duality between reality and illusion is central, for it is this tendency that the whole of Advaita Vedanta tries to destroy--the automatic and generally unquestioned mode of thinking which reifies all its perceptions in ways which are not consistent with ultimate reality.

      The Advaitins go straight to the heart of the matter by focusing on the apparent duality of dreaming and waking, showing that not only are dreams illusory, but that waking experience is like dream experience in such a way as to be equally so. It is important to understand that when the term 'illusory' or 'unreal' is used, it is used in a particular way. Illusoriness or unreality is predicated of objects in order to indicate that such objects do not have independently existing, objective reality. In normal, waking life, objects we perceive are considered external to us. The chair I sit upon is not me, but is different from me. The Advaitin point is essentially that neither the chair, nor my own limited self is a real, independently existing object. Let us examine the arguments that lead to this conception.

      It is stated that "The wise declare the unreality of all the objects seen in the dream, they all being located within (the body) and on account of their being in a confined space."4 This argument seems almost silly, as it seems obvious that objects perceived in dream are unreal/illusory, because of their very nature as dream objects5. The point being made, however, is both subtle and essential, as indicated above: it is important that the illusoriness of dream objects be established independently of the illusoriness of any other experience. If dream objects were independently real then Advaita Vedanta would never get off the ground to begin with.

      The argument shows that dream objects are illusory on account of their being in a confined space--the physical space within the dreamer (presumably the dreamer's brain?). The objects perceived in a dream, a flock of pink flamingos for example, cannot literally exist WITHIN the dreamer, for the dreamer is smaller than even a few pink flamingos. But clearly if the flamingos have any reality at all it is merely by virtue of the fact that they are dreamt. Whatever reality they have is contingent upon the reality of the dream itself, which arises from within the dreamer, thus the objects in the dream cannot be more real than the dream itself (which is why it is necessary to prove the unreality of dreams for the argument to hold up--whether or not Advaita does this is of course debatable). But since the flamingoes cannot really exist within the dreamer, then their objective reality independent of the dreamer must be false. This is to say that the reality of the flamingos as independent objects is false. Inasmuch as they exist they do so not outside of the dreaming mind, but within it.

      "On account of the shortness of time it is not possible for the dreamer to go out of the body and see (the dream objects). Nor does the dreamer, when he wakes up, find himself in the place (seen in his dream)."6 A person can go to sleep and dream of being in a place hundreds of miles away, but wakes up in the same bed that he went to sleep in. This argument is used against those who might wish to establish that dream objects are real, and that the objects do exist independently of the dreamer. Shankara points out, however, that

"Though a man goes to sleep at night he feels as though he were seeing objects in the day-time and meeting many persons. (If that were real) he ought to have been met by those persons (whom he himself met during the dream). But this does not happen; for if it did, they would have said, 'We met you there today'. But this does not happen. Therefore one does not (really) go to another region in dream."7

Thus the dream experiences, though appearing to be real to the dreamer, are not actually so.8

      In order to establish more firmly why it is that dream objects are illusory, a unique explanation is given, which is absolutely central to the entire Advaitin project:

"Different objects cognized in dream (are illusory) on account of their being perceived to exist. For the same reason, the objects seen in the waking state are illusory. The nature of objects is the same in the waking state and dream. The only difference is the limitation of space (associated with dream objects)."9

This is a radical statement. It is the very fact of being perceived that accounts for the unreality of any object, whether waking or dreaming. But why does this make the objects perceived illusory?

      The best reason seems to be that the objects in dream are illusory not just because such objects are perceived, but because all the dream objects are is perception of the object itself. Or perhaps more clearly: the perception of the object in the dream constitutes the object itself--there is no reality to the dream object apart from the perception of it. Thus there can be attributed to the dream object no reality independent of the perception of the dreamer. From within the dream, dream-objects merely seem to be objects in the normal waking sense, and are experienced to be real. From the waking state, however, dream objects are illusory, because they merely exist as the perception of them (i.e. not as an experience of something independent).

