* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Word Classes.ppt
Survey
Document related concepts
Esperanto grammar wikipedia , lookup
Old English grammar wikipedia , lookup
Yiddish grammar wikipedia , lookup
Macedonian grammar wikipedia , lookup
French grammar wikipedia , lookup
Japanese grammar wikipedia , lookup
Ancient Greek grammar wikipedia , lookup
Pipil grammar wikipedia , lookup
Romanian grammar wikipedia , lookup
Old Irish grammar wikipedia , lookup
Sotho parts of speech wikipedia , lookup
Romanian nouns wikipedia , lookup
Vocative case wikipedia , lookup
Polish grammar wikipedia , lookup
Serbo-Croatian grammar wikipedia , lookup
Transcript
Word Classes Laura A. Janda Overview 1. How are word classes determined? 2. Who needs “particles”? 3. Is the “new vocative” in Russian a unique part of speech? 2 Theoretical problems with particles as a part of speech • • • 3 Langacker (2013: 96) on parts of speech: “Traditional terms lack precise definition, are inconsistent in their applications, and are generally inadequate” Croft (2001: 63-107) Parts of speech are partly language-specific: the “same” categories might not coincide exactly across languages, though the focal points of certain categories, such as noun, pronoun, verb are typologically similar Part of speech categories can be complex and can overlap: verb ← blestjaščie glaza ‘shining eyes’ verb ← budem kušan’ki ‘we will eat’ participle non-inflected diminutive → adjective blestjaščaja pobeda ‘shining victory’ → noun ja prigotovila kušan’ki ‘I prepared food’ Theoretical problems with particles as a part of speech • • • Langacker (2013: 96) on parts of speech: “Traditional terms lack precise definition, are inconsistent in their applications, and are generally inadequate” Croft (2001: 63-107) Parts of speech are partly language-specific: the “same” categories might not coincide exactly across languages, though the focal points of certain categories, such as noun, pronoun, verb are typologically similar Part of speech categories can be complex and can overlap: Russian words classed as “particles” are particularly prone to overlap across part of speech categories 4 How do we identify parts of speech? • Formal characteristics: morphological classes, e.g., nouns inflected for case, verbs for tense and person • Distributional characteristics: e.g., adpositions contiguous with noun phrases, pronouns substitute for nouns, conjunctions bind phrases • Semantic characteristics: e.g., nouns signify entities, verbs signify situations Ideally, a classification should take into consideration all three types of characteristics 5 How do we identify parts of speech? • Formal characteristics: morphological classes, e.g., nouns inflected for case, verbs for tense and person • Distributional characteristics: e.g., adpositions contiguous with noun phrases, pronouns substitute for nouns, conjunctions bind phrases • Semantic characteristics: e.g., nouns signify entities, verbs signify situations Russian words classed as “particles” Ideally, a classification should take into consideration all three a coherent definition for formal, types oflack characteristics distributional and semantic characteristics 6 2. Who needs “particles”? • In 1985, Zwicky argued that “particle” is a pretheoretical notion that should be eliminated from linguistic analysis. We propose a reclassification of Russian particles that implements Zwicky’s directive. The so-called particles lack a coherent conceptual basis as a category and many of them are ambiguous since they can also be classed as adverbs, conjunctions, predicatives and other parts of speech. 