More Armstrong, Conceptual Analysis

Thanks for the helpful comment regarding conceptual analysis. I share the skepticism of conceptual analysis, although I must admit I have a lot more work to do in that area- I need to know more about conceptual analysis before I can sustain a legitimate critique. Here are some more thoughts on that tangent:
(1) I agree that conceptual analysis appears to be a degenerating research program, which is part of the reason why I think it might be interesting to think about conceptual analysis in Lakatos' terms. That is, let's put all of our cards on the table: What is it exactly that we are trying to explain (for example, in my post on Place I wondered what the parallel would be for explanation involved in identifying lightning with electric charges), what is the "hard core" that the different approaches are committed to, and how do the different approaches compare with one another in terms of the standards for successful and degenerating research programs.
(2) Assuming we're skeptical of traditional forms of conceptual analysis: Is there still any sense in which understanding/discovering "consciousness is a brain process" is different from understaning/discovering "lightning is electric charges", in the sense that it seems easier to understand the meaning of "lightning" than "consciousness" or any other mental concept?
(3) Again, assuming we're skeptical of traditional forms of conceptual analysis: Perhaps this needn't preclude us from giving some sort of functional account of mental states (as opposed to mental concepts), but maybe we'd need to provide different types of arguments in support of the functionalist account? That is, to what extent are functionalist accounts dependent upon conceptual analysis? And to the extent that a functionalist account is not dependent upon conceputal analysis, how does it compare with Place's identity theory?
(4) Poison vs. Lightning: Is it more instructive to think of mental concepts as analogous to lightning (with the proviso that Place is not including cognition/volition in this analogy), or as analogous to poison? I suppose the poison analogy lends itself more to the conceptual analysis project, but there are problems with the lightning analogy as well: I'm repeating myself here, but if the meaning of "lighting" was on par with the meaning of mental concepts, philosophers would not be so interested in mental concepts. Of course, it could be that the interest has been systematically mistaken (this is the conclusion of eliminative materialism, yes?), but there is still work to be done in explaining away the apparent difference.
(5) Finally, I just want to point out that the conceptual analysis of poison here seems weird, although maybe this is just a reflection of my skepticism of conceptual analysis: Is the game here to try and find necessary and sufficient conditions for poison? If so, a simple causal analysis clearly doesn't accomplish this. Do we just look up poison in the dictionary, as Armstrong seems to have done?

Kripke on the Brain

Thanks to Danny for talking this out with me today. To recap, are we settled on the following (truncated, preliminary, and rough) reading of Kripke's arguments?

1. We discover the identity of (epistemically) disparate phenomena, like lighting and electric discharge, by identifying contingent properties of the relata (lightning might've been colored red and been identical to the motion of sugar molecules, for example).
2. But if x truly is y, x is necessarily y.
3. Epistemically disparate verification conditions do not entail contingent identity.
4. What appear to be 'contingent' identities are metaphysically necessary (the contingency is merely epistemic and so illusory).

For instance:

1. If pain is a brain state/process (say, c-fiber firing), it is necessarily c-fiber firing.
2. But we can imagine c-fiber firing with no correlated pain.
3. So, pain is not necessarily c-fiber firing.
4. So, pain is not a brain process (c-fiber firing).

Alternatively, or in addition:

1. If the identity relation holds at all, it necessarily holds.
2. The a posteriori discovery of identities does not make them contingent.
3. If a = b, then a and b share all essential properties in common.
4. If pain is c-fiber firing, then pain and c-fiber firing share all essential properties in common.
5. We cannot imagine pain in a world with no creatures to experience pain.
6. We can imagine a world in which a creature x's c-fibers fire but x experiences no pain.
7. The feeling of pain is essential to the concept PAIN.
8. The feeling of pain is not essential to the concept C-FIBER FIRING.
9. So, pain and c-fiber firing fail to share all essential properties in common.
10. So, pain is not c-fiber firing.

If this is close to right, the argument is a very Cartesian one; where conceivability is awarded the status of truth-maker (or breaker). Please amend. I'd like to have an adequate synopsis of what Kripke is up to before moving on.