Puquina (or Pukina) is a small, putative language family which consists of the extinct and Kallawaya, although it is assumed that the latter is just a remnant of the former mixed with Quechuan. The Qhapaq language which was brought by the Incas with Quechua is thought to be related as well as the Leco isolate language. They are spoken by several native ethnic groups in the region surrounding Lake Titicaca (Peru and Bolivia) and in the north of Chile. Puquina itself is often associated with the culture that built Tiwanaku.
Remnants of Puquina can be found in the Quechuan and Spanish languages spoken in the south of Peru, mainly in Arequipa, Moquegua and Tacna, as well as in Bolivia. There also seem to be remnants in the Kallawaya language, which may be a mixed language formed from Quechuan languages and Puquina. (Terrence Kaufman (1990) finds the proposal plausible.)
Some theories claim that "Qhapaq Simi", the cryptic language of the nobility of the Inca Empire, was closely related to Puquina, and that Runa Simi (Quechuan languages) were spoken by commoners. The Leco language might also be related.
Moulian et al. (2015) argue that Puquina language influenced Mapuche language of southern Chile long before the rise of the Inca Empire. This areal linguistic influence may have started with a migratory wave arising from the collapse of the Tiwanaku empire around 1000 CE.
Sometimes the term Puquina is used for the Uru language, which is distinctly different.
Puquina has been considered an unclassified language, since it has not been proven to be firmly related to any other language in the Andean region. A relationship with the Arawakan languages has long been suggested, based solely on the possessive paradigm (1st no-, 2nd pi-, 3rd ču-), which is similar to the proto-Arawakan subject forms (1st * nu-, 2nd * pi-, 3ª * tʰu-). Recently Jolkesky (2016: 310-317) has presented further possible lexical cognates between Puquina and the Arawakan languages, proposing that this language belongs to the putative Macro-Arawakan stock along with the Candoshi and the Munichi languages. However, such a hypothesis still lacks conclusive scientific evidence.
In this regard, Adelaar and van de Kerke (2009: 126) have pointed out that if in fact the Puquina language is genetically related to the Arawakan languages, its separation from this family must have occurred at a relatively early date; the authors further suggest that in such a case the location of the Puquina speakers should be taken into account in the debate over the geographic origin of the Arawakan family. Such consideration was taken up by Jolkesky (op. cit., 611-616) in his archaeo-ecolinguistic model of diversification of the Macro-Arawakan languages. According to this author, the proto-Macro-Arawakan language would have been spoken in the Middle Ucayali River Basin during the beginning of the 2nd millennium BCE and its speakers would have produced in this region the Tutishcainyo pottery.