I like it, but as a strategy it sounds very delicate and is highly reliant on the highest levels of the organisation being thoughtful. Having discovered and fallen in love with Robert Kegan’s adult development model recently, I am now much more keenly aware of just how much large groups of people struggle with separating ideas from the people who advocate for them. IMO it'd only take one or two people in key positions to bring down institutionalised devils advocacy. And a lot of people - potentially in leadership positions - legitimately won't understand how having a person advocate for non-consensus ideas helps because they don't judge correctness by argument but by the number and status of the people advocating it.

The technique in the article would protect against groupthink, but I suspect the people most likely to be taken in by groupthink won't understand it and will keep making decisions based on which position has the most advocates. The article notes that advocatus diaboli of the Catholics was more akin to a muckraker looking for unknown facts than operating in a genuine position of advocacy or steelmanned argument - that'd be a more stable role in the long term but that isn't a defence against groupthink. Groupthink is fact-resistant.

Reminds me of an investing strategy I came across on Wikipedia: https://en.wikipedia.org/wiki/Contrarian_investing

> In "The Art of Contrary Thinking" (1954) by Humphrey B. Neill; considered influential by some in contrarian thinking, he notes it is easy to find something to go contrary to, but difficult to discover when everybody believes it.[2] He concludes "when everybody thinks alike, everybody is likely to be wrong."

> He concludes "when everybody thinks alike, everybody is likely to be wrong."

This is very dangerous. It's the same as saying "I am smarter than all those people who agree on X". You'd better be right.

To me, "devil's advocacy" has a negative connotation, for reasons mentioned in the "Challenges" section.

The "advocates" themselves can be just as prone to groupthink as anybody else. All of the mistakes made by a group can be made by a single person -- especially if that person is a member of a group outside the organization. All of the failure modes of reasoning apply, especially if the person is motivated by their "advocacy" position to zealously defend their idea. When that happens, you get stuck trying to please the advocate rather than actually solving the problem.

This can be handled within an organization, where the members themselves have faith in each other to be working for the common good. But I often hear from self-appointed "devil's advocates", who sometimes seem quite aware that they are wasting your time, and insisting that you are guilty of groupthink when you dismiss them.

To me, the accusation of "groupthink" has become a euphemism for "No matter what happens I'm right and everybody else is wrong. Everybody else is being irrational, which I never am. I play the devil's advocate, but I never need to listen to a devil's advocate myself."

That's not to say that "groupthink" doesn't actually happen. Quite the opposite: I tend to view the accusation of groupthink as a confession. And when it genuinely happens it's hard to see around. Adding a devil's advocate often fails to solve the problem, and can even make it worse.

In other words, there is no easy answer to making the right decisions. No procedure can solve the fundamental human problems.

Does anyone have any experience actually applying this idea? Devil’s advocates often seem to engage in a lot of rhetorical sleight-of-hand in the first place. I’d worry that a social convention of always having one would somehow cause even more argument gamesmanship.

Authentic disagreements are better, I think. Maybe inauthentic ones are better than nothing, though.

This has been my role on virtually every dev team ive been on. I don’t necessarily disagree but i do wait until most have discussed a new feature, etc, ask myself, “does this make sense in every way” and if not, I ask questions until it does, or in a lot of cases we reach understanding that it doesn’t make a ton of sense. The types of questions I ask are often subjective but disagreeing as well, like, “do we need more complexity?”. Because they are questions I avoid the disagreement stigma, ala Socrates

It's a good idea, second only to authentic disagreement, but if management does not understand or support it, you will get in a lot of trouble for being the one who disagrees. Odds are high that management seeks 100% agreement 100% of the time.

It's a catchy name, but it really comes down to, well, not drinking your own Kool Aid. :->

The project and risk management literature have plentiful examples of this sort of thinking, including performing post mortems now, before starting anything: "Assume we are 2 years out and everything failed: What went wrong? How did we fail? Why?"

One of the things I like best about my current position is that the colleague I work most closely with (he's head of development, I'm a senior technical contributor sort of thing (it varies by project/week/day/need)) do this all the time, pretty much whenever one of us proposes a change or a new feature.

We listen to the other, make sure we understand, then we critique, find flaws, consider gotchas, but reasonably, not pathologically.

It's at the point where we will each sometimes start critiquing our own ideas, with the other defending.

It helps our work immensely.

I'd place the devil's advocate into the superset of "idea generation" or "brainstorming" in general. It's only diabolical in that the idea has properties that make the modal person tend to resist and deny it. But really, it's just one possibility of many. The fact itself that it's unpopular doesn't make it more or less likely. The clever thing about brainstorming is that by thinking of a new possibility, you almost always lower the probability of the "obvious" path. This is true even if the new possibility is very unlikely. In other words, brainstorming tends to increase overall uncertainty and protects you from overconfidence.

I remember this being the only part of that World War Z movie that I thought was interesting.

(I enjoyed the book quite a bit for its variety of stories and passion for zombies during a time when zombies were in a bit of a lull.)

These strategies are employed by natural selection, immune systems, etc. indirectly through random mutations, which gives more chances of survival when the "unthinkable" happens. I always thought this explains why there's people out there who are allergic to the weirdest things.