Ruminations on scientific expertise and the ethics of persuasion

« previous post | next post »

We've had a bumper crop of recent electoral events where I live, and given that I write a good deal about language and persuasion, at regular intervals I get asked to advise on political campaigns. I always decline.

I have no trouble advocating publicly and with feeling for my own political beliefs. I also have no trouble accepting money from commercial entities (well, not usually, anyway) who want to hire me to consult on the technical aspects of their persuasion strategies. But I do get squeamish when it comes to drawing on my knowledge of language and psychology in order to tinker directly with the machinery of political messaging. It basically comes down to the fact that, in order to do so effectively, I would inevitably have to recommend—at least some of the time—the use of techniques that I would ultimately prefer not to play a prominent role in our political discourse. If you read much about political psychology and persuasion, it's hard to miss the growing pile of studies that reveal the various levers and buttons that reside in the less deliberative rooms of our minds and that can set in motion behaviors and choices all while leaving the persuadee convinced that it's his rational, thoughtful self that's been at the control panel all along. Call me old-fashioned, but I still think that the wholesale exploitation of shallow cognitive processes for political ends accomplishes no good thing for the overall health of civic life, and that thoughtful deliberation and evaluation of candidates and their ideas should drive our democratic impulses.

No doubt I have an overly sentimental and earnest view of democracy. Possibly, this is due to the historical accident of having been born the granddaughter of a Czech dissident who spent years in a Soviet prison, and the daughter of parents who left their lives and livelihoods behind the iron curtain in order to enjoy certain Western freedoms. It tends to leave you wanting democracy to actually be worth all of that adversity. And certainly, some would call me old-fashioned—for instance, among his political writings, George Lakoff has published a book titled The Political Mind: Why You Can't Understand 21st-Century American Politics With an 18th-Century Brain. In it he argues that the view of voters as rational deliberators is hopelessly outdated, and that it's not just misguided but irresponsible, if you hold sincere political beliefs, not to advocate for them using techniques that appeal to the less rational thought processes of the electorate. He may well be right. Nevertheless, my personal policy has allowed me to avoid certain scratchy internal discomforts up to this point.

But since I've been approached again this month to advise on a (highly worthwhile) political campaign, it's been on my mind. What I've been wrestling with lately is the sharp line I've drawn in my own thinking between using certain techniques for political ends, and discovering or writing about them, the latter of which I've always felt to be beyond reproach. But the distinction between the discovery and the use of questionable techniques has the potential to smudge occasionally under the exigencies of doing good science.

For example: suppose you want to study the impact of, say, false rumors or subliminal messaging on attitudes about a candidate. A standard approach might be to invent fictional candidates, apply the experimental variables, and then measure attitudes about the candidates. Fine, suppose you find an effect. (People have.) It's very hard to know if the effect would scale up to the real world.

Take, for instance, the kerfuffle that arose during the 2000 U.S. presidential election in which the Bush campaign ran a TV ad criticizing the Democrats' health care policies while the word RATS flashed on the screen for a couple of video frames, too quickly to be consciously registered by most viewers. Bush dismissed accusations of sneaky subliminal messaging as "bizarre conspiracy theories." Of course, it was wildly implausible that the image was not deliberately inserted. And there were, as far as I was concerned, reasons to believe that at least in theory, the subliminal word might affect people's attitudes. But could it in fact do so in a real flesh-and-blood election where that subtle association might easily be swamped by mounds of knowledge and information that voters have about each of the candidates?

To find out, you'd need to actually expose people to the messages during a real political campaign, which is exactly what political psychologists Joel Weinberger and Drew Westen did in a 2008 study. (Westen, incidentally, happens not to share my discomfort with blending scientific expertise and political advocacy, and regularly serves as an advisor to the Democratic Party). To make a somewhat subtle story mercifully brief, their study found that flashing subliminal images—in an experiment administered over the internet, no less—had a measurable effect on people's attitudes towards Gray Davis, who was at the time governor of California and the subject of a recall referendum. (You can read the entire study here.)

Weinberger and Westen measured attitudes, leading to an admittedly less impactful finding than direct evidence of an effect on voting behavior. But an Israeli study published in 2007 by Ran Hassin and colleagues did find that voting decisions (as self-reported in any case) were shaped in part by exposure to subliminal stimuli. The finding was that subjects who'd been exposed to subliminal images of the Israeli flag ended up voting for more moderate candidates than those who'd "seen" a fleeting non-symbolic image instead. The authors' interpretation of the study was that the symbol of the flag had activated for subjects a sense of collective national unity that aligned with mainstream, centrist attitudes. (I'm not sure I'd want to hazard a guess, though, as to whether exposure to an American flag would lead to more, or less moderate political opinions). Apparently, this was no transient effect, lasting long enough to tilt voting behavior away from more extremist positions. Along similar lines, David Beaver recently posted here about a study in which voter turnout was dramatically increased simply by tweaking the wording of a question on a survey that had been administered the previous day.

