The first time I opened my mouth in college it was on the topic of devil’s advocacy. It was September 1993, and we were having a policy discussion in my Core humanities class: the teacher asked whether we should allow people to make comments they don’t agree with, “just for the sake of argument?” No one could see what the problem with doing so might be, so I spoke up. I argued against, passionately and persuasively. I won the others over: for the rest of the year that class had a “no DA policy.”
The argument I gave was that devil’s advocacy undermines the trust on which conversation is grounded. Why should we take someone seriously if she herself doesn’t believe what she is saying? Conversational progress is predicated on sincerity and openness, because only those willing to put their cards on the table are in a position to learn—or be learned from. I wasn’t aware of the oft-tweeted phrase “the devil has enough advocates,” but it would have fit right into my rousing speech.
My case against devil’s advocacy was, however, insincere. The truth is, I was just playing devil’s advocate.
The sincere case in favor starts with the Catholic Church. The devil’s advocate (advocatus diaboli) is an official religious office, dating to the sixteenth century, tasked with arguing against the canonization of a given individual. The official name for this job was “promoter of the faith,” because such a person had, in the words of seventeenth-century historian Giovanni Papa, “the role of … searching for the truth and safeguarding the law, so that the candidates for the honor of the altars appear to be truly worthy, without stain no matter how small or the slightest uncertainty.”1
Canonization is a process of discovering people who were already saints, not deciding that someone should be “made” a saint. In order to arrive at these truths the Church employed an adversarial process called a “contradictorium” in which the devil’s advocate faced up against the saint’s advocate (also known as “God’s advocate”, or, more technically, “the advocate of the cause”). It was the job of the devil’s advocate to point out flaws in the would-be saint’s character, or to produce naturalistic explanations of purported miracles performed by that person.
One way of seeing the importance of the devil’s advocate is to look at what happened after 1983, when Pope John Paul II introduced legislation lessening the adversarial role of the “promoter of the faith”: we find a larger number of canonizations during his pontificate than during the entire four-hundred-year prior history of the canonization process. For better or worse, the process became less rigorous when the devil’s advocate was made to take a backseat.
Devil’s advocacy injects normative rigor into a group inquiry—but in quite a specific way, namely by containing a predictable well of enthusiasm in a given direction. The defense attorney in a criminal trial is a kind of devil’s advocate, in that she needn’t believe her client is innocent. Her job is to resist or militate against the case being made by the prosecutor—who is, in fact, committed to the guilt of her client. (Everybody has a right to a defense, but only those believed guilty should be prosecuted.) A courtroom, too, is, therefore, a kind of “contradictorium.”
Not every case of adversarial persuasion can be described in this way. For instance, if two contestants are vying against one another, offering competing speeches as to why they would be better at, e.g. marketing some product or serving as your contractor or running for president, neither is playing devil’s advocate. Such disputes are symmetrical, whereas a contradictorium is structured asymmetrically: there is, on the one hand, the dominant, group-espoused view, and, on the other hand, a subordinate view that needn’t be espoused by anyone, whose function is to serve as a formal check on the group’s cohesiveness.
