User:Geo Swan/the lessons of "tit for tat"?

Is there a lesson, for wikipedians, from Anatol Rappaport's "tit for tat" and "tit for two tats"?

Anatol Rappaport was a mathematician who speciailized in strategy and game theory. He had written about applying game theory to the cold war. Late in his career he was one of the participants invited to submit a computer program to a contest where entrants different computer programs were to compete in a version of "The Prisoner's Dilemma".

"The Prisoner's Dilemma" is the name of a classic scenario from game theory, where guilty suspects are faced with a choice. They know that if neither they, not their confederate confesses, and denounces the other they will each get a relatively light sentence, but the Prosecution has told him or her that if they confess, and denounce the other suspects, they will be rewarded with a lighter sentence in return for providing the evidence that will allow leveling a heavy sentence against their confederate.

Reduced to its simplest, it is a game where both players have just one move, to choose to either "cooperate" or "defect".

In the contest Rappaport and the other entrants were to submit programs that would each compete against all the other programs, one at a time, over and over again. The competing programs could each be as complicated and sophisticated as the entrants wanted. They were allowed to keep track of the previous results with each of the other competing programs, and use those results to predict, anticipate, and out-maneuver their opponents.

Anatol Rappaport's program, called "tit for tat", was, by far, the simplest entrant. And the first time the contest was run, his program got the overall highest score.

"Tit for tat" is an old-fashioned term for a single instance of responding in kind to a perceived slight.

His unsophisticated program did not try to track the long term strategy of its opponents. Instead, it would always make the cooperative move, and keep always making a cooperative move, unless that particular opponent had made the defecting move the previous round. It initially cooperated, instantly retaliated against defections, and instantly forgave a defecting program, if it started to cooperate again.

Tit for two tats was a similar program Rappaport entered when the contest was run a second time, with slightly altered rules. Sorry, I can't remember how the rules changed, but in the second contest his program would forgive one defection before it would start retaliating.

Is there a lesson, for wikipedians, from Anatol Rappaport's "tit for tat" and "tit for two tats"?

I think so.

As I understand our policies, guidelines and long-standing conventions, we are always supposed to confine our comments to the level of editorial issues. As I understand our policies retaliating against other contributors for perceived slights, and personal attacks, is very strong discouraged. As I understand it, we are strongly encouraged to take a wikibreak, or to call for the opinion(s) of uninvolved third parties, or to use other mechanisms like requests for comment, when we perceive another contributor is defecting from the civility policies and conventions we are all supposed to follow.

To repeat, we are never supposed to retaliate, to respond in kind, to perceived defectins from other contributors, but it is my impression that almost none of us immediately disengage, and use the existing mechanisms we have in place for dealing with difficult individuals.

Why? My top two guesses for this are that:

1. almost none of us has any experience in interacting under rules where we are never supposed to respond in kind;
2. the mechanisms we have in place, evolved, ad hoc, and are often ineffective, or only effective when the limited number of truly unrepentant individuals who can't cooperate have left a very considerable trail of destruction.

I have worked on a lot of artcies on controversial topics. And, although I have made a special effort to work on those articles in a way that is completely compliant with NPOV and all our other policies, the controversial nature of those topics has attracted more than my share of kooks to my activities.

I noticed, in my first years on the wikipedia, that a surprisingly number of the other contributors who turned out to be routinely uncivil to me, who seemed to be routinely wikilawyering, in a deceitful manner, and who then moved on to single out other victims, turned out to be sockpuppets.

While there is a certain satisfaction to seeing long term vandals, and sockpuppets get permanently blocked, that satisfaction is considerably lessened when that block is months or even years after they first revealed their malice to you.

Would our problems with uncivil contributors get better if we all made more of an effort to refrain from responding in kind to perceived slights? Yes, I definitely think it would.

There is a general phenomenon I have seen, both at a distance, and as a participant, where a wikipedian thinks they have been personally attacked, and decides to respond in kind, when I think any fair-minded third-party observe would agree that they had not been personally attacked, and thus what they perceived to be merely an allowable response in kind, was actually the opening salvo in a discussion that got more and more personal and less and less productive.

Mahatma Ghandi is supposed to have said the strategy of "'an eye for an eye' leaves the whole world blind."

I think our general problem with incivility would improve if we were all on watch for instances when we saw the opening salvoes in a bad interaction, and spoke up to request a stronger effort to comply with our civility policies.

I'd like to see all of our administrators review their own commitments and compliance with both the spirit and the letter of all our civility policies, guidelines and conventions. Less experienced contributors look to our administrators for an example of the kinds of behavior that is acceptable here. And, very unfortunately, while some administrators are wise, patient, thoughtful, polite individuals, who are capable of approaching each question without prejudice, our corps of administrators also includes some individuals who act as if being entrusted with administrator authority allows them to ignore all our civility conventions.

What can we do to enhance the effectiveness of our mechanisms for calling for help when contributors perceive personal attacks, and choose to forgo responding in kind?

I don't know.

epilog

It turns out that the wikipedia has an excellent article on tit for tat.

To return to the lessons, for us, from tit for tat... concerning its apparent lessons for when to retaliate? I think it would be best if none of us ever retaliated, due to what our tit for tat article called the "death spiral". Additionally, "tit for tat" was designed for interactions between exactly two parties, and we are interacting in a community of hundreds of thousands of individuals.

But I believe the exercise does have a lesson for us -- forgiveness, second chances. Everyone who we think has lapsed from our policies is not a long term committed vandal. The "death spiral" is unfortunately pretty common, and the individual we think "started it", often honestly thinks they are merely responding in kind to an earlier provocation.