New Individualistic Foundations for Economics

23
New Individualistic Foundations for Economics Author(s): Alan Nelson Source: Noûs, Vol. 20, No. 4 (Dec., 1986), pp. 469-490 Published by: Wiley Stable URL: http://www.jstor.org/stable/2214980 . Accessed: 14/06/2014 03:41 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. . Wiley is collaborating with JSTOR to digitize, preserve and extend access to Noûs. http://www.jstor.org This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AM All use subject to JSTOR Terms and Conditions

Transcript of New Individualistic Foundations for Economics

Page 1: New Individualistic Foundations for Economics

New Individualistic Foundations for EconomicsAuthor(s): Alan NelsonSource: Noûs, Vol. 20, No. 4 (Dec., 1986), pp. 469-490Published by: WileyStable URL: http://www.jstor.org/stable/2214980 .

Accessed: 14/06/2014 03:41

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Wiley is collaborating with JSTOR to digitize, preserve and extend access to Noûs.

http://www.jstor.org

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 2: New Individualistic Foundations for Economics

New Individualistic Foundations for Economics ALAN NELSON

UNIVERSITY OF CALIFORNIA, LOS ANGELES

I

Philosophical puzzles arise when attempting to understand the con- nection between explanations of individual behavior and explana- tions of group behavior in the social sciences. These puzzles have engendered a long debate over whether explanations of social behavior must terminate with an account of individual behavior (Methodological Individualism) or whether social behavior is best explained by means of properties of groups that are not understood in terms of the individuals making up the groups (Methodological Holism). In this paper, I shall examine a particularly vexing puzzle of this kind.

Modern economic theory, on the face of it, appears to be in close accord with Methodological Individualism. The theory can be thought of as very roughly divided into two branches, microeconomics and macroeconomics. Microeconomics can be provisionally regarded as dealing with the economic choices of small scale economic en- tities (like persons, households, and firms) and also with the prop- erties of markets for rather specific commodities (like avocados, bicycles, and copper). In virtue of its subject matter, then, microeconomics is naturally regarded as individualistic. Macroeconomics is, provisionally, the study of large scale economic phenomena, particularly those involving properties of broadly aggregated economic entities (phenomena such as the dynamics of the general price level and general level of employment, and also

NOUS 20 (1986): 469-490 ?C1986 by Nou's Publications

469

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 3: New Individualistic Foundations for Economics

470 NOUS

entities such as national governments and whole sectors of producers and consumers). Macroeconomics, then, is not so obviously in- dividualistic. It is, however, usually considered desirable to have microeconomics and macroeconomic theories that are mutually com- patible. More particularly, it seems desirable to be able to provide macroeconomic "foundations" for macroeconomics, and considerable research effort has been expended in attempts to do this. The reason is clear: most large scale economic phenomena are (this is an on- tological 'are') just the net effects of small scale economic phenomena. This popular view of the relationship between microeconomics and macroeconomics strongly suggests that the entire science of economics rests upon individualistic foundations.

This neat picture has been disrupted by strong arguments to the effect that even microeconomics is not an individualistic science. Some have argued that microeconomics ought to conform to Methodological Individualism, but that it does not and, hence, is not even an empirical science (e.g. Alexander Rosenberg (1983)). Others have argued (most notably Milton Friedman, e.g., (1953 and 1976)), that microeconomics should be understood as being about phenomena "in the large". According to this common and often presupposed view, the theory's apparent references to individual behavior are mere conveniences and whether seeming claims about individuals are "realistic" or not is a relatively unimportant question.

In this paper, I shall defend the thesis that modern macroeconomic theory is, at least potentially, a theory of the economic choices of individual economic agents. If this project succeeds, then making macroeconomics compatible with Methodological In- dividualism will be a reasonable goal for research into the micro- foundations of macroeconomics. After all, if microeconomics serves as a foundation for macroeconomics, then (by "transitivity of foun- dations"), the whole of economics is founded by providing in- dividualistic foundations for the former.2 I think that the arguments presented in this paper can be extended in obvious ways to issues involving individualism in other social sciences, but doing so must be left for another occasion. I shall, however, have a good deal to say about parallels between philosophical interpretations of economics and philosophical interpretations of linguistics.

Let us begin by roughly stating the individualistic attitude towards microeconomics as follows:

(1) Microeconomics is (at least in part) a science of the behavior of individual economic agents.

I think it is now possible to reconstruct the considerations that have

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 4: New Individualistic Foundations for Economics

FOUNDATIONS FOR ECONOMICS 471

led to the rejection of (1) quite succinctly. There is considerable tension between (1) and three other propositions that are widely held:

(2) Microeconomics is an empirical science.

(3) Good empirical science is able to predict and explain the phenomena in its domain.

(4) Microeconomics cannot (or, perhaps, "does not try to") explain or predict the behavior of individuals.

Most commentators have thought that the only reasonable way to resolve the tension among (t)-(4) is to abandon (1). An example of this would be replacing (1) with the claim that microeconomics is about entities like average individuals or representative individuals.

In what follows, I want to show how it should be possible to maintain robust belief in (1) and to show what further involvements this belief would entail. Any successful attempt to do this will have to show how microeconomics can be a good empirical theory, given that it is uncontroversial that the theory is presently quite useless for even potentially describing the economic behavior of particular individual agents in the way that classical physics can provide an account of the particular billiard balls on the table in the student union. In other words, it must be shown that the tension among propositions (1)-(4) can be dissipated by rejecting (4). I try to accomplish that by showing how microeconomics could become an accurate predictor of individual behavior. I argue further that in- creasing the scope of the theory in this way requires removing what have been perceived as methodological obstacles in the way of close connections between microeconomics and the psychology of in- dividuals. My strategy for attacking these obstacles is to draw analogies between interpretive issues in linguistics and in microeconomics and to exploit the closeness of the analogies to apply well known work in the philosophy of linguistics to economics.

Thus, my conclusions are conditional in nature: if our under- standing of the connection between linguistics and psychology enables linguistics to be a potentially successful empirical science of individual linguistic behavior, then the same should be true of microeconomics. It should be at least possible that microeconomics be a successful empirical theory of individual economic behavior. I think that in- teresting arguments for the plausibility of this conditional claim will constitute a contribution to our understanding of economics and to some general issues in the philosophy of science even if its antece- dent turns out to be false.

