I follow Paul Graham’s writings, because he has written some very influential works (e.g. Hackers & Painters*, among others). He has a way of, in his essays at least, writing very clearly and very well. In particular, I like how he is able to classify and organise things, and then present it as if that were the most trivial and obvious approach possible. This is very much like API design, something I tend to be doing perpetually (Keychain framework’s five year anniversary of release is coming up on April 14th!).
Anyway, his most recent is “How to Disagree“, which simply describes a hierarchy of disagreement modes, from name-calling (including the cleverly noted point on pretentious name calling, which is all too commonly accepted in engineering) up to actual refutation of the central point.
While it’s very true – and an important point to make – that most arguments never get higher than DH3 (Contradiction), if you’re lucky to get that far, he only glances across one point which I think is most critical; that most often when two people are arguing about a technical issue, they’re actually arguing different things. This happens all the time at work; it’s a trait seemingly genetically ingrained in engineers. I suspect it’s a subversive defensive mechanism, conscious or not – you can’t provide any direction refutation or counter-argument against the argument as presented, so you argue about something related that you can plausibly win, and hope no-one notices. It’s hard to believe, otherwise, that intelligent people would so conveniently and naively misunderstand each other, consistently.
The movie Thank You For Smoking presents some good examples of this; it is, after all, basically about how to argue against anti-smoking lobbyists without actually arguing for smoking.
Also, I happened to reread an earlier post on Altruism, which ends with:
The third option is pretty straight forward, but can be disproved in two ways. Firstly, observe that the two other alternatives are more explainable and more logical by their own semi-proofs. That’s not enough, though. The clincher is…
…and that’ll learn me to stop in the middle of something and not resume it ’till several days later. I have no idea what I was about to say. Hopefully it’ll come back to me at some point.
I’m still hanging out for the ending. Gah! :)
* = Notable in that it is the first such one I read, many many years ago…. five years, in fact; I still have the original web page saved in ~/Documents, dated 8th of May, 2003.