Begging the Question (2/2)
An argument begs the question, you might say, if it appeals to a premise that your opponent will - by definition as your opponent - reject.
So imagine I put forward an argument A against P. You then object to A, because of Q. Can Q beg the question? No. In fact, if it seems that Q begs the question, then A itself must beg the question. If my argument A relies on the falsity of Q, and Q is true iff P is, then my argument straightforwardly assumes that P is false before it begins.
That's too schematic, so let's have an example. I argue: Physicalism is false, because of this Zombie argument. Ah, you say, but the Zombie argument relies on this false premise: that zombies are possible. I rejoin: Your claim that this premise is false begs the question; those who accept zombie arguments are, by definition, bound to accept this premise because their argument relies on it[1].
I am, contrary to first appearances, in a poor position here. For I was the one who was meant to be arguing against physicalism in the first place. The burden of proof is on me to show that we must not accept such a theory. My argument to this conclusion cannot rely on a claim that is disputed by the people who accept physicalism. But that is precisely what my argument must do if I am able to call your rejoinder question-begging.
More generally, begging the question is when you put forward an argument that fails to move the burden of proof from one side of the dispute to the other. If I argue against abortion on the assumption that abortion is murder, then my argument transparently begs the question - those in favour of abortion are extremely likely to just deny the premise that abortion is murder. What's gone wrong here is that my anti-abortion argument assumes that the burden of proof is on you to tell me why abortion is not murder, and not on me to tell you why it is. And that's precisely what I'm meant to be doing, if I'm arguing against abortion in the first place. If the burden of proof was on you, I wouldn't need to present such an argument at all[2].
It's sometimes said that all arguments beg the question. For anyone denying P will, by definition, reject any argument for P. But this misses the point that an argument is an attempt to shift a burden of proof to your opponent. Arguments only beg the question when the burden of proof sits on the same side of the discussion for the disputed premise as much as for the disputed conclusion.
Think of it like a tennis match. A ball is on one side of the net. If I hit it, then my aim is to shift the ball to your side of the net. If I hit it too weakly, and it fails to cross the net, then my hit fails to advance the game: the ball is with me just as much as it was before my last effort to hit it. In this case, it's still up to me to hit it again, and not for you to hit it back. That means that my original hit was a pointless addition to the game.
[1] Here I should note that I take this example because it's fresh in my mind, and not because I have anything to add to all that zombie fevered discussion that's been around recently. The dialectic here is an example of the phenomena I have in mind, and not a diagnosis of that discussion.
[2] Again, note that I'm not intending to opine on abortion here. I'm merely pointing out how the dialectic might look if the burden of proof were in some place. It might be elsewhere: that's fine, and exactly my point. Where we locate the burden of proof in these discussions makes a big difference to the force of various arguments.
Begging the question and shifting the burden of proof
I'm not sure I understand the relation that you claim holds between begging the question and shifting the burden of proof. Here’s how I tend to view this issue myself.
Every once in a while, I come across a philosophy paper in which the author would declare an argument ‘question-begging’ because its conclusion implies the negation of the proposition which the author was defending in the course of the dispute. It is obvious that this complaint is invalid, since otherwise no deductive chain of reasoning could ever persuade anyone of anything. So what's wrong with this complaint?
Say proposition P is put on the table: I affirm it and you deny it. If I want to persuade you rationally to accept P, I clearly cannot rely on arguments which have P among their premises, even if P is true and even if I know that it is. But suppose P follows from Q, whose truth you haven’t so far contested. And suppose I produce an argument for Q which you find persuasive. I then point out to you that P follows from Q.
Once you realize that Q implies P, you must revise either your denial of P or your assent to Q. Suppose Q is to you more credible than ~P is. I would then have persuaded you rationally to accept P, without begging any questions. Now suppose instead that ~P is to you more credible than Q is. You will then deny Q, and at that stage a new dispute will begin, where I can no longer rely on Q in the course of my argument, even if Q is true and even if you know that it is.
I agree with Pablo. In my
I agree with Pablo. In my post 'Assessing Arguments and Begging Questions', I suggest that a question-begging argument is "one that could not reasonably sway anyone who did not already accept the conclusion (because the conclusion is transparently contained within the premises in a way which makes the reasoning vacuous)." This is similar to the definition given in your opening sentence. Either way, it doesn't seem to have anything to do with playing burden-of-proof tennis.
You write: "If my argument A relies on the falsity of Q, and Q is true iff P is, then my argument straightforwardly assumes that P is false before it begins."
But that can't be good account of question-begging, as Pablo explains. (See also my 'Arguing by Degrees'.) So long as ~Q is antecedently more plausible than ~P, a valid argument from ~Q to ~P will have some rational force (in a way that a truly question-begging argument would not).
