Response to Comments on Chapter 6

Daniel gives a very nice summary of the chapter and raises a few questions. First, he wonders whether the fact that democratic procedures might generate trust, and so compliance, is an epistemic or non-epistemic reason in its favor. Before taking that up, however, I want to quibble with his description of the options. He says that I’m mainly distinguishing between values intrinsic to a collective decision procedure on one hand, and consequentialist considerations on the other. That might not be the best way to put it, since while my theory brings in non-procedural values, it is not consequentialist. (I should say, I’m using “consequentialist,” to contrast with “deontological,” not in the broader way that some use it so that virtually any normative theory could receive a consequentialist formulation. On the latter use, there’s no interesting question about whether there is a consequentialist theory of democracy, since there would be no contrast class.) On my view, democracy is not recommended on the basis of its maximizing good consequences. It is recommended because (and when) the laws that are passed are legitimate and authoritative. They are legitimate and authoritative owing, in certain ways, to the decision procedure’s having some tendency to produce just or correct outcomes. That part is instrumental in a certain sense, but not necessarily consequentialist. The accounts of authority, legitimacy, and justice could all be deontological for all I’ve said. I’m not suggesting that Daniel misunderstood this, but my view is always at risk of being misunderstood to be consequentialist, and so I’m at pains to use that term advisedly. (I address the issue at pp. 164-167, in Chapter IX.)

So Daniel notes my central distinction between theories that limit themselves to procedural values and theories that include some appeal to the substantive quality of the decisions that are made. Then his example—that democratic procedures might engender trust and compliance—is said to find no easy home in my framework. As Ben notes, this is a similar question to the one he and others have pressed about whether democracy’s value might rest in its educative effects on the citizenry. This also requires no appeal to a tendency to make produce good collective decisions. Bringing these two questions together, the general question is whether democracy’s value might rest in good consequences other than correct decisions.

Let’s look at my favorite analogy, a jury trial, and see how a similar question looks there. I suggested that the value of a jury trial is inseparable from its having a decent tendency to make the correct decision. The correct decision, I assume, would be to convince the guilty or acquit the innocent. (We don’t simply want maximal accuracy, of course, since we want to err on the side of acquittal, but that doesn’t falsify the previous sentence.) Now suppose an alternative theory were proposed. On this theory, what matters about jury trials is not whether it gets the correct answer, but whether having that sort of system educates the jurors or engenders trust in the citizenry. We would, of course, have to admit that education and trust are good things. But without any epistemic component, a trial system would be barbaric. So even though a system might educate or engender trust, those probably wouldn’t be enough to make it legitimate or authoritative.

What we learn from this analogy is that my view, epistemic proceduralism, has nothing to fear from an account which shows that democratic procedures educate people or engender trust or compliance. But an account that proposes that those consequences of democracy are an adequate basis for its authority and legitimacy is a different matter. Why should we think that bracketing the question of the correctness of decisions is barbaric in criminal trial, but not in political decisions? It is this rhetorical question that guides my instincts about a consequentialist approach to democracy, but of course we’d have to see specific proposals in order to decide whether they can survive this challenge. I conjecture that a theory that appealed to good consequences without any appeal to correct decisions would turn out to be inadequate. And I argue, in effect, that epistemic proceduralism can suffice whether or not there are ancillary benefits such as trust and education.

Daniel’s point seems to have been only that consequentialism is an alternative approach about which I say almost nothing, and that is true. I laboriously critique theories that try to do with procedural values alone, and I claim that no adequate theory could do without an epistemic component. But I don’t show that a consequentialist approach would need such a component. I am inclined to think that it would, but this depends a lot on what counts as a good consequence. On the educative approach, I don’t see how democratic participation would be edifying to a person who thought it had no particular tendency to lead to good decisions (a point we discussed in a previous thread). As for Daniel’s trust hypothesis, we’d need to hear what accounts for the trust engendered by democratic procedures. If, as I suspect, the basis would be that they are more likely to produce substantively better decisions, then the epistemic element is required (assuming publicity, as I think we must if a democratic arrangement is being proposed), just as I claim. If there’s some other basis for the trust, we would have to hear more about it in order to evaluate that approach.