7 Particles: Russian ЖЕ (ŽE) Konečno, sgorel – nel’za že v polden’ ležat’ na solncepeke [«Domovoj», 2002] ‘Of course, you got a sunburn! You can’t ŽE lie in the hot sun in the middle of the day!’ “You are wrong! And more than that, you are capable of arriving at the correct conclusion yourself, but nevertheless you are sticking to the wrong conclusion.” McCoy (2003: 125) 8 Particles: Russian ЖЕ (ŽE) • Konečno, sgorel – nel’za že v polden’ ležat’ na solncepeke [«Domovoj», 2002] ‘Of course, you got a sunburn! You • can’t ŽE lie in the hot sun in the middle of the day!’ Small uninflected words lacking referential content (Švedova et al. 1980) Meaning: modal and pragmatic attitudes towards a proposition “You are wrong! And more than • Multifunctional allow multiple that, you are capable of arriving interpretations at the correct conclusion yourself, but nevertheless you are sticking • Overlap with other parts of speech to the wrong conclusion.” McCoy (adverbs, conjunctions, (2003: 125) interjections, predicatives) • A part of speech ? 9 Particles: small words, big problems • “The wide use of particles is a typical feature of colloquial Russian” (Vasilyeva 1972: 6) Emu ja mogu poverit’ – ‘I can trust him’ Ved’ emu ja mogu poverit’ – ‘I can trust him, you know this’ Emu-to ja mogu poverit’ – ‘I know, I can trust him’ Emu ja ešče mogu poverit’ – ‘Well, I suppose, I can trust him’ Tak emu ja mogu poverit’ – ‘so I can trust him’ Vot emu ja mogu poverit’ – ‘He is the one I can trust’ Emu ja i mogu poverit’ – ‘Therefore I can trust him’ Da emu ja mogu poverit’ – ‘Well, I can surely trust him’ Xot’ emu ja mogu poverit’ – ‘At least I can trust him’ 10 Particles: small words, big problems • Active use of particles distinguishes L1 speakers from L2 learners (Nikolaeva 1985: 7) • Relevant for other languages too. • Heinrichs, W. 1981. Die Modalpartkeln im Deutschen und Schwedischen. Tübingen. • L2 German speaker: – Bitte geben Sie mir das Buch. • L1 German speaker: – Können Sie nur vielleicht mal das Buch da geben? – Ach, geben Sie nur doch bitte mal das Buch. 11 Outline • Particles in Russian – Extent – Distribution in the corpus • Our data: 9 words – Database • Analysis – Alternative annotation scheme & guidelines • Experiments 1 and 2 – Training a tagger to disambiguate between uses 12 Practical problems with particles as a part of speech • • • • 13 Automatic Part of Speech taggers are trained on a gold standard corpus 1 Part of Speech error can foul up the parsing of a whole sentence Manning 2011: Penn Treebank of English yields 97% accuracy in automatic Part of Speech tagging, but – This yields only 57% sentence parsing accuracy! – Main culprit is Part of Speech tagging errors Accurate tagging is important not only for Natural Language Processing, but for all tools sourced by NLP: – spelling and grammar checkers – intelligent computer-assisted language learning – linguistic corpora – machine translation Practical problems with particles as a part of speech • • • • Automatic Part of Speech taggers are trained on a gold standard corpus 1 Part of Speech error can foul up the parsing of a whole sentence Manning 2011: Penn Treebank of English yields 97% accuracy in automatic Part of Speech tagging, but – This yields only 57% sentence parsing accuracy! – Main culprit is Part of Speech tagging errors Accurate tagging is important not only for Natural Language Processing, but for all tools sourced by NLP: – spelling and grammar checkers – intelligent computer-assisted language learning – linguistic corpora – machine translation Russian words classed as “particles” are the most error-prone part of speech 14 Our proposal • • • • Particle is not a valid category. Russian particles have no coherent profile. “Particle” looks like a garbage category that is used when one feels uncertain about how to classify a word. Particle is not a classification but rather a failure to classify a word. • • • It is possible to reclassify the words commonly classed as “particles”. We offer improved annotation guidelines that eliminate the class of particles altogether. Descriptively more precise analysis. 