PPPS: On Gutmann and Thompson’s Arguments that Deliberative Democrats Shouldn’t be Pure Proceduralists

Simon, thanks kindly for the useful comments. I’ll reply to all but your 3. below, to get the ball rolling. I’ll reply to your 3. in the next few days.

Prior to your 1. you write:
“Dodd then argues that since (BETA) is a reasonable conception of fair procedures, Gutmann and Thompson’s argument against pure proceduralism is unsuccessful.”

This is incorrect. I don’t argue that (BETA) is a reasonable conception of fair procedures. The related argument that I make is that at present we should endorse the following disjunction - (GAMMA): Either (BETA) is reasonable or it’s an open question whether (BETA) is reasonable. This is quite a bit weaker than arguing for (BETA) itself. I point out in the paper that we’d have to go well beyond the arguments (and project) of my paper to make a case for (BETA) itself. I do argue, on one horn of the constructive dillemma that I offer, that if (BETA) is a reasonable conception of fair procedures, then Gutmann and Thompson’s argument against pure proceduralism is unsuccessful. But the constructive dilemma doesn’t require the antecedent of that conditional to be true. The constuctive dillemma gets off the ground if we’re justified in endorsing (GAMMA).

You write:
“1. If we assume that (BETA) does express a reasonable conception of fair democratic procedures that Habermas espouses, then not very much follows about Gutmann and Thompson’s position. (BETA) must be endorsed as true if it is to do any real work. Mere reasonableness does not suffice, since Gutmann and Thompson could be correct that it is false, even if it is a reasonable claim. Thus the claim on p. 10 that Gutmann and Thompson haven’t succeeded unless they can show that (BETA) is unreasonable is too strong. If the right conception of fairness is less robust then (BETA) supposes, then it doesn’t matter if a reasonable person could reasonably endorse (BETA). It would still be the case that pure proceduralist theories would not necessarily satisfy the principle of reciprocity. So I found the argument on pp. 8-10 unpersuasive in this respect.”

This point more or less targets my (10) - the first horn of the dilemma that I offer. I say ‘more or less’ because there’s a notable distinction between talk of ‘Gutmann and Thompson’s position’ (namely that pure proceduralisms can’t seriously endorse the principle of reciprocity) and talk of ‘Gutmann and Thompson’s arguments for their position’. My dilemma targets Gutmann and Thompson’s arguments for their position - and so not the more basic point of the ultimate tenability of their position. I have, e.g., provided no arguments that someone - including Gutmann and Thompson of course - couldn’t chime in and supplement Gutmann and Thompson’s extant case with new arguments. That is, I haven’t argued that Gutmann and Thompon’s position is in principle unsupportable. Rather my case is that their position hasn’t been sufficiently supported.

Even with that said though, there’s still plenty of room to ask whether what you say here suggests a problem for my arguments. I think the answer is that it doesn’t. It’s true, of course, that even if (BETA) is reasonable (as it is on the horn of the dilemma we’re looking at), this falls far short of it being the case that (or assuming that) (BETA) is true. But it’s weighty that (BETA) is reasonable unless we have strong arguments that some competing principle is true or, alternately, far more reasonable than (BETA). I made the point in my paper that Gutmann and Thompson’s case against pure proceduralisms seems to be interested in undermining pure proceduralisms without engaging in debates about what it is for x to be a fair democratic procedure. Your point here may indirectly support my claim that Gutmann and Thompon’s efforts here weren’t successful. The obvious way for Gutmann and Thompson to try to avoid my argument is to back up and argue for a principle that competes with (BETA). But this is both a hard task and not where their arguments go - and so there’s a case to be made that, as I’ve argued, their arguments just go in the wrong direction from the start.

You write:
“2. I don’t think (BETA) can be a criterion of a fair democratic procedure for a pure proceduralist theory. A pure proceduralist theory only invokes substantive principle to provide content to its procedures. A pure proceduralist theory does not presuppose any procedure-independent standard of correct outcomes. It may be the case that any fair democratic procedure will result in mutually justifiable outcomes. But this needs to be argued, and cannot be stipulated as a conception of the procedures themselves. If we could define pure proceduralism in terms of outcomes, then every procedural theory could be redefined as a pure proceduralist theory. For instance, a cake-cutting procedure would be fair if and only if the only outcome that could emerge is one that left each person happy. A trial procedure would be fair if and only if the only convictions that could emerge were of guilty individuals. These are obviously not satisfactory purely procedural criteria of fairness. (BETA) doesn’t look any better to me.”

I take it that you’re trying to raise here an in principle problem about marrying (BETA) with a pure proceduralist theory. The basic idea is that if theory x endorses (BETA), then x can’t be a pure proceduralism - because (BETA) isn’t a pure proceduralism-friendly analysis of what it is to be a fair democratic procedure. I take it that the specific worry is this: as a conceptual point, pure proceduralisms (to be pure proceduralisms) have to define fairness in democratic procedures strictly in terms of the content (or input) of those procedures and, so, not at all in terms of the outcomes (or outputs) of those procedures.

