Experiential definitions of some emotions

Definitions of emotions are usually so difficult that it seems often we don't even try, or just assume that it's subjective or just impossible to define. Still I've made some attempts here. I don't necessarily think these definitions are accurate, but they seem like a good first step towards an accurate understanding of these ideas.

Emotions are a part of being conscious, but it's not immediately clear how they relate to the rest of our conscious experience. However, I think if we look at the fundamentals of what kind of information is available within conscious experience we can build up a definition of emotions from first principles. So, what are the kinds of information available to us when we're conscious, I think we can really narrow it down to just three things:

• Sensory information - the world interacts with our senses and those sensory organs transmit information about the world to our consciousness. This information can be experienced immediately, or it can be remembered (for now I'm going to assume that memory is part of consciousness). I'd also include things like instinctive or automatic or unlearned responses here, in a sense we're a spectator for what's happening when these responses are triggered in our bodies, and we get information about them mainly, maybe entirely, through observation.
• Pain and pleasure - unlike all other information about the world, pain and pleasure don't represent anything that exists outside our consciousness. These sensations form an additional layer of information on top of what we experience and are a key component of consciousness.
• Predictions - this might not necessarily end up being an entirely unique category, but I think it's different enough that it should be separated for now. Predictions are created from sensory information, but they're also dependent on processing in our brain, and we don't necessarily have conscious access to what's happening when our brain is making predictions, so we can think of them as a third source of information to create conscious experiences.

I'm going to try to take individual kinds of information from these three categories, and combine them to create definitions of some emotions or feelings. For example, another previous definition was for beauty - "The feeling caused by experiencing something that is unusually easy to process." Which isn't exactly an emotion, but is a definition that's based on what's happening in our brain as opposed to the exactly qualities being observed.

• Sadness - predicting that things are going to be worse than previously thought
• Happiness - predicting that things are going to better than thought
• Awe - More good/interesting information is able to be experienced than expected
• Awful - more bad information is able to be experienced than expected
• Fear - unable to predict how things will be after something bad happens
• Dread - predicting I'll feel bad when something bad happens

These are some fairly straightforward emotions, but I think looking at them as combinations of the possible types of information we can experience consciously might help us think about them in new ways. I'm not entirely certain about the above definitions, but I like the way they're working so far. There's a couple key ideas I see developing:

• Emotions are triggered or caused by a difference between predictions and observations, or by a difference or change in predictions
• We could expand the above list by dividing up the subjects and objects of a change. For example, our emotional reaction to a change would be very different if we feel like it was caused by a change in the world or caused by something a person did. We would also feel different emotions depending on if we're the subject of the change or not.

This gives us a model where emotions would make an efficient way to remember the effects of complex changes in predictions. Instead of having to remember that "when this person left it affected me in ways X, Y and Z, which changed my view of the world in ways A, B and C" we can just remember "when they left I was sad". That doesn't carry as much detail as remembering all the specific reasons it made us sad, but for evolutionary reasons just remembering that it made me sad is probably contains most of the useful information that came out of comparing the changes in predictions or view of the world that happened because of the event.

We also strongly associate certain physical responses with certain emotions (laughing, crying, smiling, etc.). I suspect that what the specific physical response is, isn't as important as having some relatively consistent and unique physical response. For these kinds of complex comparisons we can't rely on feedback from either the world or other people, that means that if we're going to have some kind of physical sensory data to associate with these emotions we'll have to create it ourselves. Emotions are useful ways to compress and save information about changes in predictions, but if we can't use that information to learn to change our behavior it's not very useful. Learning is a process of tying responses to stimulus, and the more stimulus there is and the more unique it is, the easier that process is. Crying when we're sad or smiling when we're happy generates observable stimulus that our learning systems can use to tie these emotions with the kinds of responses that are useful. And that's exactly what we see when these physical responses are studied, people will become more happy when forced to smile for example. They're reacting to the unique stimulus they use to mark those emotions and not necessarily the conscious experience of the emotion. I think this gives some evidence to believe that emotions are both created by a separate kind of system in our brains, and that they're useful for categorizing and storing certain kinds of information.