Turning to Daniel’s second concern, he writes, “Part of our evaluation of policies has to do not with its approximation of some independent standard of justice, but with the way it has been reached, and with the democratic processes that go into its implementation and administration. We tend to evaluate not policies, but procedures-plus-policies, taken as wholes.” Epistemic proceduralism clearly evaluates policies partly on the basis of the procedures they derive from: their legitimacy is tied to their coming from procedures with a certain epistemic value, roughly. Daniel adds that often a policy is evaluated partly by “the democratic processes that go into its implementation and administration.” That sounds right, in some sense. Whether a policy is a good one has much to do with how it will probably be administered and implemented. But I’m not sure how this challenges their “detachability” from the procedure in which the policy was decided upon (with due allowance for the way legitimacy depends on procedure in my view). In the procedure voters ask themselves, and think about, which policy they should collectively adopt, a question which presents itself as having correct and incorrect answers. Otherwise, what’s to think about? Certainly the right answer will be determined partly by facts about how certain institutions must operate and so on.

Turning to a third point, Daniel writes, “When a jury delivers a verdict… our acceptance of it stems from the fact that it has been constrained in…various ways. In other words, moral allegiance attaches to a process of which a jury’s deliberation is a part, rather than just to the jury’s work, narrowly understood.” Isn’t democracy probably similar? I would say yes. I think this is perfectly consistent with my view, even if I don’t say much about what the constraints are. (I discuss the question briefly at 165-66.) If liberal democracy, in the sense of majority institutions constrained in numerous ways, is the form that meets the criteria I propose, then my thesis that only some form of democracy meets them is sustained.

Ben writes, “Democratically-deciding-to-X may be better than doing Y, even though democratically-deciding-to-Y would have been even better.” Notice that I can accept this without strain. Epistocrats might know the right policies. It’s central to my view that the best policy is not legitimate if not arrived at through the right process, which epistocracy is not, and some form of democracy is. Now if you say, “but in that sense they are not the correct policy,” I would just insist on a distinction between the procedure-independent kind of “substantive” correctness or justice, and legitimacy, which is procedural. I don’t mind if we call those two dimensions of correctness, although I think it risks conflating two very different things.

Ben also writes, “I’m not sure why we should refrain from over-ruling the procedure when we think it is clearly and/or very wrong. What is supposed to be the mindset of the out-voted? Is it ‘I still believe I’m right, but may be wrong’ or ‘I still believe I’m right, but this outcome is now legitimate’?” I argue that the democratic procedure may or may not give me any reason to think I was wrong, but in any case the decision is legitimate even if “substantively” incorrect. The reason not to just ignore laws I think are wrong stems from the moral account of authority and legitimacy I offer. It’s similar in form to the reason the jailer should not ignore the trial (usually). That “usually” points in the direction of justified disobedience, something Ben asks about in passing. I don’t offer a general account of it, but I assume (and say at 111-112) that some laws are so unjust or unjust in such a way that obedience is not required and would sometimes even be wrong. And later, in Chapter X, I present an account of the epistemic value of some sharp and even subversive political practice, including but not limited to various forms of disobedience. I’ll look forward to revisiting those questions in a few weeks.

Be Sociable, Share!
This entry was posted in Posts, Reading Group. Bookmark the permalink.

4 Responses to Response to Comments on Chapter 6

  1. Loren King says:

    David: “In the procedure voters ask themselves, and think about, which policy they should collectively adopt, a question which presents itself as having correct and incorrect answers. Otherwise, what’s to think about?”

    There’s a (still murky) concern that’s been nagging at me (murky in my head, I hasten to add, not in David’s wonderfully rich book).

    The view of democracy central to the book is captured nicely in the quoted passage above: we ask ourselves about a law or policy (with an eye to “the actual collective authorization of laws and policies by the people subject to them,” p. 38), and we generally think there are better and worse answers to the question at hand. We don’t need to appeal to specific (and controversial) claims about the substance of “better” and “worse,” yet we can sustain an argument that, whatever the truth is, “some democratic arrangements perform better” at tracking it “than a random choice from the alternatives presented,” and they do so without being “too much worse than any nondemocratic arrangement” and in ways that fall “within public reason” (understood in terms of acceptability to “qualified points of view,” p. 168).

    But what’s nagging at me (and motivating me to finally jump in here, rather than earlier or, more sensibly, later) is David’s “Otherwise, what’s to think about?” question, posed rhetorically to Daniel.

    I can think of two, perhaps three, importantly distinct categories of questions that come up in modern democracies.

