Marrying within Your Class: The Graph Theory of Endogamy
Chapter 4 of The Information Class
Endogamy means “marrying within.” As such, the concept is applicable to a wide variety of groups that may practice endogamy. There is religious endogamy, racial endogamy, clan or tribal endogamy, and so forth. What makes class different is that I’m defining class partly on endogamy. This risks being circular: you marry your class, and your class is who you marry.
It’s not actually circular. But it’s subtle.
Think of an abstract graph, consisting of many nodes and many links connecting them. What descriptions of the graph might be possible? We could describe it as dense, if there are a lot of links, or thin, if there are few. We can describe it as unified, if there is a path through one or more links connecting any two nodes in the chart, or as segmented, if some subgraphs are disconnected. More subtly, there might be dense parts in the graph, such that five or ten non-overlapping paths connect any pair of points, while on the other hand, for some nodes or groups of nodes, a single bridge may be the only one to move between them.
With this in mind, think of class endogamy as a certain kind of description of the family graph of a society, where the family graph represents all persons as nodes and all sexual and parental relationships as links. Within the family graph, an endogamous group would have a higher density of links internally than externally. Metrics and thresholds would need to be defined to make this quantitative rather than merely intuitive. If endogamous groupings are identified, some would be more or less class-independent, since other factors, such as race, religion, hometowns and hobbies also affect the likelihood of two people marrying each other. Other traits would identify the endogamous clusterings as classes.
To identify classes within the family graph, endogamous clusters would first need to be catalogued, and then one would need to look at commonalities and shared characteristics that explain the endogamous pattern. Key indicators of a class would be, above all, educational and occupational similarities among members of an endogamous cluster, but many other things like tastes in clothing, food and entertainment; vocabulary and speech patterns; moral attitudes; and income and wealth could also be class markers.
A major theme of this book is that the United States, and likely the West more generally, is becoming increasingly class-stratified. Far more evidence than I could easily assemble, or than I want to take space to summarize, has led me to believe that. It includes narratives, statistics, academic papers, and much personal experience, but I don't want to exaggerate how conclusive it is. I'm far from certain that “the information class” as I define it actually fits my definition of a class very closely. I'd love to see someone compile the complete family graph and see whether there really is an information class with an emerging propensity for endogamy. It would be fascinating to really know, to have the data in hand.
Unfortunately, I don’t think anyone’s likely to build the family graph, even though marriage and birth records are public, because:
1. Social science has a strong habit of conducting its data analysis based on sample surveys, which are poorly suited to studying endogamy.
2. Laws about the privacy of Social Security Numbers and other personally identifiable information (PII) get in the way of efficient data sharing.
Also, the family graph that matters extends through past and future. Aspirations to a particular kind of family life are an important part of class identity. In a society where the information class of the 2010s arose recently out of the classless society of the 1960s, I wouldn’t expect patterns of endogamy to reach far back in time, and lots of professors probably still have cousins who are truck drivers. In future, if present trends continue, I anticipate that there will be fewer family ties across gulfs of occupational prestige.
It's ironic that social science methodologies and government data privacy rules make it hard to assess the state of, and trends in, class stratification, because past societies that in general were far less information-rich actually seem often to have had better information on this one point. They cared about genealogies, remembered them, and calibrated their expectations of a man to align with the track record of his lineage. We're unusual in our ignorance of family historiea, our own and those of others. Some websites try to close the gap, helping people learn about their ancestors. Good for them!
Despite the weakness of the data, my sense is that the evidence is strong, though not wholly conclusive, for the emergence of class stratification in my sense of the term, and that college-educated people who work with their minds using computers increasingly comprise an endogamous class. I'll proceed on that assumption.