Author Information: Boaz Miller, Tel Aviv University, firstname.lastname@example.org, The Bubble Chamber; Matthew J. Brown, University of Texas, Dallas, email@example.com; Sanford C. Goldberg, Northwestern University, firstname.lastname@example.org; Kareem Khalifa, Middlebury College, email@example.com
Please refer to:
- Fuller, Steve 2012. “Social Epistemology: A Quarter Century Itinerary.” Social Epistemology 26 (3-4): 267-283.
- Table of Contents: On the Future Direction of Social Epistemology,
Editor’s Note: Boaz Miller, Matt Brown, Sandy Goldberg and Kareem Khalifa were kind enough to allow their comments, made on both Social Epistemology’s Facebook page and in the comment section below, to be integrated and formatted on this page.
Fuller’s rhetoric [in “Social Epistemology: A Quarter-Century Itinerary”, 2012] is hyperbolic, but much of his criticism is point on. Here are a few thoughts.
1. Much of ASE [Analytic Social Epistemology] works with a thin notion of the social. It plays an old game (roughly, the post-Gettier game of arguing about knowledge from institutions on simplified cases) in a new field – the (thin) social field. The main reason that many ASE scholars do not look at real epistemic practices is that this would force them to play a new game, which is something they do not want to do. They do not really care about real socially-situated cases of knowledge, which is in itself legitimate, but has implications to the depth, empirical adequacy, and scope of their work – as I explain in (2). Kareem Khalifa has a forthcoming book review of Millar, Haddock, and Pritchard in Mind where he says similar things.
2. The combination of relying on intuitions and lacking empirical knowledge about knowledge seriously hinders the depth and scope of ASE. It leads to conservatism, and to theories of knowledge that are not empirically adequate. The situation is actually worse that Fuller describes, because Tycho’s astronomical theory was empirically adequate, while current mainstream ASE theories of knowledge are largely not. Sandy’s [Goldberg, see also “’Analytic Social Epistemology’ and the Epistemic Significance of Other Minds”] own work is excluded, but it is not (yet?) mainstream.
3. There are social epistemologists who are attentive to practice. Most of them are philosophers of science. Many of them attended the last Society of Philosophy of Science in Practice (SPSP) conference in Toronto in June (guess who didn’t come?). They include people such as Helen Longino, Heather Douglas, Matt Brown, Ben Almassi, Kareem Khalifa, Dan Hicks, Axel Gelfert, Torsten Wilholt, and me. (Please forgive me if I forgot you). Fuller used to belong to this camp, but now he writes in a more meta-critical way. Many ASE people don’t read them. Their attitude varies from indifference to dismissal — while the opposite is not true, at least with respect to subjects the practice people care about, such as trust, and expertise. Again, I don’t think that the reason that ASE does not engage this camp is merely that ASE is a young field.
4. Goldman’s “Veritistic Social Epistemology”, which is the exception to most ASE in its applied aspirations is, to a large extent, unsuccessful, for three reasons. First, in practice, there are many epistemic standards of appraisal and desiderata besides truth, which are themselves socially determined: knowledge, justification, rationality, understanding, conformity with pragmatic aims, etc. Truth is too narrow a prism to look at many problems. Second, truth is too narrow a semantic relation between theory and appearances or the world. There are other relations such as isomorphism, similarity, empirical adequacy, conformity with social values, etc. Again, which standard is required in which context is socially determined, and itself a proper subject for normative philosophical analysis. Third, truth is not easily reached, and in many cases, we need to settle for less. For these reasons, many times, a veritistic analysis of a domain or a case study ends up being either trivial or ad hoc. Calling the large social epistemic literature that attains to standards other than truth “veriphobic” is highly uncharitable.
Matthew J. Brown
Boaz, your comments strike me as spot on about ASE, at least my experience of it. I’m not as much of an expert as most people who might be reading this thread, but I’ve read Goldman, and I reviewed Goldberg’s 2007 book, and I’ve made something of a go at it. In particular, Goldberg’s rejection of Fuller’s (2012) claim that ASE’s lack of attention to knowledge in practice is “a strategy that is often defended in the name of maximum abstraction and generality” seems disingenuous, or self-deceived, or at least far too optimistic. To paraphrase Dewey, epistemology is normative because and to the extent that it is empirical. It is hard for me to find much of value in ASE for the kind of work I am interested in doing, viz, figuring out how knowledge does and ought to actually work, particularly, but not exclusively, in scientific practice. (As a Feyerabendian, though, I say let 1000 flowers bloom, so long as those flowers are willing to extend the same courtesy to my projects.)
