Currently, Páez is best considered either a language isolate or the only surviving member of an otherwise extinctlanguage family. It has often been grouped with other languages in a Paezan family, but several of these proposals are based on a historical error. Even before the discovery of the error, Campbell stated, "There is no consensus upon Paezan, and opinions vary greatly".
One of the most often repeated statements is the supposed connection between Páez and the extinct Panzaleo, formerly spoken in highlands of Ecuador. However, Panzaleo is poorly documented and the evidence for this relationship is weak and may be from language contact. Thus, Panzaleo may best be considered an unclassifiedisolate. The Andaquí isolate is often connected with Páez in a Paezan grouping. Documentation is a 20-page list of words and expressions by an anonymous author published in 1928 and another word list collected in 1854 by a priest. There are a number of similarities in vocabulary between Andaquí and Páez. In other aspects, the differences are greater.
The Coconucan languages were first grouped together with Páez by Henri Beuchat & Paul Rivet in 1910. Curnow shows this is based on misinterpretation of a Moguex vocabulary of Douay, which is a mix of Páez and Guambiano/Totoró. The error has led to subsequent classifiers to group Páez with Guambiano, missing the obvious identification of Coconucan as Barbacoan. Matteson's 1972 comparison of Páez and Guambiano vocabularies show just a 5.2% overlap, less than comparisons between Páez and Arawak, Quechua and Proto-Chibchan. Following linguists such as Matteson, Curnow, Curnow & Liddicoat, and Adelaar & Muysken, the Coconucan languages are now placed under Barbacoan. The question of connections between Páez, Panzaleo, and Andaquí remains open.
Prior to Curnow's correction, the Paez–Coconucan "family" had been connected to various other families. Greenberg included Paezan in a Macro-Chibchan stock with Barbacoan, Chibchan, Chocoan, Jirajaran, and the isolates Betoi, Kamsá, Yaruro, Esmeraldeño, Mochica, Cunza, Itonama, and Yurumanguí. Morris Swadesh's Paezan included Páez, Barbacoan, Coconucan, Andaquí, Cunza, Kapixana, and Mashubí. Kaufman's Macro-Páesan "cluster" proposal included "Paesan" –Barbacoan, Cunza–Kapixana, Betoi, Itonama, and Warao. Jolkesky argues for a connection, perhaps genetic, with the Otomanguean languages of the Pacific coast of central Mexico, especially with the Zapotecan branch, and for a possible Otomanguean substratum in a couple of families neighboring Paez.
