The Matthew effect is universally derided. Could this derision be an error? Could it be that the effect not only makes a positive contribution to the scientific enterprise, but is mandated by the reward system itself? I will propose an explanation of the effect on which the answer to both questions is yes.
The explanation rests on three premises. First is the posit that the reward system in science bestows credit in proportion to a scientist’s contribution to society. Second is the notion, not much remarked on in the science studies literature, that the value of a scientific contribution may depend in part on its epistemic standing: on the degree of trust that society can place in a set of scientific results. Third is the observation that, all other things being equal, the credibility of a scientist’s results will increase with the scientist’s eminence. Let me say something more about each of these....
The rewards of science—principally fame, respect, and the material benefits that they bring, such as research grants, laboratory space, and eager graduate students—are distributed, for the most part by scientists themselves, in proportion to the contribution thought to have been made by a scientist’s work. The larger the contribution, the more credit the scientist receives from their colleagues, hence the more fame, respect, and material goods. There is much more to say about the reward system in science, and in particular there is more to say as to what counts as a contribution (Merton 1957; Kitcher 1993; Strevens 2003). For the purposes of engaging with the Matthew effect it will be enough to point out that, when determining the credit to be bestowed on a discoverer, the discovery’s contribution to society at large counts for something. The greatness of a discovery depends, that is, not only on the pleasures found in the contemplation of profound knowledge, but in the good that it does for a population not often in a position to fathom its depths.
Medical science provides many examples of this dictum. Pasteur’s supreme standing is surely in part due to the many lives saved as a consequence, direct or indirect, of his discoveries. The status of a cure for cancer as a ne plus ultra of medical research is due not only to the understanding of the mechanisms of cellular reproduction that it would no doubt involve, but also to the extent of the misery alleviated by such a cure.
An earlier paper on the reward system (Strevens, 2003) provides a basis for a more theoretical argument that (and explanation why) credit is indexed to overall benefit to society rather than to some particular benefit. In that paper, I argue that the scientific reward system works, by fine tuning the allocation of labour among available research programmes, to maximise whatever contribution is used to apportion credit. If contribution to society determines credit, then, contribution to society is maximised. Given the further premise that the reward system is responsive to the needs of society as a whole (ibid, Section 5.2), you have some reason to expect the allocation of credit to reflect overall good.--- Michael Strevens (2006) "The role of the Matthew effect in science"SHPS.
In a Facebook discussion prompted by Shen-Yi Lao's sharing of a recent article published in PNAS (by some colleagues at the University of Amsterdam),* Edouard Machery appealed to, and reminded me of, Michael Strevens's paper.1 This paper defends (at least in the abstract) the fairness of the Matthew Effect (in the paper the effect is explained). While I will be critical, the paper is worth reading for lots of reasons (especially for its very nice discussion of the retrospective nature, or not, of the Matthew effect) not the least its clarity of prose.
The Matthew effect (described by Merton and he credits it to his wife Harriet Zuckerman) amounts to the idea "that for equally good scientific work, renowned scientists tend to get more credit than unknown scientists" (there are different formulations of the effect--again Strevens is useful). As Strevens discusses the effect violates norms of fairness and also widely shared commitments within science about how the scientific system ought to operate according to merit. In what follows I accept these norms (although I have been hesitant about meritocracy recall and here) . In addition, there is a concern that the effect has bad consequences on the reward system itself--it is seen to be corrosive. Merton himself was attracted to an unintended consequence explanation of the effect, but Strevens tries to show that "the effect not only directly does good, but does good by following, rather than by contravening, the relevant social norms of science."
Before I get to that it is worth noting explicitly an effect of the Matthew effect that Strevens is silent about, although it is completely in accord with his analysis. That is, the Matthew effect generates and entrenches a status hierarchy in the scientific credit system;this, in turn, almost certainly reinforces the other economic (etc.) hierarchical elements (recall this post) of the scientific enterprise as we find it in the real world. Even if the hierarchy were merited one may still wonder -- and I think this ought to be a methodological desideratum -- if this hierarchy (one is promoting) does not generate bad epistemic and moral consequences (servility, etc.) one might wish to avoid.
With that in place, it should be fairly clear why I think there is something Panglossian about the first premise, that is, that (a) the "reward system in science bestows credit in proportion to a scientist’s contribution to society." Obviously if that were true, and (b) one accept the 'contribution to society' as the relevant normative criterion, then Strevens has met the methodological challenge of the previous paragraph. Let me take (a) and (b) in reverse order.
Let's assume, first, for the sake of argument, that contributing to society is a good thing (it counts for something), and so let's assume the criterion is not vicious. But it is not self-evidently the right criterion. It is notable that Strevens only treats the contemplative pleasure of the scientist as the only alternative relevant concern. The sense that the scientific enterprise is a good partially (normatively) divorced from either society's benefit or the scientist's pleasure has no standing here. That's peculiar because one might have thought that status in science ought to correlate with contribution to, say, expanding (to use sociology of science lingo) the research frontier or enhancing our knowledge (regardless of pleasure or good consequences). We admired Hawking for his (relatively useless) contributions to cosmology (black-holes, etc.) not the benefit to society or his private pleasure (and perhaps we admired him too for reasons unrelated to scientific merit at all).**
Of course, one could write a whole book on how one should understand and measure contributions to society relative to variants on the theme. Why not apply a difference principle here? That's not as silly as it sounds--given that the Matthew effect entrenches hierarchy. One may well think that the just version of Strevens' own criterion is something like this: does the scientific credit system facilitate contributions to society and do these lift the boats of the the least advantaged groups in society? Asking the question is answering it: even in medicine and public health, making contributions to fight diseases that inflict the poor (and racial minorities) are often least well rewarded in recognition and research funding. Obviously I have not argued for the difference principle in health care relevant science (not my idea, Norman Daniels got there a lifetime ago). The present point is just to offer an example that Strevens's criterion is by no means to be found among the normatively more desirable ones when we reflect on the real world.
One may also wonder why from a scientific perspective contribution to society is, in fact, the right way to distribute credit internally. (Again ,note **.) So, there is a big difference between designing a grant system with public funding (which may well privilege contribution to society) and the epistemic rationales for the reward/credit system "in science." Merton himself thought (as Strevens acknowledges) that the scientific credit system was a kind of reward for risks taken. In particular, the solving of very challenging and important scientific problems.
Now clearly what counts as challenging and important is relative to epistemic and moral/social considerations, and so a scientific community (say in a very noblesse oblige environment) could decide that the important contributions are those that also contribute to society in some way. But they need not do so. In fact, to do so would violate the spirit of basic or pure research. (Of course, it is possible that such research has good social consequences; but the recognition one receives for such successful research is distinct from it.) In some fields the central or challenging problems may well be driven by society's needs, and sometimes society's needs are fairly unified. But there are plenty of applied and engineering (and biomedical) sciences that are partially captured by economic interest groups of various kinds. It is by no means obvious that what gets rewarded in such fields (fill in your favorite target.) really contributes to society as a whole. So, in an imperfect society, one can't assume that the Matthew effect will enhance a reward system in science that is normatively desirable.++
Finally, one also wonders who gets to decide what counts as a contribution to society if a society itself is seriously imperfect (as all existing societies are). For,one reason to grant the sciences normative autonomy in some sense (some other time I'll qualify this a bit more) is that science is one of the great agents of social change. For, sciences can serve all of humanity even if they undermine the ways of life of the existing societies that host them.+
1. To be honest, I have not kept up with the responses to his paper, so perhaps what I am about to say has been fully digested. Mea Culpa.
*I won't discuss this paper because it agrees with my experiences in the funding schemes they describe (and about which I have published and blogged aplenty).
**In the piece Strevens primarily focuses on medicine; I think that makes his presentation seem more natural/obvious. than it really is.
+This is not always good, of course.
++I have described elsewhere how the members of the Chicago School of Economics were very fond of Merton (professionally and personally). The present post helps explain this, in part. They share a kind of optimism about how reality approximates a free market in ideas with a proper functioning credit/recognition system.