Science is decided by committee

Scientists still do not appear to understand sufficiently that all earth sciences must contribute evidence toward unveiling the state of our planet in earlier times, and that the truth of the matter can only be reached by combing all this evidence. . . It is only by combing the information furnished by all the earth sciences that we can hope to determine ‘truth’ here, that is to say, to find the picture that sets out all the known facts in the best arrangement and that therefore has the highest degree of probability. Further, we have to be prepared always for the possibility that each new discovery, no matter what science furnishes it, may modify the conclusions we draw.

—Alfred Wegener.

We have all heard Wegener’s sad story. How all of “science” aligned against him and his bizarre, false, ridiculous, obviously false theory of continental drift. What happened, more or less, and certainly not formally, was about 100 years ago all geologists got together and voted that Wegener had lost his mind. But, of course, and in fact, they had, and from Wegener arose the fascinating study of plate tectonics.

Then there is the Rene Blondlot saga. All of “science” aligned against his, too, and his weird, silly, sad, and pathetic theory of n-rays. What happened was that about 100 years ago all physicists got together and came to the consensus that poor Blondlot had lost his mind. And, of course, he had. From Blondlot came the cautionary tale of how easy it is to fool yourself, even if you happen to be a very smart man. There are no n-rays.

I don’t want to dwell on the point here, but there is no such thing as science. There are things we know and things we don’t. There are more things we think are true, and many more we think are false. And that’s it. But the purposes of this essay, I’ll, like everybody else, use the word but leave it vague and undefined.

Now, for every Wegener, there is at least one Blondlot and certainly hordes of nameless others, each touting their own personalized, probably false theories-of-everything. What this means is that because some person touts a theory which “science” denies, it is more likely that that theory is false than it is true. Thus, it is usually rational, for example, to seek Dr Smith’s of State U.’s opinion on Joe Jones’s new theory of zero-point energy. That is, an appeal to the consensus is rational.

The opinion of a great many learned persons concentrated in one place is a good filter of nonsense and falsity. But this filter is too often applied indiscriminately and too assiduously and it often blocks truth, particularly if the truth is new and different, or it is against a vogue that has taken tight, but temporary, grip on the academic masses.

Max Planck: “A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it.”

Thus, in with the new but only after the those with the old are out. It’s often not as bleak as Planck painted it, of course. Some fields, especially when they’re young and unossified, grow by leaps and bounds, and each new idea, no matter how trivial or valid, is celebrated. It’s only after a field has had time to metastasize, that is, be formally recognized by its own separate department—complete with chairs (endowed, naturally), meetings, and new journals—that the filter becomes fully functional.

Academic freedom and its opposite

It might, then, not surprise you to learn that as a professor at a university, the locus of emanations and endless chanting about academic freedom, you cannot teach what you want. You cannot even study what you want. You can still think what you like, but, as we all know by now, you cannot always speak or write it.

What I mean by this is that peer review is not confined to accepting or rejecting journal articles. The sword is also wielded inside departments. Courses, for example, are decided upon by a committee. Many committees, actually. There is a departmental one (or more than one), then usually one at the “school”, or group-department level. There is sometimes another beyond that at the university-wide level.

Each level has to vet and approve any new course so that, among other things, it fits in among other courses, that the material is aligned with the consensus, and so on. These are reasonable goals, but the constrictions lead to tremendous inertia.

The amount of innovation (in teaching method or material) allowed in a course is inversely proportional to its difficulty. Thus, very advanced courses—seminars,1 usually taught just to graduate students and other professors—are wide open. You can teach exactly what you like, require what you like, depart on any tangent. There is little consensus about what to teach or how.

But in 101-type classes, your behavior is strictly proscribed. The book2 is decided for you, the lesson plan is decided for you, and in some places even the quizzes, homeworks, and exams are decided for you. Again, usually this is not entirely bad. The closer the field is to being driven by logic or is empirically verifiable, the more likely that the basics in that field are known, and that the optimal order and method to teach the introductory concepts have been hammered out. So, for example, all physics students should learn that F = ma, “pre-calculus” students must know that ln(exp(x)) = x, and all chemistry students should know in what way a proton, neutron, and electron are different. Which is to say, there is a consensus about what is known and what is best about the fundamentals. This obviously works against you when the fundamentals have recently changed,3 or the fundamentals are in dispute.

