APPENDIX THE HUMAN SCIENCES

Our general discussion of methodology has observed the tacit conven­tion of citing scientific examples primarily from the natural sciences, on the grounds that these examples are noncontroversially examples of science.

Both the human sciences and the natural sciences can be regarded as having theoretical and factual levels, and we would anticipate the same free play of constraints between these levels over time in both areas. To begin with, we can consider the views that the human sci­ences are separated from the natural sciences either because the ob­jects they study are different, or because there is a methodological difference in the relationship of theory to fact.

Some writers have assumed that the nature of the human being, because of free will or other properties, means that the explanatory structure of the natural sciences will be simpler than, or different from, that of the social sciences. But insofar as free will implies un­predictability, it is clear that systems in both the natural sciences and the social sciences are unpredictable, and that observational exchange of information with both physical and living systems as currently de­fined may preclude prediction. This is true even of deterministic sys­tems in the single case where measurement may have an incalculable impact on the system, as in some areas of quantum physics.

General methodological differences are usually traced to differences in the relationship between theory and fact. For example, it has been claimed that this relationship is one of explanation in the natural sci­ences, but one of understanding in the physical sciences.⁴ This cannot work as a general distinction, since both sciences depend on the no­tion of interpreting or understanding data text on our view, so that understanding is not limited to the human sciences. At the same time, explanation in terms of general laws is clearly sought by many theo­reticians in Marxism, psychoanalysis, and other disciplines clearly to be associated with the human sciences. The current failure to have achieved explanatory insights in the human sciences cannot, without further argument, be taken as an intrinsic failing. Desired but un­available explanation of data is a commonplace in the physical sci­ences.

We are not born scientists, but become scientists through the ac­culturation process described above. In this process, our understand­ing seems to engage explanation in the human and natural sciences typically in a different way. As human beings, reflecting on our own society, we tend to feel that we understand the general significance of various facets of human motivation. When we encounter an alien society, we attempt translation of their understanding of their actions into ours (which may produce an enlargement of our understanding) before we can comment on their actions. In both cases, we use this understanding to guide the process of gathering data text. If the data text is unexpected, we may well first expect that it has been gathered incorrectly. Here is a simple example. A personal but informal survey of the relevant institutions may convince us that one set of schools is better than another, a conviction that may not square with our first attempts to establish this fact quantitatively. Rather than accepting the first data as a refutation of our views, we may redesign the exper­iment to see if we can get the desired results. Now theory is con­strained by data here, for if our conviction is wrong, we will never be able to locate a plausible quantitative method to prove our opinion correct, but people with differing intuitions may spend considerable time in disagreement while they redesign experiments. This is not definitionally characteristic of the human sciences, since a similar guiding of data can be observed in medicine and biology, and is un­doubtedly active in new fields of physics, especially historically. In summary, when we acquire expertise in the human sciences, we are likely to find that we understand a theoretical level best intuitively, and we may allow this understanding to guide the search for plausible experimental or observational data.

Nearly the opposite may happen in the natural sciences, especially after an experimental tradition has embodied previous theoretical in­sights into the instrumentarium.

Logically, understanding may enter the natural sciences or the hu­man sciences at either the theoretical or the factual level, and we have examples of all of the possibilities. But there is a tendency for pre- scientific understanding, and then developed scientific understand­ing, to enter the human sciences at the theoretical level and the nat­ural sciences at the level of data text. This tendency, rather than a methodological distinction, accounts for the salient facts often taken as evidence of a methodological split. In the natural sciences, we typ­ically advance most quickly by making new instruments, or by apply­ing established instruments and techniques in new areas, and then constructing experiments in a logical space that becomes intelligible as theory advances. We may wonder whether we’ve really measured “anger” or “alienation” in a human sciences experiment, and what kind of an object the “quark” could be that we’ve postulated to explain an odd lot of data. This difference is explainable by the typical point of entry of understanding into explanation in the natural sciences and the human sciences, a reflection that eschews any intrinsic methodo­logical difference between them.

If low prior understanding calls for mathematical treatment, we can expect that data will be treated mathematically in the social sciences, and theory mathematically in the natural sciences.

There is, however, an important historical difference in the instru­mentaria of the natural sciences and the human sciences. The instru­ments of the natural sciences have in many cases replaced the organs of human sense perception. In this way, the succession of instruments defines a direction of progress. New instruments see farther, hear more distinctly, than their predecessors, and they may integrate the settled theories used to explain older data.⁵ The newer instruments trade on the success of the older ones, and because the instruments are distinct, the data domains they define are distinct also. Explana­tion seems to reach completion because the instruments define their boundaries, and then theory can exhaust the content sufficiently to reach almost certain predictive success. But in the human sciences, the same domain tends to be divided up by thought, and no new instruments have appeared to define a succession of data domains. It may seem that two phenomena have no functional relationship, like the price of potatoes in Idaho and the number of Russian missiles, but thought can always provide the possibility of a connection. Thus in thinking about humans and society, it is hard to draw a line inside the entire domain and settle for logically possible explanations of the resulting data. Human sciences have had essentially one empirical method for studying individuals in society, that of the questionnaire, and sophistication in its form hasn’t found a way of limiting the do­main of theory. Because of this, the empirical domain of the human sciences has remained nearly constant except for extension in the same conceptual space for over one hundred years, in spite of the fact that no intrinsic methodological difference can single it out as intrinsically immature or different from physics. What the human sciences require for more dramatic progress is not simply more data (of the same kind), as many empiricists have stated, but new instrumentation for obtain­ing data, or reasonable theoretical restrictions of data domain so that more exhaustive explanatory possibilities can be tried.⁶ It has not been the absence of theory, but the constraining effect of intuitive theory on the interpretation of data that has produced the current feeling of vagueness about human science explanations. Not too little theory but too much, particularly when its hold on data is so very difficult to break.

