Exploring dissimilar research topics

Question

I'm currently a graduate student, but I plan on my future career being in research and academia. I like many different related topics and I like being interdisciplinary in my topics, and I see myself changing topics and doing my research broadly. However, professors usually stick to one topic for a very long time (either their entire lives or some absurd length like eight years) and milk every subtopic within it. I cannot see myself doing that willingly, but I also hear that staying within one topic allows you to essentially "master" it, thus getting you a better reputation, which feeds into a positive cycle of more publications, conference invites, security, etc. How true is that? Could I switch between topics and expect to be fine?

Answer 1

Eight years is not an "absurd amount of time" to work on a single topic by any measure: one-year research projects are typically exploratory in nature, one could expect to knock off some low-hanging fruit but that is about it. For most projects, three to five years is a more typical research cycle, and committing this much is to be expected. Eight years is just a couple of projects done in a single field.

Now, there are options for making more frequent changes. The easiest one is assuming an instrumentalist role, mastering a set of techniques and doing projects from varying fields. It requires a lot of commitment to said set of techniques, but you can get to work on a wide variety of research. Say, if you run a lab with an advanced microscope, material scientists and biologists of all kinds might want to enter a collaboration. If you run a ML lab, you get an even wider variety of projects.

Be aware, however, that with this approach you would be limited to a certain role; you might make suggestions regarding the experiment design and refine your technique, but not much more than that.

And for theoretical research, one still has to draw from being a sufficiently good specialist in a number of fields to be able to be a generalist at all. Research questions often defy classification - the notion of a field in research is no more than a useful tool, and nothing prevents anyone from applying e.g. differential geometry in number theory. The reason we stick to these fields is that a lot of effort is otherwise spent on learning the basics: from experience, it typically takes no less than an year to just get up to speed and start generating ideas people have already published a few years ago as opposed to decades ago.

The romanticized idea of polymath geniuses who just go about and solve a bunch of open problems in all the different fields is almost entirely unrealistic. In older times, a select few could even afford to do science; now we have many bright minds working in every field imaginable. Unlike industry, academia does not care much about whether you are the best baker in town: if someone across the world has strictly better results, yours are irrelevant. To be a successful generalist in science, you still have to be good at something.