Obviously, we're entitled to draw more muscular conclusions from these latter studies than we are from experiments that use fictional candidates dressed up in stiff, made-to-order biographies. But these experiments also open an ethical can of worms: to what extent is it justifiable, in the name of research, to put people through a study that might well skew their political choices, especially when the study involves exposing them to information about which they can't possibly deliberate? The studies were clearly small enough so as to have at best a negligible effect on the outcome of the election (and the Weinberger and Westen study did provide a debriefing note immediately after the assessment measures, explaining the details of the experiment before subjects actually went to the polls). But to my mind, at least, it's not crazy to question whether studies such as these impinge on one's individual rights to make a fully-aware choice as to who should be permitted to govern you (or, for that matter, whether to vote).

It occurs to me that these concerns only arise if you strongly value the importance of individuals making an informed, reasoned choice about their governance, if you see that as an important individual right. But perhaps what studies like these tell us is that this is a misplaced value, that no such right has been violated for the simple reason that the average voter is unlikely to ever be in a position of making an informed, reasoned choice—that voters' opinions are too easily buffeted by simple heuristics, associations, or whatever. In which case, my insistence on preserving the opportunities for a reasoned choice, and my reluctance to deploy a well-equipped arsenal of persuasive tools to further the causes that I care about are entirely misplaced, and Westen and Lakoff are right, after all.

But I don't think the science has led us there, just yet. We're seeing clear evidence of limits to our deliberative powers and this is perhaps leaving some of us a bit shaken. But so far, the evidence has not yet banished reason and clear-headed evaluation entirely from the stage. And the data pointing to less conscious or rational processes in our choice-making might in the end lead us in the opposite direction; it's possible that a sober consideration of these limits might even enhance the extent to which reason tempers our decision-making—perhaps not just individually, but institutionally as well, if we can be moved by that knowledge to put in place structures that encourage deliberation rather than discourage it. I suppose so long as I see this as a possible goal, and a good one at that, perhaps even a better one than the championing of any particular political cause, I'll continue to spend my energies in studying and writing about persuasion rather than prescribing its use for specific political ends.



  1. GeorgeW said,

    October 5, 2011 @ 7:39 pm

    Thanks for your principled decision.

    I too would like to think that reason plays at least some role in our political choices. To the extent that we reduce it with surreptitious devices, democracy suffers.

  2. Jason said,

    October 5, 2011 @ 10:36 pm

    I'm afraid my own studies in psychology have led me to the position that you ascribe to Lakoff — I would call it the BF Skinner/Walter Lipmann position — that nothing resembling the reasoned, deliberative policy making process you privilege in the article exists or is ever likely to exist in the public mind, and hence the only question to be settled is what form of manipulation elites should use. The nature and history of our current society constitutes powerful emprical refutation of the idea that the public has any sensible idea what it wants or what's in its own interests.

    A lot of this has to do with the invention of television. TV the most inimical to thought and deliberation of any medium ever invented. I can no longer stand watching it, and it depresses me that TV is still the dominant way most people learn about the world outside their own neighborhood. But it goes well beyond that.

    I now believe that only restricting the vote to those with some measure of scientific training, especially those who have themselves studied the science of interpersonal manipulation ("persuasion" is too neutral a word) is likely to produce better governance. Since this is politically impossible, it follows that democracy will be eternally dysfunctional, eternally ruled by photogenic narcissistic high functioning sociopaths who control the masses with a message built around populist nostrums and xenophobic gammon, that the public debate will continue to be driven by shallow imagery and one-second soundbites, that the strategy of whipping up triumphalistic in-group boosterism and ressentiment against the other will continue to be a winner, whether the person using it is Bush, Obama or Hitler. In short, we are screwed.

  3. Jason Eisner said,

    October 6, 2011 @ 2:11 am

    But these experiments also open an ethical can of worms: to what extent is it justifiable, in the name of research, to put people through a study that might well skew their political choices …

    Yes, and thanks for the thoughtful discussion. I'm loosely reminded of Sauper & Barzilay (ACL 2009), Automatically Generating Wikipedia Articles: A Structure-Aware Approach:

    In addition to our automatic evaluation, we perform a study of reactions to system-produced articles by the general public. To achieve this goal, we insert automatically created articles into Wikipedia itself and examine the feedback of Wikipedia editors.

    In future news: "To determine whether our newly created species was viable, we released it into the wild and measured how effectively it took over its ecological niche …"

  4. maidhc said,

    October 6, 2011 @ 4:55 am

    I'm dubious about the RATS theory. Two video frames is 1/15 sec. That is well within most people's perception. And even if many people didn't notice it, it would soon show up on YouTube.

