Genealogical Trees and Networks: Insights from Evolutionary Biology

When scholars in the historical disciplines invoke the analogy of genealogy, they tend to assume one of two conceptions of lineage, depending on their goals. Either they want to trace a line of direct descent, or they want to emphasize the multiplicity and hybridity of origins. (I leave to the side for now the aspirations of critical genealogy.)

The topological difference between these forms of genealogy is that between tree and network. A tree captures the branching of a lineage, but it cannot represent the intersection of lineages, what biogeneticists call interbreeding. For that we need a reticulating network—a diagram that does not just branch, but also connects branches to each other.

Phylogeneticist and bioinformatician David Morrison explains:

The conceptual relationship between trees and networks is straightforward: trees are a subset of networks. That is, trees are networks without reticulations. . . . Thus, . . . there is not really such a thing as a ‘‘family tree’’—there are only ‘‘family networks’’, which may or may not include inter-breeding. 

(The “chain” in the figure above is simply a series of nodes, each of which could be arrived at via a tree or a network.)

In the field of phylogenetics, both kinds of network—the tree and the reticulated network—accurately represent aspects of genetic relation. The evolutionary relationship between the same set of species can be understood as a tree or as a reticulated network. The choice will be dictated by which aspects of the relations need to be emphasized for a given purpose. 

The main function of trees and reticulated networks in phylogenetics is to simplify complex network relations. Computers are capable of mapping networks of intricate complexity. But the networks they are able to produce are “often far too complex to be of much use as a representation of the data.” Trees and reticulated networks both facilitate simplified representations of genetic relations within data sets. (For examples, see the sections “Trees” and “Other Networks and Trees” in Morrison’s “Primer of Phylogenetic Networks.”)

If we apply these principles to genealogy in historical disciplines, we will recognize that, insofar as there are genetic relationships among historical phenomena, these relations are intractably complex. But instead of concluding with a demonstration of complexity, as do so many works of humanistic inquiry, we should recognize the utility of simplifying models and consider the affordances and limitations of tree and reticulated network analogies. 

Among these considerations should be whether it is more important to conceptualize continuity or change. Genealogical trees usually track the continuity of a single lineage, whereas reticulated networks track the merging of multiple lineages, a hybridization that can be continuous or disruptive, depending on the principles of legitimacy the interpreter applies. 

As Christiane Klapisch-Zuber and Andrea Worm have argued, the modern family tree presents genealogical data counterintuitively and simplistically. In a typical family tree, “I” am placed at the root, while my ancestors illogically branch out above me. Furthermore, the rootedness of the diagram belies the intricate reticulation of multiple lineages with each other. Yet high-medieval genealogical diagrams resisted this counterintuitive arrangement. 

In particular, “trees of consanguinity” were built to visualize both tree and reticulated network topology. Instructions accompanying these diagrams urged readers to imagine the nodes of the tree in motion and constantly generating new trees. In this way, trees of consanguinity converted genealogical data from simple to complex relations, much like the phylogenetic method for converting between a “parsimony tree” and a “median network” (see explanation of Fig. 7 in Morrison’s “Primer”).

An example of a “median network.”

The movement of a “median network” to a “parsimony tree.”

In subsequent posts, I will explore the affordances of trees of consanguinity for genealogical thinking.