Theory-Construction in Economic Science

  • Bernard Hodgson
Part of the Studies in Economic Ethics and Philosophy book series (SEEP)


An inquiry into the logic of theory-construction in economics portends considerably more than an exercise in logic. Formal perplexities in this science have had substantive effects. More specifically, confusion concerning the epistemic status of their explanatory hypotheses and the scientific method for validating such statements has frequently led economists to formulate and accept unsound theories. A critical examination of this methodological confusion among economists would appear, therefore, to be a prerequisite for an appreciation of the explanatory force of economic theories in general. Such an investigation is the intent of this chapter.


Income Expense Sine Milton 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    See, in this respect, J. Buchanan, “Is Economics the Science of Choice?” in J. Buchanan, What Should Economists Do? (Indianapolis, Ind.: Liberty Press, 1970), pp. 39–63, L. Robbins, An Essay on the Nature and Significance of Economic Sciences, 2nd edn. (London: Macmillan, 1935), pp. 73–75, and V. Walsh and H. Gram, Classical and Neo-Classical Theories of General Equilibrium: Historical Origins and Mathematical Structure, (Oxford: Oxford University Press, 1980), pp. 264–65.Google Scholar
  2. 2.
    See, for example, W. Baumol, Economic Theory and Operations Analysis, 4th edn. (Englewood Cliffs, N.J.: Prentice-Hall, 1977), Chap. 9, secs. 6–11, for a lucid presentation of this form of the theory of choice.Google Scholar
  3. 3.
    In the standard mathematical modelling of consumer choice theory, a further axiom asserting the “continuity” of the preference relation is included. Formally, for any bundle X, if we define A(x) as the least as good as X set and B(x) as the no better than X set, then A(x) and B(x) are closed. In verbal terms, for any two goods x, y in a bundle, X, by reducing x incrementally and increasing y in similar fashion, we can identify another bundle Y, which is indifferent to X. Such continuity of preferences is necessary in order to mathematically represent a preference ordering by an adequate utility function. However, as the continuity assumption is primarily of formal technical significance, rather than of fundamental empirical or normative importance, we shall not further address its content or implications in this study. (Similar comments apply to the frequent assertion of a “reflexiveness” axiom ¡ª i.e., for any commodity-bundle, X, X is indifferent to itself, X.)Google Scholar
  4. 4.
    See C. G. Hempel, Aspects of Scientific Explanation (New York: Free Press, 1966), esp. Ch. 12, pp. 333–76, for a canonical analysis of deductive-nomological models of explanation.Google Scholar
  5. 5.
    See W. Baumol, Economic Theory and Operations Analysis, Ch. 9, sec. 7, for an explanation of the derivation of consumer equilibrium within CCT.Google Scholar
  6. 6.
    In chapters 11, 12 and 13, we shall also examine certain implications of CCT’s inclusion within general equilibrium theory and its extension to a theory of choice under conditions of uncertainty.Google Scholar
  7. 7.
    Cliffs, N.J.: Prentice-Hall, 1966), p. 59.Google Scholar
  8. 8.
    For general critiques of the empirical adequacy and normative force of the transitivity or consistency axiom see P. Anand, “Are the Preference Axioms Really Rational”, Theory and Decision, Vol. 23, 1987, pp. 189–214, sec. 3, and R. Sugden, “Why be Consistent? A Critical Analysis of Consistency Requirements in Choice Theory”, Economica, Vol. 52, 1985, pp. 167–84.Google Scholar
  9. 9.
    v See D. Davidson, J. C. C. McKinsey and P. Suppes, “Outline of a Formal Theory of Value”, Philosophy of Science, Vol. 22, 1955, pp. 140f.Google Scholar
  10. 10.
    P. Anand, “Are the Preference Axioms Really Rational?” secs. 3.4, 3.5 and 3.6.Google Scholar
  11. 11.
    See W. Baumol, Economic Theory and Operations Analysis, p. 189, for the explanation behind this conclusion.Google Scholar
  12. 12.
    See F. Machlup, “Equilibrium and Disequilibrium”, for a general review of the use of the concept of equilibrium in economic theory in his Essays in Economic Semantics (Englewood Cliffs, N.J.: Prentice-Hall, 1963), pp. 43–72. And see F. Hahn, On the Notion of Equilibrium in Economics (Cambridge: Cambridge University, 1973).Google Scholar
  13. 13.
