REVIEWS
William H. Riker and Peter C. Ordeshook, An Introduction to Positive Political Theory Prentice-Hall, Englewood Cliffs, New Jersey, 1973, pp. 387. In seeking to pull together a comprehensive and integrated treatment of the very diverse and rapidly growing domain of work by economists, mathematicians, political scientists, and sociologists (which they embrace under the title 'positive political theory' ) Riker and Ordeshook exhibit a commendable ambition. An Introduction to Positive Political Theory does a superb job in many areas, eg. an integrative treatment of the Prisoner's Dilemma along with a discussion of externalities and market failure, a comprehensive (although perhaps unduly technical) overview of the spatial modelling literature, an innovative examination of the conditions under which public regulation is desireable, a concise treatment of the size principle with an emphasis on its empirical applications, etc. It seems clear to me that An Introduction to Positive Political Theory will be the basic textbook for political science courses in 'public choice' for some years to come. If, in what follows, I seem to be unduly critical of a book which is overall such an excellent one, I must reply that this is a tribute to my high expectations of these authors. Riker, in my view, has in the past decade had the greatest impact on the discipline of political science of any single individual - directly, through his own writing, and even more importantly, indirectly through the work of his students, work which now appears in the discipline's major journals with a regularity quite alarming to its more traditional practitioners; while Ordeshook has, together with a handful of others, following in the footsteps of Hotelling and Downs, helped create an extensive and important subdomain of political science - spatial models of electoral competition and choice. As with any text which tries to cover a burgeoning and ill-defined field, An Introduction to Positive Political Theory suffers from flaws (mostly minor) of omission and interpretation. As a textbook, it suffers from a failure to suitably define its intended audience. In the preface, the authors assert their aim is to introduce some of the recent work in positive theory to Theory and Decision 7 (1976) 231-234. All Rights Reserved Copyright 9 1976 by D. Reidel Publishing Company, Dordrecht-Holland
232
REVIEWS
"students at the junior-senior and graduate levels." For the most part (Chap. 9 is a major exception), the authors do succeed in relegating partial derivative signs, integrals, etc. to the footnotes. On the other hand, in many chapters the authors follow the strategy of first developing a theory largely in the abstract via symbolic notation before introducing more concrete examples or applications, a practice which I believe to be poor pedagogy for an introductory text. Also, in some cases (eg. the Chapter on N-person game theory) mathematics is developed, political science applications for which are not shown. More generally, this book is model rich and data poor/With rare exceptions (eg. on the size principle), the authors fail to deal with data about human behavior relevant to evaluating the descriptive/explanatory usefulness of the models being considered. This appears to be an explicit sin of commission rather than an inadvertent error of omission, since the authors do not even include data from their own well known article on voter turnout in U.S. Presidential elections as a function of perceived party differential and election closeness. I believe this omission of empirical evidence and the related failure to consider theories of human motivation which do not neatly fit a rational choice framework is an unfortunate one, in that it presents the reader with a one-sided view in which the innovativeness and comprehensiveness of the rational choice view are not really shown off to full advantage. The reader new to this material is less likely to develop an appreciation for its potential or a motivation to pursue it. Furthermore, though the book is very well written, its extensive use of symbolic notation and its very compact style of presentation make it very forbidding reading. I am skeptical that it can be used as a text at other than the graduate level except by students with reasonable mathematics or economics background. 1 Some Minor Caveats (1) Riker and Ordeshook entitle their book An Introduction to Positive Political Theory, and one might think that they had in mind a distinction between normative political theory and positive political theory similar to that commonly drawn by economists between positive economics (what is) and normative economics (what ought to be). This would be an error. On the contrary, in the authors' view, "the moral and the descriptive go ... hand in hand." However, although some mention is made of the views of traditional political philosophers such as Hobbes and Rousseau, insufficient context