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 5: New Individualistic Foundations for Economics

472 NOUS

The structure of the rest of this paper is as follows. In Part II, I argue that we do in fact want microeconomics to be, in part, a theory about individuals. I think that these arguments may have become necessary because of the prevalence of defeatist attitudes in which microeconomics is taken to fail to be an empirical science, or to be a kind of disguised account of economic phenomena "in the large". Part III lays out a program for ameliorating the dif- ficulties with interpreting microeconomics as a science of individual behavior. It will be seen that an advantage of this approach is that it involves neither overhauling the theory nor restricting its scope, unlike other proposals that have been made. It calls instead for sup- plementing existing theory and methodology with psychology. Part IV defends the new approach against philosophically inspired ob- jections and Part V makes some claims about the significance of the preceding material. Part VI is a brief summary and prospectus.

II

Although it seems transparently clear that an individualistic inter- pretation of macroeconomic theory is the natural, prima facie correct one, it is possible to marshal three positive arguments for it.

Argument 1-Systematic expositions by a wide variety of philosophers and economists almost invariably state quite explicitly that the subjects of the laws (or models) of the theory are individuals. Explanations for the economic behavior of individuals are of patent intrinsic interest-in the end, all economic choices and transactions involve individuals or very small groups. Moreover, when markets or other kinds of aggregates over individuals are to be treated by the theory, the relationships of interest among aggregates are, in principle, derivable from facts about relationships among individuals that are taken to be antecedently understood. The program of methodological individualism is an explanatory ideal. Holists do not always deny this claim; many of them believe instead that in- dividualistic theories turn out to be poor predictors, have little explanatory power, etc. That is, they believe that these theories turn out to be false theories, as a matter of empirical fact. These ob- vious considerations are made to seem nearly senseless on the sup- position that microeconomics is about aggregates in the first place.

Argument 2-The last argument had it that the theory of individual economic behavior is of intrinsic interest. One indication of this fact is that a popular (though certainly not dominant) way of inter- preting the theory is to regard it as a normative3 account of ra- tional behavior. On this view, the basic assumptions of the theory are a priori truths about rationality. For example, part of being ra-

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 6: New Individualistic Foundations for Economics

FOUNDATIONS FOR ECONOMICS 473

tional just is to have transitive preferences. Advocates of this view (its most outstanding exponent was probably von Mises; see, for instance, his (1949)), often go even farther and consider microeconomics to be synthetic a priori. We are supposed to discover these properties of rationality, now taken to be synthetic, by in- trospective investigation. Indeed, if microeconomics does have nor- mative force, then it seems that this further leap is mandatory. Other- wise microeconomics could all be done in the head and the em- pirical, "synthetic" part of the theory would be largely otiose. A purely a priori science (if such things be sciences) could make use of empirical data only insofar as it had heuristic value in directing a prior investigation.

There do not seem to be any really convincing arguments that we can carry matters as far as the unattractive commitment to the existence of a synthetic a priori science, and even if we could, it is not clear how we could be entirely convinced that theorems deduced from macroeconomic assumptions would be properly taken to have normative force. But there is some pull to the observation that microeconomics deals with behavior that is "good" in some limited sense. I think we are inclined to say, for instance, that if we are faced with a choice problem in a supermarket, then given all the relevant information, the choice dictated by macroeconomic theory will be a good one. Speaking for myself, I am even inclined to say that if I had my calculator and was in no hurry, I would try to decide by solving the mathematical problem the theory used to represent the choice problem.

None of this proves anything, but it does indicate that taking facts about individuals as irrelevant except insofar as they are of use in understanding aggregate phenomena, is an implausibly strong position. The burden of proof weighs heavily on those who impugn the status of individual agents in the theory.

Argument 3-A third reason for regarding individual agents as the objects of central interest in microeconomics is that economists generally seek individualist explanations. If the primary purpose of microeconomics was simply to predict large-scale phenomena, e.g., demand curves for a single commodity in a large market, it seems the natural way to proceed would be to fit a theory to data for the variables relevant to these phenomena. Why deal with individuals at all? Friedman alludes to the simplicity, aesthetic appeal, heuristic value, etc. of accounts of large-scale phenomena that are derived from treatments of small-scale phenomena, but this is not convinc- ing. Surely, the reason economists are not satisfied with correla- tions between quantities in the large (when these are even available)

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 7: New Individualistic Foundations for Economics

474 NOUS

is the same reason that scientists in general have for being dissatisfied with brute correlations. It is the principle of the common cause; we tend to believe that strong correlations between quantities in any domain in which a science is applicable exist because of an underlying common cause that serves as a satisfying explanation. I think that it is reasonable to suppose that economists strive to derive regularities occurring at the macro level from those at the micro level because the latter phenomena are considered more fun- damental and more easily understood (see Nelson (1984) for an extended treatment).

Thus, phenomena "in the large" are to be "explained" as resulting from "deliberate acts of choice" on the parts of individuals. This terminology appears in Friedman's work (1976, p. 35) alongside his instrumentalistic view stated above. The remainder of the im- portant chapter in Friedman's book from which these passages are drawn ("Theory of Demand") continues with the standard analysis of individual choice in terms of utility maximization. Philosophical gymnastics are required here to avoid saying that the most impor- tant theoretical constructs in microeconomics pertain to individuals- and the exercise is never described in print. Consequently, it is curious that Friedman and those influenced by him resist the natural position that microeconomics is an empirical science of individual behavior.4

The cause of their predicament is not hard to find. Microeconomics is, prima facie, about individuals, but despite its inability to account for the economic behavior of real individuals, it has virtues that are unique among the social sciences. It is unsur- passed in the generality and power of its mathematical techniques, in its scope of applicability (some even claim that some day all social science will be dominated by macroeconomic theory and its methods), and in its ability to make useful predictions about large groups of individuals. There are only two ways to resolve this perplexity. First, one could attempt to explain why microeconomics does not work for real individuals and, even better, suggest how to remedy this defect. That is what I shall attempt. Or,, one could attempt to argue that microeconomics is not really about individuals despite appearances. On this line of Friedman's, real understanding of economic phenomena requires the reinterpretation of statements that seem to refer to individuals.