To be fair, not everything is grim. There are moves within epistemology toward more empirically informed writing. The experimental philosophy movement tries to systematize intuitions, which is a step in the right direction, even if their methodology is sometimes confused. Because he writes right from within the analytic tradition, Sandy Goldberg is being read and engaged with, and he is challenging existing doctrines. And so is Miranda Frickers’ work very well accepted, and she uses examples from scientific practice and from realist prose in her arguments. People such as Kitcher, Thagard, and Lipton, are also not entirely ignored, and they have a foot in each camp.
Matt [Brown], I might recommend that you extend the same courtesy to others’ work that you rightly ask of others to extend to yours. Two of the things I find particularly saddening in the exchange to which my paper is party is (1) the degree to which both sides attribute rather uncharitable motives and rationales to the other side, and (2) the degree to which both sides seem quick to dismiss the entire tradition on the other side, when in fact it is particular examples of that tradition that they really have in mind.
Regarding (1): I welcome honest criticism; but I think it best if all sides refrain from casting the other side as “so retrograde” or as “self-deceived” or … — the list could go on. Regarding (2): insofar as there really is a tradition of “ASE,” it ought to be subject to serious scrutiny, as any other tradition ought to be subject, but to say that an entire tradition is ideologically committed to a particular view out of a motive for abstraction strikes me as not particularly attuned to the diversity of projects and interests among the relevant set of researchers. I would invite you to be part of a constructive dialogue; let’s not reinforce the very fault lines that it seems you and I both agree are destructive.
I think it’s perfectly legitimate not to be interested in social practice, and there is very good work in epistemology that is not practice oriented at all. This is even legitimate in social epistemology, though I think it produces work of limited value (I am not using “limited” as a derogatory term here). What is illegitimate in my view, is producing work that is purported to be relevant to actual practice without looking more deeply at this practice. It’s like trying to apply abstracted armchair ethics to real-world complicated moral dilemmas, which may be even dangerous at times.
Matthew J. Brown
Sandy [Goldberg], fair enough. As I prefaced my earlier comments with, my experience with ASE is limited, which surely means that I am not attuned to the diversity in that tradition, if it makes sense to even identify it as a tradition. And you are quite right that charitable engagement and constructive dialogue are to be preferred, all other things being equal.
Let me try to be a little more constructive. Fuller claims (more or less correctly, as all sides seem to agree) that ASE is not sufficiently attentive to actual knowledge practices. He adds that this is part of the Plan, which rings true to me and to Boaz Miller (in a more qualified sense). Goldberg claims that this is not, in fact, part of the Plan, just a matter of work so far left undone, and indeed, this work is required, given the aims of ASE. This is obviously an important dispute: is a lack of attention to knowledge practice a temporary or permanent problem for ASE? How are we to answer the question?
My readings and my engagement with scholars in the ASE tradition (and in most other parts of “analytic” philosophy) support Fuller’s claims. I have encountered a consistent lack of attention to epistemic practice when such attention seems obviously relevant, and an ongoing resistance to take it seriously. But as I’ve said, my experience is narrow and limited, so that’s not a very good way to go about it. Goldberg’s paper suggests a few strategies. First, since the aims or rationale for ASE require careful attention to epistemic practices, this just implies that it is not part of the Plan to have a minimal understanding of epistemic practices. I don’t believe that it implies anything of the sort. It is possible to admit that to answer these questions, we should know more about the empirical situation, while nonetheless never doing anything to investigate the empirical situation, and resisting any critiques based on empirical investigation. (I would say that this is not only possible, but in many areas of analytic philosophy, quite common.)
The reasons that Goldberg gives for why ASEs have hitherto paid little attention to practice strike me as reasons for further suspicion. Why think that there was meaningful work to be done ahead of paying attention to how knowledge works in practice, or that one could answer normative questions on the basis of a minimal understanding of the practice those questions are normative for? Perhaps the principle of charity requires us to assume that if doing X is required by the aims of Y, then not doing X cannot be a part of Y’s plan or core commitments. This seems to me like charity run amok. It requires us to assume that Y is aware of and agrees that X is required, and that Y has no additional aims or commitments that go contrary to X. We must assume that Y is not self-deceived, that Y does not suffer from a rationalistic neurosis (in Nicholas Maxwell’s sense, i.e., that it does not misrepresent its aims to itself), and so on. These sort of worries seem very much in line with what Fuller is thinking.
