There’s plenty of glib, paint-by-numbers conservative fnord-phrases in this Kay Hymowitz piece on artificial insemination—”pretending nature doesn’t exist” and other such vacuities—but conspicuous by its absence is that pro-lifer favorite “life unworthy of life.” I say this because for all Hymowitz’s criticism of policies that enable small numbers of single women to bear children via AI (she cherry-picks some very obviously unrepresentative clinics in an effort to inflate the number, incidentally), the piece seems a bit coy about acknowledging the obvious upshot of the policy changes she endorses.

By her own account, the effect of denying anonymity to donors—and certainly of making them liable for child support—would be to reduce the number of available donors. Maybe a few women do end up getting married and having a child “the old fashioned way” as Hymowitz would like. But that seems unlikely for lesbians or women who’ve chosen AI precisely because they fear they’re nearing the end of their most fertile years without having found a partner they’re prepared to have a child with. Many will probably adopt instead, which might be all to the good. Some will have friends or acquaintances willing to donate, trusting that the recipient doesn’t expect them to play the paternal role. And many will probably end up just foregoing childbearing.

The only case where the policy change unambiguously pushes outcome in the direction Hymowitz wants if for women who have (or will soon have) men with whom they’re willing to reproduce, but had nevertheless opted for single motherhood via AI. In these cases, though, even assuming one-for-one displacement of AI by “natural” children, note that it’ different kids coming into the world. That is, the AI child never exists, and instead there’s a distinct naturally conceived child who may be better off along some dimensions. There’s no AI child who’s made “better off” across the two scenarios. In any event, I find it unlikely that this is a huge proportion of AI births—since, after all, why resort to costly AI in those cases?

In the trusting-friend cases, the outcome typically ends up being de facto the same as under current law for the most part. In the adoption cases, we get possibly improved care for existing children at the cost of fewer total children. And for the remainder of cases, we just get a net decrease in the number of children.

The upshot, then, is that the policy Hymowitz advocates, in the cases where it actually effects the outcome, will most often do so by decreasing the total number of children. The vast majority of these would have been born to well-educated women with both a strong desire to raise children and the economic means to provide for them. And for these children, the relevant question is not: “Would it be better if they grew up in a household with their biological fathers?” It might, as it might be better if they grew up with superpowers on a pony farm. That is not on the menu. The relevant question is: “Would it be better if these children did not exist at all?” Hymowitz, it seems, needs to answer that yes, it would—that the world is worse for their existence.