X. Hull's Evolutionary Account of Science Noretta Koertge
Popper pointed out a rough analogy between his conjectures-and-refutation theory of science and Darwinian theories of the evolution of species through variation and natural selection. Hull, however, articulates this analogy in great detail, drawing in particular on concepts such as inclusive fitness which play an important role in sociobiology.
a. Interactors and Replicators
There are many processes in nature which could be described as cases of evolution through selection. The rock formations in Monument Valley, for example, are the result of the long term differential effects of erosion by water and wind on various soil constituents.
Biological evolution has additional structural features, many arising from the fact that typically it is the phenotype which interacts with the environment (or "struggles" for existence) while it is the genotype which is differentially replicated in later generations.
On Hull's account, in science it is ideas, concepts, theories which are differentially perpetuated through time, thus producing intellectual lineages. But it is scientists who bear these ideas and whose success or failure causes ideas to proliferate or go extinct. Unlike Popper, who describes the selection process as occurring in World-3, the domain of propositions, Hull's account locates the struggle squarely within the social realm. However, as we will see below, Hull does not deny the importance of objective properties of propositions, such as their explanatory and predictive power. Rather, his purpose is to describe the institutions which permit these properties to become manifest and to influence the selection process.
Having set up his evolutionary model in terms of ideas as replicators and scientists as interactors, Hull now asks, what is the selection mechanism which causes the differential replication of scientific ideas? Although he does not explicitly say so, it appears that Hull is really wrestling with a variant of the demarcation problem. There are many forums in which ideas compete and the popularity of political, religious, artistic conceptions waxes and wanes. What is so distinctive about the process by which scientific ideas are selected?
Hull analyzes the selection mechanism operating in science into three major components - curiosity, checking and credit. Hull spends most of his time discussing the reward system in science and the various ways scientists strive to get credit for their ideas, but his account also posits a crucial role for empirical testing. As I understand him, Hull thinks the sorts of aims and methods described by traditional philosophers of science are indeed necessary for scientific progress. What he denies is that they are sufficient. Scientific institutions are more than an external scaffolding which provide a matrix and support for the growth of knowledge; they are an integral part of the epistemological process.
b. Why Curiosity and Checking Aren't Enough
To motivate Hull's emphasis on the reward system of science, let us do a thought experiment in which we start out with isolated individuals who are well-equipped with curiosity and the propensity to subject claims to empirical scrutiny, and then assemble them into an efficient, scientific community. What motivational ingredients would we need to add? What new social norms would need to emerge?
Any philosopher who emphasizes inter-subjectivity as a feature of the empirical base of science has already introduced a social ingredient into science. On this account, Crusoe needs Friday if he is to begin to do science - not as an errand boy, but as an independent, sceptical collaborator of observation claims.
Popper adds the point that although in principle we should maintain a critical distance from our theories and sincerely try to refute them, in practice it is often easier to come up with cogent objections to other people's theories. So science will be well-served by institutions which facilitate this sort of friendly, hostile cooperation. I would add a corollary: psychological studies indicate that kibitzers, people watching others try to solve puzzles, often come up with solutions faster than the agent who "owns" the problem. So maybe science would also benefit from institutionalized kibitzing (such as research-in-progress reports).
Popper also makes the trivial-sounding observation that if we all started where Adam did, there's no reason to suppose that we'd get further than Adam did. (Or as my colleague, Ed Grant, pithily puts it, "Aristotle was no dummy.") So we want our bright, critical individuals to have easy access to the results of inquiry of both their predecessors and their contemporaries.
In our earlier discussion of Popper, we listed various types of problems which trigger inquiry - violated expectations, unexplained regularities, ununified bodies of knowledge, etc. What we took for granted in that discussion, but did not articulate, was the assumption that the flaws in background knowledge which intrigued the scientist were problems which no one at the time knew the answer to. Children can satisfy their curiosity and hone their problem-solving skills by
re-discovering Archimedes' Principle or playing with Rubic's cubes. But what we want mature scientists to do is to direct their curiosity, not just to problems which they personally don't know the answers to, but to problems which no one has solved. One tension in science education is how to teach students the skills necessary for and the satisfactions of solving problems by themselves (in which case the novelty of their solutions is unimportant) while also encouraging them to look up answers to questions in authoritative reference works and to devote their energies to working on new projects.
Satisfying one's personal curiosity does not insure communal progress. For that to occur, we need to make easily available combined communal information (e.g., libraries), we need to reward people for working on genuinely new problems (e.g., by requiring literature searches), and we need to insure that people actually publish any solutions which they obtain, instead of secretly gloating that they know something which no one else does (hence, the publish-or-perish ethos).
Hull's main project is to describe exactly how the publication-citation practices in modern science reinforce the traditional cognitive aims of scientific inquiry.
c. The Scientific Credit System
Hull emphatically debunks the romantic myth of scientist as the objective, altruistic problem-solver whose only interest is that Nature be understood (and no matter who wins the Nobel Prize for being the first to probe her inner-most secrets). Scientists, like everyone else, want credit for their successes. However, Hull just as adamantly opposes the cynical view that since scientists que scientists are motivated by mundane ambitions, the products of their inquiry have no special cognitive status. This would be like arguing that since business men and professional athletes are both "in it for the money", it makes no difference whether you put Donald Trump or Magic Johnson on the basketball court! The crucial question is not whether scientists want credit; what matters is which activities they get credit for. Do the proximate rewards reinforce the ultimate aims of science? Can scientists "do well by doing good" science?
Hull believes the publish/citation system is remarkably efficient in rewarding behaviors which result in scientific progress defined in purely cognitive terms. Let us go through the Popperian problem-solving schema laid out in Chapter II and note how it interacts with the social reward system as described by Hull. Here we will emphasize congruences between the two systems. Dysfunctional aspects of the present emphasis on credit will come later.
(i) Choice of problem: If our mundane aim is to get published in scientific journals, we should choose problems which haven't been solved yet, but are ripe for solution. (It is generally difficult to publish unsuccessful solution attempts or interim reports.) It may be wise to form a team so as to be able to tackle problems which others aren't equipped to solve in order to solve them more quickly. Of course, this means we'll have to share credit with our co-authors. We also will need to be able to assess the competence of prospective teammates.
(ii) Working out a tentative solution: Since the first publication often gets the most positive citations, we must work rapidly and secretly, especially if other individuals or teams are pursuing similar lines of inquiry. This is a time for team camaraderie, brainstorming, and constructive criticism of conjectures. We will look for promising helpful hints while refereeing ouIl NPJ§@rant-proposaAwr even their submitted journal articles (although note that in scientific journals submission dates are published). On the other hand, we will be reticent to share preliminary results with anyone who might scoop us. This will also protect our own reputations if the conjecture we're working on turns out to be way off-base.
(iii) Severe Testing: Once we have a solution which has passed preliminary appraisals, we must decide when to publish it. This is a complicated choice. The reasons for publishing as soon as possible are obvious: if we are right, we want to get credit for being first. However, as Hull emphasizes, there are also lots of reasons not to rush into print. If we are quickly shown to be wrong, our reputations are likely to suffer somewhat. But if our conjecture is correct, it will generally lead to other lines of productive research. By temporarily delaying publication, we can explore these ramifications at our leisure and publish everything at once!
Hull points out that in the past, scientists were often reluctant to make their results public. There was (and is) of course a tradition of passing on craft skills and secrets only to apprentices (e.g., alchemy, Stradavarius' violins) and until the patent-system was developed it would be silly to divulge technological innovations. [Insert story of publication/patent timing of super-conductivity and cold-fusion.]
But gentlemen scientists also buried results in desk drawers out of laziness, or caution, or failure of nerve - or because the incentives and opportunities to publish were deficient. For example, here is Dijksterhuis' commentary on his countryman, Isaac Beeckman:
"Beeckman showed the same defects in the matter of science as Leonardo da Vinci. Both were deficient in the tenacity of purpose and powers of concentration required to systematize, finish, record, and publish their inquiries, even if only in one field. Of Faraday's motto: `Work, Finish, Publish', they only took to heart the first injunction. In consequence they either did not advance science at all, or at least to a much smaller extent than they might have done.
Those of Beeckman's ideas which are going to be described here do not therefore really form a link in the chain of development under consideration. However, they are of value because they give the reader some notion of the scientific thought of a gifted man of the early seventeenth century." (The Mechanization of the World Picture, p. 330)
"We shall see more of Beeckman's independent and frequently original way of thinking later: it is to be regretted that this candle never stood on a candle-stick." (ibid., p. 333)
Perhaps we could paraphrase Hull by saying that the difference between 17th century and 20th century science is not in the quality of the candles, but in the ornateness of the candelabra.
(iv) Refutations: In order to function well, the institutions which regulate publishing in science have to balance a variety of desiderata. Science (and the public) benefits when results are published, so there must be opportunities and incentives to publish. On the other hand, it is imperative to maintain quality control over what is published, so one needs to prevail on experts in each field to take time off from their own research to referee articles. Why should they consent to undertake these time-consuming and often unpleasant activites? Well, as any journal editor knows, not everyone does consent, but there are mundane reasons for doing so. It is in every scientist's cognitive interest to keep the communal knowledge store as reliable as possible. And, as I pointed out above, it's always nice to have advance knowledge of what other people working in your area are up to. Furthermore, an excellent way to make sure your own ideas are taken seriously (thus increasing their fitness) is to eliminate, or at least point out the weaknesses in, rival viewpoints.
It is plausible to assume that most great refuting experiments in the history of science were performed by scientists who personally favored alternative hypotheses to the one which is discredited. But I know of no systematic historical studies which would illuminate this question. (And, of course, Lakatos denies that there are any refutations - except through after-the-fact christening.)
(v) Corroborations: Hull emphasizes that for maximum success in science, it is not enough just to publish - one's work must be positively cited. Further articulations of the view by oneself or one's students is one direct way to accomplish this. Writing slightly different versions of the same research directed at slightly different audiences is a quicker way! However, it generally does not pay off to "saganize" one's work by trying to appeal to a lay-audience.
d. Dysfunctional Aspects
Hull emphasizes the nice fit between what we might call the proximate mundane goals of professional success and the ultimate noble aims of the search for scientific understanding. Many professions are not so lucky - the qualities and behaviors required to be a successful politician in an age of TV elections are almost contrary to those which contribute to statesmanship. And there are fewer professional incentives for doctors to stay abreast of new medical developments (unless their patients read about them in the popular press and demand them) than there are for scientists to keep up in their fields. (There is an old joke about dermatology being the best specialty - your patients never die and neither do they get better!)
Professional sports provide another example where there is a good congruence between mundane success (reflected in salaries) and internally defined excellence (extraordinary sporting performances). It would appear that even if athletes were just in it for the money, they would have to play just as well. So it would appear that mundane motivations do not harm sports as long as there is a strong correlation between salary and batting averages.
Yet perhaps it is not just romanticism which makes us suspicious of this cozy conflation of the sacred and the secular. What if coaches become reluctant to call for a sacrifice bunt (because players want to keep their averages up)? What if salaries come to depend on a player's charismatic box-office appeal, not just on box scores? Won't people playing primarily for money be easier prey for point-shaving deals with gamblers?
The general point is this: Any congruence between careerism and love of the professional activity is precarious enough that we are ill-advised to abandon our romantic-sounding rehearsals of the internal aims of science or sport. However, in the cases of dissonance between the success structure and internal aims, such as in medicine and law, we need to focus on institutional reforms, where the direction of the reform is dictated by the internal aims of medicine (which are why society values it in the first place).
So far, we have been following Hull in stressing how well the reward system in science works. But let us now actively seek out dysfunctional elements.
[To follow:
--dysfunctions
--analysis of why science is fun
--how the values society places on scientific inquiry mesh with the mundane and internal value structure of science
--can conceptual inclusive fitness be the ultimate aim?
the demarcation problem re-visited
--what happened to progress?]