    First, there are questions of fact about policies for which preferences are widely shared or at least broadly compatible, and where most of us do (or would) broadly agree on the standards of evidence that would decide the matter (or at least, such standards are not themselves the source of deep and protracted disagreement). Most citizens would prefer the more efficient tax regime and responsive service provision scheme, other things being equal. We disagree over how best to accomplish these shared ends. If a flat tax, say, really would promote investment and lead to a point on the Pareto frontier that everyone prefers, then that’s the truth of the matter.

    David’s recurring jury analogy seems most powerful here: the legitimacy of a jury’s judgement — even if likely wrong — surely has something to do with the modest but real epistemic virtue of the procedure. Similarly, it would be a strange account of democratic legitimacy that ignored the question of whether or not democratic procedures do a better-than-random job of lurching us toward the correct answer on questions where we all want the same thing, but disagree about the most efficient and effective way to get the goods.

    But is epistemic virtue doing all the work here? For the jury case, I find it hard to ignore the fact that the participants in the procedure share a sense of what, broadly speaking, the nature of truth is in this case (guilt or innocence), and they (ideally) share a motivation to find the specific truth in this case (“this defendant is guilty” or “this defendant is innocent”).

    With that thought in mind, consider a second category of questions that often arise in democracies: these are typically moral questions, for which many parties believe they have (or would recognize) facts of the matter, but there is no convergence or complementarity of preferences. Better, think of this category as making a cut on a spectrum of moral questions, some scarcely different from first-category questions (“slavery is wrong”) but becoming more and more contentious as we move further along the spectrum away from the cut. And so while we all might say we want justice, we disagree on what that means, and as we move deeper into this category, there is less and less scope for agreement on what would count as decisive epistemic standards in determining the truth of these matters.

    Now, David’s approach so far has still shined for many of these sorts of questions, I believe: we can acknowledge the epistemic value of certain collective decision procedures without having to settle the question of what, precisely, the truths of justice and morality actually are.

    But here the jury analogy seems to me to fit less tightly. Would we still affirm the legitimacy of a jury’s judgement if they sincerely and explicitly disagreed not only on the quality, relevance, and relative weight of the evidence as it bears on the question of guilt, but also on what it means to be guilty and innocent in this case, or in any case?

    I would think we’d see that as a failure of the judge in an actual trial: the instructions to the jury ought to make it clear what standards and thresholds the law sets for guilt and innocence in the case at hand. Yet in a democracy we don’t generally consider such disputes to be failures of question-framing. Or at least, the argument that all or most questions in the second category are poorly-framed instances of questions in the first category violates the spirit of what I’ve read so far in the book, because it would seem to involve some serious epistemic position-taking (perhaps something in the spirit of Raz’s position that David decisively rejects at the end of chapter three?).

    If we admit that the questions in this second category are real (and not sloppy versions of first-category questions), then I wonder if David’s rejoinder to Daniel’s point about trust-building and educative effects is a bit too hasty? For second-category questions, we might take it to be a (legitimating) virtue of democratic procedures if they make citizens more likely to take such disputes seriously, to recognize the scope of the qualification requirements David lays out in chapter three, and to search sincerely for mutually satisfactory accommodations if such are available.

    To be sure, we can still call these effects epistemic virtues of a sort, if we think that the correct answer to second-category questions is likely to involve standards of mutually acceptable accommodation buttressed by widely acceptable justifications to bearers of qualified points of view. But note that here (and unlike the first category) the truth in question is certainly not linked to what David calls “a substantive epistemic account.” Nor, however, is it tidily linked to “formal epistemic value,” because what’s at stake isn’t just “a tendency to get things right from the standpoint of justice or common good whatever the best conception of those might be” (p. 169), but such a tendency for questions of accommodation, not questions of substantive truth.

    And if that’s the case, then I think what’s doing a lot of the legitimating work for democratic procedures applied within this second category is David’s account of qualified acceptance, not epistemic virtue per se. And that shift in the justificatory burden seems to leave some room for Daniel’s point about trust and educative effects as importantly linked to legitimacy in a way that isn’t merely parasitic upon, or secondary to, epistemic properties.

    Furthermore, if we think second-category questions are real, I wonder about a possibly distinct third cluster of questions lurking at the far end of the second category: questions for which the lack of anything like shared epistemic standards for determining truth are especially pronounced. Aesthetic questions come to mind, which might seem out of the scope of democratic decision-making, until we consider the abiding importance in public life (and especially in major cities) of architecture, grand monuments, and funding for the arts and related centers.