REVIEWS
233
is provided to permit the less sophisticated reader to be aware of the extent to which the approach advocated by the authors sheds new and important insight on traditional questions of political philosophy and democratic theory eg."What is the proper role of the state? What holds society together? What constitutes the public interest?" etc. By blurring over the traditional "is-ought" distinction, the authors also blur over both the significance and the political science/historical context of some of the work on which they report. (2) In the ethical sphere, the authors appear committed to a straightforward application of the Pareto principle in which 'fairness' norms (other than the sort posed by Arrow's conditions) do not enter. (See, however, pp. 110-111, 157-160, 2 3 1 - 2 3 9 , 2 9 1 - 2 9 2 , and 373). The term 'justice' does not appear as an index entry. This omission I regard as regretable, even though perhaps inevitable in a work which is already almost four hundred, tightly written pages long. (3) The autors have chosen to invert the usual order in which 2-person game theory is presented before N-person theory is developed. Since the presentation of the 2-person theory remains a standard one, I'm unable to understand the reasons for this order reversal. (4) The literature on choice draws nourishment from a number of traditions, including economic models of consumer and producer choice (largely oriented to behavior under conditions of certainty), models of probabilistic inference and choice (rooted in Bayes Theorem and oriented to behavior under conditions of risk), welfare economics (concerned i.a. with the existence of Pareto optimal equilibria), and, last but not least, the theory of games (oriented to behavior under conditions of uncertainty where there is strategic interaction among volitional actors). Each of these traditions copes somewhat differently with the notion of 'rational' choice. Riker and Ordeshook have, in my view, come closer than anyone else in integrating these diverse approaches, but the seams often show, and there still remain some unreconciled elements. In particular, the authors fail in Chapter 2 ('The Assumption of Rationality') to satisfactorily interrelate 'consistency' notions of rationality with 'minimax' notions of rationality and with 'expected utility maximizing' notions. (5) There are puzzling omissions in An Introduction to Positive Political Theory, e.g. Farquharson's work on sophisticated voting is treated only cursorily (pp. 9 7 - 9 9 , p. 366) and neither Rae et al.'s work on proportional representation schemes, Coleman's work on exchange models, nor Hirsch-
234
REVIEWS
man's work on exit, voice and loyalty are discussed. Of course, we can expect to see some important, very recent work on spatial modelling, modelling of coalition formation processes, and minimax regret models of voter and representative choice incorporated into the next edition. There has also been a healthy growth in empirical work in the rational choice tradition the past few years, and we hope such empirical studies (on e.g. voter rationality, parliamentary coalition patterns and payoffs, and small group decision process) will be given a central place in the next edition of An Introduction to Positive Political Theory. That there will be a second edition I have no doubt. In scope, clarity, and innovation in integrating diverse approaches, Piker and Ordeshook's work will not be easily bettered. The flaws I have pointed out are readily correctable in subsequent editions.
Department of Political Science, State University of New York, at Stony Brook Lehrstuhl fiir Politische Wissenschaft, Der Universitiit Mannheim
BERNARD GROFMAN
NOTE I More suitable as an introductory undergraduate text, although considerably less comprehensive or innovative, is Steven Brams, Game Theory and Polities, Free Press, 1975.
