I think the problem with all rationality-based models of science is that they don’t take scientific realism seriously enough. That’s not surprising given that most of them were developed by philosophers in response to radical skepticism about the world and our ability to describe it.
Hypotheses are constrained by the world in three ways: (1) the hypothesis is, in the first place, constrained by a set of initial measurements and observations; (2) the hypothesis is constrained by the tools and skills available for framing and solving the problem, which are a product of previous scientific developments, which are in turn a product of physical reality; and (3) the hypothesis is constrained by prior theory, which was also subject to constraints 1-3 during its development, and is thus likewise constrained by physical reality. Hypotheses are also subject to sociological constraints but these are ultimately grounded in physical constraints. These are strong constraints and they exist regardless of whether you introduce a global normative constraint of rationality (Bayesian or otherwise).
I think any argument for global normative constraints in science should first attempt to demonstrate that the available physical and institutional constraints are insufficient. This needs to be done with reference to science as practiced rather than other equally ahistorical formal models of science. If we take scientific realism seriously, and believe the objects of science exist, then they themselves can explain the success of science in explaining the world. (Scientists are implicitly aware of this. If you ask a scientist, “How did you reach that conclusion?” they’ll say “I did x, y and z” and list off the practical steps they took. If, however, you ask them, “What is the methodology of science?” they’ll talk about skepticism or logical positivism or Popper’s falsificationism or Kuhn’s revolutions or whatever happens to be the fad at the time.)
An analogy: If you’re the first to land on the East coast of a new island, it’s hardly surprising that you and your decedents will also be the first to discover its inland mysteries as well as the North, South and West coasts; the geography of the island, as a continuous body of land, ensures that you can travel from one point to another and we need not posit some additional normative constraint that made your people Great Explorers. Likewise, it’s reality that decides that you get Special Relativity if you turn the wheels on Newtonian dynamics enough, and not the alleged rationality of the researchers involved.
I think the problem with all rationality-based models of science is that they don’t take scientific realism seriously enough. That’s not surprising given that most of them were developed by philosophers in response to radical skepticism about the world and our ability to describe it.
Hypotheses are constrained by the world in three ways: (1) the hypothesis is, in the first place, constrained by a set of initial measurements and observations; (2) the hypothesis is constrained by the tools and skills available for framing and solving the problem, which are a product of previous scientific developments, which are in turn a product of physical reality; and (3) the hypothesis is constrained by prior theory, which was also subject to constraints 1-3 during its development, and is thus likewise constrained by physical reality. Hypotheses are also subject to sociological constraints but these are ultimately grounded in physical constraints. These are strong constraints and they exist regardless of whether you introduce a global normative constraint of rationality (Bayesian or otherwise).
I think any argument for global normative constraints in science should first attempt to demonstrate that the available physical and institutional constraints are insufficient. This needs to be done with reference to science as practiced rather than other equally ahistorical formal models of science. If we take scientific realism seriously, and believe the objects of science exist, then they themselves can explain the success of science in explaining the world. (Scientists are implicitly aware of this. If you ask a scientist, “How did you reach that conclusion?” they’ll say “I did x, y and z” and list off the practical steps they took. If, however, you ask them, “What is the methodology of science?” they’ll talk about skepticism or logical positivism or Popper’s falsificationism or Kuhn’s revolutions or whatever happens to be the fad at the time.)
An analogy: If you’re the first to land on the East coast of a new island, it’s hardly surprising that you and your decedents will also be the first to discover its inland mysteries as well as the North, South and West coasts; the geography of the island, as a continuous body of land, ensures that you can travel from one point to another and we need not posit some additional normative constraint that made your people Great Explorers. Likewise, it’s reality that decides that you get Special Relativity if you turn the wheels on Newtonian dynamics enough, and not the alleged rationality of the researchers involved.