Falkenblog

In college debate there's a strategy called 'speed and spread', where a speaker speed reads the broadest number of arguments possible in their allotted time. Watching high school and college debates use this, it seems absurd. The objective is to present the maximum number of arguments in a debate round, hoping that a few unrebutted points seem like concessions to judges. Quality and consistency, are not priorities. I suppose in debate the judges merely read the arguments, because it seems impossible to process these statements (see below).



I thought this was silly until I was in court. I was being accused of appropriating some unspecified intellectual property, and several intimations were made in the original complaint but they were not exhaustive. The problem is that litigation is very expensive (IP lawyers are $500/hour), and there's an opportunity cost because as long as you are accused of appropriating undescribed IP, no one with any money would dare hire or work with you. So, you want to get it done. However, they can say 'you took the idea of investing in low volatility stocks', and by the time you mention you did this at a prior job, they then say '...in a particular way', and add on several other items, such as the concept of 'mean variance optimization', etc. You can knock these down one by one, but the longer you play the more you lose.

The judge, meanwhile, looks at every allegation piecemeal, so that they aren't penalized for making absurd allegations, they just hope at least one remains viable. My lawyers then leveled some counter-arguments I found were irrelevant, but they noted you never know what arguments work, that is, which might resonate with the judge. Restricting one's counter-suite was suboptimal. Real life was an absurd spread debate I had to play.

I was reminded of this watching this testimony on how to reform the governmental housing system, and one Janneke Ratcliffe made several remarkable assertions, including that low-down payment mortgages on some subset of loans did above average over the past decade, implying to her that low down payments were not a problem. I'm sure there is such a subset of loans, but in the context of the fact that default rates are multidimensional this is unsurprising--you can find subets of fat men who are actually quite healthy, but on average at the margin obesity is not healthy. The point is, she made an assertion to some atypical finding, presented it as a fact that was dispositive of her assertion, and it sat out there as evidence because no one could, in the time presented, review and rebut it.

If you ever watch a speech by Noam Chomksy you'll note his discussion crosses political science, economics, sociology, history, and mentions several disparate facts in each of these domains. The key to being persuasive seems to be to have an argument that is irrefutable, because no fact is essential, and they are so disparate facts no one can refute a significant proportion of them definitively. That is, appearing to have a fusillade of facts is overwhelming. So, the big issue (eg, anarcho-syndicalism? Capitalism is evil?) is promoted via thousands of supporting facts. You don't have enough time, and no one has the breadth, to evaluate all of them, and merely refuting a finite subset leaves open the logical possibility he's correct on his big point.

If you look at debates about taxes, the stochastic discount factor, global warming, or whether the moon landings were faked, you'll find yourself unable to counter some point that helps the opposing view, reality has too many dimensions. The paper Fact-Free Learning explains this pretty well (AER, Dec 2005). The idea is that most 'facts' are really statements about relations between datapoints. As there are an infinite number of relational facts, you can’t expect to know them all even if you know all the basic data points, so certainly you can't easily rebut them all. So everyone knows only a different subset of the relations, which gives them different ‘facts’. There's no simple way to reconcile them, because there are so man facts, each supportive in their own way.

Any expert merely knows a lot of facts relevant to his prejudices, and it's not hard to make a case for anything, which is why a good lawyer always helps. It also explains why people don't converge on the truth, because once you are a valued advocate for X, the payoffs from others who like X are surely better than recognizing that X is, in fact, false (plus, presumably after so many years most of your friends and colleagues are X-lovers).