Unifying the Sciences
The Structure of Science.
by Ernest Nagel.
Harcourt, Brace and World. 618 pp. $10.00.
In the early part of this century, the leading American philosophers were visionaries. Peirce, James, Whitehead, and Dewey were intellectual adventurers whose charts of knowledge, while fuzzy and inaccurate in detail, revealed new territories and changed the proportions of the old. But American thought, like America’s role in world affairs, seems to have outgrown its youthful idealism and to have settled down to the important but unexciting task of perfecting the status quo. Pragmatism, once a revolutionary call to arms against the tyranny of religious tradition over science, now functions as a method for defending the present state of science against proposals for radical change. Ernest Nagel, one of our most distinguished contemporary philosophers, exemplifies both the virtues and the defects of the present trend toward consolidation. The publication of his monumental work, The Structure of Science, is a major event in American philosophy because, in addition to its own intrinsic merit, the book sums up and shapes into a coherent pattern the achievements of American philosophy in the last three decades. He is one of the few contemporary writers in this age of technical specialization whose breadth of learning and sense of proportion qualify him to attempt the almost impossible task of constructing a coherent picture of the world out of the rapidly growing mass of scientific knowledge. The result is an impressively balanced consolidation of the contrasting insights of two rival traditions of scientific philosophy, logical positivism and pragmatism.
The early logical positivists, who considered mathematical physics to be the paradigm of scientific method, were obsessed with a vision of knowledge as a unified deductive system; in this respect they were the modern heirs of 17th-century mechanism. James and Dewey, on the other hand—guided by the more purely empirical procedures of psychology and sociology—regarded scientific theories as instruments whose validity is limited to the particular problem-solving contexts in which they are successfully employed. Professor Nagel, who is equally at home in natural and in social science, applies pragmatic criteria to theoretical issues in physics and biology, and applies rigorous positivistic standards of clarity and precision to the evaluation of our knowledge of man.
Merely to have effected such a consolidation is in itself a remarkable achievement. However, The Structure of Science suffers from a disparity between the brilliant clarity of the individual details and the somewhat blurry quality of their over-all composition. Physical science is taken to be the central source of knowledge, from which all other rational disciplines radiate; but this view blends into a pragmatic and pluralistic one of many different kinds of scientists at work, each hammering out his own concepts and techniques in response to his special problems. Nagel finds unity in the methods of the sciences and multiplicity in their subject matter, and the result of this contrast is that the relation between method and subject matter—which one would naturally assume to be an intimate one—becomes tenuous and strained.
For example, Nagel distinguishes four different senses of “explanation”—deductive, probabilistic, purposive, and genetic—and demonstrates how they function in different types of inquiry. He argues convincingly against the view that science merely describes sequences of phenomena, and that only theology and metaphysics can provide genuine explanations of them. But on the intra-scientific issue of whether only one of his four kinds of explanation is adequate, or whether each is adequate in its own way, his position is not perfectly clear. At times he expresses a strong preference for the deductive pattern of explanation exemplified by mathematical physics, and suggests that the other three kinds are merely incomplete forms of it; in fact, his technical ingenuity as a logician is most evident in his demonstrations of how purposive explanations in biology and probabilistic explanations in social science may be expressed as partial deductions. At other times, however, he suggests that probabilistic and purposive explanations are adequate alternative methods for dealing with problems that simply do not lend themselves to deductive theoretical treatment.
In any case, Nagel is by no means an old-fashioned mechanist: he explicitly denies that all branches of physics are reducible to mechanics, and that all other sciences are reducible to physics. On most specific issues in the philosophy of science—such as how closely a theory must match the actual structure of the phenomena it explains, whether real space is Euclidean or non-Euclidean, whether the elementary particles of nuclear physics have definite spatio-temporal location, and whether invariant laws of human behavior can be discovered by scientific inquiry—Nagel argues for non-intervention by philosophical legislators and maintains that working scientists must be free to decide these issues on pragmatic grounds. But, like many logical positivists, he seems to be himself a more sophisticated kind of mechanist who would like to see the logical structure of every science resemble that of classical mechanics. And it is apparent from his consistent search for deductive structures in biology, sociology, and history that only a keen awareness of the special complexities of each branch of science inhibits him from declaring a goal of every discipline to be the logical unification of its theories within the framework of a universal science.
