This is really interesting, but I feel like you've done a good job of outlining the shortcomings of the current approach but not really considered how an alternative would work in practice. Like, say we want to encourage novelty. How do we distinguish between novel ideas that are actually right, and ideas that are wrong, or are correct but insignificant, or... . And how do we do that on a timescale consistent with the only reward we currently have, which is tenure. All of which is to say, I agree that what we have isn't great but you can see why we iterated towards that equilibrium, and what does a different equilibrium look like that recognizes novelty of significance but doesn't just reward novelty for its own sake? Because if you just reward novelty for its own sake you run the risk of researchers just basically shooting the shit, and you lose the value from incremental contributions that we're currently getting.
You may enjoy 3 slim volumes by Nicholas Rescher
The Limits of Science (1984, 1999) (references Gerald Holton)
Nature and Understanding (2000)
Epistemetrics (2006)
Really interesting, thanks for writing!
This is really interesting, but I feel like you've done a good job of outlining the shortcomings of the current approach but not really considered how an alternative would work in practice. Like, say we want to encourage novelty. How do we distinguish between novel ideas that are actually right, and ideas that are wrong, or are correct but insignificant, or... . And how do we do that on a timescale consistent with the only reward we currently have, which is tenure. All of which is to say, I agree that what we have isn't great but you can see why we iterated towards that equilibrium, and what does a different equilibrium look like that recognizes novelty of significance but doesn't just reward novelty for its own sake? Because if you just reward novelty for its own sake you run the risk of researchers just basically shooting the shit, and you lose the value from incremental contributions that we're currently getting.
The branching thing is known in economics as innovation S-curves.