REVIEWS
235
W.V. Quine, The Roots o f Reference, Open Court Publishing Co., La Salle, Illinois, 1974, 151 pp. $10.00. In 'Epistemology Naturalized', an essay published a few years before the appearance of this volume, Quine argues that epistemology should henceforth be practiced as a branch of empirical psychology. The task of the enlightened epistemologist, Quine holds, is not to justify our claims to knowledge, but to describe and explain the process by which we acquire our theories about the world. From this book one can get an idea of what epistemology practiced in the Quinean style would be like. Quine argues that the best way to study the acquisition of theories is to study the acquisition of the language in which they are couched. Then, focusing his attention on our knowledge of objects, he speculates on how we might have learned to refer to them. The book is divided into three parts. Part I is concerned with learning in general. Here Quine attempts to show how some of the notions basic to learning theory - perception, disposition, similarity, salience - can be accomodated by a psychology which is behavioristic, in the sense that it recognizes only those mental phenomena which manifest themselves in behavior, and materialistic, in the sense that it identifies all mental entities with physical mechanisms whose nature is not yet understood. Part II deals with the primitive steps of language - acquisition. At this stage, according to Quine, the child has not yet learned to refer to objects. Thus, it is only in the third part of the book that Quine gets around to speculating on how a child might master those devices in his language which give him the power of objective reference. It would be impossible in this short review to cover all the interesting issues raised by this book. I shall therefore touch on just a few points. (1) The verification theory o f meaning. In this book Quine leaves no room for doubt that he is a positivist about meaning. "The meaning of a sentence," he tells us, "lies in the observations that would support or refute it." (p. 38) This seems inconsistent with Quine's rejection of the verification theory in 'Two Dogmas', but I think the two passages can be reconciled. In the paragraph which immediately follows the sentence quoted above, Quine tells us that although the verificationists were right in saying that meaning is evidence, they went wrong in assuming that meaning and evidence attach to individual sentences. It is this point, I take it, which Quine has in Theory and Decision 7 (1976) 235 -236. All Rights Reserved Copyright 9 1976 by D. Reidel Publishing Company, Dordrecht-Holland
236
REVIEWS
mind when he rejects verifationism in 'Two Dogmas'. What he objects to is not the equating of meaning and evidence, but the assumption that the bearers of meaning and evidence are individual sentences. (2) Analyticity. On this topic there does seem to have been some change in Quine's position. He seems mellower, these days, about meaning. On p. 64, he dreams of "a scientific semantics," which would involve "retracing ... the process by which we acquire our command of ... language .... " He also offers a definition of analyticity: "a sentence is analytic if everybody learns that it is true by learning its words." (p. 79) This definition, however, seems to be subject to the same objection as the definition of stimulus analyticity in Word and Object. Obvious existential claims, such as 'There have been black dogs', or 'Something was once red', would count as analytic. (3) Ontological relativity. Quine holds that there is a difference between learning observation sentences and learning to refer to objects. Thus, in the case of either a child or a foreign adult, the fact that there is a verbal response that he makes on demand in case red is present and not otherwise does not show that he is referring to red. For we do not yet know whether he is referring to the color or predicating red of a certain object, or a part of that object, etc. What bothers me about this is that the description of what the child has learned does not seem to accord with what Quine says in Part II, where he notes the importance of salience and demonstratives for ostensive learning. Ttrere is not space to discuss the matter fully here, but consider the following example. A parent waves a red scarf in front of a child and the child, pointing, says, 'That's red'. Quine would presumably say that it is not clear whether the child is predicating 'red' of the scarf, or of one of its parts, or of one of its temperal stages, or whatever. But, given that it is the scarf, rather than one of those other things, which is salient in this case, this is not entirely plausible. Why not say, instead, that the child's 'that' refers to the scarf, and that his 'red' is predicated of it? This book has all those good qualities which we have come to expect from Quine: clear and witty writing, incisive argumentation, and, most impressive in this age of specialization, the ability to organize his individual insights so that they form a philosophical system. It goes without saying that the book is a welcome addition to the literature. EDWARD F. BECKER
