Although I feel Jonathan did an admirable job defending the Misesian methodology, he has asked me to buttress his own writings with some philosophical thoughts of my own to which I of course will oblige. Jonathan links us to a critique of the practice of praxeology by Eric Perkerson at The Social Rationalist. After some interesting thoughts on the value of knowledge for its own sake and knowledge for the sake of its application, we find Eric’s thesis put bluntly:
The process of pure deduction, as a methodology, should not be valued in this regard. As we will show, it cannot be applied because it cannot yield appropriate knowledge-as-means.
This thesis is demonstrably untrue. First, I will deal with the body of his work – which deals with this specific thesis – and I will then comment on a few other erroneous, auxiliary points Eric tries to make against praxeology. Eric begins the meat of his post with an unqualified and wrong assertion:
Pure deduction can only yield conditional knowledge. Only with foundational assumptions, or with things that are “given,” can logical deduction yield any results at all. Logic in a vacuum yields nothing. Nothing in, nothing out.
Pure deduction, as a matter of fact, can yield quite a bit of certain knowledge. The understanding of mathematics and geometry are based on pure deduction from an incontestable axiom: the law of identity. A=A is not an empirical hypothesis, nor is it an abstract illustration of philosophical confusion (see: Hegel’s “self-differentiating unity” for an example). It is an incontestable truth that we cannot conceive of as erroneous. That A=A is correct does not require continual testing or experimentation; it can and must be understood a priori. Likewise, the law of non-contradiction – that something cannot both be A and not-A at the same time – is an example of pure deduction. Contrary to Eric’s assertions, these laws do not represent condition knowledge, but the opposite: unconditional knowledge. We can know with absolute objective certainty that a thing is equal to itself, and we know this by the very definition of what we are proposing. A=A is a tautology, just as all deduction is. Whether Eric thinks logical tautology can give us usable or meaningful knowledge is another thing, but certain knowledge it definitely can. For a more obvious example of the results of pure deduction to give us a bedrock of certain knowledge, see Descartes’ cogito.Furthermore, contrary to Eric’s ex cathedra pronouncement, the law of identity is not an assumption. It is a requisite for all deduction. There is no conceivable world where A=A cannot be true, and this elevates it to the status of an incontestable axiom – a product of deduction.
For conditional knowledge to be elevated to the status of proper knowledge, we have to somehow verify that the conditions or assumptions on which the conditional knowledge is contingent actually hold true in some respect out in the external world. The conditional knowledge of the pythagorean theorem is based on the condition that one has a right triangle. Actual knowledge based on the pythagorean theorem has to come from knowledge that one is dealing with something which satisfies the definition of a right triangle. Conditional knowledge becomes proper knowledge when the conditions it is based on are known to be satisfied. How this knowledge comes to be known is a problem that is outside of the realm of pure deduction: it is necessarily empirical. This is the interesting epistemological problem.
This paragraph is untrue in all of its parts. Not only does Eric not give us a definition of “proper” knowledge (versus improper knowledge?), but he is wrong when he asserts the need to empirically demonstrate the necessary conditions for the Pythagorean Theorem. The theorem, as we all should know, says that the two sides of a right triangle, when squared and added, will equal the hypotenuse when squared. It must be remembered, that in order for this to be true, we do not need to posit the existence of any right triangles. In order for it to be true that the circumference of a circle is equal to its diameter multipled by Pi, we do not need a circle. These polygons illustrate the logical theorems, but are themselves not exactly necessary. The Pythageoream theorem and the mathematical properties of a circle are not dependent on witnessing or “having” such shapes. They are true by virtue of their tautological status. The relationship of Pi to a circle is such that the concept of “circle” – a round plane figure whose boundary (the circumference) consists of points equidistant from a fixed center – already implies the necessary mathematical unit of Pi. What empirical testing and experimentation can do is to verify whether certain theorems fit into a specific case, not whether certain theorems are true or false. Eric seems to adopt a certain mathematico-positivism, whereby the truth value of all logical claims are held conditional to some external reality. How and why Eric believes the observable world can guarantee him something other than “conditional” knowledge, I do not know. The observable world is far more unreliable than logical tautologies as far as delivering the verdict on the truth value of scientific claims, and for this reason we should be skeptical of trying to prove logic using reality. The really interesting epistemological problem, again contrary to Eric’s argument, is not a boring empirical argument, but a much deeper issue in metaphysics and philosophy of mind. How our knowledge squares with reality is not satisfied by a “go and see for yourself” attitude. It is satisfied by a rigorous metaphysical formulation that explains the essence of reality and ties in true epistemic claims about that reality. Whether reality is composed of many substances or just one, whether reality is deterministic or indeterministic, whether reality is simply “our universe” or if it consists of multiple universes; these are metaphysical and epistemological arguments that are the heart of philosophy and will not be solved by a microscope. Left unmentioned is exactly why Eric believes “reality” and the a posteriori can grant us unconditional knowledge when it is a priori knowledge that is based on necessary and universal truisms, not happenstance observations.
Conditional knowledge, without knowledge about whether or not the foundational assumptions hold, cannot be knowledge-as-means because there is then no way of connecting the real world with the abstract world in which the assumptions we have based our conditional knowledge on hold. We have shown that the process of deduction by itself, as a methodology, is useless in attaining knowledge-as-means, because it gives no way of solving this problem.
I have already illustrated that these foundational “assumptions” are not assumptions at all but necessary truisms. The laws of identity and non-contradiction are a priori true because they cannot coherently be denied. The question of how our logical knowledge connects with reality is an old philosophical question, and one that Kant has helped us understand a great deal. The Kantian framework of synthetic and analytic propositions, along with the a priori and a posteriori methods, have given us a great deal of headway into answering this question. I relate Kant to Mises here.
