Richard Petty, John Cacioppo and Martin Heesacker asked a question that turns out to have an unobvious answer: does phrasing an argument as a rhetorical question make it more persuasive?
Their 1981 experiment presented students with a proposal to institute comprehensive senior examinations at their university. Some heard arguments as statements. Others heard the same arguments as rhetorical questions — don't you think comprehensive exams would improve the quality of education here? And crucially, the proposal was framed either as taking effect next year, making it personally relevant, or in ten years, making it not.
For the low-involvement students, rhetorical questions helped — but only when the arguments were strong. With strong arguments, rhetorical phrasing increased persuasion; with weak arguments, it decreased it. For the high-involvement students, the pattern reversed: rhetorical questions reduced persuasion regardless.
The explanation is elaboration. A rhetorical question is a prompt to think about the message. When the audience was not otherwise motivated to think, the prompt increased processing — and increased processing helps a good argument and hurts a bad one, because what the listener does when processing is evaluate. When the audience was already thinking hard, the questions interrupted the processing already under way and reduced it.
Two practical conclusions follow, and they run in opposite directions.
For a speaker with a strong case and a disengaged audience, rhetorical questions are a genuine tool: they recruit the listener into doing the work, and self-generated conclusions are more durable than supplied ones. Once you have concluded something yourself, rejecting it later means contradicting yourself, which brings Festinger's dissonance pressure to bear on your side.
For a speaker with a weak case, they are dangerous — which is the empirically interesting finding, and it is a rare instance in this guide of a technique that punishes bad-faith use.
But there is a second family of rhetorical question that Petty's design does not capture, and it is the manipulative one: the question that forbids an alternative answer. How much longer are we expected to put up with this. Do you want to be the kind of person who lets that happen. Is this really the country we want to leave our children. These do not invite elaboration; they present a conclusion in interrogative clothing, and the interrogative form makes disagreement awkward because there is nothing asserted to disagree with.
The cross-examiner's version is the same device with legal force behind it, and so is the sales close that asks whether you would agree that peace of mind is worth something.
The countermeasure for the manipulative variant is to answer it. Out loud, literally, as though it had been a real question. A question designed to be unanswerable collapses when someone supplies an answer, because the answer reveals that the question had content and the content was contestable. Well — quite a lot longer, if the alternative is worse. No, I do not agree that peace of mind is priceless; I agree it is worth about eighty pounds a year.
Cross-ref: Milton model double binds