Okay, second: we can wait and see. Goldberg predicts that the work of attending to practice is or will start and become common in ASE. Miller supports this with examples. (A) Miranda Fricker is one that I would agree is exemplary, but not one I would have thought to group with the other ASEs under discussion. (Again, the problems of individuating philosophical “traditions”.) (B) Philosophers of science who ASEs pay attention to. This doesn’t really seem to me to help. But fine: if this critique IS taken up by the ASE community and we start to see work that engages carefully with practice as a major force in ASE, then I think we all agree that this would be an improvement (and that Fuller’s critique succeeds as constructive criticism but fails as damning criticism).
On another note, it seems to me that Goldberg’s responses to Fuller’s 2nd and 3rd criticisms fail to charitably engage with what Fuller is saying. In (2), Fuller objects to a form of social epistemology that has no interest or even seeks to minimize the impact of philosophers on social epistemic practice. Pragmatism is a red herring. I find it puzzling that for someone who so emphasizes the normative dimensions of epistemology, minimizing the impact of epistemology does not seem troubling. Why do we make normative claims if not to critique and improve the practice? In (3), Fuller seems to me concerned not at all with the status quo in epistemology, but rather, with the status quo of epistemic practice. So the fact that ASE is challenging the status quo in epistemology is again, a red herring. The Feyerabendian courtesy I alluded to is one in which we recognize the right of different intellectual approaches to exist and flourish, that we not use institutional power to squelch them, etc. I would happily extend that courtesy to a social epistemology whose Plan is to rigorously avoid actual practice, though I would doubt that such an approach would pay off in ways that would interest me. I would be willing to extend much more than that to ASE as Goldberg envisions it, once it has started to test its claims by engaging with actual epistemic practice. That is a tradition that I would expect fruitful dialogue with.
Matt, you ask: “Why do we make normative claims if not to critique and improve the practice?” I think that many ASE people would say that their aim is simply to analyze and make explicit the existing epistemic norms. This is, for example, why they deal so much with the norm of assertion, and pay excessive attention to knowledge attributions. A common assumption among ASE people, especially reliabilists I think, is that there is nothing fundamentally wrong with our existing epistemic norms. Of course, more attention to empirical practice may change one’s mind about this. This is why feminist epistemologists don’t think so. Philosophers of science are usually also more suspicious of existing norms, and in general come from a very prescriptivist tradition.
Sandy, I wonder how much you speak for yourself, and how much for ASE as a group, but I recognize that you are in a leadership position and have the power to set agendas for ASE, influence graduate students, journal editors, etc. So I take the purpose of your paper to be not just testifying about reality, but affecting it as well.
I wouldn’t pretend to “speak for” ASE as a group. That’s something for which I do not have, and would not claim, any authority. I do, however, think that my description of the programme for social epistemology captures something that unifies a good deal of the work that goes under the “ASE” label. This said, I am under no illusions; I suspect many who would be considered ASE theorists will disagree with some, and perhaps much, of what I have to say. So be it.
Matt, I fear I fail to appreciate the force of your defense of Fuller on the second and third points. I don’t see why characterizing epistemic norms is not itself a useful activity, whether or not anyone changes their behaviors. It would be great if they did change their behaviors, to match the norms (assuming one’s characterization of the norm is correct). Still, unless one is assuming a kind of pragmatism, I don’t see why it follows, from the fact that there are no practical changes at all following one’s characterization of the norms of the practice, that one’s characterization is ipso facto deficient. But perhaps I am missing something here.
Regarding your defense of Fuller on the third point, again I gave evidence to think that epistemic practice itself is far richer than Fuller’s (3) suggests. I take it you were unmoved by that evidence, though you don’t say why. The big question is on Fuller’s first criticism. I have had my say. I suspect that there is some talking-past-each-other for the reason you yourself anticipate; namely, that it is hard to say what is in and what is outside of “the tradition of ASE.” Indeed, it is even unclear that there is a single tradition here in the first place. I am sympathetic to that point. That is why I don’t particularly care for the “ASE” label itself. But do remember that it is not my label. I was responding to Fuller’s use of it. I would say that if there is no single tradition here, this is so much more grist for my mill, since in that case at best Fuller is saying something true about the work of a subset of those people in the so-called “analytic social epistemology” tradition, at worst he is saying something false about a tradition that is much more diverse than (1) describes.
