ARTICLE Two problems in Armenian phonology* Luc Baronian Département des arts et lettres, Université du Québec à Chicoutimi Correspondence Luc Baronian, Département des arts et lettres, Université du Québec à Chicoutimi, 555 boul. de l'Université Est, Chicoutimi QC G7H 2B1, Canada Email: luc.baronian@uqac.ca Abstract The author surveys 2 major phenomena in Armenian pho- nology. The first is schwa epenthesis in Western Armenian, which is known to break up the impressive consonant clus- ters of the language. Before looking at the synchronically epenthetic schwas in Western Armenian, it is important to first distinguish schwas that are unambiguously part of lexical representations then to recognize schwas that have led, over time, to different allomorphs. A fourth category is left to explain: schwas that break up attested clusters, but only in derived morphological environments. After listing known facts about the syllabic nature of Armenian, the author shows that motivated prosodic specifications can account for the derived environment schwas (simplify- ing somewhat the previous account by Vaux, 1998). The second phonological phenomenon surveyed is the voicing and aspiration patterns found in Armenian dialects. While the patterns neatly divide the linguistic domain into seven groups, the standard classifications rely on present‐tense formation as a primary divider between larger groups of dialects. The author, however, highlights that the phono- logical division between voicing and aspiration patterns should be considered a more fundamental one, once textbook notions of geolinguistics are applied to reduce the 7‐way division to a more ancient 3‐way division of dialects. Indeed, sound change is a more common type of innovation used for grouping related languages and * I thank an anonymous reviewer, Anaïd Donabédian and Elan Dresher for very helpful comments on this paper. On various aspects of this research as it evolved over the years, I also thank Hrayr Khanjian, Paul Kiparsky, and Bert Vaux, as well as the audi- ences at Harvard's Indo‐European Workshop (Fall 2013), MOLT 2015 (University of Toronto) and MOLT 2016 (University of Ottawa), the Berkeley *dhworom group (Winter 2017), and the UCLA Phonology seminar (Spring 2017). Received: 8 March 2016 Revised: 9 May 2017 Accepted: 13 June 2017 DOI: 10.1111/lnc3.12247 Lang Linguist Compass. 2017;11:e12247. https://doi.org/10.1111/lnc3.12247 © 2017 John Wiley & Sons Ltd wileyonlinelibrary.com/journal/lnc3 1 of 18