Life histories are complex, multidimensional entities composed of a series of ongoing, often interrelated contexts, activities, and relationships. Each life domain contributes to the context within which events in other domains are shaped, and is in turn shaped by activities or events in other domains. Capturing these two key aspects of life histories – simultaneous dependence, and multidimensionality – is an important challenge for gerontology and life course research. In this paper we introduce a graph theoretic technique for exploratory analysis of life history data that speaks to this challenge. Using the notion of life history graphs, we represent a single life history as a series of spells (activities with specified start and end points) linked by their patterns of temporal overlap. Given multiple individuals, their associated life history graphs can be directly compared using distance-based methods that assess differences in the topologies of their respective spell structures while preserving qualitative differences between life domains (e.g., non-exchangeability of job and marital spells). Predicated on overlaps across life domains, our approach is designed to incorporate both the multidimensional structure of detailed life history data and the importance of simultaneity. Given a sample of individuals from a target population, our techniques can also be used to identify and test hypotheses regarding the factors that differentiate life histories within the population in question. Using data from the United States, England, China, and Vietnam, we demonstrate how our methods allow us to make comparisons both within and across populations.