Here's an essay I just wrote; don't know if anyone cares.
Prompt: Is there a perceptual experience you could have while you are awake that you could not dream you are having? If not, what does that imply about the possibility of understanding perceptual knowledge of the world? If so, could you know on the basis of that experience alone that you are not dreaming at the moment?
Because perceptual experiences within dreams resemble perceptual experiences while awake, it’s important for a person to know what kind of perceptual experience they are having before they can justly come to conclusions about reality (assuming reality is the world independent of perception). While dreaming, there is no guarantee that the perceptual experiences a person is having have anything to do with objects in the world; if someone believes that they are currently dreaming, they cannot justify knowledge of the world based on their current (dreaming) perceptual experience. In fact, unless a person is sure that they are not having a dreaming-perceptual-experience, they cannot rely on perceptual experience as an indicator of external reality, because the experience may just represent a dream and not resemble reality at all.
We’ll assume there are two kinds of perceptual experiences (PE): dream-perceptual-experiences – perceptual experiences that happen while dreaming (DPE) (we could think of these as unreliable perceptual experiences) and awake-perceptual-experiences – perceptual experiences that happen while awake (APE) (we could think of these as reliable perceptual experiences). In order for a person to know they are having an APE, they must have a way of unambiguously distinguishing between DPEs and APEs. And since PEs are all possible experiences (either DPEs or APEs), one could conclude that a PE that is not a DPE is an APE and vice versa. So knowledge that an experience is an APE requires knowledge about the set of possible DPEs. And this knowledge (about the set) is necessary because it’s not enough for a person to guess that they are having an APE and happen to be correct – they still couldn’t justify conclusions about reality without being sure the experience is an APE, because otherwise, they couldn’t be sure the experience wasn’t unreliable.
How should a perceptual experience be defined? One way would be to call each possible state of the brain a PE, with the set of possible PEs being equal to the set of possible brain states. In this case, there are no possible PEs that one could have while awake that one could also have while asleep (assuming the brain is in a different state during sleep). But this definition doesn’t help distinguish whether an individual PE is an APE or DPE, because although the brain states are different, the subjective “feel” of different brain states may be equivalent. So even if one was having an APE (and the corresponding brain state), they couldn’t justify it being an APE, as it’s possible that the PE that they’re having is equally consistent with the “feel” of an APE and a DPE, meaning there’s nothing to distinguish them by. Defining a PE this way, then, doesn’t help anyone justify their conclusions about reality.
We could perhaps define a perceptual experience a different way – a PE is a member of the set of possible things one can feel like. In other words, there are many possible ways one could feel (the set of possible PEs), and a PE is any specific one of those ways of feeling. Using this definition, it’s possible that many different brain states (possibly while awake or asleep) could count as the same PE (as long as the different brain states make one “feel” the same way). But how do we know that there isn’t a possible DPE consistent with every PE? If all possible PEs could be DPEs, then conclusions about reality could always be questioned since the basis of those conclusions could have come from dream experience.
How, then, could a person know whether a single PE is a DPE or not? (If a PE is not a DPE, it is by definition an APE, as the set of possible PEs is the disjunction of DPEs and APEs.) That person would have to know about the set of possible DPEs, and check whether each PE they have is in that set. If it’s not in that set, the PE must be an APE, and the person could conclude that they’re not dreaming. Just knowing the set of possible DPEs isn’t enough to conclude that some perceptual experience is a reliable indication of reality, though, because until a PE comes along that isn’t in the set of possible DPEs, any perceptual experience could still be a dream. So one would have to know both the members of the set of possible DPEs and that there exists a PE that is not a DPE (and therefore must be an APE). Then they could conclude that there is at least one perceptual experience that could not exist in a dream.
Distinguishing between DPEs and APEs could also be done with knowledge of the set of APEs – if a PE isn’t a member of the set of possible APEs, it must be a DPE. But that doesn’t help anyone know which perceptual experiences reflect reality; it only allows a person to know if a perceptual experience doesn’t reflect reality. From the perspective of justifying that a PE is an APE, knowledge of the set of APEs is irrelevant – just knowing a PE is in the set of possible APEs doesn’t rule out that PE being a DPE. The only knowledge relevant to justifying conclusions about reality has to do with the set of possible DPEs, because if a PE isn’t in that set, it must be an APE.
So it seems simple. In order to know that one is not dreaming, that person must know the set of possible dream experiences and have an experience not in that set. That specific experience, then, can be thought of as accurately portraying reality. Practically, though, we run into a big problem with this view – how would any person get knowledge about the set of possible DPEs?
In order to get knowledge about this set, a person would look to experiential evidence. Maybe they’ve studied the brain, maybe a friend taught them a bit about neuropsychology, etc. But how could they conclude that the information they have about the set of possible DPEs actually came from reality as it is? How could they know they didn’t dream up the information? In order to rely on the information a person has about the set of possible dream experiences, they need to know that the information came from an APE, which they couldn’t know without already knowing the set of possible dream experiences. So they can’t justify knowledge about the set of possible DPEs without assuming the exact knowledge they are trying to justify. In other words, no one can justify that they know if they are dreaming without already assuming that they know if they’re dreaming.
And that’s just what we do. We assume that our information about reality is reliable even though we cannot know that it is. If we perfectly understood reality, we would never be wrong about it (which everyone knows is not the case). But we do have mechanisms at play – when our beliefs about reality turn out to be wrong, those beliefs are selected against (via reinforcement learning) for use in the future. Beliefs about reality that hold true tend to outperform and eventually replace beliefs that aren’t true. And that’s the story of humans – we act on our beliefs even though we can’t justify them, and spend our lives trying to better match those beliefs with reality – a task that we cannot even judge the success of.
(Note: I was thinking of dream-perceptual-experiences as all experiences that are unreliable, meaning that awake-perceptual-experiences are by definition reliable, as perceptual experience is the disjunction of DPEs and APEs. The logic of dream experiences applies to hallucinations and other perceptions of non-reality as well.)