Derivational restrictions on dorsals in Tundra Nenets Peter - PowerPoint PPT Presentation

Derivational restrictions on dorsals in Tundra Nenets Peter Staroverov Wayne State University September 30, 2017 Peter Staroverov Nenets dorsals September 30, 2017 1 / 46

Overview Overview 1 Background on Tundra Nenets 2 Problem 3 Analysis 4 Conclusion 5 Peter Staroverov Nenets dorsals September 30, 2017 2 / 46

Overview Overview Input restrictions are inherently problematic for OT (Prince & Smolensky, 2004) Peter Staroverov Nenets dorsals September 30, 2017 3 / 46

Overview Overview Input restrictions are inherently problematic for OT (Prince & Smolensky, 2004) But input restrictions are necessary (Rasin & Katzir, 2017) Peter Staroverov Nenets dorsals September 30, 2017 3 / 46

Overview Overview Input restrictions are inherently problematic for OT (Prince & Smolensky, 2004) But input restrictions are necessary (Rasin & Katzir, 2017) They provide insight into two problems with Tundra Nenets dorsals Peter Staroverov Nenets dorsals September 30, 2017 3 / 46

Overview Overview Input restrictions are inherently problematic for OT (Prince & Smolensky, 2004) But input restrictions are necessary (Rasin & Katzir, 2017) They provide insight into two problems with Tundra Nenets dorsals Derivational solution within Stratal OT Peter Staroverov Nenets dorsals September 30, 2017 3 / 46

Overview Overview Input restrictions are inherently problematic for OT (Prince & Smolensky, 2004) But input restrictions are necessary (Rasin & Katzir, 2017) They provide insight into two problems with Tundra Nenets dorsals Derivational solution within Stratal OT Opaque distributional generalizations as stem-level output constraints (Kiparsky, 2000; Bermúdez-Otero, 2001) Peter Staroverov Nenets dorsals September 30, 2017 3 / 46

Background on Tundra Nenets Overview 1 Background on Tundra Nenets 2 Problem 3 Analysis 4 Conclusion 5 Peter Staroverov Nenets dorsals September 30, 2017 4 / 46

Background on Tundra Nenets Two data sources Fieldwork on a Western dialect of TN (joint with Darya Kavitskaya) Corpus of recorded texts Nikolaeva (2014) and larkpie.net/siberianlanguages/ The patterns we identify match what’s reported in the published sources (Janhunen, 1986; Salminen, 1997, 2012; Nikolaeva, 2014) Peter Staroverov Nenets dorsals September 30, 2017 5 / 46

Background on Tundra Nenets Surface vowel inventory i i; u u; e o 2 a ‘Null vowel’ written as < ◦ > Realized as an over-short reduced vowel or release of a consonant (or not at all) Results from a vowel reduction process, roughly: / 2 / → ◦ in unstressed syllables Peter Staroverov Nenets dorsals September 30, 2017 6 / 46

Background on Tundra Nenets Surface consonant inventory labial coronal dorsal glottal p p j b b j t t j d d j stop k g P s s j z z j x x j fricative m m j n n j nasal N ts ts j affricate l l j r r j liquid w w j glide j Peter Staroverov Nenets dorsals September 30, 2017 7 / 46

Background on Tundra Nenets Syllable structure CV(CC) complex codas very limited: C+ P (for certain Cs only) Peter Staroverov Nenets dorsals September 30, 2017 8 / 46

Background on Tundra Nenets TN: levels Most alternations that we will be talking about occur in phrasal phonology Within words Across word boundaries within the same phrase (esp. in casual speech style) Peter Staroverov Nenets dorsals September 30, 2017 9 / 46

Problem Overview 1 Background on Tundra Nenets 2 Problem 3 Analysis 4 Conclusion 5 Peter Staroverov Nenets dorsals September 30, 2017 10 / 46

Problem Phrase-initial position p p j t t j k x # V * � � � No syllable-initial clusters (so # C is impossible) No [k] phrase-initially This pattern is challenging since cross-linguistically initial position is known to favor preservation of contrasts Beckman (1998); Casali (1996) Peter Staroverov Nenets dorsals September 30, 2017 11 / 46

Problem Initial obstruents: borrowed [k] adapted as [x] p p j t t j k x # V * � � � (1) Phrase-initial obstruents: a. [pedara] ‘forest’ [ti] ‘reindeer’ b. [xos ◦ ka] ‘cat’ from Russian [ "koSk@ ] [x j ino] ‘cinema’ from Russian [ "k j i"no ] Peter Staroverov Nenets dorsals September 30, 2017 12 / 46

