A meme that seemed to run through much of the comments on Jeremy’s recent post on salesmanship in science seemed to be that you could be a wonderful scientist but a terrible communicator of your science and that you would suffer for this career-wise and that would be unfair. This came as a surprise to me. I have a hard time thinking of people who I would call a great scientist but a terrible communicator. Now they may have stage fright and give a bad talk, but write great papers (or vice versa). And they may be bad networkers or bad self-promoters. But the sterotypical genius with ground breaking ideas but who drools and can’t put two words together let alone coherently communicate what they’ve done and why it is important, no. Which leads to the deeper, more philosophical question, if there is “good science” inside somebody’s head and it can’t get out, is it science? Hence the allusion in the title to the zen koan about a tree falling in the forest. Or if somebody is shipwrecked on a desert island does research for 10 years and then dies and their notes decay before they are found, have they done science?
I am certain most of you will say of course, absolutely they have done science! But I want to probe this by arguing the opposite position, that they have not done science. Of course to discuss this one must define what it means to “do science.” To my mind some of the most interesting work on this is what is known as the demarcation question – i.e. what distinguishes science from pseudo-science like ESP or alien abductions? Popper and his student Lakatos both tackled this question but came to rather opposite conclusions.
The thing Popper and Lakatos both agreed on is that real science has ideas that are testable (and specifically can be shown wrong). This is often oversimplified in popular discourse to the idea that the “scientific method” makes science what it is.In reality the scientific method is no simple formula, but a complex twisted path. Indeed the essence of that complexity is why I will argue that uncommunicated science is not science.
For Popper it was fairly black and white. A hypothesis makes a prediction which can be falsified. Repeated attempts at falsification fails.Evidence accrues to theory. A single failure throws out the hypothesis. It is at least conceivable in this view of science that you can do science by yourself without communicating. But this is overly simple. Indeed, I blame Popper for the cartoon version of the scientific method found in elementary schools around the world. As many philosophers of science since Popper have shown the real scientific process is much more complex.
Even Popper’s student, Imre Lakatos argued for a more complex view. Lakatos argued for the idea of research programs. And that specific tests of research programs can sometimes fail without requiring us to throw out the research program, so it is the total balance of evidence that counts. And Lakatos very explicitly brought in the human element. In his view, ultimately the metric of balance of evidence is what research programs scientists choose to invest their valuable time and energy in. Lakatos did not live in a classical Greek, Platonic, idealist world where perfect truth exists and advances in crystal clear steps. In short, I would say, Lakatos lived in the real world. And in the real world science is a muddy messy and ultimately social process. Kuhn and Feyerbrand later doubled down on this notion of science as a social process. It is indeed a major theme in modern philosophy of science.
Here’s a question? If you want to answer the fundamental question, ” what does science know to be true today?”, can you do it without falling back on ideas of consensus or textbooks (or Platonic ideals that can’t be objectively agreed upon in this world)? No, you can’t. There is an inner core of knowledge that “all reputable scientists agree on” but there is a fuzzy boundary that is actively debated. And how do things move from that fuzzy boundary into the “well-known core” or off into the graveyard of dead theories? It is a social process. Make no mistake the consensus can be and not infrequently is wrong. But science does have self-correcting methods for that in the long run, and they are also social.
Take the example of the 2005 Nobel prize in medicine awarded for the discovery that a bacteria caused stomach ulcers. In the early 1980s Marshall and Warren developed this theory based on correlation – H pylori bacteria were always found in individuals with ulcers. They proposed this and were roundly ridiculed. “Everybody knows bacteria can’t survive in the acid environment of the stomach”. Failure ensued – many pigs were forced to drink H. pylori and nothing happened. Then in an experiment that was probably as important for its storyline features as its decisiveness, Marshall drank H. pylori himself and instantly got gastritis (precursor to ulcers). The fact that he was brave/alturistic enough to risk his own body and that it was on a human overrode the narrative of repeated failure on pigs. We now know there is considerable variation in humans in immune resistance to H. pylori and Marshall apparently didn’t have much immunity. It is entirely possible that if his partner Warren drank the H. pylori nothing would have happened. Much follow on research and gradual convincing of the skeptics, and now we no longer blame ulcers on spicy food and a hedonistic lifestyle but treat simply with antibiotics and a Nobel prize came in 2005. At what point in that story did we “know” that H. pylori causes ulcers? And can you disentangle that from the social processes? I would say definitively not!
