EH.net is owned and operated by the Economic History Association with the support of other sponsoring organizations.
Comparative Historical Analysis in the Social Sciences
Published by EH.NET (March 2004)
James Mahoney and Dietrich Rueschemeyer, editors, Comparative Historical Analysis in the Social Sciences. Cambridge: Cambridge University Press, 2003. xix + 445 pp. $25 (paperback), ISBN: 0-521-016445-2.
Reviewed for EH.NET by Bertrand M. Roehner, University of Paris 7.
"The editors of this book have assembled an all-star cast of comparative-historical sociologists and political scientists" notes Edgar Kiser in an advance praise excerpt at the back of the book. Additional advance praise comes from Charles Tilly. In the concluding chapter, Theda Skocpol writes that she is glad indeed to add her own contribution to the "central insights proffered in this remarkable collection." One may wonder as a result what, except for providing an additional rubberstamp, a reviewer can say of a book which has received the endorsements of such distinguished scholars. The reviewer's task is all the more difficult because the book comprises twelve essays, each of which would deserve a review in its own right. As this will, of course, not be possible within the scope of this review, my first plea is to ask for the indulgence and understanding of the contributors. I must confess that the points on which I will concentrate my attention are chosen on a subjective basis. At the end of the review, I offer a suggestion which would permit replacing subjective assessments with more objective judgments.
The review is organized as follows. First I explain why the discussion of comparative analysis is also of interest to economic historians. In the second section, I briefly summarize the book's content. Then I discuss in more detail some of the proposals put forward in the papers. Finally, I try to take a broader view of the relationship between comparative studies and real-world transformations.
Relevance to Economic History
This book is about revolutions, political institutions, democracy and similar topics which have little to do with economic history. Nevertheless, I believe that it can be of great importance to economic historians. Why? To explain that point, please allow me a personal reminiscence. When I became a member of the Cliometric Society and for the first time took part in a Clio Conference at Bloomington back in 1991, I had high expectations. However, I soon discovered that comparative analysis played but a limited role in Cliometrics. It took me somewhat longer to understand why. If one admits that economic theory is able to explain each and every economic phenomenon, then the only thing which remains to be done is to illustrate the general theory through specific episodes about one industry in a given country and time period (e.g. the banking industry in the United States during the New Deal). However, if economic theory really were so highly successful and satisfactory, one would expect it to provide more accurate forecasts than it usually does. If, on the contrary, we assume that economic theory needs to be improved in some essential ways, then, comparative analysis may be the perfect tool for such a job. If so, the book under review should be a useful source of inspiration for economists who wish to use comparative analysis is an effective and systematic way.
Below I have listed, for each of the twelve essays, the name of the author(s), keywords of the title, the number of pages and extent of the reference section. To those who are familiar with this field the names of the contributors will already give a good idea of the book's content.
(1) Mahoney and Rueschemeyer, Introduction, 36 pages; references 12 pages.
(2) Goldstone, Revolutions, 50 pages; references 6 pages.
(3) Amenta, Social policy, 41 pages; references: 7 pages.
(4) Mahoney, Democracy, 44 pages; references 6 pages.
(5) Pierson, Long-term factors, 31 pages; references 4 pages.
(6) Thelen, Political institutions, 33 pages; references 5 pages.
(7) Gould, Networks, 29 pages; references 2 pages.
(8) Katznelson, Periodization, 32 pages; references 4 pages.
(9) Rueschemeyer, Single cases, 32 pages; references 3 pages.
(10) Mahoney, Deterministic/probabilistic, nominal/ordinal analysis, 36 pages; references 4 pages.
(11) Hall, Regression/prediction, 32 pages; references 5 pages.
(12) Skocpol, Concluding observations, 22 pages; references 3 pages.
Incidentally, I was surprised not to find David Laitin among the contributors. His book about Russian-speaking populations in the countries of the former USSR is a model of comparative research in terms of problem selection, project design and execution. A similar observation would apply to Michael Hechter. Of course, his work is mentioned in several of the papers, but it would have been desirable to get a real account of the rational choice approach.
Apart from the introduction and conclusion, there are three parts which are entitled: Accumulation of research (2-4), Analytic tools (5-8), and Issues of method (9-11). However, as is often the case, these subdivisions fail to reflect the real contents of the papers. For instance, there are insightful methodological observations in Goldstone and Pierson's papers as well. Moreover the "accumulation papers" provide fairly personal views rather than a survey of the topic which, of course, would have been an overwhelming task.
