Tacit Knowledge, Scientism, and Government

AnomalyUK defines the concept of tacit knowledge in a post called “The Subjective”:

The Enlightenment position is that knowledge should be objective. I think that originated in an analogy with the scientific method: the only conclusions that should be accepted are those which can be independently verified. If I say that a bird cannot live in air in which a candle has burned out, you should be able to put a bird in a jar with a candle and kill it the same way. If you can’t, then my claims are not objective, and are scientifically worthless.

The Enlightenment extended this principle to government. The decision of a government should not be made on the basis of one person’s private judgement; it should be made by a scientific process, and the reasons for making it should be objective facts that others can share.

Democracy requires that principle. Like science, democracy requires that one person’s conclusions can be replicated by another. In some cases the replication may not be contemporaneous with the actual decision, but the principle must still be that “If you knew what I know”, you would reach the same conclusion, and the politician can be judged retrospectively by that standard.

Hayek identified the problem with this approach. The main problem is “Tacit Knowledge”. Tacit knowledge is what you know, but you don’t know that you know. It is knowledge that cannot be shared just by publishing a paper, but only, if at all, by teaching a craft.

A decision that has to be justified objectively cannot rely on tacit knowledge. Tacit knowledge is, by definition, subjective. A person, who, in whatever environment, is making a decision that is going to be evaluated by others, must deliberately ignore subjective considerations — tacit knowledge — and make what seems to be the best decision without that knowledge.

This process has a catastrophic impact on personal responsibility. If I make a decision not because, based on all my knowledge objective and tacit, I think it is the right one, but rather, because it is the one I can best justify to someone else, then I am no longer responsible for the result of my decision, only for the process (sound familiar?).

If someone is responsible for the results of their decision, rather than for the process of making the decision, then they will naturally make the decision most likely to have the desired result, and they will do so based on all the knowledge they have, objective and tacit.

Objectivity applied to the natural world has had such beneficial results, in terms of uncovering scientific knowledge, that it is tempting to apply it to as many areas as possible. The confounding factor is that the subtleties of human judgment and values are too complex for us to scientifically and exhaustively define with our current state of knowledge. There is the temptation to explain social policies with simple principles and rules, in the same category as simple principles used to describe the natural world, such as Newtonian mechanics.

We should distinguish between science and “scientism.” Science is a tool. Most simply, it’s the scientific method, which just means making hypotheses in advance and using experiments to confirm or deny them. In some areas, such as engineering, the application of the scientific method has led to great benefits to civilization. In other areas, such as romance, attempts to apply science are rather undeveloped. A man with an IQ of 100 who is confident will be better at romance than a scientist with an IQ of 150 whose practical experience with relationships consists of reading evolutionary psychology papers.

The complexities of good governance are more similar to romance than to engineering. An effective civil servant or politician will not be able to write down everything he knows. He will have substantial “tacit knowledge,” embodied in well-tuned perceptive and inductive biases. To give a concrete example, consider a Foreign Service officer experienced in the art of diplomacy. There will be thousands of small signs, facts, hints, moods, expressions, and details the officer will leverage to make all sorts of daily decisions. Until we have a comprehensive, unified theory of all the details of operation of the human mind, formalizing this tacit knowledge will be impossible.

“Scientism” is science applied “in excess.” The “Rationality and Science” portion of the Less Wrong quantum physics sequence comes to mind, which makes distinctions between Bayesian rationality and science. According to Bayesian rationality, there is not necessarily a hard break between “scientific” and “not scientific.” Science is just another tool that can provide Bayesian evidence of a situation. That evidence may be weak, or may not pertain to the judgment in question. Framing is important, and context greatly influences how knowledge can be applied. When it comes to government, science provides evidence for all sorts of questions and policies, but formulating effective policies may come down to tacit knowledge and subjective value judgments which are not formalized scientifically.

Where science is overextended is where people claim that it provides a socially optimal form of governance. For instance, arguing that science “proves” democratic government is optimal. Social matters, subjective values, and politics are too complicated for there to be objective scientific answers to many questions. By adopting a probabilistic perspective about judgment, we can pick and choose which kinds of scientific conclusions are useful, without being forced to overemphasize science for its own sake.

The early understanding of science, brought about during the thick of the Enlightenment, is a more rigid and Aristotelian system of thinking than the modern, Bayesian understanding of science. Unfortunately, many scientists and political commentators have an out-of-date concept of science, thinking it is applicable to more than it really is. When this comes into contact with politics, it produces men and women overconfident in their own beliefs. They suppose that being scientific gives them definitive answers to complex questions, even when the science is not particularly pertinent to the decisions involved. This was a wide error made by the Soviet Union, which supposed it could centrally plan the economy with “scientific” principles like Lysenkoism.

Science is susceptible to confirmation bias, just like any other kind of human activity. It is possible for science to be politically driven, and people can find “scientific evidence” to selectively back nearly any political opinion.

We ought to be skeptical of attempts to apply “science” or even objective standards in general to government. Just because something is difficult to define, doesn’t mean it’s not important. Civil servants and government leaders may make important decisions based on their experience, and forcing them to have objectively-definable criteria for making all decisions removes powerful sources of evidence. Only when someone is allowed to fully invoke their tacit knowledge can they be held fully responsible for the consequences of their decisions.