Originally Posted by kenkoskinen
It is trivial to construct examples where the scientific method is suboptimal. For example the "Occam's razor" feature requires that we choose the hypothesis which matches the data and has the least number of presuppositions (simplest). But obviously sometimes the simplest hypothesis is not the correct one.
Clearly "the scientific method" is not the only possible model building scheme. At least in some cases it is not the optimal model building scheme. It's almost unimaginable that it's the "best" in any meaningful sense.
If you think I said that we subjectively observe models, then you have misread it entirely. We subjectively observe experiments. We collect this subjective data and using the sci-method produce a model of the subjective observations. As you describe we continue to refine and revise the model - but that doesn't change the fact that we are modeling based on subjective observations and comparing the results of the model to other subjective observations. I'll suggest read some D.Hume or other early empiricists philosophy.
No - you are confusing my term "model building" with creation of a single testable hypothesis. The sci-method requires the creation of a single testable hypothesis. But once verified a hypothesis is added the the sum of (tentatively) accepted hypotheses that together form a coherent model
of the physical observations.
We have one (rather fuzzy) model of physical reality that we call "science" that incorporates all tested-and-so-far-valid hypotheses.
We can never know "the real world" through our subjective senses. It's a pleasant and common fantasy to assume that our observations accumulate a picture of reality rather than a picture of subjective human observations.
I'm not interested in benefiting science at the moment - I'm interested in creating a coherent picture of exactly what science is and does. If we were to encounter some other advanced alien culture that understood how to make predictions and extrapolations about their subjective observations of the physical world - then we might find that their method of developing a model and our "science" are quite different. We might find that our monkey brains have "blind spots" that cause us to create incorrect or sub-optimal models, or perhaps there is something better than our application of probability to science.
Sience has absolutely nothing to do with "truth". The point of science is to create a model that permits prediction and extrapolation of observations from observations. If the observations are distorted or a sham - then the sci-method may still work, but it isn't modeling "reality" or "truth"; it's modeling subjective human observations.