Edit: First paragraph removed and small changes made to the rest.
I am putting forth a hypothesis is about the nature of consciousness. First I will have to tell you how I am using certain words because they are generally used in a number of ways. 'Brain' is an biological organ and it has a function, 'mind'. Mind is not an object; it is what brains do. It is not a property of the brain, let alone an emergent property, whatever that is. It is a function - so mind-is-to-brain as circulation-is-to-heart or digestion-is-to- intestine. There is one brain in any head and there is one mind being produced by that brain – not two. (Assuming sanity) The different parts of the cortex work together; the two hemispheres work together; the fore-brain structures work together with the mid-brain structures. The mind includes at least: perception, cognition, learning, intention, motor control, remembering, and most importantly, the forming a model of the environment and the person in that environment. The division between 'conscious mind' and 'unconscious mind' is meaningless. The brain does its mind-function which maintains the model. Some but not all of this model is made globally accessible to all of the brain and remembered. That edit of the model is what we experience as conscious experience, in other words, is our 'consciousness'. Consciousness is awareness not thought. Consciousness is not separate but part of a single mind-function. Now that the words are straight, I can describe the hypothesis.
How is the model edited for consciousness?
There is an attention focus that is triggered by the on-going work of the mind and the events that happen in the environment. I may concentrate on some task so that I am not conscious of other parts of the model but a loud noise will cause my attention focus to shift to the source of noise in the model. The level of attention is variable from non-existent (coma) to intense. This level depends on the signals coming from the lower parts of the brain, through the thalamus, into the cortex. A common analogue for attention is a searchlight scanning the mind-model of reality. We cannot be aware of the whole of the model at any instant of time.
How is the model formed?
The fragments for the model are forced together into a best fit global model. The perception of the various senses, inborn constraints, our understanding of the world, our memory of the previous seconds, our expectations etc. together build a cohesive model by constructing a synchronous neural activity. Fragments that cannot be fit into the model are lost from it. This is done by an almost unbelievable number of parallel, slightly overlapping feedback loops, across the cortex and between the cortex and the mid-brain (especially the thalamus). The feedback loops are much more like patch boards then like digital computers. They rattle for an instance until they find a stable synchrony. There is nothing like step-wise processes at this stage of forming a global model.
How is the consciousness edit of the model used?
There is little doubt that consciousness is useful because it is biologically expensive. Evolution will eliminate expensive functions that do not earn their keep. There are three very important processes that are carried on by the consciousness aspect of mind.
1) The working memory that holds the last few frames of consciousness is the source of episodic memory. There is an important link between consciousness and the formation of memory. We could treat working memory as part of consciousness or part of more permanent memory or even the link between them. Consciousness is in effect 'the leading edge of memory'. No conscious experience of something than no memory of it.
2) The working memory allows some cognition and learning that needs to 'juggle' information. I cannot add two digits if I cannot retain one while I perceive the other. So some thought processes are going to be in the edited model so that they are be continued through the use of working memory. This does not constitute a conscious mind that is separate from an unconscious mind. It is only that some types of thinking register bits of their progress in our awareness so they can be retrieved later.
3) Consciousness does mild prediction and therefore can register errors in perception and motor control. It takes a fraction of a second to form the conscious experience of an event. But we do not live our lives a fraction of a second late. Information from time t is used to form a model of what the world will be like a t+x and x being the time it takes to create the model and its conscious edited version, then we will seem to experience t+x at t+x. The difference between the model of t+x and what comes in via our senses at t+x is the actual error in our perception and motor control and is be used to correct the system.
These three functions seem ample to justify the metabolic expense of consciousness and rule out philosophical zombies. The functions also seem to rule out consciousness being uniquely human. 'If it quacks like a duck' logic applies to animal consciousness. If an animal appears to have a good memory of events, learns from its experiences, has smooth motor control in complex changing situations, then it is hard to imagine how this happens without consciousness including self consciousness. There would, of course, be degrees of consciousness and variations in the aspects of environment/self that would be modeled by different animals.
My answers to some problems ahead of their being asked
Most readers of this site are comfortable with the idea of the map and the territory. This post is using a very similar (maybe the same) idea of reality and model of reality. There is nothing surprising about the difference between the physical tree in the garden and an element that stands for that tree in my model of reality. It is the same idea to think about the difference between the reality-now and the model-now. The difference between my physical leg and my model leg is not difficult. We need to extent that comfort to the difference between the reality-me and the model-me. Introspection gives us awareness of our model, it is not our reality-mind but our model-mind we are turning our focus of attention on. There is a difference between reality-decisions and model-decisions. We live in our model and have absolutely, positively no direct knowledge of anything else – none ever.
I have given no evidence for the hypotheses here but for two years I have been collecting evidence on consciousness in my website, Thoughts on thoughts. My hypothesis is not that different from the one that Academian is giving in his series of posts and I do not mean mine to be in opposition to his, but to a large extent supportive. Treating consciousness as a sense is not that different from treating it as as a selective awareness. There is no need to get hung up on the words or analogies we use.
I have side-stepped the 'hard question' of how and why red is experienced as red. I have the feeling that this is a 'wrong question' but I am not sure why. It is certainly not explained by the hypothesis I have given here. All I have to say about the hard question is: “Can you think of a better way to be aware of red then the one you have?, Is there something more efficient or more vivid or more biologically functional?” In other words, “What is the alternative?” Even if you go all spiritual, that still does not explain the experience of red. Dualism does not answer the hard question either and I have not encountered any philosophy that does. If it is answered, I would put my money on a scientific, material answer.
I have not side-stepped the question of how consciousness is reduced to physics. The method is clear: reduce consciousness to biology and biology to physics/chemistry. We accept that biology is in principle reducible to physics/chemistry. We generally assume that the brain is understandable as a biological organ and so if we can assume that consciousness is a function of the brain, it is in principle reducible to physics.