Blain’s summary is accurate and helpful, and he raises several good questions. I take those up before turning to questions by the other participants.
Blain’s first worry is whether, in the absence of an account that would give us clear boundaries for the reasonable or the qualified, we are unable to go forward with this kind of approach at all. Is any way of going forward bound to be unacceptably ad hoc? I try to indicate an alternative to this defeatist position. I mention this methodological stance briefly in several places, including pp. 63-64, p. 217, and p. 286, note 3. Notice that the wish for clear boundaries is no particular support for Blain’s second concern, that the principle of acceptability should be defended by resting it on a deeper principle of respect. It might seem as though having such a deeper account would also give us the boundaries we want. But I see no general reason to think it would. Suppose the duty not to lie is based on the categorical imperative. That’s very little help in in knowing what the exceptions are to this duty, or whether (as Kant thought) there are none. Depth and specificity are quite independent features of moral theories.
I’ll turn in a moment to the question whether to rest the acceptability requirment on a principle of respect. First, let me just briefly rehearse my non-defeatist methodological response to the lack of a clear account of the boundaries of the qualified. The argument of the book needs to appeal, in a few places, to claims that some point of view either is or is not qualified. Here are some of the main ones: No invidious comparison with respect to moral wisdom is beyond qualified rejection (I actually consider slightly modifying this below). It is not disqualified to deny that a democratic arrangement would do epistemically better if the educated were given more votes. The epistemically best political arrangement, so far as the case can be made beyond qualified objection, would be recognizably democratic. I’m sure there are others, but not very many. You can look at those claims in either of two ways. One is to see them as assertions by me, backed up with little or no argument. As I remarked in an earlier reply, no theory can do without undefended assertions, so that’s no vice in itself. Call them premises. If you agree with me about those claims (which is different from thinking either of us has any account to support them) then that is all I need for the rest of my argument. It would be nice, of course, to also have the general account that draws the whole “curve” on which these can be seen to be a few data points. But that’s not a question I need to take up unless it would help me to persuade you of the several claims themselves.
But as with any undefended premises, rather than seeing them as assertions by the argument’s author, they can often be usefully seen more as parts of a conditional argument: if these premises were true then these conclusions would follow. That is part of what I am doing as well, for the following reason. I doubt that most of us have precise convictions about where the line lies between a qualified and a disqualified point of view in this context. It is useful, then, to reflect on what benefits there would be to drawing the line in a place that makes my several claims true. If we can get a good theory from drawing the line in a certain place, this might be one good reason for doing so. We often learn what premises to accept by seeing what conclusions they would support. The order of discovery can go in the opposite direction from the order of justification.
To charge those several claims with being ad hoc, then, turns out not to be obviously damaging. If those claims of mine are surprisingly convenient, it would also be a mistake to ignore just how convenient they are. I assume that the correct theory of the qualified will fit conveniently into the correct theory of democratic legitimacy and authority. This kind of convenience wouldn’t have much weight against substantial objections, but convenience itself isn’t a substantial objection. When a physicist cites the law that explains some data, she shouldn’t be too troubled if we say, mustering what irony we can, “Well, isn’t that convenient.” Nor should we be troubled if it turns out that her basis for accepting the law is, in large measure, precisely that it fits the relevant data.
Blain’s second concern is to express the suspicion that any argument for the qualified acceptability requirement could itself justify democracy. For example, don’t we need to give some deeper principle of respect? But if so, wouldn’t that principle justify democracy all by itself?
As we’ve talked about before (in the discussion of Chapter 3), I advocate a general acceptability requirement, but I decline either to offer a deeper moral basis for the principle, or (as I’ve just been discussing) to try to specify in a general way which points of view count as qualified and which do not. I also argue that the principle cannot be regarded as merely a part of a Rawlsian overlapping consensus. Since it is the principle that establishes the moral relevance of an overlapping consensus in the first place, it would have to stand on independent ground. The principle, as I formulate it, says that political justification must take place in terms that are acceptable to all qualified points of view. Partly because it is specifically about politics, it is not likely to be a fundamental moral principle, one for which no deeper justification exists. So I grant that there must be some deeper moral basis for it. But if, as I also argue, this general acceptability requirement is itself acceptable to all qualified points of view, political justification can proceed without inquiring into any deeper basis. Not only that, but the principle’s deeper moral basis might very well go beyond what all qualified points of view can accept. In that case, the deeper moral basis would not itself be available in political justification.
Blain, following Larmore and Hill, proposes going one step deeper, to a “principle of respect for persons.” Let’s suppose, at first, that such a principle would not itself be controversial among qualified points of view. Why, then, shouldn’t I take that step? First, I have no objection to calling the general acceptability requirement a certain kind of principle of respect for persons. It clearly is. But calling it a principle of respect doesn’t yet appeal to some different principle—the principle of respect for persons—upon which the acceptability requirement rests. It is useful to see that it is a kind of respect, but the question raised by Blain is whether there is something to be gained by going further and positing a principle of respect that is more basic. My answer has the three parts I have touched upon: 1. Yes, there is very probably some deeper moral basis, but 2. Since the acceptability requirement itself is acceptable to all qualified points of view, political justification can rely on that principle without needing to appeal to the deeper basis, and, 3. The deeper moral basis might very well be controversial among qualified points of view, which would, according to the general acceptability requirement, be unavailable. Maybe not, but why risk it if there’s no need, in political justification, for it? None of this is to say, of course, that philosophy must not look into the deeper basis. Philosophy can and should inquire into that, and into many other things, such as whether there is a god, whether man is essentially a political animal, whether morality derives from our rational natures or are more independent of us than that. But these are deep and controversial matters that political justification doesn’t need and normally can’t address without going outside public reason.
