In 2000, I completed the first draft of the article that later appeared as Gildea (2003), proposing the hypothesis that a cluster of Cariban languages spoken in Venezuela contain certain shared innovations that suggested a possible closer genetic relationship among them. I pointed out a number of innovative patterns that were either unique or particularly widespread in Venezuelan Cariban languages: (i) one phonological and one morphological innovation that connect all of them; (ii) one lexical innovation and four grammatical innovations that connect †Tamanaku, Panare, and the Pemóng Group; and (iii) two phonological innovations and three grammatical innovations that connect Panare and the Pemóng Group. Given that most of the Cariban languages of Venezuela have not been well-documented, I suggested that future descriptions might particularly look for evidence of these innovations, so that we could test their utility as a tool to further subclassify the languages into groups and branches.

In the intervening 15 years, a few works have engaged with the specific claims of that article, most notable among them Mattei Muller (MM 2002, MM2003, MM2005), Meira & Franchetto (2005, supplemented by Gildea, Hoff & Meira 2010), and Cáceres (2011). In this talk, I will review the current status of each claimed innovation, some of which have been convincingly challenged, others challenged but perhaps still of value, and others that appear to be real innovations that are still attested only in this hypothesized subgroup. The remainder of this abstract gives a brief synopsis of the issues to be discussed in the talk.

The two patterns that were claimed to hold for the entire Venezuelan Branch have both been critiqued. The claimed phonological innovation of Proto-Cariban (PC) *o > Proto-Venezuelan (PV) ə no longer stands: Meira & Franchetto (2005) argue convincingly that PC *o > o and PC *ô > ə, o, so that the two were already distinct phonemes in Proto-Carib. Gildea, Hoff & Meira (2010) argue that the phonetic value of *ô must have been [ə] or [ɤ], so the presence of /ə/ no longer counts as a shared innovation. Mattei Muller (MM 2002) argues that the claimed morphological innovation, Proto-Cariban Ø- ‘3’ > Venezuelan *it- ‘3’ on vowel-initial nouns and postpositions, might be valid, however: (i) it does not always occur on vowel-initial roots, and (ii) might be confused with the very similar prefix PC *t- ‘3reflexive’ (she goes on to argue that this confusion discredits the reflex of *it- identified for Ye’kwana, but cf. Cáceres’ 2011 findings for a counter-argument). Also, she suggests that the PV *it- could have originated in a third person pronoun with reflexes in Mapoyo (tëï), Yawarana (tëwï) , and Ye’kwana (tïwï).

The five patterns that were claimed to link †Tamanaku, Panare, and the Pemóng Group (TPPKM) have been critiqued or modified. Suppletion in the verb of speech (PC V_INTR *kaci > a blend of PC *kaci and PC V_TR *tarô) is also attested in Yawarana, Mapoyo, Kari’nja (MM 2002) and Hixkaryana (Meira pc). Thus, the innovation appears not to be limited to the Venezuelan Branch, and even in TPPKM, it is not clear that the suppletive pattern is actually shared, as opposed to a series of independent innovations. The innovative progressive construction (*V_INTR-nô/V_TR-rɨ pôkô) is attested also in Mapoyo and Yawarana, and perhaps in Kari’nja (MM 2002), plus in Yukpa and Apalaí, so these cases might all represent parallel innovations. The innovation of finite relative clauses has now been attested in Ye’kwana (Cáceres 2011), and Yawarana (MM 2002), but unless we can argue convincingly that the innovative constructions are actually cognate (which may be difficult, because the relativizing morphology is not obviously cognate), this would also become a case of parallel innovations. Yawarana is added to the list of languages with modern reflexes of putative innovation *-cetɨ ‘nominalizer’ (MM 2002, MM 2005 apud Colina Amaro 1991), but a source for this form has not yet been identified, which means it is still possibly a shared retention from Proto-Cariban. The innovation of a- ‘impersonal’ has not been further discussed, but the innovation of an- ‘3O.Negative’ is attested in both Kari’nja and Ye’kwana (MM 2002, the latter more fully described in Cáceres 2011).

The six patterns that purportedly link Panare to the Pemóng Group (PPKM) have not really been critiqued so far, but only four remain robust: both phonological innovations, word-final vowel deletion in nouns and subsequent debuccalization of the final consonant, are still attested only in PPKM, as is loss of the 1+2 nominal, postpositional, and verbal prefix *k(ɨ)-‘1+2’ and innovation of the verbal suffix *-nô‘1+2S’. The other two innovations are now more questionable as shared changes: the innovative use of PC *-tjapô ‘absolutive nominalizer’ (an improvement on Gildea’s 1998/2003 reconstruction as PC *-sapo) with personal prefixes and as a main clause verb is also attested in Ye’kwana (Cáceres 2011), and may be an independent innovation in Panare and Ye’kwana, as discussed in Cáceres & Gildea (2014). As for Gildea’s (2003) suggestion that PC *rɨ > PPKM Ø, then Panare spread -ŋ ‘possessed N’ to verbs by analogy, Meira & Franchetto (2005) show that there is a consistent sound change, in which PC *r > Panare ŋ /__#. This provides a much simpler explanation for PC *-rɨ > Panare -ŋ and removes a potential shared innovation.

In conclusion, some of these putative innovations now look like shared retentions, others look like parallel innovations, and still others continue to look like plausible shared innovations. The greatest contributions to this discussion have come from new data, which identify these patterns in new languages of the family. We anticipate that future descriptive work with all of the Venezuelan Cariban languages will provide still greater contributions.