Methodological Conservativism

I haven't posted in a while, and just for the sakes of keeping the blog active I am going to post some uncompleted notes for a quite a long post I was writing some time ago and now do not think I will have time to finish. So here is what I had on the idea of methodological radicalism. It may at least be useful as a collection of links to various papers relating to the proper infusion of ethical or political values into our argumentation or modes of epistemic evaluation.

Here are two senses of "method" one might use - first, a method is a procedure one uses for generating claims to endorse. So in this sense, for instance, my method for working out what the temperature is in the morning is to check the weather app on my phone. Second, a method might be a procedure one uses to justify your claims. So, in this sense, my method for justifying my beliefs about the temperature might be to tell you the above procedure and perhaps add some reasons to believe that the weather app is generally reliable about such things. These things are not wholly separable - the example chosen to illustrate that sometimes part of my method for justifying claims goes via pointing to features of the method I had for discovering it. But they also can come apart: I might come to believe a theorem after having a dream in which I was given the strong impression it was true, and only justify it through a proof at some later date.

A recent paper in Radical Philosophy has argued that feminists ought be methodological radicals, endorsing the following claim:

These feminists suggest that, if we are to combat sexist and racist social formations, we therefore need to complement our political radicalism with a methodological radicalism that involves making use of the practical power of ethically non-neutral resources, conceived as in themselves cognitively authoritative.

That is to say, the essay argues that the aforementioned feminists are correct to so think, takes it for granted (it is addressed to self-identified feminists) that we should combat sexist and racist social formations, and so argues that we should "mak[e] use of the practical power of ethically non-neutral resources, conceived as in themselves cognitively authoritative." The article pitches itself against analytical feminism, so I thought I would make note of a number of ways in which analytical philosophers (sometimes explicitly feminists, but I am assuming that analytical feminists will be allowed to draw upon the work of other analytical philosophers and still count as analytical feminists) have been willing to countenance ethically non-neutral resources, conceived as in themselves cognitively authoritative.

• The classic case is the inductive risk argument. We never settle what is true with certainty. How much evidence short of certainty, therefore, must we require in favour of a proposition before we can say that it has been sufficiently well confirmed to (at least provisionally) accept? Round my neck of the woods it is very widely believed that any answer to this question will necessitate some procedure which is not ethically neutral, since it will essentially involve deciding whether it is worse to believe a falsehood or worse to fail to believe a truth.  This is an interesting case in that it seems to touch upon both senses of method. One does in fact need to decide for oneself when one has done enough inquiry to be sufficiently convinced of whatever conclusion you have reached. Further, one must have some notion of "good enough evidence to be warranted in my conclusion" to make the case that one is justified. 

• There is also the sense in which conceptual schemes may be voluntarily adopted by an individual or community, and in so doing reflect our/their socio-political values. Inquiry necessarily involves categorising things and drawing attention to some similarities over others. In the first sense, one's method of categorisation will therefore inevitably reflect a sense of what it is important to draw attention to, which things ought be considered alike, and will in many cases by so categorising prompt responses of a certain sort. For some illuminating discussions of this sort of thing in science see here and here.  This very blog has featured discussion of some philosophers who would say that conscious choice of conceptual schemes on ethically non-neutral grounds is required of us. I have a published historical study of a certain application of this perspective, and a paper where we engage in conscious conceptual engineering for political reasons. I have seen some interpretive debate on this point, but I at least  think of something like this as lying behind Haslanger's notion of ameliorative analysis. I should note that whereas I think the inductive risk argument is very widely accepted, this sort of position is more controversial and those of a more realist bent (perhaps this, for instance) may think that there is an ethically neutral but properly favoured conceptual scheme. As is discussed in the Radical Philosophy paper, Fricker's idea of hermeneutic injustice is probably an instance of this sort.

