UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Form and function of expressive morphology: a case study of Russian 2008

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
ubc_2008_spring_steriopolo_olga.pdf [ 1.03MB ]
ubc_2008_spring_steriopolo_olga.pdf
Metadata
JSON: 1.0066282.json
JSON-LD: 1.0066282+ld.json
RDF/XML (Pretty): 1.0066282.xml
RDF/JSON: 1.0066282+rdf.json
Turtle: 1.0066282+rdf-turtle.txt
N-Triples: 1.0066282+rdf-ntriples.txt
Citation
1.0066282.ris

Full Text

       Form and Function of Expressive Morphology: A Case Study of Russian  by  OLGA STERIOPOLO      A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE STUDIES  (Linguistics)            THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  February 2008  © Olga Steriopolo, 2008 ii ABSTRACT  In this thesis, I conduct a detailed case study of expressive suffixes in Russian. I show that although the suffixes under investigation have the same function (“expressive”), they differ significantly in their formal properties. I identify two major semantic types of expressive suffixes: attitude and size suffixes. Attitude suffixes convey an attitude of the speaker toward the referent. Size suffixes both convey an attitude and refer to the size of the referent.  I argue that the two different semantic types map onto different syntactic types. Attitude suffixes are syntactic heads, while size suffixes are syntactic modifiers. As heads, attitude suffixes determine the formal properties (syntactic category, grammatical gender and inflectional class) of the derived form. As modifiers, size suffixes do not determine the formal properties of the derived form. Attitude suffixes can attach both to category-free √Roots and to categories (n/a/v), while size suffixes can only attach to a noun category.  I investigate the functional and formal properties of Russian expressive suffixes in a systematic way, which has not been done before. In doing so, I analyze how expressive suffixes pattern along several kinds of criteria (gender/class change, category change, subcategorization). An important byproduct of this analysis is that I show how grammatical gender of an expressive form can be predicted from its inflectional class (combined with animacy and natural gender of the base).  One implication of this analysis is that I show that the formal properties of expressives are no different from those of non-expressives (descriptives), as both expressives and descriptives can attach as heads or modifiers either to √Roots or categories. Another implication is that the formal criteria which I develop for a small set of expressive suffixes in Russian can be extended to set up a cross-linguistic typology of expressives. iii TABLE OF CONTENTS  Abstract……………………………………………………………………………………ii Table of Contents............................................................................................................... iii List of Tables ...................................................................................................................... v List of Abbreviations Used ............................................................................................... vii Acknowledgements............................................................................................................ ix Dedication ........................................................................................................................... x Chapter 1: Introduction ....................................................................................................... 1 Chapter 2: Expressive content and its characteristics ....................................................... 11 2.1. DESCRIPTIVE AND EXPRESSIVE CONTENT ................................................................ 11 2.2. CHARACTERISTICS OF EXPRESSIVE CONTENT .......................................................... 15 2.2.1. Independence of Russian expressive suffixes................................................. 16 2.2.2. Nondisplaceability of Russian expressive suffixes......................................... 21 2.2.3. Perspective dependence of Russian expressive suffixes................................. 23 2.2.3.1. The speaker’s attitude .................................................................................. 24 2.2.4. Descriptive ineffability of Russian expressive suffixes.................................. 26 2.2.5. Immediacy of Russian expressive suffixes ..................................................... 28 2.2.6. Repeatability of Russian expressive suffixes.................................................. 30 2.3. CONCLUSIONS ......................................................................................................... 33 Chapter 3: Background Information ................................................................................. 35 3.1. GENDER .................................................................................................................. 35 3.1.1. Grammatical gender........................................................................................ 35 3.1.2. Natural gender (sex)........................................................................................ 39 3.1.3. Common gender nouns ................................................................................... 41 3.2. INFLECTIONAL CLASS .............................................................................................. 42 3.2.1. The traditional approach to Russian inflectional classes ................................ 43 3.2.1.2. Class II ......................................................................................................... 45 3.2.1.3. Class III ........................................................................................................ 48 3.2.2. Corbett’s (1982, 1991) approach to Russian inflectional classes ................... 49 3.2.2.1. Why the properties of -išč’ favour the traditional approach ........................ 51 3.3. ANIMACY ................................................................................................................ 56 3.4. PHONOLOGICAL ALTERNATIONS.............................................................................. 58 3.4.1. Vowel alternations .......................................................................................... 60 3.4.2. Consonantal alternations................................................................................. 59 Chapter 4: How are expressive suffixes merged?............................................................. 63 4.1. INTRODUCTION........................................................................................................ 63 4.2. CHANGE IN CATEGORY............................................................................................ 64 4.2.1. Attitude suffixes.............................................................................................. 65 4.2.2. Size suffixes .................................................................................................... 67 4.2.3. Summary......................................................................................................... 69 4.3. CHANGE IN GRAMMATICAL GENDER ....................................................................... 70 4.3.1. Attitude suffixes.............................................................................................. 71 4.3.1.1. Animate nouns ............................................................................................. 71 4.3.1.2. Inanimate nouns ........................................................................................... 83 4.3.1.3. Summary...................................................................................................... 86 iv 4.3.2. Size suffixes .................................................................................................... 87 4.4. CHANGE IN INFLECTIONAL CLASS............................................................................ 93 4.4.1. Attitude suffixes.............................................................................................. 93 4.4.1.1. Change in inflectional class ......................................................................... 94 4.4.1.2. Correlation between change in class and change in gender......................... 97 4.4.2. Size suffixes .................................................................................................. 104 4.5. NON-EXPRESSIVE SUFFIXES HOMOPHONOUS WITH SIZE SUFFIXES ......................... 111 4.5.1. The suffix -išč’ .............................................................................................. 111 4.5.2. The suffix -ec ................................................................................................ 113 4.5.3. The suffix -k .................................................................................................. 115 4.5.4. Summary....................................................................................................... 116 4.6. CONCLUSIONS ....................................................................................................... 116 Chapter 5: Where are expressive suffixes merged?........................................................ 118 5.1. INTRODUCTION...................................................................................................... 118 5.2. ATTITUDE SUFFIXES .............................................................................................. 119 5.2.1. Merging with √Roots .................................................................................... 120 5.2.2. Merging with categories ............................................................................... 122 5.2.3. Merging with short names (hypocoristics) ................................................... 128 5.2.4. Summary....................................................................................................... 133 5.3. SIZE SUFFIXES ....................................................................................................... 134 5.3.1. Merging with nouns ...................................................................................... 134 5.3.2. Size suffixes do not merge with adjectives or verbs..................................... 138 5.3.3. No merging with √Roots............................................................................... 140 5.4. INTERMEDIATE SUMMARY..................................................................................... 143 5.5. CO-OCCURRENCE RESTRICTIONS ........................................................................... 144 5.6. CONCLUSIONS ....................................................................................................... 148 Chapter 6: Expressive morphology across languages..................................................... 149 6.1. EXPRESSIVE HEADS THAT CAN ONLY MERGE WITH NOUNS .................................... 151 6.2. EXPRESSIVE MODIFIERS MERGING WITH √ROOTS .................................................. 154 6.3. EXPRESSIVE MORPHOLOGY IN THE HIGHER NOMINAL DOMAIN .............................. 157 6.3.1. Expressive morphology merging with Number............................................ 158 6.3.2. Expressive morphology merging with Determiner....................................... 162 6.3.3. Summary....................................................................................................... 163 6.4. EVIDENTIALS ........................................................................................................ 164 6.5. CONCLUSIONS ....................................................................................................... 168 Chapter 7: Conclusions and topics for further research.................................................. 171 7.1. CONCLUSIONS ....................................................................................................... 171 7.2. TOPICS FOR FURTHER RESEARCH........................................................................... 172 7.2.1. Complex expressive suffixes ........................................................................ 172 7.2.1.1. The suffix -en’k/-on’k ................................................................................ 173 7.2.1.2. The suffix -išk ............................................................................................ 175 7.2.2. Delimitative verbs and delimitative prefix po- ............................................. 176 7.2.3. Further issues ................................................................................................ 178 References....................................................................................................................... 180 Appendix......................................................................................................................... 190  v LIST OF TABLES  Table 2.1 Affectionate and vulgar suffixes in Russian………………………….…. 16 Table 2.2 Size suffixes in Russian…………………………………………………. 21 Table 2.3 Characteristics of expressive suffixes in Russian……………………….. 34 Table 2.4 Attitude suffixes are expressive…………………………………………. 34 Table 2.5 Size suffixes are expressive and descriptive………………………….…. 34  Table 3.1 Russian grammatical genders (Corbett 1991:35)………………………... 36 Table 3.2 Masculine and feminine nouns, by phonological form of stem……….… 36 Table 3.3 Singular vs. plural agreement with an adjective…………………….…... 39 Table 3.4 Kinship terms………………………………………………………….… 39 Table 3.5 Personal names……………………………………………………….…..39 Table 3.6 Domesticated animals using three nouns for species, male, and female………………………………………………... 40 Table 3.7 Inflectional classes in Russian (“traditional approach”)……………….... 44 Table 3.8 Inflectional Class II (feminine, masculine, and common gender nouns).. 46 Table 3.9 Inflectional Class III ……………………………………………………. 49 Table 3.10 Inflectional classes in Russian, according to Corbett (1982, 1991)….….. 50 Table 3.11 Masculine animate and inanimate nouns (Class I)…………………….... 58  Table 4.1 Attitude suffixes……………………………………………………….… 63 Table 4.2 Size suffixes……………………………………………………….…….. 63 Table 4.3 Diagnostics for syntactic heads vs. syntactic modifiers…………………. 64 Table 4.4 Attitude suffixes (change in category)……………………………….….. 67 Table 4.5 Size suffixes (no change in category)………………………………….... 69 Table 4.6 Comparison of attitude and size suffixes……………………….……….. 70 Table 4.7 Change in category………………………………………………….…... 70 Table 4.8 Attitude suffixes (used with TYPE II nouns)………………………….…. 79 Table 4.9 Attitude suffixes (used with TYPE I nouns)……………………….……...81 Table 4.10 Attitude suffixes (used with TYPE III nouns)……………………………. 82 Table 4.11 Attitude suffixes (animate nouns)………………………………….……. 83 Table 4.12 Attitude suffixes (inanimate nouns) …………………………………….. 86 Table 4.13 Attitude suffixes (change in grammatical gender)……………….……… 87 Table 4.14 Size suffixes (no change in grammatical gender)……………………...... 92 Table 4.15 Inflectional endings in Nom. sg………..………………………………... 94 Table 4.16 Attitude suffixes (change in inflectional class)…………………….……. 97 Table 4.17 Class membership of attitude suffixes……………………………...…… 97 Table 4.18 Size suffixes (no change in inflectional class)………………………..... 106 Table 4.19 Contrasting examples (Class II and Class III nouns)…………….…….. 107 vi Table 4.20 Final consonants of Class III stems………………………….………… 108 Table 4.21 Non-expressive -išč’ and expressive -išč’………………………..……..113 Table 4.22 Non-expressive -ec and expressive -ec…………………………..…….. 115 Table 4.23 Non-expressive -k and expressive -k………………………………….... 116 Table 4.24 Syntactic heads vs. syntactic modifiers………………………….…….. 117 Table 4.25 Expressive suffixes in Russian……………………………………….…117  Table 5.1 Attitude suffixes (attachment to √Roots)………………………….…… 122 Table 5.2 Attitude suffixes (merging with adjectives)………………………….… 124 Table 5.3 Attitude suffixes (merging with verbs)……………………………….... 127 Table 5.4 Attitude suffixes (merging with categories)………………………….…128 Table 5.5 Attitude suffixes (attachment to √Roots and to categories)……….…… 134 Table 5.6 Size suffixes (merging with nouns)……………………………….…….138 Table 5.7 Size suffixes (no merging with adjectives)…..………………………… 139 Table 5.8 Size suffixes (no merging with verbs)…………………………….…… 140 Table 5.9 Size suffixes (attachment to n, a, v, and √Roots)…………………….… 143 Table 5.10 Attitude vs. size suffixes (difference in attachment sites)……………....143  Table 6.1 Distributional difference #1 between attitude and size suffixes in Russian……………………………………………….... 149 Table 6.2 Distributional difference #2 between attitude and size suffixes in Russian………………………………………………… 150 Table 6.3 Asymmetries between attitude and size suffixes in Russian…………... 150 Table 6.4 Syntactic heads that can only merge with a noun category (German).… 154 Table 6.5 Expressive morphology in Russian, German, and Halkomelem…….… 157 Table 6.6 Tongan determiners…………………………….………………….……162  vii LIST OF ABBREVIATIONS USED  a adjective ACC Accusative case adj  adjective affect affectionate anim animate ANTIHON antihonorific ASP aspect aug augmentative BP Brazilian Portuguese CAUS causative CLASS inflectional class COP copular D, DET  determiner DAT  Dative case DEFIN  definite DELIM.PREF  delimitative prefix dim diminutive DIRC  direct EMOT emotional ERG ergative EVID evidence/evidential EXPR expressive F functional projection FEM feminine GEN Genitive case INDEF indefinite INF infinitive INST Instrumental case LEX lexical prefix LOC Locative case viii MASC  masculine MASC/FEM masculine or feminine n noun NEUT neuter N.SG/Nom sg.                                 Nominative singular NOM  nominalizer NONEXP non-experienced non-EXPR non-expressive OOC out-of-control PERF perfective PL plural PREP preposition PRES present PROG progressive SG  singular SUBJ.HON subject honorific SUFF suffix SUPERLEX superlexical prefix TH   theme vowel v  verb VERB.PREF verbal prefix vulg vulgar ix ACKNOWLEDGEMENTS  I would like to thank my supervisor, Martina Wiltschko, for her huge help with this thesis. Thank you, Martina, for all your advice and suggestions for this work. Your supervision has made doing this work a true pleasure and a lot of fun. Thank you for believing in me from the start and for taking a lot of time away from your family in order to help me with this research. It is your enthusiasm that has made me fall in love with my work. Thank you very much for teaching me the “structure”; it will be my great pleasure to take it with me wherever I end up in life.  Thank you very much to my committee members, Gunnar Hansson and Hotze Rullmann for all your comments and suggestions. Thank you for your help with the difficult issues in phonology and semantics, which I would have never figured out without your help. Thank you for devoting many hours to reading this work and for giving me very detailed and helpful comments. And thank you for all your encouragement and moral support, which were so welcomed and appreciated during both difficult and happy times as I worked on this thesis.  Thank you to Henry Davis for fighting through the bureaucratic difficulties in order to enroll me in the Ph.D. program in the first place. Thank you for giving me this great chance to earn my Ph.D. degree from the University of British Columbia. I would also like to thank you for supervising my work on Nuu-chah-nulth. You have made me fall in love with this wonderful language, as well as the First Nations culture and people. Although I am not currently working on First Nations languages, I will always keep a place in my heart to appreciate their fascinating complexity.  Thank you to Rose-Marie Déchaine for all your help with many topics I worked on during my graduate program. You have always found time for me, no matter how busy you are. Your advice always pushed me in the right direction when I was hopelessly lost. It has always been a great pleasure for me to discuss my work with you. I also would like to thank you for helping me out financially during the first years in my Ph.D. program when I needed support the most. x Thank you to Patricia Shaw, Joseph Stemberger, and Douglas Pulleyblank for your huge help with my work on “Jer vowels in Russian prepositions”. I also would like to thank you for your everlasting encouragement, support and for making me fall in love with the Russian phonology.  Thank you to Lisa Matthewson for your financial support during the difficult times. Thank you for never letting me go without help, in spite of your own difficulties and many students that also relied on your support.  Many thanks to Katrina Petrik for her huge help with editing of this work. Thank you, Katrina, for devoting many long hours to proofreading and formatting this thesis and for always finding time to do all this work under the biggest time pressure.  Thank you very much to Edna Dharmaratne for helping me to survive my many years in the Ph.D. program. Edna, you have become my true friend and my guardian angel since the first day I set my foot in this department. Thank you for always being there for me, through the good and especially though the bad times. And thank you for always finding kind words for me when I most needed them. I will always remember your never-ending support and encouragement.   DEDICATION             xi            To my mother,  Elena Steriopolo                                                          Thank you, mamichka, for being there for me all these                                                               years and for your eternal love and support! 1 Chapter 1: Introduction  1.1. EXPRESSIVE CONTENT AND EXPRESSIVES An utterance can have descriptive content and expressive content. For example, the descriptive content of the English utterance in (1) is the proposition that Kresge is famous, as illustrated in (i). The expressive content conveys the negative attitude of the speaker toward Kresge, as illustrated in (ii).  (1) That bastard Kresge is famous.  i. Descriptive content:   ‘Kresge is famous’ ii. Expressive content:   ‘Kresge is a {bastard/bad in the speaker’s opinion}’ (Potts 2007:3)  To give another example, the descriptive content of the Japanese utterance in (2) is the proposition that Sam laughed (i). The expressive content conveys the positive attitude of the speaker toward Sam (ii).  (2)     Sam-ga       o-warai-ninat-ta.           Sam-N.SG    SUBJ.HON-laugh-SUBJ.HON-PAST           ‘Sam laughed’         i. Descriptive: ‘Sam laughed’        ii. Expressive: ‘The speaker honours Sam’ (Potts & Kawahara 2004:1)  Expressives such as bastard (English) and o-…-ninat (Japanese honorific) are linguistic objects that convey the attitude of the speaker toward the referent. What makes expressives particularly interesting to investigate is the fact that their meaning (“function”) appears to be similar across languages (expressive), but their syntactic structure (“form”) differs significantly from one language to the next (e.g., a separate word in English, but circumfix in Japanese).   2 1.2. RUSSIAN EXPRESSIVE SUFFIXES Russian employs many expressive suffixes to convey the speaker’s attitude. According to Polterauer (1981), there are over 30 simplex and complex expressive suffixes. For the purpose of this thesis, I only investigate simplex expressive suffixes (a brief discussion of complex expressive suffixes is given in Chapter 7). Russian has been chosen as the language of investigation because it has many suffixes that have the same function (expressive), as well as the same morphological form (suffixes). Thus, Russian is an ideal candidate to investigate whether there is a correlation between the form and function of expressives.  Expressive suffixes in Russian present certain puzzles in terms of both their functional and formal properties. For example, although they are associated with the same expressive function, there are differences in meaning. Some expressive suffixes only convey the attitude of the speaker toward the referent (3), while others can both convey the attitude and refer to the size of the referent (4).  (3)  d’ed-úl’-a                         pr’išól       grandfather-EXPR-N.SG     came       ‘Grandfather came (affectionate attitude)’         i. Descriptive: ‘Grandfather came’        ii. Expressive:  ‘The speaker feels affection toward the grandfather’  (4)  zv’er’-ók                   pr’išól       animal-EXPR.N.SG       came        ‘(The) animal came (affectionate attitude & small size of the referent)’         i. Descriptive: ‘The small animal came’        ii. Expressive:  ‘The speaker feels affection toward the animal’  In terms of their formal properties, some expressive suffixes seem to change the formal properties of the nominal base (e.g., category, gender, inflectional class), while others do not. For example, in (5), the expressive suffix -ux changes the category of the base from adjective to noun; while in (6), the expressive suffix -ok does not (the example is ill-formed).  3 (5) adj  noun        a.  gr’áz-n-ij                                     b.  gr’az-n-úx-a             dirty-ADJ-MASC.SG                                  dirty-ADJ-EXPR-N.SG             ‘dirty’                                             ‘dirty animate’  (6) *adj  noun        a.  gr’áz-n-ij                                     b.  *gr’az-n-ok             dirty-ADJ-MASC.SG                                    dirty-ADJ-EXPR.N.SG             ‘dirty’                                               ‘dirty animate’  With respect to (3)–(6), the following question arises: Are we dealing with only one class or different classes of expressives suffixes in Russian? If there are different classes, how do we distinguish them? Despite the fact that a great deal of descriptive research has been devoted to individual expressive suffixes in Russian (Bratus 1969; Dementiev 1953; Fentslova 1985; Ivanova 1965; Kolomiets 1988; Kosmeda 1999; Mandelshtam 1903; Ogol’cev 1960; Plyamovataya 1955, 1961; Polterauer 1981; Popoff-Böcker 1973; Popov 1967; Protasova 2001; Rakušan 1981; Shvedova et al.1982; Spiridonova 1999; Stankiewicz 1968; Vaseva 1977, among others), Russian expressive suffixes have not yet been analyzed in a systematic way.  Here, I systematically study the functional and formal properties associated with Russian expressive suffixes by (i) analyzing their meaning and (ii) determining whether or not they can change formal properties of the base.  1.3. FINDINGS I show that there are two major semantic types of expressive suffixes in Russian: attitude and size suffixes. Attitude suffixes only convey an attitude (examples 3 and 5 above); while size suffixes both convey an attitude and refer to the size of the referent (example 4 above).  I argue that the formal properties of Russian expressive suffixes vary along two dimensions: (i) how they attach—as a head or as a modifier—and (ii) where they attach—to category-free √Roots (in the sense of Marantz 1997; notation from Pesetsky 1995) or to categories.  4 I further argue that the two semantic types (attitude vs. size) map onto two distinct syntactic types. Attitude suffixes are syntactic heads (7a), while size suffixes are syntactic modifiers (7b).  (7) a. HEADS            X                  b. MODIFIERS             Y                      2                                        2                    X            Y                                         X           Y                     EXPRattitude                                                   EXPRsize  Attitude suffixes are noun heads that can attach to √Roots (8a) or to various syntactic categories (n/a/v) (8b). Size suffixes are noun modifiers that can only attach to a noun category (9).  (8) HEADS      a.              n                                       b.          n                      2                                          2                    n          √Root                                     n            n/a/v                  EXPRattitude                                         EXPRattitude   2                                                                                n/a/v        √Root   (9) MODIFIERS                                n                                                   2                                          EXPRsize          n                                                                    2                                                        n         √Root   1.4. BYPRODUCTS OF THE ANALYSIS In order to investigate the formal properties of Russian expressive suffixes, I analyze how they pattern along several kinds of criteria (gender/class change, category change, subcategorization). A byproduct of this analysis is that the grammatical gender of an expressive form can be predicted from its inflectional class (combined with animacy and natural gender of the base). It has been claimed in the literature that Russian grammatical gender can be predicted from inflectional class (Corbett 1982, 1991; Corbett & Fraser 2000). 5 This thesis systematically shows how this works with respect to expressive forms, which has never been done before.  Another byproduct of this analysis is that Russian expressive suffixes differ with respect to whether or not they can have descriptive content in addition to expressive content. I show that size suffixes can have both expressive and descriptive content, while attitude suffixes can have only expressive content. To the best of my knowledge, Russian expressive suffixes have never been studied before in terms of their descriptive vs. expressive content.  1.5. THEORETICAL FRAMEWORK In this thesis, I assume the Principles and Parameters (P&P) framework, which contrasts with descriptivist frameworks focusing on a particular language of investigation. The descriptivist frameworks view categorization in terms of inflection vs. derivation, but this has been proven problematic with respect to the behaviour of expressives (Dressler & Barbaresi 1994; Manova 2004; Scalise 1984, 1988; Vinogradov 1972). It has been shown in the literature that the behaviour of expressives is not wholly inflectional or derivational. In contrast, the P&P framework regards inflection and derivation not as primitives, but as derived notions, and thus, this framework can better account for the behaviour of expressives.  For the purpose of this thesis, I assume a model of grammar in which syntax and morphology are analyzed as a single engine, as for example, in the framework of Distributed Morphology (DM) (Halle & Marantz 1993; Halle 1997; Marantz 1997; Harley & Noyer 1999, 2003; Marantz 2001; Bobaljik 2002; Marvin 2002; Arad 2003; Embick & Noyer 2005; Müller 2005; Halle & Matushansky 2006, among others). The particular assumption I adopt is that words are built by the same principles as phrases and sentences—by syntactic principles.  Another assumption I adopt is in regards to the treatment of √Roots and syntactic categories. √Roots are language-specific combinations of sound and meaning, such as √break- or √cat- in English. √Roots have no category per se, but can never appear “bare”: they have to be categorized by combining with a category-defining functional head, such as the “little” n, a, or v, to form nouns, adjectives, or verbs, respectively. A single √Root can be assigned to 6 more than one category, for example: the break  (noun) in the glass and John breaks  (verb)  the glass. The category-defining functional heads are determined either by phonologically realized or zero affixes (10).  (10)                 n                   2                  n         √cat-                 -Ø  By hypothesis, the grammar makes reference to so-called dissociated features (Embick 1997, 1998; Embick & Noyer 2005). Dissociated features are absent in the syntax but are inserted at PF (Phonological Form). One example of a dissociated feature is morphological case. According to Embick (1997, 1998), morphological case is a feature that is added at PF and conditions the choice of the vocabulary item that expresses case. For example, in Latin, morphological case and number are fused in one morpheme (femin-a ‘woman-N.SG’). According to this approach, only Number (#) is present in a syntactic structure. Morphological case, however, does not figure in the syntax and is added at PF (11).  (11)    Syntactic structure for a number morpheme                           #                     2            # [+/- Pl]         n                            2                             n          √Root  Like in Latin, Russian case and number are also fused in a single morpheme, thus I assume that the structure in (11) also applies to Russian. For example, for the Russian word sobak-a ‘dog-N.SG’, I assume the following structure (12).  7  (12) Structure for sobak-a ‘dog-N.SG’                            #                      2                   #               n                  -a         2                              n           √sobak   Under the assumption that category labels are independent of √Roots, we expect two different sites for building words in the syntax: (i) words are created from √Roots, and (ii) words are created from other words. Thus, a category head X (e.g., n, a, v) may merge either with a √Root, as illustrated in (13a), or with a pre-existing word, as illustrated in (13b).  (13) a.     X                             b.    X           2                           2        X            √Root                 X           n/a/v                                                         2                                                    n/a/v       √Root  The distinction between word formation from √Roots and word formation from other words is a universal distinction, but its manifestations may differ from language to language (Marantz 2001). For example, in English, many category-defining affixes attach to √Roots. This is evident from a mechanism that Aronoff (1976) calls “truncation” (compare atroc-ious and atroc-ity1). Marantz (2001) claims that what seems like truncation reflects word formation from √Roots. The √Root atroc- creates atroc-ious in an adjectival environment (suffix -ous) and it creates atroc-ity in a nominal environment (suffix -ity) (14).  (14) a.           a                             b.         n                 2                               2               a          √atroc-                       n         √atroc-              -ous                                         -ity   1  Where ‘atrocity’ is derived from the adjective ‘atroc-ious’. 8 Marantz (2001) also shows that there are suffixes in English that can attach both to a √Root and to a category. The nominalizing suffix -or is one such example: don-or, don-at-or (derived from the verb don-ate) (15).  (15) a.    n                                    b.     n          2                                  2         n         √don-                             n            v        -or                                         -or      2                                                                   v         √don-                                                                -ate  This thesis provides additional empirical support for a distinction between word formation from √Roots and word formation from categories. I show that both structures (15a) and (15b) are productively used by Russian expressive suffixes.  1.6. IMPLICATIONS OF THE ANALYSIS One implication of this analysis is that the formal criteria which I develop for a small set of expressive suffixes in Russian can be extended to set up a typology of expressives. Despite the fact that much research has been devoted to expressives across languages (Dressler & Barbaresi 1994; Dressler 1997; Ettinger 1974, 1980; Gillis 1998; Helbig 1974; Schneider 2003; Beard and Szymanek 1988; Kalasniemi 1992; Voeykova & Dressler 2002; Volek 1987, among others), such formal criteria do not currently exist. As such, this work has the potential to trigger a whole new direction of research on expressives across languages.  Another implication concerns the question of whether the formal properties of expressives are unique/special, as opposed to non-expressives (descriptive linguistic objects). I show that expressives do not have any special formal properties that would set them apart from non- expressives. Their formal properties are exactly like those of non-expressives, because both expressives and non-expressives differ with respect to (i) how they attach (as a head or modifier), and (ii) where they attach (to category-free √Roots or categories). If the formal properties of expressives are exactly the same as those of non-expressives, we expect that one and the same linguistic object might have both expressive and descriptive content. This is exactly what we find in Russian expressive size suffixes (see §1.4 above). 9  1.7. STRUCTURE OF THE THESIS The thesis is organized as follows. In Chapter 2, I discuss expressive content and its characteristics. I study the semantic function of Russian expressive suffixes by analyzing what they contribute to meaning. I show that expressive suffixes fall into two major semantic types: attitude and size suffixes. I also discuss how attitude and size suffixes differ with respect to whether or not they can have descriptive content in addition to expressive content.  In Chapter 3, I outline the background of Russian morphology. I discuss grammatical and natural gender, inflectional class, and animacy in Russian.  In Chapter 4, I investigate how expressive suffixes attach—as a head or as a modifier. I show that attitude suffixes are syntactic heads, while size suffixes are syntactic modifiers. By doing so, I show in a systematic way how grammatical gender of an expressive form is predictable on the basis of inflectional class, animacy, and natural gender (sex). I also discuss non- expressive suffixes that are homophonous with size suffixes in Russian. I show that the non- expressive suffixes differ from size suffixes both in meaning and in syntactic structure.  In Chapter 5, I investigate where expressive suffixes attach (to category-free √Roots or categories). I show that attitude suffixes can attach both to √Roots and to categories, while size suffixes can only attach to categories. I also analyze co-occurrence restrictions in both attitude and size suffixes, and show that all predictions about these co-occurrence restrictions are borne out in Russian.  In Chapter 6, I discuss expressive morphology across languages. I show that the expected types of expressive morphology that follow from the current analysis are found cross- linguistically. I analyze expressives in the following languages: German (Germanic), Halkomelem (Salish), Brazilian Portuguese (Romance), Southern Barasano (Eastern Tucanoan language of Colombia), Welsh (Celtic), and Tongan (Polynesian). I also discuss evidentials and whether they are part of the same system as expressives.  10 In Chapter 7, I present the conclusions and discuss topics for further research, including complex expressive suffixes in Russian, the delimitative prefix po-, and further issues like stress and palatalization.  In the Appendix, I give examples of attitude suffixes which show the effects of each individual suffixes on the formal properties of linguistic objects it attaches to. Examples of size suffixes are described and analyzed in the body of the thesis.  This thesis is aimed at morphologists on both empirical and theoretical levels. It will be of interest to semanticists and typologists, as well as to anybody interested in general linguistics. The thesis contains a large amount of fascinating data that can be used by teachers and students of Russian. 11 Chapter 2: Expressive content and its characteristics  In this Chapter, I examine how to determine whether a given linguistic object (LO) is expressive or descriptive. I start with a brief discussion of the differences that set apart descriptive from expressive meanings (§2.1). I then introduce several diagnostic properties of expressive linguistic objects (LOs) as developed in Potts (2007), and I show that Russian expressive suffixes satisfy these properties. In §2.3, I present the conclusions.  