OPTSP/NBAR root ambiguity with proper names

[olzama]

In Spanish, proper names can occur with articles, so a typical way to realize a construction with a proper name is with the optional specifier rule:

At the same time, a root like this seems to be needed, for various titles and such (e.g. many examples in TIBIDABO portion with sentences of length 2, "bebe prematuro" and so on):

root_nbar := phrase-or-lexrule &
 [ SYNSEM [ LOCAL [ COORD-REL.PRED non_implicit_coord_rel,
                    AGR.DIVISIBLE -,
                    CAT [ HEAD noun & [ KEYS.KEY nom_rel ], 
                          VAL [ SUBJ < >,
                                COMPS < >,
                                CLTS < > ] ] ],
            NON-LOCAL [ REL 0-dlist,
                        SLASH 0-dlist ] ] ].

This results in some ambiguity for proper names:

I couldn't immediately figure out how to rule this out without breaking things, so, just documenting that this ambiguity is there.

[olzama]

Generally, right now the grammar attempts to distinguish between different types of nouns, some that allow optional specifier and some that require a specifier. Furthermore, some nouns are DIVISIBLE + and some are DIVISIBLE -, which determines whether they can go through the optional specifier unary rule or the so-called N-bar rule (the latter requires DIVISIBLE +). However, in the lexicon there is not a perfect assignment of nouns to the appropriate types, and perhaps many should be reassigned to the type which allows an optional specifier. Before, many sentences were covered due to various mistakes such as underspecified SPR < > in some of the phrase structure rules. After fixing, some, I lost sentences such as "Es profesor universitario". To fix that, I reassigned "profesor" to a type that allows an optional specifier. Many other nouns remain that should probably also be reassigned.

[fcbond]

We could use the Spanish wordnet to find nouns related to professions, if that would help, ...