    The temptation is to dismiss such questions as too often originating in disqualified views (thus, for instance, David’s passing example of rural voters perhaps being “irrationally fond of open spaces,” p. 221). I don’t want to get too far ahead of our discussion schedule, but my point here is not that David is necessarily wrong about rural voters or disqualified views, but that this seems to be where the legitimating action is for these sorts of questions — that is, with an account of qualified acceptance, and not primarily with the modest epistemic virtue of democratic procedures.

  2. Loren: I can’t hit everything you raise, but here are a few thoughts:

    The main disanalogy that you mention between the jury case and democracy is that in the jury case the jurors share more criteria for what counts as a correct answer. Still, in the democracy case it looks as if you are granting that people accept that there are correct answers, even if the criteria are not as widely shared. Each voter seems to think there are correct answers since each has an opinion about what it is. (I grant, of course, that in many cases there is a range of equally correct answers, but that doesn’t affect the point.) But if we must say that there are correct answers, why should we set the standard for the collective decision lower than that, settling only for “accommodation”? There might be contexts where the correct answer is an accommodation of a certain kind. That doesn’t make it any less of a correct answer. In some cases the correct answer is not an accommodation of any kind, but we must settle for the least-incorrect answer for pragmatic reasons. None of this seems to displace the idea that there are correct answers to many political questions.

    You’re right, of course, that the answers to many political questions depend on factual matters. But I’m not sure I see how that would answer my rhetorical question, “if there aren’t correct answers, what’s to think about?” The only point of thinking about the facts is because one thinks they bear on the correct answer. Should we build a bridge? The facts matter: How much will it cost? How would it affect traffic? But answers to those questions are not answers to the question of what to do. That’s a further question that one thinks about partly by thinking about these factual questions.

    On your question whether we would still think a jury’s verdict legitimate if jurors shared hardly any criteria for a right answer: I agree the answer is probably “no.” But I think my account can say this without inconsistency. We should not use juries in that case because (or at least if) there is some other procedure which can be seen (beyond qualified disagreement) to be likely to perform better. For example, we could probably do better with a single judge. In the democracy case, despite fewer shared criteria, it may yet be the best we can do, so far as can be established beyond qualified disagreement. This is why I think we demand higher accuracy of juries that we should of democracy.

    As for the questions that are more like aesthetics, we had an exchange about that in week 1. You might look back at what I say about a choice whether to put a playground or a sculpture in the town square (in my response to comments for that week).

  3. Loren King says:

    David: “On your question whether we would still think a jury’s verdict legitimate if jurors shared hardly any criteria for a right answer: I agree the answer is probably “no.” But I think my account can say this without inconsistency.”

    To be clear, I wasn’t charging inconsistency. I think the argument hangs together very nicely indeed. But I am wondering about the legitimating work being done by epistemic properties of a choice procedure as beliefs and preferences shift from ‘shared ends, contested means’ to ‘how could so many of my fellow citizens possibly be more wrong about moral truth and how to evaluate the correctness of laws and policies?’

    I’m imagining, say, a territorially dispersed minority population (pansexual polyamorous pacifist vegans, perhaps) who have some unconventional but sincerely held beliefs about moral truths and correct policies, and they have cogent arguments in support of these beliefs that meet the burdens of qualified acceptance.

    Under most plausible understandings of democratic procedures, this minority will likely never see their moral truths and correct policies reflected in public debate, let alone democratic decisions and, eventually, laws and policies (I don’t think this is entirely true: I can envision some unconventional federal democratic arrangements and novel approaches to representation a la Andrew Rehfeld that could grant considerable room for this minority to see its favoured views reflected in law and policy, but their pacifism and some of their other beliefs might still be impossible to accommodate).

    True, these citizens may readily acknowledge that democracy is sufficiently likely to avoid “primary bads” over time, compared to any conceivable nondemocratic alternative (although for some of these folks, the tyranny of militaristic, meat-based, paternalistic-monogamist, heteronormative culture may rank as pretty bad — perhaps these are cases for civil disobedience?).

    But there is an important space of (non-disqualified) truths, as understood by these citizens, that democracy is clearly not roughly tracking over time in a better-than-random way. This isn’t the failure of, say, a few laws to satisfy this group; it’s a long course of decisions over which their distinctive yet qualified views remain utterly ignored.

