A Response to Gene Callahan: Scientism In The Way Of Science

From Thinkmarkets: Scientism in the way of science by Gene Callahan.

Gene takes critics of economics to task. I misunderstand it at first, (as does Ebeling), and the ensuing commentary is worth reading.

(Originally posted there. Posted here for documentary purposes. -A nod to the few sites like Econlib that seem to think documenting one’s work this way is bad for some reason.)

I repeatedly find attacks on positions in the social sciences made based on extremely limited and, frankly, antiquated views of how the physical sciences proceed. I will give one example from a rightist criticism of a leftist view, and one that is a leftist criticism of a rightist view, to illustrate that my point has nothing to do with ideology — or perhaps, that it has to do with the way ideology can lead one to embrace flimsy criticisms of other’s positions.

The first excerpt is from Hunter Lewis’s book, Where Keynes Went Wrong:

“In chapter 15, we saw how Keynes wrote N = F(D), which means that employment, denoted N, is a function of demand. Demand however is defined as expected sales, not actual sales. We noted that expectations are not a measurable quantity and thus do not belong in an equation.”

Well, one way to measure these expectations would be to walk around and ask the entrepreneurs “How much do you expect to sell this year?” then total up those amounts. Why in the world this would not be a fine measurable quantity is unclear.

But perhaps even worse is Lewis’s contention that only a “measurable quantity” belongs in a mathematical equation. So, let us strike pi from all of our equations, and e, and, most certainly, i! All complex numbers must be banished, and negative numbers are fairly suspect as well.

Furthermore, most of the entities dealt with by modern physics are not directly measurable. Instead, what we measure is a dial reading or a trail on a photographic plate, things which require a great deal of theory to connect them to entities like electrical fields or positrons. As the philosopher Susanne Langer wrote:

The sense-data on which the propositions of modern science rest are, for the most part, little photographic spots and blurs, or inky curved lines on paper. These data are empirical enough, but of course they are not themselves the phenomena in question; the actual phenomena stand behind them as their supposed causes… we see only the fluctuations of a tiny arrow, the trailing path of a stylus, or the appearance of a speck of light, and calculate to the “facts” of our science. What is directly observable is only a sign of the “physical fact”; it requires interpretation to yield scientific propositions… and [realizing this,] all at once, the edifice of human knowledge stands before us, not as a vast collection of sense reports, but as a structure of facts that are symbols and laws that are their meanings.

(Philosophy in a New Key)

And surely this was what Keynes thought: aggregate demand may not be directly observable, but we can formulate laws by which itseffects are observable, for instance, in a recession. Now, whether he was correct or not is not my topic, but there is certainly nothing unscientific about his hypothesis.

The second excerpt is from a history of marginalism at The New School for Social Research:

“However, [marginalism’s] Achilles’ heel was the very notion of ‘marginal utility’. Marginal utility, let us be frank, is hardly a scientific concept: unobservable, unmeasurable and untestable, marginal utility is a notion with very dubious scientific standing.”

Unobservable, unmeasurable and untestable — like, say, infinitesimals in calculus! (And people like Berkeley directed just such criticism at infinitesimals and other mathematical notions.) Once again, we have some unfounded belief that scientific entities must be directly observable, rather than observed by their hypothesized effects. (And certainly the theory of marginal utility predicts many observable phenomena, such as the lack of a price for air in normal circumstances.)

[callout] … probabilism in the social sciences as we understand it … is unscientific. Not simply beause the methods are logically false and because the predictive capacity of our methods are false, but because NOT USING THEM appears to produce better results than using them.[/callout]


I think that  Gene’s argument is a bit clearer now that I have read comments by others.   And perhaps I’m adding additional vectors of inquiry rather than debating his position.

Gene’s argument is that people from the physical sciences argue that economics is not a science and counters the grounds on which their criticisms are based.  I interpreted his posting that people from the psychological school were forming the criticism against positivism in economics.  Gene’s criticisms are correct, in that mathematics relies upon incomplete approximations that are convenient contrivances, and that economic science relies upon similar assumptions, so he is attacking the physical sciences on their methods – saying their criticisms of social sciences are hypocritical.

I would argue that since the velocity of the transfer and transformation of energy in time and space is knowable, and that the same velocity of knowledge is not yet knowable, that probabilism in the social sciences as we understand it – and as I have stated below,  is unscientific. Not simply beause the methods are logically false, and because the predictive capacity of our methods are false, but because NOT USING THEM appears to produce better results than using them. And that while results in the physical sciences have neutral consequences (or perhaps do not have moral consequences – those that affect others without their consent) that consequences of failure necessary for testing in the physical sciences creates negative externalities, as well as being simply counter-productive in the social sciences.

