154 Paulette LevyGiven that in (n), the specifier xa- licenses a noun as an attribute of another noun,
in (lib), it triggers the restrictive reading of an adjective, a more revealing gloss
for the phrases containing xa- in (11) and (nb) might be 'of the X, the one that Y':
for (11), 'of the meats, the one of the armadillo', for (lib), 'of the flowers, the yel-
low ones'.^5 These specifying constructions furnish the clearest diagnostics to sep-
arate nouns from adjectives in those Totonac languages where nouns and adjec-
tives are more sharply (i.e. more syntactically) distinguished. So one hypothesis is
that the development of a specifying construction, which shares with adjectives as
attributes the function of restricting reference, is one factor that led to the gram-
maticization of adjectives as a more clearly defined part of speech in PT.2.2. COPULA COMPLEMENTS
Both NPs and adjectives can be Copula Complements (CC). The copula, in both
cases, is zero in the present, wan ('become') in all other Tense/Aspects. The un-
marked order is CC + Copula + Copula Subject, whether the subject is pronom-
inal or an NP. The examples in (13) show both NPs and adjectives as CCs, with
pronominal and NP subject in the present, i.e. with zero copula.
(13) (a) ixkdm wix (a') tli'waqi wix
ix-kam wix strong you
3POSS-sonyou
'You are his son.' 'You are strong.'
(b) ixkdm namd: cuqa'wasa (b') tli'waqi namd: cuqa'wasa
ix-kam namd: cuqa'wasa strong DEM boy
3POSS-son DEM boy
"That boy is his son.' "That boy is strong.'
(14) is from text. It shows what on the surface might appear as an NP (i.e. xa-cone-
jo DET-rabbit, 'the rabbit'), but must be analysed as a zero copula sentence without
overt subject, because it is the complement of a relative pronoun, a context that
demands a sentence and disallows either an NP or an adjective. The parallel head-
less relative with an overt copula that follows it strengthens the argument. Analo-
gous sentences with adjective CC can be constructed and are readily accepted by
native speakers.
(14) ni: ma:qaxaqxi:ya'! [ti: [xa-conejo]s]^p nidd [defante nawdn]s
ni: ma:qaxaqxi:-ya:-f ti: xa-conejo
NEC understand.it-icpL-2 REL.hum DET-rabbit
ni:-la defante na-wan-a:
NEC-possible elephant FUT-COP-ICPL
(^5) In some constructions with xa- the set is stated explicitly. In others, it is presupposed as recov-
erable. In Levy (looia) I analyse it as a semidefinitivizer, one which does not entail unique identifi-
ability of an entity, but whose use rather implies that the speaker presupposes that the addressee will
uniquely identify the set out of which a certain entity is selected.