The theory of continental drift has a very parallel but even longer story. Wegener proposed it in 1912 but was ridiculed (in part because he was a climatologist rather than a geologist). At one geological conference (I believe it was in the 1950s), they actually asked the room to stand up and walk to one side of the room or the other to vote on whether people believed in continental drift. The idea was roundly rejected (and remained out of textbooks). Then in the 1960s a number of phenomenon relating to seafloor spreading uncovered by paleomagnetism (a new technique) rapidly changed minds. And by the 1970s it was in all the textbooks. When did continental drift become a scientifictruth? and can you disentangle this from the social process? Of course not.
So if you cannot disentangle science from pseudo-science, or determine scientific progress and ultimately scientific truth separate from the social process scientists go through as a group, can a scientist do science that remains uncommunicated (or terribly communicated)? I would say no. Now you can invoke the idea of the genius but anti-social, illiterate scientist who has a translator. But really, how many examples of this can you come up with? Too much is lost in the translation. The most forceful communicator is likely to be the person with the insight.
So I would submit that science is inseparable from communication and the social processes that make it science. Yes you can be a good scientist but an imperfect communicator. Even a good scientist with noticeable flaws in some aspects of communication. Almost all scientists likely fall in this category. But can you be a great scientist who can’t communicate. If its just running around in your head and stays there, is cience occurring? I would say no.
The three main implications of this are:
- As scientists we need to take our communications skills seriously and constantly work on improving them. And as teachers and mentors we need to devote serious attention to teaching communication of science.
- And for me at least, to circle back to the original question at the top of the post, it is definitely not “unfair” that somebody who is a bad communicator does not advance in science. It is a central part of the job.
- This is the real reason I push every student to publish their dissertation even if they’ve already landed a job in .e.g. a government agency where they don’t need publications. Its not science until its published.
What do you think? Do I have a point, but I’ve gone too far? Can one do science in one’s own brain and never communicate it?
Dynamic Ecology 
I think you’re trying to have your cake and eat it, too. If science is seen as fundamentally social process, then – like you lay out – the communication between scientists is crucial part of science. But if science is truly social process, how can then an individual scientist be blamed on being bad communicator? Isn’t it also a responsibility of other scientists, who are more fluent in communicating the research, to do their part for the dissemination of ideas from other scientists? What I’m trying to say is that if science is truly a social endeavor, then it’s not fruitful to pit against different skills of scientists. But, then again, that’s too much of a idealized picture of the scientific world. As competition is high, it’s also personal clout and visibility which can bring in career opportunities and funding. Trying to simultaneously maximize personal benefits for career advancement and general advancement for science is in most cases impossible.
Nowadays assessments want to maximize the impact. But obviously, there are different kinds of impact (writing scientific papers, refereeing, editing, writing textbooks, popular science, teaching, mentoring, policy etc.) and it’s really up to the funders and research institutes to decide what kind of impact they want. That’s obviously complex political decision. And it can undermine the social nature of scientific trade – if it puts too much emphasis on career advancement of individual researchers, has a poor balance on different scientist’s skills or puts too much emphasis on one aspect of impact.
