In particular, I need to know whether it's true when one accounts for strictly pronominal plurality and for 'impure' tense markers such as Chinese aspect particle 了.
I think you just answered your own question. It's clearly not true if you do account for circumstantial features. Why should using aspect instead of tenses be an exception?
I do think your rewording is more objectively debatable:
If a language L has lexical nouns that obligatorily inflect for number, then L has verbs (lexical or otherwise) that obligatorily inflect for tense, where number and tense both admit at least two values with morphologically distinct realizations
I think Malagasy might be of interest to you as a potential counter-example:
To put these into sentences:
- Mividy zavatra aho ~ I buy something. (Buy - something - I)
- Mividy zavatra isika ~ We (inc.) buy something. (Buy - something - we(inc.))
- Nividy zavatra aho ~ I bought something. (Bought - something - I)
- Nividy zavatra isika ~ We (inc.) bought something. (Bought - something - we(inc.))
Here we have (morphological) tense marking despite the lack of distinctive plural inflection.
if a language has no plural morphology
Malagasy doesn't have plural morphology although it has pronouns which may be plural.
it has no tense marking
Malagasy does in fact have morphological tense marking, so there's a little exception there. Though I do think there's a bit of debate on whether the verb inflections show modality or tense. But I can assume this is more acceptable since you see a conjoined word rather than a distinctive character being used as a marker.
I do realize this must be less desirable given your background, but why not study this correspondence as a pattern or trend rather than a universal? I still think you might actually be on to something on a logical level.