Another old philosophy paper. The other one is from 2009.
I wrote this in 2008 about the “Cranes of Culture” chapter in Daniel Dennett’s Darwin’s Dangerous Idea. I was interested in the claim that ideas can act like replicators. They spread, compete for attention, and survive for reasons that may have little to do with whether they are true or good for anyone.
That part still holds. Successful ideas are selected for replication, not truth. It is easier to see now, after years of algorithmic recommendation, than it was when I wrote the paper.
The paper itself is too faithful to Dennett. I explain his argument at length and object at the edges. The difficult question is whether a meme can be identified clearly enough to do useful analytical work. I barely address it. I had not yet learned the difference between explaining a position competently and deciding what I thought about it.
Dawkins’s example of a bestseller list increasing the sales of books already on the list now sounds like a small model of recommendation algorithms. Dennett’s claim that minds are partly shaped by the memes that reach them also invites a comparison with language models and their training data. I am not sure whether that comparison explains anything or merely extends an already loose concept. The distinction between what spreads and what is true still interests me. It appears again in Digital priesthood and Interface friction.