Interesting post. My view (as an academic writing coach) is that you don’t know something (for scientific or academic purposes) if you can’t communicate it to other knowledgeable people. You don’t have to be a very good “communicator” (in the popular sense) to do that in most disciplines. So, yes, you should publish, but I don’t think scientists really need to take communication more seriously than the sort of communication that establishes a connection to their peers. That’s more than just knowing how to get past peer-review, let me hasten to add. But it’s not the sort of communicative savvy that gets you into the pages of Scientific American or on the stage of a TED talk. It would be bad for science, I think, if someone who’s a poor public communicator was held back in their career in science. Also, if someone’s public PR blunder cost them their job.
Yes – good distinction. Not everybody needs to be able to talk to reporter and the general public. It is the communication with peers that is central to the scientific process.
Then you have to hope some people are willing and good at communicating this on to the general public, but not everybody needs to do it (although expectations are rising in this area).
I think you are both right and wrong (how’s that for Zen). Depends what you mean by “science”. If “science” is a process by which an individual can find out things about the universe, then yes, one can do science without communicating it – when I compare loads of laundry washed with two different kinds of detergent, I’m thinking scientifically. You are quite right that this process works better with multiple people doing and comparing notes, but that’s efficiency, not definition.
But if by “science” you mean “rationally derived human knowledge”, then no, you can’t do science without communicating. In a lot of your post you do seem to be talking about that (“When did continental drift become a scientific truth?”). Data I’ve gathered remain only pencil marks on a sheet until they’re published, at which point they become a (small) addition to human knowledge. If you can’t (or won’t) communicate, you don’t add to human knowledge and thus haven’t advanced “science” in this sense.
I agree entirely with your closing points – communication is just as important to a successful scientific career as statistics, experimental design, etc etc. Which is why I’m about to go back to (ugh) proofreading The Scientist’s Guide to Writing….
I think you’ve put your finger on the heart of the matter. Is “thinking scientifically” enough to “do science”. I guess I’m arguing for a large fraction of definitions of “doing science” (including being funded by society through grants or jobs at public universities with a research component), the answer is no.
Great article and points that I think greatly emphasize the importance of collaborating in the STEM fields. I often look at this as a “dry run” when publishing. If I can’t get my point across effectively to my co-authors, then I will likely fail in communicating the point during the peer-review process.
I’ve been arguing a similar point for a few years now!
I like all of your points, but I’d also add another: people are bad at independently assessing their own work. We’re also very good at fooling ourselves into thinking we’ve covered all of our bases. Without subjecting your work to the slings and arrows of other people’s criticisms, you can’t really be sure you’ve done the analysis right. The input and critiques of other people (colleagues, peer reviewers, subsequent publications from other groups, etc etc) are vital in ensuring that you’ve actually done the analysis correctly.
Also, as an historical note, a big reason why continental drift was originally maligned is actually a good example of the above. Wegener was a climatologist, and he proposed that the continents plowed their way through the ocean crust. Oceanic crust is denser than continental crust, so geologists thought that was a ridiculous idea since continents *couldn’t* plow through the denser oceanic crust. So he had no proposed mechanism (what made continents move) and his hypothesis didn’t really make sense (continents plowing through ocean crust). Hence when Hess later found evidence that the sea floor was spreading and pushing the continents around (rather than being plowed through), it made a lot more sense.
Excellent point on a big reason why the social process is not just a way to accomplish science but is central to why science is so successful – lack of self-objectivity.
And I didn’t know the exact details on Wegener – thanks for that.
That’s interesting about Wegener’s hypothesis, I didn’t know that.
I think it would be interesting to compile cases in the history of science in which the hypothesis we now accept as correct was rejected when it was first proposed for what we would now consider to be good reasons (as opposed to, say, being rejected for contradicting the Bible). And crucially, to compare these to the reception of hypotheses *we now reject*. Including, ideally, both hypotheses we now reject *that were rejected when first proposed* as well as hypotheses we now reject that were initially accepted before being later overturned. Without the benefit of hindsight, could you tell the difference between these different sorts of cases?