A within-department consensus always exist to also ensure that the work professors do is limited. Most do not feel that it is a burden to toe the line; after all, most professors are hired to work on a specific sub-sub-area within a field, and it is this area in which they enjoy working. Academics attend meetings in their specialties and sub-specialties and the areas of work which are popular are discovered. This group-think can lead to success of the kind we have certainly seen in many fields, but it also tends to narrow the scope of new work. We have all heard somebody tell us “You’re working on that? Nobody is interested in that.” And we have taken its meaning: work on something popular or your tenure or promotion will be more difficult.

The trend towards specialization has a built-in positive feedback. The more people work in a narrow field, the narrower that area becomes, or the more likely an area splits into two or more areas which also constrict. Again, not always bad, as this can lead to rapid progress, but it clearly not the best model for all people or all fields. If you are hired to do radiative cloud modeling, for example, you are not encouraged to dabble in your neighbor’s boundary layer fluid flow problem. You certainly would receive odd looks if you were to suddenly discover an interest in, say, difference equations or philosophy. You might furtively work in these areas that are “not yours”, and you might even publish in them, but you will not receive any credit for doing so, and as I said above, papers published in other areas might even work against you: “She’s not focused” is a commonly heard phrase. Which is to say, broad curiosity is not rewarded; potentially stultifying specialization is.

On being wrong

The closer a field of study is itself to politics or any area which involves human behavior, the more the consensus acts to keep people in line than it does to promote innovation. Non-consensus ideas are not welcome. Professors holding verboten thoughts are not hired, or if they are found out, they are let go, or they even leave voluntarily, tired of the process.

Naturally, the more a field agrees on what is actually true, then the stronger the consensus is to be sought. Problem is—as you might have guessed—is that people in these human-centered fields often feel, as people in more physical fields do not, in the grip of enlightenment and so always advocate the consensus stridently. The reasons for this are obvious and well known. The solution seems to be, because people in areas which involve humans are prone to ill-informed zealousness, that they should all be taught and consistently reminded that they might be wrong. This is the reason, after all, that, on average, people involved in physical areas are humbler: they have seen and verified their failures, and they have seen and acknowledged that their predictions sometimes are a bust.

Not all who work in physical areas are so lucky as to face correction. Today, there are at least two fields in which predictions are being made that either cannot be verified or cannot be verified until quite a lot of time has passed: string theory and climatology. The best these two fields can say is “Observations we have seen are consistent with our theory.” A true, or mostly true, statement. But, and I need hardly point this out, the observations can be equally, or even more, consistent with different theories, even theories which make opposite predictions. This is why making predictions is more important than explaining what we have already seen.

In fields where making predictions is more difficult, again, the human-centered or influenced ones, the local consensus is stronger, and people in those fields look more to the past to find observations which support their views. Evidence is picked over, and the best—in the sense of most agreeable—is kept, the rest discarded or explained away. The more a field is in the grip of explanation, the stronger the consensus will be, and of course the greater the chance that there will be splinter consensuses.

This is contrasted with fields in which (verifiable) prediction is king. There may be—there certainly are—splinter groups, but people can and do swear allegiance to more than one group. The consensus in these groups is more fluid and more likely to change on short notice. If there are many factions—explanations for a phenomenon—the first from which arises a correct prediction is the one that gains the most support. If that explanation can continue to make verifiable predictions, then eventually the explanation is accepted and becomes part of the consensus.

Everybody who agrees with me, raise their hands

So far we have seen that the consensus can work both for and against what is true. This should not be surprising. Research is done by people, and people have foibles. The process, on the whole, and especially in areas which do not involve human behavior, appears to be working. It is a clunky system, but it has shown results and still has promise.

The system breaks, as it always has, when people fall in love with an idea because that idea fits in with other deeply held beliefs, or when people simply want the idea to be true. When these like-minded people form a group and then a consensus, progress is halted, or even set back. These people need more experience with failure—that is, with acknowledging failure. I have no clear idea how to do this.

Naturally everything in this essay is subject to dozens of caveats and exceptions to the rule. The general theme sticks, however: people are generally too sure of themselves.