• Related to one or both of the above, one might think that choice of deductive inference procedures is not ethically neutral. Arguably a big part of one's method (in the second sense) is what one thinks may be justifiably inferred from a given body of claims. From a certain perspective (well explained here) that may be thought of as broadly Carnapian, what is deducible from what is as much a matter of social decision (and as much open to ethico-politico evaluation) as one's inductive inference rules or one's general conceptual scheme. Some works exploring that idea can be found in this collection here. I should say in this case that while I think the general point is very well taken and I agree with it (though again it is controversial, for a good explanation of logical realism and an exploration of its consequences see here) I am yet to be persuaded of any application of it. It does not yet seem to me that we have found an area of inquiry or life where it has actually been all that politically significant or useful to adopt any deductive logic or another on recognisably partisan ethical or political grounds. Still, the idea is in there air and has received some sympathetic discussion.

• Lots of people nowadays are Bayesians, some subjective Bayesians. Even folk who are not ideologically all that committed to Bayesianism will make use of Bayesian discovery procedures or statistical methods. Or even some besides all this will find themselves having to make probability estimations that have, shall we say, a somewhat indirect relationship to any evidence they have. This means that prior probabilities in some form or another will play some role in their inquiry. This could be in either the first or second sense of method. Plenty of argumentation out there that this must or does typically involve some non-neutral resources being given a role as cognitively authoritative! See for instance here or here for discussion of these sort of ideas. That such ethically laden priors sometimes do in fact feature in inquiry in the first sense of method is not all that controversial, but whether they ought, and relatedly whether they may permissibly feature in the second sense of method, is very controversial.

• Of course there is a rather obvious sense in which non-neutral things should be treated as cognitively authoritative, which is the sense they would be authoritative for one if you thought they were truths which other truths must therefore be consistent with, perhaps especially truths discoverable in the course of the very inquiry whose method is under discussion. This can relate to the second bullet point, in that one might think that inquiry must inevitably use or (if successful) come to discover truths that can only be stated using ethically laden "thick" terms. This is to say that it would be missing something important and noteworthy if you gave an ethically neutral description of some persons, rather than describing them as courageous or oppressed or wicked or the like. A pet scholarly hobby horse of mine is that Du Bois almost certainly thought something of this form, and one can see discussion of views along these lines here and here. 

• A person's social standing and political beliefs (or preferences or affective relationships to their situation quite generally) might make them more or less reliable as a witness or provider or testimony. This is related to the point I discussed at length here so I shan't say more on it here. Taking this on board may affect who one wants to be involved in the process of inquiry, as discussed here. Related to that last, it is worth noting that the community of inquiry (since it is a social community) may have its own distinctive set of social values and internal injustices and these may make a difference to how trustworthy or reliable they are - for some related discussion to this see here, here, and here.

Let us assume that our hypothetical analytical feminist philosopher accepts all of the above uses of ethically non-neutral resources as cognitively authoritative. That is to say, they think that the concepts they use are subject to proper ethico-politico debate, that when they have enough evidence depends on what is ethically at stake, that what they may infer from what has already been granted depends on ethico-political decisions, that their prior probabilities or starting place in inquiry may and should reflect some ethico-political stances, that some of the truths they may discover or already accept and therefore take as background knowledge are moral or normatively laden truths, and that the values and participants of their inquiring community are both ethically and epistemically significant. How does such an analytical feminist compare to the ideal of the Radical Philosophy paper?

From what I gather they'd go further than the paper grants the analytic feminist can, but still not quite far enough. The paper spends some time arguing in abstract that one's method in philosophy should be responsive to ethically non-neutral reasons treated as cognitively authoritative. From where I stand this was less controversial in the relevant community than this allocation of space would suggest, but they cite some people to the contrary from within analytic philosophy and it is in any case good to rehearse argumentation just for the sakes of it being explicit. But when it comes to what it would actually look like to take on board ethically non-neutral reasons and treat them as cognitively authoritative the following ideas were put forward.