One can attempt to make the human sciences like the natural sci­ences, by looking for restricted data domains, but full sensitivity to the phenomena under investigation seems to constantly threaten to push one back against the full complexity of the human in society. In fact natural science is probably an elastic category, one that defeats any neat separation of the sciences, as Freud seems to have thought.⁷ A psychology that leaves the natural science haven of stimulus-re­sponse conditioning for slips of the tongue seems inevitably drawn into the horror of the data complexity about dreams. For individual slips of the tongue, plausible scientific explanations can be attempted, and an enlarged psychology is close to normal physical science. Dream material loses its anchor in the publicly observable event, and threat­ens to demand a new method. Freud may have thought that if the method adequate to the phenomena could be called forth, it would be seen to be a form of natural scientific explanation, but the textual evidence is uncertain. We will attempt to solidify our conception that general methodology is irrelevant to specific differences between hu­man sciences and natural sciences by looking at some specific exam­ples from economics and sociology.

Both the existence of the levels of theory and fact in the human sciences as well as our considerations of the entry level of understand­ing into the human sciences are confirmed by the historical relation­ship of economics and sociology. These disciplines are often in fre­quent contact in the study of economic man in a social setting, and we will ignore here their troubled relationships to the psychological study of the individual.⁸ The economic tradition, neoclassical and Marxist, attempts to enter the explanatory framework of the human sciences in the manner of the natural sciences, but it has been frozen in development by the lack of a suitable instrumental means of distin­guishing data domains.⁹ A long history of bitter argument between the neoclassical and Marxist traditions over the possible ideological status of the former in defense of the capitalistic status quo and its lack of historical insight is mitigated in areas like economic planning, where the nature of the data domain of current goods and services and short-term projections over this domain have forced a narrowing of formal differences in spite of philosophical opposition. What they have jointly produced is the most human of the natural sciences, or the most natural of the human sciences, but at any rate a discipline somewhat on the border of the intuitive distinction between the do­mains.

Opposed to mathematical economics, which tends to constrict to an apologetics based on appeals to some conception of theorizing in the natural sciences, is an intuitive sociology that refuses to see the eco­nomical as a natural subdomain of human science subject matter. These thinkers, historicists and holistic Marxists, find any thought of eco­nomic man to engage in principle the whole contradictory web of society, producing a domain of discussion whose complexity seems to call for a special hermeneutical or reflective method. Their entry to the human sciences is often through philosophical reflection, and the dominance of theory in their approach causes data to become a com­pletely secondary matter. From this perspective, mathematical eco­nomics is a peripheral subject of no real theoretical interest. There is thus a tension between those for whom economic man is a legitimate abstraction and those for whom total social structure is the only legit­imate subject matter of the human sciences, a subject matter that can only be illuminated, but not exhausted, in special disciplines.

Our position here is not that general methodology can resolve this dispute, but that this dispute is what should be expected in the light of the entry-level problem of understanding for the human sciences. No intrinsic reason why an instrumentarium cannot be developed for the human sciences has been found, so that this situation is the result of investment in the entry positions and a history of pushing compro­mise ideas into the other camp. We will examine this situation in the context of an important dispute between Adorno and Popper, which will have no purely methodological resolution. The difficulty of find­ing distinct data domains inside the human sciences is reflected in the instability of attempts like Max Weber’s to ground human scientific investigation in the middle ground of ideal types, or Habermas’s at­tempt to mediate between natural science and human science with a scheme of communicative competence.¹⁰ Both the holists and the mathematical economists have seen such attempts at mediation as complete failures, indicated by the fact that the mediators are not welcomed, but placed in the camp of the enemy. One might say that the lack of an appropriate data domain of intermediate scope means that these theoretical attempts are free floating, and can be pushed away by vocal adherents of each of the oppositional tendencies pro­vided by our analysis. Although we cannot resolve the dispute in terms of the analysis provided here, the structure of the dispute as we shall briefly consider it tends to support the methodological analysis that has been provided.

