Friday, December 11, 2015

Should people get credit for papers that are influential but wrong?

A colleague once told me a story about his research field.
"Ten years ago Professor X got some surprising experimental results. He then made bold claims about what this meant. Some people did not believe it. But, people then did detailed experimental and theoretical work to test his results and claims. They basically found that he was wrong but in the process they made some valuable and interesting discoveries and clarified several issues in the field. To half the people in the field he was a hero and to the other half he was a pariah."
The hero status was assigned because if he did not exist or had not made these claims, the new discoveries would not have been made (or might have been made much later).
The pariah status was assigned because he did not do careful scientific work and misled people.

How much credit should people get who open up new scientific directions with “wrong” papers or with unsubstantiated speculation?

Different people I talk to have quite different views about this.

My view is that such people should get very little credit, particularly if their work is sloppy and/or they engage in hype, self-promotion, and unsubstantiated speculation. 

On balance, I think such individuals have a negative overall influence on science. This problem has been compounded by the speculative and hype culture enhanced by the rise of the luxury journals. Rewarding people for doing bad science is just going to promote more bad science. Maybe one in fifty bad papers will have fruitful consequences. But the other 49 will waste time and resources and create confusion.

What do you think?


  1. Newtonian gravitation wasted a lot of people's time. Everyone should have immediately known that it was wrong. The optics, too. We should all be so lucky to make such wrong moves.

    A paper that needs more thought to realize that it's wrong than is given to it by a journal editor and by a referee is OK.

    One person's sloppy is another person's Pollock. If half the audience like it, you might be doing OK.

    If you're not confused, you're not doing research, you're writing it up.

    On balance, we'd all be better off dead already.

    But, as you say, different people will have different views on how to play Science. It can be beautiful and rouse our passions, but sometimes a player wins a few games disgracefully.

    A sloppy set of the proverbial, as it were.

    1. Peter,

      Thanks for the comment.
      I don't think Newton wasted a lot of peoples time. I would also say that he was correct. The theory is valid in a well defined limit.

      The "wrong" papers I am talking about are not correct in any limit or under any circumstances.

  2. I was told that back in the heyday of high-Tc, Bob Schrieffer gave an after dinner talk in which he talked about the space of theories in the cuprate world. He placed theories on a 2D map where the axes were right/wrong and interesting/boring. Presumably things like interlayer tunneling were put (by Schrieffer) in the the wrong-interesting corner and some theory which posited that high Tc was caused by some jacked up conventional coupling constant BCS-like theory were in the boring-wrong corner. I think Schrieffer felt that you should get credit as long as you were out of the boring-wrong corner. Of course you got the most credit if you were in the interesting-right corner.

    It seems like this is a good method for credit/blame, as long as you also add something like a "competency" axis. So people who are right for completely the wrong reason or by accident don't get as much credit as someone who did their experiments correctly and interpreted them fairly or as theorist didn't pull a “then a miracle happens" step in the depths of some calculation.

    I think it is fair to say that there are very prominent people in CM theory who have NEVER predicted (or even post-dicted) the results of an actual experiment correctly. But their theories are interesting and their calculations competent (in their context) and so they should get a lot of credit. The person that Ross refers to is apparently interesting, but wrong and incompetent… and there is only one thing worse than that.

    1. Feynman had it right: "If it disagrees with experiment, it’s wrong. And that simple statement is the key to science. It doesn’t make any difference how beautiful your guess is, it doesn’t matter how smart you are who made the guess, or what his name is… If it disagrees with experiment, it’s wrong. That’s all there is to it.”

  3. I tend to agree with Peter.

    However, given the fact that science progresses by falsification (not by proving something right!), being wrong has a role in the process: being proven wrong shuts doors, which is a good thing as it focuses one towards the right path.

    And there's only one way to be proven wrong: by proposing something that's wrong.
    But there is no way one could nowadays write a paper saying "one could do this (something novel) but it's wrong": if you propose something for the first time, and determine it's wrong, you don't publish it.
    Extreme example: "I propose to invert gravity. If we do that, then apples fall up. But that's wrong."
    Obviously that does not fly (pun intended).
    (It's different if you say "people do X, but if you take the logical next step to X+1(novel), it turns out to be wrong, and thus X can't be right.)

    So credit should be given for people that were wrong, but triggered ideas that advanced the field - but only if they were wrong without being stupid. I.e. without being wrong when they could have (reasonably...!) known they were wrong.

    As always the operative word is subjective (reasonably). But that's human...

    My $0.02

  4. I would say, as long the 'doing wrong' is super-personal (does not involve personal interest), it should be fine. Aristotle had been wrong in a lot of things for a very long period, but he deserved plenty of credit. No theorist can be 100% confident in his theory, even for the best minds.