I'm a little busy, so
I'm a little busy, so perhaps a more complete reply to each of you later, but:
Richard and Pablo,
The schematic example with A, P and Q seems to be a counter-example to your view, since the objection Q begs the question in your sense (it won't sway anyone not disposed to accept the implication that argument A fails), but still seems to me to be a perfectly good objection.
Further (and this certainly needs more thought), aren't claims about the antecedent plausibility of P and Q going to be problematic given that an argument is by definition a response to a disagreement in how plausible these are? Question-beggingness should be an observer-independent property of arguments, and it won't be if an argument begs the question only if its target has certain credence levels in the propositions in question.
Brandon,
Interesting thought, I'll have a think about how I can deal with hypothetically high burdens of proof.
Alex
Not Question-begging
"since the objection Q begs the question in your sense (it won't sway anyone not disposed to accept the implication that argument A fails)"
My view is not that a non-QB counterargument must potentially sway someone "not disposed to accept the implication [i.e. counter-argument]." That's clearly a hopeless view: no argument is such that it will sway those who won't be swayed by it. Rather, my view is that a non-QB counterargument must potentially sway someone who was not antecedentally disposed to accept its conclusion. Suppose I'm arguing from ~Q to ~P, when P iff Q, and my opponent believes that P. This schematic argument is NOT essentially question begging at all, since it's entirely possible that someone might satisfy the following conditions. They:
(i) initially accept P
(ii) find ~Q even more plausible than P
(iii) hadn't previously reflected on the fact that P iff Q, and hence that their credences in (i) and (ii) are inconsistent.
By revealing this inconsistency, you can rationally persuade them to give up their belief in P (or, perhaps, in ~Q). So it's not question-begging.
Observer Independence
Your last point is interesting. I was actually thinking of an observer-independent notion of credence. So: not the actual credence of the individual you happen to be talking to, but that of a hypothetical, normalized P-believer.
Really, the core of my account is the following test: could we plausibly expect a hypothetical observer to find the premises of the argument more antecedently plausible than the conclusion? The answer may vary slightly with context, just as -- I think -- whether something is question-begging may vary slightly with context (widespread background knowledge, culture, etc.). But it doesn't seem wildly relativistic in the way you suggest.
I'm afraid I'm a little lost
I'm afraid I'm a little lost here.
In the first paragraph, I don't understand what the "hypothetical normalized P-believer" might be. The thing that springs to mind is an ideally rational agent. But that will presumably imply that such an agent is already rational in the sense that they believe all of the implications of their beliefs. That in turn means that they couldn't possibly ever be receptive to any argument, regardless of whether or not it begs the question. (Another way to put that: Arguments merely demand consistency, so can't sway someone already wholly consistent.)
The more general worry is this: We're discussing the force of arguments, which are tools for evaluating belief justification. I don't see how we can shed any light on this issue by considering what ideal agents would appreciate, since those are the very people without the need for any such tool.
Alex
Right, that's why I appealed
Right, that's why I appealed to normal, not ideal, believers.
Normal as in average?
Normal as in average? Surely that can't be right: the average person might hold bizarre credences in all sorts of view. More generally, isn't this going to reduce to a simple kind of naturalistic fallacy?
To take a simple example, presumably many religious believers will have a higher credence in the proposition that God exists than in any other proposition. Presumably you don't want to say that, if such people are in the majority, then no argument with "God exists" as a premise can beg the quesion. The argument "God exists, so Atheists are Stupid" begs the question, regardless of what credence the majority has in the premise and conclusion.
Alex
No, 'normal' doesn't mean
No, 'normal' doesn't mean 'average'. (You can cut a leg off every actual dog and it won't change the fact that a normal dog has four legs. You've just made it so that no actual dogs are normal.) I'm not sure exactly how to analyze 'normal', but it doesn't strike me as a problematic notion. It's inherently normative, rather than natural/statistical, so there's no naturalistic fallacy here. (My initial invocation of a 'hypothetical' normalized agent was meant to signal this independence from actual contingencies.)
Though I'm not sure your example works anyway. Who is the target audience of the "atheists are stupid" argument? If the target is atheists, presumably even the 'average' account would count this as question-begging. But if the target is theists who are initially disposed to think well of atheists, it's not so clear that the argument is question-begging. Perhaps a committed theist really is committed to thinking that atheists are stupid (that will presumably depend on other premises of the argument besides 'God exists'), in which case it's perfectly legit to present this argument to persuade a committed theist who doesn't yet accept that conclusion.
On the abortion example
You write: "What's gone wrong here is that my anti-abortion argument assumes that the burden of proof is on you to tell me why abortion is not murder, and not on me to tell you why it is."