Now this cautious mechanism, tempered by his sensitivity to the need of each science for some conceptual autonomy—however tense at times the relation between them—does work effectively in Nagel’s account of the natural sciences. In his analysis of sociology and history (in the last third of the book), however, the fluctuation between a program of deductive reorganization of these disciplines and a pragmatic defense of their statistical and crudely empirical character is more problematic. Nagel’s impressive replies to the stock arguments against any kind of behavioristic inquiry into human conduct arc powerfully reasoned and, I believe, unanswerable. But his positive thesis, that the social sciences are capable of developing a theoretical structure comparable, in precision, logical clarity, and predictive power, to the natural sciences, seems to be far less plausible.
The key issue he considers is whether or not the social sciences are capable of developing the kind of neat deductive theories which, in the natural sciences, have enabled man to explain and predict events with ever-increasing reliability. Nagel recognizes the conspicuous absence of theoretical laws in social science, but he does not regard it as evidence of an essential difference between social and natural science. Yet he modifies this claim for social science with so many qualifications that it emerges as little more than a hope—and, finally, he limits himself to affirming that the evidence so far fails to prove the logical impossibility of theoretical knowledge of man that approaches our knowledge of the physical world.
Now it seems to me that Nagel vacillates between two views of social knowledge, each of which has great merits but only when the commitment to it is wholehearted. One, the mechanistic view, finds the trouble with the present state of social science, as Nagel frequently suggests, to be that its concepts are not defined in terms of measurable physical magnitudes. The other, the view of pragmatic pluralism, stresses “the essentially ‘practical’ nature of our current interests in social phenomena” and recognizes that the study of man is fundamentally unlike the study of nature, due to its practical orientation, the greater complexity of its subject matter, and the necessarily vague character of its concepts. Thus: “. . . there are some grounds for doubting that the social sciences are likely to refine their current distinctions beyond a certain point—a point fixed by the general character of the problems they investigate and the level of analysis appropriate for dealing with those problems.”
Should social science, then, continue to pursue its gross statistical methods, occasionally enlivened by dashes of literary metaphor (“inner-directed,” “charismatic personality,” “conspicuous consumption,” etc.), or should it construct deductive models whose concepts are operationally defined in terms of precise measurements? This is the methodological translation of the classical metaphysical question: is man an integral part of nature, or does he to some extent stand outside of nature? Either answer, the monistic or the pluralistic one, would be visionary in both a bad and a good sense. But it would chart a clear direction for social inquiry. John Wisdom once wrote that philosophy is most interesting when it is wrong, by which I think he meant that the role of philosophy, unlike that of science, is to suggest rather than to prove. If so, then Nagel’s scholarly caution reduces rather than augments his philosophical power.
The concluding chapter of The Structure of Science raises similar questions about the nature of historical inquiry. Nagel argues forcefully against writers like Isaiah Berlin, who deny that it is either possible or appropriate for historians to search for objective explanations of historical events. He agrees that there are no specific laws of history but concludes from this, not that history is not a science, but that it is an applied science which explains events in terms of laws derived from theoretical disciplines such as psychology and sociology. Even if one grants, as Nagel successfully demonstrates, that deductive explanations are not logically impossible in history, there still remains the more practical question: ought one to look for them? Nagel himself argues that it would be impractical to expect historians to satisfy the rigorous standards of natural scientific explanations, and that we must resign ourselves to the continued use of probabilistic and genetic patterns. But if the historian can get along quite well without deductive models and the physicist cannot, doesn’t this suggest a very fundamental difference in subject matter and in the method appropriate to investigating it? If it is wrong to impose a priori forms on scientific inquiry, if the scientist must be free to deal with his problems by whatever conceptual means he finds effective, then what purpose can be served by demonstrating that his results could, in principle, be recast in a form appropriate to natural science—what purpose other than to suggest that, if we had sufficient knowledge, we would be able to unify all the branches of science within a single conceptual system patterned on classical physics?
My insistence on what seem to me to be discrepancies in Nagel’s general picture of knowledge should not divert attention from the extraordinary scholarship, clarity, and logical insight that inform his detailed discussions of specific problems in each field of science. Nor can one do justice to his philosophical work without recognizing the depth and consistency of his commitment to scientific method. Nagel has devoted his career to demonstrating that in the sciences unsolved problems are evidence of vitality rather than impotence, and that the methods for solving these problems are perfectly continuous with the methods that expose them. In this fundamental purpose he has, I think, been eminently successful. His work, taken as a whole, provides an authoritative assessment and a spirited defense of scientific method against metaphysical substitutes for intellectual hard work.