15 Extent of particles in Russian Estimates of the number of Russian particles vary: • Zaliznjak (1980) designates over 100 Russian words as particles. • Nikolaeva (1985: 8) lists the following alternative counts: – 131 particles in the 17-volume Academy dictionary – 110 particles in the 4-volume Academy dictionary – 84 particles in Ušakov’s dictionary – 75 particles in Ožegov’s dictionary • Starodumova (1997: 8-9) claims that Russian is among the most “particle-rich” languages in the world, with approximately 300 particles. 16 Parts of speech in the disambiguated subcorpus of Russian Part of speech # tokens in RNC % tokens in RNC S (noun) 1,707,312 28.7% V (verb) 1,007,526 16.9% SPRO+APRO+PRAEDICPRO+ADVPRO (pronoun) 875,142 14.7% PR (preposition) 612,857 10.5% A (adjective) 506,691 8.5% CONJ (conjunction) 471,275 7.9% PART (particle) 268,139 4.5% ADV (adverb) 426,367 4.1% NUM+ANUM (numeral) 126,567 2.1% PRAEDIC (predicative) 42,318 0.7% PARENTH (parenthetical) 25,891 0.4% 8,377 0.1% 5,944,156 99.1% INTJ (interjection) Sum 17 Extent of part-of-speech ambiguity in particles Type Part of speech Unambiguous 2-way ambiguity 3-way ambiguity 4-way ambiguity 18 # of lexemes Example - 54 vot Conjunction 28 ved’ Adverb 17 ešče Predicative 7 net Interjection 8 nu Parenthetical 4 požaluj Preposition 1 vrode Adverb, Predicative 5 prosto Adverb, Conjunction 5 poka Parenthetical, Predicative 1 spasibo Parenthetical, Adverb 1 nikak Parenthetical, Adverb, Predicative 1 xorošo Parenthetical, Adverb, Conjunction 1 točno Data: 9 particles with one additional part-of-speech reading Lexeme Gloss Frequency Adverb Conjunction RNC ešče ‘more, yet’ 14,765 ADV PART tak ‘so’ 22,093 ADV PART ved’ ‘you know’ 5,149 CNJ PART slovno ‘like, as if’ 1,369 CNJ PART daže ‘even’ 8,562 CNJ PART že ‘after all’ 21,350 CNJ PART li ‘if’ 7,708 CNJ PART da ‘yes, and, but’ 12,280 CNJ PART net ‘no, there is no’ 9,786 Database: 100 random sentences for each word 900 rows in Excel spreadsheet 19 Predicative Particle PRED PART Source of data: disambiguated part of the Russian National Corpus (gold standard RNC) Can we re-tag the uses of these lexemes and avoid the tag ‘particle’? Analysis: Particle-free annotation Lexeme Gloss ešče ‘more, yet’ ADV CNJADV tak ‘so’ ADV CNJADV CNJSUB ved’ ‘you know’ ADV CNJADV CNJSUB ADV CNJADV ADV CNJCOO slovno ‘like, as if’ 20 Adverb Conjunction Predicative daže ‘even’ že ‘after all’ li ‘if’ ADV CNJCOO CNJSUB da ‘yes, and, but’ ADV CNJCOO net ‘no, there is no’ Interjection Emphasizer Question word INTJ CNJADV CNJCOO EMPH QST PRED VMOD INTJ NEST PRED INTJ Analysis: Particle-free annotation Lexeme Gloss ešče ‘more, yet’ ADV CNJADV tak ‘so’ ADV CNJADV CNJSUB ved’ ‘you know’ ADV CNJADV CNJSUB ADV CNJADV ADV CNJCOO slovno ‘like, as if’ 21 Adverb Conjunction Predicative daže ‘even’ že ‘after all’ li ‘if’ ADV CNJCOO CNJSUB da ‘yes, and, but’ ADV CNJCOO net ‘no, there is no’ Interjection Emphasizer Question word INTJ CNJADV CNJCOO EMPH QST PRED VMOD INTJ NEST PRED INTJ Analysis: Particle ЖЕ (ŽE) 1. Adverbial conjunction (ADVCNJ) – syntactically optional, usually preposed. 13% Konečno, sgorel – nel’za že v polden’ ležat’ na solncepeke. ‘Of course, you got a sunburn! After all, you can’t lie in the hot sun in the middle of the day!’ 2. Coordinating conjunction (CNJCOO) – usually postposed, obligatory for creating an explicit contrast between syntactic constituents: Satira i jumor. Odni ix rezko razdeljajut ... , drugie že vidjat v jumore ... raznovidnost’ satiry. ‘Satire and humor: some people keep them strictly distinct ... , others 6% however see humor as a form of satire.’ 