How do we discover? That’s a question that keeps coming up in my mind. There are books around that work at this point–to point to some recent examples, Ashton’s How to Fly a Horse, Kauffman’s Investigations, Johnson’s Where Good Ideas Come From. I thought it might be helpful to make a list of ways discovery happens, in no particular order–
- random observation
- systematic study
- solve a problem
- create a problem and solve it
- adapt something known in another context
- recognition by a prepared mind
- attempt to build/prototype
- careful observation and documentation
- allow to mature or develop without intervention
- become expert at something to realize what otherwise isn’t evident
- messing around
- thinking about stuff
Of these, perhaps 4, 5, and 6 might be called “research.” But exploration can involve simply traveling to places that haven’t been visited before. We discover by going to Mars (or at least sending a probe)–hey, those aren’t canals built by Martians after all! Travel is not research, though one might say one is doing “research” by looking around once one has gotten somewhere new. Similarly, one can mess around with something and call it “experimenting,” but the scientific experiment, designed to distinguish between competing hypotheses, is a narrow version of experiment. It’s just that experiment sounds more dignified and authoritative than just messing around. Systematic study, of course, is the essence of research. Arranging things in patterns helped us figure out genetics and where to look for undiscovered elements.
No doubt there are more things to add to the list. What strikes me, though, is that there are many ways to discover. The idea that research is the only/primary/best way to discover doesn’t seem all that persuasive. There are arguments around that much of the discovery in the industrial revolution came from experts in an area messing around with ideas about how to do things better or adapting what they knew from one area to another. Matt Ridley, for instance, makes an argument of this sort in The Rational Optimist. A century ago, the rising company laboratories debated whether it was better to give problems to engineers to solve, and if they couldn’t do it, hand the problem off to scientists in the backroom, or whether scientists should explore stuff and call in engineers when they had found something interesting and let the engineers work out how to use it.
During World War 2, Vannaver Bush developed a skunk works approach to innovation. He recognized that the military establishment pursued research and development drawing on the knowledge that it found useful for its purposes. Once it established its purposes, then it was efficient at prioritizing what it would consider. Bush observed that there was plenty of science that the military didn’t consider but that might create technology useful to the military–radar, say, or a monte carlo technique for predicting where bombers flying in semi-random pathways over a target might actually be, or, say, an atomic bomb. So he combined scientists, industry engineers, and “gadgeteers”–folks who could make things–and developed new technologies that the military planners could not imagine let alone justify allocating effort toward.
Bush and President Roosevelt got the idea that perhaps this same approach might be applied to civilian areas, such as medicine or communications, and for that purpose Bush advocated for the federal government to sponsor scientific research at universities outside the mission priorities of any government agency. The funding of research part worked out, but not so much the independence from government-specified priorities. The emphasis moved toward solving problems that could be defined in advance and reviewed for merit rather than to exploration or messing around or following hunches.
When university officials call for more funding for “basic research,” they repeat just a part of Bush’s argument that the government has a legitimate interest in supporting research. But their call for research doesn’t include a reason why that research should be done by faculty hosted by universities over any other venues in which to conduct research–companies, nonprofits dedicated to research, professional organizations, or labor unions. Any number of venues might engage people to explore nature or mess with technology or build wacky prototypes. According to accounts, Nicholas Tesla was a mental mess–but being part-mad apparently had a wonderful benefit in allowing him to get to ideas and technology that others simply could not grasp. Sure, it is said he didn’t accept the idea of a structured atom–but does that science matter? Or, perhaps even, accepting that science might distract from, complicate, or even eliminate lines of thinking that would otherwise be productive.
The same thing goes for technology sophistication. One gets rooted in a line of development and the cloud of adjacent possibles–the things that are relatively easy to do next as extensions of what we already have done–shift accordingly. As Neil Stephenson has wryly put it, we get our best minds working to make better spam filters. We get locked into an opportunity set and the problems we solve and the stuff we discover is directed to getting to those opportunities. The more focused and efficient we are about it, the fewer the chances to get off topic and discover something that might break away and form its own line of exploration and discovery.
If we consider the conceptual and social tools we use to promote innovation via research, we might be shocked about how narrow they are. The institution as host and owner of discoveries. The sponsored research contract. The research proposal awarded on merit assigned by a committee of reviewers. The publishable study. The study with obvious practical importance. Money coming from disinterested but politically interested if not motivated sources, such as the federal government. The idea of competing for funding based on proposals or justifying funding based on the money spent on research rather than on benefits arising from the outcomes of research. The patent as a monopoly on discovery. The idea that commercialization is necessary to use a discovery or for public benefit.
