V. Subjectivist economics *
In his essay the "The Domain of Subjective Economics: Between Predictive Science and Moral Philosophy", James Buchanan characterizes subjective economics as a discipline that "emphasizes ... the existence and the importance of the area between empirical science and moral philosophy" (Buchanan 1987, 68). Noting that his article was first published within a volume of essays in honor of Ludwig v. Mises one might wonder why Buchanan did not locate subjective economics between a posteriori empirical predictive science and a priori praxeology. Presumably he wanted to have two extremes that both make good sense. As he does not think that the claims of praxeology make good sense, he seems to take resort to moral philosophy which he "define(s) as discourse that embodies an explicit denial of the relevance of scientific explanation" (Buchanan 1987, 68).
As a moral philosopher I do not feel too comfortable with this characterization of the moral discourse. The characterization would of course fit in nicely with some of the non-consequentialist approaches to morals like the Kantian one. On the other hand, consequentialist moral philosophy of the Humean type is concerned with more or less explanatory issues about the actual workings of moral institutions, too. It is, at least in some sense, both explanatory and normative.
Such a Humean theory of the moral order tries to derive means ends relationships or hypothetical imperatives, that show how certain sets of moral institutions could be instrumental to our aims, ends, or values. As a theory of prudent behavior that recommends ways to reach given ends of individuals it has to rely on scientific explanations or on scientific laws predicting how normative proposals eventually might work out. As a normative theory it does not simply explain or predict what is or will be going on in the world from the objective point of view of the external onlooker. It acknowledges that "the observer is among the observed" (Buchanan 1987, 67) and therefore proposes improvements from the point of view of the participant in the ongoing game of life.
In the last resort, the basic outlook of such a normative theory is more or less identical with what James Buchanan calls "subjective economics". As Buchanan is a known admirer of Hume and argues in the best Humean tradition this criticism should not concern him too much. However, it still does not seem to be clear what subjective economics is and where exactly it is located on our map of human intellectual achievements. The further characterization of the subject by Buchanan himself starts from the Smithian propensity to "truck, barter, and exchange". He then goes on to insist on the importance of the active role of the human actor as opposed to passive adaptation: "The person who initially imagines some postspecialization, postexchange state and who acts to bring such a state into existence must engage in what I shall here call 'active' choice. He must do more than respond predictably to shifts in the constraints that are exogenously imposed on him." (Buchanan, 1987, 69)
This notion of active choice leads to one of two interpretations of Adam Smith's basic economic analysis. On the one hand, Smith seems to be concerned with explaining the relative exchange values of commodities. On the other hand, his explanandum are the exchange institutions themselves. Even though one may agree with Buchanan's distinction between Smith as a classical value theorist as opposed to the interpretation of Smith as an institutional or constitutional economist, it does not seem to be clear how the remark eventually would be related to the notion of active choice.
Human individuals can make active choices with regard to institutions as well as with regard to simple acts of choice and exchange under institutions. The crucial element within active choice is that of innovation. However, innovation can come about as an unintended side effect in both spheres. An institution may gradually evolve due to minor changes in individual behavior and so may specialization. Somebody might, by chance, have deviated from his own preferred mixture of privately produced goods and may afterwards find out that somebody else has deviated in a complementary way. This may be a start, a first marginal specialization in a process of ongoing further specialization -proceeding until the costs of further specialization exceed the gains. The same may apply to the gradual evolution of a social institution. This process presupposes variation, selection, and tradition but not necessarily active and intentional choices. Even animals can form traditions (cf. the well-known examples of potato washing among Japanese monkeys (makakes), patterns of bird songs, the practice of opening milk bottles adopted and spreading among certain birds in GB).
