Shortly before arriving at Chapman University in 2008 I discovered some essays by Frank Knight. Every graduate student in economics learns about or, at least, has heard references to Frank Knight, one of the original members of the "Chicago school of economics." While his book Risk, Uncertainty and Profit (1921) is well known and considered seminal for disentangling uncertainty from risk, I've met few people who have actually read the book. I have never read the book. What is equally not read from Knight's extensive body of work, but mostly unknown in modern economics, is his collection of essays entitled, The Ethics of Competition.
If I was still teaching graduate level experimental economics, the fourth essay would be required reading. The first three essays on the ethics and psychology of economics, published in the Quarterly Journal of Economics between 1922 and 1925, are equally as important and perhaps lay the necessary groundwork for understanding the fourth essay, which is entitled, "The Limitations of Scientific Method in Economics." (Notice that there is no definite article before "Scientific Method.") First appearing in a 1924 book that is quite hard to come by in hardback, the fourth essay opens with a discussion on the meaning and logic of science.
Given the demands of academic specialization, experimental economists have precious little time to ponder the philosophy of science, much less the philosophy of social science, but that doesn't deter us from calling the professional association of experimental economists the Economic Science Association nor my academic home the Economic Science Institute. Science is the name of the game in experimental economics.
So how far can we take the scientific method in economics? As an ardent existentialist, Knight says only so far: "life is at bottom an exploration in the field of values, an attempt to discover values, rather than on a basis of knowledge of them to produce and enjoy them to the greatest possible extent," and thus, as if straight out of Goethe's Faust, "[w]e strive to 'know ourselves,' to find our real wants, more than to get what we want" (p 97). Earlier in the book he succinctly rejects the first lesson of intermediate microeconomics as answering the wrong foundational question: "[l]ife is not fundamentally a striving for ends, for satisfactions, but rather bases for further striving" (p. 15).
In other words, the operative part of goods in the study of economics is not as a substantive, but rather the adjectival, aesthetical good of goods. Value isn't just a maximum willingness to pay for a good denominated in some currency, but a value judgment on the goodness of spending a sum of money this way as opposed to some other way. Evaluating what we do as good is a human universal. Linguists like Anna Wierzbicka have found that the concept of GOOD is expressible in every language ever studied and thus is presumed to be innate. It's that important to the human condition.
Knight argues that values place a limit on scientific thinking in economics because of the tendency for the scientific method to "subordinate" understanding to a "desire for control" (p. 99). A prominent industrial organization theorist once quipped to me that "experimentalists are control freaks." Indeed we are. It is standard practice to attempt to control for the propensity to cooperate in repeated interactions by ensuring no participant will ever encounter the same participant twice in a single session.
The problem is that this "scientific atmosphere obscures if it does not eclipse a considerable part of the field of values. It centres attention on the results of activity, weakening or destroying the value of the process" (p. 99). Being sociable to those we see daily on the playground is a lesson we are taught from an early age on. If people are so stubbornly determined to cooperate in repeated interactions in the laboratory, then perhaps by attempting to exert control over the results we are instead paradoxically losing control over the big picture of why and how we do good by each other.
This problem of values, however, is only a limitation insofar as it is a self-imposed constraint on the techniques we employ and the questions that we ask in economic science. The deeper constraint is a categorically different kind of problem that economic science faces as a study of human beings, the study of ourselves. In presaging Wittgenstein and Hayek, recall that this is the early 1920's, Knight questions behaviorism and the "much loose talk about the 'analysis' of mental phenomena, which...is a largely a pernicious use of words; in general these cannot be taken apart or combined experimentally, and neither 'elements' nor 'compounds' have any general marks of identification" (p. 107). (What would Knight think of social preferences?) He further recognizes that "people are far more concerned with meanings than with sense qualities, and meaning...must be suggested rather than stated" (p. 126).
Meaning must be suggested because it is "felt but not reasoned," to quote Hayek. We can't sense the 6-D waggle dance of a bee, but mathematicians can specify it precisely as a flag manifold. Human conduct is the opposite. We can sense the fairness, justice, equity, etc., of human conduct, but because of the (literally) "imperceptible differences in the situation" we cannot precisely specify what we feel (p. 114). Try to specify all the rules of fair play.
In this way discussing economics can be much more like discussing art than gravity. Take, for example, a Rothko painting. If I have to state what I see in the painting, I say I see some "rectangles" (where are the boundaries?) that appear to float and others that pulse. We cannot specify how the colors, textures, lighting, brush and sponge strokes, distance from the painting, angle of view, etc., combine to generate those feelings when in the intimate Rothko Room of The Phillips Collection. But I can suggest what it feels like. Similarly, we cannot account for all the sensory inputs that lead us to judge a situation as just but not fair, or fair but not equitable, or equitable but not just. We simply feel but not reason that it is unfair for one person to receive 80% of the pie in an ultimatum game or unjust to let a defendant off due to a lack of evidence beyond a reasonable doubt.
From this fundamental difference between the natural and human sciences Knight concludes that the "real sociology and economics must be branches of literature as much as of science. In fact they need to be both, and commonly succeed in being neither" (p. 139). Ninety years later, it is worse than commonly succeeding in being neither. A chasm separates economics from the humanities. The positive facts about how we became the most prosperous species in the history of the planet are unappreciated in the humanities, often because the facts and logic of wealth creation are regrettably unknown. And it is equally unappreciated in economics that ours is not comprised only of objective science. The economics of prosperity is the study of ourselves as human beings, and so value judgments of good have had something to do with it along the way.
Because our minds present the GOOD of economics to us in a scientifically unpinpointable way, we must engage our imaginations through literature if we are to discover the values that commingle with the scientific facts and logic of economics. Is economics a science? Yes, but it can quickly devolve without an aesthetic appreciation of the GOOD in economics. Thus, as Knight concludes (p. 139):
The first step toward getting out of this slough, we suggest, is to recognize that man's relations with his fellow man are on a totally different footing from his relations with the objects of physical nature and to give up, except within recognized and rather narrow limits, the naïve project of carrying over a technique which has been successful in the one set of problems and using it to solve another set of a categorically different kind.