REVIEWS
237
Herbert Stachowaik, Allgemeine Modelltheorie, Springer, Wien-New York, 494 pp. $ 40.00. Herbert Stachowaik's Allgemeine Modelltheorie is a first step toward establishing a comprehensive synopsis of all the fields of model theory. His approach includes, besides the mere formal concept of model theory, epistemic, pragmatic and axiological points of view. He begins with a review and benevolent criticism of the epistemic models of positivism, neo-positivism and logical positivism. Stachowiak argues that these models, even if supported by tests methods are too narrow. In order to rectify this, he proposes to introduce neopragmatic ideas into model theory, the aim being to attain, in general, a more active, decision-oriented science and scientific methodology. For this purpose, he initiates his K-system ('K' stands for kybiak, an abbreviation) which is, in reality, a tacit expression of Stachowiak's philosophical point of view. According to this K-system, each model should contain self-regulating, cybernetic factor - not necessarily a human one - a kind of extrapolated Simon-problem solver which outlines and performs all decisions and inter-actions of the system with its environment. He combines the K-factor with the following elements: (1) information and game theory, (2) cybernetics, (3) theoretical operationresearch, (4) system-theoretical decision making, and (5) a kind of prograined, hypothetical robotization. With this K-factor Stachowiak attempts to introduce operational or neopragmatic trends into all-kinds of model theory. Thus, model theory is to be changed into a kind of active, decisionoriented planning of the problem - Man vs. his environment, It is also true that, by means of this K-factor, Stachowiak introduces a highly idealized, but formally useful, operator into model theory, one which is able to adapt the function of models and systems to environmental, social and external scientific purposes. In the second chapter, the general concept of K-models is discussed. Here the representational character of models is stressed. Stachowiak argues that the diminuative, pragmatic and isopomorphic character of representation should be achieved by a maximal, structural approximation of the model to the original. It is interesting that Stachowiak's model concept is so wide that he even includes, for example, philosophical and cognitive representation of models. It becomes clear that, for the sake of this general theory, especially for the sake of the inclusion of non-scientific models Theory and Decision 7 (1976) 237-238.All Rights Reserved Copyright 9 1976 by D. Reidel Publishing Company, Dordrecht-Holland
238
REVIEWS
(M. Scheler, K. Mannheim, K. Jaspers, H. Leisegang), the author is forced to neglect the formal concept of models (whose explication is handled in only 36 pages). Nevertheless, the book represents an attempt to synthesize, in a general synopsis, the most encompassing features of all kinds of models. It puts forward new general criteria, especially a novel formalization of pragmatic criteria, and analyzes, therefore, the relationship of models and actions. Thus, it constitutes a value-theoretical foundation of model construction. As a first effort to encompass all scientific systems within a general model theory, it represents a highly systematic approach. On the other hand, it is, perhaps, not to the advantage of this book that it relies so heavily upon Martin's pragmatic logic, thus neglecting the structural, set-theoretical notations and formulations which are the usual tools of model theory. However, we should also point out that this same reliance upon pragmatic logic, while not explaining the quantitative models of science, is certainly advantageous with regard to general philosophical models. Fi~lally, we should say that Martin's methods do not seem to permit their beirt~ extended in the way which would be necessary to understand and support Stachowiak's generalizing ideas. This is not to say that the author should discontinue his use of the decision and game-theoretically oriented formalism he is quite familiar with. My suggestion is of a moderate, methodological nature, one which should help Stachowiak to develop his original and fresh neopragmatic approach into a general model theory.
Univ. of Nebraska
W. LEINFELLNER
REVIEWS
239
Edward O. Wilson, Sociobiology: The New Synthesis, The Belnap Press of Harvard University Press, Cambridge, Mass., 1975. 697 pp. $ 20.
