Readers have reacted with a well thought-out lack of enthusiasm for my attempt to define "class" as "a group of potential in-laws, people who might turn out to be ancestors of mutual descendents."
While there's a clever insight in focusing on the marriage market as central to issues of class, self-congratulation got the best of me. There appear to be at least two problems.
First, it extends the purview of class beyond any previous definition. Most notably, religion has traditionally had much to do with who-marries-whom, but to redefine religions as classes reeks of Humpty Dumptyism:
'When I use a word,' Humpty Dumpty said, in a rather scornful tone,' it means just what I choose it to mean, neither more nor less.'
'The question is,' said Alice, 'whether you can make words mean so many different things.'
'The question is,' said Humpty Dumpty, 'which is to be master - that's all.'
Similarly, region plays a banally obvious role in who marries whom (people from South Carolina are more likely to marry other people from South Carolina than people from Oregon). It doesn't help much to define, say, a South Carolina factory worker family as a different class from an Oregon factory worker family just because, say, the odds of their intermarrying is lower than the odds of the South Carolina farm family intermarrying with a South Carolina factory-owner family.
(Of course, region can overlap with class. Midwesterners are notorious for po-mouthing their origins -- I'm just a boy from a town in the Midwest so small you probably have never heard of it, Kenilworth, Illinois ... For example, the rich characters in The Great Gatsby feel slightly oppressed knowing they are from the Midwest. Texans, in contrast, seldom publicly admit feelings of class inferiority due to being from Texas.)
Second, "class" as typically used implies levels of hierarchy about which there exists something of a consensus within a society about which classes are lower and which are higher (and thus which are preferable to marry into). Traditionally, these judgments are based on a complicated combination of wealth, bloodline, moral merit, job, education and so forth, so there is much room for disagreement, but much interest in ranking people on this dimension (see Jane Austen's novels).
Still, it makes sense to visualize a particular class as spreading horizontally across the social map, while other factors that affect who marries whom like region and religion are more vertical, and thus intersect class. Thus, say, an wealthy ultra-Orthodox Jewish family and a wealthy Parsi family might belong to the same class, but aren't likely to have common descendents within the next few generations because they both insist upon endogamy.