The arguments employed by Friedman and subsequent writers are always difficult to follow; one suspects that most advocates of the reinterpretation view arrive at their destination via some short- cut. The kernel of the basic argument is this. Microeconomics'

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 8: New Individualistic Foundations for Economics

FOUNDATIONS FOR ECONOMICS 475

statements about individuals are false, but that is no problem. It is permissible for the "assumptions" of a theory to be false as long as its predictions are accurate. In the case of microeconomics, the predictions about groups of individuals are good so those are what the theory is really about. Friedman (1953) contends that all suc- cessful scientific theories, including physical theories, employ false assumptions and properly so. It has seemed to many commentators that Friedman's arguments rely on strictures that are positivistic in the worst sense of the term and on improperly drawn analogies. I agree and have little to add to the objections that others have raised (Rosenberg (1976) has an especially careful critique).

Still, even if the considerations usually urged in favor of the anti-individualist position are no good, this position might be ad- vocated as the only possibility for dissolving the conundrum posed by the joint acceptance of propositions (t)-(4). Therefore, let us now proceed with the positive project of maintaining the natural, individualistic attitude towards the subject matter of microeconomics.

III

The approach I shall defend begins by taking the pronouncements of macroeconomic theory at face value. When the theory says, for example, that the consumer has an ordinal utility function (or near- ly equivalently, technical niceties aside: a consistent set of preferences) and that his economic behavior can be deduced from this function and some boundary conditions, let us interpret this as meaning that "having" the function is part of the consumer's physical, or biological, or psychological structure. The claim that overt economic behavior is deducible from the utility function and some boundary conditions can then be interpreted as meaning that the utility func- tion plays a causal role in the etiology of the behavior. It follows that completely describing a bit of economic behavior in this way is now going to implicitly contain some psychological consequences.5 This entailment comes from the fact that the economic description of the behavior will supervene on psychological (and physical, etc.) descriptions of the same behavior.

More precisely, I want to consider the consequences of basing macroeconomic theory on a claim like:

(5) Utility functions are psychologically real in the sense that, (i) given appropriate idealizations, making a rational economic choice requires the agent's psychologically accessing the utility function as well as the economically relevant in- formation concerning prices, the budget constraint, etc.; (ii)

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 9: New Individualistic Foundations for Economics

476 NOUS

the utility function is then maximized subject to the con- straints set by this information; (iii) the agent's choice is determined by the result of this process.

There are a number of points that must be made about (5). First, how much of the process described in (5) is going to be con- sciously carried out by the agent is to be determined empirically. It is possible, in fact likely, that very little of this process is going to be introspectable. The analogy between (5) and what happens in human language use is quite strong.6 It is well known that there is overwhelming evidence to the effect that routine linguistic tasks require a great deal of processing which is not done consciously and there is no a priori reason that economic choice should be any different.

Second, the idealizations that are "appropriate" will also pro- bably have to be determined as we go along. The idealizations which turn out to result in the best theory (where the "best theory" will be chosen on the basis of the usual scientific criteria) will be the ones that are appropriate.

Third, the procedure whereby the agent "maximizes" his utility function may or may not be the mathematical technique that the economist uses to maximize utility functions. On the face of it, it seems implausible to require that all agents who make economic choices can do calculus consciously or otherwise, but we do not know until we investigate the matter. Similarly, there is no reason to decide a priori whether it is appropriate to idealize the maximizing process of economic agents as involving the use of calculus should it turn out that there is no way they really do use it.

Finally, there is some latitude available in the manner in which the outcome of this process is taken to "determine" the subsequent choice of the agent. Whether or not it will be fruitful to consider other factors which contribute to the choice as part of the theory is an open question. Once we have interpreted macroeconomic theory in this way, it becomes possible to characterize the broad outlines of a general research program for microeconomics. Let us call this research program psychoeconomics.

The principle advantage to be expected from psychoeconomics is that it makes it possible to empirically specify the mathematical forms that utility functions take thereby giving empirical content to the theory as a whole. Some of the best arguments to the effect that microeconomics is not presently an empirical science (as found, e.g., in Rosenberg (1983)) are thus circumvented. The way the func- tional forms are to be specified will be similar to the way in which many "empirical laws" (the things made familiar in the writings

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 10: New Individualistic Foundations for Economics

FOUNDATIONS FOR ECONOMICS 477

of Carnap and other philosophers of science) have supposedly been determined. Empirical laws like Hooke's Law for springs appear to be the results of inductive generalizations (although it is well known that more subtle considerations than mere curve-fitting are commonly involved).7 Similarly, the psychoeconomist will hypothesize a func- tional form for an individual's utility function. In the beginning this will no doubt be a very simple first approximation which is easy to work with. The investigator is likely to include only a few arguments in the function and assume that any other arguments will have effects that are small enough to be ignored in a first ap- proximation. For example something mathematically convenient (like this Cobb-Douglas utility function) may be chosen for an individual economic agent, i:

(6) U, =culog z1 + (1 -c) log Z2

where U. is the amount of utility, ae is a constant, z1 is the amount of food consumed, and Z2 is the amount of shelter con- sumed.8 Then, using ordinary econometric techniques, the scientist can estimate the values of the a.. The hypothesized function (6) can then be tested in routine fashion against data obtained for real economic agents and actual values of zi and Z2. In the likely even- tuality of poor agreement between the predictions of (6) and the observations, our scientist can either introduce auxiliary hypotheses to account for the discrepancy or (6) can be discarded. If the latter option is chosen, the parameters of (6) can be altered, or the z's appearing in (6) can be modified or added to, or the functional form of (6) can be changed, etc. Just about anything can turn out to be helpful in guiding the formation of hypotheses, both original- ly and in revision. None of this should seem unusual-it is what we all learn in an introduction to scientific method. After this pro- cedure is repeated many times we would expect to come up with utility functions which did an acceptable job of covering the data for given individuals and of predicting the results of future ex- periments on these individuals. If these expectations were not ful- filled, then it would be prudent to suspect that something was radical- ly wrong with the entire approach. We might even become con- vinced that postulating utility functions was not a very good start in the construction of a scientific theory of economic choice. (Of course, even more radical measures would be available).

It would be inappropriate here to try to place any a priori restric- tions on the kinds of empirical evidence which will turn out to be relevant. It may prove fruitful to run actual controlled experiments

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 11: New Individualistic Foundations for Economics

478 NOUS

on individuals in a laboratory setting, or to perform less carefully controlled experiments on individuals in the marketplace. It seems likely that introspective reports will be of some value; there are many legitimate ways of eliciting such reports, as is well known to psychologists. It is also possible that data about averages that are reconstituted from studies of aggregative phenomena will prove useful. Which of these or other procedures will prove most fruitful can usually be decided only by trying them all and finding out directly.

It is worth saying explicitly that none of this implies that it is the job of the psychoeconomist to determine the exact form of the utility function for every individual in the economy. It is to be hoped that there are significant generalizations to be made over individuals. Ideally, we hope to determine what properties of utility functions are "universal" properties which all individual utility functions share. For instance, it would be very interesting to find that all utility func- tions were Cobb-Douglas in seven important commodities, or even in seven different groups of commodities grouped into "kinds". (Similarly, it is interesting that all bodies are subject to universal gravitation.) The more of these universal properties there are, the more powerful the explanations provided by the theory will be. Moreover, a good way to check a proposed "universal utility func- tion" would be to come up with a lot of individual functions and check to see whether the proposed universal function really has features common to all the individuals. Analogously, chemistry texts have nothing to say about the chemical structure of the rock serving as a paperweight on my desk, but it had better be the case that a chemist could take that very rock and show beyond reasonable doubt that it was composed of nothing other than some of the 100 or so elements, if called upon to do so.

I think that the correct conclusion to draw from these considera- tions is this: it is plausible that the ground breaking work in con- structing macroeconomic theory should consist of experimentally determining regularities obeyed by individuals. We can then hope to generalize these results and obtain simpler and more abstract regularities that all relevant agents obey. It could then be expected that these generalizations, in turn, will explain the regularities exhibited in the behavior of particular individuals in a manner quite common in successful sciences. Consider, for example, the way that the Bohr model of the atom explains the law for the Balmer series by suggesting a causal mechanism and providing a mathematical technique for making predictions.

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 12: New Individualistic Foundations for Economics

FOUNDATIONS FOR ECONOMICS 479

IV

It should be apparent that the kind of macroeconomic research pro- gram described above is potentially fruitful. Moreover, the kinds of investigations which are proposed are quite similar to the kind of work that gets done in well established sciences like physics. That ought to weigh heavily with many economists who think it impor- tant to count their discipline as being on a methodological par with respectable, well established sciences. Given this, it is puzzling that many of these economists steadfastly adhere to an interpretation of the theory which explicitly rules out any possible contributions from psychology. I think it will be useful to attempt to solve this puzzle by trying to say what it is about the proposed research pro- gram that economists find so objectionable. It will then be possible to argue my case further by showing that their perceived fears are unsubstantiated.

The most important potential objection to the implementation of psychoeconomics is that it makes microeconomics too vulnerable to the vicissitudes of empirical discovery. In its usual formulations, the macroeconomic theory of the consumer is characterized by a number of assumptions or axioms concerning utility functions (or preferences). These assumptions often include the transitivity, quasi- concavity, continuity, etc. of utility functions. If it turned out that these assumptions were persistently and flagrantly violated by the phenomena, we would probably have to conclude that the theory was false. It is important to have some qualifiers such as 'persistently' and 'flagrantly' here because, of course, all these theories are going to have to do a good deal of idealizing if they are to help us to understand the world and a science like microeconomics which deals with human nature will probably have to do more idealizing than most. But even the most cursory, pre-theoretical inspection of economic behavior reveals that the foundational assumptions of microeconomics are persistently and flagrantly violated by the actual behavior of individual agents.9

If we believe that utility functions are more than convenient fictions, that people actually "have" them, then it is highly likely that the mathematical forms that are attributed to them are going to be closely constrained by actual economic behavior because of the role they play in the causation of that behavior. Hence, psychoeconomiscs requires that the standard macroeconomic assump- tions be answerable to a wide range of potentially observable phenomena in a more direct way than is usually supposed. The unavoidable conclusion is that if a) the possibility of employing

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 13: New Individualistic Foundations for Economics

480 NOUS

psychoeconomics is countenanced and b) its employment yields deter- minations of utility functions that violate deeply entrenched presup- positions of the theory, then c) those presuppositions should be regarded as having suffered some disconfirmation. To many, c) will sound like the denial of the possibility of microeconomics altogether. But we shall see that microeconomics is still possible.

Now let us see how to reconcile the implementation of psychoeconomics with these facts. Since the problem is supposed to arise from the inevitable disconfirmation of these principles by commonplace facts about human behavior, what is needed by way of defense is a well motivated means of accommodating the apparent disconfirmations. A defense against nearly identical charges has already been clearly articulated for another science which interacts significantly with psychology, namely linguistics. In modern transfor- mational linguistic theory, we can find ourselves in a situation quite analogous to the one in which the above examination of macroeconomic theory has landed us.

A successful linguistic theory of this kind would do two things. First, it would "cover" an important body of data by generating all and only the grammatical sentences of a language.'0 Second, it would provide an explanation of these and other data by assign- ing a structural description to each sentence that would help account for phenomena like syntactic and semantic ambiguity, intuitions of sentence similarity, complexity, etc. Very schematically, these theories work by positing both constructs that generate phrase structures for sentences and constructs that perform operations on these phrase structures. It is common to regard these constructs as internally represented by language users because this lays the foundations for the building of bridges between the linguistic theory and related theories, e.g., theories of language production and understanding in psycholinguistics, and of language acquisition in cognitive psychology.

But when we turn to the corpus of linguistic data, the theory runs into potential problems in the accounts of both language pro- duction and language understanding. Most humans produce a very substantial number of utterances that cannot be generated by the grammar, i.e. ungrammatical sentences. In fact, there are un- doubtedly a great many linguistic agents (perhaps it is most of them) who produce mostly ungrammatical sentences in normal speech. On the other hand, it is well known that there are many classes of sentences that are grammatical, but are, nevertheless, virtually impossible for any human to understand under normal circumstances.

