Is "operatively coupled to" sufficiently well established that I can use it in the claims even if I don't use it explicitly in the spec?
My understanding is that its meaning is not a term of art in patents but that it's sufficiently clear English to be used without specific definition in the spec. However, if you want to rely on a particular interpretation, I think you should include that definition in the spec.
I take some comfort in the lack of input from examiners, experienced practitioners, and newly-minted practitioners with the latest rules and such freshly pressed into their already over-stuffed minds that I'm right that it's not a patent term of art with special meaning.
I think you mentioned somewhere that you might need to distinguish some intervening elements (like I resistor, I believe) yet still cover other intervening elements -- by not reciting a "direct" connection. You might consider what one element communicates to the other that is important for your particular invention. In my practice areas (mostly software), what gets passed around the elements is often the important part. You might find that you can focus on something that is blocked by prior art intervening elements but that would be passed by intervening elements in accused devices that use the essence of your invention.
Regards.