Representing Lineage as a Graph: Modeling Complex Kinship
Abstract
The tree is an intuitive representation of lineage, yet it falls short. This paper argues that kinship is modeled more accurately as a graph than as a tree, drawing on work in mathematical anthropology and kinship network analysis [1][2][5], and on the practices of modern graph databases [4]. We present the problem at an expanded conceptual level and note that the Nasayeb platform adopts this approach at a high level, without detailing the implementation.
1. The Common Metaphor and Its Limits
A family tree is usually presented as an actual tree: a root, branches that spread in a single direction, and every node with one parent and a single path back to the root. This metaphor is elegant and easy, but it imposes constraints that do not hold up against real families. A tree, by its mathematical definition, is a structure with no cycles and a single parent per node; both of these conditions contradict the reality of kinship in many cases.
2. Where Does the Tree Break Down?
• Marriage between relatives: when two individuals descended from a common ancestor marry, a "cycle" appears in the lineage diagram, because two different paths lead to the same person. A tree does not allow this. • Multiple marriages: children are distributed across more than one family unit, which is hard to represent as a clean, single branch. • Repeated appearances: an individual may appear in two branches through marriage or adoption, so they are either duplicated (creating contradictions) or lost. • Non-parental relationships: guardianship, adoption, and social parenthood find no place in the "single parent" model.
Mathematical anthropology has shown for decades that these are not rare exceptions to be set aside, but a structural feature of human kinship, and that representing it correctly calls for graphs and groups [1][7]. Network analysis studies have likewise developed quantitative tools for studying kinship networks as graphs [5].
When we force the tree onto kinship, we erase from reality whatever does not fit the mold.
3. The Graph Model
The alternative is to stop thinking in terms of the tree and think instead in terms of the graph: each individual is a single node, no matter how many relationships they have. Nodes are connected by edges labeled with the type of kinship: paternity, maternity, marriage, and others. This representation allows cycles and multiple relationships without duplicating a node, so each person remains a single entity connected to several branches.
This conception is not merely theoretical; it is close to the models of modern graph databases, which store nodes and edges directly and optimize queries about paths and relationships [4]. It also intersects with the kinship representations used in historical demographic studies that reconstructed kinship networks from records [3][6].
4. What Does This Mean for the Family?
The practical outcome is that each individual is stored once, with their data, photos, and stories, even if they appear in more than one place on the graph. There are no duplicated records that contradict one another, and no people lost among the branches. When a user searches for a relative, they find a single, complete entity rather than scattered copies. Computing kinship relations (who is whose cousin, and to what degree) also becomes precise and automatable, because the structure reflects reality rather than a simplified mold.
5. Contribution and Application (Advanced Level)
The Nasayeb platform represents the family as a kinship graph that stores each individual once and accommodates complex relationships. The engineering details of the storage, indexing, and inference layer are part of our proprietary work, so we confine ourselves to the principle: accuracy begins with correct representation. The contribution here is a commitment to a structure that reflects the reality of kinship as it is, not as it is easy to draw.
6. Conclusion
The tree is a comfortable metaphor, but a narrow one. Human kinship, with its marriages between relatives, its multiple households, and its intertwined in-law ties, is a graph by nature. Moving from the tree to the graph is not a technical luxury, but a condition for accuracy, reliability, and preserving the whole truth for future generations.
References
- White, D. R., & Jorion, P. "Representing and Computing Kinship: A Graph Approach." Current Anthropology, 1992.
- Read, D. "Kinship Algebra Expert System (KAES)." Structure and Dynamics, 2006.
- Bouchard, G. "Population Databases and Genealogical Reconstruction." Historical Methods, 1992.
- Robinson, I., Webber, J., & Eifrem, E. Graph Databases. O'Reilly, 2015.
- Hamberger, K., Houseman, M., & White, D. R. "Kinship Network Analysis." The SAGE Handbook of Social Network Analysis, 2011.
- Wetherell, C. "Historical Social Network Analysis." International Review of Social History, 1998.
- Ballonoff, P. A. (ed.). Mathematical Models of Social and Cognitive Structures. University of Illinois Press, 1974.