First:

if we are to do justice to the real or objective texture of women’s lives, we need to explore and, where appropriate, take on board ethically-loaded perspectives

Context, I think, suggested the following readings for this. We should be unafraid of using ethically loaded thick terms, recognising their full emotive power as important for conveying what is going on here. (There is another potential reading, discussed below.) I think our hypothetical analytical feminist philosopher would already have agreed to both of these points, however. That there are moral truths to be discovered and incorporate into our background knowledge, and that thick terms may be required to adequately convey what is going on, are ideas that featured in my hasty literature survey.

Second:

the methodologically radical conviction that investigating charged perspectives opened up by forms of bias to which women are subjected is essential to efforts to get objective aspects of women’s lives clearly into focus. 

This seems to be the familiar idea that we should be aware that in deciding who to take more seriously than others (whose perspective to adopt as a sound basis for inquiry) we should think that a person's identity and political stance matter.  This also featured in the above literature survey.

Third:

if we are to pursue [liberating and rationally sound social thought] in a rationally responsible manner, we need to manifest a sensitivity to the indefinitely complex ways in which, at concrete historical times and places, different and interweaving forms of bias expose members of particular social groups to harm. 

Context suggests this was an elaboration of the first and second points, with an emphasis on the importance of contextual factors in guiding one's decisions about what the salient moral truths are or who is epistemically advantaged.

Fourth,

notice that, insofar as [Fricker] depicts us as relying in most of our discursive dealings on ‘stereotypes’, understood as empirical generalisations about behaviour whose evaluation is an ethically neutral affair – and insofar as the kind of testimonial virtue she urges is a matter of ‘neutralising [stereotypical] prejudice in one’s credibility judgments’ – she operates in the logical realm determined by a neutral conception of reason.

One of the most interesting sections of the paper is an extended critical discussion of Fricker's work, with which I was in considerable sympathy! But note that Fricker is, on this reading (not knowing enough, I will not contest it as a piece of Fricker-exegesis), not making full use of the resources of analytical feminism. It is very clearly not true that if one says that something comes down to evaluating empirical generalisations one has therefore committed to ethical neutrality - the widely accepted inductive risk argument is already enough to block that inference, but so would nearly any of the other bullet points. Second, the point about credibility judgements may relate either to the point about interlocutor selection (or community of inquiry construction) or the role of priors or prior-esque probability judgements in one's inquiry. In either case, again, analytical feminism has already developed ways of understanding the political implications of these.

Fifth,

[philosophers like Fricker who assign an important role to argumentation] are frequently helping themselves to a ‘neutral’ understanding of an argument as a proposition or set of propositions which licenses a further concluding proposition in a manner that does not depend on any tendency of the initial propositions to shape our routes of feeling. 

This was a point at which I do think there is some genuine disagreement. When I first read this piece I did not notice the last clause of this claim. And so it seemed to me to be similar to the above - there is clearly in analytical feminism room for the idea that what makes for a good argument, inductive or deductive, depends in some way upon one's political values. However, I am not aware of any defence of the idea in analytical feminism that part of what makes for a good argument qua argument is how it directs the feelings of those who make or consider the argument. Now, the quoted passage (and wider context) could be the claim that the initial premises should involve ethically laden terms which will have peculiar emotive content. In which case the analytical feminist could insist that after all they want sound argumentation, and some of the true premises one ought work from involve moral truths or must be stated using thick ethical terms.  But as we shall see, I do not think this a likely reading of the paper.

My notes ended here, with the word "Sixth", a comma, and no indication of what I thought followed. Or any means of following up on the various "Later I shall"s above. Sorry folks! I'll only note that I think that this last could be used in a more or less radical way. If the claim is that a good method ought to be one which (in addition to being methodologically good in various other ways) directs us to conclusions that appropriately engage our emotions and thus cause us to be more passionately committed to a claim (method in the first sense of the opening paragraph) then while I am not aware of this being much discussed it seems well in keeping with much of what is out there. One of these days I shall write more properly on Du Bois on art and I shall discuss similar themes then! If, on the other hand, the claim is that part of the justificatory power of an argument consists in it appropriately so engaging our feelings - and I think past me was persuaded this was the proper reading of the paper - then that it is a more significant divergence from typical argumentative canons.