We Do Know When We're Awake, But How?

The first premise of the "dream argument", namely that I don’t know that I’m not dreaming now, is tricky. When we dream, we can think that we’re not dreaming and that we're actually awake – but when we are awake and in our normal mode of comportment (i.e. not still in a dream state or under hypnosis) we don’t tend to doubt whether or not we’re dreaming.

The motivation for this premise is that in dreams we can’t tell, so therefore in reality we shouldn’t be able to tell either. If we start from the proposition that there is a difference between dreaming and reality, and we state that in dreams we are sometimes fooled into believing that we’re awake, then it seems to suggest that there is nothing to appeal to in our waking state to yield knowledge that we’re not still dreaming. The illusion of awakeness, presumably, can be generated by the brain.

But it remains a matter of empirical fact that we normally don’t think we’re still dreaming when we’re awake - indeed many of us would be willing to wager that we are actually awake when asked. Why do we have this sense of conviction? And why do we never think we’re still dreaming when we’re awake? I have a feeling there is something ontologically significant about this conviction, but that perhaps it is not graspable in the mode of propositional knowledge. I myself had an experience not too long ago where something quite strange occured in my own daily life. It caused me to ask "am I still dreaming?", which prompted me to become irrefutably aware of my own being awake. When pushed for reasons as to how I knew, I could not give any.

In any case, the point remains – it is only whilst dreaming that you can be fooled into thinking you are awake, so therefore there must be something about the dream state which enables you to be able to be fooled. Conversely, there must be something about the waking state which means that some change in circumstance along the lines of hypnosis would have to be invoked in order to alleviate the certainty of its effect. We will call, for the sake of simplicity, the condition (whatever it is) of certainty that you are awake when you are in fact awake: “X”. X therefore stands for the necessary and sufficient condition of our firm conviction that in our waking state we are awake.

When we are asleep, we lack condition X (as X is such that having it guarantees the truth and conviction that one is awake) and therefore it cannot be through X that we come to be fooled into believing we are awake when we are still in fact asleep. Now, we can come to believe in the truth of all kinds of illusions, but to mistake the entirety of your dream experience for reality is something quite spectacular! Why we become lucid in dreams is a contentious area, and I regret that I won’t be able to think about it too much here. Let it simply suffice to say that when we are in fact awake we never doubt that we are, ceteris paribus, and that we can infer from this that there is some truth-conducive condition in the waking state which resists even the doubts of the philosopher.

Now it might be objected that we’re simply helping ourselves to the conviction of the truth of the waking state, and that condition X is simply a condition which marks out the line between a dream state, and a state of dreaming within a dream. To expand this, we see that the idea rests on the objection that it remains possible that even our waking states are all dreams, and any dreams we do have are simply dreams inside of dreams.

If this objection is found to be significant, then our condition X is useless for guaranteeing awakeness as it (at best) helps us determine which level of dream-state we are in, but does not at all guarantee the truth of awakeness which we’re taking it to guarantee. However there is a looming semantic difficulty in all of this, so let’s make our terms absolutely clear. We will call A the “waking state” - so whether you believe it is a dream or not, let A stand for what we normally call the waking state. D will stand for the type of dream we have as a result of going to sleep following an exhausting A-state day. Lastly, let A* count for our hypothetical “real awakeness” – the kind of awakeness we would have should we awake from our A-state, if the A-state does turn out to be a dream after all.

So if being part of the A-state is necessary and sufficient for condition X, then as a result of this relation to the A-state it allows us to differentiate between being in it and being in the D-state if and only if we are in fact in the A-state. The answer it allows us to give when we’re asking “how do you know if you are not dreaming” is an answer to the question “how do you know that you are not in the D-state?” In which case, condition X allows us to discern between dreaming and being “awake” in so far as dreaming points to the D-state while being awake points to the A-state. However, if what you mean by dreaming is being in the A-state, the kind of state you go into when you fall asleep in the A*-state, then condition X will not be of very much help. However, positing the possibility of an A*-state is not just something we can do without requiring justification!

It’s a fact that our dreams refer to our waking state (so the content of D-states refers to and is in some sense dependent on content apprehended in the A-state) – so why is it that our A-state seems not to have any other referent? If we are inferring the possible existence of an A*-state from our experience of the relation of A-states and D-states, then why should we not require an analogous referential relationship of the sort observed between A-states and D-states in the proposed relationship between A*-states and A-states? Would we not carry in our A-state some residue of our A*-states? Would there not be some inter-referential content which we would causally depend upon experiences in our A*-state, as there is (fundamentally only) inter-referential content from our A-states when we are in a D-state? Does the absence of this referential condition not cast considerable doubt on the reasons for positing the existence of an A*-state?

I suggest that by positing the possibility of an A*-state through knowledge of the relation between A-states and D-states we also posit the necessity of the type of relation the A-state and D-state have. If this relation is not observed then inferring it by analogy will not work as the type of relation we’re proposing the A*-state and A-state to have is not analogous to the A-state and D-state. This is troublesome as it is precisely this relation which gives the notion of an A*-state its plausibility. So if the conditions which maintain between A-states and D-states are not observed between any hypothetical A*-state and the A-state then we’re not justified in believing in the existence of an A*-state by this analogous reasoning.

Being awake is not like dreaming. We experience things we’ve never experienced before in our waking state, we apprehend the content which our dream states must paint a picture with. Cutting right to the chase, we stand in an entirely different relation to beings when we’re awake. In dreaming we might experience some pretty wild stuff – like flying and other things we consider to be impossible. But everything in the dream is significant by virtue of its significance in the waking state – the world we experience when we dream is a manifestation of our implicit understanding of the world whilst being awake.

Consider how even our wildest fantasies cannot escape being a variation, or particular instance, of our fundamental grasp of being. Hubert Dreyfus makes a very similar point about how in science fiction we can only think up variations on our own world, and not something truly alien. Flying, for example, is a normative but not ontological impossibility and therefore does not count as something alien to our understanding. The things which the worlds of sci-fi are populated with are ontologically similar to our own (as in the same sorts of relationships maintain between the entities and roles of things in that fictional world than do in our own) but are nonetheless ontically distinct, as the things which fill in those roles are different. Transports become flying cars and spaceships, habitats become lunar colonies, and so on. The fundamental relationship to beings which maintains in the A-state is, I think, the clue to substantiating the "X" condition and rendering it thematically.