      The Advaitin wishes to show that waking objects are of the same nature as dreaming objects on account of the fact that they are both perceived. The argument is that because dream objects are illusory due to their perceivedness, then waking objects must also be illusory because they too are perceived. And in fact the Advaitin maintains that ultimately waking objects, like dream objects, are constituted by perception and have no independent reality apart from perception. It is essential to the main Advaitin argument that waking objects are perceived in a way that is similar enough to dream objects to establish that the illusoriness of dream objects (due to being perceived) holds for waking objects in the same manner.

      This correlation is stated thusly:

"In dream, also, what is imagined within by the mind is illusory and what is cognized outside (by the mind) appears to be real. But (in truth) both these are known to be unreal. Similarly, in the waking state, also, what is imagined within by the mind is illusory; and what is experienced outside (by the mind) appears to be real. But in fact, both should be rationally held to be unreal."10

The Advaitin argument rests upon the assumption that it really is the case that waking experience is analogous to dream experience in such a way that the 'perceivedness' argument is valid for objects in both states. But is this the case? Is it possible that the perception of dream objects is of a different order than the perception of waking objects?

      It is important to note that what the Advaitin is focusing on is the fact of perception itself as an indicator of illusoriness. This is so because in order to perceive there must exit a perceiver and a perceived. The Advaitin cannot admit the reality of such a distinction precisely because it is a duality, a clear separation and thus improper reification of two things, a subject and an object, where in reality, only the one subject exists (Brahman). As soon as there is perception of something, an illusory duality is implicitly established when in fact reality is only one--the subject without an object, the perceiver without a perceived. Thus no matter what kind of perception one is having, it is always 'off the ontological mark' so to speak, and is therefore unreal, illusory.

      But then a different sort of argument is given that is independent of the 'perception argument':

"That which is non-existent at the beginning and in the end is necessarily so (non-existent) in the middle. The objects are like the illusions we see, still they are regarded as if real."11

It is a tenet of Advaita Vedanta that for something to be ultimately real it must not partake of either creation or destruction--it must not change. No transition from existing to non-existing can be made for an object that is real, for to be real is to exist, in the complete sense of the term. If at any time an object either comes into or goes out of being, it is necessarily unreal, for to be real means to exist, and to exist means to exist as this particular thing and not as anything else; if the object changes, then it is no longer the particular thing it was, but is something different, and hence the object has made a transition from being to not-being, from being this, to being that. Hence the thing is necessarily unreal, for if it partakes of change then it is coming into and going out of being, which is the same as being unreal at all points. The point is subtle and perhaps can be approached by saying that in order for an illusion to exist there must be a 'substratum' for the illusion (as the rope is the substratum for the illusion of the snake in the classic example). The point is that the substratum of the illusion must remain the same for the illusion to exist. Burn the rope and the snake cannot be seen to exist, for without a substratum the illusion cannot take place12. But why is it that an object cannot exist for a little while and still be real during that time?

      The idea is this: that in order for something to change, there must be something else that does not change. For example, in order for me to change, there must be something about 'me' that does not change, that is constant throughout the change--otherwise the statement 'I have changed' is meaningless, because there would be no continuance between 'I now' and 'I previously'. Similarly, in order for anything to change, there must be something about the thing undergoing the change that which is constant--there must be something unchanging and real/actual/existing in order for change to exist, or absolutely no sense can be made of any distinction whatsoever. The Advaita Vedanta point is precisely that no matter where you being this analysis, you must always end up at the same place--Brahman, which is a convenient name for the very thing that does not change itself but that underlies all apparent change. It can be stated in many ways: "the only thing that doesn't change is change itself" is one kind of quaint way of getting at this difficult thought.

