Mark:
Believing that beliefs should be grounded anticipates that there is absolutely no change in anticipation if one were to change these free floating ideas. Of course this doesn't really answer your question because it just restates the definition of 'free floating beliefs' in different words. This belief actually follows from Eliezers belief in Occam's Razor, which predicts that when faced with unexplained events, if one creates a set of theories explaining these events, any predictions made by the simple theories are more likely to actually happen than predictions made by complex theories.
I'm not quite sure if Occam's Razor is an axiom of science or just yet another belief. At least there is quite a bit of support for this belief, if you look into the history of science.
Another point: I think phlogiston is a bit of a poor example. Phlogiston actually corresponds very closely with something currently believed as real: phlogiston is the absence of oxygen. Seeing it this way, it's very well possible to build a theory of phlogiston explaining and predicting nearly all observations of fire, e.g. fire releases phlogiston, and if you burn something in a confined space the air gets saturated with phlogiston and cannot take in any more, so the fire goes out. A very important argument in the debate between phlogistionians and oxygants was when experiments were done to measure the weight of phlogiston and oxygen, and phlogiston turned out to have a negative weight.
Mark: Believing that beliefs should be grounded anticipates that there is absolutely no change in anticipation if one were to change these free floating ideas. Of course this doesn't really answer your question because it just restates the definition of 'free floating beliefs' in different words. This belief actually follows from Eliezers belief in Occam's Razor, which predicts that when faced with unexplained events, if one creates a set of theories explaining these events, any predictions made by the simple theories are more likely to actually happen than predictions made by complex theories. I'm not quite sure if Occam's Razor is an axiom of science or just yet another belief. At least there is quite a bit of support for this belief, if you look into the history of science.
Another point: I think phlogiston is a bit of a poor example. Phlogiston actually corresponds very closely with something currently believed as real: phlogiston is the absence of oxygen. Seeing it this way, it's very well possible to build a theory of phlogiston explaining and predicting nearly all observations of fire, e.g. fire releases phlogiston, and if you burn something in a confined space the air gets saturated with phlogiston and cannot take in any more, so the fire goes out. A very important argument in the debate between phlogistionians and oxygants was when experiments were done to measure the weight of phlogiston and oxygen, and phlogiston turned out to have a negative weight.