Even if one admits that the plasticity of human behavior as well as the complexity of human responses to given stimuli goes much beyond anything that could be imagined among animals, this still is not sufficient to draw the principal distinction between 'natural" response behavior (natural plasticity making viable complex patterns of operant conditioning) and active choice that Buchanan has in mind. For him there is something in human behavior that cannot be approached as simply more complex animal behavior. "The residual aspects of human action that are not reducible to ratlike responses to stimuli, even in the much more complex human variants, define the do main for a wholly different, and uniquely human, science -- one that cannot, by its nature, be made analogous to the positive predictive sciences of the orthodox paradigm." (Buchanan, 1987, 78)
To make clear that he is not adverse to empirical science, per se, Buchanan continues: "There is surely room for both sciences to exist in the more inclusive rubric that we call economic theory. We must acknowledge that in many aspects of their behavior, men conform to laws of behavior so that such behavior becomes subject to scientifically testable prediction and control through the external manipulation of constraints. But we must also acknowledge that men choose courses of action that emerge only in the choice process itself. Men create value by the imagination of alternatives that do not exist followed by the action that implements the possibilities imagined." (Buchanan, 1987, 78)
Here we seem to encounter a quasi-Kantian intuition that human beings are part of two worlds. On the one hand, human beings are part of nature and accordingly there is a role for predictive science in dealing with the actions of human beings as parts of nature. On the other hand, human beings arc special because they are endowed with the faculties of reason and the imagination and thus are not merely parts of nature. As with Kant the concept of the dual nature of rational beings leads into difficulties as far as two springs of causality are assumed to exist -- a "noumenal" one, rooted in reason, and "phenomenal" one, rooted in nature. Buchanan, insisting that only man as a rational being can create the genuinely new which for the very reason of being genuinely new cannot be prediemd, is evidently on the same track as Kant. This account of an extra-natural kind of causal influence on the course of affairs of our world may be at least on face value a bit more plausible than some of the Kantian tenets. Still, it is obviously vulnerable by arguments that are well known from the discussion about free will and the Kantian notion of the autonomy of the person (cf. for an overview Watson 1982, Denett 1984 and for a discussion of related topics Baurmann 1987).
Now, Buchanan might argue that there is a weaker version of his dualism that falls short of ontological dualism. This defense is itself linked to the limits of human knowledge and the unpredictability stemming from this fact. At least to a human observer, human behavior is special for the observer, being human himself, he is not an omniscient external onlooker but rather part of the ongoing process. For the latter reason he can adopt a participant's attitude (cf. for this term Strawson 1974) towards this kind of behavior and can understand it 'from the inside'. Though it does not make much sense to say 'what if I were in the shoes of that rat in situation x ... ? it does make sense to ask oneself 'what if I where in the shoes of some other human being in situation x ... T.
In the following sections, I will try to make several suggestions about how to render more precise some of Buchanan's intuitions. The first will be taken from elementary decision theoretic modeling. The second draws attention to the fact that Buchanan's argument in favor of subjectivism is closely paralleled by some arguments supporting an (empiricist) "hermeneutic" approach to law. The controversy between predictive, in particular, American and Scandinavian "realism" and empirical normativism in legal theory is in many respects similar and even almost identical to that between subjectivism and positivism in economic methodology. In making more precise the tenets of subjectivism it may be helpful to borrow some notions from the debate in legal theory. Finally, I will argue that the main thrust of subjectivist economics may perhaps best be regarded as a philosophical endeavor that strives for a better self-understanding of what interaction in a world of rational beings might be like.
Active choices can be made only as far as individuals can exert an influence on the future course of events. Such choices are intentional and thus the ability to exert some influence must be perceived in advance. Indeed, human beings can act on behalf of a model of the situation in which their action takes place. One if not the most basic rationality principle requires that they discriminate carefully between those aspects of an action situation that they can influence and those that -according to their subjective expectations and their model of the situation -- cannot be intentionally influenced.
Using the elementary notion of a decision matrix an individual who sticks to the rationality precept mentioned before will carefully distinguish between the aspects of the situation that are beyond his control and those that can be controlled by the decision itself. Basically, the aspects that are beyond the individual's control will form the columns of his decision matrix. His probability estimates -- if he has any -- that the events forming the columns will actually occur belong in the same category. The functions ("actions") that map the columns ("states of the world") into the final outcomes as such are also beyond the individual's control. The individual cannot control the causal effects that his actions will have, given certain circumstances or states of the world. This "translation" of actions into effects is determined by laws of nature. However, the individual can control by his choices which of the functions actually will apply. Thus the individual can choose a class Of possible outcomes under the probability distribution over the states of the world he subjectively takes as given. The standard model is represented by the Matrix M of order mxn:
|
Matrix XVI |
|||||
|
M |
S1 |
S2 |
S3 |
|
Sn |
|
f1 |
f1(S1) |
f1(S2) |
f1(S3) |
|
f1(Sn) |
|
f2 |
f2(S1) |
f2(S2) |
f2(S3) |
|
f2(Sn) |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fm |
fm(S1) |
fm(S2) |
fm(S3) |
|
fm(Sn) |
This standard model of rational choice does not seem to leave much room for "active decision making" as opposed to passive forms of decision making. Every choice of a function is as active or as passive as any other -- and it is presumably more passive than active. Thus one might think modern decision theory does not leave room for the notion of active choice. However, carrying the analysis one step further suggests a natural way to explicate the concept of active choice. Alluding to the notion of a price taker one could say that human beings are not "model takers". They can actively change their model of the decision situation. Thus, instead of taking the matrix M from above they might develop the matrix M' of order hxk:
|
Matrix XVI |
|||||
|
M |
S1 |
S2 |
S3 |
|
Sk |
|
f1 |
f1(S1) |
f1(S2) |
f1(S3) |
|
f1(Sk) |
|
f2 |
f2(S1) |
f2(S2) |
f2(S3) |
|
f2(Sk) |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fh |
fh(S1) |
fh(S2) |
fh(S3) |
|
fh(Sk) |
In this matrix M' either the states of the world or the set of actions or both are different from those described in M. More precisely, the theories about what is possible or likely have changed. This is like a constitutional change that precedes the actual choices. (Changes in the subjective probability estimates and subjective values -- utility and probability being conjoined measures anyway -- are irrelevant here or they are at least of minor interest.) Thus the relevant shift involved in an act of active choice might be regarded as being more or less like a change in theory or even a kind of "paradigm shift".
