Claudio Gutierrez
In the first part of this study I have tried to clarify the role of subjectivity in the social sciences generally. There I came to the conclusion that two different aspects must be distinguished: subjectivity (or teleology or purposiveness) in the subject matter (which we decided to call subjectivity2) and subjectivity in the explanation (which we decided to call subjectivity1). The first aspect relates to the fact that the object of study in the social sciences is always an entity, man, who has purposes, knowledge, desires so that the terms "purpose," "knowledge," "desire" must appear in the social-scientist's language. These terms are best interpreted as theoretical terms, as the term "force" is interpreted in the physical sciences, due to their obvious irreducibility to pure data of empirical observation. The second aspect relates to the fact that the scientist himself, be he a social or a physical scientist, has also purposes, beliefs, passions, etc. of his own. These are determinant in the way he does science, and even define him as a professional scientist.
A conclusion that can be drawn from this is the fundamental homogeneity of all sciences, in a double sense. First, both physical and social scientists have to use theoretical terms which are not reducible to empirical sensation. Secondly, both social and physical sciences are dependent, ultimately, on subjective considerations. Theoretical terms, because of their lack of direct empirical foundation, have to rest on the creative imagination of man, on his heuristic power, and are, to that extent, subjective (in the sense of subjectivity1). The difference between, for example, "force" and "purpose" is not decisive. Both are theoretical terms which are to be operated in much the same way, namely, within a context of ultimately heuristic or subjective powers. One can point to a distinguishing characteristic of social science, however, but it seems to be rather external to the methodological problem. In fact, one can do social science about the physical scientist, as scientist, whereas one cannot possibly do physical science about the social scientist, or for that matter any other scientist, as scientist. This reflexivity, so to speak, seems to have impressed very much the writers that have occupied themselves with methodology of the social sciences, up to the point of leading them into thinking that social science is an altogether special kind of science, obeying logical canons completely different from the canons of general logic (RICKERT 21, p. 36). Now, I think that such reflexivity presents, of course, an unavoidable temptation to the methodologist. But he will do well to resist it and refrain from seeing it as anything more than an interesting, but irrelevant, characteristic, external to the fundamental epistemological problem.
Two remarks, I think, are here in point. The first one is that not making the distinction between subjectivity1 and subjectivity2 is the source of the strength that the "reflexivity temptation" seems to have. If "purpose" must have a place in social science, and who could deny that, the fact that we ourselves, in doing social science, are also purposive seems to shed all the light we need in our methodological problems. We are led into thinking that all we have to do to develop our science is to contemplate the workings of our own mind and somehow project what we find there into the outer world. I do not want to say that this approach is totally wrong. There is a grain of truth to it. But I do think that the approach will be much better if corrected, in the sense suggested, i.e., by a more nearly full analysis of the role of subjectivity in science generally, and particularly in the social sciences.
The second remark that I want to make moves, as it were, in the opposite direction. The fact that there is subjectivity at two levels of science makes for the existence of a connection between these two levels. The need for theoretical terms and the impossibility of reducing them to empirical observation makes the subject matter of the science dependent, to a very important extent, upon the heuristic creativity of the scientist. This fact alone is capable of providing the needed link between subjectivity1 and subjectivity2. Theoretical terms make sense because the theorist is trying to close the scientific system with the help of his own personality. He does this by making a bet on the "manageability of ignorance" or the prophetic power of the products of reason. Two avenues of interpretation for the role of theoretical terms may seem to open here. We can treat them as heuristic anticipations of the scientist, assumptions in which he dwells. Or we can rather see them as empirically empty concepts with only operational value which, nevertheless, point, via the principle of indiscernibles, NOTE 1 to an essential intervention of the personality of the scientist in the configuration of the subject-matter.