I think this would be interesting because there’s a tendency to read the history of science through the lens of our current understanding. So anyone who came up with an idea we now accept must’ve been a prescient genius, and anyone who opposed an idea we now accept was some combination of stupid, ignorant, blind, and biased (or at best, sadly trapped behind the blinders imposed by the then-existing paradigm).
The controversy over Darwin’s Origin is a good example here. Despite teaching a graduate seminar on the Origin, I’m still not satisfied with my own knowledge of the alternative views and the evidence for them (“special creation”, for instance), as they were seen in Darwin’s day. It’s my impression that the opposition to Darwin was for a mix of reasons. Some of which wouldn’t be considered good reasons today (contrary to the Bible). Some of which would be considered good reasons today (the mechanism isn’t well-developed or contradicts other things we think we know). And some of which are kind of borderline (e.g., objections based on Darwin purportedly having an unsound philosophy of science). Students who read the Origin today tend to be hugely impressed by all the prescient-sounding bits, to the point where they often read back into Darwin modern views he didn’t actually hold. They tend to skim over or excuse the bits that turned out to be wrong (which includes some *very* important bits!). And they tend to assume that Darwin’s opponents were the Victorian equivalent of today’s creationists, or maybe just not as knowledgeable and brilliant as Darwin. So I need to do a better job of conveying to today’s students the full force of objections to the Origin–objections that were very reasonable by today’s standards, given what was known at the time.
Fun note about prescience, Darwin and continental drift:
Darwin wrote in his journal about the 1835 Conception Earthquake the following passage:
“A bad earthquake at once destroys our oldest associations: the earth, the very emblem of solidity, has moved beneath our feet like a thin crust over a fluid…”
My old structural geology teacher used to have that plastered on her door, because it’s so perfect.
It’s amusing to read the Origin as a modern reader and instantly spot all the places where Darwin has to appeal to rises and falls in the level of the land, and to organisms floating around on glaciers, to explain data we’d now explain via continental drift.
Agreed that communicating results is part and parcel to science. However, this extreme position (“you could be a wonderful scientist but a terrible communicator of your science and that you would suffer for this career-wise and that would be unfair”) is sort of a straw man.
I’d argue that many would-be scientists that are very-good-to-excellent at other aspects of science are just mediocre-to-good communicators. In this world where <10% of grant proposals are funded and publishing in high-level journals is becoming ever more competitive, the difference between being an 'okay' writer and an 'excellent' writer could make-or-break an early career trajectory. I benefit from this, but don't think it's necessarily a good thing for the scientific enterprise.
There is no doubt that being a better communicator is better for your career. That gets into all kinds of conversations about incentives and reward systems, evaluation systems, etc. My short short version on this is much like Churchill’s quote on democracy – its a terrible system but also the best available system. I suspect we could have a long and fruitful dialogue on this and perhaps still ending up agreeing to disagree on this.
But at least in my mind I can separate “better for your career in science” and “doing good science”. What I want to argue is a deeper point than good communication is good for your career, but in fact you can’t even do good science without communicating.
Bill Wimsatt is one philosopher of science who argues that science doesn’t have a single “method”, but rather has a whole bunch of methods. All of them heuristic, rough-and-ready sorts of methods, that work well in some characteristic circumstances but fail in other characteristic circumstances. And all of which are optimized for helping imperfect, cognitively-limited, social beings learn about the world.
This is in many ways similar to Feyerabend’s views, but he maybe took these consequences a bit further. He often emphasize that “Science” isn’t a single enterprise, but separate activities that should be evaluated as such. A nice quote:
“At this point some defenders of uniformity [Uniformity of Science, my addition] rise to a higher level. Science may be complex, they say, but it is still ‘rational’. Now the word ‘rational’ can either be used as a collecting bag for a variety of procedures – this would be its nominalist interpretation – or it describes a general feature found in every single scientific action. I accept the first definition, but reject the second. In the second case rationality is either defined in a narrow way that excludes, say, the arts; then it also excludes large sections of the sciences. Or it is defined in a way that lets all of science survive; then it also applies to love-making, comedy and dogfights. There is no way of delimiting ‘science’ by something stronger and more coherent than a list.” (Against method, p 246)
Yeah, Wimsatt’s definitely no Feyerabend!