Group inquiry is subject to well-known biases that stem from the fact that individuals in the group often have stronger incentives to get along with one another than to find the truth. Groups tend to “cascade” in the direction of least resistance: in canonization, the group will tend to coalesce over the desire to find someone to venerate (the danger being: idolization), and, when a crime has been committed, the group will tend to coalesce around the desire to find someone to blame (the danger being: scapegoating). To give just one example of how such bias works: groups tend not to discuss information known to few members of the group and to overemphasize information known to much or all of them.2
Those biases might incline us to seek alternatives to group inquiry—and, in fact, alternatives do exist: to answer a given question, one can conduct empirical research, turn to an expert adviser, or consult a scholarly consensus. If the answer is spread, in bits and pieces, over members of a large group, one might seek an extra-deliberative mode of aggregating it. For example, Cass Sunstein has pointed that the various biases associated with group cascade can be avoided through prediction markets in which people place bets on the likelihood of a given outcome. Sunstein says that “prediction markets often outperform deliberating groups, simply because they are so effective at pooling dispersed information among diverse people. Indeed, prediction markets realign private incentives in a way that makes them exceptionally well-designed to reduce the problems that infect deliberating groups.”3
These extra-deliberative, technical measures—research, expert advice, scholarly consensus, prediction markets—are ways of avoiding bias. But they do so at a cost, namely by avoiding the very thinking that such bias “infects.” There are times we care not only about the correctness of the conclusion arrived at, but also about the kind of ownership we can take over that conclusion; those times, these measures won’t do. Deliberation makes us the sources of our own judgments, and this matters when the judgment in question—that so-and-so is a wrongdoer or a saint—has to ground actions of deep moral or spiritual significance, such as punishing or venerating. Those actions need to spring from an understanding the agent—whether it is an individual or a group—can full-throatedly avow.
Thinking together is riddled with pitfalls, but we can’t really claim to live together without doing it. That is why we need devil’s advocates: they safeguard group-deliberation from the inside. The devil’s advocate defends faith and justice by being in the group but not of it: by keeping the group divided against itself, she holds a space for truth against the pressure of consensus. A devil’s advocate is, for instance, well set up to hunt for as-yet unshared information, since for her the sharing of information is never an attempt to be on the same page as other people.
In principle, devil’s advocates allow us to combine the goal of figuring things out together with the goal of commitment to the truth—and they do this by functioning as a check on group consensus. In practice, devil’s advocates often fail to adhere to this subordinate “checking” role—whether it be a spiteful referee for a journal, an online troll or an attention seeking-provocateur, devil’s advocates are prone to excess. Such people are not held back and restrained by the rules of a given office, and they criticize in a way that is counterproductive and excessive.
Take me in 1993. Instead of checking the group consensus by raising some concerns about devil’s advocacy, I took it upon myself to swallow up the group with a persuasive speech. In group deliberation, there will always be those who jump at such a rhetorical opening, tyrannically exploiting the opportunity for cascade. And there will always be others who sow seeds of faction with the destructive aim of splitting the group against itself. The first indulge in an unrestrained quest for unity; the second in an unrestrained quest for division. We tend to respond to the first problem—the dangers of demagoguery—by emphasizing the value of critical distance, and the second—the dangers of sectarianism—by saying that we need to learn to get along. But it’s only if we face both problems at the same time that we see how serious our predicament is.
Whenever people must decide something together—in classrooms and city council offices, in board meetings and on the internet—we find freelance devil’s advocates doing a poor job of it. The asymmetrical structure of these informal contradictoria is fragile, quick to collapse into bad forms of agreement and bad forms of disagreement. Often these come together: oratorically speaking, the cheapest forms of unity are often those purchased by vilifying some subgroup into outsider status.
Indulgence in excess is so easy, self-restraint is so hard. If social media exaggerates these problems, that is also a kind of virtue: perhaps we have never been able to see, so clearly, what it looks like when we all try to get along. It’s not pretty. You can leave Facebook or Twitter, but you can’t really leave the ugliness behind. Technical skill is essential to informing our judgments, but it cannot be the ultimate ground of our getting along with one another. We cannot ask the authorities, experts or science to do all our thinking for us. Sometimes, we need to think as a group, and that means we cannot afford to cynically dismiss “devil’s advocacy” as a term of opprobrium. It has to become an honorific.
Image credit: xkcd
The first time I opened my mouth in college it was on the topic of devil’s advocacy. It was September 1993, and we were having a policy discussion in my Core humanities class: the teacher asked whether we should allow people to make comments they don’t agree with, “just for the sake of argument?” No one could see what the problem with doing so might be, so I spoke up. I argued against, passionately and persuasively. I won the others over: for the rest of the year that class had a “no DA policy.”
