Dick:
> And:
> >### Whether they're the same lexeme or not, they're certainly not the same
> >#concept or category, because they have contradictory properties. So
> >#when you're processing a and b you must be invoking different verb
> >#categories:
> >#(a)     I cut the bread.
> >#(b)     The bread cut easily.
> >#No category can have contradictory properties.
> >
> >The verb in (a) isa CUT and doesn't isa Middle. The verb in (b)
> >isa CUT and isa Middle. I don't see why more has to be said
> >than that. If you want to go a step further and recognize categories,
> >then the relevant categories would be CUT and a category that
> >isa CUT and isa Middle. (a) would exemplify the default properties of
> >CUT. (b) would exemplify the properties of CUT and those
> >overridden by the subcategory.
> ## Precisely. But remember that "the verb in (a)" is itself a category,
> which isa CUT:past, so the latter is a category, etc. My categories =
> concepts, and some are stored while others are computed ad hoc and
> forgotten. So what I'm saying is that the verbs in (a) and (b) must be-a
> [NB new notation for 'isa' - whadyathink?] different categories. Maybe one
> of these is just the default CUT, but the other must be a distinct
> category, CUT/middle, even if it's ad hoc.

Okay. So the issue under debate is whether for a verb lexeme X, it
is necessary to systematically relate X to X/middle, and whether
that demands a Middle-of derivational relation. You say Yes to
both. I still don't see why.

--And.