Problem Post-consonant strengthening: /x/ → [k] /x/ strengthens to [k] after a consonant /s/ also strengthens to [ts] (examples not shown) (2) Post-consonant strengthening / jar-x2na / → [ jark ◦ na ] ‘side-loc.sg’ ‘at the side’ Peter Staroverov Nenets dorsals September 30, 2017 13 / 46

Problem Voicing and postvocalic obstruents: overview [p p j t t j ] are voiced after a vowel, whereas [k] shows no similar alternations An instance of an unnatural class Mielke (2008) We’d need a disjunction ‘coronal or labial’ to characterize the undergoers of voicing Peter Staroverov Nenets dorsals September 30, 2017 14 / 46

Problem Voicing for coronal and labial stops: examples (3) Post-vocalic voicing (no similar alternations for [k]) a. Voiced stops after a vowel / ja-ta / [ jada ] ‘earth-poss.sg3sg’ ‘his earth’ [ m j ar j oj ◦ b2ni:naP ] / m j ar j oj2 p2ni2-naP / ‘bald garment-poss.pl1pl’ ‘our bald garments’ b. Voiceless stops elsewhere / jar-ta / [jarta] side-poss.sg3sg ‘his side’ / p2ni2 / [ p2ni: ] garment ‘garment’ Peter Staroverov Nenets dorsals September 30, 2017 15 / 46

Problem Status of [g] [g] is allowed in TN, but only after nasals where all obstruents are voiced and strengthened We won’t focus on NC effects here, but it’s important to explain why postvocalic voicing does not produce [g] Peter Staroverov Nenets dorsals September 30, 2017 16 / 46

Problem An aside on cluster simplification Underlying clusters escape voicing, deriving surface post-vocalic [p t k] Peter Staroverov Nenets dorsals September 30, 2017 17 / 46

Problem An aside on cluster simplification Underlying clusters escape voicing, deriving surface post-vocalic [p t k] Although I won’t exemplify all aspects of cluster simplification, we need an overview Staroverov & Kavitskaya (2017) Peter Staroverov Nenets dorsals September 30, 2017 17 / 46

Problem An aside on cluster simplification Underlying clusters escape voicing, deriving surface post-vocalic [p t k] Although I won’t exemplify all aspects of cluster simplification, we need an overview Staroverov & Kavitskaya (2017) All syllables must end in a placeless consonant, | P | or | N | , at a certain point in derivation Peter Staroverov Nenets dorsals September 30, 2017 17 / 46

Problem An aside on cluster simplification Underlying clusters escape voicing, deriving surface post-vocalic [p t k] Although I won’t exemplify all aspects of cluster simplification, we need an overview Staroverov & Kavitskaya (2017) All syllables must end in a placeless consonant, | P | or | N | , at a certain point in derivation Some consonants lose place to meet this requirement Peter Staroverov Nenets dorsals September 30, 2017 17 / 46

Problem An aside on cluster simplification Underlying clusters escape voicing, deriving surface post-vocalic [p t k] Although I won’t exemplify all aspects of cluster simplification, we need an overview Staroverov & Kavitskaya (2017) All syllables must end in a placeless consonant, | P | or | N | , at a certain point in derivation Some consonants lose place to meet this requirement Others add an inserted [ P ] Peter Staroverov Nenets dorsals September 30, 2017 17 / 46

Problem An aside on cluster simplification Underlying clusters escape voicing, deriving surface post-vocalic [p t k] Although I won’t exemplify all aspects of cluster simplification, we need an overview Staroverov & Kavitskaya (2017) All syllables must end in a placeless consonant, | P | or | N | , at a certain point in derivation Some consonants lose place to meet this requirement Others add an inserted [ P ] P or N later coalesce with a following C, if there’s a suitable target Peter Staroverov Nenets dorsals September 30, 2017 17 / 46

Problem Cluster simplification: an example Derivation: /s2s-s j 2/ → | s2P-s j ◦ | → s2ts j ◦ ‘be strong-ger.mod’ Place loss introduces [–cont] values Peter Staroverov Nenets dorsals September 30, 2017 18 / 46

Problem Cluster simplification: an example Derivation: /s2s-s j 2/ → | s2P-s j ◦ | → s2ts j ◦ ‘be strong-ger.mod’ Place loss introduces [–cont] values Second step is a single coalescence mapping: | P 1 s 2 | → [ts 1 , 2 ] Peter Staroverov Nenets dorsals September 30, 2017 18 / 46

Problem Cluster simplification: an example Derivation: /s2s-s j 2/ → | s2P-s j ◦ | → s2ts j ◦ ‘be strong-ger.mod’ Place loss introduces [–cont] values Second step is a single coalescence mapping: | P 1 s 2 | → [ts 1 , 2 ] Coalescence analysis matches what we know independently about morphological domains and process ordering in TN Peter Staroverov Nenets dorsals September 30, 2017 18 / 46