I take it as obvious that signaling is an important function in many human behaviors. That is, the details of many of our behaviors make sense as a package designed to persuade others to think well of us. While we may not be conscious of this design, it seems important nonetheless. In fact, in many areas we seem to be designed to not be conscious of this influence on our behavior.
But if signaling is not equally important to all behaviors, we can sensibly ask the question: for which behaviors does signaling least influence our detailed behavior patterns? That is, for what behaviors need we be the least concerned that our detailed behaviors are designed to achieve signaling functions? For what actions can we most reasonably believe that we do them for the non-signaling reasons we usually give?
You might suggest sleep, but others are often jealous of how much sleep we get, or impressed by how little sleep we can get by on. You might suggest watching TV, but people often go out of their way to mention what TV shows they watch. The best candidate I can think of so far is masturbation, though some folks seem to brag about it as a sign of their inexhaustible libido.
So I thought to ask the many thoughtful commentors at Less Wrong: what are good candidates for our least signaling activities?
Added: My interest in this question is to look for signs of when we can more trust our conscious reasoning about what to do when how. The more signaling matters, the less I can trust such reasoning, as it usually does not acknowledge the signaling influences. If there is a distinctive mental mode we enter when reasoning about how exactly to defecate, nose-pick, sleep, masturbate, and so on, this is plausibly a more honest mental mode. It would be useful to know what our most honest mental modes look like.
Robin assumes that anything done in public (visible to others) is for signaling, so for his assumptions, I think you're right that this is the best answer.
I'm really questioning that assumption though. I think anything we do that species with less complex social environments also do would qualify as likely not for signaling: eating, sex, anti-predatory activities, etc.
And I think there's value in distinguishing between things we do to strut (showing off the newest cell phone) and things we do because of required social signaling (mowing the lawn). Otherwise it seems too easy to say "Everything is signalling" and not really have learned much.
Well the fact of eating is clearly done mostly for non-signaling reasons, but since eating is social the details of how we eat are greatly influenced by signaling. So our beliefs about why we pick those details are unreliable, to the extent we tend to not be aware of signaling influences on behavoir.