The method that linguists have developed to deal with this group of problems is well known. A distinction is drawn between the

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 14: New Individualistic Foundations for Economics

FOUNDATIONS FOR ECONOMICS 481

linguistic competence of agents and the linguistic performance of agents. There has been much dispute over the legitimacy of this distinc- tion, but I think that it is now clear there is at least one perfectly acceptable and useful way of making it. I shall take a little space to do so because I want to scrupulously circumscribe and avoid the problems and confusion that arise when the claims being made leave even the most tenuous foothold for misinterpretation. I do not take any credit for this exposition; it seems to me to be quite explicit in any sensible treatment (e.g., Fodor (1981)).

In linguistics, a theory about a speaker/hearer's competence is a treatment of those aspects of his behavior (or potential behavior) explained by properties of a particular psychological entity, a grammar. A grammar is, in this sense, a structure which is said to be internalized by the agent." Examples of the kinds of phenomena treated of in such a theory include the ability of an agent to poten- tially produce or understand an infinite number of novel sentences despite having a finite mind, and the ability to classify sentences as grammatical or ungrammatical with exceptions only in a relatively insignificant number of anomalous cases.

A theory about an agent's performance is a treatment of his actual behavior. It explains this behavior as resulting from the in- teraction of the agent's competence with mechanisms postulated by other psychological theories that deal with how the competence is used. A performance theory, therefore, will need to include an account of at least some of the processes that eventuate in overt behavior. In the best case, a performance theory would enable us to understand those processes all the way from their genesis (which causally involves internal representations) to their culmination in overt behavior.

It is critically important to appreciate that although the com- petence/performance distinction is used to make philosophical and methodological points, attempts to draw it must be empirically in- formed. Whether the distinction can be made, and whether it will be useful if it can be made, depends on how the facts turn out. There also need not be any sense in which a performance theory is prior to a competence theory or vice versa. We can arrive at a competence theory by looking for generalizations and abstracting away from complications present in a performance theory or we can work in the other direction and regard behavior conflicting with a competence theory as the result of interference from lower level factors which can get accounted for in a performance theory. It is probably most realistic to expect these kinds of theories to develop simultaneously with advances on one front feeding back into the other. Accordingly, there is no a priori reason to rule out potential

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 15: New Individualistic Foundations for Economics

482 NOUS

contributions to linguistics (or economics) from highly performance oriented fields like artificial intelligence or from highly competence oriented fields like formal decision theory.

The kind of theoretical framework being described enables linguists to investigate the purely formal properties of the abstract grammars without constantly having to worry about apparent discon- firmation by actual behavior. The investigator working on perfor- mance theory (probably a psycholinguist) attempts to handle these by connecting the abstract competence theory with a more general psychological theory dealing with interfering factors like familiar- ity, limited memory, and the necessity of computing in real time. It may also be possible or necessary or desirable to include con- siderations like motivation, beliefs, attention, etc.

It will not be possible, however, to include too many things in this second group or linguistic behavior will no longer be sharply distinguished from any other kind of behavior and a special theory of linguistic behavior will not be possible. Wittgenstein can be in- terpreted as claiming that this is in fact the case and language must be studied not just in the context of a general psychological theory (which he also took to be impossible), but in the context of very broad anthropological considerations. Similarly, an economic per- formance theory cannot be expected to account for absolutely everything that might influence economic behavior. Just one exam- ple is that although nearly any imaginable commodity can give an economic agent some (perhaps negative) amount of utility, the theory will be virtually useless if all of these things appear essentially as arguments in the utility function. The reason is that, in that case, it will prove impossible to make reasonable hypotheses concerning the overall structure of the functions because of Duhem's problem and because of the human impossibility of investigating all the effects of the different kinds of commodities. The sort of economic theory we are considering will be interesting only if it is possible to per- form some aggregation of commodities into "economic natural kinds". It had better be possible, for instance, to treat an avocado salad consumed at 12:17 p.m. and an avocado salad consumed at 12:18 p.m. as both being instances of the commodity "avocado salad consumed in the afternoon", or even better as "salad", or better still as just "food". Whether a Wittgensteinian argument can be sensibly advanced against utility theory is, I take it, an empirical matter that can only be finally answered by producing a successful theory or by showing that a successful theory cannot be produced.

I am now going to argue that microeconomics can benefit from work that has been done by philosophers and linguists; it ought

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 16: New Individualistic Foundations for Economics

FOUNDATIONS FOR ECONOMICS 483

to be possible to connect a theory of macroeconomic performance with a more general theory of the competence of rational agents to choose in an economic setting. Economists can study the im- plications of having certain kinds of utility functions in different economic environments working either completely abstractly or with actual utility functions determined with the aid of psychoeconomists. Moreover, they need not fear its turning out to be the case that any of the functions so obtained are empirically false, or even just inaccurate generalizations of the real functions. Discrepancies be- tween the competence of economic agents based upon their ideal ability to maximize the utility functions which we now assume they really do possess, and the actual performance of these agents in the marketplace (or even in the laboratory) can be accounted for in terms of interfering factors which are irrelevant to abstract macroeconomic theory. It can be hypothesized that this interference will eventually be satisfactorily explained as resulting from things like misinformation, imperfect memory, limited calculating abilities, etc. It does not much matter whether the theory that explains the interference gets called microeconomics or psychology or anything else. Another way of putting this is that standard microeconomics can be construed as a partial theory of the economic performance of agents-a theory of the ability of an agent to rationally choose when freed of more or less irrelevant impediments (again, it must be stressed that exactly which impediments get classified as "irrele- vant" will be determined a posteriori by the best available theory and that it is impossible to contrive circumstances in which all of the impediments are actually removed). Earlier I argued that the standard theory can be, at best, a partial theory because it does not make any predictions worth noting until it is augmented with some psychologically determined specifications of the functional rela- tionships that it contains.12 The analogy with linguistic theory can be made more striking by noting that a theory of an agent's economic competence can be interpreted as a theory about the behavior of an idealized economic agent, the rational being called homo economicus, that has been the source of much debate and apparent confusion in the economics literature. Similarly, linguistic theories of the com- petence of linguistic agents are often said to be theories about the hypothetical behavior of an "ideal speaker/hearer". In each case, the ideal behavior is quite different from actual behavior.