Feyerabend goes too far for me. Indeed as fieldofbullets highlights so nicely, it is in large part because science is where social processes encouraging criticism meets empirical testing that science advances (admittedly nonlinearly) towards some underlying truth – not merely creating social constructs as Feyerabend would have it.
I’ve never really understood Feyerabend as denying that there are underlying “truths” out there for science to describe, but maybe I’m misreading him. Too me, he is mostly writing about the relativity of our knowledge (so Epistemological relativism). Overall, he’s saying little about how well scientific theories approximate reality (or ontology in general), but he doesn’t appear to he saying that there aren’t objective underlying truths. When he compares theories (e.g. Newtonian physics vs. relativity) it is often clear that he considers one theory as a better representation of reality/truth. So I would consider him a realist of sorts. To me, he is in large parts basically making the same points as Lakatos (e.g. that our knowledge cannot be disconnected from our theoretical framework, and is in that sense relative). However, he is making the point much more forcefully (and in a completely different style), and he is intentionally provocative. Granted, Feyerabend is not easy to understand, and he is often writing tangentially. I’ve also mostly read texts from around “Against method”, and his views might have changed over time. Overall, I prefer Lakatos because his criticism is more constructive, but I like Feyerabends ideas. I think it’s fair to describe the boundary between what we describe as ‘Science’ and ‘non-Science’ as fuzzy and not sharp, and his description of ‘Science’ as a collection of different activities that lack a unified method is also reasonable to me.
“This is the real reason I push every student to publish their dissertation even if they’ve already landed a job in .e.g. a government agency where they don’t need publications. Its not science until its published.”
– Very true. It can be uploaded in some of the repositories for PhD theses.
Well, except that those repositories are little read. Even less than quite low-impact journals, I’d guess.
First off, thanks Brian for taking another crack at this. I think it is important.
So here’s my take. To be a successful scientist, which I operationalize as having your work cited and the ideas therein endure in scientific media, is a product of three “meta-abilities”:
(1) the ability to ask and answer insightful and relevant questions
(2) the ability to convince any rationale member of the the community you have succeeded
(3) to ability to maximize the number of those people you encounter
Much of the g’mish associated with the last post, besides representing a classic problem of clashing definitions, reflects our anxieties as we confront the increasing importance of (3), that is, as we all struggle to find ways to get enough eyes on our work when the literature has exploded in size. “Salesmanship” as it is often used, seems entirely the domain of 3, and partially the domain of 2. It represents problems we all are struggling with:
–It is diverting what seems to be an unjust amount of effort from (1) to the formatting and reformatting of manuscripts for high profile, short-format journals, in order to maximize (3).
–It is eschewing Ecosphere because it has a lower citation rate than Journal X, when one know the only way to get Ecosphere’s citation rate up is to send your own good work to Ecosphere.
–It is not sending colleagues copies of your own work, or tweeting your own work, because that is “crass self-opportunism and salesmanship”, when you may actually be doing those colleagues a favor because they are dealing with the same firehose of information as everybody else.
We are increasingly in a world where the
p(becoming a successful scientist) ~ (1) * (2) * (3)
and where each of those components each requires its own suite of talents. We need venues, like this one, that allow folks to identify discuss, and hone these talents. We need to allot significant fractions of our national and international meetings toward the same end.
Very clearly put!
And I think you’ve independently reinvented William Shockley’s hurdle model for succeeding as a scientist: https://dynamicecology.wordpress.com/2014/01/23/william-shockley-on-what-makes-a-person-write-a-lot-of-papers-and-the-superstar-researcher-system/
Really interesting piece Brian. Here’s something else to add to the mix: where does the relatively recent phenomenon of “Professor of Public Understanding of Science” (and similar titles) fit into this? We have several such PPUSs in the UK, the most famous of whom was Richard Dawkins, until he retired. The ones I’m aware of are all scientists who have gone down the route of writing and broadcasting about science rather than doing the science themselves. What does that make them? Are they still scientists?