The argument I gave was that devil’s advocacy undermines the trust on which conversation is grounded. Why should we take someone seriously if she herself doesn’t believe what she is saying? Conversational progress is predicated on sincerity and openness, because only those willing to put their cards on the table are in a position to learn—or be learned from. I wasn’t aware of the oft-tweeted phrase “the devil has enough advocates,” but it would have fit right into my rousing speech.
My case against devil’s advocacy was, however, insincere. The truth is, I was just playing devil’s advocate.
The sincere case in favor starts with the Catholic Church. The devil’s advocate (advocatus diaboli) is an official religious office, dating to the sixteenth century, tasked with arguing against the canonization of a given individual. The official name for this job was “promoter of the faith,” because such a person had, in the words of seventeenth-century historian Giovanni Papa, “the role of … searching for the truth and safeguarding the law, so that the candidates for the honor of the altars appear to be truly worthy, without stain no matter how small or the slightest uncertainty.”1Jason A. Gray, “The Evolution of the Promoter of the Faith in Causes of Beatification and Canonization: A Study of the Law of 1917 and 1983,” Pontificia Universitas Lateranensis Facoltas Iuris Canonici (Rome, 2015).
Canonization is a process of discovering people who were already saints, not deciding that someone should be “made” a saint. In order to arrive at these truths the Church employed an adversarial process called a “contradictorium” in which the devil’s advocate faced up against the saint’s advocate (also known as “God’s advocate”, or, more technically, “the advocate of the cause”). It was the job of the devil’s advocate to point out flaws in the would-be saint’s character, or to produce naturalistic explanations of purported miracles performed by that person.
One way of seeing the importance of the devil’s advocate is to look at what happened after 1983, when Pope John Paul II introduced legislation lessening the adversarial role of the “promoter of the faith”: we find a larger number of canonizations during his pontificate than during the entire four-hundred-year prior history of the canonization process. For better or worse, the process became less rigorous when the devil’s advocate was made to take a backseat.
Devil’s advocacy injects normative rigor into a group inquiry—but in quite a specific way, namely by containing a predictable well of enthusiasm in a given direction. The defense attorney in a criminal trial is a kind of devil’s advocate, in that she needn’t believe her client is innocent. Her job is to resist or militate against the case being made by the prosecutor—who is, in fact, committed to the guilt of her client. (Everybody has a right to a defense, but only those believed guilty should be prosecuted.) A courtroom, too, is, therefore, a kind of “contradictorium.”
Not every case of adversarial persuasion can be described in this way. For instance, if two contestants are vying against one another, offering competing speeches as to why they would be better at, e.g. marketing some product or serving as your contractor or running for president, neither is playing devil’s advocate. Such disputes are symmetrical, whereas a contradictorium is structured asymmetrically: there is, on the one hand, the dominant, group-espoused view, and, on the other hand, a subordinate view that needn’t be espoused by anyone, whose function is to serve as a formal check on the group’s cohesiveness.
Group inquiry is subject to well-known biases that stem from the fact that individuals in the group often have stronger incentives to get along with one another than to find the truth. Groups tend to “cascade” in the direction of least resistance: in canonization, the group will tend to coalesce over the desire to find someone to venerate (the danger being: idolization), and, when a crime has been committed, the group will tend to coalesce around the desire to find someone to blame (the danger being: scapegoating). To give just one example of how such bias works: groups tend not to discuss information known to few members of the group and to overemphasize information known to much or all of them.2Garold Stasser, Laurie A. Taylor, and Coleen Hanna, “Information Sampling in Structured and Unstructured Discussions of Three and Six-Person Groups,” Journal of Personality and Social Psychology 57 (1989): 67.