3. Emphasizer (EMPH) – syntactically optional, follows a phrasal stress-bearing word and brings it in focus of attention Seli s kraju ― i tut že iz veščmeška Vovka izvlek butylku portvejna. ‘They sat down ― and right away then Vovka pulled a bottle of portwine out of the supply bag.’ 81% 22 Analysis: Particle ДА (DA) 1. Interjection – ‘yes’ 50% Da, zavtra ja priedu. ‘Yes, I will come tomorrow’ 2. Coordinating conjunction – ‘and’, ‘but’ 25% Ded da baba ‘grandfather and grandmother’ 3. Adverb – ‘after all, well’ 19% Da v etix pal’to pol goroda xodit! ‘Well, half of the town is wearing these coats!’ 4. Predicative – stands for entire proposition, carries stress: 3% Neobxodimo ustanovit’, želaet li on vozvratit’sja. Esli da, to kogda. ‘It is necessary to find out whether he wants to come back. If so, when’. 5. Modal verb – unstressed, used with imperatives, infinitives, present tense finite forms: Da budet svet! ‘Let there be light!’ 3% 23 Experiment 1: How badly does tagging of Russian particles perform? • Source: Russian National Corpus gold standard, using the tags manually assigned there • Database: 100 randomly-selected sentences for each of nine high-frequency particles (= 900 sentences) • Method: Hidden Markov Model (HMM), 10-fold crossvalidation, each time using 90 sentences as training set and 10 sentences as test set We measure performance as improvement in accuracy (correct guesses/total guesses) over baseline (frequency of most common tag for each word) 24 Experiment 1: How badly does tagging of Russian particles perform? • Source: Russian National Corpus gold standard, using the tags manually assigned there • Database: 100 randomly-selected sentences for each of nine high-frequency particles (= 900 sentences) • Method: Hidden Markov Model (HMM), 10-fold crossvalidation, each time using 90 sentences as training set and 10 sentences as test set We measure performance as improvement in accuracy (correct guesses/total guesses) over baseline (frequency of most common tag for each word) 25 Distribution of tags across the 900 sentences according to the RNC gold standard Lexeme ADV (adverb) CNJ (conjunction) ešče 83 tak 101 PRAEDIC (predicative) PART (particle) Total examples 17 0 100 101 ved’ 33 67 100 slovno 83 17 100 daže 16 84 100 že 6 94 100 li 18 82 100 da 54 46 100 42 100 net 58 Most common tag taken as “baseline” for each word 26 Experiment 1: So how did the HMM tagger do? Lexeme ešče tak ved’ slovno daže že li da net 27 Baseline 83% 100% 67% 83% 84% 94% 82% 54% 58% Average accuracy 70% 100% 78% 89% 90% 89% 76% 75% 89% Gain over Baseline -13% 0% 11% 6% 6% -5% -6% 21% 31% Experiment 1: So how did the HMM tagger do? еščё tak ved’ slovno Average accuracy daže Baseline že li da net 28 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Experiment 1: So how did the HMM tagger do? еščё tak ved’ slovno Average accuracy daže Baseline že li da net 29 0% Total gain: 51% These results confirm our suspicion that the tagging of Russian “particles” in the RNC gold standard is not consistent 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Experiment 2: Life without particles • Same source, database, and method as Experiment 1, but using our tags for the nine words instead of those in the RNC gold standard 30 Distribution of tags across the 900 sentences according to our tagging scheme ADV (adverb) CNJ (conjunction) ešče ADV 100% tak ADV 85% ADV 57% ADV 49% ADV 85% CNJADV 0% CNJADV 7% CNJSUB 8% CNJADV 33% CNJSUB 