2.1. DESCRIPTIVE AND EXPRESSIVE CONTENT The expressive content of an utterance conveys information about the attitudes and emotions of the speaker toward the content of the utterance (Potts 2003:1). The expressive content is usually secondary to the descriptive content of the utterance, but can have a significant impact on discourse. This is illustrated in the following examples from Japanese. In (1), the subject honorific o-…-ninat is used. The descriptive content of this sentence is that Sam laughed. The expressive content is that the speaker views Sam with honour.  (1)  Sam-ga       o-warai-ninat-ta.           Sam-N.SG    SUBJ.HON-laugh-SUBJ.HON-PAST           ‘Sam laughed         (with honorific)’    i.   Descriptive:  ‘Sam laughed’   ii.  Expressive:  ‘Speaker views Sam with honour’    (Potts & Kawahara 2004:1)  In contrast, in (2), the antihonorific -yagat is used. The descriptive content of this sentence is that Sam laughed. The expressive content is that the speaker does not view Sam with honour.  (2)      Sam-ga       warai-yagat-ta.            Sam-N.SG    laugh-ANTIHON-PAST           ‘Sam laughed        (with antihonorific)’   i.    Descriptive:  ‘Sam laughed’  ii.   Expressive:  ‘Speaker does not view Sam with honour’ (Potts & Kawahara 2004:1)  12 The honorific o-…-ninat and antihonorific -yagat in (1) and (2) are markers of the expressive content (Potts & Kawahara 2004:2). An example from English is the expressive attributive adjective damn. In (3), the expressive damn indicates that the speaker views the Republicans negatively. The Japanese and English examples show that expressives can be bound (as in Japanese) or free (as in English) forms.  (3)  Bush says the damn Republicans deserve public support.   i.    Descriptive:  ‘Bush says the Republicans deserve public support’  ii.   Expressive:  ‘Speaker views the Republicans negatively’    (Potts 2003:6)  The examples above show that expressive content can be positive or negative: the honorific o-…-ninat is a marker of positive expressive content, while the antihonorific -yagat and the attributive adjective damn are markers of negative content.  In Russian, there are a large number of suffixes that serve to convey expressive content. They can express both positive and negative attitudes. For example, in (4a), the expressive suffix -ul’ indicates that the speaker views his/her grandfather positively. The form with this suffix contrasts with the unmarked form in which expressive content is absent (4b).  (4)  a.  d’ed-úl’-a                        pr’išól           grandfather-EXPR-N.SG    came           ‘Grandfather came’   i.    Descriptive:  ‘Grandfather came’  ii.   Expressive:  ‘Speaker views grandfather positively’        b.  d’éd                             pr’išól           grandfather.N.SG            came           ‘Grandfather came’   i.    Descriptive:  ‘Grandfather came’  Similarly, in (5a), the expressive suffix -ug indicates that the speaker views the thief negatively; compare this with (5b), where expressive content is absent.  13 (5)  a.  vor’-úg-a                   pr’išól           thief-EXPR-N.SG           came           ‘(The) thief came’   i.    Descriptive:  ‘The thief came’  ii.   Expressive:  ‘Speaker views the thief negatively’        b.  vór                   pr’išól           thief.N.SG            came          ‘(The) thief came’   i.    Descriptive:  ‘The thief came’  Positive expressive suffixes in Russian indicate an attitude of affection and tenderness toward the referent on the part of the speaker (Bratus 1969; Efremova 2006; Kosmeda 1999; Shvedova et al. 1982). For this reason, I call them affectionate (affect) suffixes. Examples of these types of suffixes are given in (4 above) and (6 below). In (6a), the affectionate suffix -us’ indicates that the speaker views his/her grandmother with affection; compare this with the neutral statement in (6b).  (6)  a.  bab-ús’-a                        pr’išlá           gandmother-EXPR-N.SG    came           ‘Grandmother (affect) came’   i.    Descriptive:  ‘Grandmother came’  ii.   Expressive:  ‘Speaker views grandmother positively’        b.  báb-a                         pr’išlá           grandmother-N.SG      came           ‘Grandmother came’   i.    Descriptive:  ‘Grandmother came’  Evidence that this suffix indeed conveys the affection of the speaker stems from the fact that this affection cannot be denied. In (7a), the speaker denies his/her affection toward the referent grandmother used with an affectionate suffix. As a result, the sentence is infelicitous; compare this with the felicitous (7b), where the speaker expresses his/her affection toward grandmother.  14 (7)  a. # Já   n’e     l ’ubl’ú    svojú     bab-ús’-u.             I     not     love         self’s     grandmother-EXPR-N.SG             ‘I do not love my grandmother (affect)’  i.    Descriptive:  ‘Speaker does not love his/her grandmother’      ii.   Expressive: ‘Speaker views his/her grandmother positively’        b.   Já    l’ubl’ú     svojú    bab-ús’-u.             I      love         self’s    grandmother-EXPR-N.SG             ‘I love my grandmother (affect)’        i.    Descriptive: ‘Speaker loves his/her grandmother’       ii.   Expressive:  ‘Speaker views his/her grandmother positively’  Negative expressive suffixes indicate a vulgar attitude toward the referent (Bratus 1969; Efremova 2006; Kosmeda 1999; Shvedova et al. 1982). Such suffixes are only used informally and they are very productive in colloquial speech.2 For this reason, I call them vulgar (vulg) suffixes. For example, in (8a), the vulgar suffix -an indicates that the speaker views an old man with contempt; compare this with the neutral statement in (8b), where -an is absent.  (8)  a.  star’-ik-án                  pr’išól           old-NOM-EXPR.N.SG     came           ‘(The) old man came (vulg)’   i.    Descriptive:  ‘The old man came’  ii.   Expressive:  ‘Speaker views the old man negatively’        b.  star’-ík                 pr’išól           old-NOM.N.SG        came          ‘(The) old man came’   i.    Descriptive:  ‘The old man came’  Vulgar suffixes are normally not used in a formal setting. This is illustrated in (9). In (9a), the vulgar suffix -an is used in a formal setting, indicated by the formal form of the verb pozvol’- t’e ‘let-formal’ and by the formal pronoun Vam ‘you (formal)’. As a result, the sentence is infelicitous. In (9b), the same vulgar suffix -an is used in an informal setting, indicated by the  2  See Wierzbicka (1984:128–129) on a similar meaning of Australian depreciatives. 15 informal form of the verb smotr’-í ‘look-informal’ and by the informal pronoun t’eb’e ‘you (informal)’. The sentence in (9b) is felicitous.  (9)  a. # Pozvól’-t’e      Vám               predstáv’it’  étogo  star’-ik-án-a.            Let-formal      you(formal)  introduce      this     old-NOM-EXPR.ACC.SG                ‘Let me introduce you to this old man (vulg)’        b.  Smotr’-í,             vót    star’-ik-án,                o          kotórom  já   t’eb’é               govor’íl.             Look-informal   here  old-NOM-EXPR.N.SG  about   who         I     you(informal) told               ‘Look, here is the old man (vulg), I told you about’  In Table 2.1, I list simplex affectionate and vulgar suffixes in Russian. Affectionate suffixes -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’, -úš Vulgar suffixes -ág, -ák, -ál, -án, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx Table 2.1: Affectionate and vulgar suffixes in Russian  2.2. CHARACTERISTICS OF EXPRESSIVE CONTENT According to Potts (2007), expressive content has the following characteristics (10).  (10) Characteristics of expressive content (Potts 2007:2): 1. Independence: Expressive content contributes a dimension of meaning that is separate from the regular descriptive content. 2. Nondisplaceability: Expressives predicate something of the utterance situation. 3. Perspective dependence: Expressive content is evaluated from a particular perspective. In general, the perspective is the speaker’s, but there can be deviations if conditions are right. 4. Descriptive ineffability: Speakers are never fully satisfied when they paraphrase expressive content using descriptive, i.e., non-expressive, terms. 5. Immediacy: Like performatives, expressives achieve their intended act simply by being uttered: they do not offer content so much as inflict it. 6. Repeatability: If a speaker repeatedly uses an expressive item, the effect is generally one of strengthening the emotive content, rather than one of redundancy. 16  In what follows I discuss these characteristics in more detail and apply them to Russian data to establish that the suffixes under consideration are indeed expressive.  2.2.1. Independence of Russian expressive suffixes According to Potts (2007), expressive content is independent of the descriptive content of an utterance. Consequently, the former can be changed or removed without affecting the latter. Potts (2006:4) uses an example from Japanese to illustrate this point. In (11a), the antihonorific -chimat is used, adding expressive content (‘It sucks that I overslept’). In (11b), the antihonorific is removed, which immediately removes the expressive content of the phrase. Although the expressive content is changed, the descriptive content that the speaker overslept does not change.  (11) a.  nesugoshi-chimat-ta              overslept-ANTIHON-PAST              ‘It sucks that I overslept (with antihonorific)’    i. Descriptive:  ‘I overslept’   ii. Expressive:  ‘Speaker views oversleeping negatively’   b.  nesugoshi-ta  overslept-PAST  ‘I overslept’    i. Descriptive:  ‘I overslept’   (Potts 2007:4)  Russian expressive suffixes display the same kind of behaviour: they add content that is independent of the descriptive content. For example, in (12a), the affectionate suffix -ul’ expresses the speaker’s affection toward his/her grandmother. In (12b), the affectionate suffix is removed, which also removes the expressive content. Nonetheless, the descriptive content indicating that grandmother came remains unchanged.  (12)  a.  bab-úl’-a                            pr’išlá             grandmother-EXPR-N.SG    came               ‘Grandmother came (affect)’ 17    i. Descriptive:  ‘Grandmother came’   ii. Expressive:  ‘Speaker views grandmother positively’   b.  báb-a                       pr’išlá      grandmother-N.SG   came      ‘Grandmother came’    i. Descriptive:  ‘Grandmother came’  Similarly, in (13a), the vulgar suffix -ug expresses the speaker’s contempt toward the thief. In (13b), the vulgar suffix is removed, which also removes the expressive content. The descriptive content indicating that the thief came does not change.  (13)  a.  vor’-úg-a                 pr’išól             thief-EXPR-N.SG       came               ‘(The) thief came (vulg)’    i. Descriptive:  ‘The thief came’   ii. Expressive:  ‘Speaker views the thief negatively’   b.  vór                  pr’išól      thief.N.SG       came      ‘(The) thief came’    i. Descriptive:  ‘The thief came’  We have now seen that Russian affectionate and vulgar suffixes behave similarly to the Japanese antihonorific in (11): they contribute an expressive meaning that is independent of the descriptive content. This property of Japanese honorifics and antihonorifics has received a lot of attention in the Japanese literature (Kikuchi 1994; Tokieda 1940; Ooishi 1975), where they are identified as a case of “Taiguu Hyoogen” (Attitudinal Expressions).3 Following the Japanese tradition, I am going to call Russian affectionate and vulgar suffixes attitude suffixes, as they express the speaker’s attitude of affection or vulgarity. The attitude suffixes are defined in (14).     3  Cited by Potts (2003:4). 18 (14)  Definition of attitude suffixes:  Russian attitude suffixes express the speaker’s attitude (affection or vulgarity) toward the referent and contribute an expressive meaning to the phrase.  In addition to attitude suffixes that have expressive content, there is another type of expressive suffix in Russian: suffixes that have both expressive and descriptive content. Such suffixes express the speaker’s attitude and indicate the size of the referent (small or big) (Apres’an 1995; Kosmeda 1999; Mandel’štam 1903; Popov 1967; Spiridonova 1999; Stankiewicz 1954, 1968; Volek 1987, among others), as shown in (15). In (15a), the sentence contains only descriptive content. In (15b), the expressive suffix -ik is added to dóm ‘house’, which indicates both the small size of the house (descriptive content) and the positive attitude of the speaker toward the house (expressive content).  (15)  a.  Dóm            stoít     na     gor’é.              house.N.SG  stands   on    mountain                 ‘(A) house stands on a mountain’               i. Descriptive:  ‘A house stands on a mountain’          b.  Dóm’-ik                 stoít       na     gor’é.             house-EXPR.N.SG   stands    on     mountain                ‘(A) house (size) stands on a mountain’               i. Descriptive:  ‘A small house stands on a mountain’              ii. Expressive:  ‘Speaker views the house positively’  Example (16a) shows that dóm’-ik cannot be modified by the adjective ‘huge’; compare this with (16b), where it can be modified by ‘small’. Example (17a) shows that dóm’-ik cannot be used in a context when the speaker denies his/her positive attitude toward the referent house; compare this with (17b), where the speaker expresses a positive attitude.  (16)  a. # Ogrómnij  dóm’-ik                 stoít     na   gor’é.               huge          house-EXPR.N.SG  stands  on   mountain               ‘A huge house stands on a mountain’          b.   Mál’en’kij    dóm’-ik                stoít     na   gor’é.               small            house-EXPR.N.SG  stands  on   mountain 19               ‘A small house stands on a mountain’  (17)  a. # Já    n’enav’ížu    náš     dóm’-ik.               I      hate               our     house-EXPR.N.SG              ‘I hate our house’          b.   Já    l’ubl’u    náš      dóm’-ik.              I      love         our     house-EXPR.N.SG             ‘I love our house’  To distinguish this type of expressive suffix from the attitude suffixes described above, I call them size suffixes, because they indicate the size of the referent in addition to the speaker’s attitude. Size suffixes are defined in (18).  (18) Definition of size suffixes: Russian size suffixes express the speaker’s attitude (positive or negative) toward the referent, and they also indicate the size (small or big) of the referent. Thus size suffixes contribute both expressive and descriptive content.  In the literature, size suffixes are referred to as diminutive (dim) or augmentative (aug) (Derkach 2005; Nesset 2003; Polterauer 1981; Popov 1967; Shvedova et al. 1982; Stankiewicz 1954, 1968; Wade 2000, among others). Diminutive suffixes indicate the small size of a referent, while augmentative suffixes indicate the big size of a referent. Examples of diminutive suffixes were given in (15)–(17) above. I have shown that diminutive suffixes can express the positive attitude of the speaker; less frequently, they can also express the negative (pejorative) attitude of the speaker (Shvedova et al. 1982:210). For example, in (19a), id’éj- k-a ‘idea (dim)’ is used with the diminutive suffix -k, which indicates that the speaker views the idea negatively; compare this with the neutral sentence in (19b), where the expressive suffix is absent.  (19)  a.  mn’é    v         gólovu     pr’išlá    id’éj-k-a             me        prep    head        came      idea-EXPR-N.SG             ‘(An) idea (dim) came to my mind’               i. Descriptive:  ‘A small idea came to the speaker’s mind’              ii. Expressive:  ‘Speaker views the idea negatively’   20    b.  mn’é    v          gólovu    pr’išlá   id’éj-a             me        prep    head        came     idea-N.SG             ‘(An) idea came to my mind’               i. Descriptive:  ‘An idea came to the speaker’s mind’ (Shvedova et al. 1982:210)  Augmentative suffixes can also express the positive or negative attitude of the speaker (Derkach 2005; Nesset 2003; Schneider 2003). For example, in (20a), sobáč’-išč’-a ‘big/malevolent dog’ (translation from Derkach 2005:11) is used with the augmentative suffix -išč’, which indicates that the speaker views the dog negatively; compare this with the neutral (20b), where the expressive suffix is absent.  (20)  a.  Sobáč’-išč’-a        pr’išlá.              dog-EXPR-N.SG     came              ‘(A) dog (aug) came’               i. Descriptive:  ‘A big dog came’              ii. Expressive:  ‘Speaker views the dog negatively’       b.  Sobák-a       pr’išlá.             dog-N.SG     came             ‘(A) dog came’               i. Descriptive:  ‘A dog came’                                                       (Derkach 2005:11)  The data in (21a) illustrate that sobáč’-išč’-a cannot be modified by the adjective ‘small’; compare this with (21b), where it can be modified by ‘huge’. The data in (22a) show that sobáč’-išč’-a cannot be used in a context when the speaker denies his/her negative attitude toward the dog; compare this with (22b), where the speaker expresses a negative attitude.  (21)  a. # Mál’en’kaja   sobáč’-išč’-a        pr’i-š-l-á               small              dog-EXPR-N.SG    came               ‘(A) small dog (aug) came’          b.   Ogrómnaja  sobáč’-išč’-a         pr’i-š-l-á               huge             dog-EXPR-N.SG    came               ‘(A) huge dog (aug) came’  21 (22)  a. # Já    l’ubl’u    étu      sobáč’-išč’-u.               I      love        this     dog-EXPR-ACC.SG              ‘I love this dog (aug)’           b.   Já    n’enav’ížu    étu      sobáč’-išč’-u.               I      hate               this     dog-EXPR-ACC.SG              ‘I hate this dog (aug)’  The list of simplex size suffixes in Russian is given in Table 2.2.  Diminutive suffixes -k (allomorphs: -ok, -ek, -ik) -c (allomorphs: -ec, -ic) Augmentative suffix -išč’ Table 2.2: Size suffixes in Russian  To summarize, a difference between attitude and size suffixes in Russian is that the former have expressive content, while the latter have both expressive and descriptive content. Attitude suffixes can be removed without affecting the descriptive content of a phrase, which confirms Potts’ (2007) criterion of independence. In contrast, size suffixes have descriptive content of their own, and thus, by removing a size suffix, its descriptive content is also removed, which indicates a change in the descriptive content.  2.2.2. Nondisplaceability of Russian expressive suffixes Nondisplaceability is another characteristic that distinguishes expressive from descriptive content. Expressives predicate the speaker’s attitudes in a particular utterance situation that are valid only for the speaker at the time and place of the utterance. Consider for example the following German data from Potts (2007). In (23a), the descriptive content is Hermann believes that Hella’s dog is dead. In (23b), the expressive nominal Köter ‘damn.dog’ is used instead of the neutral Hund ‘dog’. As a result, (23b) conveys that the speaker holds Hella’s dog in low regard, in addition to the proposition Hermann believes that Hella’s dog is dead.  (23)  a.  Hermann   glaubt,      dass    Hellas     Hund    gestorben   ist.             Hermann    believes   that     Hella’s    dog       dead           is            ‘Hermann believes that Hella’s dog is dead’ 22               i. Descriptive:  ‘Hermann believes that Hella’s dog is dead’          b.  Hermann   glaubt,     dass    Hellas     Köter           gestorben   ist.             Hermann    believes   that    Hella’s   damn.dog     dead           is            ‘Hermann believes that Hella’s damn dog is dead’               i. Descriptive:  ‘Hermann believes that Hella’s dog is dead’              ii. Expressive:  ‘Speaker views Hella’s dog negatively’                     (Potts 2007:6)  This example demonstrates the characteristic of nondisplaceability because the expressive Köter ‘damn.dog’ relates directly to the utterance situation: it is the speaker who expresses a negative attitude toward the dog. This negative attitude cannot be attributed to the subject of the sentence Hermann (Potts 2007:6). It is worth pointing out that Köter ‘damn dog’ also has both expressive and descriptive content, similar to the Russian size suffixes discussed above.  Russian data with expressive suffixes also demonstrate the characteristic of nondisplaceability. This is shown in (24) for the vulgar suffix -an, which is used to refer to an old man (star’-ik-án ‘old man’). The sentence conveys a vulgar attitude toward an old man in addition to the proposition Mom thinks that Olga’s old man is a good person. The suffix -an indicates here that it is the speaker who expresses a vulgar attitude toward the old man, not the subject of the sentence ‘mom’, since mom says that the old man is a good person and thus has a positive attitude toward the old man.  (24) Predstáv’, máma   dúmajet,  čto  Ol’in     star’-ik-án                  xoróšij  č’elov’ék.        imagine    mom     thinks      that Olga’s   old-NOM-EXPR.N.SG  good      person       ‘Imagine, mom thinks that Olga’s old man (vulg) is a good person’  The same property also holds for size suffixes. For example, in (25), the diminutive suffix -k is added to the noun denoting ‘notebook’. As a result, the sentence conveys a positive attitude toward the notebook, in addition to the proposition Mom thinks that Olga’s notebook was found. The suffix -k indicates that it is the speaker who sees Olga’s notebook positively and says nothing about whether the subject ‘mom’ has a positive attitude toward the notebook.  23 (25)  Pr’edstáv’,       máma   dúmajet,  čto    Ol’ina     t’etrád-k-a                   našlá-s’.        imagine            mom     thinks      that  Olga’s     notebook-EXPR-N.SG   found-self        ‘Imagine, mom thinks that Olga’s notebook (dim) was found’  To summarize, both attitude and size suffixes relate directly to the utterance situation expressing the attitude of the speaker at the time and place of the utterance. This confirms Potts’ (2007) criterion of nondisplaceability. Note in passing that the criterion of nondisplaceability also pertain to the descriptive content of a size suffix. For example, in (26), it seems that the suffix -išč’ indicates that the speaker views the dog as being big, but does not indicate that the subject ‘mom’ also sees it as being big.  (26)   Pr’edstáv’,      máma   dúmajet,  čto   Ol’ina     sobáč’-išč’-a      mál’en’kaja.         imagine           mom     thinks      that Olga’s     dog-EXPR-N.SG  small          ‘Imagine, mom thinks that Olga’s dog (aug) is small’  2.2.3. Perspective dependence of Russian expressive suffixes Some expressives appear to contradict the criterion of nondisplaceability. In particular, Potts (2007) notes (after Kratzer 1999 and Schlenker 2003) that expressives may reflect the perspective of someone other than the speaker. This suggests that expressives do not necessarily express the attitude of the speaker but instead the attitude of the perspective bearer, which in most but not all cases is the speaker. For example, in (27), the expressive that bastard indicates the negative emotion of the speaker’s father, and not the speaker herself (Potts 2007:7).  (27)  My father screamed that he would never allow me to marry that bastard Webster.   i. Descriptive:  ‘Speaker’s father screamed that he would never allow the speaker to marry Webster’  ii. Expressive: ‘Speaker’s father views Webster negatively’ (Kratzer 1999, cited by Potts 2007:7)  The same phenomenon is found with Russian expressive suffixes; they too can express the attitude of the perspective bearer. For example, in (28), just like in the data above, the 24 expressive -an indicates the negative emotion of the speaker’s father, rather than the speaker herself.  (28)  Ot’éc    kr’ičál, čto          n’e   pozvól’it  mn’e žen’ít’s’a  na  étom         father  screamed  that (he)  not   let(fut)        me    marry      prep  this   star’-ik-án’-e.  old- NOM-EXPR-LOC.SG         ‘Father screamed that he won’t let me marry this old man (vulg)’   i. Descriptive: ‘Speaker’s father screamed that he would never allow the speaker to marry that old man’  ii. Expressive: ‘Speaker’s father views the old man negatively’  To deal with examples such as (27) and (28), Potts (2007) uses the notion of the “judge”. The judge is the person whose attitude is expressed by the expressive. Normally the judge is the speaker (in accordance with nondisplacebility), but sometimes it can be someone else. Potts doesn't discuss under which conditions the judge is not the speaker.  To summarize, Russian expressive suffixes, like English expressives such as bastard, indicate that sometimes the perspective bearer (or “judge”) is someone other than the speaker, which confirms Potts’ (2007) criterion of perspective dependence. 2.2.3.1. The speaker’s attitude With respect to both nondisplaceability and perspective dependence, the following question arises: what is the speaker’s (or judge’s) attitude directed at? In the literature, there is no uniform answer to this question. One definition of expressivity is that it is a direct expression of the emotive attitude of the speaker toward what he/she is speaking about (Bühler 1934; Jakobson 1960). Another definition is that emotional meaning reflects the personal feeling of the speaker, including his/her attitude toward the hearer or the content of speech (Leech 1974).  Here I show that, in the case of Russian expressive suffixes, the speaker’s attitude is directed at the referent of the noun that contains an expressive suffix and not at the hearer. For example, in (29a), the affectionate suffix -ul’ indicates the speaker’s affection toward 25 grandmother, but it does not indicate the speaker’s attitude toward the hearer. This is better illustrated in (29b), where the hearer is addressed in a vulgar manner using the vulgar suffix -an. Although the listener is addressed by the speaker with a negative attitude, the affectionate suffix -ul’ still indicates the speaker’s positive attitude toward grandmother.  (29)  a.  Bab-úl’-a                            pr’išlá.             grandmother-EXPR-N.SG     came               ‘Grandmother (affect) came’               i. Descriptive:  ‘Grandmother came’              ii. Expressive:  ‘Speaker views grandmother positively’          b.  Ej,        star’-ik-án,                  id’í    s’udá,       bab-úl’-a                           pr’išlá.             Hey,     old-NOM-EXPR.N.SG , come here,          grandmother-EXPR-N.SG    came             ‘Hey, old man (vulg), come here, grandmother (affect) came’  Unlike affectionate suffixes that indicate the speaker’s attitude toward the referent, not the hearer, vulgar suffixes can target the hearer indirectly. Vulgar suffixes can only be used in informal settings and thus, they belong to an informal register. Such suffixes indicate that the speaker treats a particular discourse as informal; thus, if used in the wrong (e.g., formal) register, they might affect the hearer indirectly and even provoke a negative reaction. Consequently, vulgar suffixes can indirectly target the hearer while directly expressing an attitude toward the referent. For example, the sentence in (30), used with the vulgar suffix - ox, can be uttered among close friends, but in a formal situation, it is considered rude and inappropriate.  (30)  Bab’-óx-a                          pr’išlá.          grandmother-EXPR-N.SG    came            ‘Grandmother (vulg) came’         i. Descriptive:  ‘Grandmother came’        ii. Expressive:  ‘Speaker views the grandmother negatively’  Size suffixes belong to both formal and informal registers (in this respect, they are similar to affectionate suffixes and different from vulgar suffixes). Size suffixes express the speaker’s attitude toward the referent, not the hearer. For example, in (31a), the diminutive suffix -k 26 indicates that the speaker views a dog with a positive attitude (small and nice dog). However, it does not express the speaker’s attitude toward the hearer. This is also shown in (31b), where the hearer is addressed in an irreverent manner with the vulgar suffix -an. Although the hearer is addressed with a negative attitude, the diminutive suffix -k still expresses the speaker’s positive attitude toward the dog.  (31)  a.  Sobáč’-k-a          pr’išlá.             dog-EXPR-N.SG    came               ‘(The) dog (dim) came’               i. Descriptive:  ‘The small dog came’              ii. Expressive:  ‘Speaker views the dog positively’          b.  Ej,        star’-ik-án,                  id’í      s’udá,      sobáč’-k-a           pr’išlá.             Hey,     old-NOM-EXPR.N.SG , come   here,        dog-EXPR-N.SG    came             ‘Hey, old man (vulg), come here, (the) dog (dim) came’  2.2.4. Descriptive ineffability of Russian expressive suffixes Another property of expressives is their lack of propositional content, that is, their descriptive ineffability. As such, expressives are similar to interjectives like ‘ouch’, which cannot be paraphrased with a description such as ‘I feel pain’. Their content is not a proposition, but the expression of an attitude/emotional state. According to Potts (2007), speakers are never fully satisfied when they paraphrase expressive content using descriptive terms, because paraphrases miss a wide range of emotive uses.  To investigate whether Russian expressive suffixes satisfy this criterion, I have interviewed three native speakers of Russian (non-linguists, all with higher education). I asked each of them to compare three sets of sentences containing expressive suffixes and paraphrases illustrated in (32)-(34) below. The speakers were asked: ‘What is the difference between the sentences in (a) and (b)?’ All sentences in (a) contained expressives, while all sentences in (b) contained paraphrases (information unknown to the speakers). After reading the sentences, two speakers told me that the sentences in (a) express an emotion, while the sentences in (b) do not. One speaker told me that the sentences in (a) express a ‘stronger’ emotion, compared to the sentences in (b). I conclude from these native speakers’ intuitions 27 that the speakers view expressive sentences and their paraphrases differently, with expressive sentences being more ‘emotive’. Consider now the sentences reviewed by the speakers (the speakers were given only the Cyrillic version).  In (32a), the affectionate suffix -ul’ is used with the word ‘mother’, expressing the speaker’s affection toward mother. In (32b), the suffix -ul’ is removed and the paraphrase ‘Mother who I love’ is used. As a result, the sentence in (32a) seems more emotive to the speakers, compared to the sentence in (32b).  (32)  a.  Мамуля  пришла.             Mam-úl’-a                pr’išlá.             mother-EXPR-N.SG    came               ‘Mother (affect) came’               i. Descriptive:  ‘Mother came’              ii. Expressive:  ‘Speaker views his/her mother positively’           b.  Мама, которую я люблю, пришла.              mám-a,            kotóruju   ja   lubl’ú,   pr’išlá              mother-N.SG    who          I    love       came                ‘Mother, who I love, came’  In (33a), the vulgar suffix -ug is used with the word ‘thief’, expressing the speaker’s irreverent attitude toward the thief. In (33b), the suffix -ug is removed and the paraphrase ‘The thief who I despise’ is used. The sentence in (33a) seems more emotive to the speakers than the sentence in (33b).   (33)  a.  Ворюга  пришел.              Vor’-úg-a             pr’išól.               thief-EXPR-N.SG    came                 ‘(The) thief (vulg) came’               i. Descriptive:  ‘The thief came’              ii. Expressive:  ‘Speaker views the thief negatively’           b.  Вор, которого я презираю, пришел.              Vór,             kotórogo   já    pr’eziráju,     pr’išól              thief.N.SG    who           I     despise          came                 ‘(The) thief, who I despise, came’ 28  In (34a), the size suffix -ik is used with kl’úč’ ‘key’, indicating that the key is small and nice in the speaker’s opinion. In (34b), the suffix -ik is removed and the paraphrase ‘The small and nice key’ is used. The sentence in (34a) seems more emotive to the speakers, compared to the sentence in (34b), which supports the view that diminutive suffixes are not just descriptive, but also expressive.  (34)  a.  Ключик  нашелся.             Kl’úč’-ik            našól-s’a.             key-EXPR.N.SG    found-self                ‘(The) key (dim) was found’               i. Descriptive:  ‘The small key was found’              ii. Expressive:  ‘Speaker views the key positively’           b.  Маленький и славный ключ нашелся.              Mál’en’kij   i        slávnij   kl’úč’   našól-s’a.              small           and   nice        key       found-self              ‘(The) small and nice key was found’  To summarize, Russian data with expressive suffixes seem more emotive to the speakers than their descriptive paraphrases because their content is not a proposition, but the expression of an attitude/emotional state. Thus, they satisfy Potts’ (2007) criterion of descriptive ineffability.  2.2.5. Immediacy of Russian expressive suffixes One property of expressives makes them similar to performatives. That is, just by means of uttering the expressive linguistic object, the speaker is performing a social act (such as insulting somebody). So, the act of uttering an expressive is the emotive performance. Potts (2007) refers to this property as immediacy.  A connection between expressives and speech-acts has been noted in the literature. For example, according to Tsujimura (1978), expressions like commands, prohibitions, or wishes should be treated from the attitudinal point of view. 29  Potts uses the following example to illustrate immediacy in English (35).  (35)  a.  That bastard Kresge was late for work yesterday.                i. Descriptive:  ‘Kresge was late for work yesterday’               ii. Expressive:  ‘Speaker views Kresge negatively’           b.  # But he’s no bastard today, because today he was on time.   (Potts 2007:13)  Just by saying (35a) ‘That bastard Kresge…’, the speaker expresses hostility towards Kresge, which explains why (35b) ‘But he is no bastard today..’ is infelicitous. The speaker expresses a negative attitude towards Kresge and then denies it without indicating that he/she has changed his/her mind.  According to Potts (2007), performatives exhibit exactly the same characteristic. An example of a performative is illustrated in (36) (promise). The sentence in (36a) is a promise to wash the dishes later. In (36b), the speaker denies his/her promise to wash the dishes later, which makes the sentence infelicitous.  (36)  a.  I promise that I will wash the dishes later.          b.   # But I refuse to wash the dishes later.      (Potts 2007:14)  Russian expressive suffixes display this property of immediacy as well. In (37a), the use of the affectionate suffix -ul’ indicates the affection of the speaker towards his/her mother. In (37b), the speaker denies affection without indicating that he/she has changed his/her mind. As a result, the sentence is infelicitous. 30 (37)  a.  Mam-úl’-a                   pr’išlá.               mother-EXPR-N.SG        came                  ‘Mother (affect) came’               i. Descriptive:  ‘Mother came’              ii. Expressive:  ‘Speaker views his/her mother positively’           b. #No    já   ejé   n’e  lubl’ú.                but    I    her   not  love               ‘But I do not love her’  In (38a), the speaker uses the size suffix -k, which expresses a positive attitude toward the children. In (38b), the speaker denies the positive attitude toward the children, which makes the sentence infelicitous.  (38)  a.  Segódn’a  pr’ixod’íl’i  d’ét-k’-i.              today        came          child-EXPR-N.PL             ‘(The) children (dim) came today’               i. Descriptive:  ‘The small children came today’              ii. Expressive:  ‘Speaker views the children positively’           b. #No   já    íx        t’erp’ét’  n’e   mogú.                but   I     them   bear         not   can                ‘But I can not stand them’  To summarize, both attitude and size suffixes behave like performatives and achieve their intended act simply by being uttered, which confirms Potts’ (2007) criterion of immediacy.  2.2.6. Repeatability of Russian expressive suffixes Another property of expressive LOs which sets them apart from descriptive LOs has to do with the fact that the repetition of expressives leads to strengthening of emotion rather than redundancy. Potts refers to this property as repeatability. For example, in (39a), the expressive damn is used once in the sentence; in (39b), damn is used twice; and in (39c), it is used three times: as a result, the speaker seems more and more emotional as we move down the list (Potts 2007).  31 (39)  a.  Damn, I left my keys in the car.          b.  Damn, I left my damn keys in the car.          c.  Damn, I left my damn keys in the damn car.   (Potts 2007:15)  Descriptive content differs from (39), because repetition of descriptive content leads to redundancy instead of strengthening of emotion. This is illustrated in (40), which is infelicitous because of redundant repetition of the descriptive content.  (40)  # I’m angry! I forget my keys. I’m angry! They are in the car. I’m angry! (Potts 2007:15)  In Russian, expressive suffixes demonstrate the characteristic of repeatability. Multiple expressive suffixes lead to a strengthening of emotion instead of redundancy (Mandel’štam 1903; Popoff-Böcker 1973; Popov 1967; Stankiewicz 1954, 1968; Volek 1987). For example, in (41a), the proposition ‘Nina came to mother’ is descriptive with no expressive content. In (41b), the affectionate suffix -ul’ is used to express the speaker’s affection toward Nina. In (41c), the affectionate suffix -ul’ is used twice, indicating that the speaker feels affection toward both Nina and her mother. Of the three sentences in (41), the one in (41c) with multiple affectionate suffixes would seem the most “emotional” to speakers of Russian.  (41)  a.  N’ín-a               pr’išlá  k   mám’-e.             Nina-N.SG          came    to  mother-DAT.SG             ‘Nina came to mother’               i. Descriptive:  ‘Nina came to mother’          b.  N’in-úl’-a                  pr’išlá  k   mám’-e.             Nina-EXPR-N.SG        came    to  mother-DAT.SG             ‘Nina (affect) came to mother’               i. Descriptive:  ‘Nina came to mother’              ii. Expressive:  ‘Speaker views Nina positively’  32         c.  N’in-úl’-a                  pr’išlá  k   mam’-úl’-e.             Nina-EXPR-N.SG        came    to  mother-EXPR-DAT.SG             ‘Nina (affect) came to mother (affect)’               i. Descriptive:  ‘Nina came to mother’              ii. Expressive:  ‘Speaker views both Nina and her mother positively’  In (42a), the proposition ‘The dog hurt its leg’ is also descriptive without expressive content. In (42b), the diminutive size suffix -k is used making the proposition expressive and indicating the speaker’s positive emotion toward the dog. In (42c), the diminutive suffix -k is used twice, expressing the speaker’s positive emotions toward both the dog and its leg. As a result, there is heightening of the emotional state of the speaker.  (42)  a.  U       sobák’-i          zabol’éla  láp-a.             PREP   dog-GEN.SG    hurt          leg-N.SG            ‘(The) dog hurt (its) leg’               i. Descriptive:  ‘The dog hurt its leg’         b.  U       sobáč’-k’-i4              zabol’éla   lápa.            PREP   dog-EXPR-GEN.SG    hurt           leg-N.SG           ‘(The) dog (dim) hurt (its) leg’               i. Descriptive:  ‘The small dog hurt its leg’              ii. Expressive:  ‘Speaker views the dog positively’          c.  U       sobáč’-k’-i               zabol’éla   láp-k-a.            