post-autistic economics review
Issue no. 15,  1 September 2002
article 5



    issue 15 contents                           PAE Review index                               home page




Towards a Realistic Epistemology for Economics

Claude Mouchot   (Centre Walras, Université Lumière-Lyon2, France)

© Copyright 2002 Claude Mouchot


If there is one point on which all economists would agree, it is that they will never agree.  And of course these disagreements have always existed.  There was a time when we attributed them to the “youthfulness” of our discipline, but today we are forced to admit that economics will remain eternally “youthful”.  It therefore seems critical that we abandon this “youth” fable, as well as stop aiming for the unification of our discourse; a fable and an aim that arise from comparing economics to physics.  That a century after Walras some of us still hold to the same epistemological position as he is, at the very least, surprising.

I wish to propose here a realistic epistemology for the “science of economics” that hopefully will enable us to explain our perpetual disagreements and thus the reasons why economics is pluralist.


In order to avoid the temptation to state the argument in terms of physics, let us begin by considering  in the context of economics the following key sentence from Thomas Kuhn.

Normal science . . . is predicated on the assumption that the scientific community knows what the world is like. (1970, p. 5)


One needs only to carefully re-read this sentence to realize that it will never be applicable to the world of economics nor to the social world in general.  Economists will never agree on what the economic world is like, and partly because of what such an agreement would signify.  It would mean that the ideological oppositions that have always run through the conceptions of man and society, the great visions of the world -- individualism / holism, liberalism / socialism, etc. -- would have disappeared.  Now each of us is aware that repression of these oppositions is possible only by totalitarian means, and, moreover, the worst of their kind; a totalitarianism that, with no form of explicit violence, would prohibit even the very thought of the term ‘opposite’, given that the very notion of opposition would be nonexistent!


Therefore, since these oppositions are a part of society, it follows that economics will never be a “normal science” in T. S Kuhn’s sense of the words.  The unification of economic theories will never be achieved, at least not in a democratic society.  Hence it is necessary to abandon all reference to physics and to make a new effort to work out the epistemological status of our discipline.

Economics is a Totalité


Let us begin by considering two “classical” definitions of economics.

1.       Economics is the study of humanity in the ordinary affairs of everyday life.  (A. Marshall)

Economics is the science that studies human behaviour in relation to ends and scarce resources having alternative uses. (L. Robbins)


It is clear that both definitions define not just economics but also the entire social setting.  The first definition includes, friendship, fatigue, pain, power, prestige, etc. and for many people even war.  The second definition is no less general as it pertains to all finalized action, ( Godelier pg. 19 – 20), and, moreover, is at the very foundation of the totalitarianism mentioned above.


But if on the other had one tries to narrow these definitions by in Marshall’s substituting “economic” for “ordinary” and in Robbin’s substituting “economic ends” for “ends”, then the definitions dissolve in obvious circularity.


The impossibility of separating the economic from the social, and the circularity of the definitions, ( economics is economics) which result when this impossibility is ignored, illustrate the unfeasibility of defining economics. This fact can be summed up by one word: economics is a Totalité.  Defining an object, involves distinguishing it from other aspects of reality, and one cannot distinguish an aspect of reality, if not from within the whole of which it is a part. The whole, on other hand, cannot be distinguished except to say that is the whole!  Let us insist on the lack of a definition for the word ‘whole’.  We have said that it sums up the problems posed by the proposed definitions for economics, but it does not resolves them.  It only illustrates the impossibility of definition in this case.


Here we have more or less reached a cul-de-sac: economics is economics, and not much can be added to that.  But of course it does not end here, as the existence of numerous economic theories proves.


Constructing Scientific Domains


“a science sets out to adequately delimit the problems likely to define a field of research and on which a consensus can be reached……” (Piaget 1970, p 41)

This simple sentence supports our proposition that “economics is not a science” and explains it further: given that economics remains undefined, it cannot as a whole be considered and accepted as a science.  How then, should we go about effectively delimiting these problems so as to be able to constitute a domain that can be studied scientifically?  My answer takes the form of identifying a necessary and sufficient condition:

In order to develop a scientific discourse in economics, it is necessary and sufficient to privilege certain aspects of this Totalité, to distinguish the aspects that we wish to study, in other words, to define the object of study.