Those biases might incline us to seek alternatives to group inquiry—and, in fact, alternatives do exist: to answer a given question, one can conduct empirical research, turn to an expert adviser, or consult a scholarly consensus. If the answer is spread, in bits and pieces, over members of a large group, one might seek an extra-deliberative mode of aggregating it. For example, Cass Sunstein has pointed that the various biases associated with group cascade can be avoided through prediction markets in which people place bets on the likelihood of a given outcome. Sunstein says that “prediction markets often outperform deliberating groups, simply because they are so effective at pooling dispersed information among diverse people. Indeed, prediction markets realign private incentives in a way that makes them exceptionally well-designed to reduce the problems that infect deliberating groups.”3Sunstein, Cass, “Deliberating Groups versus Prediction Markets (or Hayek’s Challenge to Habermas),” Episteme, 3(3) 2006, 192-213.
These extra-deliberative, technical measures—research, expert advice, scholarly consensus, prediction markets—are ways of avoiding bias. But they do so at a cost, namely by avoiding the very thinking that such bias “infects.” There are times we care not only about the correctness of the conclusion arrived at, but also about the kind of ownership we can take over that conclusion; those times, these measures won’t do. Deliberation makes us the sources of our own judgments, and this matters when the judgment in question—that so-and-so is a wrongdoer or a saint—has to ground actions of deep moral or spiritual significance, such as punishing or venerating. Those actions need to spring from an understanding the agent—whether it is an individual or a group—can full-throatedly avow.
Thinking together is riddled with pitfalls, but we can’t really claim to live together without doing it. That is why we need devil’s advocates: they safeguard group-deliberation from the inside. The devil’s advocate defends faith and justice by being in the group but not of it: by keeping the group divided against itself, she holds a space for truth against the pressure of consensus. A devil’s advocate is, for instance, well set up to hunt for as-yet unshared information, since for her the sharing of information is never an attempt to be on the same page as other people.
In principle, devil’s advocates allow us to combine the goal of figuring things out together with the goal of commitment to the truth—and they do this by functioning as a check on group consensus. In practice, devil’s advocates often fail to adhere to this subordinate “checking” role—whether it be a spiteful referee for a journal, an online troll or an attention seeking-provocateur, devil’s advocates are prone to excess. Such people are not held back and restrained by the rules of a given office, and they criticize in a way that is counterproductive and excessive.
Take me in 1993. Instead of checking the group consensus by raising some concerns about devil’s advocacy, I took it upon myself to swallow up the group with a persuasive speech. In group deliberation, there will always be those who jump at such a rhetorical opening, tyrannically exploiting the opportunity for cascade. And there will always be others who sow seeds of faction with the destructive aim of splitting the group against itself. The first indulge in an unrestrained quest for unity; the second in an unrestrained quest for division. We tend to respond to the first problem—the dangers of demagoguery—by emphasizing the value of critical distance, and the second—the dangers of sectarianism—by saying that we need to learn to get along. But it’s only if we face both problems at the same time that we see how serious our predicament is.
Whenever people must decide something together—in classrooms and city council offices, in board meetings and on the internet—we find freelance devil’s advocates doing a poor job of it. The asymmetrical structure of these informal contradictoria is fragile, quick to collapse into bad forms of agreement and bad forms of disagreement. Often these come together: oratorically speaking, the cheapest forms of unity are often those purchased by vilifying some subgroup into outsider status.
Indulgence in excess is so easy, self-restraint is so hard. If social media exaggerates these problems, that is also a kind of virtue: perhaps we have never been able to see, so clearly, what it looks like when we all try to get along. It’s not pretty. You can leave Facebook or Twitter, but you can’t really leave the ugliness behind. Technical skill is essential to informing our judgments, but it cannot be the ultimate ground of our getting along with one another. We cannot ask the authorities, experts or science to do all our thinking for us. Sometimes, we need to think as a group, and that means we cannot afford to cynically dismiss “devil’s advocacy” as a term of opprobrium. It has to become an honorific.
Image credit: xkcd
If you liked this essay, you’ll love reading The Point in print.