There is one point which these diverse contributions, and indeed the book itself, make palpable. In the United States, there is currently an active group of researchers who have taken up the challenge of developing comparative historical research and are leading the way. This contrasts with the situation which prevailed until the late 1960s, when most landmark contributions were made by European scholars. The American school now is a fertile soil for new ideas. Many different approaches have been tried which is the hallmark of great intellectual effervescence. Which one of these approaches is the most promising? We come back to that question is the last section.
Some Promising Proposals
Pierson makes a very interesting point about what he calls slow factors. As an illuminating illustration he considers the case of an earthquake: although the earthquake itself takes place in a very short period, it is the result of the buildup of pressure on a fault line over an extended period. Pierson points out that all too often we are tempted to look for factors that are both temporally contiguous and rapidly unfolding and ignore slow-moving, long-term factors. This, I think, is an idea of cardinal importance in economics as well. For instance, the drop in Japanese consumption during the past decade is often "explained" by psychological factors whereas it can be accounted for more convincingly by the burden of long-term mortgage charges weighing on Japanese households since real estate prices began to plummet.
Goldstone's paper contains a number of innovative proposals. Among them I single out his suggestion that in order to better understand revolutions it may be helpful first to study stability requirements. This idea seems indeed appealing. Obviously, social ties hold a society together. Yet, so far, we are unable to measure the strength of these ties. Revolutions entail a reorganization and realignment of many social ties. How can we understand revolutions without an estimate of the strength of such ties either before or after the occurrence of the revolution?
Often, for the sake of simplicity, one is tempted to ignore exogenous factors and international pressures. As pointed out by several authors, this can be very misleading. As an illustration consider again the case of revolutions (Goldstone p. 74). In his Riverside Address delivered on April 4, 1967, Martin Luther King declared: "These are revolutionary times. All over the globe, men are revolting against old systems of exploitation and oppression. The shirtless and barefoot people of the land are rising up as never before." King was a man with exceptional historical insight; was he not mislead, however, in that particular instance? In fact, his assessment was correct, but most of the revolutions he alluded to were crushed and those who succeeded (e.g. in Iran or the USSR) were hardly led by "shirtless and barefoot people." This suggests that one should perhaps study revolutionary movements independently of their successful or unsuccessful outcome. For instance, would it not be worthwhile to give the failed Taiping rebellion in China as much attention as the Revolution of 1911 which, to some extent, was engineered by foreign powers. This is what Goldstone seems to suggest when he observes that international pressures are not sufficiently described.
In her very lucid, yet at the same time highly stimulating, conclusion Theda Skocpol observes (p. 410) that if a specialized set of scholars gains sufficient leverage over academic appointments and research funding they can define scholarly progress in their own intramural terms; she adds that comparative historical social scientists have never accepted such tendencies toward academic introversion. These are strong words which the present reviewer read with much pleasure. However, if scholarly progress cannot be defined in "intramural terms" what kind of objective criterion can be used instead? Along with a proposal made by Peter Hall (p. 393), I would suggest that offering testable predictions provides a criterion which is not only objective, but indeed seems to be the only unquestionable and indisputable criterion one can imagine. Whether an argument seems plausible or elegant is ultimately irrelevant if it does not lead to testable predictions. Albert Einstein, we are told, never accepted quantum mechanics because he did not agree with its basic assumptions as recalled by his famous statement that "God does not play dice." Yet, he could not deny that quantum mechanics was able to make predictions which turned out to be accurate to the sixth digit or, more generally, to whatever precision experiments could achieve.
Keeping in mind the objective of making testable predictions, whether structural predictions or (something which usually is more difficult) forecasts about the future, has the additional advantage of providing the researcher with a useful guide, in the sense that it often calls for a simplification of the problem, a reduction to its basic elements and the elimination of unessential side effects.
I hope this review has conveyed the idea that reading this book is a must for anyone who wants to seriously engage in comparative analysis. Thanks to the internet, statistical and historical information is now available from all over the world much more easily than ever before. For comparative historical analysis, this should be a revolution of the same magnitude as the Hubble space telescope was in astronomy.
In recent years Bertrand Roehner has published Pattern and Repertoire in History (with Tony Syme, Harvard University Press, 2002), Patterns of Speculation (Cambridge University Press, 2002), and Separatism and Integration (Rowman and Littlefield, 2002). He is currently working on How to Measure Social Cohesion (written in French).