Before there was physics and before there was philosophy of science, there was of course natural philosophy. When you read Newton, Maxwell, Thomson, Hertz, Boltzmann, and other greats, you would be hard pressed to circle what’s philosophy and what’s physics in these writings. There were just certain questions – e.g. are fields real or are they accounting devices? what explains entropy increase? are space and time real? – and a range of considerations for and against. The answers to these questions would draw on conceptual distinctions and scientific knowledge, however the two were classified, because these are the ingredients needed for an answer.
Specialization into distinct disciplines obviously has many benefits, but it’s naïve to think something doesn’t get lost too. Fallen into the cracks are a huge assortment of questions like the ones I just mentioned. These questions are deep and significant, but they are largely untouched. One can give sociological explanations for this: little grant money flows to such work in physics, the “linguistic turn” in metaphysics, and more. Whatever the reasons, there are good questions that deserve answers and often answering them will require the skill of a natural philosopher, be they officially a philosopher, physicist, or mathematician.
I do think that philosophy has something to offer science. The information transfer can be “two-way.” Due to their differences in training, philosophers often look at the same questions physicists do through different lenses, and sometimes this can help. Philosophers have training in logic, philosophy of science, metaphysics, and often history of science; with the distinctions they learn and a Socratic “follow the argument wherever it leads” ethos, they can raise questions about contemporary science that don’t seem to get asked much, e.g. like whether inflation really solves, say, the horizon problem. Though some physicists I know may grimace at what I’m going to say, I do think that philosophers have helped us understand the deep commitments of spacetime theories, especially in the 1970s and 1980s, and in framing and tackling the quantum measurement problem and quantum non-locality in the 1990s.
Many scientists thought for a very long time that relativity theory vindicated a Machian understanding of spacetime. Machianism may rise again, but the lesson philosophers have taught us is that the opposite is the case. Many scientists thought that decoherence by itself solved the quantum measurement problem; now, thanks in part to distinctions made by philosophers, that position is seen as a non-starter. Inasmuch as getting a clear picture of what one’s theory is committed to is important for future progress, one might also see this work as significant for future progress, too.