One might object to this explication of the concept of 'active choice' that for a fully rational being there is no room for such shifts. The rational individual will always be bound to apply the best knowledge available to him at the time of decision. Though this is true, it should be noted also that investing in more and more precise models of situations and finer and finer assessments of possibility spaces has to stop somewhere. We never have the final knowledge of the situation we are in. We are not in the position of the omniscient onlooker. While travelling along we have only vague notions where we are within possibility space. Even though there are rules of optimal stopping, such stopping rules are no remedy for the fact that we have to act under conditions of radical ignorance. We neither can know what we once will know nor can we know what a golden innovation could be like and how it may change expected payoffs. Could we know, we would know the innovation already and it would cease to be one.
The latter remarks are quite in line with what some Of the Austrian economists seem to assume. They may be a first step to lend some plausibility to the notion of active choice from the point of view of standard decision theory. But it is still an open question what a science of active choice could be like. Evidently it cannot be a discipline that itself proposes methods to create such choices. This would be a science of innovation -- a thing that never will exist.
Subjectivist economics as an intellectual endeavor could hardly be boiled down to the statement that there are active choices which are innovative and therefore in the sense of all true innovations unpredictable. We have to took for additional ways to explicate the meaning of the term "subjective economics" and what it means to give explanations from the point of view of subjective economics as opposed to "behavioral economics" or behavioral science in general.
Here Buchanan would draw attention to the fact that in social interaction we are not interacting with nature but with other human beings (1979). With other individuals we can interact strategically whereas with nature not. This raises the question what the difference between a game against nature and a strategic game is.
Some theorists have claimed that a game against nature is a special form or specification of a game in general. Involved in such a game, one plays against a special type of player namely "nature". Other theorists have argued that a game against a rational player is a special case of a game against nature. One knows something about nature, namely that the states of the world are determined by another human actor. Both views would put games against nature and games against human actors in the same category. On this, both positions agree. There is merely disagreement about what is the more general case.
The subjectivist economist should disagree exactly on what the others agree. He should argue that the expression "game against nature'' is misleading. Of course, one can approach social interaction with another human being from the point of view of normal decision theory. In this case one would rely on normal behavioral laws in assessing how human beings in general will act in situations of a certain type. One could easily imagine a set of behavioral laws that would specify a certain probability distribution over the behavioral alternatives of another individual. These laws are behavioral laws and thus can be used to explain or to predict what the other will do in the normal scientific way. Here behavioral science is used as a tool by the rational chooser to form a model of his choice situation. This model is based on the presumption that the other individual is a part of nature and can be treated in the same predictive way as other parts of nature.
On the other hand, the decision maker in social interaction with another human being might try to take into account his own faculty to understand the decision situation the other is facing and so may the other, and so on. Then both individuals will not only be aware of one another but also that the other individual is aware of this fact, etc. They also accept that both of them have full control over their own choices -- in a certain realm at least -- and that they will rationally exert this control in forward looking choices.
Regardless of the fact that they will ascribe rationality to each other as well as freedom of choice they may well try to predict what the other is going to do. But, they will not do this in the same way and with the same methods that they use in a "game against nature". Each will try to figure out what the other is going to do and at the same time will know that the other will try to figure out what he is going to do. They will know that each of them has a model of the situation which includes the knowledge that the other has such a model containing the knowledge that ....