PREP   dog-EXPR-GEN.SG   hurt            leg-EXPR-N.SG           ‘(The) dog (dim) hurt (its) leg (dim)’               i. Descriptive:  ‘The small dog hurt its small leg’              ii. Expressive:  ‘Speaker views both the dog and its leg positively’  The diminutive size suffix -k (allomorphs: -ok, -ek, -ik) also allows for repeatability in the same word. For example, in (43), it is used twice in the same word (see §5.5 on a more detailed discussion of this kind of repeatability). In (43b), there is heightening of the emotional state of the speaker, compared to (43a).  4  There is a k~č’ alternation in this word typical for Russian, which can also be seen in the diminutive suffix -ek in (43b). Consonantal alternations will be discussed in Chapter 3. 33  (43)  a.  U       sobáč’-k’-i               zabol’éla  lápa.             PREP   dog-EXPR-GEN.SG    hurt          leg-N.SG            ‘(The) dog (dim) hurt (its) leg’               i. Descriptive:  ‘The small dog hurt its leg’              ii. Expressive:  ‘Speaker views the dog positively’          b.  U       sobáč’-eč’-k’-i                  zabol’éla   láp-a.             PREP   dog-EXPR-EXPR-GEN.SG   hurt           leg-N.SG            ‘(The) dog (dim-dim) hurt (its) leg’               i. Descriptive:  ‘The very small dog hurt its leg’              ii. Expressive:  ‘Speaker views the dog very positively’  To summarize, multiple attitude and size suffixes in Russian lead to strengthening of emotion rather than redundancy, which confirms Potts’ (2007) criterion of repeatability.  2.3. CONCLUSIONS To conclude, Russian attitude and size suffixes demonstrate the following characteristics of expressive content outlined in Potts (2007) (Table 2.3).   Attitude suffixes (affectionate, vulgar) Size suffixes (diminutive, augmentative) 1. Independence a * 2. Nondisplaceability a a 3. Perspective dependence a a 4. Descriptive ineffability a a 5. Immediacy a a 6. Repeatability a a Table 2.3: Characteristics of expressive suffixes in Russian  Attitude suffixes have expressive content, while size suffixes have both expressive and descriptive content (compare Table 2.4 and Table 2.5).  34 Attitude suffixes Expressive content Descriptive content Affectionate suffixes (-án’, -áš, -ón, -úl’, -ún’, -úr,  -ús’, -úš) a * Vulgar suffixes (-ág, -ák, -ál, -án, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx) a * Table 2.4: Attitude suffixes are expressive  Size suffixes Expressive content Descriptive content Diminutive suffixes -k (allomorphs: -ok, -ek, -ik) -c (allomorphs: -ec, -ic) a a Augmentative suffix -išč’ a a Table 2.5: Size suffixes are expressive and descriptive  Both attitude and size suffixes indicate the speaker’s attitude toward the referent, not the hearer. Vulgar suffixes belong to the informal register and can target the hearer indirectly if uttered in the wrong (formal) register. 35 Chapter 3: Background Information  This chapter presents background information on Russian grammatical gender, inflectional class, natural gender (or sex), and animacy. This background information is important to better understand the behaviour of Russian expressive suffixes. Grammatical gender and inflectional class are used as tests to determine whether a certain expressive suffix is a head or a modifier. Natural gender and animacy are also important to this discussion, because with respect to expressive suffixes, animate nouns that have natural gender behave differently from both animate nouns without natural gender, and inanimate nouns.  The structure of this chapter is as follows. In §3.1, I discuss grammatical and natural gender. In §3.2, I outline different approaches to Russian inflectional class. In §3.3., I discuss animacy. In §3.4, I describe phonological alternations relevant to expressive suffixes in Russian.  3.1. GENDER Grammatical gender serves as one of the diagnostic properties5 for the two syntactic types of expressive suffixes. The purpose of this section is to introduce the properties of grammatical gender (§3.1.1). Moreover, to fully understand grammatical gender it is important to discuss nouns with natural gender, since these behave systematically differently than nouns that do not have natural gender (§3.1.2).  3.1.1. Grammatical gender Russian has three grammatical genders: masculine, feminine, and neuter, as illustrated in Table 3.1. That grammatical gender is indeed a grammatical category is evident from the fact that it cannot be predicted by either semantic or phonological properties of the noun (Corbett 1991).  5  See Chapter 4 for a list of diagnostics. 36   Masculine Feminine Neuter č'áj        ‘tea’ vod-á      ‘water’ v'in-ó       ‘wine’ dóm      ‘house’ škól-a     ‘school’ zdán'ij-e   ‘building’ žurnál   ‘magazine’ gaz'ét-a   ‘newspaper’ p'is'm-ó   ‘letter’ Table 3.1: Russian grammatical genders (Corbett 1991:35)  With the exception of nouns that have natural gender, grammatical gender cannot be predicted from the semantics of a noun. For example, in Table 3.1, all words in the first row (čáj ‘tea’, vod-á ‘water’, and v’in-ó ‘wine’) denote drinking liquids; however, their grammatical gender differs. They can be associated with each of the three grammatical genders: masculine, feminine, and neuter, respectively. Similarly, the nouns in the second row (dóm ‘house’, škól-a ‘school’, and zdán'ij-e ‘building’) all denote buildings, but they also have different grammatical genders: masculine, feminine, and neuter, respectively. And finally, the nouns in the third row all denote readable things; and again they are all associated with different genders. This establishes that grammatical gender cannot be predicted from the meaning of a given noun.  Similarly, grammatical gender cannot be predicted from the phonological form of a stem, as illustrated in Table 3.2. For example, the stem portf’él’ ‘briefcase’, ends in a palatalized consonant [l’] and the word is masculine. The stem m’et’él’ ‘snowstorm’ also ends in [l’], but the word is feminine. Both stems d’én’ ‘day’ and t’én’ ‘shadow’ end in [n’], however, the words are masculine and feminine, respectively.  Masculine Feminine portf’él’   ‘briefcase’ m’et’él’ ‘snowstorm’ d’én’        ‘day’ t’én’      ‘shadow’ Table 3.2: Masculine and feminine nouns, by phonological form of stem  The only way to unambiguously determine the grammatical gender of a given word is on the basis of agreement (Doleschal & Schmid 2001:256). There are four agreement tests in Russian: (i) agreement with adjectives, (ii) agreement with verbs in the past tense, (iii) agreement with relative pronouns, and (iv) agreement with personal pronouns. I discuss each 37 of these tests in turn. The examples in (1) illustrate agreement with adjectives. In (1a), the masculine noun múž ‘husband’ triggers masculine agreement on the adjective xoróš-ij ‘good’. In (1b), the feminine noun žen-á ‘wife’ triggers feminine agreement on the adjective. And finally, in (1c), the neuter noun d’él-o ‘cause’ triggers neuter agreement on the adjective.  (1) Agreement with adjectives:   a.  xoróš-ij                  múž          good-MASC.N.SG    husband.N.SG (MASC)           ‘good husband’   b.  xoróš-aja              žen-á           good-FEM.N.SG     wife-N.SG (FEM)           ‘good wife’   c.  xoróš-eje               d’él-o          good-NEUT.N.SG     cause-N.SG (NEUT)          ‘good cause’  The examples in (2) illustrate agreement with verbs in the past tense. In (2a), the masculine noun dóm ‘house’ triggers masculine agreement with the past form of the verb stoj-ál ‘stood’. In (2b), the feminine noun škól-a ‘school’ triggers feminine agreement. And in (2c), the neuter noun zdán’ij-e ‘building’ triggers neuter agreement.  (2) Agreement with verbs in the past tense:   a.  dóm                         stoj-ál          house.N.SG (MASC)   stood-PAST.MASC.SG           ‘(The) house stood’   b.  škól-a                        stoj-ál-a           school-N.SG (FEM)   stood-PAST-FEM.SG           ‘(The) school stood’   c.  zdán’ij-e                 stoj-ál-o           building-N.SG (NEUT)  stood-PAST-NEUT.SG           ‘(The) building stood’  38 Another type of agreement test involves relative pronouns. Such pronouns introduce a relative clause and display agreement for gender with the head of the relative clause. For example, in (3a), the masculine noun dóm ‘house’ triggers masculine agreement with the relative pronoun kotór-ij. In (3b), the feminine noun škól-a ‘school’ triggers feminine agreement. And in (3c), the neuter noun zdán’ij-e ‘building’ triggers neuter agreement.  (3) Agreement with relative pronouns:   a.  dóm,                         kotór-ij               stoj-ál          house.N.SG (MASC)   that-MASC.N.SG   stood-PAST.MASC.SG          ‘(The) house that stood’   b.  škól-a,                     kotór-aja             stoj-ál-a          school-N.SG (FEM)  that-FEM.N.SG      stood-PAST-FEM.SG           ‘(The) school that stood’   c.  zdán’ij-e,                 kotór-oje           stoj-ál-o           building-N.SG (NEUT)  that-NEUT.N.SG  stood-PAST-NEUT.SG           ‘(The) building that stood’  Finally, personal pronouns necessarily match in gender with the noun they are coreferent with. For example, the masculine noun dóm ‘house’ requires the masculine personal pronoun ón (4a); the feminine noun škól-a ‘school’ requires a feminine personal pronoun on-á (4b); and the neuter noun zdán’ij-e ‘building’ requires a neuter personal pronoun on-ó (4c).  (4) Agreement with personal pronouns:   a.  Gd’é    dóm?                            Vót     ón.          where   house.N.SG (MASC)      here    it.N.SG (MASC)          ‘Where is the house? Here it is.’   b.  Gd’é    škól-a?                         Vót     on-á.          where  school-N.SG (FEM)        here    it-N.SG (FEM)                ‘Where is the school? Here it is.’   c.  Gd’é    zdán’ij-e?                     Vót    on-ó.         where   building.N.SG (NEUT)   here  it-N.SG (NEUT)          ‘Where is the building? Here it is.’  39 We have now established that gender is a grammatical category in Russian. Note for completeness that plural forms are not distinguished in terms of gender (Doleschal & Schmid 200; Shvedova et al. 1982; Wade 2000, among others). As Table 3.3 illustrates, the same plural agreement is used for plural nouns of all grammatical genders in Russian.  Singular Plural xoróš-ij    ‘good-N.SG (MASC)’ xoróš-aja  ‘good-N.SG (FEM)’ xoróš-eje  ‘good-N.SG (NEUT)’ xoróš-ije  ‘good-N.PL’ Table 3.3: Singular vs. plural agreement with an adjective  3.1.2. Natural gender (sex) As mentioned above, grammatical gender generally cannot be predicted on the basis of the meaning of the noun. The only exceptions are nouns that have natural gender (also called sex-differentiable nouns: Corbett 1982, 1991). A number of nouns that denote humans and some domesticated animals are sex-differentiable, as they are associated with natural gender (sex) in addition to grammatical gender. Natural gender can be seen in kinship terms (Table 3.4) and personal names (Table 3.5). For example, the noun ot’éc ‘father’ can only denote a male individual, while mát’ ‘mother’ can only denote a female individual. The name Iván ‘Ivan’ can only denote a male person, while Ánn-a ‘Anna’ can only denote a female person.  Male Female ot’éc    ‘father’ mát’       ‘mother’ sín        ‘son’ dóč’       ‘daughter’ d’ád’-a ‘uncle’ t'ót’-a     ‘aunt’ Table 3.4: Kinship terms  Male Female Iván      ‘Ivan’ Ánn-a      ‘Anna’ Páv’el   ‘Pavel’ Mar’íj-a   ‘Marija’ Pótr       ‘Potr’ Ól’g-a      ‘Olga’ Table 3.5: Personal names  Most nouns that denote animals, birds, and insects refer to members of the species in general and have no natural gender (such nouns are called non-sex-differentiable: Corbett 1982, 40 1991). For example, the nouns k’ít ‘whale’ (5a) and míš ‘mouse’ (5b) denote both male and female members of the species. Although these nouns are non-sex-differentiable, they have grammatical gender, which is seen from their agreement with adjectives.  (5)  a.  bol’š-ój              k’ít           big-MASC.N.SG   whale.N.SG (MASC)’             ‘big whale (member of the species)’         b.  bol’š-ája            míš           big-FEM.N.SG     mouse.N.SG (FEM)’             ‘big mouse (member of the species)’  If it is necessary for the speaker to differentiate between male and female animals, the words sam’éc ‘male’ and sámk-a ‘female’ are used, followed by the species in Genitive case (6).  (6)  a.  sam’éc       k’it-á               male           whale-GEN.SG (MASC)                ‘male of (the) whale’    b.  sámk-a     k’it-á               female       whale-GEN.SG (MASC)              ‘female of (the) whale’  A few domesticated animals, however, make use of different nouns to denote the species in general, and male and female individuals (Table 3.6).  Species Male Female lóšad’    ‘horse’ žer’eb’éc  ‘stallion’ kobíl-a  ‘mare’ sobák-a  ‘dog’ kob’él’      ‘male dog’ súk-a     ‘bitch’ Table 3.6: Domesticated animals using three nouns for species, male, and female  The grammatical gender of sex-differentiable nouns is always determined by their natural gender (Corbett 1982, 1991; Corbett & Fraser 2000). Thus, nouns that denote males are always masculine and nouns that denote females are always feminine. For example, in (7), the noun ot’éc ‘father’ denotes a male individual and thus is masculine. In (8), the noun mát’ ‘mother’ denotes a female individual and thus is feminine.  41  (7) a.  xoróš-ij                 ot’éc                         b. *xoróš-aja           ot’éc          good-MASC.N.SG    father.N.SG (MASC)           good-FEM.N.SG   father.N.SG (MASC)                ‘good father’                                               ‘good father’  (8)   a.  xoróš-aja             mát’                          b. *xoróš-ij               mát’      good-FEM.N.SG     mother.N.SG (FEM)          good-MASC.N.SG    mother.N.SG (FEM)               ‘good mother’                                            ‘good mother’  3.1.3. Common gender nouns We have seen that Russian nouns can be specified for grammatical gender. A subset of these nouns, namely sex-differentiable nouns, also have natural gender. In this section, I discuss another type of noun which, I propose, is associated neither with grammatical gender nor with natural gender. Such nouns are traditionally called “common gender” nouns. These nouns denote individuals, such as  ob-žór-a ‘glutton’, s’irot-á ‘orphan’, sud’-j-á ‘judge’, taratór-a ‘chatterbox’, tranž’ír-a ‘spendthrift’, za-d’ír-a ‘bully’, xanž-á ‘hypocrite’, etc. Common gender nouns are not associated with natural gender because they can denote both male and female individuals. Similarly, common gender nouns are not specified for grammatical gender, because they can trigger either masculine or feminine agreement. When common gender nouns are used to refer to a male referent, they trigger masculine agreement, in accordance with the fact that natural gender always determines grammatical gender in Russian. Likewise, when they are used to refer to a female referent, they trigger feminine agreement. Moreover, common gender nouns never trigger neuter agreement, because individuals can be either masculine or feminine in Russian. For example, in (9a), the noun ob-žór-a ‘glutton’ triggers masculine agreement with the adjective bol’š-ój ‘big’, while in (9b), the same noun triggers feminine agreement.  (9) a.  bol’š-ój                  ob-žór-a   (referring to a male)           big-MASC.N.SG        VERB.PREF-glutton-N.SG (MASC)           ‘big glutton’   b.  bol’š-ája                ob-žór-a      (referring to a female)           big-FEM.N.SG          VERB.PREF-glutton-N.SG (FEM)          ‘big glutton’  42 Similarly, in (10a), the noun s’irot-á ‘orphan’ triggers masculine agreement with the adjective b’éd-n-ij ‘poor’ when referring to a male referent. In (10b), the same noun triggers feminine agreement when referring to a female referent. If the natural gender of the referent is unknown, either masculine or feminine agreement can be used (Doleschal & Schmid 2001). Thus, in (9) and (10), either example (a) or (b) can be used if the gender of the referent is unknown (see §4.4.1.2 for a more detailed discussion of common gender nouns).  (10) a.  b’éd-n-íj                      s’irot-á           (referring to a male)           poor-ADJ-MASC.N.SG    orphan-N.SG (MASC)’          ‘poor orphan’   b.  b’éd-n-aja                    s’irot-á                       (referring to a female)           poor-ADJ-FEM.N.SG    orphan-N.SG (FEM)’           ‘poor orphan’  With respect to common gender nouns, the following question arises: Is common gender a fourth gender (in addition to masculine, feminine, and neuter), or is it simply the absence of gender? In Chapter 4, I argue for the latter―common gender is not a separate gender, but is the absence of grammatical gender on animate nouns. 3.2. INFLECTIONAL CLASS Another diagnostic property used to distinguish between different syntactic types of expressive suffixes is the inflectional class (or declension) of a noun. The inflectional class determines the relevant inflectional paradigm of a noun. This paradigm in turn consists of a set of endings, each of which is associated with certain morphosyntactic properties. For example, škól-a ‘school’ is associated with the following inflectional paradigm, which indicates that it belongs to the inflectional Class II (11).  (11) Inflectional paradigm for the word škól-a ‘school’ (Class II)         Nominative     škól-a     ‘school-N.SG’         Accusative      škól-u    ‘school-ACC.SG’         Genitive          škól-i      ‘school-GEN.SG’         Dative             škól’-e    ‘school-DAT.SG’         Instrumental    škól-oj    ‘school-INST.SG’         Locative          škól’-e    ‘school-LOC.SG’  43 In the literature, there is no agreement about the number of inflectional classes in Russian. Some argue that there are two inflectional classes (Stankiewicz 1978; Shvedova 1970; Zalizniak 1967), some argue that there are three (Durnovo 1922; Durovich 1964; Isachenko 1962; Karcevskij 1948; Shvedova et al. 1982; Stankiewicz 1968; Timberlake 2004; Trager 1940; Unbegaun 1957; Vinogradov et al. 1952, among others), some argue that there are four (Corbett 1982, 1991; Corbett & Fraser 2000; Karcevskij 1932; Müller 2005; Nesset 1994), and Jakobson (1958) has even argued that there are five inflectional classes in Russian.6  It can be seen in the references above that even the same authors sometimes vary in their views on the number of inflectional classes. For example, Stankiewicz (1968) assumes a three-class approach, while Stankiewicz (1978) advocates a two-class approach. Shvedova (1970) describes two inflectional classes in Russian, while Shvedova et al. (1982) assume three inflectional classes.  Below I discuss the two most wide-spread approaches to Russian inflectional classes: the “traditional approach” and an approach proposed by Corbett (1982, 1991). 3.2.1. The traditional approach to Russian inflectional classes According to the traditional approach, there are three inflectional classes in Russian, as illustrated in Table 3.7. Class I nouns comprise masculine nouns with a -Ø ending in Nominative (Nom) singular (sg) and neuter nouns with -o/-e endings in Nom sg. Class II nouns comprise masculine, feminine, and common gender nouns that all have the -a ending in Nom sg. And finally, Class III nouns comprise feminine nouns that have a -Ø ending in Nom sg.       6  Jakobson (1958) mentions three main inflectional classes and two subclasses in Russian, so his proposal may be classified as belonging to the three-class approach, rather than a five-class approach. 44 Class I Class II Class III  masculine end in -Ø in Nom neuter end in -o/e in Nom masculine/ feminine/ common end in -a in Nom feminine end in -Ø in Nom Singular Nominative zakón ‘law’ v’in-ó ‘wine’ škól-a ‘school’ kóst’ ‘bone’ Accusative  zakón v’in-ó škól-u kóst’ Genitive  zakón-a v’in-á škól-i kóst’-i Dative zakón-u v’in-ú škól'-e kóst’-i Instrumental  zakón-om v’in-óm škól-oj kóst’-ju Locative zakón'-e v’in’-é škól'-e kóst’-i Plural Nominative zakón-i ‘laws’ v’ín-a ‘wines’ škól-i ‘schools’ kóst’-i ‘bones’ Accusative  zakón-i v’ín-a škól-i kóst’-i Genitive  zakón-ov v’ín škól kost’-éj Dative zakón-am v’ín-am škól-am kost’-ám Instrumental  zakón-ami v’ín-ami škól-ami kost’-ámi Locative zakón-ax v’ín-ax škól-ax kost’-áx Table 3.7: Inflectional classes in Russian (“traditional approach”)  Below I discuss each of these inflectional classes in turn.  3.2.1.1. Class I Class I contains masculine and neuter nouns. In Nom sg., masculine nouns end in -Ø, as illustrated in (12)–(13).  (12)  a.  xoróš-ij                     zakón              good-MASC.N.SG     law.N.SG (MASC; CLASS I)              ‘good law’  (13)  a.  xoróš-ij                   ot’éc              good-MASC.N.SG     father.N.SG (MASC; CLASS I)             ‘good father’  In contrast, neuter nouns end in -o/-e, as illustrated in (14)–(15).  45 (14)  a.  xoróš-eje                  v’in-ó              good-NEUT.N.SG      wine-N.SG (NEUT; CLASS I)             ‘good wine’  (15)  a.  xoróš-eje                zdán’ij-e               good-NEUT.N.SG      building-N.SG (NEUT; CLASS I)              ‘good building’  According to the traditional approach, masculine and neuter nouns belong to the same inflectional class because they are associated with almost the same inflectional paradigm in the singular. They only differ in Nom and Acc cases, where masculine nouns end in -Ø and neuter nouns end in -o/-e. In the plural, however, they differ in three cases: Nom, Acc, and Gen (Corbett 1982). 3.2.1.2. Class II Class II nouns are characterized by the fact that they all end in -a in Nom sg., independent of the grammatical gender of the noun (masculine, feminine, or common). This is illustrated in (16)–(18).  (16) Feminine noun   xoróš-aja             škól-a               good-FEM.N.SG       school-N.SG (FEM; CLASS II)              ‘good school’  (17) Masculine noun    xoróš-ij                    d’ád’-a            good-MASC.N.SG        uncle-N.SG (MASC; CLASS II)           ‘good uncle’  (18) Common gender noun   a.  xoróš-ij                     sud’-j-á              good-MASC.N.SG      judge-SUFF-N.SG (MASC; CLASS II)              ‘good judge’   b.  xoróš-aja               sud’-j-á              good-FEM.N.SG        judge-SUFF-N.SG (FEM; CLASS II)              ‘good judge’  46 Note in passing that masculine nouns in Russian belong to either Class I or Class II. Feminine nouns belong to either Class II or Class III, which will be discussed later. In contrast, all common gender nouns belong to Class II (none of the other inflectional classes accommodate common gender nouns). The inflectional paradigms for Class II feminine, masculine, and common gender nouns are given in Table 3.8.   Feminine end in -a in Nom Masculine end in -a in Nom Common end in -a in Nom Singular Nominative škól-a ‘school’ d’ád’-a ‘uncle’ sud’-j-á ‘judge’ Accusative  škól-u d’ád’-u sud’-j-ú Genitive  škól-i d’ád’-i sud’-j-í Dative škól'-e d’ád’-e sud’-j-é Instrumental  škól-oj d’ád’-ej sud’-j-ój Locative škól'-e d’ád’-e sud’-j-é Plural Nominative škól-i ‘schools’ d’ád’-i ‘uncles’ súd’-j-i ‘judges’ Accusative  škól-i d’ád’ súd’-ej Genitive  škól d’ád’ súd’-ej Dative škól-am d’ád’-am súd’-j-am Instrumental  škól-ami d’ád’-ami súd’-j-ami Locative škól-ax d’ád’-ax súd’-j-ax Table 3.8: Inflectional Class II (feminine, masculine, and common gender nouns)  By far the largest set of nouns in Class II is associated with feminine gender. These feminine nouns denote either animate beings (e.g., t’ót’-a ‘aunt’) or inanimate things (e.g., kn’íg-a ‘book’). Masculine and common gender nouns comprise a much smaller group and they only denote animate beings. The only difference between masculine and common gender nouns is that Class II masculine nouns have the natural gender “male” as part of their meaning (e.g., d’ád’-a ‘uncle’, muž-č’-ín-a ‘man/male’, jún-oš-a ‘young man/male’). Common gender nouns have neither natural gender nor grammatical gender, thus they can trigger either masculine or feminine grammatical agreement.  47 Note in passing that historically all Class II nouns comprised a uniform group in terms of grammatical gender: only feminine nouns belonged to this class and consequently, all Class II nouns triggered feminine agreement, even nouns that denoted male individuals. According to Gorshkova & Xaburgaev (1981:136–137), in Russian historical texts of the 14th-15th centuries, as well as in some contemporary Northern dialects of Russian,7 all Class II nouns are still feminine (19).  (19) Contemporary Northern dialects  bol’š-a              muž-ič’-in-a         big-FEM.N.SG    man-NOM-NOM-N.SG (FEM; CLASS II)        ‘big man/male’ (Gorshkova & Xaburgaev 1981:136; dialects of Zaon’éžje)  In Contemporary Standard Russian, the equivalent of (19) is ungrammatical. The noun muž- č’-ín-a ‘man’ denotes a male individual, and since natural gender determines grammatical gender in Contemporary Standard Russian, this word can only trigger masculine agreement, as illustrated in (20).  (20)  Contemporary Standard Russian  a.  bol’š-ój               muž-č’-ín-a                big-MASC.N.SG    man-NOM-NOM-N.SG (MASC; CLASS II)              ‘big man/male’         b.* bol’š-ája            muž-č’-ín-a               big-FEM.N.SG     man-NOM-NOM-N.SG (FEM; CLASS II)               ‘big man/male’  This suggests that, at least historically, the set of nouns belonging to Class II were restricted to feminine nouns (therefore, they all end in -a), which means that the notion of “common gender” is a historical innovation. The fact that in Contemporary Standard Russian, Class II contains nouns of all three genders is probably due to another historical phenomenon when natural gender took precedence over grammatical gender in Russian.   7  Gorshkova & Xaburgaev (1981) unfortunately do not list the dialects, but only mention Zaon’éžje. 48 3.2.1.3. Class III Class III contains feminine nouns that have a -Ø ending in Nom sg.. What is special about this class is that all Class III stems end either in a soft (palatalized) consonant, or the consonants [š] or [ž], which were historically palatalized but lost their palatalization in Contemporary Russian. For example, in (21) and (22), the stems bróv’ ‘eyebrow’ and nóč’ ‘night’ end in the palatalized consonants [v’] and [č’].  (21)  kras’-ív-aja                        bróv’              beautiful-SUFF-FEM.N.SG      eyebrow.N.SG (FEM; CLASS III)              ‘beautiful eyebrow’  (22)  t’íx-aja                              nóč’              quiet-FEM.N.SG                       night.N.SG (FEM; CLASS III)              ‘quiet night’  In (23) and (24), the stems míš ‘mouse’ and róž ‘rye’ end in the consonants [š] and [ž].  (23)   bol’š-ája          míš              big-FEM.N.SG   mouse.N.SG (FEM; CLASS III)              ‘big mouse’  (24)   visók-aja       róž              tall-FEM.N.SG     rye.N.SG (FEM; CLASS III)                ‘tall rye’   Class I and Class II stems do not have such a restriction and can end in soft or hard consonants (see §4.4.2 for a more detailed discussion of this phenomenon). Class III has significantly fewer members than Class I and Class II. According to Corbett & Fraser (2000:67–68), Class I has 20,850 members, Class II has 16,050 members, and Class III only has 5,150 members, 4,300 of which are formed by the derivational suffix -ost’. The suffix -ost’ forms abstract nouns from adjectives, as illustrated in (25).  (25)  a.  stár-ij                                    b.  stár-ost’             old-MASC.N.SG                                old-NOM.N.SG (FEM; CLASS III)             ‘old’                                        ‘old age’ (Corbett & Fraser 2000:67)  49 The inflectional paradigms for Class III nouns bróv’ ‘eyebrow’, míš ‘mouse’, and stár-ost’ ‘old age’ are given in Table 3.9.    Feminine stem ends in a palatalized consonant Feminine stem ends in [š] Feminine end in suffix -ost’ Singular Nominative bróv’ ‘eyebrow’ míš ‘mouse’ stár-ost’ ‘old age’ Accusative  bróv’ míš stár-ost’ Genitive  bróv’-i míš-i stár-ost’-i Dative bróv’-i míš-i stár-ost’-i Instrumental  bróv’-ju míš-ju stár-ost’-ju Locative bróv’ míš stár-ost’ Plural Nominative bróv’-i ‘eyebrows’ míš-i ‘mice’ stár-ost’-i ‘old ages’ Accusative  bróv’-i miš-éj stár-ost’-i Genitive  brov’-éj miš-éj stár-ost’-ej Dative brov’-ám miš-ám stár-ost’-am Instrumental  brov’-ámi miš-ámi stár-ost’-ami Locative brov’-áx miš-áx stár-ost’-ax Table 3.9: Inflectional Class III  3.2.2. Corbett’s (1982, 1991) approach to Russian inflectional classes In contrast to the traditional approach to Russian inflectional classes, Corbett (1982, 1991) proposes to separate the “traditional” Class I into two different classes according to gender: Class I contains only masculine nouns, while Class IV contain only neuter nouns. This is illustrated in Table 3.10.  50  Class I Class II Class III Class IV  masculine end in -Ø in Nom feminine/ masculine/ common end in -a in Nom feminine end in -Ø in Nom neuter end in -o/e in Nom Singular Nominative zakón ‘law’ škól-a ‘school’ kóst’ ‘bone’ v'in-ó ‘wine’ Accusative  zakón škól-u kóst’ v'in-ó Genitive  zakón-a škól-i kóst’-i v'in-á Dative zakón-u škól'-e kóst’-i v'in-ú Instrumental  zakón-om škól-oj kóst’-ju v'in-óm Locative zakón'-e škól'-e kóst’-i v'in’-é Plural Nominative zakón-i ‘laws’ škól-i ‘schools’ kóst’-i ‘bones’ v'ín-a ‘wines’ Accusative  zakón-i škól-i kóst’-i v'ín-a Genitive  zakón-ov škól kost’-éj v'ín Dative zakón-am škól-am kost’-ám v'ín-am Instrumental  zakón-ami škól-ami kost’-ámi v'ín-ami Locative zakón-ax škól-ax kost’-áx v'ín-ax Table 3.10: Inflectional classes in Russian, according to Corbett (1982, 1991)  As a consequence of this analysis, the grammatical gender of a noun is predictable from its inflectional class. Thus, according to Corbett (1982, 1991) and Corbett & Fraser (2000), all Class I nouns are masculine; Class II and Class III nouns are feminine, and Class IV nouns are neuter (26).  (26) Grammatical gender is predicted from inflectional class   I class   ─> masculine  II class  ─> feminine  III class ─> feminine  IV class ─> neuter  An apparent problem for this approach is the fact that Class II nouns contain not only feminine but also masculine nouns (see Table 3.8). We can account for this behaviour if we assume natural gender always takes precedence over grammatical gender (Corbett 1982, 1991, Corbett & Fraser 2000). Consequently, any noun that denotes a male individual will be associated with masculine gender. Thus, masculine nouns with natural gender do not pose a problem for Corbett’s analysis because all of these nouns denote male individuals. However, 51 Class II common gender nouns are problematic for this approach8. They do not have natural gender as part of their semantics and in cases when the referent is unknown, their grammatical gender cannot be determined by natural gender. Corbett’s approach predicts that all Class II common gender nouns should be feminine; however, in cases when the referent is unknown, they can still trigger either feminine or masculine agreement (Doleschal & Schmid 2001) (see §4.4.1.2 for a more detailed discussion and examples). 3.2.2.1. Why the properties of -išč’ favour the traditional approach The traditional approach allows for a more straightforward account of the augmentative suffix -išč’, which I discuss in this subsection. I start with a discussion of the historical development of its formal properties. Until the 18th century, the augmentative suffix -išč’ changed the grammatical gender of a noun. It turned nouns of all three grammatical genders into neuter nouns, including those specified for natural gender “male” or “female” (Shanskaja 1961:16). For example, in (27), the noun báb-a ‘woman (FEM)’ changes its grammatical gender from feminine to neuter when -išč’ is suffixed. The evidence for the neuter gender comes from the neuter agreement with the adjective dur-n-óje ‘foolish’, and the neuter ending -e in Nom sg. In (28), the noun dóm ‘house (MASC)’ changes its gender from masculine to neuter when -išč’ is added.  (27)  Russian : 18th century  a.  dur-n-ája                    báb-a              foolish-ADJ-FEM.N.SG    woman -N.SG (FEM)               ‘foolish woman’           b.  dur-n-óje                         bab’-íšč’-e                foolish-ADJ-NEUT.N.SG   woman-EXPR-N.SG (NEUT)              ‘foolish woman (aug)’               (Shanskaja 1961:16)  (28)   a.  v’el’ík-ij                       dóm               big-ADJ.MASC.N.SG         house-N.SG (MASC)              ‘big house’        b.  v’el’ík-oje                  dom’-íšč’-e              big-ADJ.NEUT.N.SG         house-EXPR-N.SG (NEUT)             ‘big house (aug)’                                                         (Shanskaja 1961:16)   8  Corbett (1982, 1991) and Corbett & Fraser (2000) do not discuss common gender nouns. 52 In Contemporary Standard Russian, the augmentative suffix -išč’ does not change the grammatical gender of a noun. For example, in (29), the feminine noun báb-a9 ‘woman (FEM)’ remains feminine when -išč’ is suffixed, while in (30) the masculine noun dóm ‘house (MASC)’ remains masculine when -išč’ is suffixed.  (29) Contemporary Standard Russian  a.  dur-n-ája                     báb-a             foolish-ADJ-FEM.N.SG   woman-N.SG (FEM)              ‘foolish woman’           b.  dur-n-ája                       bab’-íšč’-a              foolish-ADJ-FEM.N.SG    woman-EXPR-N.SG (FEM)             ‘foolish woman (aug)’  (30)  a. bol’š-ój10             dóm               big-MASC.N.SG    house-N.SG (MASC)              ‘big house’           b.  bol’š-ój               dom’-íšč’-e               big-MASC.N.SG   house-EXPR-N.SG (MASC)               ‘big house (aug)’  In (30b), the masculine noun dom’-íšč’-e ‘house’ acquires the suffix -e in Nom sg. when -išč’ is added. This ending is, however, unexpected, as normally masculine Class I nouns have -Ø ending in Nom sg. The ending -e corresponds to the neuter ending -e of Class I nouns. Crucially, nouns suffixed with -išč’ always requires final -e in the Nom sg., independent of whether they are masculine (31) or neuter (32).  (31)      bol’š-ój             dom’-íšč’-e              big-MASC.N.SG   house-EXPR-N.SG (MASC)               ‘big house (aug)’  (32)      bol’š-óje            okn’-íšč’-e              big-NEUT.N.SG     window-EXPR-N.SG (NEUT)               ‘big window (aug)’   9  In Contemporary Standard Russian, báb-a ‘woman’ has a downgrading meaning when referring to a woman. For a neutral meaning, zénšč’-in-a ‘woman’ is used. 10  The meaning of the word v’el’ík-ij ‘big’ has changed in Contemporary Russian from ‘big/large’ to ‘great’ (referring to events, wars, etc.). For this reason, I use the contemporary word bol’š-ój ‘big’ in this example. 53 What does this pattern tell us about the classification of inflectional paradigms? According to the traditional approach, both nouns belong to the same inflectional Class I, thus it is not too surprising that we find the same final suffix. However, why the augmentative masculine noun ends in -e is unknown (one way to look at it is that it is some kind of historical remnant related to the history of -išč’).  According to Corbett, the neuter ending -e means that -išč’ is declined according to the fourth declension (Corbett 1982:222), which I call inflectional Class IV. In the case of dom’-íšč’-e ‘house (MASC)’, -išč’ changes the inflectional class of the noun dóm ‘house (MASC)’ from Class I to Class IV (hence, the neuter ending -e). Since according to this approach, Class IV comprise neuter nouns, we would expect dom’-íšč’-e ‘house (MASC)’ to be neuter. This is, however, not the case, as dom’-íšč’-e ‘house (MASC)’ is masculine and thus there is no change in gender of the base: dóm ‘house (MASC)’. Corbett (1982:222) suggests that there is no change in grammatical gender of this word, because gender is transferred from the base of the noun. As a result, dom’-íšč’-e ‘house (MASC)’ is in Class IV, but remains masculine (see Corbett 1982:222 for a discussion of a similar example, topor-íšč’-e ‘big axe’). To the best of my knowledge, -išč’ is a unique suffix in this respect.  The two approaches to -išč’ are summarized in (33) and (34).  (33)  The traditional approach (no change in inflectional class when -išč’ is used)          a.  bol’š-ój              dom                big-MASC.N.SG   house.N.SG (MASC; CLASS I)               ‘big house’           b.  bol’š-ój              dom’-íšč’-e               big-MASC.N.SG    house-EXPR-N.SG (MASC; CLASS I)               ‘big house (aug)’  (34)  Corbett’s (1982, 1991) approach (change in inflectional class when -išč’ is used)          a.  bol’š-ój              dom              big-MASC.N.SG   house.N.SG (MASC; CLASS I)               ‘big house’  54          b.  bol’š-ój              dom’-íšč’-e               big-MASC.N.SG    house-EXPR-N.SG (MASC; CLASS IV)               ‘big house (aug)’  These different approaches produce different results for this study. If the suffix -išč’ does not change inflectional class (according to the traditional approach), it behaves like a syntactic modifier. However, if -išč’ does change inflectional class (according to Corbett’s approach), it behaves like a syntactic head. These approaches are summarized in (35).  (35)  Approaches to the Russian inflectional suffix -išč’  a. Traditional approach   i. The suffix -išč’ does not change grammatical gender or class.   ii. The nature of -e is unknown b. Corbett’s (1982, 1991) approach i. The suffix -išč’ changes inflectional class to Class IV, but it does not change grammatical gender  Here I show that the approach by Corbett (1982, 1991) has two significant problems with respect to the augmentative suffix -išč’. First, I show that -išč’ does not change inflectional class when it attaches to feminine bases. And second I show that the neuter ending -e cannot indicate a change in class, because only part of the inflectional paradigm changes, and not the whole paradigm.  If -išč’ were a Class IV suffix, it would turn nouns of all grammatical genders into Class IV nouns. However, when used with feminine bases, -išč’ does not produce a change in inflectional class. For example, in (36a), the feminine noun ruk-á ‘hand’ is in Class II (-a ending in Nom sg.). In (36b), the augmentative -išč’ is added and the resulting noun ruč’- íšč’-a ‘hand (aug)’ is still in Class II.  (36)  a.  bol’š-ája             ruk-á                big-FEM.N.SG       hand-N.SG (FEM; CLASS II)               ‘big hand’  55          b.  bol’š-ája             ruč’-íšč’-a               big-FEM.N.SG      hand-EXPR-N.SG (FEM; CLASS II)               ‘big hand (aug)’  The data in (36) show that -išč’ cannot belong to inflectional Class IV, contrary to what Corbett (1982:222) suggests. Corbett makes his observation about -išč’ based on data with masculine bases, but he does not discuss data with feminine bases.  Also, if -išč’ were a Class IV suffix, all nouns formed with this suffix would inflect according to the neuter paradigm of Class IV. However, in the examples in (37)–(38), we observe that neuter and masculine nouns have different inflectional endings in Nom case of the plural form. In (37), the plural noun dom’-íšč’-i ‘houses (aug)’ ends in -i. In (38), the plural noun okn’-íšč’-a ‘windows (aug)’ ends in -a.  (37)   a.  bol’š-íje          dom’-íšč’-i               big-PL.N.SG       house-EXPR-N.PL (MASC)                ‘big houses (aug)’   (38)  b.  bol’š-íje            okn’-íšč’-a                big-PL.N.SG         window-EXPR-N.PL (NEUT)               ‘big windows (aug)’  The plural endings -i (37) and -a (38) correspond to the plural endings of masculine and neuter nouns, respectively, as illustrated in (39)–(40). This difference is unexpected under Corbett’s approach, because this approach predicts a uniform paradigm in both singular and plural forms.  (39) Masculine nouns in the plural    a.  továrišč’              friend.N.SG (MASC)                ‘friend’        b.  továrišč’-i               friend-N.PL (MASC)           ‘friends’   56 (40) Neuter nouns in the plural    a.  č’udóv’išč’-e               monster-N.SG (NEUT)               ‘monster’           b.  