The meaning of apriori
by Sam Selikoff September 29, 2010
My latest response in an ongoing discussion about the Austrian approach to the science of economics:
The Austrian response to your letter is two-fold. First, let me distinguish apriori statements from aposteriori statements, to the best of my ability; then I will present some problems with mainstream methodology. Aposteriori statements are hypotheses for which observations are necessary (although not sufficient) to determine their validity, while all other statements are apriori. The statement “2 + 2 = 4” is thus an apriori statement, because its truth is determined solely by its logical consistency with the basic axioms of mathematics. It would be ludicrous to try to falsify the statement with empirical data or through measurement. Similarly, the Pythagorean theorem is apriori, within the framework of the mathematical axioms. Measuring the three sides of a triangle could never invalidate the theorem; it is true by logical necessity.
What’s important to note is that just because a statement is apriori, and requires no observations to verify, does not make the statement meaningless or trivial. With the axioms of mathematics in place, all the ingredients were present for the discovery of the conclusions and implications of the Pythagorean theorem. But Pythagoras still had to make the deductions; they didn’t just fall out of the axioms. This method of deductive rumination – apriori reasoning – produces statements which are completely true by logical necessity. These statements apply to the empirical world whenever the conditions assumed in the reasoning are present in reality.
The Law of Comparative Advantage is also an apriori statement. David Ricardo showed that mutually beneficial exchange can occur between two people if each specializes in the task in which his labor is relatively more productive – even if one of them is more productive in both tasks in absolute terms. No data can verify or falsify this fact, and because the law is deduced from a few undeniable axioms and deductions – the knowledge that human beings act purposefully and the law of marginal utility – any conclusion in opposition to it would violate the basic tenets of logic. (Again, this is not to say that the law is obvious or trivial. Indeed, before Ricardo’s revelation, most people merely assumed that the general scarcity of resources in the world meant that any act of production or consumption by one person was completely rival and thus harmful for potential production or consumption by another.)
Does the law of association apply to reality, in a world of ‘frictions’ and ‘market rigidities’ so conveniently absent from the hypothesized construct Ricardo used to formulate his deduction? Is data gathering necessary to determine the answer to this question?
It is true that the real world is very complex; but this fact does not change the empirical applicability of Ricardo’ law. As a simple analogy, consider the law of gravity (the law itself is aposteriori, but this is irrelevant for my purposes here; I am simply looking at how laws apply to the real, complex world). In complete isolation, we know that an object upon which gravity is the only acting force will accelerate towards the earth’s surface at 9.8 meters per second per second. In our world, gravity is never the only force acting on an object. But in the presence of wind resistance and atmospheric pressure, is it not true that the force of gravity is still acting on the object, fully and completely?
One accurate way to phrase how the force of gravity acts upon objects in the real, complex world is: the force of gravity will displace an object’s acceleration 9.8 meters per second per second towards the surface of the earth compared to what its acceleration would have been in the absence of gravity. If the net acceleration of an object is not 9.8 m/s/s, we know there are other forces present. But this does not nullify the complete and full effect that gravity is having upon the object. Ceteris paribus, the object is falling 9.8 m/s/s faster than it otherwise would be. The law of gravity remains true, even in the real world.
Similarly, Ricardo’s law remains true in the real world, and no empirical data could prove otherwise. All else equal, two individuals can benefit from exchange, even if one is absolutely more productive in all tasks. Whether or not this exchange takes place at a certain geographical location and at a certain point in time is a question of history (for which data analysis can be helpful in determining) and not of economic theory. The causal effect of differences in human beings – that beneficial trade exists – is exclusively an apriori truth.
Another apriori truth of economics is that an individual will allocate goods to their highest-valued uses first, and that each additional unit of a homogenous good is less valuable to the user than the unit of his existing stock that is being applied to the lowest-valued end i.e., the marginal unit. Or, whenever there is a minimum wage set by law above the market-clearing price, involuntary unemployment will be higher than it would have been in the absence of the law. If empirical analysis reveals that at some time and place unemployment decreased following an increase in the minimum wage, this does absolutely nothing to contradict the truth that minimum wage legislation fosters unemployment. It merely tells us what we already know: that historical economic data is complex, and that many forces are present in the realm of human action.
