© 2020 Bálint Huszthy Proceedings of WECOL 2019 Is Italian Swedish? An uncommon look at Italian laryngeal phonology Bálint Huszthy Research Institute for Linguistics (Budapest) 1 Background Even if Italian and Swedish are distantly related languages, they do not show many common structural features in synchrony, except from laryngeal phonology. Laryngeal contrasts in languages challenge phonological theory, since beyond the voiced-voiceless binary distinction, three-way and four-way laryngeal systems also occur, while contrasts are not only based on [voice] but on other features like [spread glottis] and [constricted glottis], too (Iverson & Salmons 1995; Balogné Bérces & Huber 2010). Laryngeal Realism (Honeybone 2002, 2005; Petrova et al. 2006; Harris 2009; Beckman et al. 2013; etc.) sorts binary laryngeal systems into two categories, according to the markedness of either the [voice] or the [spread glottis] feature. In the category of the so-called voice languages (e.g. Slavic, Romance, Hungarian, etc.) the marked feature is [voice], and thoroughly voiced obstruents are in contrast with voiceless ones; on the other hand, in aspiration languages (e.g. most Germanic languages, Mandarin Chinese, etc.) the marked feature is [spread glottis], and the laryngeal contrast stays between voiceless unaspirated and voiceless aspirated stops (e.g. Eng. back [p] ~ pack [pހ]). In these langXages obstrXent voicing is usually passive, that is, possible only in intersonorant position (between vowels or sonorants); while in voice languages voiced obstruents have their own voice value, which is considered active and so it can spread, evoking regressive voice assimilation (RVA; i.e., regressive devoicing /bt/ → [pt] and regressive voicing /pd/ → [bd]; e.g. Slavic vodka [tk], McDonald’s [d], etc.). However, certain languages or varieties with a two-way laryngeal distinction do not properly fit with these generalizations. The classic example is Swedish, whose initial voiced obstruents are fully voiced, but voiceless stops are heavily aspirated (Ringen & Helgason 2004; Helgason & Ringen 2008). Apparently, Swedish presents laryngeal patterns typical for both voice and aspiration languages. Huszthy (2019) claims that the case of Italian is a similar phonological riddle. Italian is usually considered a true voice language in the literature, even if its voiced stops are in contrast with mildly aspirated voiceless stops, and voice spreading cannot be detected in the system. The laryngeal patterns found in Italian permit us to treat the language both as an exceptional voice language or as an exceptional aspiration language. In this paper we argue for the latter proposal, which makes Italian laryngeal phonology similar to that of Swedish. 2 The interface of Element Theory and Laryngeal Relativism The framework this study is couched in is an extension of Laryngeal Realism proposed by Cyran (2011, 2012, 2014, 2017a, 2017b), which he calls Laryngeal Relativism (LR); combined with Element Theory (ET), a subtheory of melodic representations in Government Phonology (Kaye, Lowenstamm & Vergnaud 1985; Harris 1990, 1994; Scheer 2004; Kaye 2005; etc.). From an ET-approach, the basic typology of two-way laryngeal systems is identical to that of Laryngeal Realism, expressed through unary phonological primes. Harris (1994) claims that in voice languages the element L (used for active voice in obstruents) is active in the marked series of obstruents (so this is the category of L-systems); while in aspiration languages the element H (used for voicelessness or aspiration in obstruents) is the marked laryngeal prime (so this is the category of H-systems). C\ranｶs LR is based on the idea that as long as a sXfficient phonetic distance is kept betZeen tZo sets of obstruents (e.g., voiced-voiceless, aspirated-unaspirated) to maintain phonological contrast, both the marked and the unmarked sets may receive any (more or less arbitrary) phonetic interpretation. That is, phonetic interpretation is partly phonological. He also claims that it may even be the case that two