č’udóv’išč’-a               monster-N.PL (NEUT)               ‘monster’  The data above show that the augmentative -išč’ is not a Class IV suffix that changes the inflectional class of a noun. In the case of feminine nouns, it produces no change in inflectional class. And in the case of masculine nouns, what looks like a change because of the neuter ending -e in the singular, does not indicate a change in the whole paradigm.  Based on these facts, I adopt the traditional approach that claims that both masculine and neuter nouns belong to Class I.  3.3. ANIMACY  The background on animacy and its relation to grammatical gender is relevant for this work because animate and inanimate nouns of different grammatical gender behave differently with respect to expressive suffixes in Russian. Masculine and feminine gender assignment is independent of animacy in Russian. That is, the animacy of a noun does not function as a predictor for grammatical gender: animate nouns can be either masculine or feminine (41) and so can inanimate nouns (42).  (41) Animate masculine and feminine nouns   a.  bol’š-ój                zv’ér’               big-MASC.N.SG    animal.N.SG (MASC)          ‘big animal (animate)’   b.  bol’š-ája             sobák-a           big-FEM.N.SG    dog-N.SG (FEM)               ‘big dog (animate)’  57 (42) Inanimate masculine and feminine nouns        a.  xoróš-ij                  zakón               good-MASC.N.SG    law.N.SG (MASC)               ‘good law (inanimate)’        b.  xoróš-aja                škól-a              good-FEM.N.SG      school-N.SG (FEM)             ‘good school (inanimate)’  Unlike masculine and feminine nouns, neuter nouns denote mostly inanimate things, such as v’in-ó ‘wine’ (43a) and zdán’ij-e ‘building’ (43b). There are three exceptions of neuter nouns that denote animate beings—č’udóv’išč’e ‘monster’, živótnoje ‘animal’, and nas’ekómoje ‘insect’—which are morphologically adjectives. These exceptions are not relevant to the current work. (43) Neuter inanimate nouns   a.  xoróš-eje                v’in-ó          good-NEUT.N.SG    wine-N.SG (NEUT)           ‘good wine (inanimate)’          b.  xoróš-eje                zdán’ij-e               good-NEUT.N.SG    building-N.SG (NEUT)              ‘good building (inanimate)’  Not all animate nouns are sex-differentiable in Russian. For example, the nouns zv’ér’ ‘animal’ (41a) and sobák-a ‘dog’ (41b) are animate because they denote living beings, but they are non-sex-differentiable, as there is nothing in their semantics that would indicate natural gender. Thus, when these nouns do denote an individual of a particular natural gender, the gender of the referent does not override the grammatical gender of the noun. For example, when sobák-a ‘dog’ refers to a male dog, it still triggers feminine grammatical agreement. In other words, Russian animate nouns can be either sex-differentiable or non- sex-differentiable, while inanimate nouns are only non-sex-differentiable (44).  (44)                             Animate                                                               Inanimate                                  3                                                             g                    Sex-differentiable      Non-sex-differentiable                    Non-sex-differentiable Examples: kob’él’ ‘male dog’    zv’ér’ ‘animal’                              p’is’m-ó ‘letter’                  súk-a   ‘bitch’           sobák-a ‘dog’                               zdán’ij-e ‘building’ 58  Animate and inanimate nouns are distinguished not only semantically, but also grammatically in Russian (Corbett 1980). The grammatical distinction is expressed in the inflectional paradigm of the masculine Class I nouns in the singular, and all nouns in the plural. In the singular, masculine animate Class I nouns have the same grammatical endings in Accusative and Genitive cases (ACC=GEN syncretism). In contrast, masculine inanimate Class I nouns have the same endings in Accusative and Nominative cases (ACC=NOM syncretism), as illustrated in Table 3.11. Exactly the same syncretism patterns hold for all nouns in the plural.    Class I Masculine animate Class I Masculine inanimate Singular Nominative zv’ér’ ‘animal’  zakón ‘law’ Accusative  zv’ér’-a (ACC=GEN) zakón (ACC=NOM) Genitive  zv’ér’-a zakón-a Dative zv’ér’-u zakón-u Instrumental  zv’ér’-em zakón-om Locative zv’ér'-e zakón'-e Table 3.11: Masculine animate and inanimate nouns (Class I)  To summarize, there are animate and inanimate nouns in Russian. Animate nouns are of two types: (i) nouns that have natural gender (sex-differentiable), and (ii) nouns that do not have natural gender (non-sex-differentiable). Common gender nouns are of type (ii) as they do not have natural gender. All inanimate nouns are non-sex-differentiable.  3.4. PHONOLOGICAL ALTERNATIONS Here I discuss phonological alternations that are seen in Russian data with expressive suffixes. In §3.4.1, I discuss vowel alternations; and in §3.4.2, I discuss consonantal alternations.  59 3.4.1. Vowel alternations All Slavic languages have one or two vowels that alternate with zero. Such vowels are called jer (or yer) vowels. Russian has two jer vowels: [ɛ] and [o], as illustrated in (45). Jer vowels are underlined. For comparison, (46) shows regular vowels [ɛ] and [o] that do not alternate with zero. (45) Jer vowels [ɛ] and [o]        (alternation with zero)   a. kus-ók    [kusók]   ‘piece . N.SG’       kus-k-á     [kuská]       ‘piece.GEN.SG’  b. d’én’      [d’ɛn’]   ‘day.N.SG’         dn’-á         [dn’á]         ‘day.GEN.SG’  (46) Regular vowels [ɛ] and [o] (no alternation with zero)   a. koról’      [kʌról’]   ‘king.N.SG’             korol’-á      [kərʌl’á]   ‘king.GEN.SG’  b. mudr’-éc11  [mudr’ɛc]  ‘wiseman.N.SG’      mudr’-ec-á [mudr’icá] ‘wiseman.GEN.SG’  Slavic jer vowels have received a lot of attention in the literature (Hansson 1993; Hermans 2002; Lightner 1972; Matushansky 2002; Rowicka 1999; Rubach 1986; Spencer 1985; Steriopolo 2007; Szpyra 1992; Timberlake 2004; Yearley 1995; Zoll 1998, etc.). The traditional view is that a jer vowel is realized when the next vowel is also a jer; otherwise, it is deleted. Yearley (1995) argues for a different treatment of jer vowels, suggesting that they are underlyingly mora-less vowels that are realized in order to avoid a complex coda. For example, in kus-ók ‘piece . N.SG’, the jer vowel [o] is realized to eliminate the complex coda [sk]. In contrast, in kus-k-á ‘piece.GEN.SG’, there are two syllables [kus] and [ka], neither of which has a complex coda, therefore, a jer is not realized (47). The Russian jer vowels [ɛ] and [o] are phonologically represented by the notations /E/ and /O/, respectively.  (47)  a.   kus-ók    [kusók]    UR:  /kus-Ok/    ‘piece . N.SG’        b.  kus-k-á      [kuská]    UR:  /kus-Ok-a/     ‘piece.GEN.SG’   According to Yearley (1995), although a jer is not realized (is not present on the surface), it is still present underlyingly (UR), as illustrated in (47b).  11  The suffix -ec is not a diminutive size suffix in this word, but its homophonous counterpart which means ‘person’ (see §4.5.2 on homophonous suffixes). 60  Russian diminutive suffixes -k (-ok, -ek) and -c (-ec) also contain jer vowels. As such, they have [ɛ] and [o] vowels that alternate with zero. For example, in (48a), the diminutive -ek contains the jer vowel [ɛ] in Nom case: ovráž-ek ‘small ditch (N.SG)’. In (48b), this vowel is not realized in Gen case: ovráž-k-a ‘ditch (GEN.SG)’. In (48c), the diminutive -ec also contains the jer vowel [ɛ] in Nom case: brát’-ec ‘small brother (N.SG)’. In (48d), it is not realized in Gen case: brát-c-a ‘small brother (GEN.SG)’.  (48) ɛ ~ Ø       a.  ovráž-ek     UR: /ovraž-Ok/                     b.  ovráž-k-a    UR: /ovraž-Ok-a/              ditch-EXPR.N.SG (MASC)                                         ditch-EXPR-GEN.SG (MASC)            ‘small ditch’                                                  ‘small ditch’          c.  brát’-ec       UR: /brat’-Ec/                       d.  brát-c-a       UR: /brat’-Ec-a/            brother-EXPR.N.SG (MASC)                                    brother-EXPR-GEN.SG (MASC)               ‘small brother’                                                ‘small brother’  In (49a), the diminutive -ok contains the jer vowel [o] in Nom case: gr’ib-ók ‘small mushroom (N.SG)’. In (49b), this vowel is not realized in Gen case: gr’ib-k-á ‘small mushroom (GEN.SG)’.  (49) o ~ Ø          a.  gr’ib-ók      /gr’ib-Ok/                            b.  gr’ib-k-á    /gr’ib-Ok-a/            mushroom-EXPR.N.SG (MASC)                              mushroom-EXPR-GEN.SG (MASC)               ‘small mushroom’                                         ‘small mushroom’  The diminutives -ok, -ek and -k are phonologically the same, namely /-Ok/, as the distinction “e” vs. “o” is purely orthographic here. Thus, the phonological representation for what I refer to as the diminutive suffixes -ok, -ek, and -k is just /-Ok/. The diminutives -ec and -c in (48c, d) are also the same phonologically. The phonological representation for what I refer to as the diminutive suffixes -ec and -c is /-Ec/.  3.4.2. Consonantal alternations In Russian, velar consonants and [c] undergo softening before the diminutive suffixes -k (allomorphs: -ok, -ek, -ik) and -ec (allomorph: -ic). For example, in (50b), the final [k] of the 61 stem ruk- ‘hand’ becomes [č’] before the suffix -k. In (50c), the final [k] of the diminutive allomorph -ek also becomes [č’].  (50)   k ~ č’       a.  ruk-á                     b.  rúč’-k-a                         c.  rúč’-eč’-k-a              hand-N.SG (FEM)        hand-EXPR-N.SG (FEM)       hand-EXPR-EXPR-N.SG (FEM)               ‘hand’                        ‘small hand’                     ‘very small hand’  In (51b), the final [c] of the stem ovc- ‘sheep’ becomes [č’] before the diminutive -k. In (51c), the final [k] of the diminutive allomorph -ek also becomes [č’].  (51) c ~ č’          a.  ovc-á                     b.  ov’éč’-k-a                     c.  ov’éč’-eč’-k-a             sheep-N.SG (FEM)        sheep-EXPR-N.SG (FEM)      sheep-EXPR-EXPR-N.SG (FEM)              ‘sheep’                      ‘small sheep’                   ‘very small sheep’  In (52b), the final [g] of the stem róg ‘horn’ becomes [ž] before the diminutive -ok. In (52c), the final [k] of the diminutive allomorph -ok becomes [č’].  (52) g ~ ž        a.  róg                         b.  rož-ók                           c.  rož-óč’-ek              horn.N.SG (MASC)        horn-EXPR.N.SG (MASC)    horn-EXPR-EXPR.N.SG (MASC)               ‘horn’                        ‘small horn’                     ‘very small horn’  In (53b), the final [x] of the stem br’úx- ‘belly’ becomes [š] before the diminutive -k. In (53c), the final [k] of the diminutive allomorph -ek becomes [č’].   (53) x ~ š       a.  br’úx-o                   b.  br’uš-k-ó                      c.  br’úš-eč’-k-o                belly-N.SG (NEUT)        belly-EXPR-N.SG (NEUT)    belly-EXPR-EXPR-N.SG (NEUT)                ‘belly’                        ‘small belly’                    ‘very small belly’  In some stems, such as karmán ‘pocket’ and okn- ‘window’, the final [n] is replaced by [š] before a diminutive suffix (54)─(55). In (55b), okóš-k-o ‘small window’, the second vowel [o] is a jer vowel, which is part of the phonological representation of the okn- stem itself: the stem is phonologically /okOn-/. 62  (54) n ~ š          a.  karmán                     b.  karmáš-ek               pocket.N.SG (MASC)                                 pocket-EXPR.N.SG (MASC)                ‘pocket’                                                  ‘small pocket’   (55) n ~ š        a.  okn-ó                                             b.  okóš-k-o               window-N.SG (NEUT)                                     window-EXPR-N.SG (NEUT)               ‘window’                                               ‘small window’  Some expressive suffixes, such as -ag, -ar, -on, -ug, -uk, and -ur trigger palatalization of the final consonant in the base (56). Other expressive suffixes, such as -al, -an, -an’, -ax, -ob, -ot, -ox, -ul’, -un’, -us’, -uš, and -ux do not trigger palatalization (57). This issue is discussed in Chapter 7.  (56)  a.  žád-n-ij                                       b.  žad’-úg-a        (triggers palatalization)             stingy-ADJ-MASC.N.SG                            stingy-EXPR-N.SG (MASC/FEM)            ‘stingy’                                              ‘stingy animate (vulg)’  (57)  a.  žád-n-ij                                       b.  žad-ób-a       (does not trigger palatalization)             stingy-ADJ-MASC.N.SG                            stingy-EXPR-N.SG (MASC/FEM)            ‘stingy’                                              ‘stingy animate (vulg)’  63 Chapter 4: How are expressive suffixes merged?  4.1. INTRODUCTION In this chapter, I show that the two different semantic classes of expressive suffixes (attitude and size suffixes) map onto two different syntactic classes. For convenience, I repeat the simplex attitude and size suffixes below.  Affectionate suffixes -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’, -úš Vulgar suffixes -ág, -ák, -ál, -án, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx Table 4.1: Attitude suffixes   Diminutive suffixes -k (allomorphs: -ok, -ek, -ik) -c (allomorphs: -ec, -ic) Augmentative suffix -išč’ Table 4.2: Size suffixes  I argue that attitude suffixes merge as syntactic heads (1a), while size suffixes merge as syntactic modifiers (1b).  (1) a. HEADS                    X                        b. MODIFIERS             Y                              2                                                2                            X            Y                                             X            Y                            EXPRattitude                                                  EXPRsize  The distinction between heads and modifiers lies in the projection of category features (Bierwisch 2003, Schütze 1995, Bachrach & Wagner 2007). Heads project, thus they determine the category and grammatical features of the output. In contrast, modifiers do not project, thus they do not determine the category and grammatical features of the output.  The following three diagnostics will be used to determine the syntactic type of expressive suffixes in Russian (2).   64 (2) Diagnostics (cf. Bachrach & Wagner 2007:4) Diagnostic I:  Do expressive suffixes change syntactic category? Diagnostic II: Do expressive suffixes change grammatical gender? Diagnostic III: Do expressive suffixes change inflectional class?  Expressive suffixes are classified as syntactic heads if any of the answers to (2) are affirmative. In contrast, expressive suffixes are classified as modifiers if the answers to (2) are negative (Table 4.3).  Diagnostics Syntactic heads Syntactic modifiers Do expressive suffixes change syntactic category? a * Do expressive suffixes change grammatical gender? a * Do expressive suffixes change inflectional class? a * Table 4.3: Diagnostics for syntactic heads vs. syntactic modifiers  Below I use these diagnostics to show that attitude suffixes are heads, while size suffixes are modifiers. In §4.2, I investigate whether expressive suffixes change syntactic category, and in §4.3, whether they change grammatical gender. In §4.4, I investigate whether they change inflectional class. In §4.5, I analyze non-expressive suffixes which are homophonous with size suffixes. I show that the non-expressive suffixes differ from size suffixes both in meaning and in syntactic structure. Finally, in §4.6, I present the conclusions.  4.2. CHANGE IN CATEGORY Here I apply Diagnostic I (change in syntactic category). I illustrate that attitude suffixes produce a change in syntactic category, thus they behave like syntactic heads. Size suffixes do not produce a change in syntactic category, thus they behave like syntactic modifiers.  I provide evidence that attitude suffixes can merge with any input category (n/a/v), always forming a noun, no matter what the input category is. In contrast, size suffixes can only merge with nouns, and the suffixed form remains a noun. In other words, attitude suffixes act as nominalizers (3a), while size suffixes act as noun modifiers (3b). Since it is not clear what 65 the category of a modifier is, I treat modifiers as a single sound-meaning correspondence without categorial information (3b).  (3) a. HEADS          n                   b. MODIFIERS             n                  2                                        2                n            n/a/v                              EXPRsize        n                EXPRattitude  In §4.2.1, I analyze attitude suffixes; in §4.2.2, I analyze size suffixes; and in §4.2.3, I present the conclusions.  4.2.1. Attitude suffixes The data in (4)–(6) illustrate that attitude suffixes turn adjectives into nouns. For example, in (4), žád-n-ij ‘stingy’ is an adjective, as evidenced by the fact that it contains the productive adjectival suffix -n (4a). When the attitude suffix -ug attaches (4b), the adjective turns into a noun, as evidenced by the fact that it has the nominal ending -a. The same pattern is observed in (5)–(6). Attitude suffixes consistently turn adjectives into nouns, supporting the claim that they function as heads.  (4)  a.  žád-n-ij                                     b.  žad-n’-úg-a           stingy-ADJ-MASC.N.SG                           stingy-ADJ-EXPR-N.SG (MASC/FEM)           ‘stingy’                                            ‘stingy animate (vulg)’  (5)  a.  gr’áz-n-ij                                   b.  gr’áz-n-úx-a            dirty-ADJ-MASC.N.SG                            dirty-ADJ-EXPR-N.SG (MASC/FEM)            ‘dirty’                                             ‘dirty animate (vulg)’  (6)  a.  rod-n-ój                                    b.  rod-n-úl’-a            kin-ADJ-MASC.N.SG                             kin-ADJ-EXPR-N.SG (MASC/FEM)           ‘dear’                                             ‘dear animate (affect)’  Moreover, attitude suffixes can also turn verbs into nouns, as illustrated in (7)–(9). For example, in (7), pr’i-l’ip-á-t’ ‘to cling’ is a verb, as evidenced by the fact that it contains the verbal prefix pri- (this prefix can only attach to verbs) (7a). When the attitude suffix -al attaches, the verb turns into a noun with the nominal ending -a. The same pattern is observed with different attitude suffixes -aš and -ux in (8)–(9).  66 (7)  a.  pr’i-l’ip-á-t’                               b.  pr’i-l’ip-ál-a           VERB.PREF-cling-TH-INF                       VERB.PREF-cling-EXPR-N.SG (MASC/FEM)           ‘to cling’                                        ‘clinging animate (vulg)’  (8)  a.  ras-t’er’-á-t’                               b.  ras-t’er’-áš-a           VERB.PREF-lose-TH-INF                        VERB.PREF-loose-EXPR-N.SG (MASC/FEM)             ‘to lose’                                          ‘animate who loses things (affect)’  (9)  a.  za-v’ir-á-t’                                  b.  za-v’ir-úx-a           VERB.PREF-lie-TH-INF                            VERB.PREF-lie-EXPR-N.SG (MASC/FEM)             ‘to lie’                                             ‘lying animate (affect)’  Evidence that the verbal prefix attaches before the suffix stems from the fact that the attitude suffixes cannot attach to the √Root directly. Such forms are ungrammatical, as shown in (10)–(12). These attitude suffixes can in general attach to √Roots (e.g., xník-al-a ‘complaining animate’, kut’-áš-a ‘carousing animate’, etc.), but they do not attach to √Roots in these particular cases and prefixation after suffixation presupposes that (10)–(12) should already exist.  (10)    * l’ip-ál-a           cling-EXPR-N.SG (MASC/FEM)           ‘clinging animate (vulg)’  (11)    * t’er’-áš-a           lose-EXPR-N.SG (MASC/FEM)             ‘animate who loses things (affect)’  (12)    * v’ir-úx-a           lie-EXPR-N.SG (MASC/FEM)             ‘lying animate (affect)’  Finally, attitude suffixes can combine with nouns, as illustrated in (13)–(14). Nouns that are used with attitude suffixes predictably do not change their syntactic category. For example, in (13a), the word čud-ák ‘an eccentric’ is a noun, as evidenced by the fact that it contains the nominal suffix -ak. When the vulgar suffix -in attaches, there is no change in syntactic category.  (13)  a.  č’ud-ák                                      b.  č’ud-ač’-ín-a             wonder-NOM.N.SG (MASC)                wonder-NOM-EXPR-N.SG (MASC/FEM)            ‘an eccentric’                                   ‘an eccentric (vulg)’  67 (14)  a.  kras-ot-á                                    b.  kras-ot-úl’-a              pretty-NOM-N.SG (FEM)                         pretty-NOM-EXPR-N.SG (MASC/FEM)             ‘beauty/prettiness’                           ‘pretty animate (affect)’  These findings are summarized in Table 4.4.  EXPRattitude Input  Output adjective noun verb noun -án’, -áš, -ón, -úl’, -ún’-, -úr, -ús’, -úš, -ág, -ák, -ál, -án, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx noun noun Table 4.4: Attitude suffixes (change in category)  4.2.2. Size suffixes In contrast to attitude suffixes, expressive size suffixes do not turn adjectives and verbs into nouns. For example, in (15) the adjective žád-n-ij ‘stingy’ does not become a noun when a size suffix attaches. Instead, all examples with size suffixes merging with this adjective are ungrammatical (15b, c, d). The same behaviour is shown in the data with the adjective gr’áz- n-ij ‘dirty’ (16).  (15)  a.  žád-n-ij                                     b.* žad-n-(o)k/*žad’-en-k-a             stingy-ADJ-MASC.N.SG                        stingy-ADJ-EXPR.N.SG             ‘stingy’                                           ‘stingy animate (dim)’           c. * žad-n’-(e/i)c-(a)                    d.* žad-n-išč’-e               stingy-ADJ- EXPR-N.SG                     stingy-ADJ-EXPR-N.SG               ‘stingy animate (dim)’                  ‘stingy animate (aug)’  (16)  a.  gr’áz-n-ij                                  b.* gr’az-n-(o)k/*gr’az’-en-k-a             dirty-ADJ-MASC.N.SG                           dirty-ADJ-EXPR.N.SG             ‘dirty’                                             ‘dirty animate (dim)’           c. * gr’az-n’-(e/i)c-(a)                  d.* gr’az-n-išč’-e               dirty-ADJ- EXPR.N.SG                           dirty-ADJ-EXPR-N.SG               ‘dirty animate (dim)’                   ‘dirty animate (aug)’  Similarly, size suffixes cannot attach to verbs to create new nouns (17). All the data where size suffixes merge with verbs are ungrammatical (17b, c, d). (17)  a.  pr’i-l’ip-á-t’                             b.* pr’i-l’ip-(o)k-(a)             VERB.PREF-cling-TH-INF                       VERB.PREF-cling-EXPR.N.SG             ‘to cling’                                        ‘clinging animate (dim)’ 68          c. * pr’i-l’ip’-(e/i)c-(a)                  d.* pr’i-l’ip-išč’-e              VERB.PREF-cling-EXPR.N.SG                VERB.PREF-cling-EXPR.N.SG             ‘clinging animate (dim)’                 ‘clinging animate (aug)’  I have shown that size suffixes cannot change adjectives and verbs into nouns. Below I show that they cannot combine with these categories to preserve a category. The examples (18) and (19) illustrate that size suffixes cannot attach to adjectives or verbs to mean ‘a little bit’ or ‘a lot.’ In (18b), the diminutive suffix -k is added to the adjective ‘stingy’. The resulting word does not mean ‘a little bit stingy’, but instead it is ungrammatical. The same holds for (18c) and (18d), where the diminutive -c and the augmentative -išč’ are ungrammatical with the same adjective.  (18)  a.  žád-n-ij                                     b.* žad-n-(o)k-ij/*žad’-en-k-ij             stingy-ADJ-MASC.N.SG                         stingy-ADJ-EXPR-MASC.SG             ‘stingy’                                           ‘a little bit stingy (dim)’           c. * žad-n’-(e/i)c-ij                        d.* žad-n-išč’-ij               stingy-ADJ-EXPR-MASC.SG               stingy-ADJ-EXPR-MASC.SG               ‘a little bit stingy (dim)’                ‘a lot stingy (aug)’  In (19), size suffixes are added to the verb ‘to cling.’ The resulting words do not mean ‘to cling a little bit’ or ‘to cling a lot,’ but instead are ungrammatical.  (19)  a.  pr’i-l’ip-á-t’                              b.* pr’i-l’ip-(o)k-a-t’             VERB.PREF-cling-TH-INF                       VERB.PREF-cling-EXPR-TH-INF             ‘to cling’                                        ‘to cling a little bit (dim)’          c. * pr’i-l’ip’-(e/i)c-a-t’                  d.* pr’i-l’ip-išč’-a-t’              VERB.PREF-cling-EXPR-TH-INF          VERB.PREF-cling-EXPR-TH-INF              ‘to cling a little bit (dim)’              ‘to cling a lot (aug)’  The data above illustrate not only that size suffixes are unable to turn adjectives and verbs into nouns, but that they are simply unable to combine with these categories. Although size suffixes cannot combine with adjectives or verbs, they are very productively used with nouns (20)–(22).  (20)  a.  č’ud-ák                                       b.  č’ud-ač’-ók             wonder-NOM.N.SG (MASC)               wonder-NOM-EXPR.N.SG (MASC)             ‘an eccentric’                                 ‘a little eccentric’ 69  (21)  a.  kras-ot-á                                     b.  kras-ot-íšč’-a             pretty-NOM-N.SG (FEM)                        pretty-NOM-EXPR-N.SG (FEM)             ‘beauty/prettiness’                         ‘big beauty’  (22)  a.  sos-ún                                         b.  sos-un’-éc             suck-NOM.N.SG (MASC)                          suck-NOM-EXPR.N.SG (MASC)             ‘suckling’                                        ‘little suckling’  In other words, size suffixes can only combine with nouns, producing no change in syntactic category: a noun suffixed with a size suffix remains a noun.  This evidence is, however, not fully conclusive to consider size suffixes syntactic modifiers as they are represented in (23).  (23) MODIFIER             n                       2                EXPRsize         n  The data above would be also consistent with the hypothesis that size suffixes are syntactic heads, but just do not combine with adjectives or verbs. More conclusive evidence that size suffixes are noun modifiers will be given in §4.3 and §4.4, where I show that they produce no change in grammatical gender or inflectional class of a noun.  These findings are summarized in Table 4.5.  EXPRsize Input  Output adjective *noun/*adjective verb *noun/*verb -k/-ek/-ok/-ik, -c/-ec/-ic, -išč’ noun noun Table 4.5: Size suffixes (no change in category)  4.2.3. Summary Attitude suffixes can turn adjectives and verbs into nouns. They can also combine with nouns to return nouns. In other words, no matter what the input category is, the resulting category is always a noun.  70 Size suffixes demonstrate a different behaviour. They cannot combine with adjectives and verbs, but can only combine with nouns. And when they combine with nouns, they do not change syntactic category. In Table 4.6, I compare attitude and size suffixes.   Input  Output adjective noun verb noun EXPRattitude noun noun adjective *noun/*adjective verb *noun/*verb EXPRsize noun noun Table 4.6: Comparison of attitude and size suffixes  Attitude and size suffixes have different formal properties with respect to a change in syntactic category: attitude suffixes can change a category, while size suffixes cannot (Table 4.7).   Change in category EXPRattitude a                                 EXPRsize * Table 4.7: Change in category  Based on Diagnostic I, because attitude suffixes change syntactic category, they behave like syntactic heads (24a). In contrast, size suffixes do not change a category, thus they seem to behave like syntactic modifiers (though the evidence is inconclusive) (24b).  (24) a. HEAD          n                   b. MODIFIER               n                  2                                        2                n           n/a/v                               EXPRsize        n                EXPRattitude  4.3. CHANGE IN GRAMMATICAL GENDER In this section I apply Diagnostic II (change in grammatical gender). Given the change in category that we’ve seen above, we predict that attitude suffixes should be able to change grammatical gender, while size suffixes should not. Here I show that this is indeed the case. Attitude suffixes can change grammatical gender and thus, they behave like syntactic heads 71 (25a). Size suffixes cannot change grammatical gender and thus, they behave like syntactic modifiers (25b). In the next section (4.4) I show that the change in grammatical gender is not due to attitude suffixes themselves being specified for gender, but it is the indirect result of a change in inflectional class caused by these suffixes.  (25) a. HEAD             n2[gender Y]                b. MODIFIER                    n[gender X]                     2                                                     2                n2[gender Y]     n1[gender X]                              EXPRsize            n[gender X]                  EXPRattitude   2                                                         2                        n1[gender X]    √Root                                         n[gender X]           √Root   In §4.3.1, I analyze attitude suffixes; in §4.3.2, I analyze size suffixes; and in §4.3.3, I present the conclusions.  4.3.1. Attitude suffixes As mentioned in Chapter 3, grammatical gender in Russian is largely dependent on animacy and natural gender (Corbett 1982, 1991). To show how attitude suffixes change gender, I first propose a formal implementation of gender assignment; then I analyze the change in grammatical gender. In §4.3.1.1, I discuss animate nouns; in §4.3.1.2, I discuss inanimate nouns; and in §4.3.1.3, I summarize the findings.  4.3.1.1. Animate nouns As noted in Chapter 3, Russian animate nouns denote living beings, like humans and animals. Animate nouns can be sex-differentiable and non-sex-differentiable (26). Sex- differentiable nouns are those that include natural gender (male or female) as part of their semantics. Non-sex-differentiable nouns (including common gender nouns) do not have natural gender. Common gender nouns will be discussed later in this section.  72 (26)                                            Animate                                             3                                  Sex-differentiable                Non-sex-differentiable               3                (including common gender)                              Male                  Female          Examples:      brát                 s’estr-á                   č’elov’ék                            ‘brother’             ‘sister’                     ‘person’  For example, the nouns brát ‘brother’ and s’estr-á ‘sister’ are animate and sex-differentiable. They are animate because they denote living beings. They are sex-differentiable because brát ‘brother’ can only denote a male being, while s’estr-á ‘sister’ can only denote a female being. In contrast, the noun č’elov’ek ‘person’ is non-sex-differentiable because it can denote persons of either sex.  Recall further that in Russian sex-differentiable nouns, natural gender always takes precedence over grammatical gender. Thus, a noun that has the natural gender “male” is always masculine (27). A noun that has the natural gender “female” is always feminine (28).  (27)  a.  xoróš-ij              brát                         b.* xoróš-aja               brát             good-MASC.N.SG   brother.N.SG (MASC)      good-FEM.N.SG      brother.N.SG (FEM)            ‘good brother’                                              ‘good brother’  (28)  a.  xoróš-aja               s’estr-á                     b.* xoróš-ij                 s’estr-á             good-FEM.N.SG       sister-N.SG (FEM)              good-MASC.N.SG    sister-N.SG (MASC)            ‘good sister’                                                 ‘good sister’  As Russian non-sex-differentiable nouns do not have natural gender, there is no dependency of grammatical gender on natural gender in such nouns. Like sex-differentiable nouns, non- sex-differentiable ones are either masculine or feminine; but unlike the former, the grammatical gender of non-sex-differentiable nouns seems arbitrary. For example, compare č’elov’ék ‘person’ and p’ersón-a ‘person.’ Both nouns are non-sex-differentiable because they can denote male and female persons, but č’elov’ék ‘person’ is masculine (29), while p’ersón-a ‘person’ is feminine (30).  (29)  a.  bol’š-ój           č’elov’ek                     b.* bol’š-ája               č’elov’ek             big-MASC.N.SG    person.N.SG (MASC)          big-FEM.N.SG        person.N.SG (FEM)             ‘big person’                                                  ‘big person’  73 (30)  a.  bol’š-ája            p’ersón-a                     b.* bol’š-ój            p’ersón-a             big-FEM.N.SG      person-N.SG (FEM)               big-MASC.N.SG     person-N.SG (MASC)             ‘big person’                                                 ‘big person’  Thus, sex-differentiable nouns can be masculine or feminine depending on natural gender. Non-sex-differentiable nouns can also be masculine or feminine, but their gender is not determined by natural gender (31).  (31)                                  Animate                                       3              Sex-differentiable                Non-sex-differentiable                     3                            3           Male                Female              g                         g       [MASCULINE]   [FEMININE]    [MASCULINE]   [FEMININE]   In the framework of Distributed Morphology, animacy and natural gender are analyzed as part of the semantic information of the √Root (Müller 2005). For example, the semantics of the √Root s’estr- ‘sister’ indicates that it is animate (denotes a living being) and sex- differentiable (denotes a female). When the √Root s’estr- is nominalized by combining with a functional head n, the grammatical gender of the resulting noun depends on the natural gender ‘female’ which is encoded as part of the semantics of the √Root. As the natural gender ‘female’ always determines feminine grammatical gender, the resulting noun s’estr-á ‘sister’ is feminine (32). From now on I will use the notations (animate)/(inanimate) and (male)/(female) to refer to animacy and natural gender of the √Root. In the framework of Distributed Morphology it is assumed that categorized nouns as well as categorized adjectives and verbs are stored as idioms (Marantz 2001), which ensures that a particular √Root is associated with the right grammatical features (e.g., grammatical gender, inflectional class, etc.). Therefore, under this assumption, the information that s’estr-á ‘sister’ is a noun with the grammatical feature [FEMININE] would be stored as an idiom.  (32)                                     n[fem]                                        2                                        n[fem]        √s’estr-                                           (animate) (female)  74 In contrast to the √Root s’estr- discussed above, the √Root brat- ‘brother’ can only denote a male being. Since the natural gender “male” always determines masculine grammatical gender, the resulting noun brát ‘brother’ is masculine (33).  (33)                                     n[masc]                                       2                                     n[masc]     √brat-                                             (animate) (male)  The √Roots č’elov’ek- ‘person’ and p’erson- ‘person’ do not have natural gender as part of their semantics and can denote both male and female persons. As a result, their grammatical genders seem to be assigned arbitrarily: the noun č’elov’ék ‘person’ is masculine (34a), and the noun p’ersón-a ‘person’ is feminine (34b).  (34) a.   n[masc]                            b.   n[fem]         2                               2        n[masc]   √č’elov’ek-                n[fem]      √person-                      (animate)                               (animate)  Next we turn to the formal analysis of  common gender nouns (e.g., s’irot-á ‘orphan,’ sud’-j- á ‘judge,’ koll’ég-a ‘colleague’, etc.). I propose that the formal difference between common gender nouns and other nouns in Russian is that the former are unspecified for grammatical gender, while the latter are specified. All Russian common gender nouns are animate; there are no inanimate common gender nouns. And since Russian animate nouns can be either masculine or feminine, common gender nouns can also be either masculine or feminine. Thus, I propose the following representation for nouns of common gender (35).  (35)                                    n                       ← unspecified for grammatical gender                                       2                                         n         √s’irot-                                              (animate)  To summarize, my approach allows us to understand gender assignment in Russian nouns. I identify the following three types of animate nouns in Russian:   75 TYPE I: Animate sex-differentiable nouns whose grammatical gender is determined by their natural gender (36a) TYPE II: Animate non-sex-differentiable nouns whose grammatical gender seems to be assigned arbitrarily (36b)             TYPE III: Animate common gender nouns that are unspecified for grammatical                          gender (36c)   (36)     a. TYPE I      n[masc]/[fem]                    b. TYPE II      n[masc]/[fem]         c. TYPE III       n                            2                                           2                                   2                        n[masc]/[fem]     √Root                               n[masc]/[fem]   √Root                           n          √Root                                   (animate) (male/female)                             (animate)                                (animate)   Now that gender assignment has been discussed, we can apply Diagnostic II to see if there is any change in grammatical gender when attitude suffixes are added. Here I show that attitude suffixes indeed produce a change in grammatical gender. This change involves TYPE II nouns (36b).  The noun in (37a) is animate but it is non-sex-differentiable (TYPE II). The grammatical gender assigned to this noun is masculine. When the attitude suffix -ug attaches, as in (37b), the grammatical gender changes. The resulting word is now a common gender noun that can trigger masculine or feminine agreement (MASC/FEM), as shown in (38a-b).  (37)  a.  zv’ér’                                          b.  zv’er’-úg-a            animal.N.SG (MASC)                               animal-EXPR-N.SG (MASC/FEM)            ‘animal’                                           ‘animal (vulg)’  (38)  a.  bol’š-ój                  zv’er’-úg-a              big-ADJ.MASC.N.SG   animal-EXPR-N.SG (MASC)              ‘big animal (vulg)’          b.  bol’š-ája                 zv’er’-úg-a              big-ADJ.FEM.N.SG    animal-EXPR-N.SG (FEM)             ‘big animal (vulg)’  While the common gender noun can trigger either masculine or feminine agreement, the masculine noun can only trigger masculine agreement (39). 76  (39)  a.  bol’š-ój                    zv’ér’             big-ADJ.MASC.N.SG     animal.N.SG (MASC)             ‘big animal’          b.* bol’š-ája                   zv’ér’               big-ADJ.FEM.N.SG      animal.N.SG (FEM)               ‘big animal’  According to the current proposal, common gender nouns are unspecified for grammatical gender. For this reason, they can trigger either masculine or feminine grammatical agreement. Thus, a change in grammatical gender is more accurately described as neutralization of grammatical gender: a noun that is otherwise specified for grammatical gender appears to lose grammatical gender. This is schematized in (40). The animate noun zv’ér’ ‘animal’ is specified as grammatically masculine. When the attitude suffix -ug merges, as in (40b), grammatical gender is neutralized. As a result, the newly formed noun zv’er’-úg- a ‘animal (vulg)’ is unspecified for gender and can therefore trigger masculine or feminine agreement.  (40)      a.      n[masc]                 2                n[masc]    √zv’er’-                            (animate)               b.        n2                              ← unspecified for grammatical gender             2         n2            n1[masc]             -ug         2                         n1[masc]  √zv’er’-                                    (animate)  This phenomenon is not restricted to grammatically masculine nouns; it is also found with feminine nouns, as shown in (41). The Russian noun tvár’ ‘animal’ is animate and non-sex- differentiable (TYPE II): it does not denote natural gender as part of its semantics. The grammatical gender of this noun is feminine as evidenced by the obligatory feminine agreement with adjectives (42). In (41b), the attitude suffix -uk is added, which neutralizes the grammatical gender. The resulting noun tvár’-úk-a ‘animal (vulg)’ is now in common gender and as such can trigger both masculine and feminine agreement (43a-b) 77  (41)  a.  tvár’                                            b.  tvar’-úk-a             animal.N.SG (FEM)                                   animal-EXPR-N.SG (MASC/FEM)             ‘animal’                                           ‘animal (vulg)’  (42)  a.  bol’š-ája                 tvár’             big-ADJ.FEM.N.SG    animal.N.SG (FEM)             ‘big animal’           b.* bol’š-ój                  tvár’                big-ADJ.MASC.N.SG animal.N.SG (MASC)                ‘big animal’  (43)  a.  bol’š-ój                   tvar’-úk-a             big-ADJ.MASC.N.SG  animal-EXPR-N.SG (MASC)             ‘big animal (vulg)’          b.  bol’š-ája                  tvar’-úk-a             big-ADJ.FEM.N.SG   animal-EXPR-N.SG (FEM)             ‘big animal (vulg)’  The analysis of this pattern is schematized in (44)–(45). The simplex noun is associated with feminine gender. When the attitude suffix -uk merges as in (45), the grammatical gender is neutralized: the resulting noun is unspecified for grammatical gender.  (44)               n[fem]                  2                 n[fem]      √tvar’-                           (animate)  (45)                 n2                              ← unspecified for grammatical gender             2         n2             n1[fem]             -uk          2                          n1[fem]   √tvar’-                                    (animate)  Note that neutralization of grammatical gender is not restricted to examples where the base is a simplex noun. It is also found when the attitude suffix combines with a complex noun. For example, in (46a), the noun dur-ák ‘stupid animate’ is formed by means of the productive nominalizing suffix -ak. This complex noun is grammatically masculine, as evidenced by the agreement patterns in (47). When the attitude suffix -in is added as in (46b), the newly 78 formed noun is a common gender noun. Consequently, this noun is compatible with either masculine or feminine agreement (48a-b).  (46)  a.  dur-ák                                         b.  dur-ač’-ín-a             stupid-NOM.N.SG (MASC)                       stupid-NOM-EXPR-N.SG (MASC/FEM)            ‘stupid animate’                               ‘stupid animate (vulg)’  (47)  a.  bol’š-ój                     dur-ák             big-ADJ.MASC.N.SG   stupid-NOM.N.SG (MASC)             ‘big stupid animate’           b.* bol’š-ája                 dur-ák                big-ADJ.FEM.N.SG   stupid-NOM.N.SG (FEM)               ‘big stupid animate’   (48)  a.  bol’š-ój                     dur-ač’-ín-a             big-ADJ.MASC.N.SG   stupid-NOM-EXPR-N.SG (MASC)             ‘very stupid animate (vulg)’           b.  bol’š-ája                   dur-ač’-ín-a              big-ADJ.FEM.N.SG    stupid-NOM-EXPR-N.SG (FEM)              ‘very stupid animate (vulg)’  The proposed structures for (47) and (48) are shown in (49) and (50), respectively. Suffixation of the expressive suffix results in a lack of grammatical gender.  (49)               n[masc]                 2                n[masc]     √dur-               -ak        (animate)  (50)                 n2                              ← unspecified for grammatical gender             2         n2             n1[masc]              -in         2                          n1[masc]     √dur-                         -ak         (animate)  To summarize, Russian attitude suffixes neutralize grammatical gender of TYPE II nouns (animate, non-sex-differentiable). Nouns used with attitude suffixes always end up having common gender, regardless of the grammatical gender of the input (Table 4.8). The 79 following suffixes all display this behaviour, which is exactly the group of suffixes that changes syntactic category (with the exception of the suffix -an) (compare with Table 4.4).  EXPRattitude Input  Output -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’, -úš, -ág, -ák, -ál, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx animate, [MASCULINE]  animate, [FEMININE] animate, unspecified  animate, unspecified Table 4.8: Attitude suffixes (used with TYPE II nouns)  The suffix -an is exceptional in this respect, because there is no direct evidence that it can change a TYPE II animate noun into a common gender noun. There are just a few examples when this suffix attaches to animate nouns, and in all these examples both the input and the output are masculine nouns (51)–(52).  (51)  a.  brát                                         b.  brat-án             brother.N.SG (MASC)                         brother-EXPR.N.SG (MASC)             ‘brother’                                     ‘brother (vulg)’  (52)  a.  star’-ík                                   b.  star’-ik-án             old-NOM.N.SG (MASC)                      old-NOM-EXPR.N.SG (MASC)            ‘old man’                                     ‘old man (vulg)’  For completeness, I show that nouns of the remaining two types (TYPE I and TYPE III) do not change grammatical gender (and grammatical agreement) when attitude suffixes attach. This is expected, however, because the gender of TYPE I nouns (animate, sex-differentiable) is always determined by their natural gender (male/female) and TYPE III nouns (common gender) are unspecified for gender in the first place, and so there could be no change. I discuss both types of nouns in more detail below.  TYPE I nouns contain natural gender as part of their semantics. As natural gender always determines grammatical gender in Russian, male nouns are always masculine and female nouns are always feminine. When attitude suffixes merge with these nouns, the nouns remain semantically male or female, and therefore, there is no change in grammatical gender (including neutralization). This shows that only grammatical gender can be neutralized, but not natural gender. This is illustrated in (53)–(56) below.  80 In (53), the noun s’estr-á ‘sister’ denotes a female individual and therefore its gender is feminine. When the attitude suffix -ux attaches, the complex noun remains semantically female, and thus, it is grammatically feminine. In (54), the noun sín ‘son’ denotes a male individual and therefore its gender is masculine. When the attitude suffix -ul’ is added, the resulting noun remains semantically male and thus, it is grammatically masculine. The same pattern is found with all TYPE I nouns; more examples are given in (55)–(56).  (53)  a.  s’estr-á                                        b.  s’estr-úx-a             sister-N.SG (FEM)                                      sister-EXPR-N.SG (FEM)             ‘sister’                                             ‘sister (vulg)’  (54)  a.  sín                                               b.  sin-úl’-a             son.N.SG (MASC)                                       son-EXPR-N.SG (MASC)             ‘son’                                                ‘son (affect)’  (55)  a.  d’év-a                                          b.  d’ev-áx-a             girl-N.SG (FEM)                                         girl-EXPR-N.SG (FEM)             ‘girl’                                                ‘girl (vulg)’  (56)  a.  pár’en’                                        b.  parn’-úg-a             guy.N.SG (MASC)                                       guy-EXPR-N.SG (MASC)            ‘guy’                                                ‘guy (vulg)’  The proposed structures for (53) and (54) are given in (57) and (58), respectively. In (57), the attitude suffix -ux does not neutralize grammatical gender of the noun s’estr-á ‘sister.’ The natural gender ‘female’ still determines feminine grammatical gender. In (58), the attitude suffix -ul’ does not neutralize grammatical gender of the noun brát ‘brother’ because it denotes a male, which determines masculine grammatical gender.   (57)                             n2[fem]                        2                     n2[fem]       n1[fem]                         -ux         2                                     n1[fem]    √s’estr-                                             (animate) (female)  81 (58)                               n2[masc]                        2                     n2[masc]      n1[masc]                          -ul’        2                                     n1[masc]   √sin-                                             (animate) (male)  To summarize, attitude suffixes do not produce a change in the grammatical gender of TYPE I nouns because these nouns are associated with natural gender (male or female), which in turn determines grammatical gender (Table 4.9).  EXPRattitude Input  Output -án’, -áš, -ón, -úl’, -ún’, -úr,  -ús’, -úš, -ág, -ák, -ál, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx animate, male, [MASCULINE]  animate, female, [FEMININE] animate, male, [MASCULINE]  animate, female, [FEMININE] Table 4.9: Attitude suffixes (used with TYPE I nouns)  Next we turn to TYPE III nouns (common gender). Under the current proposal, these nouns are unspecified for grammatical gender and can therefore trigger either masculine or feminine agreement. When attitude suffixes attach to such nouns, they remain unspecified for gender, and therefore, there is no change of grammatical gender. For example, in (59a-b), the noun s’irot-á ‘orphan’ is a common gender noun that can trigger either masculine or feminine agreement. When the attitude suffix -in attaches as in (60), no change in gender appears: the resulting noun can still trigger either either masculine or feminine agreement.  (59)  a.  bol’š-ój                    s’irot-á             big-ADJ-MASC.N.SG  orphan-N.SG (MASC)             ‘big orphan’           b.  bol’š-ája                 s’irot-á              big-ADJ-FEM.N.SG   orphan-N.SG (FEM)              ‘big orphan’  (60)  a.  bol’š-ój                    s’irot’-ín-a                 big-ADJ-MASC.N.SG  orphan- EXPR-N.SG (MASC)              ‘big orphan’  82       b.  bol’š-ája                   s’irot’-ín-a             big-ADJ-FEM.N.SG   orphan- EXPR-N.SG (FEM)             ‘big orphan’  The structure for (60) is given in (61). The attitude suffix -in merges with a common gender noun unspecified for grammatical gender. The resulting noun is also unspecified for grammatical gender.  (61)                 n2                              ← unspecified for grammatical gender             2         n2             n1                      ← unspecified for grammatical gender             -in          2                           n1          √s’irot-                                        (animate)  To summarize, attitude suffixes do not change grammatical gender of TYPE III nouns (common gender) because these nouns are unspecified for gender in the first place, and thus, no change can take place (Table 4.10). This is the same set of attitude suffixes that change syntactic category (compare with Table 4.4). The suffix -an is not listed in the table because it rarely attaches to animates, and when it does, the output is always TYPE I nouns (see (51)– (52) above).  EXPRattitude Input  Output -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’, -úš, -ág, -ák, -ál, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx  animate, unspecified animate, unspecified Table 4.10: Attitude suffixes (used with TYPE III nouns)  To conclude, attitude suffixes produce a change in grammatical gender of TYPE II nouns (animate, non-sex-differentiable). When attitude suffixes merge with TYPE II nouns, the resulting nouns turn into TYPE III nouns (animate, common gender). When attitude suffixes merge with TYPE I (animate, sex-differentiable) or TYPE III (animate, common gender) nouns, there is no change in grammatical gender (Table 4.11; change is indicted in bold).  83  Input  Output TYPE I (animate, sex-differentiable) TYPE I (animate, sex-differentiable) TYPE II (animate, non-sex-differentiable) TYPE III (animate, common gender)   EXPRattitude TYPE III (animate, common gender) TYPE III (animate, common gender) Table 4.11: Attitude suffixes (animate nouns)  4.3.1.2. Inanimate nouns In this section I show that attitude suffixes can also change the grammatical gender of inanimate nouns. Inanimate nouns can be associated with masculine, feminine, or neuter grammatical gender (62). For example, žurnál ‘magazine’ is masculine, gaz’ét-a ‘newspaper’ is feminine, and p’is’m-ó ‘letter’ is neuter.  (62)                                        Inanimate                                              9                              masculine     feminine         neuter         Examples:      žurnál        gaz’ét-a        p’is’m-ó                             ‘magazine’ ‘newspaper’ ‘letter’  When attached to inanimate nouns, the majority of attitude suffixes (except -an) derive feminine nouns, regardless of the grammatical gender of the input. The attitude suffix -an forms masculine nouns, regardless of grammatical gender of the input.  I start by analyzing attitude suffixes that form feminine nouns. Take for example the attitude suffix -in. As the data below indicate, -in can attach to nouns of all grammatical genders (masculine, feminine, neuter). But no matter what the gender of the base, -in derives a feminine noun as illustrated in (63)–(68). The masculine noun ovrág ‘ditch’ becomes feminine (64); the feminine noun jám-a ‘ditch’ remains feminine (66); and the neuter noun bolót-o ‘swamp’ becomes feminine (68). In the previous section I showed that -in can neutralize gender, resulting in a common gender noun. With this respect, the following question arises: why does -in form feminine nouns with inanimates, but common gender nouns with animates? In §4.4 I will show that expressive suffixes are not specified for 84 gender; instead, they are specified for inflectional class, and gender is predictable on the basis of inflectional class combined with animacy and natural gender.  (63)                  n2[fem]              2         n2[fem]         n1[masc]         -in          2                         n1[masc]       √ovrag-                                         (inanimate)  (64)  a.  ovrág                                          b.  ovráž-in-a             ditch.N.SG (MASC)                                   ditch-EXPR-N.SG (FEM)             ‘ditch’                                            ‘ditch (vulg)’  (65)                      n2[fem]                 2           n2[fem]          n1[fem]            -in          2                                n1[fem]     √jam-                                           (inanimate)  (66)  a.  jám-a                                          b.  jám’-in-a             pit-N.SG (FEM)                                          pit-EXPR-N.SG (FEM)             ‘pit’                                                ‘pit (vulg)’  (67)                 n2[fem]                        2         n2[fem]          n1[neut]                -in          2                             n1[neut]      √bolot-                                           (inanimate)  (68)  a.  bolót-o                                        b.  bolót’-in-a             swamp-N.SG (NEUT)                                swamp-EXPR-N.SG (FEM)             ‘swamp’                                          ‘swamp (vulg)’  Most attitude suffixes (except -an) display exactly the same pattern: they form feminine nouns irrespective of the gender of the base, as illustrated in (69)-(72).  (69)  a.  sm’éx                                          b.  sm’ex-ot-á             laughter.N.SG (MASC)                            laughter-EXPR-N.SG (FEM)            ‘laughter’                                         ‘laughter (vulg)’  85 (70)  a.  skúk-a                                         b.  skuk-ot-á             boredom-N.SG (FEM)                               boredom-EXPR-N.SG (FEM)             ‘boredom’                                        ‘boredom (vulg)’  (71)  a.  stíd                                              b.  stid-úx-a              shame.N.SG (MASC)                                 shame-EXPR-N.SG (FEM)              ‘shame’                                           ‘shame (vulg)’  (72)  a.  kómnat-a                                     b.  komnat-úx-a             room-N.SG (FEM)                                      room-EXPR-N.SG (FEM)             ‘room’                                              ‘room (vulg)’  The only suffix that does not share this behaviour is the suffix -an, which always forms nouns of masculine gender. To illustrate how -an forms masculine nouns, let us look at the data below. In (73), the noun gub-á ‘lip’ is feminine. When the attitude suffix -an attaches, the resulting noun gub-án becomes masculine. In (74), the noun púz-o ‘belly’ is neuter. When -an attaches, the resulting noun puz-án also becomes masculine. In (75), the noun lób ‘forehead’ is masculine. When the suffix -an attaches, the resulting noun lob-án remains masculine. The reason why -an differs from other attitude suffixes is discussed in §4.4.1.2.  (73)  a.  gub-á                                          b.  gub-án             lip-N.SG (FEM)                                          lip-EXPR.N.SG (MASC)             ‘lip’                                                 ‘animate with distinct lips (vulg)’  (74)  a.  púz-o                                          b.  puz-án             belly-N.SG (NEUT)                                   belly-EXPR.N.SG (MASC)             ‘belly’                                             ‘animate with distinct belly (vulg)’  (75)  a.  lób                                              b.  lob-án             forehead.N.SG (MASC)                           forehead-EXPR.N.SG (MASC)             ‘forehead’                                       ‘animate with distinct forehead (vulg)’  The proposed structures for (73)–(75) are shown in (76)–(78), respectively.  (76)                            n2[masc]                            2                  n2[masc]      n1[fem]                  -an        2                                    n1[fem]      √gub-                                                (inanimate)   86 (77)                n2[masc]                  2             n2[masc]       n1[neut]             -an         2                          n1[neut]     √puz-                                     (inanimate)   (78)                 n2[masc]                    2                n2[masc]      n1[masc]                -an        2                           n1[masc]    √lob-                                       (inanimate)  To conclude, attitude suffixes change the grammatical gender of inanimate nouns. Most attitude suffixes (except -an) form feminine nouns, regardless of the gender of the input. The attitude suffix -an forms masculine nouns, regardless of the gender of the input (Table 4.12).   EXPRattitude Input Output inanimate, [MASCULINE] inanimate, [FEMININE] inanimate, [NEUTER] inanimate, [FEMININE] -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’, -úš, -ág, -ák, -ál, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx inanimate, [FEMININE] inanimate, [FEMININE] inanimate, [MASCULINE] inanimate, [MASCULINE] inanimate, [NEUTER] inanimate,[MASCULINE] -án inanimate, [FEMININE] inanimate,[MASCULINE] Table 4.12: Attitude suffixes (inanimate nouns): change is indicted in bold.  4.3.1.3. Summary Affixing an attitude suffix to a given base results in a change in grammatical gender. This change partly depends on animacy and natural gender of the √Root. The change is seen in animate non-sex-differentiable nouns (TYPE II), which become common gender nouns (TYPE III). The change is also seen in inanimate nouns, which become either feminine (with the majority of attitude suffixes), or masculine (with the attitude suffix -an) (Table 4.13). In order to see this systematicity in the behaviour of attitude suffixes, the behaviour of gender 87 with different noun types should be analyzed. To the best of my knowledge, this kind of analysis has never been done before.  EXPRattitude Input  Output animate (non-sex-differentiable), any input animate (non-sex-differentiable), unspecified -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’, -úš, -ág, -ák, -ál, -ár, -áx, -íl, -in, -ób, -ot,  -óx, -úg, -úk, -úx inanimate, any input inanimate, [FEMININE] -án inanimate, any input inanimate, [MASCULINE] Table 4.13: Attitude suffixes (change in grammatical gender): change is indicted in bold.  According to Diagnostic II (change in grammatical gender), attitude suffixes behave like syntactic heads because they produce a change in grammatical gender (79).  (79)    HEAD             n2[gender Y]                     2                n2[gender Y]     n1[gender X]                  EXPRattitude   2                        n1[gender X]    √Root   Based on the findings above, the following question arises: Why do we get nouns with different grammatical genders when attitude suffixes attach? If attitude suffixes were themselves associated with the category “gender”, we would expect that all nouns used with these attitude suffixes should have the same grammatical gender. On the other hand, if attitude suffixes are not associated with the category “gender”, what determines a change in gender? In §4.4, I shed some light on these questions.  4.3.2. Size suffixes Unlike attitude suffixes that can change the grammatical gender of a noun, size suffixes do not produce a change in gender. According to Diagnostic II, this means that size suffixes behave like syntactic modifiers (80).  88 (80) MODIFIER           n[gender X]                     2           EXPRsize            n[gender X]                                       2                       n[gender X]           √Root  Size suffixes can attach to both animate and inanimate bases of all grammatical genders with the same result: no change in gender. Let us first look at TYPE I nouns (animate, sex- differentiable). As I discussed above, these nouns contain natural gender (male or female) as part of their semantics. When size suffixes attach to TYPE I nouns, there is no change in grammatical gender. For example, in (81), the noun brát ‘brother’ is sex-differentiable because natural gender “male” is part of its meaning. As males are always masculine, this noun is assigned masculine grammatical gender. When the size suffixes -ik and -ec are added to this noun, there is no change in gender. The resulting nouns brat’-ik ‘brother (dim)’ and brát’-ec ‘brother (dim)’ are still masculine.  (81)                   n[masc]                2      -ik/-ec             n[masc]                               2                             n[masc]     √brat-                                      (animate) (male)  (82)  a.  brát                                               b.  brát’-ik             brother.N.SG (MASC)                                brother-EXPR.N.SG (MASC)             ‘brother’                                            ‘brother (dim)’           c.  brát’-ec            brother-EXPR.N.SG (MASC)             ‘brother (dim)’  Another example of a sex-differentiable noun is illustrated in (83). The noun s’estr-á ‘sister’ contains the natural gender “female” as part of its meaning. As females are always feminine, the noun s’estr-á ‘sister’ is assigned feminine grammatical gender. In (84b), the size suffix -ic is added and the gender remains feminine.   89 (83)                 n[fem]              2               -ic             n[fem]                      2                          n[fem]       √s’estr-                                   (animate) (female)  (84)  a.  s’estr-á                                          b.  s’estr’-íc-a             sister-N.SG (FEM)                                       sister-EXPR-N.SG (FEM)             ‘sister’                                               ‘sister (dim)’  Let us now look at TYPE II nouns (animate, non-sex-differentiable). Nouns of this type do not contain natural gender as part of their meaning. I show that when size suffixes attach to TYPE II nouns, there is no change in grammatical gender. For example, in (85), the noun zv’ér’ ‘animal’ is non-sex-differentiable because it can denote both male and female animals and it is associated with the masculine grammatical gender. When the size suffixes -ok and - išč’ merge with this noun, there is no change in gender. The resulting nouns zv’er’-ók ‘animal (dim)’ and zv’er’-íšč’-e ‘animal (aug)’ are still masculine (86a-c).  (85)                 n[masc]                    2           -ok/-išč’         n[masc]                      2                          n[masc]     √zv’er’-                                        (animate)  (86)  a.  zv’ér’                                           b.  zv’er’-ók             animal.N.SG (MASC)                                 animal-EXPR.N.SG (MASC)             ‘animal’                                            ‘animal (dim)’          c.  zv’er’-íšč’-e             animal-EXPR-N.SG (MASC)             ‘animal (aug)’  Another example of a non-sex-differentiable noun is given in (87) with the feminine noun ríb-a ‘fish’. When the size suffixes -k, -ic and -išč’ merge with this noun, there is no change in gender. The resulting nouns remain feminine (88a-d).     90 (87)            n[fem]               2        -k/-ic/-išč’   n[fem]                      2                    n[fem]       √rib-                                 (animate)  (88)  a.  ríb-a                                            b.  ríb-k-a             fish-N.SG (FEM)                                         fish-EXPR-N.SG (FEM)             ‘fish’                                               ‘fish (dim)’          c.  ríb’-ic-a                                         d.  ríb’-išč’-a             fish-EXPR-N.SG (FEM)                             fish-EXPR-N.SG (FEM)            ‘fish (dim)’                                      ‘fish (dim)’  Next I turn to TYPE III (animate, common gender) nouns. I show that size suffixes do not produce a change in the grammatical gender of these nouns either. For example, in (89), the noun s’irot-á ‘orphan’ is of common gender: it can trigger either masculine or feminine agreement. When the size suffix -k merges with this noun, there is no change in gender. The resulting noun is still of common gender (90)–(91).  (89)  a.  s’irot-á                                                 b.  s’irót-k-a             orphan-N.SG (MASC/FEM)                                  orphan-EXPR-N.SG (MASC/FEM)             ‘orphan’                                                     ‘orphan (dim)’  (90)  a.  b’éd-n-ij                       s’irot-á             poor-ADJ-MASC.N.SG    orphan-N.SG (MASC)             ‘poor orphan’   b.  b’éd-n-ája  s’irot-á   poor-ADJ-FEM.N.SG  orphan-N.SG (FEM)    ‘poor orphan’  (91)  a.  b’éd-n-ij                  s’irót-k-a                  poor-ADJ-MASC.N.SG  orphan-EXPR-N.SG(MASC)             ‘poor orphan’      b.  b’éd-n-ája                  s’irót-k-a     poor-ADJ-FEM.N.SG   orphan-EXPR-N.SG(FEM)    ‘poor orphan’  The proposed structure for (89) is shown in (92).  91 (92)           n                               ← unspecified for grammatical gender             2     -k            n                        ← unspecified for grammatical gender               2                    n            √s’irot-                                  (animate)  So far I have analyzed different types of animate nouns. I illustrated that there is no change in grammatical gender of animate nouns when size suffixes are added. Below I deal with inanimate nouns and show that size suffixes produce no change in grammatical gender of these nouns either. To illustrate this, let us look at inanimate nouns of different grammatical genders (masculine, feminine, and neuter). For example, in (93), the noun l’és ‘forest’ is masculine. When the size suffixes -ok and -išč’ are added, there is no change in grammatical gender. The resulting nouns are all masculine (94a-c).  (93)                 n[masc]                    2              -ok/-išč’      n [masc]                      2                          n[masc]      √l’es-                                      (inanimate)  (94)   a.  l’és                                            b.  l’es-ók              forest.N.SG (MASC)                               forest-EXPR.N.SG (MASC)             ‘forest’                                          ‘forest (dim)’           c.  l’es-íšč’-e              forest-EXPR-N.SG (MASC)             ‘forest (aug)’  In (95), the noun róšč’-a ‘grove’ is feminine. When the size suffix -ic is added, there is no change in grammatical gender. The resulting noun remains feminine (96a-b).  (95)           n[fem]              2         -ic             n[fem]                      2                   n[fem]        √rošč’-                                 (inanimate)  92 (96)  a.  róšč’-a                                         b. róšč’-ic-a             grove-N.SG (FEM)                                     grove-EXPR-N.SG (FEM)             ‘grove’                                             ‘grove (dim)’   In (98), the noun bolót-o ‘swamp’ is neuter. When the size suffixes -c and -išč’ are added, there is no change in grammatical gender. The resulting nouns are all neuter (99a-c).  (98)           n[neut]              2     -c/-išč’          n[neut]                2                    n[neut]     √bolot-                                (inanimate)   (99)  a.  bolót-o                                        b.  bolót’-c-e              swamp-N.SG (NEUT)                                swamp-EXPR-N.SG (NEUT)              ‘swamp’                                           ‘swamp (dim)’           c.  bolót’-išč’-e             swamp-EXPR-N.SG (NEUT)             ‘swamp (aug)’  To summarize, size suffixes produce no change in grammatical gender of both animate and inanimate nouns (Table 4.14). This property follows if we assume that size suffixes are syntactic modifiers.  EXPRsize Input  Output animate, male, [MASCULINE] animate, male, [MASCULINE] animate, female, [FEMININE] animate, female,[FEMININE] animate, [MASCULINE] animate, [MASCULINE] animate, [FEMININE] animate, [FEMININE] inanimate, [MASCULINE] inanimate, [MASCULINE] inanimate, [FEMININE] inanimate, [FEMININE] -k/-ek/-ok/-ik, -c/-ec/-ic, -išč’ inanimate, [NEUTER] inanimate, [NEUTER] Table 4.14: Size suffixes (no change in grammatical gender) 93 4.4. CHANGE IN INFLECTIONAL CLASS The third property that is predicted by this analysis is that attitude suffixes should be able to change inflectional class, while size suffixes should not (Diagnostic III). Here I show that this is indeed the case. Attitude suffixes can change inflectional class, thus they behave like syntactic heads (100a). In contrast, size suffixes cannot change inflectional class, thus they behave like syntactic modifiers (100b).  (100) a. HEAD         n2[class Y]                  b. MODIFIER                n[class X]                          2                                                 2                n2[class Y]     n1[class X]                               EXPRsize       n[class X]             EXPRattitude  2                                                   2                                 n1[class X]        √Root                                     n[class X]        √Root   In §4.4.1, I analyze attitude suffixes; in §4.4.2, I analyze size suffixes; and in §4.4.3, I present the conclusions.  4.4.1. Attitude suffixes The majority of attitude suffixes (except -an) form nouns of inflectional Class II, regardless of the inflectional class of the input. The attitude suffix -an forms nouns of inflectional Class I, regardless of inflectional class of the input. In view of this, the following question arises: What is the proper representation of attitude suffixes, that is, what is their lexical entry? Here I propose that the majority of attitude suffixes are specified for inflectional Class II in their lexical entry (101a), while -an is specified for inflectional Class I (101b).  (101)          a.              n2[class II]                          b.                      n2[class I]                     2                                              2                     n2[class II]       n1[class X]                               n2[class I]       n1[class X]                  EXPRattitude     2                                 EXPR-an   2               (except -an)  n1[class X]        √Root                                 n1[class X]        √Root  I also show that attitude suffixes are not specified for grammatical gender, but instead that grammatical gender can be predicted from the inflectional class of the attitude suffix. In other words, the change in grammatical gender discussed above (§4.3) falls out directly from the change in inflectional class. In §4.4.1.1, I illustrate how attitude suffixes change inflectional 94 class; and in §4.4.1.2, I show that a change in inflectional class is correlated with a change in grammatical gender.  4.4.1.1. Change in inflectional class Recall the distribution of inflectional classes in Russian. I only review the distribution in Nom sg., because all examples cited below are in this case (see Chapter 3).   Class I Class II Class III  Singular masculine (end in -Ø in Nom sg.) neuter (end in -o/e in Nom sg.) masculine/feminine/ common (end in -a in Nomsg.) feminine (end in –Ø in Nom sg.) Nominative zakón ‘law’  v’in-ó‘wine’ škól-a ‘school’ kóst’ ‘bone’   Table 4.15: Inflectional endings in Nom sg.  Class I and Class II both contain masculine nouns, but they differ in their inflectional endings. Class I masculine nouns have -Ø ending in Nom sg., while Class II masculine nouns have -a ending. Class II and Class III both contain feminine nouns, but they also differ in inflectional endings. Class II feminine nouns have -a ending, while Class III nouns have -Ø ending. Neuter nouns belong to Class I and have -o/-e endings. Common gender nouns only belong to Class II and therefore have -a ending.  I start with a discussion of the attitude suffixes. Most attitude suffixes (except -an) can attach to nouns of any inflectional class and always return a Class II noun. For example, in (102), st’íd ‘shame’ is a masculine noun that belongs to Class I, which is evident from the -Ø ending in Nom sg. When the attitude suffix -ob is added, the inflectional class changes. The newly formed noun is now in Class II (-a ending in Nom sg.). The noun also changes its gender to feminine (103a-b).  95 (102)                       n2[class II]                  2                n[class II]    n1[class I]                     -ob     2                              n1[class I]    √stid-  (103)  a.  st’id                                             b.  stid-ób-a                shame-N.SG (MASC; CLASS I)              shame-EXPR-N.SG (FEM; CLASS II)               ‘shame’                                           ‘shame (vulg)’  If the base noun is a masculine noun of Class II, such as páp-a ‘dad’ in (104), the actual change in class is vacuous––we cannot see it, as a Class II noun ‘changes’ into a Class II noun. The resulting noun remains masculine (105a-b).  (104)                        n2[class II]                   2                n[class II]      n1[class II]                        -us’     2                                n1[class II]   √pap-  (105)  a.  páp-a                                            b.  pap-ús’-a               dad-N.SG (MASC; CLASS II)                 dad-EXPR-N.SG (MASC; CLASS II)               ‘dad’                                               ‘dad (affect)’  Finally, if the base noun belongs to Class III, like the feminine noun tvár’ ‘animal’ in (106), suffixation of the attitude suffix -uk will also result in a change in inflectional class. The newly formed noun belongs to Class II (and thus has -a ending in Nom sg.). And the resulting noun also changes its gender to common gender (107a-b).  (106)                       n2[class II]                  2              n2[class II]     n1[class III]                   -uk         2                               n1[class III]   √tvar’-  (107)  a.  tvár’                                            b.  tvar’-úk-a               animal.N.SG (FEM; CLASS III)           animal-EXPR-N.SG (MASC/FEM; CLASS II)               ‘animal’                                          ‘animal (vulg)’  In sum, no matter what the inflectional class of the input (Class I, Class II, or Class III), these attitude suffixes always form nouns of Class II. 96  We now turn to the attitude suffix -an. In contrast to the majority of attitude suffixes, it consistently forms nouns of Class I. For example, the Class II noun gub-á ‘lip’ in (108) changes its inflectional class when the attitude suffix -an is added. The resulting noun belongs to Class I (-Ø ending in Nom sg.) (109a-b).  (108)                       n2[class I]                  2               n2[class I]     n1[class II]                    -an       2                              n1[class II]     √gub-  (109)  a.  gub-á                                             b.  gub-án               lip-N.SG (FEM; CLASS II)                           lip-EXPR.N.SG (MASC; CLASS I)               ‘lip’                                                     ‘animate with distinct lips (vulg)’  In (110), the Class I noun brát ‘brother’ (-Ø ending in Nom sg.) predictably does not change its inflectional class when the attitude suffix -an is added. The resulting noun is still in Class I (111a-b).  (110)                      n2[class I]                  2               n2[class I]     n1[class I]                    -an        2                           n1[class I]      √brat-  (111)  a.  brát                                                b.  brat-án               brother N.SG (MASC; CLASS I)               brother-EXPR.N.SG (MASC; CLASS I)               ‘brother’                                            ‘brother (vulg)’  Thus, no matter what the inflectional class of the input (Class I or Class II), the suffix -an always forms nouns of Class I. To the best of my knowledge, there are no examples where the suffix -an would attach to Class III bases.  To summarize, the majority of attitude suffixes (except -an) form nouns of Class II, regardless of the inflectional class of the input. The attitude suffix -an forms nouns of Class I, regardless of the inflectional class of the input (Table 4.16).  97  EXPRattitude Input  Output Class I Class II Class II Class II  -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’,  -úš, -ág, -ák, -ál, -ár, -áx, -íl, -in, -ób,  -ot, -óx, -úg, -úk, -úx Class III Class II Class I Class I Class II Class I -án Class III N/A Table 4.16: Attitude suffixes (change in inflectional class): change is indicated in bold.  This behaviour follows under the assumption that attitude suffixes are syntactic heads and are themselves associated with an inflectional class. The class membership of individual attitude suffixes is illustrated in Table 4.17.  