Another strong analogy emerges when we consider that linguists and philosophers often argue that the proper data base to be accounted for by a competence theory is linguistic intuitions. The ability to account for this kind of data is sometimes said to provide

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 17: New Individualistic Foundations for Economics

484 NOUS

"internal" evidence for the theory. Thus, on this view, a sentence's grammaticality is not to be determined by observing uncontrolled linguistic behavior to see whether the sentence is uttered or not. Instead, a subject is confronted with the sentence by an experimenter and asked to answer "yes" or "no" to the question "Is this sentence grammatical?" The subject is encouraged to consider his answer carefully resorting, if necessary, to pencil and paper or other helpful devices. The same strategy might be assayed by those interested in economic theories of competence. The appropriate data base might be taken to be experimental subjects' well considered intuitions of economic rationality in given choice contexts. For example, if one wants to see some actual cases of rational choice one should not observe behavior in the marketplace, but instead should present sub- jects with choices in a more idealized laboratory setting. Subjects can be allowed to make the choice in whatever manner they see fit even if this involves something like their estimating their own utility functions and maximizing them by solving a calculus pro- blem. This outlook seems especially attractive in economics insofar as microeconomics is supposed to include a general theory of ra- tional choice.'3 Surely one constraint on such a theory is that its results conform to a satisfactory extent with our pretheoretical in- tuitions of what counts as being economically rational. This would seem to indicate that there is at least some basis for thinking it sound macroeconomic methodology to avoid straying too far from pretheoretical intuitions.'4

For present concerns, the important thing about the state of the art in economics and in psychology is that the performance theories required to complement the theory of economic competence are at least potentially available. Much work has been done attemp- ting to constrain performance theories with competence theories, and vice versa, in linguistics (see, e.g., Berwick and Weinburg (1984), and Fodor, Garrett, Walker and Parkes (1980), and the citations therein) and some psychologists are at least guardedly optimistic about the prospects for psychological (i.e. performance) theories of rational choice that utilize maximizing strategies (see, e.g., Rachlin, Battalio, Kagel and Green (1981)). Naturally, if it turns out that such theories are not forthcoming, then serious reconsiderations will be in order, but we are far from being able to fully evaluate the potential of this kind of psychological theorizing.

V

Some restraints on the correct application of psychoeconomics should be noted. We must be careful not to push the competence/perfor-

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 18: New Individualistic Foundations for Economics

FOUNDATIONS FOR ECONOMICS 485

mance distinction too far. It is tempting to assume that the con- siderations reviewed above show that linguistics (or microeconomics) is free to completely ignore behavioral data after all. One might argue as follows:15

Why not make the theory as abstract as we like? We are still free to suppose that the abstract entities which are postulated are im- plicated in the causation of actual behavior, but at a very abstract level. It is, in fact, quite impossible for this very abstract theory to be disconfirmed empirically because we can always assume that the psychological mechanisms mediating the causal chain are extremely complicated. It is the job of a performance theory to spell out just how the constructs chosen by the theorists of the abstract figure in the causation of behavior.

It is not difficult to show that this position is too strong ifthe empirical nature of the science in question is to be retained. We can grant that well established theoretical entities can be extremely resistant to disconfirmation when there is good reason to believe in the ultimate success of a performance theory which explains away the troublesome cases. Still, we can readily imagine cases in which low level behavioral data could have an evidential bearing on even the most abstract theoretical entities. Suppose that we have two grammars which make exactly the same predictions. If one of these grammars is very easily embedded in a well-confirmed learning theory (a performance theory with respect to the grammar) and the other resists attempts to fit it into other existing theories we obviously have good grounds for preferring the more compatible theory. Or consider this convincing polemic (presented in Fodor (1981)). Imagine a history in which a Ptolemaic astronomer had responded to the discovery of Venu- sian phases by claiming that astronomy is only concerned with abstract entities which account for planetary positions. This astronomer could have claimed that information about planetary phases was irrelevant to theories meant to account for the positional properties of these abstract constructs and ought to be accounted for in a "performance" theory.

Not many astronomers would have been convinced by this argu- ment; a heliocentric theory that accounted for planetary positions and phases was, everything else being equal, simply a better theory. Similarly, if a linguistic construct is postulated that is, on the face of it, incompatible with behavioral data while there is an alternative construct which does the same job but is compatible with the data, then that would seem to constitute a powerful piece of evidence for the compatible construct. Such evidence can be found. Some is presented against the psychological reality of definitional seman-

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 19: New Individualistic Foundations for Economics

486 NOUS

tic representations and for something like meaning postulates (Fodor, Garrett, Walker and Parkes (1980), Fodor, Fodor, and Garrett (1975), and Martin (1978)).16

There does not seem to be any philosophical reason why this kind of evidence should not be effective in economics as well. For instance, the standard theory claims that all utility functions are quasi-concave. But if this turned out to be incompatible with behavioral data and there was no performance theory available to account for this, but there was a convincing way of showing how a class of utility functions which violated this constraint meshed with a well-confirmed psychological theory, then it seems this state of affairs would have to count against the proposition that utility func- tions are quasi-concave. It may not count decisively, but we can imagine how this kind of evidence might be multiplied until the case was decisive, especially in the presence of a competitor that conformed to the constraint. It is worth stressing again that confor- mity with the data is not the only criterion for theory choice. There may be cases in which some accuracy can be exchanged for simplicity, confirmability, etc.

The upshot of this discussion is that the basic assumptions of macroeconomic theory are, after all, subject to disconfirmation. Nevertheless, something has been gained for economics by invok- ing the lessons from the philosophy of linguistics. Its basic assump- tions can be well insulated from disconfirmation without complete- ly closing all the windows of vulnerability. This should not bother anyone very much. If the assumptions employed are good ones, i.e. if economic agents have utility functions which really do con- form to the assumptions, then it is highly unlikely that there will be any other descriptions of utility functions which pair up with convincing performance theories to give better predictions of the actual behavior of agents.