Neoclassical economics is similar to physics in that its theoretical units are all alike, economic human beings who are maximizing ex­pected profit or expected utility. This means that by assessing the objective situation in which an individual finds himself or herself, and assuming some form of rationality postulate, we can predict that the individual will act to maximize utility, the firm to maximize profit, and so forth.¹¹ It may then be said that such optimizing individuals frequently have only one option open to them, or at least very limited options. In other words, if the situation is described as one of perfect competition involving goods that can be arbitrarily divided, and so on, and price and cost curves are well defined, and we assume a producer is out to maximize profits, understands the situation as de­scribed, and is rational (acts appropriately), we can predict the output level the producer will seek. In such situations, there may be only one such point on the cost curve representing optimization. We ex­pect the producer to be near this point when we investigate empiri­cally. This turns the decision maker into a cipher or a computer, and the logic of the decision lies fully in the situation. With these con­straints, explanation of rational action can be more or less fitted into the standard explanatory framework of the natural sciences.

Neoclassical predictions have frequently failed in practice, a point taken by institutionalists, Marxists, holists, and other critics to indi­cate a failure in modeling.¹² The reply of neoclassical is methodolog­ically sound. Falsity and failure have not proven devastating in natural science provided that predictions are close enough to reality to be helpful. Further, where optimization cannot be traced to a single em­pirical variable in practice, the theory can be revised by a function maximizing at least two factors, a complication that parallels the ad­justment of natural science theory to data. Kornai, for example, has proposed splitting the analysis of the firm into a part concerned with the production of real goods and services and a control part that proc­esses information. In this view, the dubious neoclassical assumption that information spreads instantaneously can be combined with a the­ory of information flow that would allow the behavior of firms to fit empirical data more closely.¹³ Other attempts have tried the addition of an explicit psychological component to handle motivational com­plexity.¹⁴ But perhaps the most interesting attempt to increase the sophistication of neoclassical theory is the addition of class theory as a determinant of action, a move that brings mathematical Marxism into potential methodological congruence with neoclassical theory.¹⁵

Some critics of neoclassical theories have pointed out that micro­economics and macroeconomics in their current form can’t be easily reconciled, that there are divergencies between neoclassical econo­mists on specific problems in microeconomic theory, and so forth. At this level of abstraction, such observations are not likely to refute neoclassical outlooks, although they may stimulate the rate of research and development. The objections cited are of a piece with the atti­tudes underlying the unified science hypothesis. There is no a priori reason to suppose that all of economics would have to be integrated into one logically consistent economic theory before it was methodo­logically respectable. As we have seen, the reputation of physics as a paradigm science has not been tarnished by the fact that its theories are hardly integrated into one logically consistent theory.

The relationship between Marxist and neoclassical economic theo­ries has been illuminated in recent years as the result of a controversy about the internal consistency of neoclassical theories based on the work of Sraffa.¹⁶ In fact Sraffa’s contention that physical production plus the wage rate determines value quantities and profits and prices without there being a causal connection between value quantities on the one hand and the latter two on the other is a powerful criticism of a number of economic theories, neoclassical as well as Marxist. As against neoclassical doctrine, this critique says that the physical de­scription of productive mechanisms sets a range of possible rates of profit, with the actual rate of profit being set by the actual wage rate. The actual wage rate, however, is an outcome of class struggle, which is not a concept in the neoclassical domain. Neoclassicism is therefore inconsistent in the sense that its set of concepts is not sufficient to explain what it wishes to explain. (Economists use the notion of in­consistency in this way.) It can be argued in return that this criticism applies primarily to the version of neoclassical theory used for simpli­fied exposition in which individuals (their preferences and endow­ments) are analytically separated from firms and their technology. In general equilibrium theory, however, one describes a system in which factor prices (including labor prices) are determined partly by individ­ual preferences for leisure rather than for work, and individual pref­erences for immediate rather than delayed consumption. General equilibrium theory is a theory in which everything determines every­thing else, so to speak, and in which initial endowments and prefer­ences can’t be separated from technological possibilities. The theory is then perhaps not inconsistent, but it can’t explain how these en­dowments and preferences come into being. One then confronts di­rectly the question of whether historical changes of certain kinds are an appropriate object of economic inquiry. Because of the technical problems involved, as well as the remote empirical consequences of the theory, we cannot pursue this further here except to note that the charge of inconsistency doesn’t seem to terminate the tenability of neoclassical theory.

We shall not develop formal Marxism here except to say that in newer, more mathematical versions, it can enter into explicit contro­versy with neoclassical theories. The resistance to class theory, in view of the fact that the concept of class is methodologically merely a theoretical posit, can arise because the theoretical use of class notions threatens a charge of complete theoretical blindness for neoclassical theorists, who have simply failed to notice a fundamental variable if Marxist class theory is true in some form. Neoclassical economists have also found various historical prognoses involved with class theory to be empirically dubious. The dialogue produced by Sraffa’s critique is sufficient to indicate that these theories are roughly in the same methodological space, although it has taken this work to ascertain that fact with relative assurance. Holistic alternatives to economic theories relying for methodological insight on the natural sciences will now be considered.

The criticism of various historical schools of economics is that the method of economic abstraction produces economic theories that do not fit social reality, and perhaps cannot fit social reality. Insofar as the historical school simply attacks the fit of abstract theories to real­ity, it cannot legitimately draw the conclusion that the axiomatic method is wrong. It can only show that the axioms have been poorly chosen. To argue that the laws of physics are general, but the laws of economic development are peculiar to countries, and are therefore historical, so that economic laws are not equivalent to physical laws in terms of generality does not show a methodological difference between eco­nomics and physics. Rather than review a lot of bad arguments, we will present the following overview. The historical school began in the nineteenth century by simply eschewing abstract theory, and by confining itself to a close description of economic phenomena and their history from which it was attempted to induce “trends” on the basis of which economic recommendations could be made to prevent specifically anticipated human misery. In this procedure theory is not separate from fact, but both are seen as aspects of the particular his­torical data. As time went by, the abstract theories of neoclassical economics began to become more and more descriptively adequate with the development of econometrics. The response of the historical school was then to move from a detailed study of appearances to the view that economics must deal ultimately with a “hidden reality.” Then the mathematical side, no matter how well it fits theories to the restricted reality that it can measure, must be wrong about the nature of man in a total societal setting. This modern form of historicism must depend on a special method, dialectical or hermeneutical, to allow the reality hidden from mathematical abstraction to come into view. The first clear point at which a special method for economics and sociology has seemed to many to appear is with the methodology of verstehen and ideal types discussed by Weber.

The value dispute in whose context Weber attempted to synthesize natural science methodology and historicism arose in a quite definite context involving the Verein fur Sozialpolitik (a professional organi­zation), and the question of how, as a professional body, its internal theoretical disputes and the investigations of its members should be conducted. Again, the origins of economics and sociology as we know them are involved in the dispute, but the focus of the problem is how technical problems are to be approached. Weber’s contribution to the 1909 meeting of the Verein, as well as his call in his lecture for a value-free sociology as the legitimate pursuit for members of the Ve- rein, is the starting point for a long (and still continuing) discussion of the place and meaning of values for sociological and economic theory that is known as the value dispute. As Menger had earlier attempted to separate economic theory, economic history, and economic policy, Weber separates what he calls social science from social politics. Social science is to be value free, although it has a special method dependent on verstehen and on the consideration of ideal types. Because of the complicated, eclectic views of Weber, he has for various reasons been normally primarily associated with the positivistic tradition in the German view of sociological history, but with the historical tradition in the American and English views of sociological history. Weber was concerned to discuss subjectivism and rationality in the context of the historical school. In Weber’s opinion, the historical school had as­sumed that because human action was “subjective” in nature, it would possess irrational or arational characteristics. Weber thinks that “free” subjective action can be analyzed in terms of chains of rationality, chains linking motives or purposes to the means of satisfying them. “Irrational” behavior is insane behavior, or the behavior of people whose motives or purposes we do not understand. Once we have understood motives and purposes (through a process of verstehen), we will expect rational agents to act in accordance with certain laws of behavior linking these motives and purposes to the possible means of satisfying them. The use of verstehen is not simply to suggest hy­potheses that can then be understood and validated independently of verstehen in the context of justification, as the functionalist’s or posi­tivist’s tame version of verstehen would suggest.¹⁷ Verstehen is used to develop hypotheses that, when understood, can be put to an em­pirical test only in terms of this understanding. This was Weber’s concession of the historical side.

On the other hand, Weber did not see these hypotheses as subject to the test of crucial experiments prevalent in the area of the natural sciences.¹⁸ Weber maintained that the propositions of economic and sociological theory involve concept constructions peculiar to the social sciences. As theoretical propositions they would be “exact,” as are the propositions of the natural sciences, but they present a model of ideal­ized behavior rather than a law. The ideal types involved in these propositions are an enhancement or sharpening of reality, as are the propositions of the natural sciences. Humans plan actions using these ideal concepts, but their planning may go astray through biased or wrong calculation, or through failure to anticipate some real world interference. The propositions underlying rational action are not hy­potheses under which real actions are subsumed, nor do they stand for statistical averages of real actions. Rather they outline the course that rational human action would take in certain fixed circumstances if it were directed toward an unambiguous goal and if it were not disturbed by error and other irrational features. Weber thus thinks that the social sciences involve a separate methodological scheme, but that this scheme can be put into a fixed relationship with the meth­odological scheme of the natural sciences and brought under precise methodological control. But without an empirically attached data do­main, this conception can slide either into natural science or into philosophical holism when other writers fail to see it as an attempt to locate a unique subject matter for the human sciences.

It’s pretty clear that if the impact of verstehen is watered down, it is possible to read Weber as a neoclassical apologist and even as a Popperian, as many German scholars have done. As a result of the same fact, one can read Weber as a neoclassical apologist and primi­tive Popperian who is largely intelligible but has a slight eccentricity regarding the method of verstehen, as American and English scholars have done in making Weber a representative of the historical school.¹⁹ Weber quite clearly had no idea that his methods would be applied to trivial subject matter. He had in mind that sociology should inves­tigate the general cultural significance of social and economic struc­tures in total human communities, as his wide-ranging studies of In­dia, China, and Europe indicate. Weber thought the national limitations of the historical school could be transcended by his method, and that his method would result in no simple-minded reduction of history to economic forces, but a wide-ranging yet precise characterization of such cultural factors as religion.

