Who Decides What is True?

How Does Science Proceed from Hypothesis to Fact to Common Knowledge?

Simulation of the asteroid impact that killed the dinosaurs


There’s no Supreme Court of Truth, no supreme authority that affixes an imprimatur of “scientific fact”.

Yet we believe many things to be true which we could not have known about without science. It’s obvious that science can draw conclusions which are effectively certain, but it’s less than obvious how this happens.

For example, we believe that the Milky Way is the vast cloud of solar systems of which our own is a member, itself one of a vast number of galaxies in the universe. If you encounter someone whose beliefs about reality require this to be false, (a flat earth for instance) you are justified in dismissing those beliefs without further consideration. A worldview based on something contradictory to established fact is one that is not viable.

But this fact, though common knowledge nowadays, was unheard of two centuries ago (when, apparently, Immanuel Kant first proposed it). How did what was once a wild speculation turn into an established fact?

It Matters

It turns out that this isn’t merely an academic question.

We’re in an increasingly complex world, which requires increasingly complex collective decisions. If we avoid magical thinking, it seems clear that getting the facts right is, while hardly sufficient, at least necessary.

Nowadays, the public is constantly subjected to “zombie arguments”, recycled scientific controversies, long-settled within science, which retain their currency in the public consciousness. This confusion derails efforts to act effectively, even when the information available may be unambiguous within science.

The public is also somewhat confused about how science establishes facts. Some believe that once a paper passes peer review, it is established as “true”, on an equal epistemic footing with every other paper unless proven “false”. “False” papers are then considered a mark of shame, of negligence if not actual dishonesty. “Bad” papers are considered “refuted” or even “debunked”.

This is actually not far from the truth in publications in some engineering fields. One reports on actual achievements; one avoids lies and error. There’s little room for honestly, competently wrong in reporting the efficacy of an invention.

In a science in its most delightful phase, though, when much of consequence is known and much remains unknown, a number of different explanations for phenomena may be considered and defended. Barking up the wrong tree is a perfectly legitimate contribution, as much can be learned from considering whether the prey (the bird of Truth?) is there or elsewhere.

But eventually, truth does emerge.

A Recent Progression from Hypothesis to Common Knowledge

We can consider a recent example — the speculation that an asteroid strike was responsible for the Cretaceous-Paleogene extinction, first proposed by Luis and Walter Alvarez in 1980. There was plenty of back and forth in the literature and at paleontology conferences for thirty years, but this theory was eventually considered established.

Interestingly, there was actually a sort of formal declaration of victory in this spectacular case, a paper jointly authored by many of the most respected practitioners of relevant sciences. One thing to take from this is that there are papers and there are Papers; some are more influential than others, based on the scope of the research and the backgrounds of the authors. Often the most important watershed papers have multiple authors from multiple research groups. As in the asteroid case, they sometimes exist merely to ratify a consensus that has emerged within a field. Such a report (and arguably IPCC reports on climate change are in this class) cannot create a consensus. It can merely muster the evidence for its acceptance.

The consensus emerges within the relevant scientific community, which then presents it to the world.

Formalization of Consensus Follows Creation of Consensus

Massively credentialed review articles, and the next step up, formal consensus documents, appear mostly in cases where there is a community of practice or a segment of the public that is interested in the results.

But consensus appears first, sometimes invisibly. Important generalizations can emerge without representation in formal communications at all.

It’s notable how informal the consensus process can be.

For instance, during the (all-too-brief from my point of view) period when I was privileged to attend paleoclimate lectures at the University of Chicago, it became entirely clear to me how central CO2 greenhouse warming variation has been to the climate trajectory of the Earth. I found myself making the claim that “you really can’t sensibly paleoclimate without the greenhouse effect of CO2 — it just would drain all the sense out of the field” in an online conversation, and was challenged to defend it to a reasonably scientifically literate non-climatologist. I struggled and came up empty.

Fortunately, some years later, Richard Alley stepped up to the plate, in this, one of my favorite climate lectures of all time: “The Biggest Control Knob: Carbon Dioxide in Earth’s Climate History”. (Set aside an hour to watch it if you have any interest in climate of any sort!) So it’s not as if I didn’t have a point. I just couldn’t find a credible source for the claim, a claim which from where I was sitting “everybody knows”.

It hadn’t ever been formally asserted in peer review. Yet it was, and is, an established fact as far as science is concerned.

Consensus as a Social Process

When we consider how a hypothesis becomes accepted as a fact, it seems there are two phases to consider: 1) how does a genuine scientific consensus emerge within the community of experts and 2) how is society to recognize it when it does?

I argue that both phases are neither as purely logical as some would hope, nor as purely social as some adamantly insist.

In fact pure evidence is weighed in a highly social process.

Such weighing still works well enough in the sciences, though it is under systematic assault. But the second phase, the transition of socially relevant scientific consensus to public consensus, appears to be breaking down.

Politics Disrupts the Consensus Process

Once a certain level of coherence of evidence for a proposition exists in a real science, it is as good as true for any practical purposes.

