Natural Experiments of History

Jared Diamond,James A. Robinson

• Publisher: Belknap Press
• Year: 2012
• ISBN:
  9780674076723
• Topic: History
• Subtopic: Historiography
• Index: History

Some central questions in the natural and social sciences can't be answered by controlled laboratory experiments, often considered to be the hallmark of the scientific method. This impossibility holds for any science concerned with the past. In addition, many manipulative experiments, while possible, would be considered immoral or illegal. One has to devise other methods of observing, describing, and explaining the world.In the historical disciplines, a fruitful approach has been to use natural experiments or the comparative method. This book consists of eight comparative studies drawn from history, archeology, economics, economic history, geography, and political science. The studies cover a spectrum of approaches, ranging from a non-quantitative narrative style in the early chapters to quantitative statistical analyses in the later chapters. The studies range from a simple two-way comparison of Haiti and the Dominican Republic, which share the island of Hispaniola, to comparisons of 81 Pacific islands and 233 areas of India. The societies discussed are contemporary ones, literate societies of recent centuries, and non-literate past societies. Geographically, they include the United States, Mexico, Brazil, western Europe, tropical Africa, India, Siberia, Australia, New Zealand, and other Pacific islands.In an Afterword, the editors discuss how to cope with methodological problems common to these and other natural experiments of history.

1

Controlled Comparison and Polynesian Cultural Evolution

PATRICK V. KIRCH

In early January of 1778, Captain James Cook, in command of HMS Resolution and Discovery, was sailing through uncharted waters in the central North Pacific Ocean, en route to the coast of New Albion, as the Pacific Northwest was then called. The Admiralty had instructed Cook to replenish at Tahiti, an island he already knew well from two previous voyages, then to go northward in search of the fabled “northwest passage.” On January 18, the Resolution’s lookout spied a high island to the northeast; a second volcanic peak was soon discerned to the north. The following day Cook and his crew made “first contact” with one of the most isolated societies on earth—the Polynesian inhabitants of Kaua‘i, one of the Hawaiian Islands.

Cook was no stranger to Polynesia. He had first gone to Tahiti a decade earlier, at the behest of the Royal Society of London, to observe the June 3, 1769, transit of Venus across the sun. That mission accomplished, Cook extended his explorations to other islands of the Society archipelago, followed by an unprecedented circumnavigation of New Zealand. In 1772 the Admiralty dispatched him again to the Pacific, to determine whether or not the long-hypothesized continent of Terra Australis actually existed. In addition to taking his ships farther south than any man had gone before, Cook explored and mapped more of Polynesia, including the Tuamotu Islands, Tonga, the southern Cook Islands, Easter Island, and the Marquesas.

After a decade of sailing throughout the central Pacific, mapping the islands, and observing their inhabitants, Cook had acquired considerable knowledge and insight into the peoples we now group together under the rubric “Polynesian.”¹ The first thing to catch his attention when the Kaua‘i islanders’ canoes came alongside the Resolution was that their speech was clearly a variant of the language spoken in Tahiti, more than 2,700 miles to the south. On the eve of his departure from Kaua‘i to continue his voyage on to New Albion, Cook penned these words in his log: “How shall we account for this Nation spreading it self so far over this Vast ocean?”² He was astounded that people speaking clearly related languages, and by inference sharing a common origin in the not-too-distant past, were distributed from New Zealand to Easter Island and now to his newfound archipelago in the North Pacific. Geographically, Cook calculated, this “Nation” was spread over “an extent of 60° of latitude or twelve hundred leagues north and south and 83° of longitude or sixteen hundred and sixty leagues east and west.” Cook, one of the great explorers of the Enlightenment, was confronting a great puzzle of human history. The question of Polynesian origins, and the history of their subsequent dispersal and cultural differentiation, are problems that have ultimately yielded to the methods of controlled comparison.

