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.