According to classical game theory each player will make his predictions about what the other one is going to do from the point of view of the model of the situation the other one is supposed to entertain (the model is common knowledge). This means that each tries to put himself into the shoes of the other one. Both try to emulate the choice situation of each other. Though this is a way to make predictions, the process in which predictions are created is one of making fictitious rational decisions instead of directly predicting such choices. The predictions are not based on behavioral laws about what other persons regularly will do but on the model of the situation the other person is supposed to entertain. It is assumed that all participants in the game are endowed with the same rationality. Though game theory is concerned with analyses of what is rational in certain situations and the concept of what is regarded as rational changes within its own progress it assumes throughout that -whatever may be our concept of rationality -- all players are endowed with this faculty in the same way. In this sense classical game theory clearly is counter to fact from the outset Thus the whole game theoretic endeavor is characterized by a side constraint of theorizing that strictly separates it from empirical studies.
Of course, should one use such a model for predictive purposes in an empirical sense, then one may stipulate an empirical or behavioral law that people act rational on what they themselves perceive to be the incentives of the situation. From the example of classical game theory -- and Buchanan always argued in favor of applying game theory as opposed to other paradigms of economic theory to social interaction -it is also clear that subjectivist economics, to a large extent, may be a speculative scrutiny of counterfactual "what-if" questions. Such a speculative study will not directly lead to empirically testable results. But, as the example of game theory shows, it may be helpful in rendering hypotheses that could be so tested. For instance, often the "rationale" of certain institutions can be fully understood only against the background of deviations from a model of perfect rationality -- and actually our understanding of the rationale will change if our understanding of rationality changes. The why and how of the deviations is of interest and also whether the alleged deviations arc deviations at all.
This interpretation of classical game theory as a paradigm example of a subjectivist approach to human behavior shows that subjectivism and the rejection of mathematical method in no way are logically related. However, this does not imply that only such a speculative and highly mathematicized theory as classical game theory could form an example of the subjectivist approach to human behavior. To the contrary. Jurisprudence in the spirit of modern "empiricist hermeneutics" -- an Anglo-Saxon movement that is quite different from the less sober minded German variants (cf. Albert 1988 for a criticism of applying "German hermeneutics" to economics) -- is close kin to subjectivist economics. It may facilitate our understanding of subjectivist economics if we take a look at some of the discussions of analytical jurisprudence first and only then return to the example of game theory.
Legal theorist Neil McCormick characterizes his own approach to human behavior as hermeneutic. He says: "By a hermeneutic approach I mean one which seeks to explain human actions, practices, etc. through an interpretation of the meaning they have for those who take part in the actions, practices, etc." (McCormick 1980, 29) This is a clear statement of the basic aims of hermeneutics. Human actions are to be explained by referring to the intentions and perceptions the actors have in the action situation. Still, it is not completely clear what it means to "explain" human actions in terms of what they mean "for those who take part in the actions, practices, etc.". Do the intentions of the actors as such explain what is going on in the world of overt action or does an acceptable explanation in the last resort require that we add a law of nature such that human beings in general will act according to their intentions in and perceptions of situations?
In the first case the actions of human beings are explained without reference to a behavioral law or, for that matter, a law of nature. From a methodological point of view it does not seem too convincing to talk of an explanation and at the same time to insist that there is no reference to an explaining law. An explanation based on an empirical law and one that is no-. based on such a law are of different types. To claim that the latter is of a superior type -- as some of the adherents of pure hermeneutics are inclined to -- seems to be even more misleading.
As long as there is no law involved the explanation is simply deficient from the point of view of the standard covering law model of scientific explanation. From such a methodological point of view it is merely a sketch of an explanation. There may be different or secondary forms to which we apply the same term "explanation". This is not illegitimate. However, it is necessary that we carefully distinguish between the two types of explanation described before. A fullfledged empirical explanation besides motives always must refer to a law of nature linking motives and subsequent action. In an -.mpirical sense the motives as such do not explain the acts of the individual. However, it may be true that the behavioral laws needed to produce a more complete explanation are in many cases trivial. This leads to the second case or to what may be called empiricist hermeneutics as opposed to pure hermeneutics.
In this case one acknowledges that a law of nature is necessary for any explanation that is not merely a sketch of what may become an explanation. At the same time it may be true that for many explanations in the realm of human behavior the law involved may be such a general and trivial one as the one stating that human beings respond to what they subjectively regard as the exigencies of the situation they are in. Then, as Popper would say, the "logic of the situation" carries most of the explanatory weight. Again, even this notion of the "logic of the situation" has a subjective and an objective -- behavioral -interpretation.
