Excerpt from change request document:
The languages were probably misanalyzed as dialects of a single language because, as Breen admits "Garrwa and Wanyi are superficially very similar" (2003: 426). He concludes an extensive comparison of the two languages by saying "The many correspondences in grammar, especially the pronoun inventories, confirm the conclusion that these languages are closely related, but at the same time the substantial differences in grammar suggest that they must be mutually unintelligible -- closely related languages rather than dialects of a language" (2003: 454).
Garrwa and Wanyi have about 50% shared lexicon, either identical words or confirmed cognates. Nominal morphology is more or less similar but by far the most significant difference is the verbal system. Wanyi has no evidence of the clitic system that Garrwa uses to express TAM, and prefers a richer verb conjugation to express these instead. Belfrage (2003) corroborates these conclusions of Breen, and while he revises the verb analysis so that the two languages resemble each other even further, he maintains the assertion that the languages are distinct.
Perhaps contributing to the confusion is the asymmetrical distribution of the two: they both occupy areas about the same size, but per Walsh (1981) Garrwa had "300+" speakers while Wanyi had a questionable "10?". Moreover, Walsh recognizes the dichotomy and goes so far as to list Garrwa and Wanyi as distinct subgroups within a larger family. As Breen (2003) explains, the situation is more complex than it seems on the surface. Garrwa in fact has two or three true dialects--Western Garrwa and Eastern Garrwa, the latter of which may or may not be distinguishable from another variety called Kunindirri (2003: 426).
In terms of higher-level organization, Austin mentions that "The two closely-related languages Garawa and Wanyi are spoken in an area immediately north of Wagay. In vocabulary and morphology they are very different from all the languages discussed above, and are classified as members of a distinct language family by O'Grady et al" (1981: 327). Breen calls this family "Yanyi," explaining "I have used the name Yanyi (the word for 'language' in these languages) for the group for a number of years... and propose this as a name for the group that makes no implications regarding its status and does not imply primacy for either language" (2003: 426).