Now the main conclusions argued for in this paper can be summarized.

(i) Economists have resisted (or not even considered) augment- ing standard macroeconomic theory with psychoeconomics because it appears to render absolutely basic presuppositions of the theory, for example that agents always make choices that maximize their utility functions, sitting ducks for empirical disconfirmation.

(ii) A similar situation obtains in linguistics, but linguists have, by and large, not completely scorned contacts with psychology. They understand the connection between the two sciences in terms of a competence/performance distinction. This enables them to mediate explanations of linguistic behavior in terms of abstract linguistic con-

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 20: New Individualistic Foundations for Economics

FOUNDATIONS FOR ECONOMICS 487

structs with less abstract accounts of how agents actually use these constructs to produce the behavior.

(iii) The same solution to the problem in microeconomics looks like a potentially valuable option. It could enable us to find room for psychoeconomic research into the mathematical forms of realistically interpreted human utility functions, and for the study of idealized features of these utility functions (traditional macroeconomic theory). It will be up to the theory of economic per- formance to account for deviations from the behavior that we expect given the form of a well-confirmed individual utility function.

(iv) The basic presuppositions of macroeconomic theory are susceptible to empirical disconfirmation. I hope that is has become clear, however, that this is no cause for immediate concern. On the contrary, it should be a reason for considerable optimism. It turns out that these presuppositions need not be a priori dicta con- cerning the way that the data must be read by economists. They can have empirical content and this means that questions concern- ing their a priori methodological validity need not arise. Since they are at the core of the theory, however, they will be quite well in- sulated against empirical disconfirmation. Should the phenomena prove to be so recalcitrant that our confidence in these basic prin- ciples becomes shaken, then perhaps it will be time to shop around for a radically different kind of theory of macroeconomic behavior, but such a horror story seems like little more than a story at this point. And finally,

(v) In its present form, microeconomics does not look like an em- pirical science partly concerned with individual behavior. Its lack of a genuine psychoeconomic treatment of individuals makes it look more like a branch of mathematics. 7 This defect (assuming, as many economists do, that it is a defect) can be removed through the prac- tice of psychoeconomics. The relationship between psychoeconomics and the ordinary theory is at least as sound in principle as the rela- tionship between psycholinguistics and linguistics. Furthermore, pro- ceeding in this way would make it possible to preserve the con- siderable achievements of microeconomics, quite unlike the sugges- tions made by many other critics of the theory.

VI

I shall conclude by emphasizing that this paper defends the cogency of a program for developing microeconomics. It cannot be argued that this program must succeed, or even that it is likely to succeed. That is good. I have tried to show how the theory can be an em- pirical theory of individual behavior; empirical theories are, by

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 21: New Individualistic Foundations for Economics

488 NOUS

nature, falsifiable. I do, however, think that the approach that has been presented is cause for some optimism, and that it is the only approach that has any chance of retaining the bulk of current macroeconomic theory as an empirical science of individual behavior.

If the program I have defended turns out to be an empirical failure, it would still have important and instructive consequences. First, with respect to economics, such a failure would provide strong evidence, perhaps overwhelming evidence, against the possibility of individualistic foundations for microeconomics and, hence, for economics in general. Since economics is such a highly developed social science, this would in turn create a very strong presumption against the possibility of individualistic foundations for any other social science. Second, with respect to other sciences employing a competence/performance distinction, the failure of the program of psychoeconomics would throw the remaining problems into sharp relief. The close analogies between economics and, for example, linguistics might give us reason to expect this kind of theorizing to fail in linguistics as well.18

This kind of speculation, however, concerns the future course of scientific progress and it is, therefore, outside the scope of philosophy.

NOTES

'Colloquia partly based on the material in this paper were held at the following in- stitutions: University of Illinois at Chicago, Northwestern University, University of Pittsburgh, University of Vermont, UCLA, UC San Diego, and the University of Southern California. In each instance, I benefitted from the discussion. I am especially grateful to Paul Teller for detailed written comments.

2The precise nature of the relationship between macroeconomic and macroeconomic phenomena and their associated theories is, however, the source of many difficult philosophical problems that are not wholly relevant to present concerns; see Rosenberg (1976) for a general discussion and Nelson (1984) for details.

3Here I am not using 'normative' to indicate the presence of moral considerations as economists sometimes do; I mean to imply a value judgment only in the sense that decision theorists do when they refer to their work as a "normative" theory of human choice behavior.

4Friedman (1976, p. 33) is thereby led to advocate the abandonment of the distinction between micro- and macroeconomics in favor of a distinction between price theory and monetary theory, but that position is more idiosyncratic and not widely adopted by economists.

5One interesting point that cannot be discussed here is that economists often, as a matter of deliberate choice, explicitly rule out consideration of the interface between economics and psychology. Here, I shall restrict myself to the observation that the kinds of theorizing that even these economists engage in overstep the bounds of isolationism towards philosophy. This behavior is not peculiar to economists. Philosophers have often noted divergence between interpretive pronouncements and theoretical practice in the work of physicists and psychologists as well. The writers of texts in these sciences will often espouse operationalism in the in- troduction and then begin chapter one with a discussion of how instrumental apparati often fail to accurately measure real magnitudes.

6My statement of (5) is actually based on a similar proposition that appears in an article on the interpretation of linguistic theory (Fodor, Fodor, and Garrett, 1975). Much will be made of this analogy in what follows.

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 22: New Individualistic Foundations for Economics

FOUNDATIONS FOR ECONOMICS 489

7Empirical laws also tend to become rationalized by being derived from "theoretical laws". In this respect, empirical laws are probably not like utility functions; the latter are not going to be "derived" from anything except, perhaps, very general psychological laws (which are not on the scientific horizon).

8Another kind of idealization is going on here insofar as commodities like food and shelter are aggregates of many specific commodities. Perhaps the actual utility function that we attempt to approximate will have an argument for every commodity in the consumer's choice set, although this possibility is fraught with some obvious problems.

9Quirk and Saposnik (1968) note that it is not possible to preserve as much as the minimum shreds" of transitivity without essentially incorporating stochastic elements in

the theory and even then, transitivity is violated. See also the discussion in Cohen (1981) and the accompanying commentary. Cohen's paper is useful because it deals in a nontechnical way with some of the important experimental literature like the famous work done by Kahneman and Tversky. I should point out that Cohen's conclusions are very different from mine despite some terminological similarities.

