Underdetermined March 12, 2021


I never have a pass on making a bad argument. No scholar should claim more certainty than is warranted by the evidence.

Explanations are easy, but sufficient explanations are hard. We can show why something might have caused something else, but showing that it did? Not so easy.

Even a scientist must take care. WVO Quine suggested, and I think he was correct, that many ideas in science are underdetermined by the evidence presented. The amount of certainty claimed is too great and there are logically incompatible theories possible that equally explain the data for any given theory.  This does not mean any speculation is equal to a theory that scientists have tested and carefully examined. Other theories are possible, but that does not mean there are other equal theories right now!

Quine’s idea matters, though. Even science must be careful about claiming too much for a theory or law and science is the most certain way of knowing we have. This is not so bad! Certainty and the hubris that goes with it is the only loss. Philosophers cannot, of course, comment on which particular theory of science better fits the evidence. They can note when a theory or idea is radically underdetermined by the evidence presented. If one says “X is true!” then there should not be obvious counter-examples or undue certainty.

The need for epistemic humility is even greater when we do cultural analysis in the humanities, particularly when our conclusions agree with the prejudices of our peers. Taking on people one does not like, and one’s colleagues do not like, is dangerous as overstated conclusions will be harder to spot.

The medium matters. Naturally, a simple editorial cannot take much into account and will tend to be punchier than a book. Still even then, nobody should claim too much. Some kinds of atheism have a problem with objective morality, but not all kinds do.  Christian apologists have a habit of claiming too much for an argument, but critics do as well. A book, or an academic paper, must do better. Great claims require careful examination of causal connections.

Piling up correlations only helps if obvious counter examples are explained. What if you showed, in a narrow field of data, that one idea correlated with a bad outcome! Imagine looking at a scandal and finding in it the confirmation of your own views. If only they had listened to you, then much evil would have been averted!

This happened recently in the Catholic church when people found confirmation of all their prior beliefs when things went bad. Yet when things were bad in similar ways in Hollywood with very different people,  a new set of assumptions, totally incompatible with the first, were confirmed to a new set of writers as the cause of the vice. This goes on forever, so long as basic plausibility is met, not a hard task for a clever rhetorician.

What accounts for each, what is common to all cases? We need not say, if we limit our little monograph to one community and to monomaniacal, singular, solutions that fit the intended readership. Blame the Marxists to the conservatives, the Evangelicals to the liberals.

And after all, many are helped by calling out bad people, since evil is really evil and any answer to why, even one radically underdetermined by the evidence, will comfort the hurting.

If anyone points out that, correlation is not, after all, causation, then we might appear (God forbid!) to be denying the problem.

Yet many bad things seem correlated with causes: As we eat more ice cream, drowning deaths increase.


Perhaps, there is another reason, a slightly deeper one related to rising temperatures, that would explain why eating ice cream and drowning (say in a pool) would be correlated without one being the cause of the other. Recently we had two kids get sick in the same class, but investigation showed that the cause of either student getting sick was not the other one. What seemed obvious, was wrong. This is difficult to accept if there is an ideological commitment to the correlation being caused by something we think (already) is bad.

Imagine though that there was an ideological group opposed to the consumption of ice cream. They might write on “Drowning and Ice Cream” and, having given a reason for the pain of the bereaved, sooth many people. To challenge the correlation is to appear insensitive to the drowning victims.

The argument is still bad and the “cause” still underdetermined by the evidence. Comfort is good, but false comfort will not last. A brilliant scholar may suggest, tentatively, a connection, a cause, for a social phenomena. There are so many possibilities that one must dare a theory! If I were to state;”atheism is the cause of genocide,” list the many examples that exist where atheist regimes committed genocide, but then ignore the many examples where theism led to genocide, then I would have gone too far.  Atheism and theism both seem to be capable of genocide.

Modesty is the best policy. One tell is always how much the work reads like a screed.

This is not an age where philosophical modesty pays. We are tempted, instead of presenting a modest hypothesis, to argue that those Trump voters, or those Biden voters, are the problem for sure.

Do some stand explaining history crying: “Pay attention to me?”

Browse Our Archives

error: Content is protected !!