Is novelty overrated in science?

11 September 2017

Interview with

Barak Cohen, Washington University

Each month in the eLife Podcast we devote some time to trying to look beyond just the results of scientific endeavour and to consider also the practice and some of the societal aspects of science. Previously, we’ve heard about diversity in the discipline, how women and early-career scientists are faring in the modern scientific era, and how as a community we need to avoid distorting the facts. This time, speaking with Chris Smith, Washington University’s Barak Cohen has something he needs to get off his chest...

Barak: There are a bunch of reasons I wrote this. First, the accumulating number of rejection letters that I was getting for both my grants and papers which were stamped as being not novel enough, so that’s just to admit to you that I'm a human being and fallible. But the other reason that I wrote the piece is that I'm becoming more and more interested in issues around the philosophy of science and in particular issues about why science seems to work so amazingly well especially relative to other human endeavours.

Chris: Let’s talk about this rejection letters first.

Barak: Sure.

Chris: Let’s get that out of the way. What is the scale of that problem in the US, and also beyond the US, worldwide?

Barak: Well, my impression is that it’s large. I mean, if I'm representative of a typical scientist then I would say, probably at least half of the rejection letters I get coming back are lack of novelty as part of the reason. In terms of grant proposals, I would say it’s got to be at least in the 50% range. I see that from the other side because of course, as a working scientist, I not only submit grants and papers, I also have to review papers and grants. When I'm sitting in the room on a review panel, the review panel talks a lot about novelty and novelty is a big reason that grant proposals get rejected these days.

Chris: But why shouldn’t we be striving for novelty because isn’t that partly why we’re all in business? We want to know how the world works. We want to discover new things, so it is important that we prioritise that.

Barak: Yes and that science has been amazingly good at uncovering novel phenomena. If you just look at the pace at which new discoveries are being made in molecular biology and have been being made in molecular biology over the last 50 or 60 years, the pace of novel discovery is incredible. And that is without an explicit emphasis on doing novel research. I guess part of what my piece is about is that there really isn’t a need to explicitly reward proposals that are novel. Novel lines of research are important, but it is also important to explore things deeply. So, with a finite amount of effort, you have to apportion that effort in ways that both open up novel lines of research and encourage people to probe deeply into those phenomena.

Chris: Do you not think that there might be a historical bias here in the sense that a lot of the novelty was novel because it was easy to stumble upon, relatively speaking? And that now we’re in that law of diminishing returns where the curve has begun to flatten and were having to invest more and more, and more to fill in the gaps because the landmark leaps have been made.

Barak: Always a dangerous thing to say but I do sort of think that. When a field is young - and molecular biology and molecular genetics is still a pretty young field - in the beginning, there are lots of foundational principles to uncover and big discoveries are made quickly and rapidly, and then you do get to a position of diminishing returns. If you're thinking about it that way, if you're thinking that you know, in order to be successful, I have to uncover a foundational principle. I guess, part of what I’d like to argue here is that you can be a very successful molecular biologist by taking principles that we qualitatively know about already and putting them in a quantitative framework that makes them predictive and useful.

Chris: Well go on. Explain what you mean by quantitative model then. Give me a tangible example of how that would apply to molecular biology.

Barak: One of the things that I work on in my own is transcriptional regulation. If you ask me how does a gene get turned on and off at the right time and the right place, the language that I'm going to be using is mostly qualitative. And so, it’s very difficult to take that qualitative understanding and turn it into a quantitative prediction.

Chris: Do you think that we need to be prescriptive about that or will it not just happen anyway because as people are opening up new avenues and discovering new things? they will gently fill in the gaps behind.

Barak: I think we do have to be prescriptive and encourage people to think about quantitative models because it does not seem to happen naturally in molecular biology. My adviser used to tell me that, “you can conquer the world as long as you have an internal control. You can always tell where there's a difference.” But what was missing in that statement is that it might matter whether it’s a two-fold difference or a ten-fold difference, and how we incorporate those sorts of thinking into model building. I don’t think that really comes naturally to molecular biologists and particularly in the types of training that graduate students go through during their PhD.

Chris: They’re not talking much about philosophy either and there's a big chunk of that in the article you’ve written. I find it quite refreshing.

Barak: Yeah. Well I think maybe there's a reason the degree that we get is called doctor of philosophy. It’s not actually called doctor of biology or doctor of chemistry. It’s called doctor of philosophy and I think the reason for that is that the really important skills that graduate students learn as they go through their PhD are not actually how to do a PCR reaction or how to write a particular piece of code. It’s how to breakdown a problem and break it into manageable pieces and how to know when you’ve collected data that really answers that problem and how to synthesise data into a coherent argument. Those are really the important skills that you'll learn as a PhD student and I would say that those come straight from philosophy.

Comments

Add a comment