Does it take a theory to beat a theory?

4 Dec, 2021 at 15:55 | Posted in Economics | 4 Comments

George Stigler quote: Theories are not rejected by cirsumstantial evidence: it  takes a...If the substantive theories and methods you use, given their ontological presuppositions, are appropriate to the nature of those aspects of social reality you are addressing, then fine. The only methodological ‘dictum’ I support is tailor methods to the nature of the phenomena you are addressing. Certainly, I would offer no ontological critique. Specifically, if the five sets of properties you list do characterise the phenomena you address, and your methods do not carry ontological presuppositions that are so different as to be contradictory (for example that require of the phenomena that you are addressing that they take the form of isolated atoms [causal factors that have the same independent, invariable effect no matter what the context or anything else that is going on]) then fine. It remains the case though that almost all methods of mathematical modelling employed in modern economics (not matter how simple of complex, linear or non-linear, stochastic or dynamic, simulative, predictionist or otherwise) do carry such presuppositions. Where/if bits of reality so conform to a system of isolated atoms then limited use of such methods might be fine. Of course, you suggest, indeed, that this is your approach. The mainstream of modern economics (and let us face it the only coherent account of the modern mainstream is a reliance, indeed dogmatic insistence on, methods of mathematical modelling) determines its methods a prior independently of the natures of the material that are to be addressed. That’s different. And it’s no wonder that the project results in little insight …

I believe you have often explicitly written that “it takes a theory to beat a theory”, and you see that as a challenge to people like Lars and myself to engage in substantive economics. I do not think it is. Yes, it takes a theory to beat a theory, but the theories being contrasted must be comparable. The continuing problems of modern economics do not reduce to this particular mathematical model or that particular model but the whole modelling emphasis. The commonality of all such modelling is the widespread commitment to deductivist forms of reasoning/explanation that, to be relevant, presupposes a social reality composed largely of closed systems of isolated atoms. This is the ontological theory of the modern mainstream whether recognised or not. It is this theory to which we must apply your dictum ‘it takes a theory to beat a theory’. And that, or so I claim and argue, is precisely what critical realism (CR) does. Thereafter very many (often competing) theories of substantive phenomena can be constructed all of which draw on CR (just as many substantive economic modelling exercises abound). If CR is correct then almost all modelling exercises will be irrelevant (so CR is efficacious), whereas those underpinned by CR will be merely more of less correct or false. It is not up to Lars and myself to produce these accounts (even though I have produced some). I would claim that figures as diverse as Marx, Keynes, Veblen, and Hayek have all in their own ways both criticised deductivism (as vulgar economics, pseudo-science, neoclassical economics and scientism, respectively) and its ontology, and produced (very different) substantive accounts consistent (I argue) with CR.

This of course feeds into the reason that Lars, and also myself, are almost always negative in referring to the mainstream. From the perspective we adopt, it is simply a huge error to adopt unthinkingly – and especially to insist that we all do so — methods that carry ontological presuppositions that rarely if ever hold in the social realm. So, a largely critical orientation to the mainstream (as here understood) is surely warranted. But I myself can be and am very positive towards anything more relevant.