In the first case, the explanation may be built on the logic of the situation as it is perceived by the subject whose actions are to be explained. This is a truly subjectivist account of the antecedent premises of an explanatory argument. It reconstructs these premises "from the inside" first and only afterwards bridges the gap between the premises and overt behavior by some "bridge law" or other. This bridge law connects subjective perceptions of situations with predicted overt behavior. As a law it is itself beyond the control of human actors and thus has the same status as other "laws of nature". Insofar as the explanation is based on a subjective model of the situation in the antecedent clause of the law such an explanation is completely in line with what has been said before about the hermeneutic approach to human behavior. Still, one may wonder how the subjective point of view of the acting individual can be reconstructed by another individual without taking recourse to some objective or at least intersubjectively accessible indicators of what the actor himself may perceive. Such indicators will characteristically stem from some knowledge of the Personal history of the individual whose behavior is under scrutiny (cf. Vanberg 1989). The indicators can be indicative only because there are laws of learning theory and cognitive psychology that somehow relate observation statements about past events with some theoretical term s.
The behavioral science opponent of the subjectivist economist might want to eliminate "intervening variables" like motives and corresponding theoretical terms and laws altogether. He could suggest to start from the objective or intersubjectively accessible facts of the environment and to relate these facts directly to reactions in overt behavior. Such an objective reconstruction of the logic of the situation will rely, in the last resort, on observable "incentives" and try to eliminate intentions of the actors from the theoretical picture of the action situation. The motives, aims and values of the acting individual are left in the dark of what essentially is treated as a black box. (From a philosophy of science point of view the Beckerian methodological ideals behind such an approach may seem surprisingly close to old fashioned observational positivism and thus quite besides the point. But, in the philosophy of language we have both tendencies: on the one hand a view of speech acts that is referring to intentions of the actors, on the other hand a view that argues that the meaning of a speech act is not dependent on the intentions of the actor or the addressee of the act but exclusively on the objective features of the action situation including its history and context.)
As far as the distinction between subjective and objective views of the logic of situations is concerned, none seems to be superior per se. I think that those economic theorists who claim that an objectivistic reconstruction of the logic of situations that is entirely based on observables is a theoretical ideal are right. However, I do not believe that these adherents of Gary Becker type methodologies are pursuing a fruitful line of research if they start out for their methodological ideal in our present state of knowledge -- and perhaps in all conceivable future states. I do not want to engage here in the controversy about whether human beings are a part of the rest of nature or not. I am quite willing to admit that they are. I do not accept, however, the assumption that we are in a position yet to apply our normal scientific instruments that are adequate for the study of other parts of nature (including other living creatures) to the study of human behavior. The careful reconstruction of the logic of situations from the subjective point of view of the actors involved in such situations is the best we can do. We will fail to throw some light on an essential part of human interaction if we treat the human mind as a black box and try to rely on such behavioral laws that directly connect observable stimuli with overt behavioral outcomes.
Within an explanatory approach that allows for intervening theoretical entities like motives subjective reconstructions of the action situation may lead to improved predictions provided that appropriate bridge laws can be found. Though I agree with Buchanan and many philosophers of science that it may not be the only relevant test for the quality of our explanations to scrutinize whether or not they can be symmetrically transformed into predictions it does not seem to be wise to insist that subjectivist economics cannot be helpful as an intermediate step within the process of forming better predictions. There seem to be some quite trivial though fundamental reasons, however, why, regardless of their potential role in predictions, hermeneutic studies of human behavior form a subject of interest in their own right and why they are so persistent in arguments about social theory. To these reasons I will now turn.
Though there are no clear cut borderlines between animals and human beings, the behavior of human individuals is characterized by a much higher degree of plasticity than that of other species. Humans can learn faster and they can adapt more easily to -.he varying circumstances of their environment. They are endowed to a higher extent with what we have come to regard as the higher faculties of the mind Due to imagination, foresight and understanding they can make deliberate forward looking decisions that are based on extended models of the choice situation. Starting from these models they can behave in very complex and often unpredictable ways and thus seem to be "free to choose".