It is ironic that Fuller, who talks about the elusiveness of consensus and traditions as a fictional objects retroactively constructed by historians for the sake of their arguments, talks about ASE as a single robust tradition.
Matthew J. Brown
Boaz, the claim that ASE is interested in explicating EXISTING norms of the practice is inconsistent with Sandy’s excuse for why ASE has not so far paid attention to the actual practice: That ASEs are interested in the normative and not the descriptive. That kind of normative work is impossible without careful attention to the practuce itself. On the other hand, the idea of articulating norms that aren’t just implicit in the actual practice without some connection to critique and reform of the practice simple strikes me as incoherent. If this is a commitment to some kind of pragmatism, so be it, though it certainly isn’t the sort of straw man pragmatism that Sandy mentions in his paper.
In (many) ASE people’s eyes, they are concerned with actual practice, and they think that intuitions on simplified mundane cases provide excellent evidence on actual practice. They regard themselves as empirically informed. I personally think that there are many problems with this approach: (1) mundane testimonial cases and their likes, e.g., asking a stranger for directions, constitute a small and unrepresentative part of all knowledge dissemination and generation social exchanges; in particular, they do not represent the scientific practices; (2) simplified cases tend to idealize away or push into the background real-world difference makers, which are highly theoretically significant; (3) analytic philosophers’ intuitions are spoiled and tainted by their previous training and indoctrination; (4) they are further typically uninformed by the empirical knowledge referred to in (1) and (2), which makes them even less reliable. But this is not how they see things.
Let me begin with a worry about ASE that I discuss in review of Haddock, Millar, and Pritchard’s Social Epistemology (forthcoming in Mind), as a way of framing the issues that I see emerging from this discussion.
ASE tends to trade in the following:
(P1) All entities with epistemic property F have epistemic property E;
(P2) X is a social entity with F; so
(C) X is E.
If this ‘epistemic explanation’ happens to be the best explanation of an epistemic phenomenon, then great. But I think it a misnomer to say that this is social epistemology. (Toy example: let F be a reliable belief-forming process; E, warrant; X, testimony) Rather, we are simply applying traditional epistemology to epistemic phenomena that are only incidentally social. (That X is a social entity plays no essential role in the explanation.) So this is one way of unpacking what Boaz means when he says that ASE uses a ‘thin notion’ of the social.
For instance, consider Sandy’s contribution to this volume. Sandy argues that a knower is justified in believing that ~p if she: (a) is justified in believing that had p been true, then the her sources would have found out and successfully communicated to her that p, and (b) gains no information that p from her sources. While talk of communication may appear to spare this from the aforementioned criticism, this is not so. The counterfactual dependence between the proposition’s truth and the source’s indications (broadly construed) drives the account, and the source’s indication being another person’s communication is incidental. Hence, at best, Sandy’s social considerations do epistemological mop-up work.
Let me offer a rather different way of conceiving of social epistemology. Social epistemology should provide social explanations of epistemic phenomena (SEEP). A SEEP might be characterized as follows:
(P1) All entities with social properties S have epistemic property E;
(P2) X is an entity with S; so
(C) X is E.
(Forgive the Hempelianism — it simply makes for a tidy presentation. The key point is that a social property is showcased in the explanans, and an epistemic property is the star in the explanandum. That may be cashed out in a non-DN, e.g. causal, idiom)
Ideally, these should be the best explanations of the epistemic phenomenon in question. Now let me say that analytic epistemology (social or otherwise) often exhibits many virtues that we seek in an explanation: it aims to have broad scope, invokes relatively few parameters (J,T,B, anti-Gettier), and is fairly stringent and precise. However, it falls short on other dimensions of explanatory evaluation, such as empirical support and lack of ad hocness, and I think that these are precisely the points on which Boaz and Matt have pushed.
