Wicked and tame, another face for determinism versus probability

I came across the term "wicked problem" some long while ago. Got around to investigating it just recently. It is the term used to cover what I have been calling complex, dynamic, nonlinear systems with hidden feedback mechanisms. Systems which are poorly understood and difficult to manage.

Atul Gawande in Something Wicked This Way Comes provides a good discussion.

In 1973, two social scientists, Horst Rittel and Melvin Webber, defined a class of problems they called “wicked problems.” Wicked problems are messy, ill-defined, more complex than we fully grasp, and open to multiple interpretations based on one’s point of view. They are problems such as poverty, obesity, where to put a new highway—or how to make sure that people have adequate health care.

They are the opposite of “tame problems,” which can be crisply defined, completely understood, and fixed through technical solutions. Tame problems are not necessarily simple—they include putting a man on the moon or devising a cure for diabetes. They are, however, solvable. Solutions to tame problems either work or they don’t.

Solutions to wicked problems, by contrast, are only better or worse. Trade-offs are unavoidable. Unanticipated complications and benefits are both common. And opportunities to learn by trial and error are limited. You can’t try a new highway over here and over there; you put it where you put it. But new issues will arise. Adjustments will be required. No solution to a wicked problem is ever permanent or wholly satisfying, which leaves every solution open to easy polemical attack.

The original paper is Dilemmas in a General Theory by Horst W.J. Rittel and Melvin M. Webber. The abstract of their paper is:

The search for scientific bases for confronting problems of social policy is bound to fail, because of the nature of these problems. They are "wicked" problems, whereas science has developed to deal with "tame" problems. Policy problems cannot be definitively described. Moreover, in a pluralistic society there is nothing like the undisputable public good; there is no objective definition of equity; policies that respond to social problems cannot be meaningfully correct or false; and it makes no sense to talk about "optimal solutions" to social problems unless severe qualifications are imposed first. Even worse, there are no "solutions" in the sense of definitive and objective answers.

What Rittel and Webber appear to focus on is what I have distinguished, analogizing to physics, as problem solving based in Newtonian mechanistic determinism versus problem solving that is based in Maxwellian probability and uncertainty.

But I like the succinctness of wicked versus tame. I am not aware that the terminology is widely used but it is pertinent to a wide realm of argument. I think many people struggle to keep the two frames (Newtonian and Maxwellian) in play at the same time and end up defaulting to one (usually mechanistic determinism) or the other.

Take gender pay discrepancies as an example. Many feminists argue that there is a pay differential between men and women something on the order of 70 cents to the dollar, i.e. women are paid 70 cents for the same work as men. This position has long been debunked. When you add up all the money women earn, it is indeed about 70% of that which men earn. But the simple aggregation masks multiple issues that drive the difference such as education choice, industry choice, hours worked, continuity of work, etc. When you look at single men and single women of the same age without children, they earn the same amount.

If the position is refuted, why does the canard remain in circulation, and why is it so passionately believed?

I think the answer resides in the Newtonian versus Maxwellian framing.

It is indisputably the case that across an economy with some 120 million workers that there are instances of discrimination based on sex but also on race, ethnicity, age, class, region, accent, religion, orientation, manners, and myriad other factors. Not only are there instances of individual discrimination but there will also be instances of systemic discrimination. Surely that bares out the feminist position? If you frame things solely in mechanistic determinism terms, then the answer would have to be yes. Discrimination at the individual level must mean that the system is discriminatory.

The Maxwellians, in defending the system averages, argue that you can have instances of discrimination at the individual level that won't show up at the system level if the instances of individual discrimination are random. Say there are 10,000 people in an industry, all single and without children and of the same education attainment and of the same work patterns. Yes, you might be able to come up with a 100 instances of clearly discriminatory behavior against women in terms of compensation. But if there are also a 100 instances of clearly discriminatory behavior against men in terms of compensation, then the two populations cancel out. At a system level, there is no measurable evidence of discrimination.

Advocates who are Newtonian thinkers will find instances of discrimination and from those instances, mechanistically and deterministically conclude that the entire system is discriminatory. It isn't.

The Maxwellians would concede that there might be individual instances of discrimination, which ought to be stamped out, but that as long as instances of discrimination are non-material and randomly distributed, then you can still have an overall system that is clearly nondiscriminatory.

Wicked versus tame problems are real by their nature, regardless of the terminology used to define them. But they also exist because of human inclination to default to mechanistic determinism when more often, with complex systems, probability and uncertainty prevail.