But while we have good reasons for believing this [action axiom] to be true, it is by no means true on the basis of pure deduction alone. Fundamentally, this is an empirical claim. We can certainly conceive of a world where humans are nothing more than mere animals, reacting on the basis of instinct alone. We also know that there are human beings in our world who do not act, who have fallen into persistent vegetative states. But if one tries to solve this problem by appealing to definition, by claiming that human beings aredefined by their ability to act, then one is at once at a loss to explain how they can know that the beings around them who look similar and with whom they seem to communicate are in fact human beings in the sense that they act.
Eric is clearly confused. In the first place, it is true that Mises writes “Man acts,” and we are tempted to run out into the world and see if this is true. But it would be more accurate for Mises to have written “Agents act.” Despite some sloppy terminology from the early 20th century, we can still recover much of Mises’ formulation. For instance, it is an incontestable truth that agents act – this is in fact a prerequisite for all other definitions of agency (argumentation, reflective thinking, the pursuit of virtue, etc.). Action is necessary to the concept of agency and cannot be done without. Secondly, to tie in to my earlier point, we do not need to posit the existence of any agents to arrive at a theoretically true economic science. All we need to say is that if agents exist, then they must act. If they act, they must choose. If they choose, they must exhibit preferences. And also, if they choose, they must forgo any competing alternatives. This illustrates the theoretical principle of opportunity cost – that the cost of any action is the value of one’s next most favorable alternative. This logic is sound irrespective of if there are any agents in which it could be manifest. Praxeology does not give us knowledge about what humans or non-humans will or can do, but about categories of action that manifest themselves in the economic sphere of agents. If there are creatures who act, the whole body of praxeology applies to them in all its parts.
Far from being the end of the matter, the action axiom is but the first of many assumptions needed to fuel the engine of deduction that produces the claims embodying Austrian economics.
The action axiom is not an assumption, but an incontestable truth. Nobody can deny that man (or agents) act because a denial is itself an action and the opponent would be caught in a performative contradiction. The action axiom is thus immediately and obviously true. Mises makes a few auxiliary empirical claims (such that labor constitutes a disutility and that we live in a world of scarcity, among others), but these are not exactly necessary to erect the body of praxeological truth. There are no assumptions in Austrian economics. There are premises and there are deductions.
Very little, and possibly nothing which could be classified as “useful” economic knowledge-as-means can be derived from simply the action axiom alone.
An obvious counterexample is the principle of diminishing marginal utility, which is proved on praxeological grounds. Hoppe writes:
[T]he law of marginal utility follows from our indisputable knowledge of the fact that every actor always prefers what satisfies him more over what satisfies him less, plus the assumption that he is faced with an increase in the supply of a good (a scarce mean) whose units he regards as of equal serviceability by one additional unit. From this it follows with logical necessity that this additional unit can only be employed as a means for the removal of an uneasiness that is deemed less urgent than the least valuable goal previously satisfied by a unit of such a good. Provided there is no flaw in the process of deduction, the conclusions which economic theorizing yields must be valid a priori. — Hans-Hermann Hoppe, The Economics and Ethics of Private Property (Auburn: Ludwig von Mises Institute, 2006), p. 278
To put it in terms for non-economists, Hoppe says that every actor prefers that which satisfies him more over that which satisfies him less. When you add an additional unit of a good to an already existing stock of the same good, this additional good must be put to a use less valuable than the previous units of the stock. Thus, the first sack of grain to a pioneer farmer will be put to a more productive use in satisfying his desires than his second sack of grain, which will still be more productively utilized than his third sack of grain, etc. This all follows from the praxeological fact that we prefer things which satisfy our desires over things that satisfy them less. This proof does not require empirical testing, experimentation, or falsification – but logic.
“Useful” knowledge, for people acting in the real world, must be knowledge about the real world.
This is the great value in Kant’s synthetic a priori. It is synthetic – in that the concrete matter of inquiry is a real world phenomenon – and yet it is learned a priori – by logic. We can learn “useful” knowledge about the real world without acquiescing to microscopes and hypotheses. Economics is a subject dealing with real-life content: value, choice, cost, profit, and loss. These are the categories of action that we can understand, as I illustrated above, using a purely deductive methodology.
The remainder of the critique is focused on the difference between methodologies. Our author makes us aware that the methodology of physics is different than the methodology of history. Surprisingly, he draws no conclusions from this on the question of what methodology can give us knowledge in economics and proceeds to throw up his arms on the debate, flatly declaring that:
“In the final analysis, the methodology of Austrian-economics-as-practiced, as opposed to the stated methodology of Austrian economics, is a synthesis of casual empiricism, in that the assumptions necessary for the reasoning are simply taken for granted, and logical deduction.”
Clearly Ludwig von Mises’ thousand-page treatise on economic theory and methodology skipped right over our author’s head when considering source material to verify that the “Austrian-economics-is-empiricism” story is true. The ideas that differentiate the Austrian school from others – from a time-preference theory of interest, to the catallactic function of profit and loss, to the Mises-Hayek business cycle theory, to name a few – are not products of casual empiricism, but rigorous deduction. While not all Austrians are Misesian purists, the Austrian school by and large holds and practices a substantially different mode of science. We are not empiricists with some mathematics pasted for a backbone. This critique would hold fine for the Neoclassicals – who do not attempt to justify their assumptions on cardinality, indifference, continuous functions, among other aberrations – but it looks absolutely ridiculous when charged against the Austrians who make economic methodology a prime concern of their study.