I got this, just as I was sitting down to work on a paper about syntactic and semantic approaches to uncertainty and unawareness. Here's a link to a current draft

Not Meehl related, but translation as a foundational building block of meaning has been a recent obsession of mine. Adding your paper to my queue. Thanks!

This paper takes me back -- all the ideas related to connecting lattice-theoretic and probabilistic notions of uncertainty, plus the references to Halpern, Gilboa-Schmeidler, etc.

So when there's stuff like the AI Impacts survey asking people to estimate probabilities, and then they do statistics over those probabilities, is that doing type 2 over type 1? Is there a name for this... technique 🤔

By the way, at one point Meehl recommends two chapters from Vic Barnett’s book on comparative statistical inference as a good reference on the interpretations of probability. Check them out if you haven’t already, they are excellent.

Yeah, Barnett's book is excellent cover to cover. Have you read Hacking's "Logic of Statistical Inference?" Also great at picking at the dappled nature of inference. And I also really like Wesley Salmon's discussion of the problem of probability in "The Foundations of Scientific Inference."

I've leafed through Hacking. Currently making my way through Mary Hesse's "Structure of Scientific Inference," which is (like Hesse in general) hugely underrated.

What I like about Barnett, Hacking, and Salmon is I walked away more confused but also more relaxed. The statistical dogmatists are much more stressful to read.

I got this, just as I was sitting down to work on a paper about syntactic and semantic approaches to uncertainty and unawareness. Here's a link to a current draft

https://www.dropbox.com/scl/fi/clso6ejrmpgg3f6eciip2/Translation-10.pdf?rlkey=h4pg70g04k86f5sxchm95ym93&dl=0

Not Meehl related, but translation as a foundational building block of meaning has been a recent obsession of mine. Adding your paper to my queue. Thanks!

This paper takes me back -- all the ideas related to connecting lattice-theoretic and probabilistic notions of uncertainty, plus the references to Halpern, Gilboa-Schmeidler, etc.

We’ve been working on this for a long time, and there are still basic issues unresolved. But I think we have made some progress here.

So when there's stuff like the AI Impacts survey asking people to estimate probabilities, and then they do statistics over those probabilities, is that doing type 2 over type 1? Is there a name for this... technique 🤔

I think it's called "Doomerism."

I guess Bayesians would say that both are equally fundamental to understanding the unknown.

Frequentists would agree! The two differ in how you move between the two. More on this topic coming later this week.

edited Jul 8As I was listening to that lecture, I kept nodding in agreement because I articulated more or less the same dichotomy here: https://realizable.substack.com/p/probabilities-coherence-correspondence.

By the way, at one point Meehl recommends two chapters from Vic Barnett’s book on comparative statistical inference as a good reference on the interpretations of probability. Check them out if you haven’t already, they are excellent.

Yeah, Barnett's book is excellent cover to cover. Have you read Hacking's "Logic of Statistical Inference?" Also great at picking at the dappled nature of inference. And I also really like Wesley Salmon's discussion of the problem of probability in "The Foundations of Scientific Inference."

edited Jul 8I've leafed through Hacking. Currently making my way through Mary Hesse's "Structure of Scientific Inference," which is (like Hesse in general) hugely underrated.

What I like about Barnett, Hacking, and Salmon is I walked away more confused but also more relaxed. The statistical dogmatists are much more stressful to read.