    [(js) You can find the ad here. The timing hovers right around the perceptual threshold. I've shown it quite a few times to large groups of students and have generally found: When I don't tell them to look for anything unusual, no one is aware of seeing it. When I tell them to look closely for a briefly-flashed message, maybe 10% are. When I tell them what word it is, roughly 20%. And if I tell them where to look for it, that number goes up to about half the group (after repeated viewings by this point).]

  5. Allienne Goddard said,

    October 6, 2011 @ 7:06 am

    I agree with Jason, except that the state of politics cannot be blamed on television. It has always been like this: a small number of intellectuals rebuke the masses for being ignorant of political realities, while the elites rule by dividing the electorate. It was supported by radio before television, and newspapers before radio. It will be supported to an ever increasing degree by the Internet. This is what makes FDR such an interesting anomaly in US history; a member of the elite introduced socialist and regulatory reforms, which rapidly became highly popular, in response to an economic catastrophe. A president of the United States kept his promises to the poor who elected him. Remarkable. I don't expect to ever see such a thing in my lifetime.

    As for whether or not it is ethical to use methods besides rational argument in politics? Appeals to emotion can be ethical, in my view, if they are appeals to decent emotions. More cynical methods (voter suppression, lies, rumor-mongering, psychological manipulation) ought not be used, unless one thinks winning is important. If one believes that not using these methods results in political losses that have a serious negative affect on the lives of the majority of people here and in the world, then one might well be better off using them. On the other hand, there are so few decent people on either side that it hardly makes a difference who gets elected. So, perhaps one might as well hold on to one's principles.

  6. GeorgeW said,

    October 6, 2011 @ 7:50 am

    @Allienne Goddard: "Appeals to emotion can be ethical, in my view, if they are appeals to decent emotions"

    Subliminal rats are okay if advancing one's personal sense of decency? But aren't "decent emotions" subjective?

    George Lakoff in "Moral Politics" demonstrates that there principled positions on both the left and the right.

  7. Andy Averill said,

    October 6, 2011 @ 8:38 am

    Oh dear, well I hate to sound like the most cynical person in the group, but I'm not even convinced that "rational" decision making leads to good governance. Read David Halberstam's The Best and the Brightest for an account of how a group of Harvard-educated advisors around JFK merrily led us into the hell of Vietnam. And there are many other such examples.

    Whatever illusions people may once have had about democracy (which incidentally was viewed with deep suspicion by most of the Founding Fathers), at this point the best argument in favor of it is that it evens the playing field better than any other system. It's at least theoretically possible that the various distortions of reality to be found in the minds of the populace at large can cancel each other out.

  8. Harold said,

    October 6, 2011 @ 9:34 am

    Andy, Machiavelli thought the same thing about democracy:

    "I discussed above at length on this subject, where it was demonstrated that the People are less ungrateful than Princes. And as to prudence and stability, I say, that a People is more prudent, more stable, and of better judgment than a Prince: And not without reason is the Voice of the People like the Voice of God . . . . . And if . . . they err in things concerning bravery, or which appear useful, a Prince also errs many times in his own passions, which are much greater than those of the people. . . . And a People will never be persuaded that it is better to bring to that dignity a man of infamous and corrupt habits: to which a Prince may be persuaded easily and in a thousand ways. It will also be observed that when a people begin to hold a thing in horror, they remain in that opinion for many centuries, which does not happen with a Prince. . . . In addition to this, it will be seen that the Cities where the people are Princes, make the greatest progress in the shortest time and much greater than those who have always been under a Prince." –Machiavelli, Discourses on Livy

  9. David Eddyshaw said,

    October 6, 2011 @ 9:54 am

    The cynical have something to explain:

    The world's democracies are the pleasantest places to live, by and large. (I grant you the correlation is imperfect, but it is surely very strong. If you dispute this, I doubt whether you have ever lived in a non-democratic state.)


  10. Ran Ari-Gur said,

    October 6, 2011 @ 11:02 am

    @David Eddyshaw: Perhaps I should leave it for someone cynical to answer your question, but . . . I believe the usual explanation is that the growth of a strong middle class leads to democratic reform, rather than the other way around.

  11. David Eddyshaw said,

    October 6, 2011 @ 12:13 pm

    You would then presumably expect a state that has achieved democracy to deteriorate rapidly, given that democracy is such a dysfunctional system.

    One can, I suppose, think of examples, (Weimar … though there the catastrophic worsening was surely connected with the very *loss* of democracy) but there seems to be no such universal doom. True believer cynics will, I expect, ask me to trust that time will tell.

  12. David Eddyshaw said,

    October 6, 2011 @ 12:32 pm

    Machiavelli (no wet liberal, I believe) has nailed it. Many thanks, Harold.

  13. maidhc said,

    October 9, 2011 @ 5:28 pm

    I watched the clip. "Rats" is very obvious.