Given Sandy’s conceding that Fuller’s First Criticism is more or less correct, it appears that ASE’s social explanations are not well supported by the extant evidence. We may try to diagnose why ASE has not yet held itself to a higher standard of empirical adequacy, but I think this leads very quickly to the rather unfriendly “diagnoses” (e.g. of self-deception) that Sandy rightly thinks are counterproductive to conversation. So, let us simply register that if ASE aspires to provide SEEP’s, it should develop a more thoroughgoing empirical program to complement its otherwise fruitful theoretical program.
Nevertheless, we can start to sketch the contours of that empirical program. In particular, there must be some theory-neutral way of determining whether or not an epistemic property (E, in the schema above) obtains. This is precisely where issues of the normative and the descriptive intersect. The issue is normative because epistemic properties — such as truth, justification, reliability, etc — are what make our beliefs good. The issue is also descriptive in that we need to figure out the conditions wherein these epistemic properties are instantiated. In other words, we are not asking, “Ought this X be an E?” Rather, we’re asking “Is this X an E?” This was another point of disagreement between Matt and Sandy, and I hope this helps to clarify the interaction of the normative and descriptive.
In particular, Sandy, I understood you to be saying that ASE’s use of philosophical intuitions was the best theory-neutral means of ascertaining whether an individual bears an epistemic property. That’s not obvious to me. Philosophers of science often defer to the reliability of scientific methods to determin e— defeasibly — if an epistemic property obtains. (My heuristic is something like the following: if p passes the latest t-test and I can’t think of any confounds that have not been controlled for, then p is justified.) If, broadly speaking, the most plausible reason to think that scientists judge something to be epistemically significant is because they had incomplete information, overlooked something, or engaged in faulty reasoning, then philosophers of science can criticize the practice. By contrast, practitioners of ASE take intuitions as their chief theory-neutral arbiters, but I think for many outside of ASE (of which I include Boaz, Matt, and myself) question just how theoretically innocent these intuitions are.
It seems to me that where scientific expertise conflicts with philosophical intuition, it will almost always be the case that scientific expertise determines how we should proceed—even in the case where the intuition is correct! How is this possible? Suppose that scientific practice suggests p and philosophical intuition not-p. Generally, there is nothing further to explore with philosophical intuition — it is brute. By contrast, scientific practices are usually complex states of affairs that can be unpacked. Take something like the data that scientists gather. Even though data are generally thought to be brute, the sampling methods, forms of data reduction, etc. are all things that can be scrutinized to determine whether the data are reliable. So, should we opt for not-p, the reason is not because the intuition provided very strong evidence, but because we found something deficient in the scientific practice.
What of cases where there is no scientific expertise on which to lean? Then perhaps we lean on intuitions, or perhaps we opt for some other evidence, e.g. experimental philosophy or experimental results from a non-philosophical field (e.g. psychology). Indeed, I suspect that with many ASE topics (testimony, disagreement, etc.) there may be experiments that actually involve reasoning tasks involving small groups, as opposed to the standard x-phi practice of reading of vignettes. This would surely provide better theory neutral measures of an epistemic property than running a thought-experiment by a handful of fellow philosophers.
Finally, note that the chief explanatory factor in a SEEP is a social property. (If “social property” sounds weird, replace “property” with “factor,” “structure,” “mechanism,” etc.) So a good producer of a SEEP will have to be well versed in a wide variety of social properties, presumably by being a skillful user of at least one social-scientific framework. Let us even construe a social-scientific framework broadly, e.g. by including history as a social science. It is unclear to me that ASE’ists are conversant in this way. So, this is another sense in which ASE has a thin notion of the social.
So, if I am right, ASE has three major challenges, which I’ll restate along with some proposed solutions.
First, ASE should try to explain epistemic phenomena in a manner where social factors are not just incidental. I suggest deploying more SEEP’s. Second, ASE must develop an empirical program that has theory-neutral means of ascertaining whether certain epistemic properties obtain. I suggest a more deliberate fusing of ASE with the philosophy of scientific practice and with experimental disciplines. Third, ASE must develop a richer repository of social concepts, ideally by engaging the social sciences more regularly. Perhaps more ASE’ists should engage the philosophy of social science (it’s worth noting that virtually all of the leading philosophers of social science — Alison Wylie, Jim Bohman, Paul Roth, Mark Risjord, David Henderson — are conversant in ASE and the philosophy of scientific practice, so I think they would be great folks to bring into this discussion).