You can do that, but that's surely not the best way to do it, unless the concepts don't have anything to do with each other aside from being part of that field of study.
If it lets me avoid a circular definition, I’m willing to tolerate a bit of ugliness in my definitions (at least, provided that it can be shown to be equivalent (under assumptions of some results that can also be proven under the awkward form) to a nicer framing).