Apriori statements, then, are necessarily true (assuming no flaw in the logical reasoning leading to their deduction). Using mathematical modeling and statistical analysis to verify them is not only unnecessary but absurd. There is no reason to empirically test a logical certitude.
Positive economics is devoid of apriori statements. In fact, it forms aposteriori statements, which it then falsifies through comparisons with historical data. It is true that deductive statements formed from the standard assumptions of rationality involve the use of logic; but as logic is not sufficient for establishing these statements’ validity, any similarity between them and the deductions of mathematics, for example, is completely illusory.
If you buy the axioms of Austrian economics, then, apriori reasoning produces statements which are necessarily true. The method of apriori reasoning is particularly efficacious for the study of economics, as all economic data are complex and incomplete. It would be insurmountably difficult to develop an audio theory of acoustics by listening to an orchestra – but that is what economists attempt when using statistics to tease out causality from complex historical information. The logical limitations that experiments place on our understanding of causal events in the natural sciences do not apply to economics. We cannot conduct controlled experiments with purposive human actors. This is a central flaw of using observations to establish economic truth.
Additionally, the assumptions of consumer behavior made in the mainstream treatment – devised solely to allow for the use of functional analysis – conflict with reality. There is no reason to think that human beings’ value scales are smooth along the continuum of the real number system. In this world, humans are only given the opportunity to act upon discrete values. Far from being harmless, the assumption of continuity of preferences is both anti-realistic and unscientific. Even if we were to accept the hypothesis of continuity, though, the ever-changing subjective preferences of individuals cannot be represented by numerical functions. Functional analysis implies mutual determination, equality, and the existence of mathematical constants. In contrast, human action is a tale of cause and effect, where action always and everywhere involves the choice of one state of affairs over another, and where no quantitative constants exist.
Economics is the study of the universal laws of human action. History cannot prove or disprove any general statement; on the contrary, it is only through the lens of an apriori theoretical framework that historical events can be properly understood. ‘Measuring’ the elasticity of demand for labor for a definite city during a definite time period does nothing to develop universal truths about human interactions; it merely establishes a historical fact.
In contrast to empirical trial and error, the method of apriori reasoning is how Austrian economists develop the universal laws of economics. Austrians begin with the apriori axiom that human beings act purposefully, that is, that they employ means to achieve ends. The truth of this statement can be seen upon reflection. Any attempt to deny the statement that human beings act purposefully involves action which is directed toward the purpose of disproving the statement; that is, the denial of purposive action itself involves purposive action, so that no data or observation is required to verify the statement. Taking the statement as the main axiom of economics, we can proceed as follows: since all action is aimed toward ends, and the means (goods) employed to achieve these ends are scarce, inherent in action is the choice of one end over another. The value of the end foregone is the cost of the action taken, and every action thus involves the expectation that the psychic profit from the end pursued is greater than its cost (why else would somebody act?). Goods are valued based on the ends which they are employed to achieve, and goods of equal serviceability are applied to their highest-valued uses first. To think otherwise, that is, to think that homogenous goods are not applied to their highest-valued uses first, would contradict the very fact that human beings act purposefully. Hence, each additional unit of a homogenous good provides the user with less utility (because the end to which the new unit will be applied is necessarily less valuable than the previous marginal end), and with the introduction of exchange, we have deduced the law of demand. No measurement or data set could rebut this, because a denial of the law of demand implicitly involves a denial of the axiom that human beings employ means to achieve ends.
Austrians believe that historical data is useful for the study of economic history – but not for the development of economic theory. Representing the sentiments of many empiricists, Milton Friedman remarked that he is a ‘slave to the data’, and that economists should go where historical data leads them. Friedman argued that disagreements about apriori assumptions concerning human beings can never be settled, so economics should discard them completely. I have demonstrated that the action axiom is undeniable. My question is, why should economists start with one hand tied behind their backs? As scientists who study human interaction, we know something about human beings: that they behave purposefully. If there are apriori deductions which follow from this basic axiom – deductions which are true by logical necessity – why should we ignore them?
As a short concluding note, my knowledge of mainstream/neoclassical economics is still far from complete, and I am learning more about it every day in graduate school. It saddens me, however, that so few economists are aware of the Austrian methodology. Based on my individual research thus far, it seems to me a truly superior approach that has been unjustifiably relegated to the dustbin of economic thought.