But that's not what's wrong at all. The problem is instead what you highlight in your previous sentence, namely, that "those in favour of abortion are extremely likely to just deny the premise that abortion is murder"
What's wrong, in other words, is that their premise is no more plausible than the conclusion. Most importantly: nobody is likely to accept the premises who does not already accept the conclusion. And that's what it is to be "transparently question-begging". It's not nearly so transparent where the burden of proof lies in such a case, but that doesn't matter because it's a separate issue from whether an argument is question-begging.
Burden of Proof and Begging the Question
Very interesting suggestion. I'm inclined to think that burden of proof is not a norm for arguments but a norm for discourse that requires the agreement of both parties. Thus it's easy to identify burden of proof in cases of law, or formal debate, because they accepted a certain standard and distribution of proof when they entered into that particular kind of argument. But outside of such artificial cases, I don't think there's a real foundation for claiming that one side or the other has a burden of proof unless they have explicitly accepted it. Even if this weren't so, however, there seems to be good reason for distinguishing failure to meet the burden of proof from question-begging: I can accept, for instance, an artificially high burden of proof, for the sake of argument, or because we are talking about an area in which we need to be especially careful. It doesn't follow from this that if I fail to meet that burden of proof that my argument was question-begging: it may be a perfectly reasonable and straightforward argument, just not one up to that standard.
I'm inclined to think that the best word on begging the question is still Aristotle; on Aristotle's view:
(1) There are conclusions for which question-begging-style arguments are perfectly legitimate (although perhaps not always useful), namely, self-evident ones.
(2) This is the problem with question-begging: it involves treating a non-self-evident claim as if it were self-evident.
(3) Thus an argument begs the question when one could not reasonably expect to know at least one of the premises without already knowing the (non-self-evident) conclusion.
(4) And the proper response to a false accusation of begging the question would be to show that there are reasons not dependent on the conclusion for accepting the premises.
Fallacy
The system appears to be specifically trying to discourage misleading as opposed to any other objective. Many if not most typical examples of fallacies have persuasive power, they just overestimate it for example:
"If I have the flu, then I have a sore throat. I have a sore throat. Therefore, I have the flu."
The fallacy rules are a request to restate that argument so that it will not mislead, for example with the conclusion "therefore I may have the flu". Generally this amounts to a sacrifice regarding persuasive power.
Brandon,
I don't think we want to give up on the ability to ascribe burden of proof to discourse where it is in dispute. It would seem to be useful.
If there is no current default rule then as a community we should flesh one out. In discourse between two people you could place the default burden of proof on the side that appears to have an unfair advantage, the side which is most able to make a proof or the side that initiated the dispute.
Default Burdens of Proof
Anonymous,
It certainly can be useful, which is why default burdens of proof are artificially created for certain types of inquiry, e.g., courts of law. But this is only possible because parties interacting in a court of law -- lawyers and judges and so forth -- implicitly agree on the terms going into the inquiry; and there appears no way that we "as a community" can require default burdens of proof in general, because most of our inquiries and arguments do not occur in artificial contexts. (And I think we should be at least wary of attempts to impose such default burdens. After all, one of the common complaints in feminist and race theory is that default burdens of proof are designed to the convenience of the people in power, who try to impose it on everyone. And they are right that this has at least sometimes genuinely occurred, e.g., in the U.S. with black testimony in the white South.
Yes we should definitely be
Yes we should definitely be wary - there are so many vested interests involved. But leaving it ambiguous seem to be part of the issue.
Default burdens of proof emerge in spite of us not formalizing the burden of proof rules. They also tend to end up being debated specifically in the context of the debate using the compass that always points away from you.
To use your example, we might have benefited if Southern Whites had had an additional reason to question how they distributed the burden of proof.
Default burdens of proof, I
Default burdens of proof, I think, are taken to emerge because people analogize arguments they are having to formal debate. But what makes the application of the analogy inevitably ambiguous -- and I do think that it is inevitable -- is that the natural arguments lack the artificial aspects (rules, agreements, etc.) that make it possible to assign a definite burden of proof. In other words: to have default burdens of proof, you need authoritative conventions to which one side can unambiguously appeal.
Of course, it is entirely possible for people to build their own set of conventions impromptu as the argument goes, and many natural arguments involve this. But I don't think there's any basis for trying to force someone to accept any kind of burden of proof that they don't want to accept. There are certainly burden-of-proof rules that would make considerable sense for certain kinds of argument, should the participants agree on them; but I doubt that there are any that apply to every kind of argument. So I think the best one can hope for in this direction is simply to lay out the sort of rules that make sense for particular kinds of argument, along with the reasons for those rules, and put it out there as a sort of argumentative advisory.
If the advisory is the best
If the advisory is the best we can do, then I'll take it.