Classifiers are more than classes, more than 20 maybe
They are semantically determined (classes are somewhat too, but classifiers do that clearly)
Classifiers of the same root can be changed, they are not a property of the word but of its meaning.
There is no agreement between the classifiers of the head and dependents.
These features can of course give different definitions, but those look like classes.
I've never seen the number matter in determining classifierness...
That said, these classes are semantically determined, very clearly.
There is no agreement between the classes of the head and dependents.
Classes of the same root can be changed - though this is very class-like (at least, in the Bantu languages it's like that), so I'm not sure this is phrased correctly.
If one wants to say "the loaf of bread", you'd use the contained class, with a marker to indicate change of class. Unless it's incorporated, in which case you don't. This applies to other "classes" as well.
the spider = odáú'n /o.dāw.n̩/ (high tone on daw, low tone elsewhere)
I see (something living) = netoitsé /nɛ.toj.tsɛ̄/
ne "1st p." + ta "theme applic." + o "AS class" + eits "see" + e "epenthetic"
the bread = lhapara (lha "RF class" + para "bread")
loaf of bread = íúbara (il "CO class" + class change deaspiration + para)
I see bread (lots of it) = netleitsé para (ne + ta + lha + eits + e)
I see loaf of bread = ndíleitsé para (ne + ta + il +eits +e)
They may not be classifiers, but there are classifierish behaviors. Do I just call them whatever I want?