Thanks for this post! I was impressed by how you managed to make it substantive, nuanced, concise, and funny all at once.
There seem to be at least two major types of outside views / reference classes that one could bring to bear on this question:
Historically, what proportion of societies have been proper democracies?
Or related things like “At what proportion of years since civilization emerged has proper democracy been as common as it is today”, or “What has been the average proportion of people living in proper democracy since civilizatin emerged?”
Historically, how many trends that lasted a couple centuries stuck around?
Or related things like “Historically, what is the average length of time for which a trend lasts, given that it’s already lasted a couple centuries?”
You argue convincingly that the first type of reference class bodes poorly for democracy. But regarding the second type of reference class, you seem to only say:
We know, though, that centuries-long social trends often reverse themselves.
But you don’t give examples or statistics, and “often” is consistent with (for example) “5% of a very large set, thus totally tens of instances”. And I know of at least some other social trends that lasted a couple centuries and then much longer, including lasting to the present day (I mainly have in mind things discussed in Henrich’s WEIRDest People in the World).
So I feel like it’d make sense to look closer at that second type of reference class. I also feel like, until we do so, it might be premature to predict that “[one thousand years from now, given certain specified conditions] there is something like a 4-in-5 chance that the portion of people living under a proper democracy [would] be substantially lower than it is today”. (To be clear, I mean that it might make sense to stick a bit closer to 50⁄50 until we look closer at the second type of reference class; I’m not saying it was premature of you to make any quantitative prediction, and I appreciate you doing so.)
This also makes me think that it might be very useful for someone to:
Compile a dataset of “civilization-related trends” (could be social, moral, economic, political, etc., just not things like biological or physical trends)
Classify them in ways that helps us get a sense of how relevant one trend is to another
Thereby come up with base rates for the persistence of certain types of trends
At first glance, I’d guess that:
Any random EA should be able to do a better-than-nothing first pass at this in just a week or so
That first pass might already be quite useful
It could then be refined and expanded over time
Does anyone know if something like this has been tried? Would it be tractable and useful? I guess maybe there’s some relevant work in the literature on cultural persistence / persistence studies (I haven’t looked into that myself)?
Hi Michael, I think this is a great comment! I would be really interested in a rough ‘civilizational trends database’ or anything that could help clarify what a sensible prior for social trend persistence would be.
I’m not exactly sure how this would work, but one trick might be to pick a few well-document times/regions in world history and try to log trends that historians think are worth remarking on. For example, for the late Roman Empire, the ‘religious trends’ subset of the database would include both the rise of Christianity (ultra-robust) and the rise of Sol Invictus worship (not nearly as robust). Although, especially for older periods, shorter-lived trends might be systematically under-discussed/under-recorded.
Thanks for this post! I was impressed by how you managed to make it substantive, nuanced, concise, and funny all at once.
There seem to be at least two major types of outside views / reference classes that one could bring to bear on this question:
Historically, what proportion of societies have been proper democracies?
Or related things like “At what proportion of years since civilization emerged has proper democracy been as common as it is today”, or “What has been the average proportion of people living in proper democracy since civilizatin emerged?”
Historically, how many trends that lasted a couple centuries stuck around?
Or related things like “Historically, what is the average length of time for which a trend lasts, given that it’s already lasted a couple centuries?”
You argue convincingly that the first type of reference class bodes poorly for democracy. But regarding the second type of reference class, you seem to only say:
But you don’t give examples or statistics, and “often” is consistent with (for example) “5% of a very large set, thus totally tens of instances”. And I know of at least some other social trends that lasted a couple centuries and then much longer, including lasting to the present day (I mainly have in mind things discussed in Henrich’s WEIRDest People in the World).
So I feel like it’d make sense to look closer at that second type of reference class. I also feel like, until we do so, it might be premature to predict that “[one thousand years from now, given certain specified conditions] there is something like a 4-in-5 chance that the portion of people living under a proper democracy [would] be substantially lower than it is today”. (To be clear, I mean that it might make sense to stick a bit closer to 50⁄50 until we look closer at the second type of reference class; I’m not saying it was premature of you to make any quantitative prediction, and I appreciate you doing so.)
This also makes me think that it might be very useful for someone to:
Compile a dataset of “civilization-related trends” (could be social, moral, economic, political, etc., just not things like biological or physical trends)
Classify them in ways that helps us get a sense of how relevant one trend is to another
Thereby come up with base rates for the persistence of certain types of trends
At first glance, I’d guess that:
Any random EA should be able to do a better-than-nothing first pass at this in just a week or so
That first pass might already be quite useful
It could then be refined and expanded over time
Does anyone know if something like this has been tried? Would it be tractable and useful? I guess maybe there’s some relevant work in the literature on cultural persistence / persistence studies (I haven’t looked into that myself)?
Hi Michael, I think this is a great comment! I would be really interested in a rough ‘civilizational trends database’ or anything that could help clarify what a sensible prior for social trend persistence would be.
I’m not exactly sure how this would work, but one trick might be to pick a few well-document times/regions in world history and try to log trends that historians think are worth remarking on. For example, for the late Roman Empire, the ‘religious trends’ subset of the database would include both the rise of Christianity (ultra-robust) and the rise of Sol Invictus worship (not nearly as robust). Although, especially for older periods, shorter-lived trends might be systematically under-discussed/under-recorded.
I think the closest things we’ve got that’s similar to this are:
Luke Muehlhauser’s work on ‘amateur macrohistory’ https://lukemuehlhauser.com/industrial-revolution/
The (more academic) Peter Turchin’s Seshat database: http://seshatdatabank.info/