1Incidentally, these seminar courses are often taught “off the books” by the professor. Meaning they do not always count towards their official teaching load. Credit for students taking seminars is usually limited, too.

2The difference between the 101-books used in these courses is driven more by economics and fad than by fact or material.

3This happened in physics about 60-70 years ago, and is happening in statistics now.


  1. ” … people are generally too sure of themselves.”

    Except for us lowly Engineers, of course. We unceasingly devise ever more clever methods, processes, and procedures to check that we have a correct answer. We are truly where the rubber meets the road, or the cables meet the bridge, or the wings meet the air, … .

  2. Briggs


    Who but engineers experience and acknowledge failure on a daily basis?

  3. I would like to introduce some nomenclature relative to the models. methods, and applications of computer software into this discussion.

    Dr. Briggs said:

    Today, there are at least two fields in which predictions are being made that either cannot be verified or cannot be verified until quite a lot of time has passed: string theory and climatology.

    I will focus on the word ‘verified’.

    In the scientific and engineering software world we assign meanings to two different words relative to ensuring correctness of models. methods, and applications of computer software; Verification and Validation. (Actually there are several words and phrases in use, but I’ll focus on these two.) As a zeroth-order introduction:

    Verification are procedures used to ensure that the equations are solved correctly.

    Validation are processes used to ensure that the correct equations have been solved.

    Verification is basically comprised of review plus mathematics procedures and these are independent of the intended applications of the models, methods, and software. These procedures focus on the documentation and coding of the model equations and the solution methods used for these. In many cases, Verification can be reduced to a mathematics problem.

    Validation is comparisons of results, as calculated by the software and its application procedures and users, with data. Validation focuses on the physical phenomena and processes contained in the mathematical models.

    Validation of a typical General Circulation Model, GCM, relative to the long-term Global Average Temperature is considered by many to be un-attainable, maybe even impossible. While this view might or might not be correct, the following observations are indeed correct.

    Verification of all software, and I do mean each and every piece, used in the Climatology Communities is in fact possible. The fact that apparently little of the software, if any, has been subjected to Independent Verification is what many engineers (and maybe scientists) find to be extremely distressing. So far as I know this situation relative to software the results from which have the potential to significantly affect the health and safety of the public, is unprecedented on the entire planet. Although far far outside my areas of expertise, I’ll speculate that even computer software used in String Theory can be Verified.

    Validation of many of the individual mathematical models of physical phenomena and processes that make up the complete model can in fact also be Validated. We do it all the time in engineering software.

    Software Verification and Validation (V&V) and Quality Assurance (SQA) is SOP in every other endeavor for which the health and safety of the public are potential issues. Every one. The Climatology Communities seem to me to be getting a pass on these critical issues. Application of V&V and SQA to even the most simple and straightforward methods seem to be ignored.

    I have found it very interesting that whenever I attempt to bring up this subject I discover there are two groups that enter the discussions. One group that uses the processes and procedures on a daily basis and know the value added attained by their use. And a second group that has never used them and declares, “These processes and procedures can’t be applied to our software. It’s impossible.”

    Much more on these critical issues can be found in this book by Pat Roache. Additional, and more recent, information is available in reports and papers from a group at Sandia National Laboratories. This Google search will lead you to tons o’ info. Or go over to and check out the Information Bridge and you’ll find free electronic copies of almost all the Sandia reports.

    Extended discussions of several of these issues are available in this post. While you’re over there kindly take some time to look around and you’ll find other discussions of Verification, Validation, and Software Quality Assurance along with some related stuff.

    “Software development, especially for applications that have vanishingly small probabilities but extremely significant consequences, is a cruel, cruel mistress” me 2008

  4. Briggs


    You’re a treasure. Thanks for all the (necessary) clarifications.

  5. Noblesse Oblige

    It is an epiphany indeed that fields where prediction is most difficult tend to generate the strongest consensus. Let me add a third field to the examples of string theory and climatology: evolution. I run into many people who do not believe in evolution who also viscerally object to global warming. They believe that both are examples of science run amok and that a powerful group of arrogant ‘them’ are at work to force both down our throats. They are conspiracy theorists who won’t be convinced by fact. BTW these people do not deny that changes take place within species via evolution; they challenge whether evolution is responsible for the generation of NEW species. They do not necessarily accept the alternative of devine intervention but typically just acknowledge that they don’t know. And they resent being characterized as ‘kooks’ or ‘deniers.’

