Definitions of race range from those who say it does not exist – that it's purely a social construct – to those who insist it has a biological warrant. All sorts of semantic problems arise, let alone political ones: if there is no such thing as race, what does “diversity” mean? If on the other hand race is biological, surely that is morally irrelevant?

I’m wary, by default, of “social construct” arguments; they usually only shift the issue elsewhere - failing, or not attempting, to answer why certain “constructs” repeatedly emerge in the first place. Also, something being a social construct doesn’t necessarily mean it exists any less; language is surely a "construct", but nobody would deny its existence.

The biological notion of race has something in it, vaguely, but is very limited. Meeting another person, almost anyone has the sensation of knowing – however irrelevantly – whether the other belongs to the "same race” as themselves; there are globally recognisably traits. Beyond that, what can biology explain? Apparently little.

Recently I came across another definition, worthy of note and perhaps plausible (from an article by Steve Sailer, which I won’t summarise – read it and decide for yourself):

“A racial group is an extended family that is inbred to some degree”.

"Inbred", obviously, is not meant pejoratively in the sense of within-a-family; the "extended family" here is a community, somehow defined and perhaps vast. I think the definition is better if phrased in less loaded language:

"A race is any group with a tendency to endogamy".

It is interesting if only by containing the possibility of a feedback between cultural and biological causes. The question is: what are the boundaries that define the in - or out - of inbred?

These boundaries probably have nothing at all to do with biology. They might be religious, territorial, traditional, or whatever else; ethnocentric in some general sense. Over generations, this process might have the effect of converging the gene pool of the group towards certain defining biological characteristics, which may themselves become a factor determining the inbreeding boundary. Such boundaries are of course porous, more-or-less vague, but they do seem to exist.

Religions certainly encourage inbreeding (to understate the matter). Marriage partners must typically be co-religionists. Aside from the scripture and doctrine edifying this, a modern manifestation is the proliferation of religious dating sites: Christian, Muslim, Jewish, etcetera. If these breeding restrictions apply strictly enough and for long enough, presumably a correlation between culture and biology can under certain conditions develop; religion and "race" become less distinguishable. Especially in religions that do not go in for proselytism. In religions that do so extensively - Christianity and Islam - any genetic convergence is probably countered by the ongoing import of new genes.

Interestingly, while race-based (by which I mean ethnic/biological) dating sites (Asian, Arabic, Black, White) are complemented with all combinations of inter-racial services, if you search for inter-faith dating, you only get pages on how to deal with its "problems".

So religious boundaries for in-breeding seem the strongest by far. In America for instance:

Martin Luther King Jr. famously declared that "11 o'clock Sunday morning is the most segregated hour of the week ... And the Sunday school is still the most segregated school." That largely remains true today. Despite the growing desegregation of most key American institutions, churches are still a glaring exception. Surveys from 2007 show that fewer than 8% of American congregations have a significant racial mix.

Religious doctrine and niche dating services - by prescribing or promoting breeding within the group - can be understood both as part of race's “social construction” and as a cause or clustering of biological characteristics (a "cline"). The word becomes flesh.