The Quality of the Archaeological Record
by Charles Perreault
University of Chicago Press, 2019
Cloth: 978-0-226-63082-3 | Paper: 978-0-226-63096-0 | Electronic: 978-0-226-63101-1
DOI: 10.7208/chicago/9780226631011.001.0001


Paleobiology struggled for decades to influence our understanding of evolution and the history of life because it was stymied by a focus on microevolution and an incredibly patchy fossil record. But in the 1970s, the field took a radical turn, as paleobiologists began to investigate processes that could only be recognized in the fossil record across larger scales of time and space. That turn led to a new wave of macroevolutionary investigations, novel insights into the evolution of species, and a growing prominence for the field among the biological sciences.

In The Quality of the Archaeological Record, Charles Perreault shows that archaeology not only faces a parallel problem, but may also find a model in the rise of paleobiology for a shift in the science and theory of the field. To get there, he proposes a more macroscale approach to making sense of the archaeological record, an approach that reveals patterns and processes not visible within the span of a human lifetime, but rather across an observation window thousands of years long and thousands of kilometers wide. Just as with the fossil record, the archaeological record has the scope necessary to detect macroscale cultural phenomena because it can provide samples that are large enough to cancel out the noise generated by micro-scale events. By recalibrating their research to the quality of the archaeological record and developing a true macroarchaeology program, Perreault argues, archaeologists can finally unleash the full contributive value of their discipline.


Charles Perreault is assistant professor at the School of Human Evolution and Social Change at Arizona State University.


“In this cogently argued book Perreault exposes the shaky foundations of many if not most archaeological claims about the past and proposes a new research agenda that plays to the enormous potential of the archaeological record to reveal large-scale patterns in space and time.”
— Stephen Shennan, University College London Institute of Archaeology

“Archaeologists have long acknowledged that the archaeological record is not the same as an ethnographic record, yet they apply microscale, ethnographically based explanatory models and theories to the archaeological record. In firm but not shrill language, and by rigorous analyses of the influences of the generally low-scale resolution and dimensionality of the archaeological record, Perreault demonstrates the underdetermined nature (weaknesses) of much modern archaeological research and argues convincingly for ‘recalibrating’ archaeological methods and theories to the macroscale qualities of the archaeological record. This volume is among the top five must-read books to appear since the 1980s.”
— R. Lee Lyman, University of Missouri–Columbia

“This book moves the epistemology of the historical sciences in general and archaeology in particular to a new level of sophistication. Perreault’s grasp of ‘nature’s messy experiments’ and his analyses of theories of verification are simply brilliant. His call for ‘macroarchaeology,’ or the search for macroscale phenomena in the archaeological record, optimistically defines our discipline’s future. This book should be mandatory reading for any serious theorist of cultural evolution theory.”
— Charles Stanish, University of South Florida

“Perreault’s thesis is both apt and extremely important for archaeologists to address. Thoughtful engagement with it could transform the field, which would make this book the agent of transformation that it deserves to be and that archaeology needs.”
— Michael J. Shott, University of Akron

"Perreault's review is thorough and salutary, usefully setting out the empirical basis for what is often assumed. . . . A call to focus on the macroscale is welcome; analyzing change over long timescales is a contribution exclusive to archaeology."
— Susan Greaney, Times Higher Education

“For its critical analysis and corrective approach, this volume is to be commended… The Quality of the Archaeological Record presents a precise and critical commentary on archaeological interpretation that will be of value to students and professionals of archaeology alike.” 
— Canadian Journal of Archaeology



- Charles Perreault
DOI: 10.7208/chicago/9780226631011.003.0001
[Equifinality;Underdetermination;Hypothesis testing;Experimental sciences;Historical sciences;Simulations;Ethnographic analogies;Epistemology;Laboratory methods;False results]
The way archaeologists test their hypotheses undermines their capacity to make valid inferences about the human past. Historical sciences seek to explain contemporary traces in terms of past causes. Archaeologists settle on explanations when they are consistent with their data, irrespective of whether or not there are alternative explanations that are also consistent with the data at hand. The test of consistency leads to a confirmatory bias and leads to underdetermination and wrong results. In contrast, successful historical sciences work by formulating multiple mutually exclusive hypotheses and finding a smoking gun that will discriminate between them.Smoking guns must be found in nature: computer simulations, mathematical models, ethnographic analogies and experiments are not sources of smoking guns. Unlike experimental scientists, historical scientists such as archaeologists cannot use laboratory methods to manufacture new empirical evidences or to shield themselves from false positive or false negative results. This dependence on the quality of the archaeological record is not trivial. It means that the archaeological record dictates what can be learned about the past. Because of the quality of the archaeological record, there are research questions that we will never be able to answer beyond any reasonable doubts. (pages 1 - 22)
This chapter is available at:
    University of Chicago Press