10% CNJADV 51% ved’ slovno daže že li ADV 23% da ADV 19% net 31 PRAEDIC (predicative) INTJ (interjection) EMPH (emphasizer) QST (question word) INTJ 1% CNJCOO 15% CNJADV 13% CNJCOO 6% CNJCOO 6% CNJSUB 0% CNJCOO 25% EMPH 81% QST 71% PRAEDIC 3% VMOD 3% NEST 60% PRAEDIC 10% INTJ 50% INTJ 30% Experiment 2: So how did the HMM tagger do? Lexeme ešče tak ved’ slovno daže že li da net 32 Baseline 100% 85% 57% 51% 85% 81% 71% 50% 60% Average accuracy 100% 90% 80% 68% 86% 76% 85% 79% 89% Gain over Baseline 0% 5% 23% 17% 1% -5% 14% 29% 29% Experiment 2: So how did the HMM tagger do? еščё tak ved’ slovno Average accuracy daže Baseline že li da net 33 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Experiment 2: So how did the HMM tagger do? еščё tak ved’ slovno Average accuracy daže Baseline Total gain: 127% More than twice the gain over baseline as in Experiment 1, despite much more complex tagging scheme že li da net 34 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Conclusions 35 • Can we eliminate particles from the part-of-speech classification in Russian? – Yes, “particle” is not a classification but a failure to classify a word. – It is possible to reclassify the words commonly classed as “particles”. • What are the practical benefits of this approach? – Particle-free annotation, where all categories are meaningful and useful for further applications. – Analysis that is descriptively more precise. • Our methods – Usage-based analysis of corpus data: 9 high-frequency “particles”. – Experiment: training an automatic tagger to disambiguate uses. 3. Is the “new vocative” in Russian a unique part of speech? 36 THEORY PART ONE: What is a vocative? A call for attention that hasn’t gotten much attention 37 Vocatives have been largely ignored “even though they are amongst the most basic and earliest acquired structures of language, vocatives have hardly ever been discussed in all their facets from a linguistic point of view” Sonnenhauser & Hanna 2013: 3 Among the few studies: Fink 1972, Zwicky 1974, Levinson 1983, Greenberg 1996 However, this situation seems to be changing: Sonnenhauser & Hanna 2013, Hill 2014, Julien 2014 38 Is a vocative a case form of a noun? YES, the vocative: – Can have distinct morphological form (Kiparsky 1967) – Can often be replaced by nominative, which is a case – Can show agreement within the NP (Hill 2014, Julien 2014) – Can be syntactically integrated via a Vocative Phrase (Hill 2014) NO, the vocative: – Is not syntactically integrated into a clause (Isačenko 1962) – Has peculiar restrictions that do not apply to other cases (functional, lexical, morphological, phonological, cf. Andersen 2012) – Diachronically behaves differently (cf. Bulgarian & Macedonian lost all cases but kept vocative) 39 Is a vocative a form of a verb? YES, the vocative: – Marks 2nd person (Fink 1972) – Shares features with imperatives (Jakobson 1971, Greenberg 1996) – Possessive predicational vocatives (Din idiot! ‘You idiot! [lit. Your idiot]’ ≈ Du er en idiot ‘You are an idiot’, cf. Julien 2014) NO, the vocative would be a mighty defective verb… 40 Is a vocative a form of a verb? YES, the vocative: From the perspective of Russian and Saami,(Fink interpretation as a verb – MarksNorth 2nd person 1972) is unlikely, so imperatives we won’t pursue that 1971, – Shares features with (Jakobson Greenberg 1996) avenue here – Possessive predicational vocatives (Din idiot! ‘You idiot! [lit. Your idiot]’ ≈ Du er en idiot ‘You are an idiot’, cf. Julien 2014) NOTE: association with possessive NO, the vocative would be a mighty defective verb… 41 Is a vocative another part of speech? Andersen (2012) argues that the vocative is a separate part of speech, based