(I believe I understand how to discover the formula for that velocity, and how to know it, but not what it is, and that someone more intelligent, and most likely younger than I am will be required to solve it. But at least google is accumulating the data needed to determine it.)

Original Reply:


Well, I think the argument against the use of models is different from that which you’re stating.  There are three or four major lines of argument in your posting all making assumptions about ‘science’ and the scientific method.

Marginal utility is an expression of the relativity and subjectivity of value, and the plasticity of utility, and the dynamic variability of value in real time.  This creates a set of variables that lead to the effective uniqueness of each object in time for many (if not all) objects, which in turn leads to the categorical error of aggregation when applied to quantities, each of which includes necessary errors due to aggregation.  And this error of aggregation is the reason for non-prediction. And therefore non-prediction is caused by the very reasons austrians stated.  That in the aggregate much of this can be modeled, is true, at least for many commodities.

Objects in physical space have a prior course.  So do human events.  We can measure the delta in the course of physical events, but CANNOT measure the delta in the course of social events. That is the simplest statement of the problem.  It is that social events CANNOT be measured because they are temporally unique.

And further, Marginal utility is absolutely testable (and has been.) So I don’t understand, or rather, you could be making any number of points, and its unclear which.

Marginal utility is a categorical description of a visible, measurable process, whenever that process results in an exchange (at least.)  True, we cannot know the opportunity costs paid by individuals, but we can measure whenever they do act to exchange goods or services.

Instead, the criticism of models is not on grounds of material measurement of transactions, but that :

a) empirical models in the social sciences are not predictive and are even inversely predictive in relation to their utility in time.

b) that they are consistently not predictive (although they are descriptive of the past) and therefore false, and

c) that as demonstrably false, they are unscientific.  That due to subjectivity and innovation, plasticity of utility, and the resulting heterogeneity of capital, and asymmetry of information, shocks and the vicissitudes of time, they are logically destined to be false. (ie: it is not the use of measurement, it is the use of measurement to determine causality – not correlation but causality – that is scientific.)

d) that the use of false, non-predictive, arguments are used to justify implementing dangerous risk-accelerating unscientific policy.

e) that we cannot  model what might have been, had we not used false models to enact policy, and therefore calculate the real cost of policy. (ie: we cannot compare what might have been with what has come to pass, and sum our costs plus our profits.)

f) that by implementing such policies we expose ourselves to  and indeed, encourage greater risk. Ie: the austrian business cycle of booms and busts.

g) that it appears, that in history, whenever the commercial sector grows faster than the state can regulate it or redistribute the capital from it, and form a predatory bureaucracy upon it, the results for the entire society, at least narratively if not certainly empirically, seem to be better than those where state intervention has occurred.

Meaning that the Austrian criticism is that the use of the calculus of measurement in heuristic social processes will result in non-prediction and exacerbation of risk — or at least, such models will be limited to prediction based on the asymmetry of information discovered by the act of building the model, but not of the asymmetry of information yet to be developed by innovation or shocks, and therefore undiscoverable by the process of building a model.

I would argue that Keynes covered these problems in A.T.O.P. Although I am not a scholar of his work.  And that austrians agreed with him on many of those positions.  But the PRACTICAL matter is that the profession is heavily invested in a technology that demonstrably does not work, yet is relied upon for policy decisions every day.

Models are a superior means of describing causal processes where language and human limits to conception fail. However, tehy rely upon a mathematics derived from the much more simplistic physical sciences.  And until we can measure the ‘natural forces’ of men’s mental capacity, which are largely the properties of memory in time when in the presence of vast information, we have no formulae by which we can call our efforts sufficiently scientific rather than simply a convenient means of toying with economies against the will of those struggling with knowledge and capital to avoid and circumvent all that toying.

So either I don’t understand, or I do not think your criticism is founded. The people that criticize empiricism may not be using a substantive foundation either, and may justify sentiments and intuitions with false appeals to reason that they do not fully understand. But I do not see how your criticism is logical in the context.

Looking forward, solving the problem of induction instead of relying on (false) equilibria relies that we understand, and develop a formulae what might best be called ‘velocity’, which is the rate of innovation given the limited ability of the human mind to make ‘jumps’.  Therein is a formula of greater importance than E=mC^2.  And because that velocity can be known, probabilism will have a rational boundary, rather than the irrational boundary we have conveniently constructed out of historicist necessity.

I hope I have been sufficiently cogent on a subject of complexity that has admittedly exhausted many of our best minds.  And apologize in advance for my failures.


Leave a Reply