At the moment I'm having a problem with Popper's critique of Historicism :

My understanding of science goes like this :

  • I conjecture that there are certain types of things in the world : eg. massive bodies.
  • I conjecture laws appertaining to them, eg. Force = mass times acceleration

Both of these conjectures are "blind" ie. there is no logic to guide me as to what I should conjecture.

They do have deductive consequences, and if I find observations that are incompatible with them, I'm obliged to revise my original conjectures.

Now this I'm happy with. But suppose instead of conjecturing types and laws such as "massive bodies" I conjecture a "trend" eg. "popular music is getting better all the time"

It also has deductive consequences which can falsify it : bad records.

If the nature of rationality is blind conjecture and willingness to reject beliefs when confronted by contradictory evidence, doesn't the trend case fit just as well as the science case? Shouldn't it be equally rational to structure one's set of beliefs in terms of unfalsified predictions based on conjectured trends, as unfalsified predictions based on conjectured types or laws?

Of course, one of the reasons that I, as a Popperian, am questioning the critique of historicism is because I'm also considering the status of sociological theories taking the form FeudalismCapitalismInformationalism

The difference between "trend" and "law" is not due to any particularly inherent "level" of shared subjectivity in either, but rather the specificity of the definition (which may lead to the former). A "trend" is surely, by its (man-defined) nature, an overview of a particular domain, a summarisation of what could be considered to be a series of minute "laws", inasmuch as a law is effectively subjective.

For example, you can conjecture that you do or don't like a particular song. You can also conjecture, or rather define a "genre" that that song fits into. This genre consists of many songs, and as such loses it specificity. You can dislike a song of that genre, yet still like the genre as a whole. Such is the relation between "law" and "trend". If suddenly the majority of songs coming out within that genre you disliked, then either you would tend towards a dislike of the genre (a "trend" being disputed through quantitative opposing "laws"), or you would further specify your "genre" to distance those songs that you do like from those that you don't.

Questions arising...

  1. What about conjectures that change without any basis? To continue the song theme, is there a rational reason why you may dislike a song on first hearing, yet may grow to like it after several days? Should this affect your opinion of other songs? Or is each conjecture its own instance, subject to its own rationality, despite sharing the same domain as another?
  1. If a "trend" can be specified further into distinct "laws", and the former is effectively subjectively similar to the latter, then how can you ascertain the constituent specifics of a "law"? Is a "law" simply something that has no opposing conjecture, while a "trend" takes into account both (or more)?

Apologies if these are obvious answers... it's now 1am and it's been a long day and the fairy cows are singing at me.

[2003-07-16 01:00 - scribe]

See also :

*(WarpLink) TrendWatching