Harassment is defined by its effect on the person addressed, not by the speaker's intent

The argument

Proponents of this position begin with how harassment is understood everywhere else. In workplaces, schools, housing and conduct rules generally, behaviour counts as harassment when it is unwanted and when it leaves the person subjected to it degraded, intimidated or unsafe. The question asked is what the conduct did to the person on the receiving end, not what the person doing it says they meant. A manager who insists their remarks were affectionate does not make them acceptable by insisting it, and nobody finds this a strange way to proceed. Applied to the street, proponents argue, the same test gives a clear answer. Remarks shouted at a stranger about their body are not solicited, cannot be declined in advance, and are commonly described by those who receive them as unpleasant, diminishing or frightening rather than flattering. Proponents point out that people alter their routes, their clothing and the hours at which they go out in order to avoid them, which is not what anyone does to avoid a compliment. On the ordinary test, conduct that produces those effects and was not invited is harassment. The argument is therefore one about consistency rather than about severity. Proponents accept that a shout in the street is usually less serious than a sustained campaign at work, and say the comparison was never the point. What they deny is that the setting creates an exemption: the same conduct in an office would be recognised immediately, and the only thing changed by moving it outdoors is that the person affected has no one to report it to. Calling it harassment, on this view, simply applies the standard already in use.

Premises

Counter-arguments

Opponents reply that the ordinary test was built for settings a person cannot leave and in which the parties stand in some relation of power, such as a workplace, and that importing it into the street makes any unwanted remark between strangers harassment. They add that an effect-based test makes the offence depend on the sensitivity of the hearer, so identical words are an offence with one passer-by and a compliment with the next, which they regard as an unworkable standard for public conduct.

Rejecting the premises

[Rejecting P1] The settings named are ones a person cannot leave and in which the parties have continuing dealings, which is what makes an effect-based test workable there. The street shares neither feature, so the test does not transfer with its justification intact. [Rejecting P2] Intention is not the whole of it, but discarding it entirely leaves no way to tell a clumsy approach from an attempt to intimidate, and conduct rules elsewhere do not in fact discard it. [Rejecting P3] Commonly is doing a great deal of work here. Reactions to the same remark differ widely, and the premise generalises from those who object to everyone addressed. [Rejecting C] A test that classifies conduct by the hearer's response cannot serve as a rule for public behaviour, since complying with it would require knowing in advance something only the hearer knows.