4 ms·

True, but we've come to seek more than just predictive models, we want understanding and "truth".

lmm 11y ago
I'd say the opposite; we've come to realise that accurate predictions are the truest benchmark and notions of understanding can often be misleading. But supposing you're right, so what? Unless you'd prefer a more understandable but inaccurate model, there's still no better alternative to dark matter or dark energy.
- TheOtherHobbes 11y ago
  You need deep understanding to develop the next set of accurate models and - perhaps - turn them into useful technology.
  - lmm 11y ago
    Agreed, but I don't think going back to theories that don't explain all current observations is a fruitful path there.
daveguy 11y ago
Until you have a better predictive model there's no way to tell what you had before wasn't the "truth".
- yxhuvud 11y ago
  Yes there is. Models can be disproved by empiricism without needing another model. However, even if a model doesn't work all the time, it doesn't mean it is useless for prediction. Newtons laws is still in use, for example.
  - Retra 11y ago
    Newton's laws weren't disproved, they were just shown to be less general than originally assumed. Quantum theories usually assume Newton's laws to be true axiomatically, otherwise the quantum theory would be under-specified and practically useless.
  - darkmighty 11y ago
    He means the most common definition of truth is 'The best predictive model' (in the scope of the problem). I would add the principle of parsimony, truth is 'The simplest available formulation of the best predictive model'. In both cases you cannot know that no better predictive models exists (and hence if your truth is absolute), even if you experimental error is zero (which is essentially impossible in most sciences unless you grossly overfit), since there may be unobserved phenomena as well.
bweitzman 11y ago
Truth is just a sufficiently correct predictive model

lmm 11y ago

I'd say the opposite; we've come to realise that accurate predictions are the truest benchmark and notions of understanding can often be misleading. But supposing you're right, so what? Unless you'd prefer a more understandable but inaccurate model, there's still no better alternative to dark matter or dark energy.

TheOtherHobbes 11y ago
You need deep understanding to develop the next set of accurate models and - perhaps - turn them into useful technology.
- lmm 11y ago
  Agreed, but I don't think going back to theories that don't explain all current observations is a fruitful path there.

daveguy 11y ago

Until you have a better predictive model there's no way to tell what you had before wasn't the "truth".

yxhuvud 11y ago
Yes there is. Models can be disproved by empiricism without needing another model. However, even if a model doesn't work all the time, it doesn't mean it is useless for prediction. Newtons laws is still in use, for example.
- Retra 11y ago
  Newton's laws weren't disproved, they were just shown to be less general than originally assumed. Quantum theories usually assume Newton's laws to be true axiomatically, otherwise the quantum theory would be under-specified and practically useless.
- darkmighty 11y ago
  He means the most common definition of truth is 'The best predictive model' (in the scope of the problem). I would add the principle of parsimony, truth is 'The simplest available formulation of the best predictive model'. In both cases you cannot know that no better predictive models exists (and hence if your truth is absolute), even if you experimental error is zero (which is essentially impossible in most sciences unless you grossly overfit), since there may be unobserved phenomena as well.

bweitzman 11y ago

Truth is just a sufficiently correct predictive model