This is an interesting claim. But it’s really only of categorizational significance. If we grant your conceptual point, then (if (BETA) accurately characterizes Habermas’s view) Habermas’s view shouldn’t be categorized as a pure proceduralism. The same would hold of, if I remember my theories correctly, each of the other views that Gutmann and Thompson call a ‘pure proceduralism’ and target in their work. Nothing of philosophical consequence follows from this though. It’s probably true (well, maybe) that Gutmann and Thompson’s arguments would work better if they were read as targeting merely theories that satisfy your suggested constraint. But Gutmann and Thompson don’t target merely theories that satisfy your constraint. Indeed it seems clear that all of the theories that they explicitly mention as targets don’t satisfy your constraint. My paper concerns whether Gutmann and Thompson’s arguments work against the specific theories (especially Habermas’s) that they target - and so against ‘pure proceduralisms’ on Gutmann and Thompson’s (perhaps flawed) used of that term.

I think you’re raising (at least) two issues in the last paragraph of your comment above:

-A- One is whether pure proceduralists can advance an analysis of ‘x is a fair democratic procedure’ that defines fairness in democratic procedures in terms of the outcomes (or outputs) of those procedures.

-B- Another is a challenge for a pure proceduralist to explain for any given procedure that s/he calls fair why it is that that procedure guarantees reciprocal outcomes.

These are both very interesting issues. Here are some thoughts on each:

Re -A-: I don’t see an in principle problem with pure proceduralists defining fairness in democratic procedures in terms of the outputs of those procedures. Pure proceduralisms can take the project of analysing ‘x is a fair democratic procedure’ to flow more or less analogously to the project of analysing, say, ‘x is a top-quality engine’. Suppose we are trying to give an analysis of ‘x is a top-quality engine’. Part of what our engine analysis will appeal to will be things that occur when the engine is turned on and is running (e.g., gas effiicency, noise, power). But part of our engine analysis may well appeal to outcomes of the engine’s having been turned on (e.g., pollution). A top-quality engine will both run efficiently while active and not produce a bunch of pollutive outcomes (…and not produce things that slowly build and contribute to pollutive outcomes down the line). In the case of engines it seems very plausible that the best way to try to resolve bad outcomes may be to try to refine the engine in some way - the idea being that the best method of resolving the bad outcomes mightn’t be to create a second device that chews up the engine’s pollution, but instead to try to trace back the causal line and refine the engine.

In principle there’s no reason why we can’t operate on a similar assumption in democratic theory. We can, just as in the engine case, generate at the level of theory a set of criteria for being a fair democratic procedure, some or all of which are outcomes of procedures. Then if problems in outcomes arise such that the criteria go unsatisfied, the program can be (analogously to trying to refine the engine) to return to the procedures to try to prevent such outcomes from arising. There are pragmatic reasons to go this route of returning to procedures, just as there are pragmatic reasons to return to working on the engine. But we don’t need to get at these to address the mere idea of appealing to outcomes.

As for whether a theory that makes this sort of appeal to outcomes still qualifies as a bona fide pure proceduralism, that’s a largely semantic issue. If we opt for G&T’s analysis of ‘x is a pure proceduralism’, a corrolary of which is that pure proceduralists hold that the only constraint on what policies can be adopted in a political community is that the policies are outcomes of fair democratic procedures, then there’s no in principle problem. On G&T’s analysis of ‘x is a pure proceduralism’ one way we’ll get differentiation between pure proceduralisms is by how they analyse ‘x is a fair democratic procedure’, esp. with seemingly no option being restricted in G&T’s analysis of making some appeal to outcomes when proposing different analyses of ‘x is a fair democratic procedure’. So there’s scope for both ‘restricted-to-inputs’ and ‘involving-outputs’ pure proceduralisms if we operate with G&T’s analysis. ‘Involving-output pure proceduralisms’ can seemingly take the intuitive engine analogy given above a part of their motivation for taking their preferred direction.

Re -B-: I think there are a couple of descriptively similar challenges for pure proceduralists that we might run together. Challenge -B- is this: For a pure proceduralist to explain for any given procedure that s/he calls fair why it is that that procedure guarantees reciprocal outcomes. This challenge actually can be met merely by means of definitions. Any procedure that a pure proceduralist (or at least the kind of Habermasian pure proceduralist I’m most interested in) rightly calls ‘fair’ is guaranteed to be one that only involves reciprocal outcomes because that’s just part of what these pure proceduralists mean when they call something ‘fair’. If it turns out that a procedure that was dubbed ‘fair’ initially instead results in some outcomes down the line that violate reciprocity, then it was just false that the procedure was actually fair. Challenge -B- is analogous to the following: For an engine mechanic to explain why it is that any engine that s/he rightly calls ‘top-quality’ is guaranteed to be one that is efficient and not heinously pollutive. Any engine that an engine mechanic rightly calls ‘top-quality’ is guaranteed to be efficient and not heinously pollutive because that’s just part of what it means for it to be top-quality. If it turns out that the engine is actually heniously pollutive, then it was just false by definition to say that it was a top-quality engine.

