Homeostasis and embodiment: 

Homeostasis is the purpose and key criteria of all representational  systems. Everything else is a side effect. 

Homeostasis is a representational proposition; it is not a feature of physics.

there are a variety of different forms of representation.  x ; y    (x y) ; z  or (x y) -> z    x = x;y  

And when we think about how representation happens, we find it is a deeply arbitrary.  We can make representations of anything.  This is the nature of improvisation and creativity.  However, the relationship between representations does not seem to be so arbitrary.  the larger structure of representations must make some kind of sense.  

Representations which do not make sense are illusions, or fantasies.  A unicorn is a fantasy.  We certainly know what one is, but we do not find any unicorns in our environment.  We find horses and goats with horns, but not uni-horned hoofed animals.  Nor do we find pegasus, or flying carpets, or conscious machines.  While all kinds of ideas can created, we cannot make all ideas connect to other ideas in the way we want them too.  We can assert that ideas connect together, but it is what other representational connections are associated to an assertion that confirm whether an idea is true, or whether a representation makes any kind of sense. 

And this is an implicit feature of representation making.  What constitutes an idea, and specifically an instance of an idea, is how the idea connects to the universe of other ideas.  We know what a unicorn is, but there is no instance of a unicorn.  For there to be unicorns eating grass or cupcakes in fields would require many other corollary ideas to relate to the fact (and idea) of unicorns.  But these connections do not exist as ideas or representations connected to unicorns.  We can see an instance of a unicorn as art, but not as an actual hoofed animal that lives on farms eating hay, or in forests eating grass or cupcakes. 7
[Did you know unicorns eat cupcakes?  http://www.youtube.com/watch?v=GXVSNPIWFo0&feature=player_embedded]

This connecting of representations is what gives us a sense of reality, for what it is.  Ideas gain more "reality" when they have more connections to other ideas.  All of the representational ideas, which certainly includes mathematics and logic acquire meaning, importance, and power because those ideas connect to other ideas in ways which themselves are meaningful.  This process is the same as how we establish truth.  Because truth is another idea. 

For a representation to be true, it must have some corollary representation or representational process which asserts it unambiguously.  For instance, "If I drop a rock, it will fall.  that statement is true because of gravity."  This basic construction is how all of mathematics and logic work.  Propositions must be accessible through transformations following a fixed set of rules or axioms for them to be true.  

This web of interaction is how we assert truth, or validity, or reality, or meaningfulness to individual representations.  The representations which "are" true are ones which reflect some other representation which is itself established to be true.  Just as the above (x ; true) reflects the (x ; y) and both reflect the true value established by some web of conditions of reflection and representations.  

Now of course these webs can be shown to be false, but what produces falsity is contradicting representations.  When some data element that is putatively (a representation itself) subject to some representational construct, and it contradicts that construct, say through the sensory representations and perceptions we make, then we have a conflict in the representational web.  Scientific advancement shows this process of conflicting representations very well.  When some data or fact conflicts with or sits outside of the web of assertions and demonstrations of the connected representations, this is when we engage in the process of developing new connections and relationships between our representations to connect the contradictory or anomalous fact.  To account for, to embed the excluded or conflicting data or fact or representation into the web and determine how this new data is also true.  Usually this activity involves creating exceptions or special descriptions, but eventually, a contradicting web of representations is eventually discarded for a more comprehensive interconnected web of representations. 

But why should development of interconnection between ideas and representations happen?  And why should it matter whether the ideas we have are "true" or not?  

Because these conflicts force the production of new ideas and representations in organism which experience the representations in conflict. The purpose of a representation system is to maintain the homeostasis of the organism, and when representations are in conflict, this is an indicator that the organism is not functioning correctly because it's representations produce conflicting responses.   And conflicting responses for moving organisms is a prescription for death.   the organism that cannot successfully flee a predator, because it cannot maintain a coherent map of its environment will be eaten much more quickly than one which makes rapid choices to evade a predator.  Also, the organism that makes the correct choices during evasion or pursuit to ensure it's continued survival, a homeostatic imperative, will survive and prosper over organisms which do not.  

The homeostatic imperative of an organism thus drive it's representational responses to representational conflicts.  this also explains why established players do not disrupt their own economic advantages when faced with contradictory data or theories People who have relied upon certain ideas, certain representations to build careers and lives that maintain and improve their own homeostatic functioning will require some sufficiently powerful homeostatic response that is self-originated to change their own minds.  This is as true in politics and art as it is in science and technology.  

This homeostatic imperative of an organism will bind the organism to representational biases which it has found successful and allow new organism which are not bound to successful representations to adopt new representations in changing conditions.  

But why should this be?  Because representations are embodied in the organism itself.  the cells and networks of the nervous system are where representations are established.  The complex development of these representations which are constantly being made and unmade produce a structure that serves the homeostatic imperative.  If these structures do not form, the organism will simply die. 

the organisms that cannot form, and sufficiently adapt (remake) it's representations to changing inputs and conditions is less well served than more creative and adaptive organisms.  Of course some adaptive behaviors may put an organism at risk, and thus overly creative and adaptive organisms may be at risk of termination.  These survival and homeostatic "pressures" constantly affect the representational system of the organism.  But the homeostatic imperative can only work using a nervous like system, a network of signaling and structures signaling cells because that organism embodies representations with it's nervous system.

