Considered as heuristics, or even adaptations, specific signaling behaviors don’t necessarily distinguish the situations where the behaviors can be observed by others and those where they can’t be. Behaviors that take contextual efficacy into account are more complicated, and likely won’t be developed unless the cost of idle signaling is high. The choice to contextually choose to implement a behavior should be considered a part of that behavior.
Thus, privacy of behavior isn’t a very good criterion: given behavior may be just a context-insensitive side effect of some other signaling behavior, and as such won’t be the best thing to do in that situation, still optimized for signaling.
Instead, it might be better to look for behaviors (globally, including the choice to perform in context as part) that have low signaling payoff and high personal cost of getting wrong. For behaviors that are never observed by others, the heuristics is that signaling payoff is (almost) zero, but the problem with this heuristics, as was pointed out is other comments, is that almost any behavior can be observed one way or the other by other people, even through the effect of self-signaling.
I see one possible goal of looking for behaviors untainted by signaling: they allow to elicit personal preference more clearly. Thus, it’s instructive to start from preferences, and look for decisions that go strongly for or against these preferences (thus having high personal cost of getting wrong), without significantly influencing the person’s image, with both factors only weakly depending on context.
This generates the following examples. Important food choices, if your social circle doesn’t care about food. A not-extraordinarily-looking hobby, if you are not involved in a community of same-theme hobbyists. Topics for which you surf the web. Choice of books to read, among image-similar classes. Job choice, if it isn’t impressive for other communities you are member of. More generally, a choice of community to participate in (among image-similar options).
This shows another pattern: even if a choice is important for signaling, there usually exist subsets of options for this choice, for which it’s unimportant for your image which particular elements get chosen. Then, the choice among these options in a subset may be clear of signaling, even if the global decision isn’t.
Considered as heuristics, or even adaptations, specific signaling behaviors don’t necessarily distinguish the situations where the behaviors can be observed by others and those where they can’t be. Behaviors that take contextual efficacy into account are more complicated, and likely won’t be developed unless the cost of idle signaling is high. The choice to contextually choose to implement a behavior should be considered a part of that behavior.
Thus, privacy of behavior isn’t a very good criterion: given behavior may be just a context-insensitive side effect of some other signaling behavior, and as such won’t be the best thing to do in that situation, still optimized for signaling.
Instead, it might be better to look for behaviors (globally, including the choice to perform in context as part) that have low signaling payoff and high personal cost of getting wrong. For behaviors that are never observed by others, the heuristics is that signaling payoff is (almost) zero, but the problem with this heuristics, as was pointed out is other comments, is that almost any behavior can be observed one way or the other by other people, even through the effect of self-signaling.
I see one possible goal of looking for behaviors untainted by signaling: they allow to elicit personal preference more clearly. Thus, it’s instructive to start from preferences, and look for decisions that go strongly for or against these preferences (thus having high personal cost of getting wrong), without significantly influencing the person’s image, with both factors only weakly depending on context.
This generates the following examples. Important food choices, if your social circle doesn’t care about food. A not-extraordinarily-looking hobby, if you are not involved in a community of same-theme hobbyists. Topics for which you surf the web. Choice of books to read, among image-similar classes. Job choice, if it isn’t impressive for other communities you are member of. More generally, a choice of community to participate in (among image-similar options).
This shows another pattern: even if a choice is important for signaling, there usually exist subsets of options for this choice, for which it’s unimportant for your image which particular elements get chosen. Then, the choice among these options in a subset may be clear of signaling, even if the global decision isn’t.