Tony Lawson / RWER Blog


  1. @ JS
    My earlier comment was a protest against the negativity and emptiness of “social ontology” and Critical Realism.
    Syll and Lawson admit that they are reluctant “to engage in substantive economics”. This makes it very difficult to fathom the meaning, implications and utility of their verbiage.
    A rare exception was Lawson’s recent discussion of the nature of money which he produced as an “illustration” of “social positioning theory”. But sadly Lawson could only reach the ridiculously incorrect and useless conclusions that:
    1. In modern economies money consists of “private/commercial bank debt and central bank debt”.
    2. “Cash, comprising central bank notes and coins, is never money”.
    More generally, I am unaware of any useful practical results from Critical Realism.
    Instead of engaging in substantive economics, CR writers choose the easy path of criticizing mainstream economics. However, they exaggerate and distort the arguments. For example, Lawson alleges that economics presumes “a social reality composed largely of closed systems of isolated atoms”. However, I can’t find any economist who holds this absurd view.
    Contradicting Lawson are numerous statements by eminent economists, eg:
    “economists study the actions of individuals, but study them in relation to social rather than individual life ; they therefore concern themselves but little with personal individual peculiarities of temper and character. They watch carefully the conduct of a whole class of people, sometimes the whole of a nation, sometimes only those living in a certain district, more often those engaged in some particular trade at some time and place”.
    Alfred Marshall – Principles of Economics 8e 1920, Bk I, Ch.2, §7.
    Moreover, Lawson is plainly wrong in claiming that economics is only concerned with closed systems.
    Examples of open systems include the discussions of are exogenous exports, animal spirits, confidence, government expenditures and wages in Keynes’ General Theory, chapters 10, 19 etc .
    The negativity and emptiness of Lawson and Syll’s arguments stem from their Critical Realist philosophy. This philosophy is based on the extraordinary claim that a “deeper reality” exists beyond everyday human experiences and beyond the reach of science.
    This is an empirical claim but it lacks any supporting evidence. There is zero physical evidence, zero testimony from travellers into “deeper reality”, zero communications with the inhabitants, zero information from supernatural beings, and zero valid evidence from dreams, fantasies or from UFOs.
    Using their own peculiar language, the unwarranted theoretical ontological presumptions of Syll, Lawson, Bhaskar and other so-called “realists” lack any “theoretically and empirically adequate data generating process“, lack “external validity”, and lack any “export licence”.
    CR philosophy leads to distorted views about economic methods. For example Lawson believes that “If CR is correct then almost all modelling exercises will be irrelevant“. Deductive methods/modelling and to inductive methods/econometrics are both scorned because they are not focussed on the “underlying structures/forces/powers/ mechanisms” which supposedly exist in “deep reality”.
    Lawson and Syll rely almost entirely on abduction, aka retroduction or “Inference to the Best Explanation”.
    However, this is not a really a method for inference because there are no principles regarding the difficult issues regarding data and the criteria for deciding what is best.
    – Simplicity?
    – Comprehensibility?
    – Empirical Adequacy?
    – Cognitive Fit?
    – Explanatory power?
    – Minimal changes in scientific belief compared with previously held theory?
    – Intellectual pedigree?
    – Aesthetic beauty?
    – Consistency with political, philosophical and religious beliefs?
    – Anti-racism?
    – Wokeness?
    How should we measure and weight such criteria? And who should decide such matters?
    Most scientists prefer inductive methods, eg maximum likelihood given the available data. However, abduction may have an very important role at the outset of statistical reasoning in generating hypotheses. (This is the historical meaning of the term “abduction” in philosophy, but it is not the meaning of Critical Realists.)
    “All the devices for the discovery of the relations between cause and effect… have to be used in their turn by the economist: there is not any one method of investigation which can properly be called the method of economics; but every method must be made serviceable in its proper place, either singly or in combination with others.
    Alfred Marshall – Principles of Economics 8e 1920, Bk I, Ch.3, §1.

  2. I am a bit surprised to read such commentary still. Lawson has argued (not asserted) over and again that the ontology of the modelling mainstream is worlds of isolated atoms. You offer no argument you just dismiss the idea. Can we have an argument?
    Worse, having dismissed Lawson’s reasoned assessment you then focus on induction and deduction, the two methods consistent with presupposing atomism, and simply provide a quoted from a presumed authority to imply that together they are sufficient. Lawson has also provided an argument (not a quote from some authority) against restricting the focus to deduction and induction too, also over and again. Thus, in his almost 25-year old Economics and Reality, Lawson defended the causalist, non deductivist, perspective he was defending arguing that it

    “is important to recognise, therefore, that the essential mode of inference […..] is neither induction nor deduction but one that can be styled retroduction or abduction or ‘as if’ reasoning. It consists in the movement, on the basis of analogy and metaphor amongst other things, from a conception of some phenomenon of interest to a conception of some totally different type of thing, mechanism, structure or condition that, at least in part, is responsible for the given phenomenon. If deduction is illustrated by the move from the general claim that ‘all ravens are black’ to the particular inference that the next one seen will be black, and induction by the move from the particular observation of numerous black ravens to the general claim that ‘all ravens are black’, retroductive or abductive reasoning is indicated by a move from the observation of numerous black ravens to a theory of a mechanism intrinsic (and perhaps also extrinsic) to ravens which disposes them to be black. It is a movement, paradigmatically, from a ‘surface phenomenon’ to some ‘deeper’ causal thing.”

    It is too easy (though consistent with the usual standards of so many modern economists) to just dismiss an argument” (perhaps quoting some authority) rather than with engaging it.

    • Sorry I am not good at this. The comment (I am now commenting on) was on Kingsley Lewis’s comment (below)

  3. Contrary to Lawson’s verbiage, economics does not presume “a social reality composed largely of closed systems of isolated atoms”.
    Critical Realist philosophy is irrelevant.
    The role of deductivism and modelling in economics is an ancient debate which was resolved by Alfred Marshall over 130 years ago:
    “You [Neville Keynes (& Lars Syll)] make all your contrasts rather too sharply for me. You talk of the inductive & the deductive methods: whereas I contend that each involves the other … I think the right order is first to emphasize the mutual dependence of induction & deduction, & afterwards to show in what kinds of inquiry the economist has to spend the greater part of his time in collecting arranging & narrating facts, & in what kinds he is chiefly occupied in reasoning about them & trying to evolve general processes of analysis & general theories .”
    Alfred Marshall – Letter to Neville Keynes (about 1898?) quoted in:
    Equilibrium and Evolution Alfred Marshall and the Marshallians by Neil Hart (auth.) (

Sorry, the comment form is closed at this time.

Blog at
Entries and Comments feeds.