We can leave it as an open question here whether or not human behavior, in the last resort, is completely determined by general laws of nature. Even if freedom of the human will does not reach beyond the bounds of natural laws, and even if all human choices could be predicted in detail by an omniscient being, it still would make sense to distinguish free choices of human beings from choices that are following a fixed behavioral program. The springs of human action are so complex that the prediction of future choices will always remain quite uncertain. Even if the laws of nature should provide a fixed program of human behavior, the program is not known. Due to the complexity of its causes human action will always seem as very flexible and thus "free" even though we may thoroughly believe that it is ultimately governed by the laws of nature in general.
Still, this weak sense of freedom does not really capture the way human individuals perceive themselves. A complex process driven by chance events might be as unpredictable as human behavior but nobody would be inclined to apply the term "free" to it. We do not think that unpredictability of behavior amounts to the same thing as freedom. As human beings we cannot avoid to form a self-image that deviates from the notion of a very complicated natural process or "machine". We are naturally inclined to conceive ourselves as actors who themselves exert control over the choice of perceived alternatives. Forming this self-image we give up the perspective of the external observer of human behavior and take the internal point of view of the behaving individual. We switch between the attitudes of the onlooker to that of the participant.
From the internal point of view of the decision maker flexibility of behavior gains a different quality. Though he may acknowledge that his behavior is the outcome of the operation of the laws of nature and thus in a sense beyond his control as an acting person, he will subjectively feel that he can exert some control over his own actions. He adopts the view of the participant in the game of life who makes decisions and not that of the onlooker who observes them happening (cf. again Strawson 1974 and Baurmann 1987). As a rational person he will go even further and perceive himself as somebody who has full control over the choice of alternatives open to him.
Taking an internal point of view does not preclude the rational actor from adopting the point of view of an external observer. He can adopt both points of view. This may be useful even with respect to his own behavior. For instance, there may be acts that the chooser can control now but not in the future. The so-called Ulysses problem is an example for such a situation. Ulysses decides that he will be bound to the ship because he predicts from the point of view of an external observer of his own behavioral tendencies that he will be better off without full control over his choices in a future situation. It is important to note however that making the decision to be bound to the ship he regards himself as having full control over that choice. In deciding he does not predict anything -- he virtually makes something that at least he himself could not predict in the full sense of that term.
The rational human actor will carefully discriminate between aspects of choice situations that he can influence by his choices and those that he cannot influence. The distinction between what he can and what he cannot influence by his own choices roughly amounts to the distinction between choices and side constraints of choice. If a rational actor as an external observer of his own behavioral technology gets to know part of his behavioral program then he may take it as a constraint of his choices. At the same time, he can react on this knowledge and thus show that the fixed behavioral program is not a fixed program at all or at least he can try to "outsmart" it like Ulysses and alter the side constraints of later decisions by prior ones.
In the case of Ulysses who is virtually tied to the mast the commitment to later behavior is perfect. Ulysses knows in advance how he will act after he has made his "binding decision". He has excluded all later choices by a prior commitment. But, of course, there may be situations that do not allow for such perfect commitments. The decisions that prior selves make, leave some choices open to future selves. A prior self can only bring about deliberate changes in the expected payoffs of future decision makers. In case of such an imperfect commitment, a present self can know what future decisions are going to be only to the extent that present self can predict the reactions of future selves (cf. also commitments and rules).
To accomplish that task, the present self must try to predict the reactions of future selves to changes in side constraints. Here, the present self must necessarily take recourse to the laws of nature that govern human behavior in general. Only insofar as side constraints and future behavior are beyond the control of future selves, a present self can control the behavior of future selves. Only in these respects, future selves are bound by the commitments of the present self.
As long as the rational chooser takes the internal point of view to his decisions, he cannot know how he will decide in the future. Though he might quite well know how he will in general respond to the exigencies of situations he always will perceive himself as one who can act otherwise when he actually has to make a certain decision. From the internal point of view he makes decisions. Making a decision is different from predicting one. One can predict a future decision from an external point of view. But nobody can make a future decision before the time of the decision has come. A specific act can only be performed in the specific action situation.
Though the decision maker can predict his own likely responses to the incentives of a future specific decision situation and can accordingly try to manipulate or to influence these incentives, this will not eliminate his self-perception as a decision maker. If a present self tries to anticipate the decisions of future selves from their internal point of view, the present self engages in a simulated decision making process in which the decision maker has full control over the alternatives and picks one according to his criteria of assessment. These criteria are evaluative, not predictive. Still, even such a procedure could become part of a prediction if the present self who anticipates the evaluation process of the future self from the inside so to say, adds the law of nature that, in general, human beings act according to what their criteria of evaluation recommend as optimal.
