Flood and Dresher is the wrong reference here. Their experiment didn't show anything about expected utility one way or the other. The important challenge to EU, also around 1950 was Allais.
And there's lots of subsequent work (including by me and by Allais) developing generalizations of EU that are more consistent with the empirical evidence.
Allais, M. (1953) Le Comportement de l’Homme Rationnel devant le Risque: Critique des Postulats et Axiomes de l’Ecole Americaine. Econometrica, 21, 503-546, http://www.jstor.org/stable/10.2307/1907921.
probably the last major paper in econ to be written in French
My generalisation in Quiggin, J. (1982) A theory of anticipated utility. Journal of Economic Behavior and Organization, 3, 323-343, http://ideas.repec.org/a/eee/jeborg/v3y1982i4p323-343.html. is now generally referred to as rank-dependent utility, and is, I think, still the most popular
That was incorporated in Kahneman & Tversky's 1979 prospect theory to form cumulative prospect theory, for which Kahneman got the economics Nobel award and Tversky a rare posthumous mention
But after the thoughtfull plea for mathematical pluralism, I would, in David Deutsch spirit sum it up with: Thus we get nearer to the ever elusive Truth.
This is a good argument. Sometimes the obstacles to pluralism go the other way too.
Most people dislike mathematics and quantitative thinking -- I believe most are capable of it but just find it taxing and perhaps emotionally painful -- and some of them simply refuse to engage with quantitative arguments or models. Often they start with valuable critiques like those you bring up, but then conclude that all mathematical modeling is entirely worthless in these areas.
Should it be the responsibility of algorithms to minimize the harm, for say something like social media addiction? Are incentives of companies aligned with it?
Scientists are already losing their standing as authority in the current American political environment. The reality is that lot of us are better scientists than politicians, so when scientists try to be politician and make authoritative demands for policies, it will likely backfire even if our intentions are good and the outcome is highly likely to be beneficial for everybody involved regardless of political affiliation. That being said, learning to communicate what we know to the right people (and people in general) is always harder than just being the one that says “we should just do this”.
Flood and Dresher is the wrong reference here. Their experiment didn't show anything about expected utility one way or the other. The important challenge to EU, also around 1950 was Allais.
And there's lots of subsequent work (including by me and by Allais) developing generalizations of EU that are more consistent with the empirical evidence.
Send me references! I'd love to read them.
The classic critique of EU is
Allais, M. (1953) Le Comportement de l’Homme Rationnel devant le Risque: Critique des Postulats et Axiomes de l’Ecole Americaine. Econometrica, 21, 503-546, http://www.jstor.org/stable/10.2307/1907921.
probably the last major paper in econ to be written in French
My generalisation in Quiggin, J. (1982) A theory of anticipated utility. Journal of Economic Behavior and Organization, 3, 323-343, http://ideas.repec.org/a/eee/jeborg/v3y1982i4p323-343.html. is now generally referred to as rank-dependent utility, and is, I think, still the most popular
That was incorporated in Kahneman & Tversky's 1979 prospect theory to form cumulative prospect theory, for which Kahneman got the economics Nobel award and Tversky a rare posthumous mention
Thank you!
I don’t think we can say there are many truths, but we can say there are many lies. To be fair to the definition of truth, there can only be one.
Kudos!
But after the thoughtfull plea for mathematical pluralism, I would, in David Deutsch spirit sum it up with: Thus we get nearer to the ever elusive Truth.
This is a good argument. Sometimes the obstacles to pluralism go the other way too.
Most people dislike mathematics and quantitative thinking -- I believe most are capable of it but just find it taxing and perhaps emotionally painful -- and some of them simply refuse to engage with quantitative arguments or models. Often they start with valuable critiques like those you bring up, but then conclude that all mathematical modeling is entirely worthless in these areas.
> Algorithms cause harm
Should it be the responsibility of algorithms to minimize the harm, for say something like social media addiction? Are incentives of companies aligned with it?
Scientists are already losing their standing as authority in the current American political environment. The reality is that lot of us are better scientists than politicians, so when scientists try to be politician and make authoritative demands for policies, it will likely backfire even if our intentions are good and the outcome is highly likely to be beneficial for everybody involved regardless of political affiliation. That being said, learning to communicate what we know to the right people (and people in general) is always harder than just being the one that says “we should just do this”.