In his extensive glossary E. O. Wilson characterizes sociobiology as "the study of the l~iological bases of all social behavior." Soeiobiology presents a program for research and an encyclopedia of recent work in this area. It seems likely to fix its title as a disciplinary label. Sociobiology as a research program is designed to overcome two broad criticisms that Wilson advances against previous work. First, much of the research of the last decade provides narrow theories of quantitative data about social groups without offering the requisite higher level account of these results. Sociobiology seeks to establish a fundamental, general theory which will explain these low level theories by appeal to principles derived from population genetics and behavioral ecology. It is in this respect that the book differs, for instance, from I. Eibl-Eibesfeldt's Ethology: The Biology o f Behavior (second edition, Holt, Rinehart and Winston, 1975)which provides a discussion of many of the same topics. Confident of the prospects for obtaining a fundamental sociobiological theory, Wilson predicts that ethology will be increasingly absorbed on the one side into neurophysiology and on the other into sociobiology and ancillary disciplines. Second, Wilson holds that where more general theories haven been offered, they have often been supported by the 'advocacy method' which focuses on 'literary content' without adequate convern for scientific falsifiability. The book is divided into three parts. Part I sets forth the program of sociobiology, its basic concepts, and relevant techniques from population biology and behavioral ecology. Wilson stresses the importance of group (especially kin) selection in solving the central theoretical problem of sociobiology: namely, how altruism can be adaptive. He argues that kin selection explains altruism and thereby provides the groundwork understanding the 'evolutionary origins of ethics' and the tension between individual and group interests. The second part of the book (almost half the text) provides a detailed analysis of methods, results and controversies attendant to such standard topics as group size, communication, aggression, territory, dominance, roles, sexual bonds and parental care as they occur across a broad range of phylogenetically diverse animals. In each case he tries to account for the evolutionTheory and Decision 7 (1976) 239-241. All Rights Reserved Copyright 9 1976 by D. Reidel Publishing Company, Dordreeht-Holland
240
aEvI~ws
ary import of these social phenomena and thereby to sketch the shape and application of general sociobiological theory. Part III examines sociality of four conspicuously social kinds of animals: colonial invertibrates, social insects, nonhuman mammals and man. Wilson is concerned to indicate how group (kin) selection explains inportant differences in social behavior among them. In the final chapter Wilson briefly considers the relationship of sociobiology to sociology. He maintains that sociology is in need of reconstruction on a neurobiological foundation. As it now stands sociology is still at the 'natural history' stage and will be amenable even to a low level 'phenomenological theory' only when it is merged with cultural anthropology, social psychology and economics. Once this transformation has taken place, sociology can be included in the 'new synthesis' and sociobiology will be able to reconstruct the evolutionary history and adaptive significance of the mechanisms revealed. The sociobiologist embraces a very broad concept of sociality which construes a society as 'a group of individuals belonging to the same species and organized in a cooperative manner'. Correlatively, related concepts have been extended to admit, for example, communication and socialization by chemical means. The strength of such a broad characterization lies in the encouragement it gives to the development of a nascent science. Disciplinary imperialism can serve to redirect and develop an area of scientific inquiry. Wilson's call for a neurobiology to augment sociological theory has the virtue of avoiding the objections which can be raised against alternative 'advocacy' approaches which attempt to establish claims about the genetic bases of social behavior, such as aggression or dominance, solely by pointing to similarities in behavior between human beings and other animals. A neurobiology which shows, for example, the relationship between episodes of aggressive behavior and hormone levels provides a groundwork which supports the claim that this behavior is genetically based via selction for hormone producing capacities. The weakness of the sociobiological approach to sociality lies in the real distinctions it may adumbrate. We have no assurance that a fundamental theory in sociology at the neural level will be forthcoming, especially if such a theory is expected to provide biological bases for all social behavior. Wilson has not produced an account, for example, of the evolutionary significance of role and status in human society and he admits that there is troubling counterevidence. He is left with the suggestion that "the influence of genetic factors
REVIEWS
241
toward the assumption of certain broad roles cannot be discounted." The relationship between these broad roles and roles as more normally discussed in sociology is presumably to be settled by future research. To the extent that a fundamental sociological theory with neurobiological underpinnings remains dilly a promise, the impact of sociobiology on the methodology of the social sciences remains in doubt. It is unclear as well what effect sociobiological understanding of altruism and the evolutionary origins of ethics wouls have. We cannot blithely move from the explanation of the origin of certain patterns of behavior or belief to the claim that they are justified. Of course, as Wilson points out, an adequate sociobiology might constitute part of the 'factual' basis for normative decisions about the shaping of future society. Sociobiology is already a prominent addition to the growing literature in biosocial science and has provoked substantial controversy, especially among sociologists who are understandably suspicious about being colonized by biology. University o f Nebraska at Omaha
DAVID W. PAULSEN