Is the prolepsis (notion, anticipation, preconception, etc) of justice simply the concept that one neither harms nor inflicts harm?
At first blush, I would say not, that it varies from group to group and person to person.
My current opinions on this topic:
I agree with Godfrey's take with this modification: I don't think anticipations contain "black letter law." I think "black letter law" ("contracts to neither harm nor be harmed are good") is a conclusion that is reached only after rationally processing raw input data.
I think that in large part because preconceptions / anticipations are arranged in the canon as if they are parallel to or at the same level as to (1) the 5 senses, and also to (2) the feeling of pain and pleasure. Both of those function automatically, pre-rationally, and are not operating through rational processing and opinion.
So the parallel I would assert is that there is a "faculty of preconceptions" just like the faculty of sight and the faculty of hearing and the faculty of feeling pain and pleasure, all simply reporting and not processing data.
If that is correct, then the proper analogy would be that due to your faculty of preconceptions you have the ability to perceive that certain relationships and dealings with people are significant enough to perceive them as falling into a general category of relationships that we choose to call "justice." The particular preconception is a recognition that in the dealings of the people involved there is a relationship that we can expect to be repeated over and over and thus needs evaluation. It would be the faculty of pain and pleasure which then processes the data about the relationship and decides whether we feel pleasure or pain at the observation of the relationship. And of course we wouldn't notice it at all if we didn't have sight or hearing to make the observation in the first place.
So this is where I think DeWitt is on the right track by saying that the canonical faculties all work together, all simultanously, all serving as "reactions" which happening automatically and without (at that first stage) evaluation by reason or opinion.
It's only after all of those canonical faculties weigh in that we have the raw facts which our minds would then process and develop opinions and rational (or irrational, as the case may be) conclusions about what we "think" is going on in the situation we're considering.
And in support of this analysis I would cite the other clear text about anticipations, involving anticipations of the gods, where Epicurus seems to be saying that the false opinions about the gods also involve the processing of anticipations. So it's never anticipations themselves that are right or wrong or good or bad, but only our processing of them in the mind that leads to opinions about right or wrong or good or bad.
Anticipations wouldn't be "right" or "wrong" anymore than data from the eyes or ears is "right or wrong." What they could be is "more or less accurate" in the same way that someone's vision or hearing could be operating poorly, but even in those situations where there is some defect in the mechanism of seeing or hearing or anticipating, that doesn't mean that the reports from those faculties are "wrong" - it just means that they are less true to the facts being observed at that particular moment.
That's the way at present I interpret the direction DeWitt was going, which I think is correct. But the main and absolutely imperative point would be to reject that anticipations contain "opinion" or rational processing of any kind. If you go down the road of thinking that anticipations contain "conclusions" or "opinions" then that removes their canonical authority of being automatic and pre-rational, and puts you into a Platonic position of saying that there is some mechanism of transmitting "ideas" that are inherited at birth. Lots of things are inherited at birth, including apparently things we view as "instincts," but I think Epicurus ruled out the view that "ideas" are inherited from birth.