EXPRattitude Lexical entries  -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’,  -úš, -ág, -ák, -ál, -ár, -áx, -íl, -in, -ób,  -ot, -óx, -úg, -úk, -úx Class II  -án Class I  Table 4.17: Class membership of attitude suffixes  4.4.1.2. Correlation between change in class and change in gender Here I address the question of how grammatical gender is changed in Russian. In §4.3.1, I showed that there is variation in grammatical gender of nouns used with attitude suffixes. Here I argue that this variation is determined by the inflectional class assigned by an attitude suffix. In other words, attitude suffixes are specified for inflectional class in their lexical entries and grammatical gender falls out directly from this inflectional class.  Recall the variation in grammatical gender we have observed above. When attitude suffixes are added to TYPE I nouns (animate, sex-differentiable), the resulting nouns are masculine or feminine, depending on the natural gender of the nouns they attach to (112)–(113).  (112)  a.  brát                                                   b.  brat-úx-a               brother.N.SG (MASC; CLASS I)                    brother-EXPR-N.SG (MASC; CLASS II)                  ‘brother’                                                   ‘brother (vulg)’  98 (113)  a.  s’estr-á                                              b.  s’estr-úx-a               sister.N.SG (FEM; CLASS II)                        sister-EXPR-N.SG (FEM; CLASS II)                  ‘sister’                                                      ‘sister (vulg)’  In contrast, when attitude suffixes are added to inanimate nouns, the resulting nouns are always feminine (114).  (114)  a.  gólod                                             b.  golod-úx-a              hunger.N.SG (MASC; CLASS I)                 hunger-EXPR-N.SG (FEM; CLASS II)                  ‘hunger’                                               ‘hunger (vulg)’  When attitude suffixes are added to TYPE II nouns (animate, non-sex-differentiable), the resulting nouns become unspecified for gender (common gender nouns) (115).  (115)  a.  zv’ér’                                             b. zv’er’-úx-a               animal.N.SG (MASC; CLASS I)                 animal-EXPR-N.SG (MASC/FEM; CLASS II)              ‘animal’                                               ‘animal (vulg)’  When attitude suffixes are added to TYPE III nouns (animate, common gender), the resulting nouns are still in common gender (116).  (116)  a.  kóp-a                                              b.  kop-úx-a               dig-N.SG (MASC/FEM; CLASS II)                 dig-EXPR-N.SG (MASC/FEM; CLASS II)                  ‘slow animate’                                           ‘slow animate (vulg)’  As the data (112) and (113) show, the same attitude suffix -ux can produce nouns of different grammatical genders: masculine, feminine, or common gender (unspecified for gender). This means that the attitude suffix cannot be specified for grammatical gender in its lexical entry. If it were, then we would expect all nouns suffixed with this attitude suffix (except for nouns that have natural gender) to be of the same gender. If it were specified for masculine gender, the data (114b) would be unexpected: that is, we would not expect the suffixation of -ux to derive a feminine noun from a masculine noun. Similarly, if -ux were specified for feminine gender, then common gender nouns, like (115b) or (116b) would be feminine in cases where the natural gender of the referent is unknown. However, in such cases, these nouns are still unspecified for gender and thus, can trigger either feminine or masculine agreement (Doleschal & Schmid 2001).  99 I conclude that attitude suffixes are not specified for grammatical gender at all. They are, however, specified for inflectional class (§4.4). Observe that although the nouns in (112b)– (116b) used with the suffix -ux have different grammatical genders, they are all in Class II (have -a ending in Nom sg.).  I propose that the variation in grammatical genders falls out directly from the inflectional class of attitude suffixes. In what follows, I show how this arises. I start with a discussion of the attitude suffixes that assign inflectional Class II, and then I analyze the attitude suffix -an which assigns inflectional Class I.  Russian Class II nouns are either animate or inanimate (117). Animate nouns are either sex- differentiable or non-sex-differentiable. Both sex-differentiable and non-sex-differentiable nouns can be either masculine or feminine, the difference being that the grammatical gender of sex-differentiable nouns is determined by their natural gender “male” or “female”. Inanimate nouns of this class are all feminine.  (117)                                                          Class II                                Animate                                                               Inanimate                           3        Sex-differentiable              Non-sex-differentiable           3    Male                    Female      g                            g [MASCULINE]      [FEMININE]  [FEMININE]           (unspecified)     [FEMININE]    d’ád’-a               t’ót’-a      p’ersón-a                 s’irot-á            kn’íg-a     ‘uncle’                ‘aunt’         ‘person’                 ‘orphan’            ‘book’  It follows from my proposal that an animate noun used with an attitude suffix that derives a Class II noun can be either masculine or feminine, while inanimate nouns can only be feminine. This is indeed what we find. For example, in (112) above, brát ‘brother’ belongs to Class I (-Ø ending in Nom sg.). When the attitude suffix -ux is added, this inflectional class changes to Class II (-a ending in Nom sg.). Since brát has the natural gender ‘male’, the resulting noun brat-úx-a ‘brother (vulg)’ is masculine. Thus, knowing the animacy, natural gender, and inflectional class of the noun, it is possible to predict its grammatical gender. 100 That is, if a noun is animate, male, and belongs to Class II, its grammatical gender is always masculine (118).  (118)  a. (male) and [class II] → [masculine]              b.                n2[class II]               ← predicted grammatical gender [masculine]                  2                n2[class II]    n1[class I]                      -ux       2                            n1[class I]     √brat-                                     (animate) (male)  Similarly, in (113) above, s’estr-á ‘sister’ belongs to Class II (-a ending in Nom sg.). When the attitude suffix -ux is added, there is no change in class. The resulting noun s’estr-úx-a ‘sister (vulg)’ is still in Class II. Since s’estr-a has the natural gender ‘female’, the resulting noun s’estr-úx-a ‘sister (vulg)’ is feminine. Here again, knowing the animacy, natural gender, and inflectional class of a noun, it is possible to predict its grammatical gender. So, if a noun is animate, female, and belongs to Class II, its grammatical gender is always feminine (119).  (119) a. (female) and [class II] → [feminine]              b.                 n2[class II]                ← predicted grammatical gender [feminine]                  2                n2[class II]     n1[class II]                      -ux        2                             n1[class II]   √s’estr-                                      (animate) (female)  In (114) above, the same attitude suffix -ux is added to the inanimate noun gólod ‘hunger’. The noun gólod ‘hunger’ belongs to Class I (-Ø ending in Nom sg.). When the attitude suffix -ux is added, this inflectional class changes. The resulting noun golod-úx-a ‘hunger (vulg)’ is now in Class II (-a ending in Nom sg.). Since inanimate Class II nouns are all feminine in Russian, here again it is possible to predict grammatical gender from inflectional class. Thus, if a noun is inanimate and belongs to Class II, its grammatical gender is always feminine (120).  101  (120) a. (inanimate) and [class II] → [feminine]              b.                 n2[class II]                ← predicted grammatical gender [feminine]                   2                n2[class II]    n1[class I]                     -ux        2                            n1[class I]    √golod-                                         (inanimate)  Finally, in (115) above, zv’ér’ ‘animal’ belongs to Class I (-Ø ending in Nom sg.). When the attitude suffix -ug is added, there is a change in inflectional class. The resulting noun zv’er’- úg-a ‘animal (vulg)’ is now in Class II (-a ending in Nom sg.). Since zv’er’ is animate but non-sex-differentiable, its grammatical gender cannot be determined by its natural gender. As a result, when it becomes a Class II noun, its grammatical gender is unspecified, which accounts for its status as a common gender noun. Thus, if a noun is animate, non-sex- differentiable, and belongs to Class II, its grammatical gender is unspecified and can be either masculine or feminine (121).  (121) a. (animate) and [class II] → unspecified gender              b.                n2[class II]                ← unspecified grammatical gender (common gender)                  2                n2[class II]   n1[class I]                     -ug       2                          n1[class I]       √zv’er’-                                         (animate)  Next we turn to the remaining attitude suffix -an which assigns inflectional Class I. Russian Class I nouns can be animate or inanimate. Animate nouns are masculine, while inanimate nouns are either masculine or neuter (i.e., they can have either masculine or neuter grammatical agreement) (122).  (122)                                             Class I                                                  3                                        Animate                   Inanimate                                        g                        3                                [MASCULINE]   [MASCULINE]       [NEUTER]                                        brát                l’és                  pól’-e                                 ‘brother’         ‘forest’               ‘field’ 102  The attitude suffix -an attaches to both animate and inanimate nouns of different classes. As a result, it produces a change in both animacy and the class of a noun. The resulting words are always animate Class I nouns. For example, in (123), gub-á ‘lip’ is an inanimate Class II noun (-a ending in Nom sg.). When the attitude suffix -an is added, the resulting noun gub- án ‘animate with distinct lips’ becomes animate and changes its class to Class I (-Ø ending in Nom sg.). In (124), púz-o ‘belly’ is an inanimate Class I noun (-Ø ending in Nom sg.). When the attitude suffix -an is added, the resulting noun puz-án ‘animate with distinct belly’ becomes animate, but it remains in Class I.  (123)  a.  gub-á                                             b.  gub-án               lip-N.SG (FEM; CLASS II)                         lip-EXPR.N.SG (MASC; CLASS I)               ‘lip’                                                   ‘animate with distinct lips (vulg)’  (124)  a.  púz-o                                            b.  puz-án               belly-N.SG (NEUT; CLASS I)                     belly-EXPR.N.SG (MASC; CLASS I)               ‘belly’                                                ‘animate with a distinct belly (vulg)’  Because the suffix -an always produces a Class I animate noun, regardless of the class or animacy of the base, I propose that it is specified for both animacy and Class I (125).  (125)                                n2[class I]                        2                     n2[class I]           n1[class I]/[class II]                      -an                  2                 (animate)   n1[class I]/[class II]   √Root   Now I show how it is possible to predict grammatical gender of nouns used with the suffix -an. In (123), gub-á ‘lip’ is an inanimate feminine noun that belongs to Class II. The suffix -an turns this inanimate noun into animate and changes Class II to Class I. This combination of animacy and Class I automatically changes feminine gender to masculine, because animate nouns of Class I are all masculine (126).   103 (126) a. (animate) and [class I] → [masculine]             b.                     n2[class I]                 ← predicted grammatical gender [masculine]                    2                n2[class I]       n1[class II]                     -an          2                  (animate)  n1[class II]  √gub-                                          (inanimate)  In (124) above, púz-o ‘belly’ is an inanimate neuter noun that belongs to Class I. The suffix -an turns this inanimate noun into an animate noun, but it remains in Class I. Here, too, the combination of animacy and Class I automatically changes neuter gender to masculine, because animate nouns are all masculine in Class I (127).  (127)                         n2[class I]                ← predicted grammatical gender [masculine]                    2               n2[class I]        n1[class I]                     -an          2                (animate)  n1[class I]   √puz-                                          (inanimate)  To summarize, I have argued that attitude suffixes are specified for inflectional class, which is consistent with their status as a syntactic head. The majority of attitude suffixes (except -an) are specified for Class II (128a). The suffix -an is specified for both animacy and Class I (128b).  (128)          a.              n2[class II]                          b.                          n2[class I]                     2                                                 2                     n2[class II]       n1[class X]                               n2[class I]           n1[class X]                  EXPRattitude     2                                 EXPR-an        2                                   n1[class X]        √Root                    (animate)   n1[class X]        √Root   Knowing animacy, natural gender, and inflectional class of a derived expressive noun, it is possible to predict its grammatical gender, which accounts for the variation in grammatical genders observed in §4.3. In other words, under the current analysis, grammatical gender is not specified in the lexical entry of attitude suffixes, but instead, it falls out directly from the inflectional class of the suffix. 104  4.4.2. Size suffixes Contrary to the attitude suffixes that can change inflectional class, size suffixes cannot change inflectional class (in nouns of Class I and Class II). According to Diagnostic III, this is consistent with their behaviour as syntactic modifiers (129). Class III nouns demonstrate a different behaviour that will be discussed later in this section.  (129) MODIFIERS      n[class X]                     2              EXPRsize       n[class X]                                   2                                   n[class X]       √Root  To illustrate this point, consider the following data. In (130), the noun č’elov’ék ‘person’ belongs to Class I (-Ø ending in Nom sg.). When the size suffixes -ek and -išč’ are added, there is no change in class. The resulting nouns are still in Class I (131b-c).  (130)                      n2[class I]                  2          -ek/-išč’         n1[class I]                                 2                             n1[class I]   √č’elov’ek-  (131)  a.  č’elov’ék                                       b.  č’elov’éč’-ek               person.N.SG (MASC; CLASS I)                person-EXPR.N.SG (MASC; CLASS I)               ‘person’                                             ‘person (dim)’            c.  č’elov’éč’-išč’-e12               person-EXPR.N.SG (MASC; CLASS I)              ‘person (aug)’  Similarly, in (132), the noun xl’éb ‘bread’ is in Class I (-Ø ending in Nom sg.). When the suffix -ec is added, there is also no change in class. The resulting noun remains in Class I (133b).   12  Recall that the neuter-like ending -e in such nouns does not mean that the nouns are neuter (Chapter 3). č’elov’éč’-išč’-e ‘big person’ is a masculine noun because it patterns with the masculine version of the Class I pattern, rather than with the neuter version of the Class I pattern. 105 (132)                         n2[class I]                   2                -ec            n1[class I]                                  2                                 n1[class I]    √xl’eb-  (133)  a.  xl’éb                                            b.  xl’éb’-ec               bread.N.SG (MASC; CLASS I)                  bread-EXPR.N.SG (MASC; CLASS I)               ‘bread’                                               ‘bread (dim)’  The same pattern is observed with Class II nouns. For example, in (134), the noun ruk-á ‘hand’ belongs to Class II (-a ending in Nom sg.). When the size suffixes -k and -išč’ are added, there is no change in class. The resulting nouns are still in Class II (135b-c).  (134)                       n2[class II]                  2            -k/-išč’         n1[class II]                                2                              n1[class II]    √ruk-  (135)  a.  ruk-á                                              b.  rúč’-k-a               hand-N.SG (FEM; CLASS II)                     hand-EXPR-N.SG (FEM; CLASS II)               ‘hand’                                                ‘hand (dim)’            c.  ruč’-íšč’-a              hand-EXPR-N.SG (FEM; CLASS II)              ‘hand (aug)’  In (136), the noun lúž-a ‘puddle’ is in Class II (-a ending in Nom sg.). When the size suffix -ic is added, there is also no change in class. The resulting noun lúž-ic-a ‘puddle (dim)’ remains in Class II (137b).  (136)            a.            n2[class II]                    2                 -ic             n1[class II]                                 2                               n1[class II]    √luž-  (137)  a.  lúž-a                                            b.  lúž-ic-a               puddle-N.SG (FEM; CLASS II)                 puddle-EXPR-N.SG (FEM; CLASS II)               ‘puddle’                                            ‘puddle (dim)’  106 To summarize, size suffixes do not produce a change in Class I or Class II nouns, which is consistent with their behaviour as syntactic modifiers (Table 4.18).   EXPRsize Input  Output Class I Class I -k/-ek/-ok/-ik, -c/-ec/-ic; -išč’ Class II Class II Table 4.18: Size suffixes (no change in inflectional class)  Class III nouns, however, show a different behaviour. When size suffixes merge with these nouns, their inflectional class changes to Class II. For example, in (138), nóč’ ‘night’ is a feminine Class III noun (-Ø ending in Nom sg.). When the size suffix -k attaches, the newly formed noun is in Class II (-a ending in Nom sg.). In (139), kr’ép-ost’ ‘fortress’ is in Class III. When the size suffix -c attaches, the newly formed noun belongs to Class II.  (138)  a.  nóč’                                             b.  nóč’-k-a               night.N.SG (FEM; CLASS III)                   night-EXPR-N.SG (FEM; CLASS II)               ‘night’                                                ‘night (dim)’  (139)  a.  kr’ép-ost’                                       b.  kr’ep-ost-c-á                stong-NOM.N.SG (FEM; CLASS III)       stong-NOM-EXPR-N.SG (FEM; CLASS II)                   ‘fortress’                                            ‘fortress (dim)’  What accounts for this behaviour of Class III nouns? Does this behaviour mean that size suffixes are syntactic heads associated with inflectional class of their own, just as attitude suffixes are? If size suffixes were associated with inflectional class, they would always produce nouns of the same inflectional class, as attitude suffixes do. For example, the size suffix -k would always produce nouns of Class II and thus, it would be able to turn a Class I noun into a Class II noun. But as the data above show, this is not the case. As I showed above, Class I nouns remain in Class I when a size suffix attaches. I propose that the change in inflectional class from Class III to Class II has nothing to do with the syntactic properties of size suffixes. Instead, I argue that this change is determined by the phonological properties of Class III nouns.  107 As observed in Thelin (1975), there is a systematic correlation between the final consonants of a feminine stem and its inflectional class. A “stem” is traditionally understood as a √Root + derivational and/or modifying suffix, excluding an inflectional ending (140).  (140) Root + suffix + inflectional ending                 Stem  For example, in (141), the stem consists of the √Root kr’ep-, the derivational nominal suffix -ost’, and the modifying suffix -c. The stem does not include the inflectional nominative singular ending -a.  (141)    kr’ep-ost-c-á             stong-NOM-EXPR-N.SG (FEM)                ‘fortress (dim)’  Thelin notes that feminine stems can end in a “hard” (non-palatalized) or “soft” (palatalized) consonant (e.g., /n/ ~ /n’/, /t/ ~ /t’/). Most consonants can be hard or soft, but c, š, ž are only hard, while j, č’, šč’ are only soft. If the final consonant of the stem is c, j, or the hard member of a hard-soft pair, the noun belongs to Class II (e.g., pt’íc-a ‘bird’, all’éj-a ‘alley’, stran-á ‘country’). If the final consonant of the stem is š, ž, č’, šč’ or the soft member of a hard-soft pair, the inflectional class cannot be predicted. Below I list some contrasting examples from Thelin (cited in Corbett 1982:213). The final consonant of the stem is indicated in bold (Table 4.19).  Class II Class III p'ésn’-a   ‘song’  žízn'   ‘life’ grúš-a   ‘pear’ túš   ‘ink’ dáč’-a   ‘country house’ nóč'   ‘night’ Table 4.19: Contrasting examples (Class II and Class III nouns)  Thus, based on Thelin’s generalizations, the difference between Class II and Class III stems is that Class II stems can end in a hard or soft consonants, while Class III stems can only end in a soft consonants (including č’, šč’ that are always soft) or the hard consonants š, ž. The final consonants of Class III stems are summarized in Table 4.20.  108  Soft  Hard Final consonants of Class III stems  t', d’, n’, s’, z’ č’, šč’ š, ž Table 4.20: Final consonants of Class III stems  With respect to Table 4.20, the following question arises: what do soft consonants and hard consonants š, ž have in common? Under Clements & Hume’s (1995) version of feature geometry, front vowels/glides including the secondary palatalization aspect of palatalized consonants, and are represented as having a [coronal] place node containing the [–anterior] ([–ant]) feature, situated underneath their VPlace (vocalic place) node. According to this feature-geometric model, both palatalized consonants and the hard consonants š, ž share the [–ant] feature. This means that all Class III nouns in Russian contain [–ant] at the end of the stem. One way to account for this is to assume a floating [–ant] morpheme that marks Class III as such.13 For example, under this assumption, the stem of the Class III noun króv’ ‘blood,’ consists of the √Root krov- and the [–ant] morpheme (142).  (142) króv’           króv+[–ant]           ‘blood’  If [–ant] is a floating morpheme and not part of the √Root, we would expect to find the √Root krov- without palatalization. This is indeed what we find in Russian. For example, in the adjective krov-áv-ij ‘bloody,’ the √Root krov- ends in a hard consonant /v/. More examples illustrating that there exists a separate floating [–ant] morpheme that marks Class III nouns are given in (143)–(144) (compare a and b).  (143)  a.   vís’                             b.  vis-ot-á                height+[–ant]                    height-NOM-N.SG                ‘height’               ‘height’  (144)  a.   glúb’                           b.  glub-ók-ij                depth+[–ant]                     deep-ADJ-MASC.SG                ‘depth’               ‘deep’   13  Many thanks to Dr. Gunnar Ólafur Hansson for his help with the phonological aspect of this thesis. 109 The assumption that Class III stems end in the [–ant] morpheme is also supported by historical evidence. In the pre-history of Slavic, all Class III nouns ended in /i/, which caused historical palatalization of the preceding consonant. In the course of history, /i/ turned into a so-called jer vowel and eventually disappeared in this position (Hermans 2002; Rubach 1986; Yearley 1995; among others). In modern Russian, this suffixal vowel is no longer present, but we can see the traces of it in the [–ant] feature of Class III stems.  The representation for Class III nouns is given in (145). In this representation, Class III nouns have an internal structure consisting of a √Root and a floating [–ant] morpheme. This means that all Class III nouns are morphosyntactically derived.  (145)                            n[class III]                                 2                                n[class III]   √Root                                [–ant]  Let us now come back to the problem discussed above: size suffixes turn Class III into Class II nouns. As I suggested above, this is related to the phonological properties of Class III nouns. When the size suffixes -k (allomorphs: -ok/-ek/-ik) or -c (allomorphs: -ec/-ic) merge with Class III nouns, the stem no longer ends in [–ant], but instead it ends in a hard consonant of the suffix. For example, in (146), the stem is noč’-k; it ends in /k/, a consonant which is not [–ant]. In (147), the stem is kr’ep-ost-c; it ends in /c/, a consonant which is likewise not [–ant] (and is in fact [+ant]).  (146)  nóč’-k-a            night-EXPR-N.SG (FEM; CLASS II)            ‘night (dim)’  (147)  kr’ep-ost-c-á            stong-NOM-EXPR-N.SG (FEM; CLASS II)               ‘fortress (dim)’  Since the stems above do not end in [–ant] anymore, the newly formed nouns nóč’-k-a ‘night (dim)’ and kr’ep-ost-c-á ‘fortress (dim)’ cannot belong to Class III either. The only class in which they can belong now is Class II, because it is the only class besides Class III that 110 contains feminine nouns. Thus, by changing the final consonant of the stem, the inflectional class also changes. The augmentative suffix -išč’, which does end in a [–ant] consonant, almost never attaches to Class III nouns; for example, when it is added to the nouns nóč’ ‘night’ or kr’ép-ost’ ‘fortress,’ the resulting data are ungrammatical (148)–(149). Thus, there is no evidence here to suggest that there is a change in inflectional class.  (148) * noč’-išč’-(a)            night-EXPR-N.SG            ‘night (aug)’  (149) * kr’ep-ost-išč’-(a)            stong-NOM-EXPR-N.SG               ‘fortress (aug)’  To the best of my knowledge, there is only one word in which -išč’ attaches to a Class III noun: von’-íšč’-a ‘stench’. Here the inflectional class changes from Class III to Class II, which is unexpected under the current hypothesis. Since the suffix -išč’ ends in [–ant], we (wrongly) predict no change in class. On the other hand, it is unclear whether in this particular word, -išč’ is indeed a size suffix. The Contemporary Explanatory Dictionary of Russian (Efremova 2006) lists von’-íšč’-a ‘stench’ as being a vulgar noun, while vón’ ‘stench’ is not vulgar (150). The added meaning of vulgarity is not typical for the augmentative size suffix -išč’ (compare with (151), where -išč’ indicates a large size and has no vulgar meaning). Because of lack of data, it is hard to say whether the current hypothesis is incorrect because it cannot account for (150), or whether there is something special about the particular (exceptional) example. In any case, it remains unclear why the augmentative - išč’ cannot attach to Class III nouns, and why in the only case when it does attach to a Class III noun (150), it has a vulgar meaning.  (150)  a.  vón’                                             b. von’-íšč’-a               stench.N.SG (FEM; CLASS III)                stench-EXPR-N.SG (FEM; CLASS II)               ‘stench’                                              ‘stench (vulg)’  (151)  a.  vólk                                              b.  volč’-íšč’-e               wolf.N.SG (MASC; CLASS I)                     wolf-EXPR-N.SG (MASC; CLASS I)                ‘wolf’                                                 ‘wolf (aug)’ 111  To summarize, I have suggested that Class III stems end in a floating [–ant] morpheme that marks the nouns as Class III. This idea requires further research that goes beyond the scope of this thesis, but hopefully it contributes toward understanding why Class III nouns change their inflectional class for Class II when a size suffix attaches.  4.5. NON-EXPRESSIVE SUFFIXES HOMOPHONOUS WITH SIZE SUFFIXES In Russian, there are non-expressive suffixes that are homophonous with the expressive size suffixes we discussed above. I argue that size suffixes and their homophonous counterparts differ not only in meaning, but also in their syntactic structure. Unlike size suffixes that are syntactic modifiers, non-expressive homophones are syntactic heads, as illustrated in (152).  (152) a. MODIFIER             Y                       b. HEAD            X                               2                                      2                              X          Y                                    X            Y                              EXPRsize                                                           non-EXPR  In §4.5.1, I investigate the non-expressive suffix -išč’; in §4.5.2, I investigate the non- expressive suffix -ec; and in §4.5.3, I investigate the non-expressive suffix -k. Finally, in §4.5.4, I summarize the findings.  4.5.1. The suffix -išč’ The non-expressive suffix -išč’ means ‘place or site’. I treat this suffix as non-expressive, because it does not convey any information about the attitudes or emotions of the speaker, and therefore, it has no expressive content (see Chapter 2). For example, the sentence in (153) has the descriptive content ‘I saw a site of fire,’ but the expressive content is absent.  (153)  Ja    uv’íd’el    požár’-išč’-e            I      saw          fire-PLACE-N.SG    i. Descriptive: ‘I saw a site of fire’  Unlike the augmentative suffix -išč’ which is a syntactic modifier, the non-expressive suffix 112 -išč’ has the properties of a syntactic head. The first piece of evidence stems from the fact that it can change syntactic category. For example, in (154), the non-expressive -išč’ acts as a nominalizer: it attaches to a verb and forms a noun with the meaning ‘place to run away (shelter)’. More data are given in (155).  (154)  a.  u-b’ež-á-t’                                     b.  u-b’éž-išč’-e              VERB.PREF-run-TH-INF                              VERB.PREF-run-PLACE-N.SG (NEUT; CLASS I)                 ‘to run away’                                    ‘shelter’  (155)  a.  pr’i-b’ež-á-t’                                b.  pr’i-b’éž-išč’-e              VERB.PREF-run-TH-INF                             VERB.PREF-run-PLACE-N.SG (NEUT; CLASS I)                ‘come running’                                  ‘refuge’  The second piece of evidence comes from the fact that the non-expressive -išč’ can change the inflectional class and grammatical gender of a noun. For example, in (156), the non- expressive -išč’ changes the inflectional class from Class II to Class I and changes the grammatical gender from feminine to neuter. Here grammatical gender of the derived form cannot be predicted from the inflectional class of the suffix under the traditional approach to Russian inflectional classes (see Chapter 3). The reason for this is that Class I contains both masculine and neuter nouns and just knowing the inflectional class of the suffix, it is not clear what the gender of the derived form is. It is worth noting that although grammatical gender of a derived expressive form can be predicted from the inflectional class of the expressive suffix (see §4.4.1.2), it cannot be predicted from the inflectional class of a non- expressive suffix like -išč’ in (154)–(155).  (156)                                                        n2[class I] [neut]                                               2                                    n2 [class I] [neut]      n1[class II]  [fem]                                        -išč’                    2                                                         n1[class II]  [fem]      √konopl’-  (157)  a.  konopl’-á                                     b.  konopl’-íšč’-e               hemp-N.SG (FEM; CLASS II)                    hemp-PLACE-N.SG (NEUT; CLASS I)                  ‘hemp’                                                 ‘place for gathering hemp’  In (158), the non-expressive -išč’ does not change the inflectional class, but it changes the grammatical gender from masculine to neuter (159a-b).  113 (158)                                         n2[class I]  [neut]                                   2                       n2[class I]  [neut]         n1[class I]  [masc]                                 -išč’                 2                                               n1[class I]  [masc] √požar-  (159)  a.  požár                                               b.  požár’-išč’-e               fire.N.SG (MASC; CLASS I)                             fire-PLACE-N.SG (NEUT; CLASS I)                  ‘fire’                                                    ‘site of fire’  To summarize, the non-expressive suffix -išč’ ‘place/site’ is a syntactic head, as it can change syntactic category, inflectional class, and grammatical gender (160).  (160)  HEAD                              X[class I] [neut]                                        2                            X[class I] [neut]         Y                                       -išč’  Thus, the non-expressive suffix -išč’ ‘place/site’ and the expressive augmentative suffix -išč’ have distinct syntax, as illustrated in Table 4.21. This homophony between expressive size suffixes and their non-expressive counterparts is quite wide-spread in Russian: every expressive size suffix has a non-expressive homophone (see also §4.5.2 and §4.5.3 below).  Non-expressive -išč’ ‘place/site’ Expressive -išč’ ‘augmentative’ Head Modifier Table 4.21: Non-expressive -išč’ and expressive -išč’ 4.5.2. The suffix -ec The non-expressive counterpart of the expressive diminutive suffix -ec means ‘person’. This suffix is non-expressive, because it has no expressive content, as illustrated in (161).  (161)  gór’-ec                        pr’išól           mountain-PERS.N.SG    came    i. Descriptive: ‘A mountain dweller came.’  Unlike the diminutive -ec, which is a syntactic modifier, the non-expressive -ec is a syntactic head. The first piece of evidence stems from the fact that it can change syntactic category. 114 For example, in (162), the non-expressive -ec attaches to a verb and forms a noun. In (163), the non-expressive -ec attaches to an adjective and also forms a noun.  (162)  a.  č'it-á-t’                                         b.  č’t’-éc               read-TH-INF                                                read-PERS.N.SG (MASC; CLASS I)               ‘to read’                                          ‘reader’  (163)  a.  u-pr’ám-ij                                    b.  u-pr’ám’-ec               VERB.PREF-stubborn-MASC.N.SG      VERB.PREF-stubborn-PERS.N.SG (MASC; CLASS I)              ‘stubborn’                                        ‘stubborn person’  The second piece of evidence comes from the fact that the non-expressive -ec can change inflectional class and grammatical gender of a noun. For example, in (164), the non- expressive -ec changes inflectional class from Class II to Class I, and it changes grammatical gender from feminine to masculine (165a-b).  (164)                                        n2[class I]  [masc]                                    2                      n2[class I] [masc]       n1[class II] [fem]                                    -ec          2                                        n1[class II] [fem]     √gor-  (165)  a.  gor-á                                           b.  gór’-ec               mountain-N.SG (FEM; CLASS II)       mountain-PERS.N.SG (MASC; CLASS I)                  ‘mountain’                                      ‘mountain dweller’  To summarize, the non-expressive suffix -ec is a syntactic head, as it can change syntactic category, inflectional class, and grammatical gender of a noun (166).  (166)   HEAD                             X[class I] [masc]                                        2                             X[class I] [masc]      Y                                          -ec  Thus, the non-expressive suffix -ec ‘person’ and the diminutive expressive suffix -ec have distinct syntax, as illustrated in Table 4.22.    115 Non-expressive -ec ‘person’ Expressive -ec ‘diminutive’ Head Modifier Table 4.22: Non-expressive -ec and expressive -ec 4.5.3. The suffix -k The non-expressive suffix -k means ‘female.’ I treat this suffix as non-expressive, because it has no expressive content (167).  (167)  Ja     uv’íd’el    vnúč’-k-u            I       saw           grandchild-FEM-ACC.SG     i. Descriptive: ‘I saw a granddaughter’  Unlike the diminutive suffix -k, which is a syntactic modifier, the non-expressive -k is a syntactic head. The evidence stems from the fact that the non-expressive -k can change inflectional class and grammatical gender. For example, in (168), it changes inflectional class from Class I to Class II and changes grammatical gender from masculine to feminine (169a- b).  (168)                                            n2[class II]  [fem]                                       2                          n2[class II] [fem]        n1[class I]  [masc]                                   -k                 2                                           n1[class I]  [masc]    √stud’ent-  (169)  a.  stud’ént                                        b.  stud’ént-k-a                 student.N.SG (MASC; CLASS I)             student-FEM-N.SG (FEM; CLASS II)                    ‘student’                                          ‘female student’  To summarize, the non-expressive suffix -k that means ‘female’ is a syntactic head, as it can change inflectional class and grammatical gender of a noun (170).  (170)  HEAD                              X[class II] [fem]                                       2                               X[class II] [fem]    Y                                      -k  Thus, the non-expressive suffix -k ‘female’ and the diminutive expressive suffix -k have distinct syntax, as illustrated in Table 4.23. 116  Non-expressive -k ‘female’ Expressive -k ‘diminutive’ Head Modifier Table 4.23: Non-expressive -k and expressive -k  4.5.4. Summary The difference between expressive size suffixes and their non-expressive homophones is not only of a semantic nature, but also of a syntactic one. Expressive size suffixes are syntactic modifiers (171a), while their non-expressive homophones are syntactic heads (171b). A conclusion that we can draw from this is that homophones are linguistic objects that are not just different in meaning, but also different in syntax, leaving just the sound the same.  (171) a. MODIFIER             Y                       b. HEAD             X                               2                                      2                              X          Y                                    X            Y                              EXPRsize                                                           non-EXPR  4.6. CONCLUSIONS In this Chapter, I used the following diagnostics to determine the morphosyntactic types of Russian expressive suffixes (172).  (172) Diagnostics (cf. Bachrach & Wagner 2007) Diagnostic I:  Do expressive suffixes change syntactic category? Diagnostic II: Do expressive suffixes change grammatical gender? Diagnostic III: Do expressive suffixes change inflectional class?  According to these diagnostics, expressive suffixes are syntactic heads if any of answers to (172) are affirmative. Expressive suffixes are syntactic modifiers if all the answers to (172) are negative (Table 4.24).    117  Diagnostics Syntactic heads Syntactic modifiers Do expressive suffixes change syntactic category? a * Do expressive suffixes change grammatical gender? a * Do expressive suffixes change inflectional class? a * Table 4.24: Syntactic heads vs. syntactic modifiers  Based on (172), I argued that the differences between attitude and size suffixes in Russian are syntactically conditioned. I showed that attitude suffixes are syntactic heads (173a), while size suffixes are syntactic modifiers (173b).  (173) a. HEAD                  X                        b. MODIFIER               Y                              2                                                2                            X            Y                                             X           Y                            EXPRattitude                                                  EXPRsize  Attitude and size suffixes are repeated for convenience in Table 4.25.   affectionate -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’, -úš Attitude suffixes  vulgar -ág, -ák, -ál, -án, -ár, -áx, -íl, -in, -ób,  -ot, -óx, -úg, -úk, -úx  diminutive  -k (allomorphs: -ok/-ek/-ik), -c (allomorphs: -ec/-ic) Size suffixes augmentative  -išč’ Table 4.25: Expressive suffixes in Russian  118 Chapter 5: Where are expressive suffixes merged?  5.1. INTRODUCTION As noted in Chapter 1, I assume an approach to grammar in which words are built in the syntax from two different sites: (i) category-free √Roots (1a), and (ii) √Roots that have already been categorized (1b) (Halle & Marantz 1993; Halle 1997; Marantz 1997, among others). In (1a), a category head X merges with a category-free √Root. In (1b), a category head X merges with already categorized nouns (n), adjectives (a), and verbs (v). The main difference between these two structures is that in (1a), X attaches before a syntactic category is formed, while in (1b) X attaches after a syntactic category has been formed.  (1) a.      X                        b.         X          2                           2        X         √Root                      X           n/a/v                                                          2                                                         n/a/v        √Root  The distinction between word formation from √Roots and word formation from categories is universal, but its manifestations may differ from language to language (Marantz 1997). Here I show that Russian expressive suffixes provide empirical support for these two sites of word formation  I demonstrate that Russian attitude suffixes can merge either with category-free √Roots (2a) or with categories (2b). In contrast, Russian size suffixes can only merge with categorized nouns (3).  (2)         a.                      n                               b.             n                                  2                                     2                                  n          √Root                                n            n/a/v                          EXPRattitude                                EXPRattitude  2                                                                                      n/a/v          √Root 119 (3)                                  n                                  2                       EXPRsize      n                                         2                                       n           √Root  The chapter is organized as follows. In §5.2, I investigate where attitude suffixes are merged. In §5.3 I investigate where size suffixes are merged. In §5.4, I give an intermediate summary of the findings. In §5.5, I discuss co-occurrence restrictions and empirical predictions made by this analysis, and I show that these predictions are borne out in Russian. Finally, in §5.6, I present the conclusions.  5.2. ATTITUDE SUFFIXES In Chapter 4, I showed that Russian attitude suffixes are syntactic heads. The next question that arises is where these heads attach. In other words, what is the locus of merge for these suffixes? Do they merge with √Roots or syntactic categories? I argue that attitude suffixes can merge at both sites: with √Roots and with categories. In the structure (4a), an attitude suffix merges with a category-free √Root. In (4b), an attitude suffixes merges with the syntactic categories n/a/v.   (4)   HEADS     a.            n                             b.               n                                  2                                     2                                n           √Root                              n            n/a/v                             EXPRattitude                             EXPRattitude  2                                                                                       n/a/v          √Root  This section is structured as follows. In §5.2.1, I investigate whether attitude suffixes can merge with √Roots, and in §5.2.2, whether they can merge with categories. In §5.2.3, I show that they can merge with hypocoritics which are truncated first names. In §5.2.4, I present the conclusions.  120 5.2.1. Merging with √Roots Evidence that attitude suffixes merge with category-free √Roots stems from the absence of category-forming morphology inside the attitude suffix. That is, in order for a √Root to be categorized as a noun, adjective, or verb, it often requires a category-forming affix. For example, in kras-ot-á ‘prettiness/beauty (noun)’, the nominal suffix -ot is a category-forming suffix that turns the √Root into a noun (see §5.2.2 for evidence that kras- is indeed a category-free √Root). Thus, if an attitude suffix merges directly with a √Root, there can be no category-forming morphology between the √Root and the attitude suffix (5).  (5)                                 n                                  2                                n           √Root                             EXPRattitude  The data in (6)–(8) show that attitude suffixes can indeed merge with category-free √Roots. The (a) examples in (6)–(8) contain adjectives with a category-forming adjectival suffix -n and the adjectival ending -ij (masculine, singular). In the (b) examples, attitude suffixes merge with √Roots without the adjectival morphology. The resulting words are nouns with expressive meanings (vulgar or affectionate, depending on the meaning of the attitude suffix). As such attitude suffixes function as nominalizers.  (6)    a.  žád-n-ij                                       b.   žad’-úg-a             stingy-ADJ-MASC.N.SG                            stingy-EXPR-N.SG (MASC/FEM)             ‘stingy’                                             ‘stingy animate (vulg)’  (7)    a.  stráš-n-ij                                      b.   straš-íl-a              ugly-ADJ-MASC.N.SG                               ugly-EXPR-N.SG (MASC/FEM)              ‘ugly’                                                ‘ugly animate (vulg)’  (8)    a.  krás-n-ij                                       b.   kras-úl’-a            pretty-ADJ-MASC.N.SG                             pretty-EXPR-N.SG (MASC/FEM)             ‘pretty (archaic)’14                            ‘pretty animate (affect)’   14  This meaning is archaic in Contemporary Russian and can be seen, for example, in the idiom krás-n-a d’év’- ic-a ‘pretty girl’ and the words kras-ot-á ‘beauty’, kras-áv’-ec ‘pretty man’, krás’-i-t’ ‘to adorn’. In Contemporary Russian, the adjective krás-n-ij is used with the meaning ‘red’. 121 Evidence that žad- ‘stingy’ is indeed a √Root stems from the following considerations. First, it doesn’t exist as an independent word; second, it can form the basis for nouns (9a, b), verbs (9c), and adjectives (9d).  (9)  a.  žad’-úg-a                                       b.  žad-ób-a             stingy-EXPR-N.SG (MASC/FEM)           stingy-EXPR-N.SG (MASC/FEM)            ‘stingy animate (vulg)’                       ‘stingy animate (vulg)’        c.   žad-n’-é-t’                                      d.  žád-n-ij            stingy-ADJ-TH-INF                                       stingy-ADJ-MASC.SG           ‘to become stingy’                              ‘stingy’  Crucially, when attitude suffixes merge with the √Root žad- they form expressive nouns. This is shown in the structure in (10). The notation # is used to indicate Number.  (10)          #            2          #             n                                  žad’-ug/ob/in-a ‘stingy animate (vulg)’         -a       2                 n           √žad-          -ug/-ob/-in  More evidence that attitude suffixes can indeed merge with category-free √Roots is given below. The (a) examples in (11)–(13) contain verbs with a verbal conjugation marker -a (glossed as a Theme vowel, following Halle & Matushansky 2006) and the infinitival ending -t’. In the (b) examples, attitude suffixes merge with √Roots without the verbal morphology. The resulting words are nouns with expressive meanings. Here, attitude suffixes also function as nominalizers.  (11)  a.  vr-á-t’                                         b.  vr-úš-a             lie-TH-INF                                                    lie-EXPR-N.SG (MASC/FEM)            ‘to lie’                                              ‘liar (affect)’  (12)  a.  xáp-a-t’                                       b.  xap-úg-a             grab-TH-INF                                               grab-EXPR-N.SG (MASC/FEM)             ‘to grab’                                         ‘grabber (vulg)’  (13)  a.  igr-á-t’                                         b.  igr-úl’-a             ply-TH-INF                                                  play- EXPR-N.SG (MASC/FEM)               ‘to play’                                           ‘playing animate (affect)’  122 V(o)r- ‘stingy’ is a √Root because it can form the basis for both nouns (14a, b, c), and verbs (14d).  (14)  a.  vr-úš-a                                           b.  vr-ún            lie-EXPR-N.SG (MASC/FEM)                    lie-NOM-N.SG (MASC)           ‘liar (affect)’                                          ‘liar’         c.  vr-ál’                                                d.   vr-á-t’            lie-NOM-N.SG (MASC)                             lie-TH-INF           ‘liar’                                                       ‘to lie’  When attitude suffixes merge with the √Root v(o)r- they form expressive nouns. This is shown in the structure in (15).  (15)         #            2           #            n                                        vr-úš-a    ‘liar (affect)’          -a       2                 n            √v(o)r-                -uš  To summarize, attitude suffixes can merge with category-free √Roots (Table 5.1).  EXPRattitude EXPRattitude + √Root -án’, -áš, -ón, -úl’, -ún’-, -úr, -ús’, -úš, -ág, -ák,-ál, -án, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx  Table 5.1: Attitude suffixes (attachment to √Roots)  5.2.2. Merging with categories I now show that attitude suffixes can also merge with categories. Evidence stems from the fact that attitude suffixes can attach after categorizing morphology has been already added to the √Root. In particular, they can merge with nouns, adjectives, and verbs. This is illustrated in (16).      123 (16)                                n                                   2                               n               n/a/v                      EXPRattitude      2                                       n/a/v          √Root                                category affix  In the (a) examples in (17)–(19), we find adjectives with a category-forming adjectival suffix -n. In the (b) examples, attitude suffixes merge outside of this adjectival suffix. The resulting examples are grammatical nouns with expressive meanings.  (17)  a.  žád-n-ij                                       b.  žad-n’-úg-a               stingy-ADJ-MASC.N.SG                           stingy-ADJ-EXPR-N.SG (MASC/FEM)               ‘stingy’                                             ‘stingy animate (vulg)’  (18)   a.  gr’áz-n-ij                                       b.  gr’az-n-úx-a               dirty-ADJ-MASC.N.SG                              dirty-ADJ-EXPR-N.SG (MASC/FEM)               ‘dirty’                                               ‘dirty animate (vulg)’  (19)   a.  rod-n-ój                                       b.  rod-n-úl’-a              dear-ADJ-MASC.N.SG                               dear-ADJ-EXPR-N.SG (MASC/FEM)               ‘dear’                                                ‘dear animate (affect)’  We have seen in §5.2.1 that the attitude suffix -ug is able to directly merge with the √Root žad- ‘stingy’ to form the noun žad’-úg-a ‘stingy animate (vulg)’. Here we see that the same suffix can also merge with the categorized adjective žad-n-ij ‘stingy’. In (17b), the category- forming adjectival suffix -n is present in the derivation žad-n’-úg-a ‘stingy animate (vulg)’, which means that here an adjectival category is formed first, and then the attitude suffix -ug attaches (20). Although pairs like žad-n’-úg-a ‘stingy animate (vulg)’ and žad’-úg-a ‘stingy animate (vulg)’ are present in the language, it is unclear why in other data an attitude suffix can combine either with a √Root, or with a category (but not both). For example, the pair rod-n-úl’-a ‘dear animate (affect)’–*rod-úl’-a ‘dear animate (affect)’ does not exist in Russian.       124 (20)                       #                          2                          #             n                               žad-n’-úg-a ‘stingy animate (vulg)’                      -a        2                               n            a                                -ug     2                                         a           √žad-                                        -n  To summarize, attitude suffixes can merge with categorized adjectives (Table 5.2).  EXPRattitude EXPRattitude + a -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’, -úš, -ág, -ák,-ál, -án, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx  Table 5.2: Attitude suffixes (merging with adjectives)  Next I show that attitude suffixes can also merge with verbs. In (21)–(23), attitude suffixes merge with categorized verbs. Evidence that the input for suffixation is indeed a verb stems from the fact that these forms contain the verbal prefixes pr’i-, ras-, and za-.  (21)  a.  pr’i-l’ip-á-t’                               b.  pr’i-l’ip-ál-a             VERB.PREF-cling-TH-INF                        VERB.PREF-cling-EXPR-N.SG (MASC/FEM)             ‘to cling’                                          ‘clinging animate (vulg)’  (22)  a.  ras-t’er’-á-t’                               b.  ras-t’er’-áš-a             VERB.PREF-lose-TH-INF                         VERB.PREF-lose-EXPR-N.SG (MASC/FEM)               ‘to lose’                                            ‘animate who loses things (affect)’  (23)  a.  za-v’ir-á-t’                                  b.  za-v’ir-úx-a             VERB.PREF-lie-TH-INF                             VERB.PREF-lie-EXPR-N.SG (MASC/FEM)               ‘to lie’                                              ‘lying animate (affect)’  In Russian, verbal prefixes merge before expressive suffixes. The evidence for that is twofold. First, as I mentioned in Chapter 4, the nouns *l’ip-ál-a ‘clinging animate’, *t’er’-áš- a ‘losing animate’, and *v’ir-úx-a ‘lying animate’ do not exist in Russian. Second, in Russian, verbal prefixes do not form verbs from categorized nouns. For example, if we try to form verbs from the categorized nouns žad’-úg-a ‘stingy animate (vulg)’, straš-íl-a ‘ugly 125 animate (vulg)’, or kras-úl’-a ‘pretty animate (affect)’, all resulting forms will be ungrammatical (24)–(26).  (24)    * pr’i/ras/za-žad’-ug-a-t’           VERB.PREF-stingy-EXPR-TH-INF             ‘to become a stingy animate’  (25)    * pr’i/ras/za-straš-il-a-t’           VERB.PREF-ugly-EXPR-TH-INF             ‘to become an ugly animate’  (26)    * pr’i/ras/za-kras-ul’-a-t’           VERB.PREF-pretty-EXPR-TH-INF             ‘to become a pretty animate’  The prefixes pr’i-, ras-, and za- are so-called “lexical” prefixes as opposed to prefixes that are called “superlexical” (Babko-Malaya 2003, Dimitrova-Vulchanova 1999, Fowler 1996, Rojina 2004, Svenonius 2004). The difference between lexical and superlexical prefixes is that the former are syntactic heads that originate in the verbal domain, while that latter are syntactic modifiers that originate outside of the verbal domain (Svenonius 2004). The lexical prefixes can combine with superlexical prefixes. In such cases, the superlexical prefixes always appear outside of the lexical prefixes (Svenonius 2004:207). This is illustrated in (27).  (27)  a.  vi-brás-iv-a-t’                               b.  po-vi-brás-iv-a-t’             LEX-trow-SUFF-TH-INF                            SUPERLEX-LEX-trow-SUFF-TH-INF               ‘to throw out’                                    ‘to throw out one by one’ (Svenonius 2004:207)  The evidence that the prefixes pr’i-, ras-, and za- are indeed lexical comes from the fact that they can combine with the superlexical prefix po-. In these cases, the superlexical prefix attaches outside these prefixes (28)–(30).15  (28)  a.  pr’i-l’ip-á-t’                                 b.  po-pr’i-l’ip-á-t’             LEX-cling-TH-INF                                       SUPERLEX-LEX-cling-TH-INF             ‘to cling’                                           ‘to cling one by one’  15  Many thanks to Dr. Sergei Tatevosov for these data.  126  (29)  a.  ras-bros-á-t’                                 b.  po-ras-brás-iv-a-t’              LEX-trow-TH-INF                                       SUPERLEX-LEX-trow-TH-INF              ‘to throw’                                         ‘to throw one by one’  (30)  a.  za-v’ir-á-t’                                    b.  po-za-v’ir-á-t’              LEX-lie-TH-INF                                           SUPERLEX-LEX-lie-TH-INF              ‘to lie’                                              ‘to lie one by one’  Based on the fact that pr’i-, ras-, and za- are so-called lexical prefixes, I assume the following structure (31) for them.  (31)                       #                          2                         #            n                                   2                                  n             v                                        2                                           v           √Root                                  pr’i-/ras-/za-  The structure for the example (21b), repeated in (32b), is shown in (32c). Since verbal prefixes merge before expressive suffixes in Russian, the prefix pr’i- merges with the √Root l’ip-, forming a verbal category v (32c). The attitude suffix -al attaches after the verbal category has already been formed.  (32)  a.  pr’i-l’ip-á-t’                               b.  pr’i-l’ip-ál-a             VERB.PREF-cling-TH-INF                        VERB.PREF-cling-EXPR-N.SG (MASC/FEM)             ‘to cling’                                         ‘clinging animate (vulg)’         c.                    #                          2                         #            n                               pr’i-l’ip-ál-a ‘clinging animate (vulg)’                         -a       2                                  n             v                             -al        2                                           v           √l’ip-                                         pr’i-  To summarize, attitude suffixes can merge with categorized verbs (Table 5.3).  127 EXPRattitude EXPRattitude + v -án’, -áš-, -ón, -úl’, -ún’, -úr, -ús’, -úš, -ág, -ák,-ál, -án, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx  Table 5.3: Attitude suffixes (merging with verbs)  Next I show that attitude suffixes can also merge with categorized nouns. The (a) examples in (33)–(35) show nouns with the category-forming nominal suffixes -ot, -ak, and-ik. In the (b) examples, attitude suffixes merge outside these nominal suffixes. The resulting words are nouns with expressive meanings (vulgar or affectionate, depending on the meaning of an attitude suffix).  (33)  a.  kras-ot-á                                          b. kras-ot-úl’-a               pretty-NOM-N.SG (FEM)                                pretty-NOM-EXPR-N.SG (MASC/FEM)               ‘beauty’                                                 ‘pretty animate (affect)’  (34)  a.  č’ud-ák                                            b.   č’ud-ač’-ín-a               wonder-NOM.N.SG (MASC)                       wonder-NOM-EXPR-N.SG (MASC/FEM)               ‘an eccentric’                                        ‘an eccentric (vulg)’  (35)  a.  star’-ík                                            b.  star'-ik-án               old-NOM.N.SG (MASC)                                  old-NOM-EXPR.N.SG (MASC)               ‘old man’                                                ‘old man (vulg)’  Consider the example (33). The √Root is kras- ‘pretty’; in (36), a noun (36a), a verb (36b), and an adjective (36c) are all formed with the same √Root kras-. The √Root kras- does not exist as a separate word *kras ‘pretty’ in Russian.  (36)  a.  kras-ot-á                                        b.  krás’-i-t’               pretty-NOM-N.SG (FEM)                               pretty-TH-INF               ‘prettiness/beauty’                                 ‘to adorn (make pretty)’          c.  kras’-ív-ij                pretty-ADJ-MASC.SG                ‘pretty/beautiful’  The suffix -ot is a productive nominal suffix that forms feminine nouns denoting abstract concepts (Efremova 2006). This is illustrated in (37).  128 (37)  a.  bistr-ot-á                                           b.  kras-ot-á               quick-NOM-N.SG (FEM)                          pretty-NOM-N.SG (FEM)              ‘quickness’                                             ‘prettiness/beauty’          c.  dobr-ot-á                kind-NOM-N.SG (FEM)               ‘kindness’  When the attitude suffix -ul’ is used with kras-ot-á ‘beauty’, it merges outside the nominal suffix: kras-ot-úl’-a ‘pretty animate (affect)’. Since the nominal suffix is still present in the resulting noun, this means that -ul’ merges only after the noun category has already been formed (38).  (38)                       #                          2                        #             n2                              kras-ot-úl’-a ‘pretty animate (affect)’                       -a         2                                   n2             n1                               -ul’       2                                        n1           √kras                                          -ot  To summarize the findings so far, attitude suffixes can merge with all syntactic categories n/a/v in Russian (Table 5.4).  EXPRattitude EXPRattitude + n EXPRattitude + a EXPRattitude + v -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’, -úš, -ág, -ák,-ál, -án, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx    Table 5.4: Attitude suffixes (merging with categories)  5.2.3. Merging with short names (hypocoristics) In this subsection I present additional evidence that attitude suffixes can merge with categorized nouns. The evidence stems from so-called hypocoristics. These are short forms of first names that are formed by truncation of their corresponding full forms. For example, in (39) the masculine name Bor’ís ‘Boris (full form)’ is truncated to Bór’-a ‘Boria 129 (hypocoristic)’. In (40), the feminine name Ól’g-a ‘Olga (full form)’ is truncated to Ól’-a ‘Olia (hypocoristic)’. More examples of hypocoristics are given in (41) and (42).  (39)  a.  Bor’ís                                             b.   Bór’-a              Boris.N.SG (MASC; CLASS I)                Boria-N.SG (MASC; CLASS II)             ‘Boris (full form)’                               ‘Boria (hypocoristic)’  (40)  a.  Ól’g-a                                             b.   Ól’-a             Olga-N.SG (FEM; CLASS II)                       Olia-N.SG (FEM; CLASS II)               ‘Olga (full form)’                                ‘Olia (hypocoristic)’  (41)  a.  P’ótr                                                b.  P’ét’-a             P’otr.N.SG (MASC; CLASS I)                      Petia-N.SG (MASC; CLASS II)               ‘P’otr (full form)’                                ‘Petia (hypocoristic)’  (42)  a.  Nad’éžd-a                                       b.   Nád’-a             Nadezhda-N.SG (FEM; CLASS II)               Nadia-N.SG (FEM; CLASS II)               ‘Nadezhda (full form)’                          ‘Nadia (hypocoristic)’  The full forms of first names are used in formal registers, while hypocoristics are used in informal registers. For example, in a teacher-student relationship, it is considered inappropriate if a student addresses a teacher with a short form. According to Stankiewicz (1968:146), hypocoristics have the function of “familiar and intimate terms of address or reference”. They are very productively used among friends and relatives.  Most hypocoristics have a monosyllabic shape (C)(C)VC-, and some have the disyllabic shape (C)VCVC- (Stankiewicz 1968:146). Hypocoristics with an initial cluster are not common, and those beginning with a vowel (VC- or VCVC-) are primarily feminine. All hypocoristics have the stress on the final (or only) syllable of the truncated stem.  All hypocoristics belong to the inflectional Class II. Thus, if a hypocoristic is formed from a full name belonging to Class I, the inflectional class changes from Class I to Class II. For example, in (39) and (41) above, the full names Bor’ís and P’ótr belong to Class I (-Ø ending in Nom sg.). Once the hypocoristics Bór’-a and P’ét’-a are formed, they switch to Class II ( -a ending in Nom sg.). If full names are already in Class II, there is no change in inflectional class, (40) and (42) above. 130  Hypocoristics are used very productively with attitude suffixes, which add expressive meanings. The expressive meanings are vulgar or affectionate, depending on the meaning of the attitude suffix. Attitude suffixes can only merge with monosyllabic hypocoristics. For example, in (43)–(45) below, the attitude suffixes merge with monosyllabic hypocoristics Bór’-a, Ól’-a, and Nád’-a. All resulting examples are grammatical in Russian. In (46), the attitude suffix merges with a polysyllabic hypocoristic Volód’-a. The resulting example (46b) is ungrammatical in Russian. Interestingly, there is no such prosodic restriction when attitude suffixes merge with non-name bases, for example, komnat-úx-a ‘room (vulg)’, dorog-úš-a ‘dear animate (affect)’. This difference in prosodic restrictions between names and common nouns is an interesting topic of further research.  (43)  a.  Bór’-a                                              b.  Bor’-ús’-a              Boria-N.SG (MASC; CLASS II)                  Boria-EXPR-N.SG (MASC; CLASS II)             ‘Boria’                                                    ‘Boria (affect)’  (44)  a.  Ól’-a                                                 b.  Ol’-ún’-a             Olia-N.SG (FEM; CLASS II)                            Olia-EXPR-N.SG (FEM; CLASS II)               ‘Olia’                                                     ‘Olia (affect)’  (45)  a.  Nád’-a                                             b.  Nad’-úx-a             Nadia-N.SG (FEM; CLASS II)                    Nadia-EXPR-N.SG (FEM; CLASS II)               ‘Nadia’                                                   ‘Nadia (vulg)’  (46)  a.  Volód’-a              Volodia-N.SG (MASC; CLASS II)              ‘Volodia (hypocoristic from the name Vdad’imir)’   b.* Volod’-úx-a        Volodia-EXPR-N.SG (MASC; CLASS II)       ‘Volodia (vulg)’  When attitude suffixes (with the exception of -an) merge with hypocoristics, there is no change in inflectional class. Hypocoristic+attitude suffix formations belong to the same inflectional class, Class II, as hypocoristics (43)–(45). However, when the attitude suffix -an merges with hypocoristics, there is a change in inflectional class. Hypocoristics+an formations are always in Class I. This is consistent with my proposal in Chapter 4 that 131 attitude suffixes are syntactic heads associated with an inflectional class of their own. As I argued in Chapter 4, all attitude suffixes (except -an) are associated with Class II, while the suffix -an is associated with Class I. Thus, the majority of attitude suffixes can change the inflectional class of the base to Class II, while -an can change the inflectional class of the base to Class I. The examples in (47) illustrate how the inflectional class of the first name can change multiple times, depending on whether a hypocoristic is formed with or without an attitude suffix. In (47a), the full name N’ikoláj belongs to Class I (-Ø ending in Nom sg.). In (47b), the hypocoristic Kól’-a is formed and the inflectional class changes from Class I to Class II (-a ending in Nom sg.). In (47c), the attitude suffix -an merges with the hypocoristic and the inflectional class changes again, this time from Class II to Class I.  (47)  a.  N’ikoláj                                          b.  Kól’-a            Nikolaj.N.SG (MASC; CLASS I)                   Kolia-N.SG (MASC; CLASS II)               ‘Nikolaj (full form)’                              ‘Kolia (hypocoristic)’          c.  Kol’-án             Kolia-N.SG (MASC; CLASS I)            ‘Kolia (vulg)’  With respect to these data, the following question arises: what is the syntactic structure for hypocoristic+attitude suffix formations? I propose that in such formations, attitude suffixes merge with a categorized noun formed by means of truncation. In other words, I propose that a syntactic head triggers truncation, which occupies a particular hierarchical position in the syntactic tree. Since attitude suffixes attach after truncation has applied, they merge with a category formed by truncation. This is illustrated in (48). In this structure, a hypocoristic is formed by truncation of a full name (n2), then an attitude suffix attaches to an already truncated noun (n3).  (48)                                #                                  2                                #             n3                                           2                                      n3              n2 (hypocoristic)                             EXPRattitude      2                                                 n2            n1 (full form)                                            truncation 132  The evidence suggesting that a syntactic head triggers truncation is a change in inflectional class. All hypocoristics formed by truncation are in Class II, regardless of the inflectional class of the input. Thus, I propose that a head triggering truncation is associated with inflectional Class II of its own. For example, consider the structure in (49) for the data in (47): the hypocoristic Kól’-a is formed by truncation of the full name N’ikolaj. As I mentioned earlier, N’ikolaj belongs to Class I. When the hypocoristic Kól’-a is formed, the inflectional class changes from Class I to Class II. This means that a noun head triggers truncation, as only noun heads are capable of changing the inflectional class of the base. The attitude suffix -an merges after the hypocoristic has been formed. Since -an is a noun head associated with inflectional Class I (see Chapter 4 for argumentation), the inflectional class of this formation changes again. The resulting expressive noun Kol’-án (vulg) is now in Class I.  (49)                                        #                                          2                                     #                n3[class I]                                     Kol’-án (vulg)                                   -Ø             2                                             n3[class I]       n2[class II]                           Kól’-a (hypocoristic)                                            -an          2                                                    n2[class II]        n1[class I]                                                                        4                                                                    N’ikolaj                     N’ikolaj (full form)  Another piece of evidence that a noun head triggers truncation stems from common nouns. The truncated forms of common nouns also involve a change in class. Truncated forms are always in Class II, just like hypocoristics. For example, in (50a), the full common noun p’etúx ‘rooster’ belongs to Class I (-Ø ending in Nom sg.). In (50b), the truncated form p’ét’- a ‘rooster (truncated form)’ belongs to Class II (-a ending in Nom sg.), which illustrates a change in class. In (51a), the common noun pár’en’ ‘guy’ is in Class I. In (51b), the truncated form pár’-a ‘guy (truncated form)’ is in Class II.  (50)  a.  p’etúx                                           b.  p’ét’-a             rooster.N.SG (MASC; CLASS I)             rooster-N.SG (MASC; CLASS II)               ‘rooster (full form)’                          ‘rooster (truncated form; colloquial)’ 133  (51)  a.  pár’en’                                          b.  pár’-a             guy.N.SG (MASC; CLASS I)                  guy-N.SG (MASC; CLASS II)               ‘guy (full form)’                               ‘guy (truncated form; colloquial)’ (Stankiewicz 1968:142)  The proposed structure for (50b) is given in (52). A noun head (n2) triggers truncation and introduces inflectional Class II. This accounts for a change in class from Class I to Class II when the truncated noun p’ét’-a ‘rooster’ is formed.  (52)                                              #                                                2                                             #             n2[class II]                    p’ét’-a ‘rooster (truncated)’                                            -a         2                                                  n2[class II]     n1[class I]                                                                  4                                                                p’etux                p’etúx ‘rooster (full form)’  To summarize, the analysis of truncation supports the current proposal that attitude suffixes can merge with a noun category. I argued that a noun head triggers truncation in Russian. Therefore, when attitude suffixes merge with truncated forms, they merge with already categorized nouns, as illustrated in (53).  (53)                                  #                                    2                                #               n3                                 (expressive form)                                             2                                       n3                 n2[class II]            (truncated form)                             EXPRattitude          2                                                  n2[class II]        n1[class X]   (full form)  5.2.4. Summary Russian attitude suffixes provide empirical evidence for the assumption that words are built in the syntax from two different sites: (i) √Roots, and (ii) categories n/a/v (Table 5.5).  134 EXPRattitude EXPRattitude + n EXPRattitude + a EXPRattitude + v EXPRattitude + √Root -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’, -úš, -ág, -ák,-ál, -án, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx     Table 5.5: Attitude suffixes (attachment to √Roots and to categories)  5.3. SIZE SUFFIXES  As I argued in Chapter 4, size suffixes are syntactic modifiers. But we have not yet addressed the question as to what they modify. That is, we need to establish the locus of merge for size suffixes. I now show that, unlike attitude suffixes, which can merge with √Roots and with categories, size suffixes can only merge with categorized nouns (54).  (54)   MODIFIER             n                                  2                      EXPRsize        n                                          2                                           n           √Root  In §5.3, I investigate whether size suffixes can merge with nouns. In §5.3.2, I examine whether they can merge with adjectives and verbs. In §5.3.3 I investigate whether they can merge with √Roots. Finally, in §5.3.3, I present the conclusions.  5.3.1. Merging with nouns The data in (55)–(59) illustrate that Russian size suffixes can merge with nouns. The (a) examples contain nouns with the nominal suffixes -ot, -ak, -in, -un, and -b. In the (b) examples, size suffixes attach outside these nominal suffixes. In (55b), the augmentative suffix -išč’ attaches outside of the nominal suffix -ot. And in (56b), the diminutive suffix -ic attaches outside of the nominal suffix -b. Similar examples are given in (57)–(59).  (55)  a.  kras-ot-á                                          b.  kras-ot-íšč’-a               pretty-NOM-N.SG (FEM)                                pretty-NOM-EXPR-N.SG (FEM)               ‘beauty’                                                 ‘beauty (aug)’  135 (56)  a.  prós’-b-a                                         b.  prós’-b’-ic-a               ask-NOM-N.SG (FEM)                                     ask-NOM-EXPR-N.SG (FEM)               ‘request’                                                ‘request (dim)’  (57)   a.  sos-ún                                             b.  sos-un’-éc               suck-NOM.N.SG (MASC)                                suck-NOM-EXPR.N.SG (MASC)               ‘suckling’                                              ‘suckling (dim)’  (58)  a.  č’ud-ák                                            b.  č’ud-ač’-ók               wonder-NOM.N.SG (MASC)                       wonder-NOM-EXPR.N.SG (MASC)               ‘eccentric animate’                                ‘eccentric animate (dim)’  (59)  a.  zaráp’-in-a                                      b.  zaráp’-in-k-a16               scratch-NOM-N.SG (FEM)                             scratch-NOM-EXPR-N.SG (FEM)                 ‘scratch’                                               ‘scratch (dim)’  The proposed structure for (55) is given in (60). As I showed in §5.2.2, the word kras-ot-á ‘beauty’ contains the category-free √Root kras- and the nominal suffix -ot. When the augmentative suffix -išč’ is used, it merges outside the nominal suffix. This means that a noun category is formed first, and then the augmentative -išč’ attaches as schematized in (60).  (60)                        #                           2                          #               n                              kras-ot’-íšč’-a ‘beauty (aug)’                        -a          2                                 -išč’          n                       kras-ot-á ‘beauty’                                          2                                         n         √kras                                           -ot  As another example, consider the structure of (56), given in (61): the word prós’-b-a ‘request’ contains the √Root pros’- and the nominal suffix -b. When the noun prós’-b-a ‘request’ is used with the diminutive -ic, the diminutive suffix attaches outside of the nominal suffix: prós’-b’-ic-a ‘request (dim)’. Since the nominal suffix -b is still present in the derivation, it means that a noun category is formed first, and only after that the diminutive attaches (61).  16  Here the suffix -in is a non-expressive homophone of the expressive suffix -in. The non-expressive -in is a nominalizer that means ‘result of action’. 136  (61)                        #                           2                           #              n                              prós’-b’-ic-a ‘request (dim)’                       -a         2                                -ic            n                       prós’-b-a ‘request’                                          2                                           n         √pros’                                          -b  That pros’- is indeed a √Root can be seen in (62): this √Root can serve as the basis for both nouns (62a) and verbs (62b). However, the √Root pros’- does not exist as an independent word. That -b is indeed a nominalizing suffix can be seen in (63), where it derives nouns from various √Roots.  (62)  a.  prós’-b-a                                         b.  pros’-í-t’               ask-NOM-N.SG (FEM)                                     ask-TH-INF               ‘request’                                                ‘to ask’  (63)  a.  prós’-b-a                                        b.  mol’-b-á               ask-NOM-N.SG (FEM)                                    pray-NOM-N.SG (FEM)               ‘request’                                                ‘prayer’           c.  xod’-b-á               walk-NOM-N.SG (FEM)               ‘walking’  Additional evidence that size suffixes merge with categorized noun stems from truncated nouns. As I suggested in §5.2.3, truncation is associated with a nominal head of its own. This makes the following prediction: since size suffixes merge with nouns, they should also be able to merge with truncated nouns. This prediction is borne out. In (64)–(67) I show that the diminutive suffix -k merges with both truncated names (hypocoristics) and with truncated common nouns. In (64)–(66), -k merges with the hypocoristics Kól’-a, Bór’-a and Ól’-a; the resulting data are all grammatical nouns with diminutive meaning. In (67), the diminutive -k merges with a truncated common noun p’ét’-a ‘rooster’.  