Thursday, September 19, 2013

Cute, cuddly....and off the mark

Well, the gag-writers on the editorial staff of the entertainment publication called Nature have done it again!  This cover, like all of their covers these days, is arrayed with cute tag-lines for the stories within.
Awww, how cute!
The cuddly picture is of some furry little mice helping each other out--cooperating.  Heavens! So anti-Darwinian of them!  Richard Dawkins must be crying in his beer!  This story is based on the very latest hot major discovery in evolutionary science: cooperation exists!  And now we know why--it's oxytocin!  (Single gene evolution; what else would you expect for a Nature story?)

This finding (and we have no way or reason to question its validity) may be viewed as a refutation of the harsher Darwinian world of relentless competition, asserted by generations of deep evolutionary thinkers, many of whom were wont to popularize pared-down ideas of what is really much more complex nature.  Until recently, any cooperation has been viewed as just a touchy-feely way to describe what is really competition in disguise.  But now it's cooperation we have to explain.

Whether or not, or to what extent, cuddly or combative characterize the true state of Nature, we seem to be going like a pendulum from one view to another.  Each leads to its set of cover stories, each touted as a major discovery.  Each the Truth du Jour.

Of course, Nature is a business and a very successful one at that.  But if they were more about science rather than sales, perhaps the pendular nature of science would itself be the story, rather than the position of the pendulum at any given time.

As we repeatedly comment on, a similar oscillatory Truth-O-Meter is routine in biomedical and other aspects of genetic research, where every story is the Big Story and is paradigm-shifting.  This partly reflects the careerist and other rather venal interests of our time, but it also reveals the realities of the kinds of data we have available to us today, that is, the state of the science itself.

What the real story is, and what the best science should be about
The actual story that should be getting our attention is what one might call the 'meta-story', the story about the stories being published every day.  That story is the fact that we have new Truths du Jour every day.  If science were what we scientists tend to blare so confidently that it is, then we should not be reversing our views every few days.  The pendulum should come to a near-stop.

The fact that we can shift from the view that all life is driven by Hamilton's rule (be nice, but only selfishly), to the latest Wilson (EO and David Sloan) view (that cooperation is what life's about), or that coffee prevents, or coffee causes disease, and so much else like it--that changeability itself is the story.  When our knowledge is so fickle, supposedly each time supported by funded (and presumably well-designed) and ever-larger technically sophisticated studies, then there is something rotten in the state of Science.

We obviously do not have an adequate understanding--an adequate theory--of the nature of life.  The main job of science is, at least as we teach to our students, to understand the truth of nature--the laws of nature, not just to teach the current fads of scientists.  If truth is our aim, we will never get all the way as there's always more to learn that we don't yet know.  But we should at least have a firm grasp on seemingly simple questions such as whether our mission in life is to kill or cuddle.  Or whether coffee, eggs, or a daily drink are good or bad for your health.  Or which gene raises and which reduce your risk of being a great athlete or getting colon cancer.

Instead of boasting of each new ephemeral finding and assigning cosmic import to it, we should be worried that we're throwing money and intellectual energy away on routine, if technically sophisticated, incremental but far off the mark research.  The push to find what amount to superficial answers, in the huge operation that is modern science, should be changed to a strong push to bypass the superficial and instead to find and understand the deeper truths.  Many of us, perhaps especially those of us who are more senior, tend to blame the tenure and research-funding pressures, as being materialistically short-sighted.  We may just be reactionary cranks, but whatever the reason, a greater stress on the basic epistemology, our most general and strongest ways of understanding living nature, should be a primary objective, and what we stress to our students.

Perhaps one might think some exception would be in order, for applied fields like medicine.  But the see-saw patterns in medicine show that even there, where the idea that research is not about abstract theoretical generalities but how to stop a specific disease, requires a much deeper basic understanding, and less reliance on sampling and statistical analysis, than we have now.


Jim Wood said...

Ken, this post raises a number of important issues, but right now I'll just comment on two of them. First, given the almost complete reliance of biomedical research on a faulty NHST framework, an inconsistent, ever-oscillating Truth-o-Meter is exactly what you'd expect. Second, one of the most damaging influences of Thomas Kuhn has been the widespread belief that the only science worth doing, career-wise, is "paradigm-shifting" science(Gawd, I loath that term!). Everything else is (sniff) "normal" science and worthy only of drones. That belief has largely killed replication, which is the capstone of sound scientific my humble opinion.