In assessing Weber’s notion of value freedom, it is essential to re­member that Weber was trying to carve out a conception of econom­ics and sociology as sciences.²⁰ Weber accepted a fact-value dichotomy on philosophical grounds, in which facts could be grounded in expe­rience but values remain ungrounded. As long as sociologists and economists linked fact and value in their work, they could not achieve the sort of agreement that was characteristic of the natural sciences. This follows from the fact that value disagreement couldn’t be ration­ally resolved within their framework. It doesn’t follow from this that values won’t appear in economics and sociology. Rather, given certain values, appropriate lines of action will be objectively discussable within these disciplines. In the same way that a physicist’s personal feelings about physical particles should not matter in physics, an economist’s feelings about, say, socialism should not figure in his or her scientific analysis of socialism—and to conceal such involvement deliberately would be the mark of a scoundrel. The sociologist or economist must therefore externalize values, express them precisely, and study their relationship to each other and to fact. In this way, these disciplines can be said to have a detached and objective subject matter like other scientific disciplines. The eclectic genius of Weber is evident in the doctrine of value neutrality. Where many historicists thought that in­creasing technological power would fuse values and differences of opinion into a postideological society, Weber agreed with Marxism that only an increasingly intensified class conflict could be expected, an opinion apparently confirmed by the events of the First World War. But as against the prevailing Marxism of his day, Weber held that one’s thought need not be determined by class standpoint and associated set of values, but that a scientist could externalize class conflict and study it from a value-neutral point of view.

Critical theorists and some Marxists could argue that fact and value are not separable as Weber assumed, that values are not independent of rational grounding as Weber thought, and that the method of ver- stehen can only yield an analysis of the misleading appearance of con­scious thought and social totality where there are class conflicts, and not an analysis of social reality.²¹ The critical theorists have held all of these positions, and have hence lumped Weber with Popper (and even the positivists) as having an insufficiently reflective and self-crit­ical social philosophy. By way of anticipation, the following remark by Habermas concerning Parson’s appraisal of value freedom in Weber is crucial:

Please allow me a final intellectual historical observation. Parsons has claimed that Max Weber’s teaching is a development towards bringing about the end of ideology. Weber is said to have broken the trilemma of historicism, utilitarianism, and Marxism, and to have led the way into the free field of discussion beyond the European fronts of civil war. I envy our American colleagues their political traditions which permit such a generous and (in the best sense of the word) liberal interpretation of Max Weber. We here in Germany, who are still seeking for alibis, would only too gladly follow them. But Weber’s political sociology has had a different history here. At the time of the First World War he outlined a sketch of a Caesar-like leader-democracy on the con­temporary basis of national-state imperialism. This militant latter- day liberalism had consequences in the Weimar period which we, and not Weber, must answer for.... Viewed in the light of the history of influences, the decisionist element in Weber’s so­ciology did not break the spell of ideology, but strengthened it.²²

In the Positivismusstreit, or positivist dispute, there is the major discernible problem of whether or not German sociology is to be dom­inated by American empirical sociological methods and theory, which were clearly formed by positivist criteria of what is to count as sci­entific.²³ There is also the problem of whether German sociology is to be dominated by reflex of mathematical Marxism. This debate from the 1960s is of great clarifying potential for the view we are consid­ering of the relationship between the human and natural sciences. Adorno argues in this dispute for an autonomous (German) sociology, to be erected in some way at least partly on the insights of German idealist philosophy, autonomously because this reflective idealist com­ponent would not allow it to be merged with, or controlled by, re­stricted empirical domains. On the other hand Popper, who is the major antagonist of Adorno in the early stages of the debate, even if not a classical positivist, is clearly not attuned to any possibility that scientific method is other than perfectly general, that is, common to all scientific disciplines and having the same force in all social settings. Thus we have in Popper’s view the idea that the method of sociology and economics involves a rationality postulate distinguishing it from the method of the natural sciences, but also a view that a general methodological scheme can be worked out for all of the sciences that is valid everywhere and explicitly depends on a single model of sci­entific explanation. This attempt to ground German sociology on the manifestly successful practice of American sociology, as interpreted by Popperian methodology, is clearly resisted by Adorno. The title of Adorno’s earliest essay in the positivist dispute, “Sociology and Em­pirical Research,” suggests by itself that sociology and empirical re­search are two distinct things. The dispute between Adorno and Pop­per thus represents the clash of general methodology with the view that the human sciences must be coerced toward holism by their data.

We will now initiate a discussion of the positivist dispute in some detail, at least the initial confrontation between Adorno and Popper. Popper thinks that science can be demarcated from philosophy and metaphysics, that science is the only hope for the future of mankind, that science (if retained in a proper context) yields an optimistic fore­cast for the future of mankind, that science at a time has a structure that can be clearly articulated by philosophers, and that the relevant parameters of science (explanatory and corroborative) can be calcu­lated. In these respects he shares important doctrine with the positiv­ists, even though he rejected their key explications of the doctrine of confirmation. Adorno regards as positivistic any naive starting point in the accumulation of data that holds that the relevant data can be described in terms of existent unreflective categories. Some of the dispute between Adorno and Popper about positivism is thus merely terminological, and we must circumvent this if we are to come to terms with their more crucial confrontation.