This last statement may need an illustration. Consider, for instance, the concept of "perfect rationality" in economic theory. As W. D. Lamont very well puts it: "To assume, for purposes of theoretical analysis, that the person entering into an economic relation is acting with a coldly perfect rationality is the only assumption which could properly be made. . . ." (LAMONT 55, p. 41). Alternative assumptions, which are not impossible to image, will have in common that they are indiscernible among themselves, in the methodological sense. There is no sufficient reason to postulate one of them rather than any other. Only the perfect-rationality assumption, even if not empirically founded, as good theoretical term, can be told apart from the other alternatives. This is why we prefer it to all the others. The intervention of the scientist as a person in this regard is, of course, the methodological decision of taking the "unfounded" theoretical concept rather than doing without any concept for the purpose of closing the scientific system. What all this means is, of course, that in the last analysis the positivist interpretation of theoretical terms is reducible to this non positivist or "heuristic" interpretation.
In my opinion theoretical terms like "purposiveness," "perfect rationality," "foresight" in economics, as in any other field, are amenable to be rightly treated in either of the two ways which have been described. If we choose to treat them as heuristic anticipations, they are the non-formal context in which the scientist dwells in order to operate with his scientific material. If we rather choose to treat them as well-delineated formal concepts, with the aim of closing the system, they are the token representation of the assumptions and may be well construed as ideal types. In the former case they will be pointing to the subjectivity of the explanation. In the latter case they will be in the operational side, where subjectivity is not immediate since objectivity is the intended aim. Both interpretations are possible; both are necessary. No one can do the job of the other and their principles are essentially dissimilar. In particular, one cannot expect that informal assumptions be able to do the job reserved for operational concepts. The one thing one can certainly not do with non formal assumptions is to operate with them. We dwell in them or in something else. But we do not operate with non formal assumptions, as assumptions.
The proposal implicit in this reasoning suggests the convenience of assuring a fair degree of flexibility to the operational side of theory, undisturbed by any illegitimate direct intervention of the postulational side. I think that much of the confusion that one finds in the methodology of economics comes from the fact that such degree of flexibility in the operation of formalism is not usually allowed. Flexibility will be guaranteed if the distinction between the two theoretical levels, the one of non formal assumptions, the other of abstract types and operational concepts, is more generally recognized. It should become apparent that a single set of assumptions can be dwelt in while operating with alternate models. If there should be at all a distinction between informal assumptions and "operational machinery," they must be recognized as essentially different and not fitted for each other's job. In particular, paramount assumptions must be conceived as ultimately unformalizable, even inarticulable, because they are the atmosphere, as it were, in which formal models are possible. Their formalization, possible if we dwell in higher assumptions, is not even a requisite of scientific activity. The only thing that is necessary is that one recognizes the need of having some theoretical terms representing the assumptions at the operational level, so as to always assure the logical closure of the system. NOTE 2
Assumptions, of their very nature, are less "replaceable" than are models, they are more permanent and more resistant to change. Models, on the contrary, tend to be movable, and several of them can be operated upon, even simultaneously and without inconsistency, under the same set of assumptions. As an illustration, take the contrast between the concepts of microeconomics and those of macroeconomic. The latter are not logically derivable from the former in any strict sense. NOTE 3 Nevertheless, they somehow correspond to the same basic unformalizable assumptions of utility theory. In the face of facts like this, one would tend to view models as sort of conventions or useful devices (FRIEDMAN 53, p. 15). From my methodological perspective this consideration is, however, not meaningful. "Conventional" makes sense only as opposed to "natural" or "unconventional." But here we have theoretical concepts, the only ones one could possibly afford in the situation, and they are called "conventional" without opposing this property of theirs to anything else. Is it the intention to say that they are not true? If we believe what they say and have enough ground for so believing, they are as true as they could possibly be.
The rendering of models as conventions makes sense only in an epistemology that makes a sharp distinction between "historical material" and "theory" (as filing system). Within that positivistic methodological framework the problem arises as to how to make formal predictions about purely empirical history. The interpretation of models as conventions is a way out, although it poses more problems than it solves. I think in particular that the way in which the relationship between "assumptions" and "hypotheses" is treated in such a scheme is unfortunate. They are not distinguished as belonging in essentially different levels. Rather, the assumptions are said to be false (unreal) while the hypotheses are supposedly true (confirmable by their implications). NOTE 4 The approach fails to reveal the most important feature of the assumptions, namely, that they must be informal or ultimately unformalizable. On the other hand, I do not consider "history" and "theory" as the only two possible "states" of knowledge. As a matter of fact, I think these are the only impossible states, being only imaginary limits of an infinite gradation, even if these imaginary limits are taken by the positivist as reality itself. But none of them does actually exist; they are only polar positions or exhaustion points of epistemological paradigms. The intermediate positions, in contrast, are the real knowledge we can aspire to have: the abstract models, of systems of such, together with the theoretical terms. They make for logical closure of the system and represent within it the unformalizable conditions of all scientific thinking.
