Michael White over on the Scientific Blogging Site posted an article in which he exposes the largely mythical nature of the narrative of the scientific underdog repressed by an entrenched scientific establishment. I urge people to read his article but if you are too pressed for time, I’ll quote White on what that narrative precisely consists of. I am quoting White but keep in mind the context: White is critical of this narrative.
The narrative goes like this:
1. The famous, brilliant scientist So-and-so hypothesized that X was true.
2. X, forever after, became dogma among scientists, simply by virtue of the brilliance and fame of Dr. So-and-so.
3. This dogmatic assent continues unchallenged until an intrepid, underdog scientist comes forward with a dramatic new theory, completely overturning X, in spite of sustained, hostile opposition by the dogmatic scientific establishment.
White’s description of the underdog narrative seems to me to be right on the money. In his article, he discusses the mythical nature of the narrative very convincingly. Again, I urge people to read his article. What follows here is my reaction to the myth. Some of what I point out what also pointed out by White in his article so this is largely written in agreement with White but I discuss some of the same points differently than he does, I skip over some of the points he discusses in his article and I add some aspects to the discussion.
The way I understand it, all new proposed scientific hypotheses have to go through a period during which they are put to the test. During that time period, the hypotheses spread throughout the scientific community and are critically assessed. The hypotheses are usually reworked on the basis of those critical assessments and in this way become more apt to explain observations. Although initially the majority of scientists take a cautiously skeptical stance towards new hypotheses, if these hypotheses are able to weather skepticism and are found through experiments to explain the available observations, then scientists change their minds and the hypotheses become established scientific theories. This is precisely how science should operate: new hypotheses should not be accepted until they’ve been thoroughly tested. Pardon the analogy with software development but I would say that new hypotheses have to pass through an alpha and beta stage during which they are debugged by the scientific community until they finally they become “gold” and are ready for general use by everyone. This is normal and desirable. So in a sense, all scientists have been “underdogs” because their hypotheses had to pass through a debugging phase before they were accepted. If it is accepted that the status of “underdog” is normal, then the “underdog myth” loses its power because it relies on the fact that only some people are by force of dogma made into underdogs.
Moreover, the fact that it takes time to change accepted theories does not make currently accepted theories “dogma”. Usually, theories that are currently accepted by the scientific community have proved themselves able to explain current observation and in some cases have predictive power. These theories often fail in some circumstances, which is evidence that the theories need replacement or refinement but this is not evidence that the accepted theories need to be mindlessly discarded. Consider the following scenario. An accepted theory cannot account for cases A and B but can account for cases X, Y, Z. Then a new theory is proposed which accounts for A but not for B, X, Y, and Z. Ok, so case A is taken care of by the new theory. Great! But what about the other cases? If the new proposed theory is just accepted because it is “new” (and new is always good, right?) and takes care of a case that the old theory could not handle, then on the whole science has regressed. Rather than simply accepting the new theory, attempts must be made to see whether the new theory can be adapted to also account for the cases the old theory was able to handle. Determining what the new proposed theory can do, what cases it can handle, what predictive power it possesses, how it can be adapted — all of this takes a considerable amount of time and effort. The debugging process can be slow but that is not sign of “dogma” or “hostility”, it is just evidence that the problems being investigated are complex. This complexity is in fact inevitable: if any scientist is going to contribute to the advancement of science, forcibly, in this day and age, that scientist is going to be striving to solve complex problems. As White puts it, the simple problems, the low-hanging fruit, have already been picked by previous generations of scientists:
It’s a community where everyone wants to come up with the next big theory that overturns long-held beliefs. But that’s hard to do, especially in fields where all the low-hanging fruit has been picked over by really talented people for decades or centuries.
So it should not be surprising nor should it be attributed to malice if a new proposal takes time to be accepted or if it is eventually rejected. Just about all scientists, even those who are considered “geniuses”, have at some point or another proposed theories that the current scientific community does not accept. And I am not talking about cases like Newtonian mechanics which is good in some contexts but is not generalizable but cases where a scientist who is considered a major contributor to the field has been completely and provably wrong. Even the “geniuses” have been proved wrong sometimes.
