3 ms·
I wish it was that clear cut. This is a tricky business in the sciences. Often, the quantity you are trying to validate is unobservable. In this case, I suspect
by arcanus 11y ago
I wish it was that clear cut. This is a tricky business in the sciences. Often, the quantity you are trying to validate is unobservable. In this case, I suspect (I'm not a squishy scientist) it is because evolution has such long time scales.
In other fields, it can be simply because it is prohibitively difficult or expensive to instrument the quantity of interest. This is extremely common, even in the 'hard' scientific fields and engineering. You very, very seldom have all the data you want.
This is all a long winded way of saying that confirmation of a hypothesis by mathematical models certainly is a mechanism by which science creates a stronger degree of belief in said hypothesis. Would it be a full fledged validation of the model? Probably not. But it is absolutely useful evidence.
Full disclosure, I am a computation scientist.
- dekhn 11y agoI'm a computational scientist too, but those mathematical models don't provide confirmation. It certainly can help build evidence, and when a model starts to make predictions outside of the model, then we really can a lot of confidence in it. But a model is just a model, and unless the universe is truly mathematical in structure (not a completely crazy idea), and our models match that structure, then they can't really confirm anything.
- arcanus 11y agoAll good points, and I think we are entering into philosophy around what we mean by 'confirmation'. I believe that models are certainly wrong. I personally doubt we have any models that actually reproduce the underlying mathematical structure of the universe (if it even exists, as you noted). ...But some models are useful. And when viewed from that perspective, a mathematical model can provide evidence that a hypothesis accurately explains an observed phenomenon, given a certain set of assumptions. I view validation of a model hierarchically, where a untested hypothesis gains credibility by increasing tests of the validity of its predictions. Direct experimental confirmation is the strongest, but this is almost always impossible, and so weaker forms of testing are necessary. Computational modeling certainly provides a weaker form of confirmation, from this perspective.
- Double_Cast 11y agoI think Popper's concept of Falsificationism is best put by Asimov in his essay The Relativity of Wrong [0]. The thesis is that Science never has the perfect model. Instead of directly arriving at The Truth, it continuously refines imperfect models by reducing error. In this view, the difference between "confirm" and "provide evidence for" is simply a matter of degree. [0] http://chem.tufts.edu/AnswersInScience/RelativityofWrong.htm http://chem.tufts.edu/AnswersInScience/RelativityofWrong.htm