Flexibility of models should be understood as essentially coupled to modifiability of the epistemological "distance" at which the models are seen from the assumptions. We have considered previously that it is in the nature of models to be able to function as assumptions with respect to other pieces of information if the scientist chooses to dwell in the (otherwise a) model. We saw that, within certain limits, the assumption also can be objectivized as a model if the scientist chooses–and can–dwell in higher assumptions. Now, this modifiability of the intellectual focus in the assumption-model relationship provides a new opportunity for scientific decisions. One has not only to choose between models; one must also choose the "right focus" for the theoretical purposes at hand. New possibilities of success, and failure, necessarily arise. In particular, it is now possible to get too close to the empirical limit of the system, with the corresponding poverty of scope. Or on the contrary, too far removed from it, with the consequence of an "inflation" of theory that might impair the effectiveness of the system.
A good example of what has just been said and which I am going to develop in the sequel, is the case of the concept of "satisfaction." What we can call the economist's interpretation of it understands it as a (formal) way of comparing the different elements of income NOTE 5, as the token for the (informal) "sense of direction" that must be postulated for the operational concept of "substitution" to make sense. What we can call the utilitarian's interpretation, on the other hand, tends to understand it as a way of bypassing the fundamental variety of human wants and aims and reducing them to some (objective) state of satisfaction or happiness. The defect of the latter position, to my mind, is that its assumptions are "too far removed" from the phenomena under consideration. "Satisfaction" itself is made an object (model) rather than a postulate. NOTE 6 Let us refrain from construing man as a being in search of the end satisfaction, as the utilitarian pleasure-and-pain machine. Let us assume instead, with narrower focus, that men are after many different things. Then we will better understand man's passion for objective freedom of choice and for all that can contribute to further that freedom, especially the magic of economic power.
As Lamont aptly puts it, economizing means the ability to hold together, in an unitary consciousness the totality of diverse demands and the attempt to obtain the greatest possible realization of that totality (LAMONT 55, p.55). Paradoxically, the utilitarian's conception, by insisting on a unified object of desire, leads to a piecemeal approach to economic calculation. One is envisioned as deciding on alternative courses of action that are, as it were, cut-off from all context. The economist's conception, which respects the fundamental plurality of objects of desire, leads to the full consideration of all future choices represented by the marginal utility of money. NOTE 7 To express it again in Lamont's words: "The individual comes to every situation with more in mind than the ostensible alternatives between which he has to choose. There is in the background of his thought, the sense of 'all the rest' of the things which are in actual or potential demand by him. . . ." (LAMONT 55, p. 55). This is the sense of the economic value of things. Its discussion we now must enter.
NOTE 2 The problem of the possible exhaustion of alternative assumptions has already been dealt with. In particular, it is my opinion that the assumption of purposiveness in the subject matter of economics is ineradicable. The importance of "purposiveness" as an abstract model is another question.
NOTE 3 "No proof is available that . . . macroeconomic assumptions cannot be deduced from macroeconomic ones. But there is also no proof that the deduction can be effected, and there is at least a presumption that it cannot be. . . ." (NAGEL 61, p. 544).
NOTE 4 See in this connection a very interesting discussion about Friedman's "principle of unreality" (HAGEL 63).
NOTE 5 "The significance of utility (more properly "satisfaction") is that it provides the correct common denominator of the infinitely diverse components of income. . . ." (KNIGHT 51b, p 221).
NOTE 6 The case may seem not to be relevant to present-day discussions of methodology of economics. Nevertheless, I think it worth mentioning since something like the utilitarian's view still lurks behind many objections to a more professional conception of value in terms of "purchasing power". We will come back to this later.
NOTE 7 "The price of a commodity tends to equal its marginal utility, as measured in terms of money, i.e., relatively to the marginal utility of money to its purchaser" (HENDERSON 63, p. 41).