Let me here open a parenthesis about a phenomenon related to the “underdog myth”. It could perhaps be labeled the “underdog theorem” (pardon my abuse of the word “theorem”) and it goes like this: “I propose X but the scientific establishment does not accept it. However, see how [insert famous scientist name here]’s theory was first rejected by the scientific community and then accepted. When [same scientist as above] proposed his theory he was an underdog victimized by the scientific establishment. I too am an underdog victimized by the scientific establishment. This proves that the hypothesis X which I propose, just like [same scientist as above]’s theory is in fact valid.” Scientists used for illustration are often Darwin or Einstein. The content of “X” is usually statements like “there is a method for building a perpetual motion machine”, “UFOs are spaceships from another planet”, “unicorns are real” or “meditation will allow you to walk through walls and fly”, etc. This “theorem” is invalid at so many levels that it is mind boggling but one interesting feature is that if the “theorem” were valid logic, then X, irrespective of its content, would always be valid. Either the scientific establishment accepts X right away and thus X is scientifically valid. Or the establishment rejects X, which is ipso facto evidence of victimization, which by the “underdog theorem” proves that X is valid. Either way X is valid and the victim is a great unsung genius.
As for the appeal of the myth, White points out how it is an attempt to get an emotional rise:
Beware the underdog narrative in science journalism. This narrative severely misrepresents how science really works. It’s designed to elicit our sympathy for a not-yet-established theory, maybe one that is socially attractive, and to arouse our indignation against the staid community of eggheaded scientists. This underdog narrative plays on our emotions, it makes for a good read, and helps us feel good about ourselves when we stand up for our convictions.
I would add that the underdog myth is also a way to make science articles more interesting. It is not by accident that science is often called “hard science”. It is hard for science journalists to accurately understand the science and report it correctly. The masses are not much interested in the scientific details which are the focus of debate because they are hard to understand. An article which focuses on the actual science rather than the human interest angle is much harder to produce and much less likely to find a large audience. The “low-hanging” fruit here is the human interest angle, science be damned! Since journalism is largely just a form of entertainment — and I do not think science journalism is any different — it is not surprising that science journalists employ dubious tactics to make their articles more entertaining.
So what does this have anything to do with peer reviews? Some questions have been bothering me during the past few months. What is the role of academic publishing? What is the value of peer review? What should be published? What should be rejected? I’ve read some published work that I thought was deficient in some areas. There certainly has to be some standard below which articles should be rejected. However, I don’t think there is a solid line which demarcates what is publishable from what is not. Because of this fuzziness, White’s article prompts me to think that peer reviewing should err on the side of accepting rather than rejecting articles. I realize this may come as “duh!” to publishers but I’m not a publisher and I do not have the interests of a publisher. There’s also a reflexive angle to this because I’m the first reviewer of my own work: at which point is my work publishable? If the bar is set too high, then interesting ideas which have not been fully worked out but which would benefit from exposure to the wider community do not get a chance to get that exposure. These ideas die in the womb, so to speak. (Pardon the gruesome image this way of speaking may suggest.) So the front door to the community has to be opened enough so that even if hypotheses do not fully explain everything they should explain, they can still resonate fruitfully throughout the community. Some of them will be modified, survive and become standards in their fields and others will be found to be so deficient that they will be pushed out the back door.
Let me conclude by returning to White’s article. He uses the example of Dr. Roughgarden to illustrate a case where the “underdog myth” was invoked. (Read White’s article for the full scoop.) Dr. Roughgarden’s theories having been in fact accepted for publication in “prestigious journals” is good evidence that there is no dogma operating here. Dogmatic attempts to suppress new challenging theories would not even allow for publications of those theories in prestigious journals. The peer review process would automatically shut the door on them. This is precisely why I think the peer review process needs to err on the side of accepting rather than rejecting. Such bias is necessary to prevent dogma.