andyandy, what would you think of this scheme as a working definition of consciousness? (Adapted from Dennett)
Scenario A: A computer, light sensor, and monitor are rigged up so that when a beam of light in the red spectrum hits the sensor, the monitor displays "I see a red light". A series of different colored lights are flashed. The monitor is blank until a red one is lit, at which point the monitor displays "I see a red light".
Scenario B: I'm asked by an experimentor to look at a screen and say when I see a red light. A series of different colored lights are flashed. I remain silent until a red one is lit, at which point I say "I see a red light".
Why do we think of only B as involving response by a conscious entity?
It has nothing to do w/ response to environment, or discrimination, or following orders of course -- it's because we know that there's an "experience" of "seeing a red light" for the human. We feel instinctively (or argue logically) that there's not for the machine.
So I take your OP to mean something like "At what point did some critters stop merely reacting chemically to the environment and start having some inner experience, a subjective 'feeling' of pain, fear, happiness, etc?"
Tough stuff, since, as others have pointed out, we can't get at that "experience", that "feeling".
But still, to me this is the crux of the issue, not some reflective sense of "I", which may come much later in terms of evolutionary development. To me, it matters if an animal has something like my experience of pain, rather than, say, a computer simulation.
Why? For the same reason we ask questions like "
Can Fetuses Feel Pain?". For the same reason it mattered whether Terry Schiavo did or didn't have any higher brain function. For the same reason it matters whether comatose patients can sense their environment.
It matters to me if, say, a cricket is just a bundle of wires like a machine, or if a fish is experiencing pain in something like the way I do.
It's just my opinion, but I think Dennett is very likely to be right when he proposes that a dual brain structure underlies the emergent phenomenon that we call conscious experience or feeling. In short, the model is: Build brain A (brain stem, lymbic system, cerebrum) to process input, then build brain B (cerebellum) to "live inside" brain A. That is, brain B receives processed data from brain A, so that brain A's output becomes brain B's environment. The feedback loops between these "2 brains" produces the sense of the world "being like" something.
If that's correct, then the bug my cat is chasing is just an organic machine, but my cat is a "feeling" animal. The bug has no experience of panic, but my cat has an experience of excitement.