I think in many branches of physics theory and experiment have a closer relationship than is widely perceived. For example, it is not particularly uncommon for a granular matter theorist to turn to experiments but still continue contributing to theory papers (though in joint projects with former colleagues). I've seen quite a lot of experimental biophysicists proposing a theory model for some results they measured (or doing measurements to prove a model they themselved proposed) - something one would probably file under "theory", usually. And a neuroscience "experimental physics" professor comes to my mind who investigates learning and adaptive behavior on robots - which certainly has an experimental blend, but also contains a large fraction of what a theoretical physicist would spend his time with.
The perception of theory and experiment being detached (which they really shouldn't be) possibly comes from the big fields (solid state physics, particle physics, possibly even astrophysics) where a huge number of people works on essentially the same thing and, as a result, work becomes extremely specialized. And from the "all physics is Strings and Cosmology" Pop-Sci culture, of course.