It seems that at one point, science was also a matter of messing around, failing, causing accidents, and the like. Much less given to systematic study. That’s the crazy science that Swift works over in Book 3 of Gulliver’s Travels. But perhaps “basic” research is way more than the systematic study of nature without an immediate practical or commercial purpose. Maybe basic research can’t really involve much in the way of systematic study, other than perhaps carefully observing and documenting stuff that everyone else thinks is wacko. We might propose, then, that if research can be described in a written proposal suitable for review by a committee of semi-leading experts, then that research isn’t basic enough. By the time something can be seriously proposed, it is already so steeped in a prevailing set of expectations that no matter its apparent audaciousness, it is unlikely to discover anything much beyond confirming one’s expectations.
And this is where John Ioannidis and others have taken on the problems in published science–almost all of such published work is the result of this now standard approach to discovery as a form of proposed, reviewed, contracted, and institutionalized activity. Ioannidis and others have found that most scientific research published even in elite journals turns out to be wrong–not superseded by new work, but found to be faulty–in experimental set up, in data handling, in data selection, in accounting for error, in integrity in reporting, in analysis. Studies that report a biological or psychological effect often cannot be replicated, or the effect apparently diminishes when the experiments are repeated. Even where clinical studies turn out to be correct, Ionannidis and others argue that many of these correct studies (replicable, not obviously faulty) aren’t helpful in the clinic. They report accurate results that can’t be meaningfully used. So we have a heap of publications, most dutifully reviewed, edited, laid out, published, archived, and indexed–but not much that’s reliable or useful.
There are proposals around that argue that institutional research must change. That institutional scientists must be better managed–whipped into a passionate frenzy by stories of suffering children or organized by ex-generals into a force to attack disease or led by activists who can rally the public to change things. Others suggest that scientists should overhype or even doctor their findings to motivate governments to provide more funding and to ensure that the public continues to see their work as a priority over other things–such as messing around or repairing roadways or hiring more fire fighters. You know, the cause of science is so noble that it’s okay to deceive ignorant politicians and the general population to ensure that science remains well funded–that is, especially, institutional science wrapped up in proposals reviewed for merit by panels of scientists also competing for the same sort of funding.
It’s not that there isn’t room for competed research proposals from institutionally controlled investigators. That’s fine, as far as it goes. My concern is that this activity, made big, draws activity away from all the other ways we might discover. People who would mess around end up sucked into working on institutionalized projects. People who do mess around are ignored because they lack institutional credentials. Vannevar Bush argued that government money for research (“the free play of free intellects”) should run through an organization reporting to the President and accountable to Congress, but otherwise independent of government, to be led by an “outstanding” person and advised by a committee of citizens “of broad interest in and understanding of the peculiarities of scientific research and education.”
And in this we have the problem of government funding. Who decides what to fund? How? Even funding everything augments activity everywhere. Funding both solar and clean coal gives neither an advantage. Funding solar over hydro or nuclear gives solar an advantage in funding, but it doesn’t mean solar research will make discoveries that dramatically improve the output of solar cells or make economical the production and service life of a solar panel. There are many ways to discover. Federal funding draws thousands of people into contracted research. In doing so, it determines what gets studied, what gets published, and whether anyone should care, other than to advance their careers and obtain further funding. Those other ways don’t get the same support. People migrate to where the funding is. What we might discover changes accordingly, even if there’s no obvious trace left by what doesn’t happen that could and still might.
In Contact, Carl Sagan pits government officials against a corporate technologist tycoon. Contact is fiction of course, not history–but it is the tycoon that is willing to fund work that to the government scientist would destroy one’s career. In Sagan’s account, of course, the discovery comes from the wacky observing by Ellie Arroway, supported by her tycoon patron, who ends up having the prepared mind to be looking for a signal and recognizing it. We might think of the story as a moral fable regarding discovery–that what we desire might not be best found in institutions, or in competing proposals, or in government priorities, no matter how many well credentialed committees meet to approve these priorities.
Just saying. We have our habits, our institutions, our efficient methods of conducting what looks, smells, and behaves like “research.” It’s just that there are many more available ways to discover, ways to fan out and consider different framing, different exploration, different messing around than we have created (or self-organized) since the federal government stepped so strongly into the support of “research.”