It’s my pleasure to share this symposium, which discusses a recent paper by Duncan Pritchard’s (UCI). The paper is named Socially Extended Scientific Knowledge’, and it’s published in Frontiers in Psychology, SI, ‘Distributed and Embodied Cognition in Scientific Contexts’. The symposium includes commentaries by Mirko Farina (Innopolis University), Orestis Palermos (Cardiff University), and Mark Sprevak (University of Edinburgh). It also includes Pritchard’s replies to comments.
Précis of ‘Socially Extended Scientific Knowledge’
By Duncan Pritchard
My paper is essentially an exercise in philosophical cartography (or philosophical hygienics, depending on your point of view). That is, I am trying to chart out the conceptual space in a particular area and in the process show that there are more dialectical options than one might have hitherto imagined. My specific concern is with a kind of social knowledge that is characteristically found in collaborative scientific enterprises, though the conclusions I argue for would equally apply to any variety of social knowledge that had the relevant properties. The issue in hand is how the extended cognition program plays out when applied to social scientific knowledge.
According to extended cognition, in the right conditions a subject’s cognitive processes can incorporate factors that are outwith that subject’s brain and central nervous system, including features of the subject’s environment, such as technology. I don’t argue for extended cognition in the paper, but rather treat it as part of the accepted theoretical background. My interest is rather in how such a research program plays out in terms of knowledge, particularly social scientific knowledge. On the general question of knowledge, I argue that the kinds of conditions under which extended cognition is plausible—which involves the external factors in question becoming highly integrated with the subject’s on-board cognitive faculties—are also conditions that would make extended knowledge (i.e., knowledge that is the result of extended cognitive processes) plausible. In this vein, I offer a virtue-theoretic account of knowledge that can accommodate extended knowledge. Very roughly, what matters for knowledge is the explanatory role that one’s cognitive agency plays in one’s cognitive success. In particular, so long as one’s cognitive agency is playing a significant explanatory role in accounting for one’s cognitive success, then one is in the market for knowledge. Crucially, however, where one’s cognitive agency incorporates external factors that are integrated with one’s on-board cognitive abilities (i.e., where one’s cognitive agency is extended), then the knowledge that results is extended knowledge.
One advantage of thinking of extended knowledge in this way is that it enables us to distinguish between genuine technologically extended knowledge and merely technologically-facilitated knowledge. In the former case, the technology is a proper part of the extended cognitive process, while in the latter case the technology is merely facilitating one’s cognition without being a proper part of the relevant cognitive process. Such a distinction is clearly important, as not every reliance on technology amounts to extended cognition. For example, we don’t want to treat every instance where a scientist successfully employs a scientific instrument as being a bona fide case of technologically extended scientific knowledge. What marks the distinction, on the virtue-theoretic view, is whether the technology is sufficiently integrated with the subject’s on-board cognitive processes such that it is a proper part of the subject’s extended cognitive agency that is accounting for her cognitive success, or whether it is merely carrying an explanatory load in this regard as something external to her cognitive agency.
Technologically extended knowledge is clearly individual knowledge, even despite the external contribution of the technology. Where things get interesting from a social epistemic perspective is when one turns to socially extended knowledge¾i.e., where the external factors in question are social rather than non-social (such as technological). This is of course particularly relevant to the kind of collaborative epistemic enterprise commonly involved in scientific inquiry. In this regard I urge a three-level distinction between kinds of social scientific knowledge. The first two levels reflect the distinction we’ve just drawn between technologically extended knowledge and technologically-facilitated knowledge, except with that distinction now applied to social knowledge. Just as this distinction is important in cases of technologically extended knowledge, so we also need to draw a parallel distinction with regard to socially extended knowledge. In particular, the presence of social factors playing an explanatory role in scientific cognitive success doesn’t entail that genuine socially extended scientific knowledge is on display, as it further depends on whether these social factors have been properly integrated within the relevant cognitive processes so that they are proper parts of these cognitive processes. If they haven’t, then all that is being manifested is socially-facilitated scientific knowledge. I suggest that many cases that are offered of socially extended scientific knowledge might in fact be better understood as merely socially-facilitated scientific knowledge. Just as we don’t want any successful use of an instrument by a scientist to thereby qualify as technologically extended scientific knowledge, so we don’t want any successful scientific reliance on social factors to thereby qualify as socially extended scientific knowledge.
One mistake that I think some commentators make is to suppose that socially extended knowledge must thereby be a kind of irreducible group knowledge (distributed knowledge, as I call it). This is a tempting inference to draw, but I think it should be resisted. It’s tempting because we might suppose that if the production of knowledge is genuinely via a socially extended cognitive process, then there must be a sense in which this is group knowledge rather than individual knowledge. The inference is too quick, however, as I show that just as technologically extended knowledge is nonetheless individual knowledge, so socially extended knowledge can also be individual knowledge. That is, it doesn’t follow from the fact that the external factor that is integrated into one’s cognitive processes is social rather than merely technological that the knowledge that results is distributed rather than individual. We thus get a further distinction between the second and third levels of social scientific knowledge, which concerns (individual) socially extended scientific knowledge and distributed scientific knowledge.
A key benefit of this exercise in philosophical cartography is that it reminds us to be more careful about how we are understanding social scientific knowledge (and, for that matter, other kinds of social knowledge). On the one hand, we should be wary of quickly concluding from the fact that social factors are involved in the production of scientific knowledge that we are dealing with a bona fide case of socially extended scientific knowledge (as it might just be merely socially-facilitated scientific knowledge). On the other hand, even when we have a genuine case of socially extended scientific knowledge, it doesn’t immediately follow that this knowledge must therefore be distributed knowledge rather than individual knowledge. As I explain, I think this second distinction is particularly important in the scientific case, as there might be reasons for thinking that the structures of scientific inquiry¾particularly in terms of the distribution of credit when things go right, and the distribution of blame when they do not¾ensure that scientific knowledge, even when socially extended, is nonetheless typically individual in nature.
Duncan Pritchard
University of California, Irvine