It is generally considered methodologically advisable to exploit alternative ways of dealing with some subject matter, and the point is to do with validity and, to some extent, efficiency. There is always a risk that a particular way (method) of addressing some range of phenomena comes with certain inherent biases, gaps or otherwise confounding/misleading/distorting aspects, or there may simply be a less demanding option. This is something one may hope to control for, to some extent at least, by trying different ways of addressing the issue to see whether things look the same regardless of approach.
A non-negligible portion of work in philosophy are what we might call heavily “key notion”-oriented. By this I mean that the focus is much on expressions involving what is considered central, seemingly subject matter-determining terms of the domain at hand - ‘causation’, ‘knowledge’, ‘explanation’, ‘function’ -  and the “method” relies heavily on eliciting verdicts about the appropriateness of asserting sentences involving those key notions in various circumstances. Manners of expression are tools we use and this one-sidedness then appears less than optimal form the point of view of the above methodological considerations regarding validity/efficiency. We are less likely to discover problems in such respects if we always operate with the same tools.
I will argue for the need for alternative manners of charting well-trodden territories; epistemology without knowledge-talk, metaphysics without cause-talk, philosophy of science without explanation-talk…etc. As for efficiency, the view – a seeming pragmatic presupposition of the practice itself - that extant manners of doing things are less burdensome than getting target terms out of the way cannot be claimed to be empirically well-grounded, and can be suspected of self-justifying bias. As for validity, one may find that when target terms are removed some constructs introduced due to the application of those notions will be found to lack independent justification in the context.
I will argue the general point and provide some examples for illustration.