      Thus, if anything has a beginning and an end, it cannot be ultimately real, but must be illusory, for the only thing that can be ultimately real is the thing that stands outside of the realm of change. As both dreams and waking experience are seen to change, both are therefore illusory. This concept is perhaps the most difficult to grasp. The person that wishes to say that dreams or illusions must be real on some level is in a sense correct, and is shown that what is real about the illusion is the unchanging Brahman. It is important to see that this is not an essentialist doctrine that finds in each thing a different essence, but rather one that finds in everything only Brahman. This is not a reductive argument, either, at least in the normal sense, for if anything is reduced it is reduced to "the all"!

The Nature of Dreams

      Is this Advaita Vedanta account of the dreaming and waking states adequate? Are there arguments against the Advaita Vedanta conclusion that dreaming and waking are identical except for the limitation of space associated with the dream state? If there is a great enough difference between the way in which waking objects are perceived and the way in which dream objects are perceived, then the Advaitin conclusion that reality is illusory because of its perceived nature will be shown false.

      Anyone wishing to make a hard distinction between dreaming and waking is forced to deal with the Advaitin point that it is only from the waking state that dreams are seen to be illusory, and that there is no way to tell from within the dream itself that we are dreaming. If at any time our experience could be 'just a dream', then there is no firm basis upon which we may say that there is a distinction between the two states (because what we think of as reality can always collapse back into a dream state). Thus, in order to show that there is a difference between the waking and dreaming states, it is imperative that we be able to show that we are awake, for it is only from the awake state that our knowledge (of the difference between waking and dreaming) is possible. Let us examine this problem more closely.

      In order to be convinced that we are not presently dreaming, some form of criterion must be used to distinguish between dream experience and waking experience. At this point, all we have is our experience, and any criterion that we can use is necessarily a part of our experience. Some possible type of criterion are: intuitive feelings, logical or rational arguments, inferences, deductions, criterion based on laws of physics, psychology, or metaphysics, coherence, complexity, and correspondence with other's experience.

      But as soon as we choose any of the above criterion and apply it to our present (presumably waking) experience, we see that we could actually be dreaming that we are using such criterion to prove that we are awake, and in fact we could be dreaming that the criterion is satisfied! Any validation I wish to carry out for the purpose of determining if I am awake is doomed to failure from the start, because I can merely dream that such validation takes place successfully. Is the distinction between waking and dream experience really dissolved so easily? We shall see.

      I wish to argue that in making this sort of 'receding frame' argument, a crucial distinction has been glossed over: the fact that what applies to the content of a dream does not necessarily apply to the actual ontological situation of the dream itself, and that we must maintain a distinction between what the dream actually is and what is happening in the dream. Assuming that there is a difference between dreaming and waking, there are two possibilities: we can apply the criterion from within the dream, or we can apply it while we are awake. When we apply a criterion to determine whether or not we are awake from within a dream, then we are not actually applying the criterion, but are merely dreaming that we are applying the criterion--the application of the criterion is taking place only on the level of the dream content, and not to the ontology of the dream-nothing is actually being tested, and since it is the ontological status of our experience as either waking or dreaming that is at stake, if the experiment fails to apply on this level then it is no experiment at all.

      Just because the content of the dream can include a successfully completed 'reality-testing' experiment, we are not therefore banned from ever knowing if we are awake or not. In order for it to be possible that we can determine if we are awake with any certainty, four things are required: first, an actual (ontological) distinction between waking and dream states must exist. Second, there must exist some criterion that actually tests for such a distinction. Third, not only must the criterion actually be testable, but we must be able to know the results of such testing.13 And lastly, reality must not be 'dream-like'. This is to say that reality must follow logical laws, physical laws, and laws of cause and effect--reality must be 'reality-like', so that a true testing situation is possible.14

      The receding frame argument fails because it does not satisfy count #2. When the reality test is carried out in the dream, it is not actually testing what it claims to be testing (the ontological status of the experience) and is actually testing nothing, as it is merely a dream experience which is on par with any other dream experience in terms of the relation between the content of the dream and the ontology of the dream. This does not mean that dreams cannot be an effect of physical states, just that from within the dream we do not have access to the true ontology of the situation by virtue of the fact that the dream content is different than the ontology of the dream. This is not to deny that the dream content itself is ontologically real-just that because of the peculiar ontological situation of the dream content we cannot know from inside the dream if we are awake with surety.

