Nadahup languages

The Nadahup languages, also known as Makú (Macú) or Vaupés–Japurá, form a small language family in Brazil, Colombia, and Venezuela. The name Makú is pejorative, being derived from an Arawakan word meaning "without speech". Nadahup is an acronym of the constituent languages.

The Nadahup family should not be confused with several other languages which go by the name Makú. There are proposals linking this unclassified language with Nadahup, but also with other languages.

External relationships
Martins (2005: 342–370) groups the Arawakan and Nadahup languages together as part of a proposed Makúan-Arawakan (Nadahup-Arawakan) family, but this proposal has been rejected by Aikhenvald (2006: 237).

Epps and Bolaños (2017) accept the unity of the four Nadahup languages, but do not consider Puinave to be related.

Language contact
Jolkesky (2016) notes that there are lexical similarities with the Arawa, Guahibo, and Tupi language families due to contact. A discussion of lexical and phonological correspondences between the Nadahup (Vaupés-Japurá) and Tupi languages can be found in Jolkesky and Cabral (2011). Nadahup languages also have various loanwords from Tucanoan languages and Nheengatu.

Languages
Nadahup consists of about four languages, based on mutual intelligibility. Nadeb and Kuyawi, Hup and Yahup, and Nukak and Kakwa, however, share 90% of their vocabulary and are mutually intelligible, and so are separate languages only in a sociolinguistic sense. These four branches are not close: Although the family was first suggested in 1906, only 300 cognates have been found, which include pronouns but no other grammatical forms.

Nadëb may be the most divergent; of the other languages, there is disagreement on the placement of Nïkâk. Martins (1999) propose two classifications, pending further research:


 * Martins, proposal A


 * Martins, proposal B

However, Epps considers Hup and Yahup to be distinct languages, and maintains that the inclusion of the poorly attested Nukak and Kakwa has not been demonstrated and is in fact highly dubious:


 * Epps

Jolkesky (2016)
Internal classification by Jolkesky (2016):

( = extinct)


 * Puinave-Nadahup
 * Nadahup
 * Nadëb
 * Nadëb do Rio Negro
 * Nadëb do Roçado
 * Hup-Dâw
 * Dâw
 * Hup
 * Hupda
 * Yuhup
 * Puinave-Kak
 * Puinave (Wãnsöhöt)
 * Kak
 * Kakwa
 * Nukak

This classification is also repeated in Nikulin (2019).

Typology
Dâw and Hup—especially Hup—have undergone grammatical restructuring under Tucano influence. They have lost prefixes but acquired suffixes from grammaticalized verb roots. They also have heavily monosyllabic roots, as can be seen by the reduction of Portuguese loan words to their stressed syllable, as in Dâw yẽl’ "money", from Portuguese dinheiro. Nadëb and Nïkâk, on the other hand, have polysyllabic roots. Nïkâk allows a single prefix per word, whereas Nadëb, which lies outside the Vaupés language area, is heavily prefixing and polysynthetic: Up to nine prefixes per word (which is highly unusual for the Amazon), with incorporation of nouns, prepositions, and adverbs.

Genetic relations
Rivet (from 1920), Kaufman (1994) and Pozzobon (1997) include Puinave within the family. However, many of the claimed cognate sets are spurious.

Henley, Mattéi-Müller and Reid (1996) present evidence that the Hodï language (also known as Yuwana) is related.

Puinavean forms part of a hypothetical Macro-Puinavean family along with the Arutani–Sape families and the Máku language.

Macro-Puinavean is included in Joseph Greenberg's larger Macro-Tucanoan stock, but this is universally rejected. Another spurious larger grouping is Morris Swadesh's Macro-Makú.

Vocabulary
Loukotka (1968) lists the following basic vocabulary items for the Macú languages.

Proto-language
For a list of selected Proto-Eastern Makú reconstructions by Martins (2005), see the corresponding Portuguese article.