The analysis of switching between external and internal point of view may be easily extended to situations in which many individuals interact. Here again two approaches to the analysis of the interaction are possible. On the one hand, we can adopt the objective attitude of the external observer of the behavior of others. The problems we face are strictly problems of prediction. The predictions of other actors' behavior are based on behavioral laws that are beyond the control of these actors. On the other hand, we can analyze what is going on from the internal point of view of various participants.
The relevant question to the behavioral scientist is whether predictions based on a process combining hermeneutic as if decision making with a simple law of nature, will lead to better predictions in general than a process that is based on laws of nature which are directly relating external stimuli to overt reactions. There is a set of simple laws of learning theory that may be used to address such problems. These behavioral laws may under certain circumstances prove quite accurate. However, the antecedent clauses will be so complex and the complete history of the individual decision maker so much unknown that this approach in general will not be too reliable. Otherwise behavioral scientists among the economists would have been much more successful in manipulating human interaction and in stirring it towards some predefined ends than they actually were.
In any case, the thoroughgoing subjectivist presumably must go one step further and, for the sake of consistency, should reject prediction of empirical phenomena altogether as tests of the quality of his specific theories. Contrary to v. Mises' views on the subject the subjectivist then does not claim superior insights stemming from a priori knowledge. From his point of view, the evaluations and the goal directed behavior of man as opposed to rats or other animals are in a principal sense, if only to a certain extend, unpredictable for -- as I should like to, add here -- fellow men. To support this tenet with some arguments, one may point to a classical discussion within the field of jurisprudence. About the turn of the century German legal theorists, philosophers and sociologists of law discussed the question of whether or not fixed legal rules could ever capture fully normative intuitions. These scholars wondered whether rules applying to "cases" that are characterized in general descriptive terms could be so well specified that they would lead to adequate results in every instance. The problem was that in formulating general rules one has to capture evaluations by descriptive terms characterizing the cases to which certain normative consequences will apply. Many of the scholars eventually came to the conclusion that such a general programming of social reactions as incorporated in rules would always lead to normatively unacceptable results in some cases that in principle could not be foreseen at the time of rule enactment. Though they did not come to the conclusion that for this reason general legal rules should not be enacted they were trying to draw attention to the fact that the definiteness of the legal rule, however well it might be specified, always could be gained only at a cost.
For the present discussion the problem of justice involved here is not particularly relevant. However, the thesis that a rule attaching consequences to descriptively characterized cases will in principle lead to normatively inadequate results is. For, should this claim in fact hold good then it would tell us something about the working of the human mind. The future evaluations human actors might apply to social situations and specific precepts of correct action would be unpredictable to a certain extent. In every program specifying the reactions of individuals there may be an implicit clause of the type "except for unforeseeable circumstances". It may be unconceivable ex ante what the exceptions to the rule may be and still, from the normative point of view, it may be pretty obvious ex post that a specific case should be decided contrary to the descriptive content of the rule.
To be more specific, in German law for instance there was a well specified rule defining theft. It was mainly related to the concept of taking away a removable thing from somebody else. Later on, electric power became widely used. Somebody evidently engaged in "theft" of electricity. He took away electric power without paying for it. From a normative point of view it seems to be pretty obvious ex post that this kind of acts should be treated in the same way as thefts of the traditional kind. Still, "stealing power" cannot be described as taking away a removable "thing". So it could not be treated as theft if one wanted to stick to the program laid down in the penal code.
Quite analogous to the example of taking away electric power it may be impossible to construct a set of laws containing descriptive antecedent clauses which are related to some predicted action or other. For, in the last resort individual values, which guide individual choices and value related dispositions may be so complex that they cannot be predicted in all detail. There simply may be no rules which, for every instance that might occur in the course of time, describe the content of those evaluations and intentions that guide individual behavior. Still, emulating the internal point of view of another individual may provide some understanding of the logic of the situation and the likely intentions and preferences of rational individuals in such a situation and thus is part of "what economists should do".
We always look at the social world from both points of view that of the external onlooker and that of the participant. This may not be the full-fledged Kantian dualism of man as guided by the principles of reason and the laws of nature at the same time. However, we might come quite close to such views. Even if rationality, in the last resort, is a natural phenomenon, even if thought processes might be subject to the same laws of nature as other processes and could be understood by an omniscient external observer in all their details without referring to the internal point of view of the thinking individuals and their mental states, even then we are forever locked into the processes of our mind. Therefore we almost necessarily take a deep rooted interest in the workings of our minds including those aspects that are beyond our explanatory capacities. In particular, classical game theory is of great value in this context because it is all about our minds, about reasoning processes rather than about prediction and explanation. In a Non-Rawlsian sense the equilibria of game theory have been reflective equilibria all the time and the study of solutions to games was concerned with a priori theories of rational behavior.