on evidence from the Russian “new” vocative Segue to Russian 42 PART TWO: Russian “new” vocative (1) Dissyllabic hypocoristics. (Костя! ⇒) Кость!, (Надя! ⇒) Надь!; (2) Dissyllabic diminutives. (Ванька! ⇒) Ваньк!, (Машка! ⇒) Машк!; (3) Hypocoristics of three or more syllables. (Наташа! ⇒) Наташ!, (Сережка! ⇒) Сережк!, ...; (4) Kinship terms. (дядя! ⇒) дядь!, (мама! ⇒) мам!,...; (5) Patronymics. (Андреевна! ⇒) Андреевн!, (Николаевна! ⇒) Николавн!,...; (6) Name + patronymic. (Анна Иванвовна! ⇒ Анн Ванна! ⇒) Анн Ванн!, (Марья Александровна! ⇒ Марь Санна! ⇒) Марь Санн!, ...; (7) Common nouns. (девушка! ‘Miss’ ⇒) девушк!, (хозяйка! ‘hostess, landlady’ ⇒) хозяйк!, (ребята! ‘boys, guys’ ⇒) ребят!, (девчата! ‘girls’ ⇒) девчат!, .... BOLDFACE marks items that conflict with Russian phonotactics (Examples from Andersen 2012) 43 Final voicing seems to be on its way out, and not all speakers accept vocatives PART TWO: Russian “new” vocative with consonant clusters that are unusual for Russian (Daniel’ (1) Dissyllabic hypocoristics. (Костя! ⇒) Кость!, (Надя! ⇒) Надь!; 2009) (2) Dissyllabic diminutives. (Ванька! ⇒) Ваньк!, (Машка! ⇒) Машк!; (3) Hypocoristics of three or more syllables. (Наташа! ⇒) Наташ!, (Сережка! ⇒) Сережк!, ...; (4) Kinship terms. (дядя! ⇒) дядь!, (мама! ⇒) мам!,...; (5) Patronymics. (Андреевна! ⇒) Андреевн!, (Николаевна! ⇒) Николавн!,...; (6) Name + patronymic. (Анна Иванвовна! ⇒ Анн Ванна! ⇒) Анн Ванн!, (Марья Александровна! ⇒ Марь Санна! ⇒) Марь Санн!, ...; (7) Common nouns. (девушка! ‘Miss’ ⇒) девушк!, (хозяйка! ‘hostess, landlady’ ⇒) хозяйк!, (ребята! ‘boys, guys’ ⇒) ребят!, (девчата! ‘girls’ ⇒) девчат!, .... BOLDFACE marks items that conflict with Russian phonotactics (Examples from Andersen 2012) 44 Features of Russian “new” vocative that Andersen uses to argue that vocative is not a case • Pragmatic restrictions: Function is primarily pragmatic, not syntactic • Lexical restrictions: Only with forms of address • Syntactic restrictions: Syntactically independent of sentence • Morphophonological restrictions: Limited to words ending in – a with penultimate or prepenultimate stress • Phonological peculiarities: Formed by truncation, resulting in word-final consonant clusters (lacking vowel insertion) and voiced final consonants not otherwise tolerated in Russian • Strong association with diminutives (themselves peculiar) 45 Features of Russian “new” vocative that Andersen uses to argue that vocative is not a case • Pragmatic restrictions: Function is primarily pragmatic, not syntactic • Lexical restrictions: Only with forms of address But Syntactically do these (and other) of sentence • Syntactic restrictions: independent arguments against calling • Morphophonological restrictions: Limited to words ending in – vocative a case really hold? a with penultimate or prepenultimate stress • Phonological peculiarities: Formed by truncation, resulting in word-final consonant clusters (lacking vowel insertion) and voiced final consonants not otherwise tolerated in Russian • Strong association with diminutives (themselves peculiar) 46 Pragmatic restrictions Other examples of cases used to express pragmatic functions: • Virile vs. deprecatory endings for Npl in Polish (Janda 1996) -owie ’-i <soft cons + -y> -y [honorific virile] [neutral virile] [non-virile] profesorowie profesorzy profesory ‘professors’ [wow, I respect those men!] ‘professors’ [they are men] ‘professors’ [they are wimps, not men] • Ethical dative (Janda 1993, Janda & Clancy 2006) Cz: Pustila jsem dceru na hory a ona ti si mi zlomila nohu! ‘I let my daughter go to the mountains and she you-DAT self-DAT me-DAT broke leg!’ [you-DAT: I’m telling you, can you believe it?!] [self-DAT: It’s her leg, she did it to herself.] [me-DAT: Just imagine what this means for me, I’m going to suffer for this!!] 47 Pragmatic restrictions Other examples of cases used to express pragmatic functions: • Virile vs. deprecatory endings for Npl in Polish (Janda 1996) Notice that the ethical datives are also -owie ’-i <soft cons + -y> -y [honorific virile] [neutral integrated virile] [non-virile] not syntactically into the profesorowie profesorzy profesory sentence ‘professors’ [they ‘professors’ [wow, I ‘professors’ [they respect those men!] are men] are wimps, not men] • Ethical dative (Janda 1983, Janda & Clancy 2006) Cz: Pustila jsem dceru na hory a ona ti si mi zlomila nohu! ‘I let my daughter go to the mountains and she you-DAT self-DAT me-DAT broke leg!’ [you-DAT: I’m telling you, can you believe it?!] [self-DAT: It’s her leg, she did it to herself.] [me-DAT: Just imagine what this means for me, I’m going to suffer for this!!] 48 Lexical restrictions There is plenty of evidence of lexical restrictions on case forms: • Loc 2 in Russian, as in в снегý ‘in the snow’: – this case ending is largely restricted to nouns designating concrete locations (“жесткая локализация”, Plungjan 2002) – an alternate ending for 148 nouns, primarily monosyllabic masculine animate nouns with mobile stem stress (Janda 1996) • Gen 2 in Russian, as in выпить чаю ‘drink (some) tea’: – used with only about 1% of masculine inanimate nouns – used with nouns referring to quantifiable substances (Worth 1984; Janda 1996) – currently productive, cf. абрикотин ‘apricot liquor’, анилин ‘aniline’, асбест ‘asbestos’… 49 Morphophonological restrictions There is plenty of evidence of morphophonological restrictions on case forms: • Loc 2 in Russian, as in в снегý ‘in the snow’: – primarily monosyllabic masculine animate nouns with mobile stem stress – there are ten nouns with polysyllabic Nsg forms, but most of these derive from monosyllabic stems via pleophony (bergъ > берег, берегý), diminutive formation (бок, бокý ‘side’ has diminutive бочок, бочкý), or prefixation (cf. порт, портý ‘port’ and аэропорт, аэропортý ‘airport’) (Janda 1996) • 50 NPl –á, as in берег, берегá ‘bank’ – possible only for nouns with accentual patterns that permit end stress in the N(A)pl as opposed to stem stress in the singular (only two exceptions to this rule – two nouns with fixed end stress: рукав, рукавá ‘sleeve’; обшлаг, обшлагá ‘cuff’) – also restricted largely to words that result from pleophony or partially imitate the segmental phonology of pleophonic forms (потрох, потрохá ‘entrail’; соболь, соболя ‘sable’) (Janda 1996, Worth 1983) Phonological peculiarities There is plenty of evidence of phonological peculiarities in case forms: • Bethin 2012 “Reduction of unstressed /o/ and /a/ to [ɐ] or [ə] after non-palatalized consonants and to [ɪ] after palatalized ones in Contemporary Standard Russian (CSR) is systematic. But in certain inflectional suffixes [ə] occurs instead of the expected [ɪ] after palatalized consonants.” – For example, the last vowel in дядя ‘uncle’ should be [ɪ], but it is [ə], despite the fact that this runs counter to prevailing иканье in Contemporary Standard Russian 51 Diachronic peculiarities Cases are lost in different orders, and it is not really true that vocative was preserved while all other cases were lost in Bulgarian & Macedonian • • • 52 The vocative is marginal and optional in both Bulgarian (Girvin 2013) and Macedonian (Friedman 1993) The original Slavic vocative was lost in some Slavic languages where all other cases were preserved (Russian, for example) There is just a lot of variation that doesn’t tell us anything about whether or not vocative is a case References 53