Popper later complained that his remarks opening the 1961 Tu­bingen conference, the first round of the dispute, were not explicitly discussed by Adorno even though he presented them as twenty-seven theses and invited Adorno’s explicit agreement or disagreement.²⁴ What this means is that Adorno did not explicitly agree or disagree with Popper’s theses. Adorno could not have done this, since he had good grounds for supposing that the resulting debate would have been biased toward Popper’s position as a result of his acceptance of Popper’s lan­guage as the basis for discussion. On the other hand, it is not true that Adorno did not reply to any of these theses. Adorno quite ex­plicitly repudiated many of them. Let us begin by looking at Popper’s theses. The first three collectively state that we have constantly in­creasing knowledge but that we are also always ignorant. In the fourth and fifth theses Popper takes the origin of knowledge to lie in the selection of problems. He must, of course, take some such tack, since according to his doctrine of falsifiability, knowledge cannot originate in observation, and as conjectured it is always capable of refutation. Thesis five suggests that we can draw a distinction between significant and insignificant problems, that is, that we can tell which problems are the ones most worth discussing. Popper believes we are easily able to determine what our problems really are, so to speak. This thesis raises a direct confrontation with the thrust of critical theory. First, it lacks a reflective component. It may be that at least one source of knowledge is reflection on the problem space in an effort to determine what our problems really are, and a reflection that may involve considerable social scope. The epistemological position de­fended above is that significance can’t be determined unambiguously at a given time, so that Popper’s easy assurances here are problem­atic. Second, to see the state of science, including the social sciences, as a more or less historically contingent set of isolated problems that should be, and can be, solved independently on a piecemeal basis is once again to adopt a stance that seems to need tempering through reflection. Popper does not brood over the idea of attempting to make sense out of the total nature of the problem set in the way that critical theorists do. Popper is content to highlight the sort of short-term problems that are at the forefront of empirical research. Third, Pop­per finds success in the social sciences to be directly proportional to the “honesty,” directness, and simplicity” with which problems are tackled. Again, this seems naive. As honesty is not to be confused with the intention to be honest, and directness and simplicity become more and more obscure as one contemplates what they might mean in this connection, this advice seems to come down to the idea that one roll up one’s sleeves and get to work, that is, to nothing very clear at all in this context. By contrast, patience, deviousness, sophis­tication, reflection, and even other properties may take on a virtue of their own. Thus Adorno is staggered by what he sees as the naivete implicit in this advice, as seems clear from his remarks:

But the cognitive ideal of the consistent, preferably simple, mathematically elegant explanation falls down where reality it­self, society, is neither consistent, nor simple, nor neutrally left to the discretion of the categorical formulation.... Popper ob­jects to the cliche that knowledge passes through a series of stages from observation to the ordering, processing, and systematization of its materials. This cliche is so absurd in sociology because the latter does not have unqualified data at its disposal but only such data as are structured through the context of societal totality. To a large extent, the alleged sociological ignorance merely signifies the divergence between society as an object and traditional method.²⁵

Perhaps this is sufficient to indicate that Popper and Adorno do not always talk past one another. Although there is the view that all of our knowledge is conjectural in Popper, so that any of it can become falsified, there is no suspicion that it could simply fail to fit the out­lines of social reality.

We will not examine the rest of Popper’s theses in detail. Many of them oppose historical determinism, on the one hand, and relativism on the other, particularly as exhibited in some forms of the sociology of knowledge. Adorno could agree largely with these criticisms, since he considers himself neither a determinist nor a relativist, and con­siders the sociology of knowledge to trade on the unreflective cate­gories of the conscious mind. Popper’s more positive description of his own methodology for the social sciences is of more interest. In Popper’s framework, the crude unified science view of positivism is replaced by a view in which the natural and social sciences share a mode of deductive explanation, except that the social sciences must make use of a rationality postulate to make social scientific explana­tions applicable to idealized patterns of human behavior. Popper is thus the heir to Weber in attempting a special explanatory principle in the social sciences, while looking to the natural sciences as the arbiter of general methodology. If the rationality postulate must occur as a premise of explanation in the social sciences, however, there is little point in viewing natural and social science explanations as having a common form, since the rationality postulate is neither a theory nor a fact in the sense usually given in expositions of the deductive model of explanation. Its compatibility with neoclassical economic explana­tions, however, is manifest. Popper is opposed to explanation in terms of individual psychological states, or in terms of forms of life. He holds that rationality transcends a cultural setting, and that institutions or social groupings cannot have causal force in determining human ac­tion. Functionalist or historicist explanations may seem to work when people are acting in a fixed way over a period of time. But they cannot explain change in society. (On the surface, Popper has taken nearly the obverse position to that of mathematical Marxists, and the ease with which this can be done supports our contention that these posi­tions are methodologically similar.) Only individuals can change insti­tutions in Popper’s view, so that change is always the result of some set of individual actions. But an individual’s actions can be explained if and only if they are not unique, that is, if and only if they are not dependent on the unique thoughts of the individual making the de­cision, but can be understood as rational by others. Thus, to explain individual action, we describe choice situations and the decisions that “rational” individuals would make in these situations. This is an ideal type of explanation, variations from which are to be explained by hu­man psychological quirks, that is, miscalculations, bias from emotion, and so on. Clearly, standard neoclassical explanations of consumer behavior fall under the rubric of such situational analysis. Popper sees the main task of the social sciences as dealing with problems that arise when the actual outcomes of action have awkward consequences:

And this remark gives us an opportunity to formulate the main, task of the theoretical social sciences. It is to trace the unintended social repercussions of intentional human actions. I may give a simple example. If a man wishes to buy a house in a certain district, we can safely assume that he does not wish to raise the market price of houses in that district. But the very fact that he appears on the market as a buyer will tend to raise market prices.²⁶

The conservative nature of this outlook is nearly obvious. For exam­ple, a number of unfortunate experiences with urban renewal projects has brought to light the fact that people may react negatively to re­locations that seem to improve their physical environment enor­mously. The reasons are obviously various and somewhat mysterious, but they are expressed as a felt loss of significance or identity, or a shift in the amount of “essential” versus “nonessential” transportation time.²⁷ This behavior can’t be usefully set aside as irrational, although early planners did so. Now Popper could accommodate the idea that perhaps the situation could be described or redescribed, so that “es­sential” and “nonessential” transportation time could be objectively assessed and new rational explanations derived, but such examples seem to show that he downplays the role of psychological feeling too much, that his methodology contains no hint about when new descrip­tions of situations are required to preserve rationality, and that his methodology is consequently conservative in some root sense. Among other problems, the mere tracing of unintended consequences leaves the entire question of what to do about them to decisional choice that must be made outside of science.

These remarks do not refute Popper. He avoids the pitfalls of naked empiricism, argues against the reduction of the sciences to physics, is sensitive to various problems of the relationship of theory to obser­vation and of fact to value. His revolutionary doctrine of fallibilism for scientific thinking is coupled, however, with his doctrine of piecemeal social engineering, and here, I believe, is where the root issue be­tween Popper and critical theory arises. Popper believes that Euro­pean liberal democracies represent the best and least repressive his­torical form of human society, and the only form of society that is, in the long run, compatible with the scientific tradition that he admires. By keeping the context alive, one can keep science alive, and use its discoveries to solve human problems at a rate sufficient to have a chance of reducing human misery. At the same time, he has no real point of leverage against those who don’t accept his perception of European history, those who see humanitarianism within Europe and America purchased at the cost of barbarism elsewhere, those who find scientific knowledge currently an artificial human product geared to the manipulation of people and things, and those who see a sinister form of domination backed by scientific manipulation as the immedi­ate form of European and American culture. These sharply divergent attitudes are, in fact, held by critical theorists, and this helps to ex­plain the mutual hostility that one can feel between Popper and the critical theorists. Popper’s theory of the social sciences, in placing choice of values and resultant change in the category of decision, has no resources for dealing with sharply divergent attitudes.

Adorno is simply opposed to methodology in Popper’s sense of the term. Methodology is a theory of how the scientist ought to conduct his or her work, and as such its mode of operation is to endeavor to make science conform to its theoretical precepts. But this, for Adorno, is backward. No theory (and no concept) of discursive scientific lan­guage can ever fit reality, in this case social reality, exactly. There­fore, reality must be given ontological and epistemological priority in reflection, and as we come to uncover reality through a dialectical process, we must simply adjust method and theory to fit what is un­covered.²⁸ The idea that there is a general methodology of science wrongly idealizes science, and makes it seem a flawless route to an assurance of understanding reality, which it is not. The point of view involved can only aid science to be a tool of domination in the wrong hands. Popper’s methodology constantly places reality on a Procrus­tean bed of falsifiability. Popper finds the wrong measure of science because he overvalues philosophy, which has always tended to give the primacy to theory. The holistic contrast thus depends on a data domain, but one that is as broad as possible within our current con­ceptual resources.

This general assessment of Popper is to be found at the end of the first paragraph of Adorno’s reply to Popper, where Adorno says that he is interested in the concrete mode of procedure of sociology, and not in the deductive structure of the cultural presentation of its re­sults, as the logic of sociology.²⁹ Adorno makes no distinction between discovery and prediction in that he believes discovery to have a logic in which the object of study is allowed to reveal itself. As a result, a good theory can be construed as a description of a “hidden” reality, but not a perfect description of that reality. Adorno makes it clear immediately that he takes Popper to have dealt with only part of sociology.