Ken Weiss said...

The issue related to your first point is that the issue is not a secret, yet people absolutely refuse to pay any attention to it. Among other things, scientists and their media parasites thrive on claiming definitiveness for their current results (but, admitting they aren't perfect, demand 'further research needs to be done' to refine the finding.

Second, I couldn't agree more about the damage that Kuhn has done with his phrase 'paradigm shift', so uncritically accepted so widely because it allows everyone to fancy that s/he is involved in creating one.

But I'm not sure I agree with your statement about it's implication. At least, I think what has happened overall has largely done the opposite, by leading to endless replication and repetition because it's easy to argue that one needs to improve the existing findings with yet another study. This leads to the massive uncritical, ultra-safe, long-term, replication studies that eat up so much research funding.

Going along with that, any study that is not a replication attempt, that is, is 'innovative' in that sense, has scant chance of receiving research funding. It's too risky and those doing conventional (replication type) studies guard the funds tightly.

There are other issues about what 'replication' is possible in areas like genetics, so you may be thinking of other areas where different sorts of replication serve just as you say.

I do, however, agree that there is at the same time a kind of denigration of 'normal science', but Kuhn actually stressed its vital importance because normal science--clinging to current accepted views and practice--is what gradually forces the innovative rethinking, because normal science can't go beyond the paradigm.

In statistical areas that you and we usually deal in, appropriate kinds of replication and sampling and statistical analysis are fundamental. I assume that's basically what your last sentence was saying.

Ken Weiss said...

I would add this sort of after-thought. Kuhn was pointing out aspects of the sociology of science, as a corrective to the prior view that science was all an objective march to truth. Kuhn's points had been made (and Kuhn knew about it but didn't give credit) by Ludwik Fleck in the 1930's in relation to the understanding of syphillis and of human anatomy.

I haven't looked at his book recently, but I don't recall his delving much into the epistemology by which the currently operative paradigm is investigated. That would get into issues of now 'normal' science was being done, and I think would involve another layer of 'paradigm' (e.g., sampling and statistical analysis and assumptions), but also might help explain why a paradigm can hold on tightly--if the methods that are part of operating practices aren't powerful or definitive enough, or if the methods themselves aren't being challenged enough.

Anne Buchanan said...

Just a note that Jim has expounded at length here on MT on the significance test issue. Well worth reading.

Alan Packer said...

It’s fair enough to want to criticize hype, glib or misleading headlines, and uninformative illustrations, but in this case you do so at the expense of a very interesting and important piece of science.

To begin with, the paper from the Malenka lab is not really about ‘cooperation vs. competition’. It’s about the identification of a neural circuit that contributes to the ‘reward’ associated with social attachment or bonding, a phenomenon that has extensive roots in the scientific literature quite independent of any of the controversies surrounding the arguments of Richard Dawkins. It also has nothing to do with ‘single gene evolution’—the paper in fact outlines the interactive effect of oxytocin and serotonin in different regions of the mouse brain as a linchpin of the neural circuitry underlying social reward. And in any case the paper does not claim that this interaction provides a complete explanation. It does, however, mark an important advance in our understanding of the brain circuits that underlie social interactions.

Your piece criticizes the attention paid to “ephemeral findings” that are “incremental” and provide “superficial answers”. More power to you. But the paper by Dolen et al. isn’t any of those things, furry little mice notwithstanding, and I hope your readers who are interested in cutting edge neuroscience will give it a careful read. [full disclosure: I work at the foundation that supported some of the work in this paper]

Ken Weiss said...

I'm sorry if you felt we were ourselves off the mark in the way we treated the paper. I could offer various views on that, but we used it as a foil for making the point about the uncritical acceptance of pendulum like findings splashed on the cover of major journals...and I don't think backing off of that point is called for.

In the extreme, a relatively simple sociality mechanism would be quickly interpreted in competitive terms by the likes of Dawkins, Dennett (and the younger EO Wilson). And Nature chose this as the cover paper, and chose to present it as we described, I think.

Anyway, we didn't mean to suggest, and don't think we did suggest, that there was anything wrong with the paper, and regret it if we did.

Attention should be focused (agreeing or not) with our intended broader point about how science is working these days.