This condition is sufficient.  For example, if I define my object as the study of how an individual allocates his income among different expenditure alternatives, I end up with the marginalist’s consumer equilibrium theory.  Or if I define it as the study of the conditions under which the division of income between wages and profits is made in a capitalist economy, I end up with a production cost theory, either neo-Ricardien or Marxist, depending on what further hypothesis are made.


This condition is also necessary.  It follows from the demonstrated nature of Totalité.  Economics is economics, and nothing more can be said about it without specifying the discourse, that is, without prioritising certain aspects of this whole.  All positive economic discourse is simply the expression of a certain point of view on economic reality, a point of view that consists of defining an object within a whole and constructing for this object a scientific theory.


All the major approaches to economics involve the construction of a scientific object focused on a particular aspect of the whole. Therefore, classical, neoclassical, Marxist, and Keynesian theories, are sciences highlighting certain aspects of economic reality, while equally neglecting others. We may say that each of these sciences “cuts out” its object within the whole of the social and economic reality. This can be represented diagrammatically as follows:



                                                            Classical Economics



                                                                                          Keynesian Economics



  Economics as a Whole



                                                                                 Neoclassical Economics

                                    Marxist Economics



Each theory neglects certain aspects of reality:


  • The Keynesian theory has no means of explaining value.


  • Classical and Marxist theories fail to account for market prices (apart from what can be explained from simple common sense according to which market prices depend on the interaction between supply and demand).


  • “this theory [neoclassical] lacks a simple answer to the question: Why are salaries and profits what they are? This is an interesting issue when we are considering the allocation of revenue amongst social classes, and these social classes are not variables that can be explained by the neoclassical theory.” (Hahn 1972)


  • The phenomena of power - more or less taken into account by Keynes (Real demand is the actual power held by entrepreneurs to reach an equilibrium characterised by underemployment.) and by Marx (the struggle of classes) - do not exist in the classical theory and are explicitly rejected by the neoclassical theory (all agents are equal).

Finally, one should note that each of these theories prioritises certain aspects of reality at the expense of others, constitutig ipso facto an ideological characteristic: partisans of these theories did not choose them for ‘objective’ reasons in the natural sciences’ meaning of this word, rather they chose them because they reflected the political values that they themselves were in favour of.


You will have noticed that in the preceding diagram, all the theories intersect at the same point, this being the economic situation under consideration.  Furthermore, it is my contention that these theories have one characteristic in common: although they privilege different logics, each describes an aspect of every economic situation.  For example, all the theories have elements for explaining the current unemployment condition in France.  Is it not E. Malinvaud who thinks this unemployment is partly classical and partly Keynesian?


To accept this is also to accept that the logic of a theory ends where that of another begins.  Consequently, if we prioritise one logic over another and insist on pushing it to its limits, we end up with absurdities, in theory as well as in practice.  For examples:


  • It is impossible to reject the consumer’s utility maximisation hypothesis because in a market economy the consumer’s equilibrium clearly represents an aspect of his/her behaviour.  But when pushed to its theoretical limits it logically leads to a general equilibrium which, however, can never exist in economic reality.  And if pushed to its practical limits, it gives rise paranoid financial markets.
  • It is impossible to reject J. M Keynes logic on effective demand. But if pushed to its limits it would give rise to a constant increase in inflation rates, as well as to the return of hyper– liberalism.
  • It is impossible to reject the Marxist logic according to which the structure of the system of production, the relations of production that determine “the places and functions in which the individuals are but only the occupants”.  The French production system has a shortage of over two million places and functions.  But pushed to the limit, this logic theoretically gives rise to an eternal structure and to what in practice would amount to communism.


Thus, each theory illuminates only one aspect of reality, and each theory should allow a place for other theories to succeed which will illuminate other aspects of reality.  The challenge then becomes the harmonisation of these numerous theories.