To me the topic of public understanding is a really important one, but separate from the one I’m asking.
My quick take on public understanding is that we need a spectrum from “pure scientists” to “pure policy & communication” to people in between. And we should value all of the continuum.
One thing that strikes me while reading this post, its comments as well as the earlier post of Jeremy and the comments to that one, is the limited extent to which the role of the audience is discussed. The audience figures a bit in the discussions about Wegener and Marshall and Warren, but up to now I miss an explicit discussion and reflective view on ourselves as audience of other scientists.
If science is a social process (which I strongly believe) communication is indeed a central part of doing science, as Brian so clearly argues. But the discussion up to now revolves around the capabilities and qualities of the scientist that brings the information, his/her “salesmanship”. However, for salesmanship to be effective, the audience needs to be receptive to this salesmanship. If “overselling” of scientific findings occurs, it tells something about the selling scientist, but also about the audience that buys his/her arguments. I believe we are all human and therefore we like hearing convenient over inconvenient findings. Salesmanship hence goes hand in hand with receptiveness. If some scientist stands up and will tell us that we have been following the wrong scientific track for quite a long time, it is not an easy message to receive, how convincing the evidence and how skilled the scientist is in communicating the message.
Rather than bemoaning the increasing role of “salesmanship” in science, I am more inclined to turn around and reflect upon my own role as a member of the audience in such communication. The increasing role of bandwagons and fashion in science I therefore find more worrisome than the increasing role of salesmanship. Brian points out that science does have self-correcting methods that in the long run will change established consensus that is simply wrong, but the long run can be sometimes very long (I can not help thinking about Gallilei’s heliocentric world view in this respect).
And the most intriguing question for me personally is which ideas or consensus’ do we accept at present despite clear, convincing and well-communicated evidence to the contrary, simply because the message is too inconvenient to listen to.
Several very interesting points. Communication is definitely a two way process. I quite agree. The audience is an active part of it.
Your last question is a fun one. My money is on “we’re in the middle of the 6th major mass extinction” but only time will tell.
I think that you’ve made some excellent points about the role of the “audience” in science communication, much of which I agree with. However I wonder about your comment regarding “The increasing role of bandwagons and fashion in science”. Is there any evidence that “bandwagons and fashion” have an increasing role, or have increased in their frequency, compared to 20 or 50 or 100 years ago?
It’s a genuine question, and not an easy one to answer (to begin with, how does one quantify “bandwagons and fashion”?). But from what I know of the history of science it seems to me that there’s always been certain research themes that have been fashionable, and jumped on as a bandwagon, and then become less fashionable after a while. Perhaps it’s only more apparent now because we “do” more science and communicate it more widely?
I had the same thought. Definitely tricky to get at. Best you could do (and this is highly imperfect) might be to look at citation patterns of papers. Are certain papers (the bandwagon-starters) that quickly accumulate lots of citations, then suddenly stop accumulating them, becoming more common?
Looking at rises and falls of keywords might be another approach. But are bandwagons in ecology always associated with new terminology?
Perhaps I was not sufficiently careful using the words “bandwagons and fashion”, I should have described better what I meant. My idea is that there is substantial emphasis on “novelty” nowadays. Take for example, the European Research Council, the most prominent funding body of science in Europe at present, handing out substantial and prestigious research grants. They clearly express that their aim is to finance “novel and breakthrough science”. It is my impression that this emphasis on “novelty” works out in such a way that long-standing, unsolved, but very fundamental questions are deemed not sufficiently interesting to investigate and finance, as they are already studied for such a long time.
Pingback: روایتگری در علم – سیتپـــــور