If it should be objected here that only the formation of empirical hypotheses is a legitimate object of science in general, and of economics in particular, I should regard this tenet as quite dogmatic. Non-empirical endeavors can be interesting, can be rationally pursued, and can be highly though only indirectly relevant to what is going on in the world. If one likes to subsume these efforts under the heading of metaphysics -- as for instance: Popper would -- this is alright. But, as Popper, one should be aware then that this kind of metaphysics as an intellectual endeavor may be at least as relevant as empirical science.
Not only our descriptive theories about social reality but also counterfactual theories of what rational behavior is or conceivably might be are part of reality itself. Though nobody may ever behave perfectly rational there should be those, who as a matter of fact, explore what perfectly rational behavior and a world of rational beings may be like. Even though nobody may ever apply these criteria perfectly in social reality, knowledge of the criteria may well influence behavior. Our theories of rational behavior of interaction between rational individuals may indirectly influence what is going on in the world. These theories are not normative in the sense that they develop prescriptions of clever behavior. They rather explore the faculties of individuals' minds under the presumption that there are no restrictions on rationality, on information processing, etc. This is counterfactual and it is speculative. But it is interesting and it is important as part of what one might regard as the human process of self-education in a process of theoretical evolution.
Subjectivist economics and game theory provide frameworks of criteria as well as paradigmatic examples of accepted argument such that speculation may become controlled and accessible for rational criticism. Nobody needs to accept the frameworks of speculation or the rules of these games. There are neither logically nor empirically compelling reasons to do this. Still, we may have good reasons to play rather one than another game of argument about rational interaction. To provide these reasons, however, is another game. In the present clarification of aspects of subjectivism in economic theory and social theory in general, this problem cannot be attacked. Suffices it to note that subjectivist economics may be a non-empirical, non-predictive, non-explanatory, non-normative, partly mathematical study of our self-understanding as rational beings who interact with other rational beings. If somebody might object then that the enterprise is a philosophical one this would not bother me nor should it bother Buchanan. Spelling out the implications of the assumption of purposefully rational behavior of individuals who strategically interact with other rational individuals and who acknowledge that they do is interesting in its own right even if it lacks explanatory and predictive implications.
It is striking how close to Kant's ideas about a world of rational individuals the views of an economist like Buchanan are. It is not only true that a deep respect for the autonomy of the individual drives Kantian political theory as well as economic theory of the Buchanan type. It is also true that both are skeptical about the human faculty to live according to the precepts of reason (cf. Kant's often neglected งง 41-44 in the metaphysics of morals). Still, they both regard it as rational to hope that human beings can improve their own fate by rational considerations. In a sense we are under a moral obligation not to give up our hope to live according to precepts of reason because this may be the only reason to hope that more reasonable sets of rules will eventually emerge and guide our actions. Economic reasoning of the subjectivist variety itself is viewed as one of the forces within the process of improving the human condition.
As Robert Tollison once remarked in a discussion, there are basically two varieties of economists: those who simply look at the world as it is from the point of view of the detached observer and those who hope to make the world a better a place. James Buchanan definitely is of the latter kind. For a subjectivist philosopher economist this somewhat "unrealistic" attitude seems to be perfectly alright. Adopting an objective empirical science attitude he may be skeptical about prospects of improving social and in particular legal institutions. Still, perceiving himself as a participant of what is going on in the world he may feel that any hope for rational improvement presupposes that we first explore what conceivably could be achieved under the presumption of rational behavior. Kant would have characterized this kind of presumption as "transcendental". Economists feel uneasy if their enterprise is characterized by such terms. But they need not quarrel much about terminological issues. They must point out only that contrary to Kant's views the rationality concept underlying subjectivist economics is a strictly instrumental one. Looking for the reason of rules we have to focus on interests and on hypothetical imperatives of prudent behavior in pursuing our interests then. The last paper of this volume will give a critical account of Brennan's and Buchanan's constitutional economics as an interest based theory.
<< References >>
|
Kliemt, Hartmut: Papers on Buchanan and related Subjects. Munich (Accedo Verlagsgesellschaft), 1990: p. 96 122. (Studies in Economics and Social Science (SESS), Vol. 1 (1990)) |
[Top]