    It should not be assumed here that we are agreeing with the neo-classicist choice of a simple, precise concept of equilibrium in preference to factual accuracy, nor with the uses to which this concept is put in CCT, including its normative deployment. Indeed, we will take serious issue with this concept below, especially in Chapter 7.Google Scholar
  14. 14.
    See, for example, R. Rudner, Philosophy of Social Science, pp. 59–63. For a moderate view of the limited explanatory power of an important idealized theory, see D. M. Hausman “Are general equilibrium theories explanatory?” in D. M. Hausman (ed.), The Philosophy of Economics: An Anthology (Cambridge: Cambridge University Press, 1984), pp. 344–59. Hausman’s general methodological perspective is more fully elucidated in his recent book, The Inexact and Separate Science of Economics (Cambridge: Cambridge University Press, 1992).Google Scholar
  15. 15.
    See C. G. Hempel, Aspects of Scientific Explanation, p. 248, who labels this criterion his R3 condition of adequacy for the components of a scientific explanation.Google Scholar
  16. 16.
    s is indifferent between A and B = df. not (S prefers A to B) and not (S prefers B to A).Google Scholar
  17. 17.
    Presentation of a dispositional analysis of wanting and believing would take us somewhat afield. For such an analysis of the former concept see the illuminating discussion by R. Brandt and J. Kim, “Wants as Explanations of Actions”, Journal of Philosophy, Vol. LX, 1963, pp. 425–35. For an investigation of a dispositional construal of belief, along with the attendant difficulties involved, see, for example, J. J. Leach, “Explanation and Value Neutrality”, British Journal for the Philosophy of Science, Vol. 19, 1968, pp. 93–108, sec. 4. The concept of knowledge in CCT, it should be noted, is to be defined in terms of belief.Google Scholar
  18. 18.
    R. Carnap, “Testability and Meaning”, Philosophy of Science, Vol. 3, 1936, pp. 419–71, secs. 8–10.Google Scholar
  19. 19.
    It is imperative to notice that (neo-classical) economists explicitly define the choice of X in terms of, or identify the choice of X with the actual observable purchasing of X. In other words, economists deal with what one might call external choice rather than with the more traditional philosophical construal of choice as an internal, mental event which functions as an antecedent condition to overt action. This point will be further discussed in Chap. 4.Google Scholar
  20. 20.
    Rudner, Philosophy of Social Science, p. 58.Google Scholar
  21. 21.
    Hutchison, The Significance and Basic Postulates of Economic Theory, p. 41.Google Scholar
  22. 22.
    Of course, the option remains for economic theorists to forego the attempts, via idealizing assumptions, to formulate completely universal laws, by adopting the methodological alternative of employing statistical-probabilistic generalizations as explanatory principles. In that event, an inductive-statistical model of economic explanation might be attempted wherein the explanans, in virtue of its probabilistic laws, provides good, inductive support for the explanandum. However, neo-classical economists, at least in the construction of theories of choice, have generally demurred from presenting their explanatory laws in probabilistic form, preferring, as we have seen, when confronted with recalcitrant empirical evidence, to construct universal generalizations for empirically unactualizable “ideal cases”, which hypotheses usually incorporate “ceteris paribus” clauses in their formulations. Neo-classical theorists have defended this strategy by claiming a preference for the superior simplicity of generalizations of universal form rather than for the increase in descriptive truth that would accrue if such generalizations were stated in probabilistic form. One should observe here that even the von Neumann-Morgenstern theory of choice under conditions of risk (J. von Neumann and O. Morgenstern, Theories of Games and Economic Behaviour, 2nd edn. (Princeton, N.J.: Princeton University Press, 1947), employ strictly universal generalizations. Consider, for instance, the rationality principle of this theory asserting roughly that all rational consumers choose that course of action which maximizes their expected utility. It is clear that the probability factor in such an hypothesis is located in the meaning of the term “expected utility” (utility of the outcome weighted by its probability), not in the form of the principle, which remains universal. For a general discussion of logical features of expected utility theory, see chapter 12 below.Google Scholar
  23. 23.
    In this case, G would violate Hempel’s R4 criterion of “factual correctness” for the constituents of a scientific explanation. See his Aspects of Scientific Explanation, pp. 248–49.Google Scholar
  24. 24.