The perspective I bring to this volume on the use of comparison in historical studies is that of an anthropologist who has spent several decades studying the ancient societies and cultures of Polynesia—those myriad islands and archipelagoes lying within the vast triangle subtended by New Zealand, Hawai‘i, and Rapa Nui (Easter Island). As Cook discovered, Polynesia is united by a common linguistic heritage. Archaeology has subsequently shown that Polynesia comprises a historically coherent cultural region because its varied cultures all share many features owing to a common origin in the first millennium B.C. For this reason, Polynesia has more than once been regarded as an ideal region in which to undertake comparative analysis. A number of classic works in anthropology applied such a comparative approach, including Marshall Sahlins’s study of the differentiation of Polynesian social structures in relation to environmental differences on islands and Irving Goldman’s analysis of “status rivalry” as a key to understanding differences in Polynesian cultures.³ In the realm of material culture, differences in Polynesian sailing canoes, barkcloth manufacture, and stone adze technology have similarly been subject to comparative study.⁴ Douglas Oliver ventured well beyond Polynesia to incorporate Melanesian, Micronesian, and Australian cultures into his magnum opus on Oceania.⁵ Historical linguists, for their part, have used their own specialized methods of phonological and lexical comparison to reconstruct much of the Proto Polynesian vocabulary.⁶

My own interest in Polynesia derives from my primary academic specialization of prehistoric archaeology (or “anthropological archaeology” as many label our field, in part to distinguish it from “classical archaeology,” which focuses on the Greco-Roman world). But while I have invested much energy excavating the detailed material evidence by which we can date and define the outlines of Polynesian history before the arrival of Europeans and advent of historical documents, I regard such fieldwork as just part of a larger process of historical research and understanding. This is because I firmly believe that the comparative analysis of multiple prehistories can tell us something more profound about human cultures and their longterm development. Over the years I have thus come to regard myself as an “historical anthropologist” and have increasingly drawn on a range of multidisciplinary evidence that includes not only archaeological finds, but historical linguistic data, comparative ethnographic studies, and paleoecological and paleoenvironmental research.

I should disclose another facet of my epistemological foundation, which is that I regard historical anthropology as a “historical science,” in the sense that Stephen Jay Gould and Ernst Mayr wrote of historical as opposed to experimental sciences.⁷ (Hence I reject a postmodern perspective that regards all constructed “stories” of the past as equally valid.) Indeed, I regard archaeology’s role in the science of human history (or “cultural evolution”) as analogous to paleontology’s role in the science of biological evolution. Both fields uncover the physical evidence of long-term change, cultural on the one hand (artifacts and the detritus of human occupation), biological on the other (bones, exoskeletons, and other fossils). But we can make sense of this evidence only when it is incorporated into a broader paradigm. Much work is now in progress to provide such a paradigm for cultural evolution, a review of which would go well beyond the scope of my essay.⁸

But to return to the concept of comparison, it should be noted that this idea is essential to any historical science, including historical anthropology, because we cannot run “experiments” on cultural evolution or long-term change in human cultures and societies. As Mayr astutely points out, the historical (or “observational”) sciences discovered an alternative to experimentation by seeking “natural experiments.” No natural experiment is more famous than Darwin’s finches in the Galapagos Islands, which provided him with crucial evidence for the theory of evolution. As Mayr has stated: “Much progress in the observational sciences is due to the genius of those who have discovered, critically evaluated, and compared such natural experiments in fields where a laboratory experiment is highly impractical, if not impossible.”⁹ Not surprisingly, perhaps, many of the most famous natural experiments involve islands and archipelagoes.

Polynesia offers just such a set of natural—or in this case cultural—experiments for understanding fundamental processes of historical change on the timescale of one to three millennia. Several factors make the Polynesian islands and their societies a nearly ideal region for comparative historical analysis. First, the islands themselves vary in ways that posed significant adaptive challenges for their human colonizers. They range in size from diminutive (a few square kilometers) to near continental (New Zealand), in form from coral atolls to volcanic high islands of varied geologic ages, and in other aspects of climate and marine and terrestrial resources. Second, all of these islands were discovered and populated by people whose origins can be traced back to the same founding group of Eastern Lapita voyagers who arrived in the Tonga-Samoa region around 900 B.C.¹⁰ Thus the later descendant societies can be compared in terms of those aspects of their cultures that are retentions from the ancestral group, as opposed to aspects that are innovations, or derived traits. And third, the Polynesian societies witnessed by Cook and other Enlightenment explorers at the end of the eighteenth century displayed a remarkable range of variation in their degrees of sociopolitical and economic complexity, from simple chiefdoms in which there was little status differentiation to large polities incorporating tens of thousands into highly structured and hierarchical social formations. Thus Polynesia presents a remarkable opportunity to carry out comparative analysis of social and cultural change within a group of historically related peoples.