Adorno can accept the verbal idea that sociology is based on prob­lems, expressed in Popper’s early theses, but he would not, of course, recognize the same problems as the focus of development. The pri­macy of the object, in the case of sociology a contradictory society, is crucial.³⁰ Although contradictory, society is determinable. What Adorno means becomes clear in the course of his essay. Current society is a contradictory mixture of modes of production, psychological types, traditions, that can only be elucidated by a historical understanding, and cannot be adequately understood in a consistent theory. A dia­lectical or critical theory cannot be broken down into subcomponents. In other words, to simply study the family in modem society in terms of its statistical appearance is pointless, because the family is what it is in modern society because of the interaction of family possibilities with the rest of society. In other societies, the family is a different institution, and understanding the current family involves these pos­sibilities and their historical link to the present. Popper’s methodol­ogy supposes that you can successfully theorize society into logically independent subunits, but Adorno’s view is that nothing approximat­ing the truth could come from this approach. This point would have to be thrashed out through specific lines of research, and a great deal of Adorno’s output elsewhere consists in showing the inadequacy of sociological research that doesn’t see the whole society reflected in the cultural object of immediate study:

There are sociological theorems which, as insight into the mech­anisms of society which operate behind the facade, in principle, even for societal reasons, contradict appearances to such an ex­tent that they cannot be adequately criticized through the latter. Criticism of them is incumbent upon systematic theory, upon further reflection but not, for instance, upon the confrontation with protocol statements.... If, in the last analysis, one does not wish to confuse sociology with natural-scientific models, then the concept of the experiment must also extend to the thought which, satiated with the force of experience, is projected beyond the latter in order to comprehend it.³¹

Adorno does not believe that the experimental method can simply be imported into sociology, since the total society cannot appear in the experimental setup. The society must be thought through if the ex­periment is to have any coherent meaning. Although Adorno agrees with Popper on the verbal formulation that the critical method is es­sential to science, he means by the critical method working through the apparent phenomena to see a reality not covered by theory, while Popper tends to believe that the phenomena themselves can be used to criticize theory. Adorno cannot limit criticism to sociological state­ments, since criticism must be of the society that ultimately is the object of sociological inquiry. Criticism cannot have logical consis­tency as a goal, since logically consistent theories and data may not be accurate representations of society. Adorno’s methodological ar­gument is that society cannot be comprehended by generalizing from a study of its parts in the manner of the natural sciences:

In sociology one cannot progress to the same degree from partial assertions about societal states of affairs to their general, even if restricted, validity, as one was accustomed to infer the charac­teristics of lead in general from the observation of the character­istics of one piece of lead.³²

A random sample of lead will have the same properties as other sam­ples of lead, that is, as lead in general has. Lead is a homogeneous class. But the subjective reactions of people or individual institutions will not have the same properties as society in general. A unique society is the focus of sociological attention. Thus we cannot ascend from individual cases to society, except in a trivial way. Society is already present in the individual cases. Opinion research, by accept­ing the ideological self-consciousness of individuals that they are in­dependent of one another and of society, makes a fundamental error:

For the findings of what is called—not without good reason— ‘opinion research’ Hegel’s formulation in his Philosophy of Right concerning public opinion is generally valid: it deserves to be respected and despised in equal measure. It must be respected since even ideologies, necessary false consciousness, are part of social reality with which anyone who wishes to recognize the latter must be acquainted. But it must be despised since its claim to truth must be criticized. Empirical social research itself be­comes ideology as soon as it posits public opinion as being abso­lute.³³

This position avoids any lazy call for more data or more theorizing, but it eschews methodology in general as offering no clear guidelines for development, since it is caught in the picture of natural science experimentation. Its proposal that significant data be sought within the context of a coherent view about social reality, however, is quite consistent with a dialectical view of scientific development, and of a piece with the dialectical view that has been proposed here to account for the development of the physical sciences.

In the confrontation between Adorno and Popper, two positions are represented that tie a philosophical theory to a coherent critique of empirical research. Adorno felt free, for example, to constrain data on the authoritarian personality according to his reflective position on Fascism, and he was able to offer sharp criticisms of specific instances of opinion research. Popper’s position represents a philosophical le­gitimation of the domain of data regarding the public manifestations of individual maximization of utility, an aspect of human individual behavior that can be clearly isolated in some situations and, as a con­sequence, is a coherent domain of investigation. Between these po­sitions, as they have been generalized above in the discussion of entry points of understanding, economics and sociology have not located clear data domains marked out by instrumentaria to which rival the­ories can attempt adaptation. Theories of intermediate domains in sociology have been linked by special methods to their own data, with the result that the relationship of the theories to each other becomes unmanageably vague, and data are used primarily in discussion of the theory to which they are related at their point of origin. This confu­sion of sociologies is the result of attempting to extend natural science models into domains of intermediate scope.³⁴ On the other hand, at­tempts to extend the scope of reflective sociology by subdividing the domain of social totality have produced normative social philosophies divorced from the constraint of specific data. The continuation of the positivist dispute indicates this situation nicely. Habermas and Albert achieve a partial rapprochement of the conflict between Adorno and Popper, but at the expense of clear lines of criticism of specific re­search.³⁵ This development, too complicated to follow closely in a book already of inordinate length, indicates once again that science is absent where the borders of data domains are not clearly posted.