I don't think it's productive to determine the part of speech of this word - it's neither, since it modifies neither a noun nor a verb. necesse/necessum est, opus est, ūsus est and oportet are all predicative expressions that are semantically indivisible, forming the predicate together. If you classify necesse as an adjective, you'd have to classify opus, ūsus as adjectives—afaik using magnum opus est to mean this is impossible, it will mean "the task is big", so you have to say valdē opus est— ...and then perhaps analyse oportet and the rest of Latin verbs as containing a head - the ending - stuck to an adjectival base (hey, there are plenty of languages where adjectives = verbs!). At least in my ignorance of generative syntax I don't see what would stop you. In short, here seems to be where traditional grammar becomes useless.

Pinkster's 2015 Oxford Latin Syntax Vol. 1 p. 622 suggests that these are in fact imperative argument clauses since they regularly lack the subordinator ut , apparently with only a single instance of nē (surprising!).

I don't think it's productive to attempt to determine the part of speech of this word - it's neither, since it modifies neither a noun nor a verb. necesse/necessum est, opus est, ūsus est and oportet are all predicative expressions that are semantically indivisible, forming the predicate together. If you classify necesse as an adjective, you'd have to classify opus, ūsus as adjectives—afaik using magnum opus est to mean this is impossible, it will mean "the task is big", so you have to say valdē opus est— ...and then perhaps analyse oportet and the rest of Latin verbs as containing a head - the ending - stuck to an adjectival base (hey, there are plenty of languages where adjectives = verbs!). At least in my ignorance of generative syntax I don't see what would stop you. In short, here seems to be where traditional grammar becomes useless.

Pinkster's 2015 Oxford Latin Syntax Vol. 1 p. 622 suggests that these govern imperative argument clauses aka 'final' clauses since they regularly lack the subordinator ut (necesse est faciās ~ volō faciās), apparently with only a single instance of nē (surprising!).

I don't think it's productive to determine the part of speech of this word - it's neither, since it modifies neither a noun nor a verb. necesse/necessum est, opus est, ūsus est and oportet are all predicative expressions that are semantically indivisible, forming the predicate together. If you classify necesse as an adjective, you'd have to classify opus, ūsus as adjectives - afaik *magnum opus est is ungrammatical, you have to say valdē opus est. ...and then perhaps analyse oportet and the rest of Latin verbs as containing a head - the ending - stuck to an adjectival base (hey, there are plenty of languages where adjectives=verbs!). At least in my ignorance of generative syntax I don't see what would stop you. In short, here is where traditional grammar becomes useless.

Pinkster's 2015 Oxford Latin Syntax Vol. 1 p.622 suggests that these are in fact imperative argument clauses since they regularly lack the subordinator ut , apparently with only a single instance of nē (surprising!).

I don't think it's productive to determine the part of speech of this word - it's neither, since it modifies neither a noun nor a verb. necesse/necessum est, opus est, ūsus est and oportet are all predicative expressions that are semantically indivisible, forming the predicate together. If you classify necesse as an adjective, you'd have to classify opus, ūsus as adjectives—afaik using magnum opus est to mean this is impossible, it will mean "the task is big", so you have to say valdē opus est— ...and then perhaps analyse oportet and the rest of Latin verbs as containing a head - the ending - stuck to an adjectival base (hey, there are plenty of languages where adjectives = verbs!). At least in my ignorance of generative syntax I don't see what would stop you. In short, here seems to be where traditional grammar becomes useless.

Pinkster's 2015 Oxford Latin Syntax Vol. 1 p. 622 suggests that these are in fact imperative argument clauses since they regularly lack the subordinator ut , apparently with only a single instance of nē (surprising!).