To point out that Polynesia presents an ideal region for comparative analysis is one thing; to develop a methodologically rigorous approach to comparison is another. For a start, such an approach must be capable of discriminating between cultural traits that are shared homologies, those that are derived or innovated (analogies), and those that were borrowed (synologies).¹¹ My colleague Roger Green and I have developed such a carefully structured method of comparative historical analysis, which, following an original proposal by anthropologist Evon Vogt, we call the “phylogenetic model.” The full exposition of the phylogenetic model, and its essential corollary, the “triangulation approach,” is presented elsewhere.¹² Here I briefly summarize the key elements of our approach, which is essential to the comparative analysis presented in the second part of this chapter.

The phylogenetic model is based on the recognition, first made by Kim Romney with respect to the Uto-Aztecan cultures of the New World, that in many parts of the world groups of related cultures (and frequently this relationship is most clearly indicated by the fact that they all belong to a single linguistic family) share a common history or “phylogeny.” In other words, the similarities shared by such cultures are homologous. Peter Bellwood has recently argued that rapid population or “demic” expansions of agricultural groups in various regions during the mid- to late Holocene has led to such a pattern of historically related language–culture groups covering significant parts of the world.¹³ Examples include the Bantu-speaking peoples of sub-Saharan Africa, the Uto-Aztecan peoples of Mesoamerica and western North America, and the speakers of the extensive Sino-Tibetan, Austro-Asiatic, and Austronesian language families of East and Southeast Asia. Polynesia, which is one branch of the larger Austronesian expansion, is therefore just one of many instances where a phylogenetic model may be fruitfully applied in historical comparative analysis. Because of its discrete island geography, however, with the implication of decreased contact and relative isolation after initial dispersal and settlement, the Polynesian case is ideal for working out methodological principles of a phylogenetic approach to cultural history.

The phylogenetic model uses a sequence of methodological steps to work out the specific history of cultural evolution and differentiation within a group of related cultures (what Romney called a “segment of cultural history”). Beginning with the mapping of the geographical distribution of such a group—for which one has hypothesized such a homologous history—the key first step is to apply the methods of historical linguistic analysis to the set of languages spoken by the members of these cultures, in order to derive a “family tree” or phylogeny of the historical relationships. Although Vogt originally advocated the use of lexicostatistics and glottochronology,¹⁴ such “phenetic” methods are often not capable of revealing the true phylogenetic relationships between the languages in question, and it is thus preferable to apply the traditional “genetic comparative approach” in historical linguistics. This classic comparative method results in a “family tree”–type model of language differentiation.¹⁵ Such a tree or phylogeny provides a model for historical relationships and for the branching or splitting process of linguistic (and related cultural) differentiation over time. The methods of lexical and semantic reconstruction can also be applied once this phylogeny has been developed, in order to reconstruct in some detail the protolanguage and protoculture of the original founding group (in this case, Proto Polynesian language and Ancestral Polynesian culture), as a baseline from which later change and divergence took place.

The phylogenic tree resulting from such historical linguistic analysis should of course be considered a model (a complex set of interrelated hypotheses) subject to cross-checking on independent evidence. Such cross-checking can be done by turning to the data provided by archaeology. Does the archaeological record of material culture correspond to the branching pattern suggested by the linguistic record? For example, are sequences of changes in Polynesian pottery, stone adzes, and fishhook styles over time consistent with a model of cultural differentiation such as that depicted in a family tree of the Polynesian languages? In the Polynesian case the correspondence is very good, lending additional confidence in the phylogeny thus proposed. Moreover, archaeology has the capacity to directly date (through radiocarbon and other dating methods) sets of archaeological assemblages that can be correlated with specific branches and protolanguage stages on the linguistic model. Thus archaeology allows us not only to independently test the linguistic model of cultural ...