A challenge that’s related to -B- is -B*-: For pure proceduralists to explain for any given procedure why it is that we have good (or bad) reason to believe that it is apt to result only in reciprocal outcomes. This is a hard challenge and meeting it will have to be done, of course, by assessing any particular procedure on its own merits in the context in which it’s to be applied. But it’s important that pure proceduralists can fail repeatedly in these sorts of justificatory efforts, just as the mechanic can fail repeatedly in justifying belief that such and such particular engine actually will always be non-heniously pollutive. None of these failures suggest or entail that the pure proceduralist (or the mechanic) should drop her/his standard of what it is to be a fair democratic procedure (or top-quality engine).

…so I still haven’t got to your comment 3. in your original comments. I will! But your last comment raised a bunch of interesting issues that I wanted to look at bit at first.

Simon, at last a response to your 3. - well to your 3(b) since it’s the main point to address:

You write
“Two points are worth mentioning here. First, I was confused by Dodd’s target, i.e. the claim on p. 4:
(6) “Some fair democratic processes can result in mutually unjustifiable policies …”
It struck me that a much more feasible target was the inference on p. 3 from:
(3) “Some fair democratic procedures are reciprocal only if they incorporate Bsubstantive principle”
to
(5) “Only democratic theories that include B-substantive principles can seriously endorse the principle of reciprocity.”

Reading p. 3, I had thought that the strategy would be to concede to Gutmann and Thompson that some fair democratic procedures would have non-reciprocal outcomes, but then argue that other fair democratic procedures would not. These other procedures would then not require B-substantive principles, and hence there would be at least one version of pure proceduralism that escapes their criticism. But Dodd instead wants to argue for the much stronger and much less plausible claim that no fair democratic procedure can result in non-reciprocal outcomes. I don’t see why the weaker and more plausible claim would not suffice as a response to Gutmann and Thompson.”

This is all fair enough. One could try to undermine G&T’s arguments by taking, presumably, any of several weaker paths. But those paths aren’t really open to someone with my background views about how we should analyse ‘x is a fair democratic procedure’. Once one takes on the heft of a Habermasian analysis here, then the most straightforward option is to work arguments around an endorsement of the strong claim that no fair democratic procedure can result in non-reciprocal outcomes. There are big questions, of course, of the ‘Why take on that analysis?’ variety. But the main point to make here is that if one backs that analysis, then the natural target to take on in G&T’s argument is their (6).

“Second, we still have a problem in showing that at least one fair democratic procedurewould not result in non-reciprocal outcomes. Even if we stipulate that permissible reasons in public deliberation must be consistent with a conception of persons as free and equal, it could still be the case that outcomes were non-reciprocal. Participants in deliberation could be mistaken, for instance, about a whole variety of moral and empirical claims in such a way that the outcome of the procedure was not mutually justifiable. All this depends, of course, on what is meant by the horribly nebulous principle of reciprocity and the idea of mutually justifiable outcomes. But if we take some public policies as required by that principle – e.g. healthcare for all – then it is surely possible that any fair democratic procedure defined in purely procedural terms could fail to issue in that policy. At any rate, I don’t see that Dodd has offered us any reason to think that there is at least one fair democratic procedure that can guarantee mutually justifiable outcomes. The stipulation in (BETA) that democratic procedures are fair if and only if they do result in mutually justifiable outcomes is singularly unhelpful.”

There are several interesting points in this paragraph. Some of the ideas link up with comments you made above in the comment thread. So my responses here will be at times be similar to some things I said in my comment above - that is, in my second comment in this thread.

First, pure proceduralists don’t need to be able to prove that there is some procedure that won’t result in non-reciprocal outcomes if it is adopted. Whether any given procedure will or won’t result in non-reciprocal outcomes is straighforwardly an empirical matter, something that we can at best make predictions about and then test. The issue here is analogous to demanding an engine mechanic or designer to prove that some particular engine is a top-quality engine before it’s been on the road for long. At the level of theory (or design) the task is to determine and design what seem to make for the best procedures (or engines). One claim of the pure proceduralist (at least the sort of pure proceduralist I’m most interested in) is that if our predictions do go wrong and the procedures end up resulting in non-reciprocal outcomes, then those procedures didn’t satisfy a necessary condition for being fair democratic procedures. This claim doesn’t require anything about having the ability to make infallible predictions about what procedures will or won’t result in non-reciprocal outcomes.

Second, it’s certainly true that both the principle of reciprocity and the idea of mutally justifiable outcomes need to be made far more precise. These are troubles not just for pure proceduralists, of course. All deliberative democrats should be interested in precisifying these things - and more generally everyone interested in public reason or public justifications. I wanted to bracket aside these issues in the present paper. But it’s very problematic that it’s relatively easy to walk through, e.g., the literature on deliberative democracy and find a thorough-going committment to the principle of reciprocity … without finding thoroughly precise descriptions or analyses of what the content of that principle is.