    Antecedent condition (5), requiring that an agent be psychologically and physically capable of performing a certain action, plays a central explanatory role. This condition is especially crucial in securing a universal form for the statement of G’; for the comprehensiveness of (5) enables it to render innocuous numerous putative counter-examples to the truth of G’. One should emphasize, in particular, that (5) includes the factor that action A be physically possible for S to perform. In other words, such external, non-subjective factors of an agent’s situation as the interference of other agents and the obstructions of the physical environment are prevented by condition (5) from furnishing counter-examples to G’.Google Scholar
  25. 25.
    I realize that the meaning and tenability of the analytic-synthetic distinction has been subjected to considerable controversy, especially since the work of W. V. O. Quine (see “Two Dogmas of Empiricism”, in his From a Logical Point of View, 2nd edn. (Cambridge, Mass.: Harvard University Press, 1961), Chap. 2, pp. 2046.) The assumption of such a distinction will, however, facilitate our philosophical analysis of CCT and it is hoped that the meaning of the concepts “analytic” and “synthetic” in our investigation will be clear from the context in which they appear.Google Scholar
  26. 26.
    For a different analysis of the “analyticity” or empirical testability of economic theories, reaching the conclusion that neo-classical economics should be classified as an unfalsifiable division of applied mathematics, see Alexander Rosenberg, “If Economics is not a Science, what is it?” Philosophical Forum, Vol. 14, Summer, 1983, pp. 296–314. Rosenberg returns to this topic in Chapter 8 of Economics ¡ª Mathematical Politics or Science of Diminishing Returns (Chicago: University of Chicago Press, 1992), pp. 228–54.Google Scholar
  27. 27.
    Seep. 14 and note 19 above.Google Scholar
  28. 28.
    “Equivalent counterpart” in the sense that the content of the axiom implies that it is functioning as a particular application of condition (1) of G’ to economic phenomena. Hence we are not using “equivalence” in the logical sense of mutual implication.Google Scholar
  29. 29.
    I say “completes” rather than constitutes a counterpart condition for we might alternatively construe the knowledge constraint of CCT as being a component of an ability condition. Lack of such knowledge would, as it were, function as a psychological impediment to choice, thus failing to satisfy the consumer freedom condition. One might also want to incorporate the condition of adequate income under the umbrella of consumer freedom as lack of sufficient financial resources would obviously function as an external, objective obstacle to choice. In sum, the presence of consumer freedom comprehensively entails that the action (particular purchase) is both psychologically and physically possible for the agent to perform. Nevertheless, I have sometimes presented the knowledge and income constraints as though they were distinct to that of freedom for the reason that such a division is usually employed by economists in formulating CCT, although, strictly speaking, they can be included within a comprehensive definition of consumer freedom.Google Scholar
  30. 30.
    L. von Mises, Human Action: A Treatise on Economics, (New Haven: Yale University Press, 1949), p. 858.Google Scholar
  31. 31.
    For this interpretation, see especially L. von Mises, The Epistemological Problems of Economics (Princeton, N.J.: Van Nostrand, 1960), Ch. I.Google Scholar
  32. 32.
    Hutchison, The Significance and Basic Postulates of Economic Theory, p. 27.Google Scholar
  33. 33.
    Ibid., Chap. 2, sec. 3.Google Scholar
  34. 34.
    L. Robbins, An Essay on the Nature and Significance of Economic Science, pp. 7879.Google Scholar
  35. 35.
    Ibid., pp. 86–87, these “inner experiences” being, for Robbins, the “valuations of the individual”.Google Scholar
  36. 36.
    Lange, “The Scope and Method of Economics”, in The Review of Economic Studies, Vol. XIII, 1945–46, pp. 20ff.Google Scholar
  37. 37.
    For Rosenberg’s original empiricist perspective see his Microeconomic Laws: A Philosophical Analysis (Pittsburg, University of Pittsburg Press, 1976); for his later view of neo-classical theory as applied mathematics see his “If Economics is not a Science, what is it?” Philosophical Forum, Vol. 14, Summer, 1983, pp. 296–314, and his Economics ¡ª Mathematical Politics or Science of Diminishing Returns, (Chicago: University of Chicago Press, 1992), Chap. 8, pp. 228–54.Google Scholar
  38. 38.
    See M. Friedman, “The Methodology of Positive Economics”, in his Essays in Positive Economics, (Chicago, Ill.: University of Chicago Press, 1953), pp. 3–43. The critical literature on the position expressed in Friedman’s article is now voluminous. Rather than adding to this volume with a direct appraisal of my own, I have preferred to let the development of my own views on the methodology of economics stand in contrast to those of Friedman. In this context, Chapters 2, 10, 11 and 12 of this study are especially relevant. And for a perceptive assessment of recent views on the general issue of “realism”, in economics, see D. M. Hausman, “Problems with Realism in Economics,” Economics and Philosophy, Vol. 14, No. 2, 1998, pp. 185–213.Google Scholar
  39. 39.