Incidentally, the conclusions I draw in this paper are meant to apply to macroeconomic treatments of behavior under risk and uncertainty though it would make the exposition too tedious to make these applications explicit.

10I am going to try hard to skirt controversial matters in the interpretation of linguistic theory. I think the points I want to make are not sensitive to how these controversies are settled and, furthermore, I am only concerned to show that whatever interpretation of linguistic theory turns out to be correct will also be available to interpreters of economic theory.

"'Internalized' means something like mentally represented. For expository purposes, this whole paper assumes that the notion of mental representation is scientifically respectable. If this assumption turns out to be false, then this paper still makes the interesting point that microeconomics, on the most reasonable interpretation, will suffer from the kinds of shortcomings that contemporary linguistic theory suffers from. Much of the confusion spawned by the competence/performance distinction results from preconceived objections to mental representation. I do not propose to defend this well entrenched part of modern cognitive psychology here. I shall mention, however, that it has proven especially counterproductive to refer to mentally represented grammars as "tacitly known". This terminology should be avoided.

'2Moreover, predictions about what humans will actually do (as opposed to just specifying what it would be rational to choose) will require at the very least the services of a limited performance theory. At the very least, these predictions will require enough of a perfor- mance theory to show us how to get from the results of a person's exercising his competence to his "yes" or "no" responses to actual choices. There are no direct means of accessing an agent's competence. This is, perhaps, one root of Quinean arguments against the cogen- cy of mentalistic linguistic theory.

'3This raises again the question of a normative interpretation of microeconomics. Should the theory be regarded as an account of the economic behavior agents ought to exhibit if they are to be rational? This paper is neutral on that question, but I do think that an affir- mative answer would need to rely on a psychologistic analysis of rationality and such analyses are all presently subject to grave difficulties. The same should be true of attempts to apply the competence/performance distinction (or the notion of "implicit theories") when theoriz- ing about moral judgments.

'4An analogous consideration motivates much of the current discussion of the degree of significance that "folk psychology" has for psychological science.

'5This is a somewhat simplified version of the argument to be found in Katz (1981, pp. 83-92 and 95-114).

'6It is encouraging that these three studies employ distinct experimental paradigms thus testifying to the fecundity of the general framework. Fodor, Fodor and Garrett (1975) relies mostly on reaction times, Martin (1978) on commonsense judgments of complexity, and Fodor, Garrett, Walker and Parkes (1980) on judgments of semantic relatedness. I see no reason why psychoeconomists should not be clever enough to devise similarly diverse means for subjecting theories of economic competence to the rigors of empirical testing, although I have not been clever enough to propose any. Some work of this kind is being done and

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions

Page 23: New Individualistic Foundations for Economics

490 NOUS

reported on in the journals, but it is too early to judge its significance to the theses put forth in this paper.

"This explains why Rosenberg's arguments (1983) that it is a branch of mathematics seems to carry so much weight.

18A potential disanalogy that might save this kind of linguistics in the face of the failure of psychoeconomics is that there may be a linguistic faculty that is modular even if there is no such faculty for economic choice. See Fodor (1983) for a review of the relevant psychological literature. It would seem that the kind of competence theories I have discussed are theories about modular faculties. Therefore, if I am right, the fate of empirical faculty psychology is entwined with interpretive questions that are of great philosophical interest.

REFERENCES

Battalio, R., Kagel, J., Rachlin, H., and Green, L. (1981) "Commodity Choice Behavior with Pigeons as Subjects," The Journal of Political Economy. 89:67-91.

Berwick, R. and Weinberg, A. (1984). Grammatical Constraints on Linguistic Performance, (Cam- bridge: MIT Press).

Cohen, LJ. (1981) "Can Human Rationality be Experimentally Disconfirmed?" 7he Behavioral and Brain Sciences. 4:317-330.

Fodor, J. A. (1981) " Introduction: Some Notes on What Linguistics is About, " in N. Block (ed.), Readings in Philosophy of Psychology, Vol. 2 (Cambridge: Harvard University Press: 197-207).

Fodor, J.A. (1983) The Modularity of Mind. (Cambridge: MIT Press). Fodor, J., Bever, T., and Garrett, M. (1974) The Psychology of Language. (New York:

McGraw-Hill). Fodor, J., Garrett, M., Walker, E., and Parkes, C. (1980) "Against Definitions," Cognition.

8:263-368. Fodor, J. D., Fodor, J A., and Garrett, M. F. (1975) "The Psychological Unreality of Semantic

Representations," Linguistic Inquiry, 6:515-531. Friedman, M. (1953) "The Methodology of Positive Economics," in Essays in Positive Economics.

(Chicago: University of Chicago Press). Friedman, M. (1976) Price Theory. (Chicago: Aldine). Katz, J. (1981) Language and Other Abstract Objects. (Totowa: Rowman and Littlefield). Martin, E. (1978) "The Psychological Unreality of Quantificational Semantics," in C.W.

Savage (ed.), Minnesota Studies in the Philosophy of Science, Vol. IX. (Minneapolis: Univer- sity of Minnesota Press).

Nelson, A. (1984) "Some Issues Surrounding the Reduction of Macroeconomics to Microeconomics," Philosophy of Science. 51:573-594.

Quirk, J. and Saposnik, R. (1968) Introduction to General Equilibrium Theory and Welfare Economics. (New York: McGraw-Hill).

Rachlin, H., Battalio, R., Kagel, J., and Green, L. (1981) "Maximization in Behavioral Psychology," The Behavioral and Brain Sciences 4:371-388.

Rosenberg, A. (1976) Microeconomic Laws. (Pittsburgh: University of Pittsburgh Press). Rosenberg, A. (1983) "If Economics Isn't Science, What is It?", Philosophical Forum.

14:296-314. Von Mises, L. (1949) Human Action: A Treatise onEconomics. (London: Hodge).

This content downloaded from 188.72.126.118 on Sat, 14 Jun 2014 03:41:45 AMAll use subject to JSTOR Terms and Conditions