Friday, April 27, 2012

Metaphysics in science, Part I: "Call us when you actually find something."

What constitutes a 'finding' in science? 
It is supposed to be a discovery of something about Nature, rather than, say, just an 'idea' about Nature.  It's supposed to be real and not a matter of metaphysics.  Metaphysics had a long history in philosophy, when philosophy was the lead-in to what we call science today.  In today's sneering world of science, science is fact, and metaphysics is made-up Blarney rather than stuff that's real.

But what is a 'finding' today?  Administrative interests are constantly on the prowl for results that their company, institutes, portfolio, or clients can use to help lobby for more funding, and the news media aren't far behind.  But everybody's busy, so what counts in science in this sense is something with a melodramatic picture that you can say faster than the word 'science', and that will grab the interest of someone whose attention span doesn't go beyond a Tweet.  Some populations are known to anthropologists by names like 'the basket weavers' -- we'll be known as 'the boasters'.

This attitude is everywhere and seen all the time, and it's detrimental to good science.  Cakes take a certain time to bake and not all food is fast food.  Science has to bake to come out as good as it should be, and can be.  Quick answers yelled from rooftops (of  Nature's offices) and rushed out of the oven for quick display purposes are notorious for false starts and hyped findings that are not confirmed later, for various reasons noted a few years ago by John Ioannidis. 

By Hooke or by crook science in the 21st century
When Robert Hooke first turned the microscope's eye onto nature, we got the first glimpse of things that had previously been impossible to see.   He documented many things in his 1665 Micrographia, like details of insect bodies, and the uneven surface of polished pins. Hooke turned micrographia into visible-scale drawings, and made flea hairs visible....and fleas were important!  We were naive then, and learned a tremendous amount from the new lens on the world.  Blowing things up was legitimate.



Today we live in a similar era, when every tiny finding, visible only through a massively-humongously-parallel-generation sequencer, is blown up--that is blown out of proportion, in our puffery laden, lobbying,  PR-driven world.  Unlike Micrographia, however, not all of today's fleas are actual 'discoveries' in the same sense.  Yet, the PR machine wants to report 'findings', and anything that can be claimed to be one (with a nice figure) is going to be trumpeted.

If it's more than 140 characters, it's not real!
We're pressured only to consider simplistic sound-byte-sized results as true 'findings',  or be embarrassed if we haven't got a slew of papers reporting our sound-byte sized things in 'high impact factor' journals, or hyped by the New York Times or the BBC. Apparently 'just' understanding Nature, most of whose traits are subtle and not melodramatic, isn't real science.  Hype is what the public is sold, it's what's sold on television and on front pages, and it's basically all that congressional staffers and their like are told about.

Now, this might be OK if the recipients of the hype--such as policy makers who have to cough up the funds--weren't so inundated by Everything They See is Phenomenal that they know full well how to ignore most of it.  Or, worse, if they actually don't know that all they see are snow jobs, we're in deep trouble.  So our version of micrographia has become the blowing up of mainly trivial things we hadn't seen before, and making them sound as if they were previously hidden giants.

Two negatives do not make a positive, but one does!
Consistent with all of this is the notorious under-reporting of negative results.  It's worse than unethical in the drug trial realm, because it leads to obvious bias, unjustified profiteering, and actual harm (sometimes lethal) to patients.  Sometimes it's intentional, but even when it's just that investigators think negative results are not worth reporting, or the 'premier' journals don't think it's worth bothering about (i.e., won't sell copy) and won't publish them, it's harmful because it's systematially biasing and misleading to science.

Take GWAS.  People do publish the results, but many of those who aren't boasting of their purported revolutionizing success (because of the 'positive' findings), are bemoaning the failure of GWAS.  "Well, see," they say, "you never find anything!"  GWAS are a failure!

Nobody thinks more than we do that GWAS and its 'next generation' successors are overselling in a bad way and often for bad reasons.  Nonetheless, and we've said this before, GWAS have been a fine success!  The 'negative' findings (no real blockbuster genes, but instead many tiny genetic contributions to risk), are not a negative but a positive: they positively tell us how complex nature is.  They are findings!

Findings about the real nature of Nature may be dramatic, as Hooke found in his day.  They do occasionally turn up in our own day and that makes science fun and interesting.  But most of Nature is complex and not amenable to quick-fix answers.  For a real appreciation of the object of science to plumb the truths of Nature, complexity itself, difficult to work out or explain (say, in evolutionary terms), should be thrilling enough: the grain without being blinded in a blizzard of chaff.

A subtle, nuanced, careful approach to science is often being overlooked these days because of the obsession with the current idea of what constitutes a 'finding'. Whether this state of affairs is just part of the game in a complex middle-class culture, or has tragic implications for the kind of work that could be done but isn't is anybody's guess.  Fixing the system would take major reform, and may not be in the cards given the nature of our society.

No comments: