Observations on Graphing Prehistory

Despite years of graphing culture change using different types and styles of diagram, there is minimal discussion of graph grammar—how to construct an effective and efficient graph, and how to decipher a graph of change. Part of the difficulty attending graph decipherment resided in (and continues to reside in) unclear distinction of transformational change from variational change. Models reflecting the former tend to be commonsensical and are similar to Petrie’s classic sequence dating graphs. The difficulty of graph decipherment is exacerbated by parsing temporally continuous variation into discontinuous spatio-temporally bounded units known as artifact types, cultures, phases, periods, stages, etc. These units are reified and (implicitly) conceived as real entities to be discovered for want of a well-developed theory of change and an attendant ontology of how continuously variable phenomena should be parsed into types for analysis. Archaeologists did perfect models of diffusion—the movement of culture traits (ideas or norms manifest as artifact types) across space over time—and built models of how it should be reflected in the archaeological record. A majority of introductory archaeology textbooks published since 1965 typically present graphs of culture change in the form of a spindle graph but with minimal discussion of graph grammar. Texts on regional or continental prehistory typically summarize culture change in spatio-temporal rectangle diagrams, which for pedagogical reasons may be reasonable. A few spindle graphs have been published in other disciplines and, like archaeological spindles, display temporally shifting frequencies.

Paleontological, Palynological, Biological, and Physical Anthropology Graphs

The earliest paleontological spindle graphs appear in the 1830s and 1840s, and are of a different style and diagram different kinds of data (absolute frequencies of taxa or kinds) than the earliest archaeological spindle graphs. Palynologists regularly produce so-called pollen diagrams, left-justified spindle graphs, that display temporally varying frequencies of pollen of each of several different plant species. These first appeared in the 1910s, and are of a different style than early twentieth-century archaeological spindle graphs, although the data graphed by the two are similar (relative frequencies of specimens of each of several kinds of phenomena). Biologists used spindle graphs during the early twentieth century to plot both the phylogenetic history of taxonomic families and orders, and frequencies of individual organisms representing different plant and animal species found in different habitat types. Differences in styles of biological spindle graphs and early archaeological spindles suggest the idea for the latter was not in the biological sciences. Physical anthropologists “seriated” biometric data, but early twentieth-century textbooks do not include spindle graphs.

Shopping Around and One-Off Graphs

Given explicit recognition between ~1915 and the 1930s that certain artifact types display unimodal frequency distributions over time, archaeologists initially presented tables of those frequencies but by the 1930s were experimenting with different types of graphs to present visual images of culture change. The lack of familiarity with graph theory and graph grammar meant numerous kinds of graph were published, often only once each as researchers sought effective (readily deciphered) and efficient (minimal ink and space) graph forms. These experimental graph types range from fairly simplistic to complex and virtually indecipherable. Lack of decipherability and errors in some graphs reflect poor understanding of the principles of graph construction and the precise nature of what a graph type is meant to illustrate. The analytical focus on culture history and recognition that artifact form varied along both time and geographic space led to some efforts to incorporate all three dimensions—form, time, space—into some graphs. It is not surprising that in the search for a useful graph type, the one-off graphs variously implied transformational, variational, and a combination of variational and transformational evolutionary change.

Introduction

Graphs of quantitative data are analytical tools that facilitate visual thinking. In many disciplines, the use of graphs was preceded by tables summarizing quantitative data. Graphs known by North American archaeologists as “battleship curves” are temporal frequency distributions of relative abundances of specimens in each of several artifact types. They are unimodal frequency distributions known as spindle graphs. In the early 1950s, it was suggested that the idea of spindle graphs was borrowed by archaeologists from paleontology. Archaeologists occasionally used bar graphs and line graphs to diagram change in artifact inventories in the early twentieth century. The questions addressed in this volume are: (i) did North American archaeologists borrow the idea of spindle graphs from paleontology, and (ii) what was the frequency of use by North American archaeologists of each of the various graph types to diagram culture change during the early and middle twentieth century?

Early History of Archaeology Graphs

The earliest archaeological spindle graph was published in 1883 by natural historian and avocational archaeologist Charles C. Abbott. Evidence that he obtained the idea from paleontology, which first published spindle graphs in the 1830s and 1840s, is circumstantial at best, and differences in graph styles weigh against such borrowing. Several spindle graphs published in the 1890s and early 1900s by archaeologist William Henry Holmes either depict his views on inevitable progressive evolution—a theory rapidly falling from anthropological favor—or were so speculative as to likely have had little influence on the discipline. During the first couple decades of the twentieth century, physicist/geographer/anthropologist Franz Boas (often referred to as the father of anthropology) published numerous line graphs of quantitative data. He influenced archaeologists Leslie Spier and Manual Gamio who used line graphs to display temporally varying frequencies of artifacts. About the same time, the wife and husband team of Madeleine Kidder and Alfred V. Kidder published several line graphs of relative frequencies of pottery types against stratigraphic provenience, seemingly largely as a result of Madeleine’s influence because Alfred never again published such a graph and instead favored phyletic seriation graphs of a type reminiscent of Sir William Flinders Petrie’s sequence dating graphs from the turn of the century.

