I agree with you on this issue, and I suspect education is at least partly responsible for the 'definitions first' approach. When teaching a topic in math, it

I agree with you on this issue, and I suspect education is at least partly responsible for the 'definitions first' approach. When teaching a topic in math, it is usual to start with definitions, and this is often also the case with science, but that is not how progress is made: more typically, the definitions follow from learning more about what starts as a vague or speculative concept, or was not even imagined before the studying was done (entropy, for example). Philosophy is not immune to the problem; it is not that uncommon for me to find an author basing an argument on an ontology or a metaphysical claim that I suspect could be, at some future time, empirically refuted, or where some thing is "in a sense" like something else, without establishing that the sense in which they are similar is relevant (for example, there is a sense in which learning how to ride a bicycle is like learning how to perform addition, but it would take more than this sense to conclude that riding a bicycle is, to any useful degree, just a matter of learning a set of propositions.) As you say, definitions can be used to avoid an issue, and while I agree with Dennett on many things, I feel that he and his fellow compatibilists are doing this with free will, which I consider to be the true 'hard problem' of consciousness.