Or, the generative effect of the reification of rules of behavior.
1) Tipping after service can be defended on a rational basis if we assume repeated play. That is, if you expect to return to the place and encounter the same service person again, a post-service tip on this occasion could be reasonably expected to induce or maintain good service next time.
2) if it becomes common knowledge that tips are standard behavior, then a post-service tip might become necessary to avoid being the subject of retribution on subsequent visits.
3) But there is more to it than this. People will have an emotional reaction to not leaving a tip, as if they will be shamed or as if they are doing something slightly immoral - physiological signs include nervousness, increased heart rate, etc. Also, others will start to enforce compliance to the standard even when it is irrelevant to their own interests. Claims about tipping are conducted in moral language, using phrases like 's/he deserves it'. Appeals will be made to the potential non-tipper appearing cheap (i.e. not generous) - a status or identity challenge. Worse, these appeals can come from inside the potential non-tipper without needing to be explicit.
4) So, behavior that could have been initially started off on a rational basis is maintained and enforced and perpetuated via completely non-rational mechanisms.
5) In a situation where repeat business is unlikely, tipping is irrational, unless it is done to avoid retribution from one's peers.