    In any case, in contrast to climatology, evolution science is to be credited on two fronts: (1) They do not attempt to mathematically model the future course of evolution; and (2) Research on fundamental biological mechanisms for species change is still not settled (spontaneous mutation doesn’t seem to cut it) and goes on to this day, avoiding the dreaded consensus. Contrast this to the climate modelers who make more runs of ever more complex models without a fundamental theory of accuracy for the models themselves. And this while avoiding the committment to the thing that models do: prediction for verification/validation. And at the same time they announce that they have all the fundamental mechanisms in hand, leading to positive greenhouse gas feedbacks but none for solar variablility and ocean cycles. Arrogance plus confirmation bias run wild. The anti-evolutionists have a point.

  6. Please expand climatology to ALL environmental sciences. They are ALL so filled with junk models, glaring anomalies, and utter lack of validation that it is a wonder the practitioners are called “scientists” at all.

    One thing you can say about paradigms: shift happens. The old guard have habit of dying off on a regular basis, so even if they hold claptrap theories dear, those theories get regularly buried along with their proponents. It might behoove Science to retire any and all scientists over 40-years-old, like baseball players past their prime, but then who would write the text books? The younger set does the science, the older set writes about it.

    This points out another difference between religion and science. Religion is about tradition, science (good science) rejects tradition. Which is why good scientists should always be religious, too. They need to understand that tradition has an important social function outside the laboratory.

    Another problem facing Science is that it has been institutionalized. Every problem you note is related to the defects of institutions. Outside of institutions the limitations are much less onerous. As a culture we tend to reject non-institutional learning and research, but that’s a mis-perception and and a hang-up we should shrug off. Free the body, and the mind will follow.

  7. Sylvain

    I raise my hand.

  8. Luis Dias

    These people need more experience with failure—that is, with acknowledging failure. I have no clear idea how to do this.

    Excelent text by mr Briggs and also an excellent comment by Dan Hughes. To answer the above question, there’s nothing that you can do. Reality will do it. Eventually, climatologists will fail. And faced with the danger of extinction of their field, they will have to improve their models. Because the climate is so long, I think we are into the beggining of a long history of climatology before a real “consensus” of how the climate “really” works is settled.

  9. My dear Mr. Briggs, hyperbolic or not, you go too far by miles in saying there’s no such thing as science. That’s a dangerously relativistic thing to say, and I do hope it’s beneath you. On the contrary, in fact, science is relatively easy to define: systematic knowledge gained by reason, in accordance with the law of non-contradiction (i.e. logic), and based upon observation. But science, as we’ve discussed before, is hierarchically dependent upon an even more fundamental discipline, and that discipline is epistemology.

    Likewise Mr. Hughes thoughtful discussion of validation and verification, both of which words, even in the narrow context he’s using them, are epistemolgic in the ultimate purport. Indeed, they are a species of the genus epistemology.

  10. Briggs

    Thin man,

    Oh, my, no. I am as far from being relativistic as it is possible to be. And it is because of epistemology that I make this statement.

    Your definition is incomplete. I’d say something more like this “knowledge gained by reason, in accordance with logic, based upon intuition and assisted by observation.” That is, there are certain things we know are true based on no evidence but our intuition (the a priori), and others things that are derived from that (logic, math, some ethics, theology, etc.), and others still assisted by observation (everything else).

    I agree that the word science is useful as shorthand for certain fields of study. But knowledge is knowledge, regardless of the subject. In the end, what matters is whether a thing is true or false, and that’s it.

    We probably aren’t too far away on this.

  11. This post, by a Certified Climatologist, concerning enormous consequences from a failure, says this near the end:

    It is worrying to think that there are supposedly qualified engineers out there who don’t know this stuff like the back of their hands…

    Yep, same goes for Certified Climatologists and the fundamentals of model, methods, and software development.

Leave a Reply

Your email address will not be published. Required fields are marked *