      In a dream, where there are no laws that need to be followed, an experiment can give any result, and either be convincing or not, all with no effect on the true ontology of the situation. Thus when we are dreaming, we can never accurately determine if we are dreaming or if we are awake, because we could just be dreaming that we are satisfying whatever criterion we choose to use to determine that we are awake.

      However, if there is a difference between waking and dreaming, then when we apply the criterion while we are awake, our results will be worthwhile, for they will tell us something about our state. The objection can be made that in order to know whether or not the results of our experiments are valid we would need to first know if we are awake or not--but this is the very thing we wish to test! This objection is quite valid, but in the end loses its power from a realist stance, due to the fact that when we apply the criterion in waking life, reality is such that it will give consistent answers regardless of our perception of those answers. We can always convince ourselves that we are in one state or the other, irrespective of any information we may be privy to, or it is possible that our test may not be scientifically valid and thus gives bogus results, but that does not change the fact of the matter, the ontology of the situation.

      What is interesting is that from a waking standpoint we have explanations of dreams, while from within the dream we have no explanation of waking reality. In particular, we know that dreaming takes place in what as known as R.E.M. sleep, in which beta-waves are dominant in the brain. When we are awake, our brains are also mostly in beta.

      This in fact serves to explain quite a bit. The similarity of the perception of waking objects and dream objects is directly correlated to the brain states associated with the perception, and the fact that in both cases the brain activity is almost identical. When an object is perceived, whether in a dream or in the waking state, the brain functions primarily in beta. The same areas of the brain show activity in a dream as in waking life. What then is the difference? Precisely that when we are dreaming, the stimulus that gives rise to perceptions of objects is generated almost entirely from within the brain itself, which means that the "input" (so to speak) is not mediated by the senses. This accounts for the ability of dreams to ignore laws of causality, rationality, and physics. In our imaginations (where the stimulus for perception is also internal) we are able to circumvent such waking laws, and in our dreams the case is similar. Objects perceived either in dream or imagination are illusory on account of the fact that they exist only for the perceiver, as they arise in the mind of the perceiver and not outside of it. By contrast, objects perceived in waking experience are not illusory on account of the fact that the stimulus for their perception arises externally to the perceiver. There seems to be a strange and unavoidable property of reality: it unwaveringly presents itself to us when we take the time to look, in a way that is mediated through our senses, to be sure, but still in a consistent way, full of details that are discovered rather than invented. The fact that we cannot choose to consciously perceive red while looking at green grass, in conjunction with the fact that we can explain exactly why this is so, lends great credence to the idea of a reality that exists independently of our perception of it.

      But does this mean we can know we are awake? Isn't it still possible in any given circumstance that we are actually dreaming, convincing ourselves that we are awake? It must be granted that this is always a possibility--but just a possibility, and a slim one at that. Perhaps the best test to see if we are awake comes when we do not wake up. It happens to be a fact that dreams always end, and not only are we returned to waking experience, but we are returned to waking experience that is commensurable with the laws of contradiction, physics, etc. and which in addition is always the same reality that we left when we first went to sleep. No matter how detailed or long a dream may be subjectively, we always wake up, and upon waking realize that we were only dreaming. Even if we cannot prove that we are awake now, all we need to do is wait. The longer we do not wake up, the more likely it is that we are not dreaming. Combine this with other tests of reality over a period of time and it becomes more and more likely that we are able to know if we are awake, as the probability that we would dream each and every time, out of all the infinite possibilities, that we are awake (when we are in fact dreaming) is extremely improbable.

