More on late-term abortion: Supervenience

I posted in January about the statements of John Harris, a British philosopher who outraged many by suggesting that there is no morally significant difference between a late-term abortion and infanticide. The more I think about this issue, the more explaining I feel like the pro-choice camp has to do. In Jamie's class today, we discussed supervenience, in the context of the supervenience of the ethical on the descriptive. A common analogy to draw is one to that of the supervenience of biological properties on chemical ones. Necessarily, if A and B are identical with respect to all chemical properties, then they are identical with respect to all biological properties. I'm pretty sure this is more or less uncontroversial. That's why I was surprised to realize how directly it refutes the sometimes-cited pro-choice understanding of life as beginning at birth. Consider two identical fetuses in identical wombs. Mother A is drug-induced to enter labor at time t-1, while Mother B does not enter labor until t-3. At t-2 (say, one minute before t-3), Baby A has been born and Fetus B has not -- there is only a geographical difference between the two -- there is no chemical difference. Therefore, by supervenience, there is no biological difference. Therefore, since 'is alive' is a biological property, it cannot be the case that A is alive and B is not. This means that life does not begin at birth. I know that you don't have to think it does to be pro-choice, but if late-term abortion is morally permissible and infanticide is not, then this difference requires an explanation. (For the record, I think that life begins gradually, that fetuses are morally significant from the point at which they can experience pain but are much less morally significant than adult humans, that infanticide harms the victim only very slightly more than late-term abortion does, and that infanticide is much less morally reprehensible than the killing of either a five-year-old or an adult.)