Cyclicity and prosodic misalignment in Armenian stems

Abstract

Phonological processes are often sensitive to morphological, prosodic, and derivational structure. In terms of derivational structure, a common factor are strata or levels, as in Lexical Phonology (Kiparsky 1982) or Stratal OT (Bermúdez-Otero 2018). Two commonly argued strata are the stem-level and word-level cophonologies which are morphologically triggered. In this paper, I argue that Armenian has cyclic processes which follow this stratal model. However, I also show that Armenian phonology utilizes a prosodically-triggered cophonology. This cophonology is triggered by the prosodic misalignment between the morphological stem (MStem) and syllable boundaries. This occurs before vowel-initial inflection. I argue that this misaligned MStem is parsed into a sublexical prosodic constituent, the Prosodic Stem (PStem: Downing 1999a). This PStem-level cophonology applies between the stemlevel and word-level cophonologies.

Isomorphism between orthography and underlying forms in the syllabification of the Armenian schwa

“…This is contra most previous work which assumed that deletion is the default, and that the use of a schwa was a last resort(Vaux 1998;Khanjian 2009;Dolatian 2021a). To my knowledge,Hammalian (1984:94,130ff) is the only source to treat deletion as a last resort.…”

“…There are various morphological factors that condition the application of vowel reduction (Dolatian 2021a). An empirical question is whether the reduced schwa in words like [tʰəχt-iɡ] is epenthetic (Vaux 1998;Khanjian 2009) vs. the output correspondent of the destressed high vowel (Խաչատրյան 1966, 1988Khachaturian 1985).…”

Prosodically-conditioned relative clause extraposition in Armenian

“…For words (1), stress is generally word-final. It is on the final vowel if it is a nonschwa, otherwise on the rightmost non-schwa vowel (Vaux 1998;132;Dolatian 2021).…”

Armenian (Yerevan Eastern Armenian and Beirut Western Armenian)

“…The first- and second-person possessor suffixes /-s/, /-t/ (Yerevan), /-tʰ/ (Beirut) appear directly after vowel-final stems, but are preceded by a schwa when they are used with consonant-final stems. The presence of schwa in these alternations is not predictable from the phonology alone (Dolatian 2021): high vowels do not generally alternate with schwa when they are unstressed (see the Vowels table for examples), and /-s/ can occur in other word-final coda clusters such as in /ˈ t͡sɑχs/, /ˈd͡zɑχs/ ‘cost’ without schwa (see Syllable structure ). Moreover, some free-standing forms like Yerevan /vəˈkɑ/ and Beirut /vəˈɡɑ/ ‘witness’ have no related forms without schwa (Vaux 1998), and our speakers indicate that it is impossible or unnatural to elide the schwa ( * vkɑ, * vɡɑ) in isolation or in citation form.…”

“…Stem-final /-kʰ/ is an exception to these generalizations about coda clusters: it can occur after any singleton consonant or two-consonant cluster without any degree of schwa epenthesis, regardless of sonority, in both dialects (see Vaux 1998: 83; Dolatian 2021 for phonological analyses). Figure 9 shows spectrograms of /pɑɾtkʰ, bɑɾdkʰ/ [pɑɹ̝̊tkʰ, pɑɹ̝̊tkʰ] ‘debt’ from SK and HD, which each show three final consonant articulations — a spirantized /ɾ/, a /t, d/ constriction and release, and a final aspirated /kʰ/ release — with no acoustic evidence for schwa 11…”