- Charles Perreault
DOI: 10.7208/chicago/9780226631011.003.0002
[Equifinality;Scope;Sampling interval;Resolution;Data quality;Temporal scale;Spatial scale;Undetermination]
A theory that describes the pathways by which processes can be archaeologically equifinal is developed. There are four aspects of the quality of any set of empirical data: scope, sampling interval, resolution, and dimensionality. The scope refers to the total amount of space and time that is represented in a data set. The sampling interval denotes the interval of time or space that separates the analytical units. The resolution is the amount of space and time that are represented within each unit. And dimensionality describes the independent variables of an object of study that have been measured. Each one of these four aspects of the quality of a dataset can lead to equifinality. Reducing equifinality can be done by improving the quality of the data. But it is always possible that the smoking gun that would resolve a question has been destroyed and is forever lost. When this happens, there is only one thing left to do: to abandon the study of those processes that are equifinal and focus instead on the those that operate over temporal and spatial scales that are similar to that of archaeological data, and where the complexity is matched by the data’s dimensionality. (pages 23 - 39)
This chapter is available at:
    University of Chicago Press

- Charles Perreault
DOI: 10.7208/chicago/9780226631011.003.0003
[Resolution;Palimpsests;Analytical lumping;Time-averaging;Diversity;Richness;Rates of change;Deposition;Sedimentation;Disturbance;Site-formation processes]
To identify what processes can be studied we need to understand the forces that shape the quality of the archaeological record. Here are examined the forces that mix archaeological material–i.e. that lump in the same analytical unit material associated with activities that took place at different points in time and in space. They include site formation processes such as cultural and natural deposition processes (e.g. the discarding of objects by ancient people, site reoccupation, reuse, sedimentation, erosion); disturbance processes (e.g. plowing, trampling, moving water, burrowing animals) and analytical processes (the lumping of remains due to imprecision in dating techniques, or by archaeologists during excavation or the construction of cultural time periods). The effect of these forces is also discussed: mixing results in palimpsests and inflates the size and the composition of assemblages; skews relative frequencies, inflates measures of diversity, richness and variance, confounds association and correlations between types of objects; and reduces apparent rates of change. Overall, the forces of mixing decrease the resolution of archaeological data, leading to space- and time-averaging. By destroying existing patterns and creating new ones, they impact every aspect of the archaeological record from which archaeologists draw inferences. (pages 40 - 79)
This chapter is available at:
    University of Chicago Press

DOI: 10.7208/chicago/9780226631011.003.0004
[Information loss;Preservation;Deterioration;Sampling intervals;Site frequency;Sedimentation;Surface cover;Temporal range;Rates of change;Pull of the recent]
I examine the forces that lead to the loss of information and archaeological material.The forces of loss include preservation loss and observation loss. Preservation loss describes the physical remains that did not preserve or that have been damaged to such an extent that the information-bearing traces have been obliterated. Observation loss refers to physical remains that are preserved in the archaeological record but that have not been discovered or noticed by archaeologists. The cultural practices of ancient people, deterioration, decay, sedimentation, surface cover and field excavation techniques can all lead to both types of loss. Loss impacts the archaeological record in many ways. It creates a “Pull of the Recent” by preferentially decreasing the number of old archaeological sites, affect the size, and the composition of assemblages. It increases the sampling interval of the archaeological record to orders of magnitude ranging from 100to 103years, leads us to underestimate the temporal range of cultural traits, can make sudden cultural change appear gradual, slows down apparent rates of change, and limits our capacity to control for covariates. (pages 80 - 111)
This chapter is available at:
    University of Chicago Press

