Dear James,If the idea that an experience cannot be spread over lots of cells needs an explanation I would recommend chapter 6 of William James's Principles of Psychology on 'The Mindstuff Theory'. Lots of other people have made the point but James uses simple ordinary language to reach the conclusion that an experience spread out over the various parts of the brain would not be 'a physical fact'. For anyone not used to tracing causal chains in biology as part of scientific research it might not be immediately apparent what is meant by this but to me as a biologist it seems clear and cogent.

Before trying to unpack 'physical fact' let me just deal with the neurons firing. You need lots of neurons to fire to produce enough independent signals to generate a rich experience. However, these signals will cause an experience in something that they all signal to. Unless they all signal to the same thing there will be nothing that gets all the signals so nothing to experience their combination. The firing itself cannot be an experience because by 'the experience had by X' we mean the way the world influences X, not the way X influences the world. 'Cell firing' is an account of how a cell influences the world.

James Blackmon has a nice essay somewhere on the net about the paradox of multiple cells being involved in an experience. If you suggest that parallel events in cells A,B,C,D,and E in a brain together constitute an experience then there is no principled reason to exclude cells G and H in that brain or cells P and Q in the brain of the person in the next house. They are all spatially separate events. The standard thing is to then say that events in A,B,C,D and E are 'together' because of some pattern of connection between these cells. But connection is only relevant to sequences of events, not parallel events. Moreover, if you try to add together sequences of events in connected cells you get overdetermination and an infinite regress in time. You cannot add the signal from A to B to the one from B to C because the one from B to C will be dependent on the previous one from A to B - and so on for ever. 

It might seem arbitrary to raise these objections, although for many people William James's way of saying this is crystal clear (I often feel that my trying to add ends up less clear). However, it is not arbitrary. It reflects an aspect of physics that is so fundamental and intuitive that nobody talks about it - what is called locality. Physics presumes that there are events that connect in dynamic, or causal, sequence. Parallel events are not connected in any sense. Special relativity showed that this is not just a way of looking at things. It is an empirical fact. In a certain sense there is no ontology beyond epistemology - all that exists is passage of information. 

Classical Newtonian physics does not stipulate the grain of 'events' because its maths assumes that it is always dealing with aggregates of events that if you investigate more deeply will always prove to be composed of finer scale events, on to the infinitesimal. So although you can treat the earth as having one 'centre of gravity' the assumption is that this is a mathematical trick and that it reflects the separate effects of an infinite number of material components each acting exactly where they are. Newton was embarrassed by the fact that this action did not in fact appear to be local - it could work 93 million miles away - and again it took Einstein to restore locality with a better theory. Leibniz realised that actual reality ought to have a, monadic, grain of real events. He could not work out how to describe these mathematically but he saw that their way of relating would be quite different from intuitive ideas of mechanical causation. Quantised physics has given us the grain - the quantised mode of excitation - and agrees that the way events relate is very unexpected (for all except Leibniz).

So at last the common sense idea that all events happen at the place and time that they happen and always in a specific sequence now has a rigorous mathematical base. It is a bit odd because each event is in fact extended in domain. However the relation between events is always determined by field values at defined points in spacetime and the sequence of relation between events is unique. We can in fact forget all the difficult maths of quantum field theory and just be reassured that at root the common sense idea that there are real chains of real events in physics is well grounded.

This is all relevant to experience because in physics all attempts to study how the world works involve chains of events, the last of which is a human experience. As people like Wigner pointed out, you might say the last event is the activation of a sensor in a videocamera, but to use that in physics someone has to look at the display screen. To interpret any observation we make the assumption that all events relate locally - and note that this has to include the final event of experience. If any event in the chain is allowed not to have an address in space and time the whole theory crashes like a house of cards. You have an absurdity like the fact that all conclusions from a syllogism with two incompatible premises are true. If you think about it we do normally insist that the experience is local. There is no point in trying to observe an eclipse of the sun a week late or on the wrong continent. All predictions for observations have a prescribed location in space and time. Normally we do not bother to stipulate more precisely than where the observer's body has to sit but for clinical neurology we do. We predict that an observer located in the brain will not observe a tap on the finger if the median nerve has been severed. All neurology assumes that locality applies deep into the central nervous system. Descartes used that to work out that observers ought to be in the pineal. He included a false premise (the apparent need for a unique rather than a paired site) but otherwise had the physics right.

The weird thing is that when neurobiologists start to write about consciousness they throw the law of locality out of the window. The say Descartes was wrong. They suggest that experience somehow emerges from lots of firing events in lots of places at once. This is completely incompatible with physics in any known context - for the reasons above. The reason why they do this is that they cannot see how experience could possibly be local, because of the same false premise that Descartes introduced. They assume that there is only one copy of experience at a time. And since we have good reason to think that experience occurs in both sides of the cortex and in several places in each it might seem that it has to be smeared over lots of cells. But the simple alternative is that there are lots of copies, each in an individual cell.

William James considered this idea in chapter 6. It is not at all new. However, he argued, reasonably on a Newtonian basis, that even a cell could not host one local event. Events had to be infinitesimal, or at least belong to atoms. However, the day has been saved by the demonstration in modern physics that individual events can actually have very large domains. Quantised units are not tiny. If anything they fill the universe - but all their relations have precise addresses in spacetime. The individual computational events in brains are instances of post-synaptic integration in individual neurons. There is good reason to think that in quantised condensed matter physics these can be real individual events. We want events that are big enough to get about 1,000 degrees of freedom in the input of relevance to signalling but not involving separate components that have to be seen as in sequence (thus more than one event). This is the obvious scale to choose.

Viewed in this way brain function is remarkably easy to understand. Each cell has a rich input that it experiences and a single spike output that contributes to the experiences of the next cells along. It could not really be anything else.

Another quick summary for you.

Best wishes

Jo