Materials and Methods

To determine the origin of archaeological spindle graphs, and to track the frequency of use of each of several types of graph used to diagram culture change, a sample of North American archaeological literature was examined. Numerous series of monographs and volumes of journals in both the archaeological and the paleontological literature were inspected. If a graph of biological (paleontological) or cultural (archaeological) change was included in a publication, that piece of literature was recorded along with the type of graph included. To record such data, a classification of graph types was developed based on categories of statistical graphs (e.g., bar graph, line graph, pie graph, time range, spatio-temporal rectangle). More than 900 pieces of literature on North American archaeology published between ~1880 and ~1960 were inspected, and more than 450 pieces of literature on paleontology were inspected. Because different graph types are constructed under different guidelines, they require an understanding of graph grammar—the rules for constructing, deciphering, and interpreting graphs.

Culture Chronology and Change

Archaeology emerged as part of the general discipline of anthropology in North America, the overall focus of which for the first five or six decades of the twentieth century was to write the history of the culture of each group of native North American people. The goal of writing a culture’s history could only be accomplished by placing artifacts in a chronological sequence, which demanded a chronometer. It was not always possible to refer to stratigraphic superposition, so various techniques of seriation—arranging artifacts based on their formal attributes in what was believed to be a chronological order—were invented and used. The results of the seriation techniques and stratigraphic superposition studies were initially summarized in tables of artifact frequencies but eventually were graphed in several ways. Interest in culture chronology and change among North American archaeologists has extended throughout the twentieth century and into the twenty-first century.

An Evaluative Framework

North American anthropologists and archaeologists have long confused the Midas-touch-like transformational evolution of Lewis Henry Morgan, Edward B. Tylor, and Herbert Spencer with the variational evolution of Charles Darwin. Following Franz Boas, evolution as a theory of change was allegedly discarded by North American anthropologists and archaeologists at the beginning of the twentieth century. However, they used the term “development” instead of “evolution” and spoke of culture change in evolutionary terms, often mixing elements of the theories and ontologies of transformational and variational evolution. Documenting culture change under variational evolution demands the materialist paradox be circumvented. This paradox highlights the question: How do we measure change in continuously variable phenomena? Paleontologists adopted the approach that each population of organisms is polymorphic; individuals are members of the same species but formally variable. Paleontologists compare central tendencies of temporally sequent populations, or chronospecies. Archaeologists who undertook frequency seriations adopted an approach that focuses on morphospecies, forms or types that occur in two or more temporally sequent populations. The occurrence of multiple types per temporal period highlights the variation upon which a sorting mechanism such as selection works, and the occurrence of one or more types in each of two or more temporally sequent assemblages provides evidence of connection between them required of studies of change. Recognizing that graph types, how phenomena are parsed into types, and theories of change are mutually influential allows evaluation of archaeological graphs of change in terms of their implied theories and ontologies.

The Fourth (and Influential) Introduction of Spindle Graphs

Close examination of James A. Ford’s self-reported 1952 history of how he developed the centered and stacked bars style of spindle graph for which he is famous indicates he likely invented this kind of spindle graph with a bit of assistance from his colleagues George Quimby and Gordon Willey. In the 1930s, his diagrams of culture change were spatio-temporal rectangles or bar graphs; his first centered and stacked bars spindle diagram appeared in the 1949 published version of his doctoral dissertation. That graph style was picked up by American Southwest archaeologist Paul S(ydney) Martin that same year; Martin had, like many of his colleagues, initially used line graphs and bar graphs to illustrate culture change. Subsequently, numerous individuals adopted Ford’s centered and stacked bars form of spindle diagram. During the 1950s in Europe, French Paleolithic archaeologist François Bordes adopted ogive or cumulative relative frequency curves as a graphic means to compare assemblages of lithic tools. Quickly adopted by many European archaeologists, this graph type was only occasionally used in North America. After Ford, most graphs diagramed variational evolutionary change.

Archaeology Spindle Graphs in the 1930s

During the 1930s, archaeological spindle graphs in the form of seriograms (straight-sided spindle graphs) were published. Three of these represent the investigator’s suspicions about culture change rather than being strictly empirical. Stylistically, seriograms were seldom subsequently published, suggesting these graphs minimally influenced later researchers. By the 1920s, based on the basically unimodal frequency distributions observed in frequencies of specimens of various pottery types in the American Southwest, anthropologists had begun to suspect there were so-called stylistic pulses reflecting the vogue or popularity of particular kinds of artifacts. Explanations fell back on probability theory, likely as a result of the influence of Franz Boas’s statistical reasoning; kinds of phenomena simply should display unimodal temporal frequency distributions given probability theory. Although conceptually unsophisticated, graphic models of these stylistic pulses published by anthropologists in the 1920s took the rough form of spindle graphs and represent a then unrecognized nod to the theory of variational evolution. These spindle graph models may be the ultimate source of archaeological spindle graphs, but these models were a bit difficult to decipher. Many graphs of culture change appearing in the 1920s and 1930s imply variational evolution.