Representations cannot be extrinsic to the organism, but must be facts and features of the organism itself.  If the representations are extrinsic, then extrinsic "rules" guide the making and unmaking of representations.  And that sort of model leads us down a rabbit hole of duality and extrinsic influences that themselves have extrinsic influences etc. 

Further, representation making must always be going on in the organism to maintain the survival and homeostasis of the organism… lest you get run over by a car, or forget where or what your car is and cannot make your way to the safety of your home and perish in the desert after you had to stop for a pee.  Which is a reminder  that whenever we have representations and meaning we also have confusion.  Errors and confusion are as much a part of representation making as consistency, coherence, and truth.  We can and do make confusing an erroneous representations and coherent and truthful ones.  

the principles of confusion and coherency exist as features of any representational system that embodies representations as means of homeostatic survival.  thus a machine consciousness must also embody it's own representations to understand the homeostatic imperatives that give rise to confusion and the importance of coherency.  If confusion and coherency are extrinsic principles, and not intrinsic to organism survival, then how could organisms function?  Where do confusion and coherency come from?  Certainly not from some extrinsic "force" which affects the physical phenomena of molecules and cells and cell structures to make them "coherent" or "confusing".  

Coherency and confusion are the result of the functioning of cells and networks of cells themselves.  And these principles are stigmergic conditions which alter the interaction of molecules in cells and thus produce changes or adherence to existing representations and functioning.  Basically, conflicts in a cellular network are also representational conflicts.  And representational conflicts are conflicts in the cellular network.  

[The fact that altering the cellular and molecular function of a nervous system alters representation, meaning, and coherence is well demonstrated with hallucinogenic drug use and incidence of brain damage.  The corollary is also true, the religious experience, the enlightenment and uplift or fear that a book or movie produces, the experience of moving art, the thrill(good or bad) of getting naked around other people are examples of representational phenomena producing obvious molecular changes.  But more deeply, the power of education and learning opportunities also produce changes to the development of the central nervous system, and thus of the body, and even of events in the world.  The athlete changes their body through the basic idea at getting better at some skill.  The chemist changes the chemical environment through their experimentation.  Manhattan project participants produced atomic explosions all in the pursuit(an idea) of an idea. ]

The conflicts we see in networks would be halting conditions, loops, failures to complete loops, and thrashing of signals.  When these events occur, there must be other cells and network structures which reflect the nature of these problematic cells and those networks are responsive to those conditions and initiate signals to the affected networks to alter their structure and possibly individual cell functioning.   Simply altering the structure of networks between cells can resolve loop issues, halts, and thrashing.  This process of network development is learning.   And this kind of adaptive response can only be achieved in an embodiment model.  

This is not development.  Development is a process that is driven by the chemistry of the cells themselves and is well described by the processes of systems biology.  The development of visual and auditory cortexes is a morphological feature driven by the physical phenomena of systems biology.  the fact cells make connections to each other is a fact driven by the systems biology as well.  but above that level, once representational structures are made by the organism at large, then we have the systems biology driven activity, that produces a layer of representational learning.  And representations, because they are structures of cells, have a causal effect on the molecular physics in cells as different cells and different cell structures impact the environment of molecules to produce different kinds of molecular phenomena in and on the network structured cells.   

It is the network structures of individual cells (not just neurons) which both instantiate representations and ideas, but which also induce causal action down into the molecules of the cells in that structure.  Thus a change in the function of the structure, produces a change to the molecular activity of individual cells in that structure.  At the individual cell level, the cell's homeostasis will dominate all of it's behavior, but small individual fluctuations will happen.  It is the aggregation off these fluctuations, it is the coherence of incoherence of these fluctuations which produce the outputs of from nervous system movements.  

For instance, the development of aggregate and fine muscle control requires the development of neural networks which signal muscles in aggregate and discrete ways.  Destroy the networks which instantiate motor control, or destroy the networks which direct motor control and either of these will alter muscle control and movement.  This is exactly the phenomena we see in stroke and brain damage  [Descartes Error, Antonio Damasio 1994] [see: Biology Seminar Series: Franz Goller, Professor, Department of Biology, University of Utah "Motor control of birdsong: Peripheral answers to central questions"]

The kinds of network structure processes that manage motor control drive the phenomena of regulating hormones, of vision, of proprioception, of emotional control, and the other nervous system processes of managing a body.  And these network structures can be developed and manipulated through representational activity, ie. learning. This is the whole point of a nervous system, to respond more quickly to the environment through change of the organism.  Most typically this involves motor changes, but it also involves other kinds of changes. The organisms that can most quickly adapt to changing environmental conditions will most likely survive.

The instantiation of representations as network structure processes of an organism is how representations are embodied.  The purpose of the embodiment of representations is to maintain homeostasis of the organism.  The representational and cellular systems both serve the homeostatic imperative of the organism.  This embodiment of representation occurs all the way down into the cell's molecular process.