I’ve got some further thoughts on Underwood et all, You say you found a revolution, over at The Stone and the Shell.
What is time?
What I think is that the questions raised ultimately go very deep. One place they end up is the nature of time and thus of history. One view (and I believe it has a name) is that time is just an ‘empty’ framework in which things happen. But another view (again with a name) sees time as arise from/with causal processes at work in the world. On this view the emergence of long ¬ whatever that means – unidirectional chains of events might reflect a new causal order.
And isn’t that, after all, how we (for some unspecified value of “we”) think of the emergence of humankind and culture, that it marks the emergence of ontologically new phenomena in the universe?
It’s very easy, too easy, to think of the data as just, well, you know, data. We just collect the ‘texts’, whether literature or music or whatever, work on them, and draw conclusions. We know, of course, that the data is evidence about a historical process – that’s why it interests us in the first place – but we don’t necessarily think ‘through’ the data to the underlying causal mechanisms. The interesting thing about what Jockers did is that, in effect, by examining similarity he was in some sense able to examine the dynamics of the system without having any explicit temporal information in his data set. Thus the fact that his analysis produced a result that maps onto time quite neatly is interesting and revealing.
It tells us something quite profound about those dynamics. But what? (See the next section.)
Mauch et al. didn’t undertake that kind of analysis. They’ve got temporal information in their data. But there’s no reason one couldn’t analyze their data in the way Jockers did his. They started with c. 30 features divided between harmony and timbre. Why not deal with them in the same way Jockers dealt with his novels data? Temporal direction should pop out of that analysis the way it did out of Jockers’.
And then, lurking around the corner, we’ve got the specter of Whig history. I understand that Underwood et al. do not in fact believe that history is unidirectional; their remarks just reflect what seems to be in the data they’re examining (Underwood communicated this in private email). The Whig interpretation of history, however, assumes that history has a direction.
But that, of course, is not all it assumes. It also assumes that history is moving from a primitive state to a more advanced state, one that is better (in every way) than earlier states. Of course, Underwood et al, neither said nor implied anything about cultural progress. I’m just pointing out that, the moment we assert that history is unidirectional, Whig history is laying in wait for us.
While Whig history is suspect as a privileged reading of the historical record, we should remind ourselves that it is not a completely delusional reading of that record. There does seem to be a direction to (at least certain ranges of) historical phenomena. How do we account for that?
That, I fear, is a deep question, and it isn’t going to be answered anytime soon.