    The problem (if there is a problem) isn’t a failure of epistemic proceduralism per se in such cases, but there’s potential for some weirdness. Again, in the case of “primary bads” there is no reason for minority members to dispute the modest but real epistemic credentials of the democratic procedure. Yet over time, outcomes seem to pull this group in the opposite direction for just those beliefs about truth and correctness that distinguish them as a group. On that issue subspace (over time and given some reasonable expectations about the likelihood that majority opinions will not be wildly swayed by minority efforts), there is little to suggest “the procedure’s generally acceptable tendency to make substantively correct decisions.”

    I think your conjecture offered to Daniel — “a theory that appealed to good consequences without any appeal to correct decisions would turn out to be inadequate” — is still secure in the face of this hypothetical case, but I am more hesitant about the claim “that epistemic proceduralism can suffice whether or not there are ancillary benefits such as trust and education.”

    In a modest sense, this second conjecture remains secure: yet again, the pansexual polyamorous pacifist vegans may be moved by the epistemic virtue of democratic features in tending to avoid “primary bads.” But the claim that democracy is sufficiently likely to track their favoured accounts of moral truth and correct policy seems less clear as a ground for legitimacy. At the very least there’s an asymmetry to be considered here: the majority can count on democratic procedures not only to tend away from “primary bads” with sufficient reliability, but also to track certain goods. For members of my hypothetical minority, only the first expectation is reasonable.

    They (and we) might well be moved, however, by an argument that their continued participation in the public life and political processes of their democracy enriches public debate within the terms of qualified acceptance, engenders tolerance across (qualified) differences, and may transform all parties (vegans and monogamists alike) in unobjectionable, even attractive ways (attractive, that is, given our commitment to a standard of qualified acceptance: they might broaden their sense of, and desire to understand, views that fall within the scope of that standard).

    If that argument persuades, then I wonder if qualified acceptance is doing as much (perhaps much more) legitimating work than epistemic performance of democratic procedures, in such cases?

    I suppose that federalist solutions to any such quandaries (ie. of qualified but really unpopular minority positions) can still be thought of as legitimate insofar as the solutions refine democratic procedures to better track beliefs about truth and correctness for these qualified minorities (supposing still, of course, that such qualified minorities could ever exist – seems to me that they could, but I guess that’s arguable). So the fact that we can imagine democratic arrangements that are more responsive to a range of qualified claims about truth and correctness supports epistemic proceduralism.

    But then I wonder if epistemic proceduralism would commit us to specific institutional designs that made democratic procedures more likely to track the truth with sufficient likelihood for distinct qualified groups, compared to less responsive designs?

    Or would an argument that emphasized an account of qualified acceptance and educative/transformative consequences, rather than epistemic performance, be an acceptable rationale for, say, not transforming the US into a very different federal system to accommodate pansexual polyamorous pacifist vegans?

    If so, then it seems there might well be some space to argue for an account of democratic legitimacy (or perhaps an extension of an account) that turns on qualified acceptance and consequences of democratic participation, rather than epistemic properties of the procedure (even granting that political choices in these cases are about what is correct). I’m curious about what epistemic proceduralism might say in situations where a case for legitimacy can be made to members of a minority group who may reasonably doubt the epistemic promise of extant procedures, given their notions of truth and correctness.

    Perhaps it suffices that democracy is sufficiently likely to avoid “primary bads”? But then there’s the answer that alternative institutional designs could do that and more, epistemically speaking, and so epistemic proceduralism may commit us to some rather demanding institutional specificity — specificity that could be avoided by a less-epistemic account of legitimacy that may well suffice to convince the sorts of citizens who (hypothetically) are raising the problem in the first place.

  4. Ben Saunders says:

    I need to read Loren’s response when I have more time, but a quick thought on David’s aside “I grant, of course, that in many cases there is a range of equally correct answers, but that doesn’t affect the point”.

    It may not affect the point here, but I think it is important and gives more place to fairness than David allows. Suppose two of us have to distribute five beans between us. I guess we could rule out (4,1) or (0,5) [either way round]. It seems that (3,2) and (2,3) are both ‘equally correct’, however. (And, of course, the levelling down solution of (2,2) may also be a reasonable one).

    Since it’s important to me whether I get 2 or 3, I think it’s important that the decision process not only ensure one of these correct answers, but gives me a fair chance of getting what I want.

Leave a Reply