    A place for contingency in the actual existence of particular states-of-affairs within a theoretical framework whose basic explanatory principles are necessarily true a priori is a theme stressed even by classical rationalist philosophers. See, for example, G. W. Leibniz, “Letters to Queen Charlotte of Prussia” (1702), in Leibniz Selections, ed. P. P. Weiner (New York: Charles Scribner’s Sons, 1951), pp. 363–64.Google Scholar
  40. 40.
    Hutchison, The Significance and Basic Postulates of Economic Theory, p. 34.Google Scholar
  41. 41.
    Ibid., p. 162.Google Scholar
  42. 42.
    See especially W.V.O. Quine, “Two Dogmas of Empiricism”.Google Scholar
  43. 43.
    The epistemological roots of this sort of a priori principle can be found in the neoKantian pragmatism of philosophers such as C. I. Lewis. See, for example, his Mind and the World Order (New York: Dover, 1929), Chap. 8. For a more recent and precise explication see the work of W. Sellars, for example, “Is there a Synthetic a Priori?”, in his Science, Perception and Reality (New York: Humanities Press, 1963), pp. 298–320, or, Science and Metaphysics (London: Routledge and Kegan Paul, 1968), Chap. IV, pp. 91–115.Google Scholar
  44. 44.
    Admittedly, the meaning of the notion expressed by such phrases as “conceptual scheme”, “conceptual structure”, etc. has remained notoriously vague in recent philosophical literature. However, it is intended that the illustrations examined herein, for instance, the action-theoretic framework for interpreting human behaviour, and Newtonian mechanics as a scheme for conceptualizing the motion of material bodies, will furnish sufficient indication of the meaning of “conceptual scheme”, etc.Google Scholar
  45. 45.
    Conceptual truths“ should be distinguished from ”logical truths“ or even ”strictly analytic truths“. For logical truths are better confined to substitution instances of the axioms and theorems of formal logic, while strictly or ”merely analytic“ statements are those which can be translated into such substitution instances with the help of appropriate definitions. Rather, our conceptual truths G’ and G” are cases of what we have called relativized synthetic a priori principles. See R. Brandt and J. Kim, “Wants as Explanations of Actions” for a related notion of what these authors call a synthetic, but quasi-analytic principle. To quote one of their examples (p. 427), “If daydreaming about P is pleasant to X, then X wants P”.Google Scholar
  46. 46.
    Some, like Gilbert Ryle of course, The Concept of Mind (London: Hutchison, 1949), would deny that there really is an ontological commitment to “inner events” even in our “action-theoretic” discourse, as a perspicacious analysis of ordinary action-talk would show. I disagree with Ryle’s view, but a presentation of my reasons for doing so is beyond the scope of this inquiry.Google Scholar
  47. 47.
    See, for example, W. Sellars, “Philosophy and the Scientific Image of Man” in his Science, Perception and Reality, Chap. 1, pp. 1–40, J. Margolis, Philosophy of Psychology (Englewood Cliffs, N.J.: Prentice-Hall, 1984), Chap. 2, pp. 8–33, and Paul M. Churchland, “Eliminative Materialism and the Propositional Attitudes”, in his A Neurocomputational Perspective: The Nature of Mind and the Structure of Science, (Cambridge, Mass.: MIT Press, 1995) Chap. 1, pp. 1–22.Google Scholar
  48. 48.
    I. Newton, Mathematical Principles of Natural Philosophy and His System of the World (1687), trans. by A. Motte, ed., F. Cajori (ed.), (Berkeley, California: University of California Press, 1960), p. 13.Google Scholar
  49. 49.
    For instance, as E. Nagel observes in The Structure of Science (New York: Harcourt, Brace and World, 1961), p. 181, “¡­ we might now adopt the first axiom as the criterion for the equality of temporal periods ¡ª two times being defined as equal if during them a body moving under the action of no forces covers equal distances along a straight line”.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2001

Authors and Affiliations

  • Bernard Hodgson
    • 1
  1. 1.Department of PhilosophyTrent UniversityPeterboroughCanada

Personalised recommendations