In the ordinary course of science, the blind alleys are quietly abandoned, and the more stubborn advocates of doomed theories, who refuse to adapt to a new consensus, are allowed to age into emeritus status with a modicum of dignity.

This social process is adequate and, at least arguably, civilized. For most purposes it works.

Recently, though, when policy is at stake and interests are threatened by consensus science, the scientific consensus is challenged. Then, less scientifically influential, less effective scientists can become important.

People who continue to advocate theories that are essentially outside the scientific community’s collective view of plausibility can suddenly attain a market value. That is, if a scientific result suggests a policy change, those whose interests are directly threatened are motivated to find credentialed scientists who appear to disagree.

Of course, credentials are not really a proof of credibility. While the most active participants tend to have excellent credentials, getting published is more a matter of tirelessness and persistence than of brilliance.

And so rather than fading into obscurity, the careers of some second-rate scientists get a boost. They are formed into a sort of ad hoc committee of denial.

Agnotropist Science

I don’t say this because they are “denialists” in the usual sense. They are not advocating a position in defiance of evidence. “Denialists” in that sense are their customers. You will not find the climate scientists most beloved by the Republicans in congress taking strong political positions, though their actions are invaluable to those that do.

You find them, instead, asserting that “science does not know”, that many things are “unclear”, that there is great “uncertainty” (a word whose technical and nontechnical meanings they happily conflate).

I call this approach to science “agnotropism” (following Oresekes’ definition of the study of cultivated ignorance “agnotology”, the study of confusion). It is adaptive for people with more credentials than talent to take positions that science “does not know” this, that or the other. Essentially their product is confusion. They find assertions of confusion and inconclusiveness attractive. It gives them a willing and enthusiastic audience, and a global platform, even though what they are offering is really the opposite of science.

It’s really not that hard to construct a null result if you set your mind to it. If all else fails, you can always reduce the size of your dataset!

It’s woefully outside the norms of science to say something like this, and it’s woefully ad hominem, but in essence the central fact of the matter is that they are not very good.

They are not very good. They don’t share in the consensus because they don’t participate in the consensus process. They aren’t part of the consensus process because their contributions are internally incoherent as well as at odds with the extant consensus.

The traditions of science are clear. One feels sorry for their students and postdocs. One silently turns away. One is scrupulously polite at receptions and meetings but has lunch with someone else.

Science’s Habit of Polite Indifference is Failing Society

Agnotropism rewards production of papers and positions that are, really, the opposite of science. They celebrate null results, and happily confuse lack of a proved connection with proof of non-connection. They do this in a way that is logically vapid but rhetorically compelling. Loose journal standards let this through.

The few people producing this nonsense are celebrated. The thousands participating in a vigorous and sound scientific process generally, with a few exceptions, attain to far less attention from the public and the policy sector.

If you were to poll the public looking for someone who could name a climate scientist, the names of the dozen chief purveyors of their own honest bafflement would be at least as likely to come up as the names of any of the most brilliant contributors to the science.

Dozens of half-baked arguments over the years are enough to sink a scientific reputation within science, but can even serve to promote it with the public.

We have to go outside our deeply bred politeness and forgiveness and avoidance of the ad hominem. Some people are just not worth paying attention to. Science knows this, but is too polite and circumspect to say it.

Consensus Studies Confuse the Issue

When you are a legitimate participant in a vibrant science, you know what the science says. The idea of polling the scientists or of counting publications, perhaps politically useful in counteracting the science-of-bafflement strategy, implicitly leads to a perception that consensus is a rare phenomenon in science.

In fact, it isn’t, because the consensus process isn’t just about certainty.

There is an ongoing consensus about what is certain, what is likely, what is possible, what is “maybe close to the truth” or “maybe part of the story”. There is rarely much disagreement on the big picture that divides a mature science.

Peer review isn’t just about publications. People who respect facts want to earn the respect of other people who respect facts. It’s intensely social. Consensus is the zeitgeist that emerges within specific scientific communities. We don’t spend our days thinking about this consensus, any more than a fish spends its time thinking about water. Consensus is the water in which we swim.

Consensus Isn’t Enough

The press, the public, and the policy sector have to make decisions based on real evidence. Yet we live in a world where the technologies of fake evidence are well developed. How can we collectively develop antibodies to bogus arguments?

I really don’t know. I would like to point out, though, that an excessive dependence on the consensus process is as dangerous as ignoring consensus altogether. I don’t care what the consensus among homeopaths is.

More crucially, I don’t care what the economic consensus is among people who believe not only that growth is indefinitely sustainable but that such growth is inevitable. (Much of the economic analysis of climate change is based on exactly this assumption, even though that is the key question at issue!)

We need ways to establish the credibility of a field and the credibility of its critics. I’m not sure I have any proposals.

But just as the emergence of consensus within science is a combination of the social and the logical, I think the same must be true of some future democratic world that takes better decisions than our own.

Those whose focus is on “cultural cognition”, who basically advocate the reduction of science to politics, and those whose focus is on logical argument who basically advocate the reduction of politics to science, each tell only a small fraction of the story. If there’s an answer, it’s far more complex than either of those.


Also published on Medium 

This article has been translated into Portuguese.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.