    Of course I was paying attention though. Most people don't pay that much attention to television ads, and I wonder if a large group of students is that much better.

  14. Dominik Lukes said,

    October 9, 2011 @ 8:07 pm

    The paradox of your preference for "rational" decision making is that it is stated in profoundly emotional terms. You have thus given up your argumentive advantage.

    You treat rationality as if it were an entirely unproblematic notion that leads to predictably felicitous results. But as other commenters pointed out, that is not so. Throughout history rational individuals arrived at contradictory solutions to problems. That is because even perfectly logical operations still operate on fuzzy concepts whose fuzziness is influenced among other things by emotions. You cannot even understand concepts like these without some kind of affective involvement.

    Moreover, appeals to rationality are themselves a part of the argumentative arsenal, and are used to elicit emotion.

    Equally, it is not clear where manipulation begins and where it ends. Is presenting one's views with infectious zeal and passion less manipulative than using emotional examples, crafting framing or even lying? Ultimately, it is the duty but also inevitable destiny of each argumentative interlocutor to put forth their best case for the position they advocate.

    [(js) You make some extremely reasonable points. I would certainly never want to say that it's appropriate to banish emotion from political discourse—"rational" approaches that do so are likely to be incomplete or even veer off into the inhumane (as one might argue is the case for Plato in his Republic). But in politics, I think emotion should serve as a companion to reason—to make valid arguments more compelling and easy to grasp, to focus attention on the issues that matter most to the well-being of citizens, and so on. It's when emotion becomes untethered from reason that I think we run into trouble, or when it's used to obscure valid conclusions in favor of conclusions that fail to align with reality. A somewhat similar view is articulated at length by Martha Nussbaum in her book Poetic Justice: The Literary Imagination and Public Life, where she argues for the role that emotion and empathy have to play in judicial decisions.

    In the domain of persuasion, it may well be that emotional involvement can often serve to enhance reasoning for the audience of the message. For instance, there's a very reliable effect in the persuasion literature showing that increasing the personal relevance of a message heightens people's ability to distinguish strong from weak arguments, presumably by dialing up the motivation to process the message deeply. (This effect has been documented at length by Richard Petty and John Cacioppo, and many others as well.) I suspect there's an affective component to this, and that you'd find similar results with messages that induce a state of empathy for others, though I don't know of any work that has specifically looked at this. On the other hand, the literature is also riddled with examples of techniques that achieve the effect of dampening discrimination between strong and weak arguments. I think we can learn something from this work that is relevant for the ethics of persuasion. But perhaps this is fodder for future posts.]

  15. Dan M. said,

    October 9, 2011 @ 9:39 pm

    maidhc, do note that the youtube rendering of the video may be quite different from a television rendering of it. Video compression can significantly change how an object (such as the word "rats") moves between frames. On the other hand, the majority of digitally broadcast television use the same basic algorithm (although it differs a lot from the one on youtube). But I think by far the most important point is that in youtube, you can rewind and check what you think you saw, whereas in broadcast, it's simply implausible (even if true) to believe that you've just watched a subliminal message.

    [(js) Indeed, and one should also not under-estimate the power of priming and higher-level expectations on perception. On that front, Michael Sherman offers this amusing dismantling of bogus claims about satanic verses embedded backwards into song lyrics, by showing just how easy it is to hear them once you know what they are.]

  16. maidhc said,

    October 10, 2011 @ 9:29 pm

    When I think of "Rats" my first two associations are Jimmy Cagney and Marvin the Martin.

  17. John Cadigan said,

    October 14, 2011 @ 2:34 am

    As this is my first post, I'd like to say that I enjoy the languagelog quite a bit.

    However, I'd like to add an addendum to the description of Lakoff's ideas about arguments in politics. Effective framing isn't only used to convince the audience about one proposition (eg. global warming vs. climate change), but to also reflect a system of values. Lakoff himself often recalls how Reagan's political adviser discovered that even though people who took a survey were opposed to Reagan's specific policies, they still wanted to vote for him. Most people employ the type 1 reasoning when voting (how many of you pull out each candidate's platform and calculate how well they match your beliefs?). This appears depressing at first, but it is actually more effective to elect someone who reflects your values because specific policy proposals often have to achieved through compromise. Furthermore, values suggest what the politician might do in a future crisis.

    If you use incorrect framing, you are misrepresenting your thoughts ("I'm against tax relief" still activates the idea that taxes are a pain/burden OR "We need to stop climate change" [We should stop the climate from changing [there is no implied agent as in 'global warming' or 'climate disruption']).

    Often Lakoff's suggestions are somehow viewed as ways to exploit that second half of the imagined reason/emotion dichotomy (the Wason task and myriad of framing effects studies suggest otherwise!), but they are just hallmarks of clear, cogent communication.

RSS feed for comments on this post · TrackBack URI

Leave a Comment