Blain raises the interesting question whether the moral basis for the acceptability requirement, whatever it is, might provide a more direct justification for democracy than I’ve offered. If there is a fundamental principle of morality then I suppose that every moral principle’s ultimate basis would provide a moral basis for every other moral principle. Maybe the structure isn’t that simple. In any case, the acceptability requirement itself, whatever its deeper basis, says that political justification—which would include the accounts of democratic authority and legitimacy—cannot appeal to just whatever the moral truth might be (it might be religious, it might not, etc.). So even if there is a basis for the acceptability requirement that also provides a basis for democracy, that would not give us what we need according to the principle itself unless that deeper basis is itself acceptable to all qualified points of view. It would be troubling if there were no basis for democracy that did not appeal to such controversial matters, but the argument of my book is that there is.
I turn to comments by the other participants:
Ben, I don’t think I’ll venture an opinion on the kind of weighted voting in the EU case you describe, except to say that if it literally makes it formally impossible for Luxembourg to ever have a decisive effect then arguably it is a fiction to say they have a vote at all. Beyond that, my argument doesn’t have any implications for weighting schemes like that. As for what Plato would say about giving the uneducated equal votes: the fact that he would say it is neither here nor there. The question is whether his epistocratic idea is beyond qualified objection. Or: do you agree with me that objections to the epistemic claims of that approach should not be regarded as disqualified? (That’s not the question of whether someone might disagree with me. Of course many people will.)
Jonathan’s question about latent and conjectural objections requires a longer answer. He wonders why, if latent and conjectural objections are qualified as objections to epistocracy of the educated, they wouldn’t also be qualified as objections to the allegedly above-random accuracy of any democratic arrangement. The challenge wouldn’t be compelling in such an abstract form. The charge I credit against epistocracy of the educated concerned selection effects, and so we should frame the challenge in those terms. If all are allowed to vote, there isn’t a selection effect of the same kind (as Ben points out). But as Ben also notices, there are selection effects in who chooses to vote. I think Jonathan would be right to say that this must raise the same sorts of latent or conjectural objections.
There are a few different ways one might try to fix this. If mandatory voting would actually lead to universal voting, that would be one possibility. But it would probably just produce a new selection effect about who complied and who didn’t. An alternative approach might be this: choose a random sample of the voting population where that sample happens to fall entirely within the set of people who have actually voted. (See who votes. Let a computer choose random sets of an appropriate size until it finds a random set that includes only people who voted, and only count those votes.) We might suppose that now there is no selection effect. But this strikes me as shaky. It would be a little like noticing that the random sample happened to include only white Protestant lawyers, and concluding nevertheless that since it was randomly produced there is no selection effect. Bad argument. If we were initially concerned that voters would self-select in a demographically troubling way, then we should get no comfort from having a random procedure choose people within that potentially self-selected set.
So, I’m inclined, tentatively, to prefer the following answer. Whether through legally required voting or by some other means, turnout must be high enough that any selection effect must be very small. There is no a priori way to set that level of turnout, but I assume it must be much higher than is normally seen in large elections. It would probably have to include a great majority of eligible voters, but one would want to decide in an actual case, with all its contingent circumstances, whether a given level of turnout allows a reasonable or qualified worry about an epistemically damaging selection effect. But, for example, if turnout were 99%, and there were no evident correlation with obvious factors such as race, sex, class, or religion, the danger of selection effect might be too small to be reasonably pressed. Perhaps turnout needn’t even be that high.
A similar move could, of course, be made in the case of extra votes for the educated, however. Suppose that 99% of eligible voters met the education criterion, and there were no evident correlations with troubling factors such as race, sex, etc. Along with other circumstances, this might disqualify any worry about an epistemically destructive selection effect. In that case, this approach that I am tentatively taking would have no objection to giving more votes to the educated. This would involve modifying my view that no invidious comparison would be generally acceptable. But it is hardly rule by an educated elite. And it has no application in conditions where the relevant education is not quite equally available to all, and so could probably not be used to support plural voting in any presently existing society.
Two further thoughts: First, since this assumes that education is more or less available to all (otherwise the demographic problem would be obvious) I believe such a system would count as recognizably democratic. By analogy, in paradigmatic democracies voters must often register. Unregistered voters are, in a sense, disenfranchised. But so long as registration is sufficiently equally available, (and supposing, contrary to actual fact, that there aren’t other grounds for worrying about a big selection effect) this doesn’t by itself render the procedure undemocratic. Giving fewer votes to a few who don’t meet certain generally acceptable educational criteria, but who could have done so without special hardship, looks to me rather similar.
Second, the epistemic gain of such a practice might be so small that there would be sufficient non-epistemic reasons (such as the stability that might be damaged by such an invidious comparison) in favor of opting for equal universal suffrage. This might move away from the epistemically very best method (within public reason) but the criterion I favor requires only that it be the best, or nearly the best, because the very legitimacy and authority of the system could not plausibly hang on very small epistemic differences between arrangements.