This book is difficult and I have many other things I could do with my time. So why am I bothering?
Well, one answer is 'I got nerdsniped and now I'm stuck here'. I think it goes a bit deeper though. It looks like I've been jumping around incomprehensibly between topics recently (visual programming, romanticism, Derrida...), but there is a weak thread running between them. I don't understand it too well yet, but it's something like this:
- Derrida has this interesting way of using the words 'writing' and 'speech', which I first learned about from Christopher Norris's talk Derrida and Oralcy: Grammatology Revisited. 'Writing' refers to the structured, repeatable element of language, and 'speech' to the meaningful, expressive element.
- Romanticism is obsessed with the 'speech' side of this dichotomy and ignores the 'writing' side. E.g. art should be a pure expression of the nature of the artist, rather than conforming to objective rules of 'good art'. Contextual detail is more important than grand theories. Order and discipline are out, uncontrolled outbursts of emotion are in. Etc etc etc.
- Derrida spends a bunch of time in Of Grammatology talking about Rousseau. Rousseau is more Enlightenment-era himself, but had a very Romantic-flavoured aesthetic and ended up being a big influence on the movement. Rousseau has this story about how music should be a pure outpouring of emotion and Derrida just completely takes his argument apart in a very satisfying way (the Norris talk explains this well).
- I think this stuff is compelling to me because I have a lot of sympathy for the romantic side of this argument, so when Derrida pulls it apart I can see how I've got myself into similar confusions to Rousseau in the past. (It probably helps that Derrida isn't just trashing the romantic side, he has a lot of sympathy for it himself. The Newton Garver intro to Speech and Phenomena makes the point that Derrida was reading Rousseau specifically because he takes the contextual side of language seriously, rather than reducing it to a logical grammatical structure.)
- Where this comes up for me is in understanding mathematics, rather than understanding language. Like many people I've got very frustrated with the awfulness of a lot of mathematics education and a lot of technical explanations in research (insert obligatory link to Research Debt here). A major part of this awfulness is an overemphasis on rote teaching the structural side - how to carry out the steps of a particular calculation - independent of any understanding of what it means or why you'd care.
- This is really annoying! Buuut... this annoyance can tip into a sort of pure romanticism where only the 'meaningful' component of mathematical thinking is taken seriously.
- Does this matter? Maybe this 'romantic' view is a useful enough corrective to the shut-up-and-calculate one that I should just put up with it. I think it is worth exploring where it goes wrong, though.
- Now finally I'll get to the visual programming link. Visual programming is interesting because it attracts people with a very strong 'romantic' orientation, who are frustrated with all this opaque useless symbol pushing. Bret Victor as obvious example. (Also got a sampling of a few of these people in the HN thread.) I think there are cases where this enthusiasm for 'keeping in the meaning' ends up running into tension with visual programming still needing to have some kind of structural elements, repeatable symbols and fixed meanings and so on. I was originally meaning to think about this when I did the HN folk wisdom post, but I got really bogged down in working through all the HN comments, and then it was so confusingly popular, and I completely lost track of what I was meaning to do with it.
Wow, this exercise was actually well worth doing. That's not really 'why do I care about Derrida?' so much 'why do I care about everything that I'm On About at the moment?' I think this is a genuinely interesting train of thought, but I need to get clearer on things. In particular, I don't have any very compelling argument right now for why the 'romantic view' has costs, just a few scattered ideas for places to dig for one. Maybe it just doesn't matter that much. I'm still interested in getting clearer on this stuff regardless, but it would be nice if this actually led somewhere useful.
A few more stray thoughts:
- Timothy Gowers is interesting as he has a really nuanced, later-Wittgenstein-inspired view that sort of threads between the extremes - cares about intuitive motivation but takes structure seriously. Here's an example. It could be worthwhile for me to summarise his view so that I understand it more clearly.
- Bret Victor has this old post in his 'Kill Math' series that's particularly uncompromising in its romantic orientation, and advocates for pretty much getting rid of symbolic manipulation altogether (he does soften it with a bunch of qualifiers at the end). There's a good response to it by Evan Miller called, inevitably, Don't Kill Math. I started making some notes on this pair, would be good to go back to it.
- Something about 'explorable explanations'. It can be hard to 'keep the meaning in'... the topic is important to the person making the explanation, but not enough context gets translated to the person interacting with it so it becomes a bit of a lifeless unmotivated exercise in fiddling with parameters and watching things change on the screen. A bit like when you see some 'interactive exhibit' at the museum as a kid and just mindlessly smash all the buttons. Not sure exactly where this fits in but it's related.