Finding a common structure for these discourses


The impossibility of a logical harmonization


Talking about the logical harmonization of two or more theories, means synthesizing them into one main theory that encompasses them all. Within this main theory, the constituent theories would be considered as special cases, arising for instance, within certain specified parameters. However, such a harmonization is not possible.


This impossibility has been historically proven: the multiplicity of economic theories is itself a form of concrete proof.  If their logical harmonisation were possible, we would at least expect that some of the theories would have been unified. This impossibility is also inscribed within the very idea that we are proposing, i.e., that of economics being a Whole within which each theory constructs its very own scientific object. The results obtained will no doubt depend on the constructed object, and there is no reason whatsoever why they should hold true for the objects of other theories.


The possibility of having contradictory results is, indeed, very high. A good example is that of excess supply in the market. Before the Keynesian period, economists claimed that lowering prices was all it took to attain market equilibrium, be it the market of a particular good or the global market. But although true in the former case (micro), it sometimes proved false in the latter case (macro), even if each producer behaved “rationally” by lowering the price. This is an example of how sometimes decisions taken by a group of rational individuals can lead to global irrationality.


Diverging scientific objects, disjointed theories, partial discourses, disharmonious and at times contradictory logic -- despite all these, we still have to continue. And so to explain and understand the economic reality as well as act on it, it is necessary for us to harmonize the different discourses in question. Given that logical harmonization is impossible, we need to find other means of harmonizing these theories.


A reasonable harmonisation


Given that we already have access to an abundant literature on the current economic reality, as well as the desire to act on it, how then can we harmonize the numerous different theories in the absence of all other forms of unification? 


All economic policy seeks to modify reality and, in principle, to modify it in a precise direction with a view to realizing the particular values -- solidarity, justice, equality, etc. -- that each individual regards as “good reasons” for acting.  Of course it goes without saying that wanting something and the ability to have it are two different things. Therefore, we still need to find mechanisms that will enable us to create a “better” reality (with respect to the values we hold); hence the necessity of a coherent analysis of the situation to be modified.


It is precisely at this point that the multiplicity of theories poses a problem: faced with different explanations, it is evident that one explanation and only one has to be chosen over the rest. If not, the actions taken may prove totally ineffective, like accelerating and breaking simultaneously. So what should be the criteria for this choice?


Let’s review what was said earlier.  In principle what we are saying is that whatever the concrete economic problem under consideration (unemployment for instance), it can be explained within a number of different logical structures, belonging to different theories.  Consequently, the choice of criteria although easy in the abstract may be difficult to implement in reality.  The politician considers and attempts to understand today’s dominate logic as it pertains to the real world situation under consideration, so as to decide according to that logic the measures to be undertaken.  But the currently dominate theory may not fit the current situation.  For example,  the current talk in favour of supply-side oriented policies aimed at increasing investment makes no sense today in France where both personal saving and the self-financing of business enterprises are very high.


Good reasons, careful consideration, decision making, choice of policy, temporary conclusions (because tomorrow, today’s conclusions will be revised as the dominate logic will not be the same); this set of elements defines what ‘reason’ means, way beyond and above simple rationality, and takes into full account individual freedom which, though limited, is part of the reality that enables man to confront reality.


I have tried to clarify the reasons as to why economists perpetually differ in opinion.  I have noted that each economist chooses the theory that supports the values that (s)he holds. It therefore follows that these differences of opinion among economists are only the reflections of the underlying political dispute.  Our discipline, being far removed from the scientific status of the natural sciences and whose limits have been revealed, ought to be re-named “Political Economics”.





Hahn, Frank (1972), The Shares of Wages in National Income : an Inquiry into the Theory of Distribution, London, Weidenfeld and Nicolson.

Kuhn, Thomas (1970), The Structure of Scientific Revolutions.  Chicago: University of Chicago Press.

Mouchot, Claude (1996), Méthodologie économique, Paris, Hachette.

Piaget, J. (1970), Épistémologie des sciences de l’homme, Paris, Gallimard.




Claude Mouchot, “Towards a Realistic Epistemology for Economics”, post-autistic economics review, issue no. 15, September 4, 2002, article 5,