DOI: 10.7208/chicago/9780226631011.003.0005
[Regional databases;Temporal sampling intervals;Spatial sampling intervals;Spatial gap;Resolution;Time;Data quality]
The sampling interval and the resolution of the archaeological record in peer-reviewed journal articles and regional databases is measured. These sources complement each other: journal articles are a good proxy for the quality of the data used by archaeologists, whereas regional databases are representative of the quality of the archaeological record itself. I show that archaeologists have to contend with units that are, on average, separated by hundreds of years and kilometers. The vast majority of the sampling interval and the resolution observed in journal articles and regional databases are greater than one human generation, and usually in the order of 102and 103years. Sampling interval and resolution of an order of magnitude of 101years do exist, but they are the exception, not the rule. Spatially, the gap between archaeological sites is in the order of 101to 102kilometers. I also explore the relationship between the age of archaeological deposits and the quality of the record. While archaeologists already know that the quality of the record degrades with time, my analyses allow us, for the first time, to move beyond an intuitive understanding of this principle and to describe it quantitatively. (pages 112 - 134)
This chapter is available at:
    University of Chicago Press

DOI: 10.7208/chicago/9780226631011.003.0006
[History of archaeology;Anthropological archaeology;Uniformitarianism;History of paleontology;Research program;Equifinality;Archaeological theories;Individuals;Agency]
This chapter examines whether the current research program of the discipline matches the quality of the archaeological record and argues that most processes studied by archaeologists operate over a decade or less. This is 2–3 orders of magnitude faster than the sampling interval and the resolution of archaeological data. This has three consequences. First, most archaeological results are wrong. The chance that an archaeological interpretation, picked among dozens of equifinal alternatives, is valid is vanishingly small. Second, most archaeological research is also unneeded. The short-scale processes studied by archaeologists are borrowed from other disciplines, such as cultural anthropology. These disciplines do not need archaeology to confirm or disprove their ideas. Third, archaeological theory is balkanized. The archaeological literature is crowded with a daunting number of theories and claims that are mutually exclusive. New theories and processes are added to the literature faster than old ones are eliminated. Archaeologists are ignoring the equifinality problem for historical reasons that are outlined here. This was further amplified by the way archaeologists understood uniformitarianism, a human-centric view of the world, and the way archaeologists test hypothesis. Paleontologists, faced a similar problem years ago and solved it by changing their research problem. (pages 135 - 160)
This chapter is available at:
    University of Chicago Press

DOI: 10.7208/chicago/9780226631011.003.0007
[Research program;Cultural history;Macroarchaeology;Biogeography;Climate;Low-pass filter;Noise cancelation;Global database]
A research program that takes full advantage of the quality of the archaeological record eliminates the study of most microscale processes — those that are observed within a human lifetime and that operate at the hierarchical scale of the individual — because the archaeological record is not a suitable medium to study them. Instead, an appropriate research program focuses on: (1) cultural history and (2) macroarchaeology, the search for macroscale patterns and processes in the global archaeological record. Archaeologists can also make unique contributions to the social sciences by studying macroscale processes that operate at a hierarchical level well above that of the individual, that cannot be seen within the span of a human lifetime, but that become visible only when looked from an observation window thousands of years long and thousands of kilometers wide. The archaeological record has the scope necessary to detect macroscale phenomena because it can provide samples that are large enough to cancel out the noise generated by microscale processes. In order to discover macroscale principles affecting human history, archaeologists need to build a global database of archaeological types. Such database can act as a low-pass filter that cancels out the noise generated by microscale factors. (pages 161 - 189)
This chapter is available at:
    University of Chicago Press

- Charles Perreault
DOI: 10.7208/chicago/9780226631011.003.0008
[Cultural anthropology;Programmatic agenda;Macroarchaeology;Macroscale patterns;Macroscale processes]
For years, archaeologists have used the archaeological record as if it was a window on the past. As if they could look through it, like an observer behind a one-way mirror, and study past human societies the way cultural anthropologist would do. In doing so, archaeologists have uncritically borrowed a programmatic agenda that was designed by, and for, researchers who study humans in the present-time and using data that has a scope, a sampling interval, a resolution, and a dimensionality that is orders of magnitude different than what archaeologists have access to.Unlike the study of cultural history, macroarchaeology, that is, the search for mega-scale patterns and processes in the archaeological record, is uncharted territory. The current theories of human culture have virtually nothing to say about what mega-trends could exist in the archaeological record, or about what mega-scale drivers, such as climate or biogeography, might have shaped the course of human history. (pages 190 - 194)
This chapter is available at:
    University of Chicago Press

Appendix A. A Formal Model of the Effect of Mixing on Variance

Appendix B. Source of Time Intervals and Time Resolutions from Journal Articles