      In the end the Advaitin's best, and (ironically) worst argument for the illusory nature of both waking and dream experience is the argument about the illusory nature of change. This is where Advaita Vedanta leaves science behind and leaps into speculative metaphysics. The argument is powerful because of its simplicity, and goes straight to the heart of the matter in a very direct and fierce way by denying the reality of change altogether. At the same time, in order to remain consistent, the Advaitin must claim that not only is change illusory, but that change does not exist in the first place. This claim is the most difficult to sustain of perhaps any claim, as it is impossible to deny that something is happening--that experience exists, and therefore change. It is not enough for the Advaitin to say that such perceptions are illusory, for this would be to fall into15 the trap of reifying the illusionary status of perception, and open the doors to a circular argument ad infinitum as seen previously.

      The arguments in the Advaita Vedanta tradition serve to show us that in many respects, dreaming is just like waking. It shows how our perception plays a large role in our experience of objects, and that what we might be tempted to call 'external' may in fact be 'internal'. It seems that there is never a one hundred percent accurate way to know at any given time if we are awake, but this says more about our ability to convince ourselves that something false is true than it does about the actual ontology of our experience.


1 It is a testament to the rootedness of our common-sense view of dreams as somehow a special case of experience defined in reference to an assumed background of waking experience that we have a name for a dream experience (a dream) while there is no corresponding term for a particular waking experience--one can be asleep and have a dream, or one can merely be awake. It is not necessary to speak of being both awake and as having 'awake' experiences at the same time. Back
2 The Mandukya Upanishad with Gaudapada's Karika and Sankara's Commentary. Translated by Swami Nikhilananda. Vedanta Press. 1995. p86 (Hereafter referred to as MU) Back
3 Due to the reifying nature of language, and self-referential nature of this kind of statement, an infinite regress is set up (just like in the problem of the receding frame which I will discuss later): the duality between reality and illusion is itself illusory, but by saying this, it is assumed that the duality is unreal in relation to what is ultimately real--non-duality. But then of course one has implicitly set up another duality (one level removed), which is itself also illusory, by the precept of the statement itself: the duality of reality and illusion is illusory, and so then the illusory nature of the duality of reality and illusion is also illusory, which is again illusory, and so on ad infinitum. This serves to further the Advaitin point, which can circumvent this sort of circularity by stating that one cannot in the first place distinguish between reality and illusion. In addition to this, the infinite progression of identifying the fact that 'this too, is illusory', ultimately requires something that is not illusory in order for any of the steps to be valid--i.e. reality. The infinite series can be avoided by stating that there is only reality, and therefore there is not even the possibility for illusion--only reality is. Back
4 MU p.86 Back
5 Notice that this type of sentiment, like many other common-sense sentiments, are seen to be tautologies upon close analysis. Back
6 MU p.88 Back
7 MU p.88-89 Back
8 There is an interesting flaw in this argument, to the effect that the dreamer's dream is merely another reality, which the dreamer really visits. The Advaitin would merely reply that there cannot exist a duality of equally real realities; there can logically only be one reality, Brahman. Back
9 MU p.90 Back
10 MU p.97 Back
11 MU p.92 Back
12 MU p.107, 124 Back
13 This criterion is somewhat strange, and normally would be included with precept #2, but in this case it must be stated separately. This is so because we are making a division between the ontological situation and the epistemological situation of the dream. It would be a Kantian-type move to say that we can never know if we are awake or not, but that this has no bearing on the fact of whether we are awake. Precept #4 assures us that we will not fall into this sort of trap, and will in the end be able to have knowledge of whether we are awake or dreaming. Back
14 If it was the case that reality was 'dream-like,' then the question "Am I dreaming or am I awake?" would never arise in the first place, because there would be no conflicting types of experience that needed to be reconciled. That there are different types of experience (reality-like and dream-like) that need to be examined is either a testament to the mass confusion of people in general (who, if there were not a difference between waking and dreaming experience, would be engaging in quite a delusion, arbitrarily making distinctions between experiences that were of the same type), or to the fact that philosophers have too much free time on their hands to create entertaining but misleading logical loopholes surrounding the issue at hand. Back

Back to Seth's Homepage.