137 (64)  a.  Kól’-a                                             b.  Kól’-k-a             Kolia-N.SG (MASC; CLASS II)                     Kolia-EXPR-N.SG (MASC; CLASS II)            ‘Kolia (hypocoristic)’                            ‘Kolia (dim)’  (65)  a.  Bór’-a                                              b.  Bór’-k-a              Boria-N.SG (MASC; CLASS II)                 Boria-EXPR-N.SG (MASC; CLASS II)            ‘Boria (hypocoristic)’                            ‘Boria (dim)’  (66)  a.  Ól’-a                                               b.  Ól’-k-a             Olia-N.SG (FEM; CLASS II)                            Olia-EXPR-N.SG (FEM; CLASS II)               ‘Olia (hypocoristic)’                             ‘Olia (dim)’  (67)  a.  p’ét’-a                                               b.  p’ét’-k-a              rooster-N.SG (MASC; CLASS II)                rooster-EXPR-N.SG (MASC; CLASS II)            ‘rooster (truncated form)’                     ‘rooster (dim)’  The proposed structures for (64) and (67) are given in (68) and (69), respectively. In these structures, truncation is associated with a noun category (n2). The diminutive -k merges with the noun category formed by truncation.  (68)                                        #                                          2                                      #              n2                                            Kól’-k-a (dim)                                     -a          2                                               -k            n2[class II]                          Kól’-a (hypocoristic)                                                         2                                                    n2[class II]      n1[class I]                                                                      4                                                                    N’ikolaj                   N’ikolaj (full form)  (69)                                       #                                         2                                         #              n2                                             p’ét’-k-a ‘rooster (dim)’                                     -a         2                                             -k              n2[class II]                     p’ét’-a ‘rooster (truncated form)’                                                         2                                                   n2[class II]         n1[class I]                                                                    4                                                                  p’etux                       p’etúx ‘rooster (full form)’  To summarize, size suffixes can merge with categorized nouns in Russian (Table 5.6).  138 EXPRsize EXPRsize + n -k (-ok, -ek, -ik); -c (-ic, -ec); -išč’  Table 5.6: Size suffixes (merging with nouns)  5.3.2. Size suffixes do not merge with adjectives or verbs In this section I show that size suffixes cannot merge with adjectives or verbs. For example, if size suffixes were able to merge with adjectives, we would expect adjectival morphology to appear between a √Root and a size suffix (70). In this structure, an adjective is formed first, and then a size suffix attaches; however, as the Russian data show, the structure (70) is unattested.  (70)                       *#                           2                           #              a                                   2                            EXPRsize       a                                           2                                            a         √Root  Examples (71a) and (72a) show adjectives with the category-forming adjectival suffix -n. In (71b–d) and (72b–d), size suffixes are added outside of the suffix -n: the resulting forms are all ungrammatical. In (71a), the word žád-n-ij ‘stingy’ contains the √Root žad- and the adjectival suffix -n. In (71b), the augmentative suffix -išč’ is added outside of the adjectival suffix, but this form is ungrammatical. In (71c), the diminutive suffix -k (allomorphs: -ok, -ek, -ik) is added outside the adjectival suffix and again the resulting form is ungrammatical. In (71d), the diminutive suffix -c (allomorphs: -ic, -ec) is added outside of the adjectival suffix, which also results in ungrammaticality.  (71)  a.  žád-n-ij                                          b.* žad-n’-išč’-ij               stingy-ADJ-MASC.N.SG                                  stingy-ADJ-EXPR-MASC.N.SG               ‘stingy’                                                  ‘stingy (aug)’         c.* žad-n-(o/e/i)k-ij                             d.* žad-n’-(i/e)c-ij                stingy-ADJ-EXPR-MASC.N.SG                     stingy-ADJ-EXPR-MASC.N.SG               ‘stingy (dim)’                                        ‘stingy (dim)’  139 (72)  a.  gr’áz-n-ij                                       b.* gr’az-n’-išč’-ij               dirty-ADJ-MASC.N.SG                                    dirty-ADJ-EXPR-MASC.N.SG               ‘dirty’                                                    ‘dirty (aug)’          c. * gr’az-n-(o/e/i)k-ij                            d.* gr’az-n’-(i/e)c-ij                dirty-ADJ-EXPR-MASC.N.SG                         dirty-ADJ-EXPR-MASC.N.SG                ‘dirty (dim)’                                           ‘dirty (dim)’  To summarize, size suffixes cannot merge with categorized adjectives (Table 5.7).  EXPRsize EXPRsize + a -k (-ok, -ek, -ik); -c (-ic, -ec); -išč’ * Table 5.7: Size suffixes (no merging with adjectives)  Next I turn to the question as to whether size suffixes can merge with verbs. If size suffixes were able to merge with verbs, we would expect verbal morphology to appear between a √Root and a size suffix (73). In this structure, a verb is formed first, and then a size suffix is attached. However, this structure is unattested in Russian. As the data below illustrate, size suffixes do not merge with categorized verbs.  (73)                       *#                           2                           #              v                                   2                               EXPRsize       v                                           2                                             v         √Root  Examples (74a) and (75a) show verbs with the verbal prefixes pr’i- and ras-. In (74b–d) and (75b–d), size suffixes are added to these verbs, which produces ungrammatical forms. In (74b), the augmentative suffix -išč’ is added to the verb pr’i-l’ip-á-t’ ‘to cling’. The resulting form is ungrammatical. In (74c), the diminutive suffix -k (allomorphs: -ok, -ek, -ik) is added to this verb, which also produces ungrammatical forms. In (74d), the diminutive suffix -c (allomorphs: -ic, -ec) merges with this verb. The resulting form is again ungrammatical.  140 (74)  a.  pr’i-l’ip-á-t’                                b.* pr’i-l’ip’-išč’-a-t’             VERB.PREF-cling-TH-INF                         VERB.PREF-cling-EXPR-TH-INF             ‘to cling’                                           ‘to cling (aug)’          c. * pr’i-l’ip-(o/e/i)k-a-t’                 d.* pr’i-l’ip’-(i/e)c-a-t’                 VERB.PREF-cling-EXPR-TH-INF            VERB.PREF-cling-EXPR-TH-INF             ‘to cling (dim)’                                 ‘to cling (dim)’  (75)  a.  ras-t’er’-á-t’                                b.* ras-t’er’-išč’-a-t’              VERB.PREF-loose-TH-INF                        VERB.PREF-loose-EXPR-TH-INF               ‘to loose’                                           ‘to loose (aug)’          c. * ras-t’er’-(o/e/i)k-a-t’                  d.* ras-t’er’-(i/e)c-a-t’                VERB.PREF-loose-EXPR-TH-INF            VERB.PREF-loose-EXPR-TH-INF              ‘to loose (dim)’                                 ‘to loose (dim)’  To summarize, size suffixes cannot merge with categorized verbs (Table 5.8).  EXPRsize EXPRsize + v -k (-ok, -ek, -ik); -c (-ic, -ec); -išč’ * Table 5.8: Size suffixes (no merging with verbs)  5.3.3. No merging with √Roots Thus far we have established that size suffixes can only merge with nouns, not with verbs or adjectives. In light of the fact that attitude suffixes can merge with √Roots (§5.2.1), the question arises as to whether size suffixes can merge with category-free √Roots as well. In this subsection I demonstrate that they cannot. If size suffixes were able to attach to √Roots, we would expect them to appear inside category-forming suffixes. In (76), a size suffix attaches to a √Root before a category is formed. However, as the data below indicate, this is not the case in Russian.  141 (76)                       *#                           2                         #           n/a/v                                   2                              n/a/v       √Root                                           2                                  EXPRsize    √Root  In (77c, d), size suffixes attach inside the nominal suffixes and all resulting forms are ungrammatical. In (77a), kras-ot-á ‘beauty’ contains the √Root kras- and the nominal suffix -ot. In (77b), the augmentative suffix -išč’ merges outside the nominal suffix. This produces a grammatical form. However, in (77c), -išč’ merges inside the nominal suffix, and the resulting form is ungrammatical. In addition, the example in (77d) shows that -išč’ cannot merge with the √Root even when the nominal morphology is absent. The structures for (77b) and (77c) are given in (78a) and (78b), respectively.  (77)  a.  kras-ot-á                                   b. kras-ot-íšč’-a             pretty-NOM-N.SG (FEM)                       pretty-NOM-EXPR-N.SG (FEM)             ‘beauty’                                         ‘beauty (aug)’           c.* kras’-išč’-ot-a                         d.* kras’-išč’-a                pretty-EXPR-NOM-N.SG (FEM)         pretty-EXPR-N.SG (FEM)               ‘beauty (aug)’                                ‘pretty (aug)’  (78)   a.                   #                              b.                    *#                           2                                             2                        #               n                                            #               n                      -a           2                                -a          2                                  -išč’          n                                              n           √kras                                          2                                   -ot        2                                         n         √kras                                         -išč’         √kras                                           -ot                     kras-ot-íšč’-a ‘beauty (aug)’                *kras-íšč’-ot-a ‘beauty (aug)’  Further examples illustrate this same behaviour of the size suffixes -ic and -ok are shown in (79) and (80).  142 (79)  a.  prós’-b-a                                    b.  prós’-b’-ic-a            ask-NOM-N.SG (FEM)                             ask-NOM-EXPR-N.SG (FEM)             ‘request’                                         ‘request (dim)’          c.* pros’-ic-b-a                             d.* pros’-ic-a             ask-EXPR-NOM-N.SG (FEM)              ask-EXPR-N.SG (FEM)             ‘request (dim)’                                       ‘ask (dim)’  (80)  a.  čud-ák                                        b.  čud-ač’-ók             eccentric-NOM.N.SG                               eccentric-NOM-EXPR.N.SG            ‘an eccentric’                                    ‘an eccentric (dim)’          c.* čud-ok-ak                                c.* čud-ok              eccentric-EXPR-NOM.N.SG                  eccentric-EXPR.N.SG             ‘an eccentric (dim)’                          ‘an eccentric (dim)’  The data in (81)–(82) again show that size suffixes cannot merge with √Roots. Examples (81a) and (82a) show adjectives with the category-forming adjectival suffix -n. If size suffixes were able to merge with √Roots, they would appear inside the suffix -n. However, this is not the case. In (81b–d) and (82b–d), size suffixes merge inside -n, which produces ungrammatical forms.  (81)   a.  žád-n-ij                                            b.* žad’-išč’-n-ij               stingy-ADJ-MASC.N.SG                                  stingy-EXPR-ADJ-MASC.N.SG               ‘stingy’                                                  ‘stingy (aug)’          c. * žad-ok-n-ij                                       d.* žad’-ic-n-ij                stingy-EXPR-ADJ-MASC.N.SG                      stingy-EXPR-ADJ-MASC.N.SG               ‘stingy (dim)’                                        ‘stingy (dim)’  (82)  a.  gr’áz-n-ij                                         b.* gr’az-n-išč’-ij               dirty-ADJ-MASC.N.SG                                    dirty-EXPR-ADJ-MASC.N.SG               ‘dirty’                                                    ‘dirty (aug)’         c. * gr’az-ok-n-ij                                    d.* gr’az’-ic-n-ij                dirty-EXPR-ADJ-MASC.N.SG                        dirty-EXPR-ADJ-MASC.N.SG                ‘dirty (dim)’                                          ‘dirty (dim)’  To summarize, Russian size suffixes can merge with nouns, but they cannot merge with adjectives, verbs, or √Roots (Table 5.9).  143 EXPRsize EXPRsize + n EXPRsize + a EXPRsize + v EXPRsize + √Root -k (-ok, -ek, -ik); -c (-ec, -ic); -išč’  * * * Table 5.9: Size suffixes (attachment to n, a, v, and √Roots)  5.4. INTERMEDIATE SUMMARY To summarize the findings in §5.2 and §5.3, the following differences in attachment sites of attitude and size suffixes emerge. Attitude suffixes can attach to √Roots and all syntactic categories (n/a/v), while size suffixes can only attach to nouns (Table 5.10).  Suffixes EXPR  + n EXPR + a EXPR + v EXPR + √Root EXPRattitude      EXPRsize   * * * Table 5.10: Attitude vs. size suffixes (difference in attachment sites)  Thus, we observe an asymmetry in attachment sites of syntactic heads (attitude suffixes) and syntactic modifiers (size suffixes) in Russian. Syntactic heads can merge at various sites (83), while syntactic modifiers can only merge at one site, namely with a noun category (84). This asymmetry will be discussed in more detail in Chapter 6, where I will show that it is not an inherent property of modifiers that they cannot attach to √Roots.  (83)   HEADS     a.          n                             b.                n                                  2                                     2                                n           √Root                              n            n/a/v                          EXPRattitude                                EXPRattitude  2                                                                                       n/a/v          √Root  (84)   MODIFIER             n                                  2                      EXPRsize        n                                          2                                           n           √Root  144 5.5. CO-OCCURRENCE RESTRICTIONS The current analysis makes the following predictions. First, since attitude suffixes are noun heads that can merge with √Roots, and size suffixes are noun modifiers, the latter should be able to merge with nouns formed by attitude suffixes. In other words, the sequence √Root+attitude+size suffix should be grammatical (85a). In contrast, the sequence *√Root+size+attitude suffix should be ungrammatical, since size suffixes cannot merge with √Roots (85b).  (85)                      a.               n                               b.                *n                                      2                                          2                             EXPRsize         n                                         n          √Root                                                  2                      EXPRattitude     2                                               n          √Root                             EXPRsize          √Root                                    EXPRattitude  In this section I show that these predictions are borne out. In (86b), the attitude suffix -ul’ merges with the √Root kras- ‘pretty’. In (86c), the size suffix -k (allomorphs: -ok, -ek, -ik) merges outside the attitude suffix; the resulting example is grammatical. In (86d), the attitude and size suffixes appear in reverse order, with the size suffix merging inside the attitude suffix. The resulting form is ungrammatical.  (86)  a.  krás-n-ij                                          b.  kras-úl’-a               pretty-ADJ-MASC.N.SG                                pretty-EXPRattitude-N.SG (MASC/FEM)               ‘pretty (archaic)’                                 ‘pretty animate (affect)’          c.  kras-úl’-k-a               pretty-EXPRattitude-EXPRsize-N.SG (MASC/FEM)                 ‘small pretty animate (affect+dim)’   d.* kras-(o/e/i)k-ul’-a   pretty-EXPRsize-EXPRattitude-N.SG (MASC/FEM)   ‘small pretty animate (dim+affect)’  The structures for (86c) and (86d) are shown in (87) and (88), respectively. In (87), the attitude suffix -ul’ merges with the √Root kras-, forming a noun kras-úl’-a ‘pretty animate (affect)’. The size suffix -k merges with a categorized noun.  145 (87)                                        n                                  kras-úl’-k-a ‘small pretty animate’                                      2                                    -k            n                             kras-úl’-a      ‘pretty animate’                                                  2                                               n             √kras-                                          -ul’  In (88), the size suffix -k merges with the √Root kras- first, and then the attitude suffix -ul’ attaches. This produces an ungrammatical form.  (88)                                    *n                               *kras-(o/e/i)k-ul’-a ‘small pretty animate’                                        2                                       n              √kras-                                  -ul’          2                                                  -k           √kras-  More examples are given in (89) and (90) to illustrate this behaviour of attitude and size suffixes.  (89)  a.  vr-á-t’                                             b.  vr-úš-a             lie-TH-INF                                                       lie-EXPR-N.SG (MASC/FEM)             ‘to lie’                                                ‘liar (affect)’          c.  vr-úš-k-a                                       d.* v(o)r-(o/e/i)k-uš-a              lie-EXPR-EXPR-N.SG (MASC/FEM)        lie-EXPR-EXPR-N.SG (MASC/FEM)             ‘small liar (affect+dim)’                     ‘small liar (dim+affect)’  (90)  a.  sm’éx                                              b.  sm’ex-ot-á             laughter.N.SG (MASC)                                 laughter-EXPR-N.SG (FEM)            ‘laughter’                                            ‘laughter (vulg)’          c.  sm’ex-ot-íšč’-a                               d.* sm’ex’-išč’-ot-a              laughter-EXPR-EXPR-N.SG (FEM)          laughter-EXPR-EXPR-N.SG (FEM)            ‘laughter (vulg+aug)’                           ‘laughter (aug+vulg)’  Another prediction of the current analysis concerns the co-occurrence of more than one size suffix. Since size suffixes are syntactic modifiers, they should allow “repeated application” (the term used by Scalise 1984:133). In other words, repetition of the same size morpheme 146 should be grammatical in Russian (91a). In contrast, attitude suffixes are syntactic heads and therefore, they should not allow for repetition of the same morpheme (91b).  (91)           a.                 n                                b.                  *n                               2                                             2                       EXPRsize        n                                          n               n                                           2                          EXPRattitude     2                                   EXPRsize       n                                       n              √Root                                                                                        EXPRattitude  Here I show that this prediction is borne out. Expressive modifiers allow repeated application, while expressive heads do not. In (92b), the diminutive size suffix -ek merges with the noun ovrág ‘ditch’. And in (92c), the same size suffix is used again. The result is grammatical, with a meaning of ‘very small ditch’. As expected (Chapter 2), the diminutive meaning is intensified.  (92)  a.  ovrág                                                  b.  ovráž-ek             ditch.N.SG (MASC)                                                  ditch-EXPR.N.SG (MASC)            ‘ditch’                                                          ‘small ditch (dim)’         c.  ovráž-eč’-ek17            ditch-EXPR-EXPR.N.SG (MASC)           ‘very small ditch (dim+dim)’  The structure for (92c) is given in (93). The size suffix -ek merges with the noun ovrág ‘ditch’. Then the second -ek merges with the noun+-ek complex.   (93)                                       n                              ovráž-eč’-ek ‘very small ditch (dim+dim)’                                      2                                  -ek             n                         ovráž-ek ‘small ditch (dim)’                                                  2                                               -ek           n                                                            4                                                                                 ovrag              ovrág ‘ditch’   17  In this word there are phonological k~č’ and g~ž alternations discussed in Chapter 3.  147 In contrast to size suffixes, attitude suffixes do not allow repeated application. In (94b), the attitude suffixes -ul’ merges with the noun kras-ot-á ‘prettiness/beauty’. In (94c), the suffix -ul’ is merged again, which produces an ungrammatical word.  (94)  a.  kras-ot-á                                          b.  kras-ot-úl’-a              pretty-NOM-N.SG (FEM)                                pretty-NOM-EXPR-N.SG (MASC/FEM)             ‘prettiness/beauty’                                 ‘pretty animate (affect)’           c. * kras-ot-ul’-ul’-a               pretty-NOM-EXPR-EXPR-N.SG (MASC/FEM)              ‘pretty animate (affect+affect)’  The structure for (94c) is given in (95).  (95)                                      *n                          *kras-ot-ul’-ul’-a ‘pretty animate (affect+affect)’                                      2                                     n            n                            kras-ot-úl’-a      ‘pretty animate (affect)’                                      -ul’      2                                               n              n                    kras-ot-á ‘prettiness/beauty’                                         -ul’           2                                                       n           √kras-                                                      -ot  To summarize, the current analysis correctly predicts the co-occurrence restrictions of attitude and size affixes. First, attitude suffix can merge inside size suffixes (96a), while size suffixes cannot merge inside of attitude suffixes (96b).  (96)                      a.               n                               b.                *n                                      2                                          2                             EXPRsize         n                                         n          √Root                                                  2                      EXPRattitude     2                                               n          √Root                             EXPRsize          √Root                                    EXPRattitude  Second, size suffixes allow repetition of the same morpheme (97a), while attitude suffixes do not (97b).  148 (97) a. MODIFIER        n                               b. HEAD         *n                               2                                             2                       EXPRsize        n                                          n               n                                           2                          EXPRattitude     2                                   EXPRsize       n                                       n              √Root                                                                                        EXPRattitude  5.6. CONCLUSIONS Inasmuch as the proposed analysis is successful, we have seen evidence for the two sites of word formation: (i) words are formed from √Roots, and (ii) words are formed from categories. Attitude suffixes in Russian clearly illustrate this point, as they productively create nouns from both √Roots (98a), and categories (98b).  (98) HEADS      a.           n                               b.             n                                  2                                     2                                n          √Root                               n            n/a/v                             EXPRattitude                               EXPRattitude  2                                                                                             n/a/v          √Root  Unlike attitude suffixes that can merge at two different sites, size suffixes can merge only at one site, namely with a noun category (99).  (99)  MODIFIER              n                                  2                   EXPRsize         n                                          2                                            n           √Root  149 Chapter 6: Expressive morphology across languages  Under the current analysis, the syntax of Russian expressive suffixes varies across two dimensions: (i) how the suffixes merge (as a head or as a modifier) and (ii) where they merge (with √Roots or with categories). With respect to the locus of merge, I argued that attitude suffixes are syntactic heads (1a), while size suffixes are syntactic modifiers (1b).  (1) a. HEAD                      X                        b. MODIFIER               Y                              2                                                2                            X            Y                                              X          Y                             EXPRattitude                                                 EXPRsize  With respect to the mode of merge, I argued that attitude suffixes are nominalizers that can merge either with √Roots (2a) or with categories (2b). In contrast, size suffixes are modifiers that can only adjoin to categorized nouns (3).  (2) a. HEAD + √ROOT                       n                b. HEADS + n/a/v                n                                         2                                                 2                                       n          √Root                                         n          n/a/v                                        EXPRattitude                                               EXPRattitude   (3)  MODIFIER + n                                                  n                                                                          2                                                              EXPRsize      n  Thus, one distributional difference between attitude and size suffixes in Russian is that attitude suffixes are noun heads, while size suffixes are noun modifiers (Table 6.1).   Suffixes HEADS MODIFIERS EXPRattitude -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’,  -úš, -ág, -ák, -ál, -án, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx a * EXPRsize -k/-ek/-ok/-ik; -c/-ec/-ic; -išč’  * a Table 6.1: Distributional difference #1 between attitude and size suffixes in Russian  150 Another distributional difference is that attitude suffixes can merge with √Roots and with categories (n/a/v), while size suffixes can only merge with a noun category (Table 6.2).   Suffixes Merge with √Roots Merge with categories EXPRattitude -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’,  -úš, -ág, -ák, -ál, -án, -ár, -áx, -íl, -in, -ób, -ot, -óx, -úg, -úk, -úx a  a EXPRsize -k/-ek/-ok/-ik; -c/-ec/-ic; -išč’  * a Table 6.2: Distributional difference #2 between attitude and size suffixes in Russian  There are two generalizations about Russian expressive suffixes that emerge here. First, the same set of syntactic heads can merge both with √Roots and with categories. Second, modifiers can only merge with categories, and thus there are no √Root modifiers in Russian (Table 6.3).   HEADS MODIFIERS Merge with √Roots ? Merge with categories -án’, -áš, -ón, -úl’, -ún’, -úr, -ús’, -úš, -ág, -ák, -ál, -án, -ár, -áx, -íl,  -in, -ób, -ot, -óx, -úg, -úk, -úx -k/-ek/-ok/-ik; -c/-ec/-ic; -išč’ Table 6.3: Asymmetries between attitude and size suffixes in Russian  The current analysis raises the following questions: i)  Can we find syntactic heads that can only merge with √Roots and others that can only      merge with categories? ii)  Is the logical possibility that is missing in Russian (expressive modifiers merging with √Roots) accidental or systematic? If it is accidental, we predict it should be attested cross linguistically. iii)  Are expressives restricted to the lower nominal domain (√Roots and nouns)? There is no principled reason why this should be the case, and thus we predict that expressives should be found in the higher nominal domain (Number and Determiner).  151 I will address each of these questions in turn. This chapter is structured as follows. In §6.1, I deal with expressive heads that can only merge with nouns. In §6.2, I deal with expressive modifiers merging with √Roots. In §6.3, I show that expressive morphology is attested in the higher nominal domain. In §6.4, I discuss evidential markers that are closely related to expressives. Finally, in §6.5, I present the conclusions.  6.1. EXPRESSIVE HEADS THAT CAN ONLY MERGE WITH NOUNS In this section, I address the first question: Can we find syntactic heads that can only merge with √Roots and others that can only merge with categories? The current analysis leads us to expect a two-way split in syntactic heads: (i) heads that merge with √Roots and (ii) heads that merge with categories.  Here I show that the two-way split is indeed attested across languages. It has been shown that German diminutive suffixes are heads that can only merge with the noun category (Wiltschko 2006). Empirical evidence is drawn from the Standard German diminutive suffix -chen and the colloquial Austrian diminutive suffix -erl. Examples with both suffixes are given in (4) and (5).  (4) German diminutives -chen and -erl    a. Baum b.  Bäum-chen (standard German) c.  Baum-erl (colloquial Austrian)   tree  tree-DIM   tree-DIM   ‘tree’       ‘(cute) little tree’  ‘(cute) little tree’ (Wiltschko & Steriopolo 2007:1)  (5) German diminutives -chen and -erl       a. Flasche     b.  Fläsch-chen (standard German) c. Flasch-erl (colloquial Austrian)   bottle       bottle-DIM  bottle-DIM             ‘bottle’       ‘(cute) little bottle’      ‘(cute) little bottle’ (Wiltschko & Steriopolo 2007:1)  That German diminutive suffixes are syntactic heads is evidenced by the following facts. First, they determine the grammatical gender of the resulting category. Diminutive suffixes 152 always form neuter nouns, independent of the gender of the base. For example, in (6a), the noun Baum ‘tree’ is masculine. In (6b) and (6c), the diminutive suffixes -chen and -erl form neuter nouns.  (6) masc  neuter        a.  der               Baum           b.  das             Bäum-chen           DET.MASC   tree                 DET.NEUT   tree- DIM           ‘tree’                                  ‘(cute) little tree’   c.  das             Baum-erl   DET.NEUT   tree-DIM        ‘(cute) little tree’  (Wiltschko & Steriopolo 2007:2)  (7) fem  neuter        a.  die             Flasche         b.  das              Fläsch-chen           DET.FEM    bottle     DET.NEUT   bottle-DIM           ‘bottle’                               ‘(cute) little bottle’   c.  das             Flasch-erl       DET.NEUT   bottle-DIM        ‘(cute) little bottle’ (Wiltschko & Steriopolo 2007:2)  Second, German diminutive suffixes appear to function as classifiers: they can turn a mass noun into a count noun. For example, in (8a), the noun Wein ‘wine’ is a mass noun. In (8b) and (8c), the diminutive suffixes -chen and -erl are added, forming count nouns: Wein-chen and Wein-erl ‘portion of wine’.  (8) mass noun  count noun       a.  viel      Wein                 b.  viele         Wein-chen           much    wine                     many.PL    wine-DIM           ‘much wine (mass)’          ‘many portions of wine (count)’    c.  viele         Wein-erl    many.PL    wine-DIM    ‘many portions of wine (count)’  153 To summarize, German diminutive suffixes behave like syntactic heads because they change the formal properties of the base (Wiltschko 2006), as illustrated in (9).  (9)     HEAD                      X                              2                            X            Y                        German DIM  With respect to the locus of merge, German diminutive suffixes attach to categorized nouns. Evidence is drawn from diminutive suffixes that appear outside of nominal morphology. For example, in (10b), the nominal suffix -nis is added to the adjective ge-heim ‘secret (adj)’, creating a noun Ge-heim-nis ‘secret’. In (10c), the diminutive suffix -chen is added outside the nominal suffix, forming a diminutive noun Ge-heim-nis-chen ‘(cute) little secret’. In (10d), the diminutive suffix is added inside the nominal suffix, which produces the ungrammatical form *Ge-heim-chen-nis.  (10)  a.  ge-heim                           b.  Ge-heim-nis             PREF-secret                          PREF-secret-NOM            ‘secret (adjective)’                ‘secret (noun)’          c.  Ge-heim-nis-chen           d.* Ge-heim-chen-nis                PREF-secret-NOM-DIM               PREF-secret-DIM-NOM            ‘(cute) little secret’                ‘(cute) little secret’  The fact that the German diminutive suffix appears outside and not inside of nominal morphology indicates that it attaches after a nominal category has been formed (11).  (11)                                  n2                                 Ge-heim-nis-chen ‘small secret’                                     2                                      n2            n1                         Ge-heim-nis ‘secret’                               -chen     2                                              n1             a                                           -nis           4                                                      ge-heim  The data above illustrate that German diminutive suffixes attach to nouns. But is this the only category that diminutive suffixes attach to? The examples in (12)–(13) illustrate that the 154 answer to this question is affirmative: German diminutive suffixes can only merge with nouns. In (12), the diminutive suffixes -chen and -erl are added to the verb les-en ‘to read’. The resulting examples (12b) and (12c) are ungrammatical. In (13), the diminutive suffixes are added to the adjective schön ‘beautiful’. The resulting examples (13b) and (13c) are also ungrammatical.  (12)  a.  les-en                        b.* les-chen                             c.* les-erl             read-INF     read-DIM                                          read-DIM            ‘to read’                         ‘to read (diminutive)’           ‘to read (diminutive)’  (13)  a.  schön                     b.* schön-chen                       c.* schön-erl             beautiful  beautiful-DIM                                  beautiful-DIM            ‘beautiful’                      ‘beautiful (diminutive)’          ‘beautiful (diminutive)’ (Wiltschko & Steriopolo 2007:5)  To summarize, like Russian attitude suffixes, German diminutive suffixes are syntactic heads. But unlike Russian expressive suffixes, they can only merge with a noun category (14).  (14)  HEAD                       n2                                     2                                 n2              n1                          German DIM  Now we can fill in the first gap in the classification of expressives: syntactic heads that can only merge with a noun category are attested in German (Table 6.4).   HEADS Merge with nouns German DIM suffixes (-chen, -erl) Table 6.4: Syntactic heads that can only merge with a noun category (German)  6.2. EXPRESSIVE MODIFIERS MERGING WITH √ROOTS Here I discuss the second question: Is the logical possibility that is missing in Russian (expressive modifiers merging with √Roots) accidental or systematic? I show that it is 155 attested cross linguistically, which points to the conclusion that the gap in the classification of Russian expressives is accidental.  It has been argued in the literature that Halkomelem diminutive prefixes (formed by means of reduplication) are √Root modifiers (Wiltschko 2008). Examples are given in (15)–(16).  (15) Halkomelem diminutive reduplication         a.  q’á:mi                         b.  q’á-q’emi                girl                                    DIM-girl              ‘girl’                                ‘small girl’      (Wiltschko 2008)  (16) Halkomelem diminutive reduplication          a.  músmes                       b.  mú-mesmes             cow                                 DIM-cow            ‘cow’                               ‘small cow’  (Wiltschko 2008)  Diminutive prefixes in Halkomelem never change the formal properties of the base (Wiltschko 2008). For example, in contrast to the German diminutive suffixes discussed above, they never turn mass nouns into count nouns (17)–(18).  (17) count noun = count noun          a.  s-páth                             b.  s-pi-páth             NOM-bear                            NOM-DIM-bear            ‘bear (count noun)’            ‘small bear (count noun)’  (18) mass noun = mass noun          a.  s-peháls                          b.  s-pi-peháls             NOM-wind                            NOM-DIM-wind             ‘wind (mass noun)’           ‘little bit of wind (mass noun)’ (Wiltschko & Steriopolo 2007:3)  Since Halkomelem diminutive prefixes do not produce a change in the formal properties of the base, they act as syntactic modifiers (19).  156 (19) MODIFIER              Y                          2                        X            Y                    Halkomelem DIM  With respect to the locus of merge, Halkomelem diminutive prefixes adjoin to category-free √Roots. The evidence stems from diminutives that appear inside nominal morphology. For example, in (20b), the diminutive reduplicant qí- is inside the nominal prefix -s, forming a diminutive noun s-qí-qewàth ‘small rabbit’. In contrast, in (20c), qí- is merged outside the nominal prefix, and the resulting form *qí-s-qewàth is ungrammatical.  (20)  a.  s-qewáth                   b.  s-qí-qewàth c.* (s)-qí-s-qewàth             NOM-rabbit                    NOM-DIM-rabbit       NOM-DIM-NOM-rabbit            ‘rabbit’                           ‘small rabbit’       ‘small rabbit’  The fact that the Halkomelem diminutive prefix appears inside and not outside nominal morphology indicates that it attaches before nouns are formed (21).  (21)                                 n                                   2                                 n          √qewath                        s-qí-qewàth ‘small rabbit’                             s-        2                                             qi-         √qewath  The structure (21) makes the following prediction: since diminutives modify √Roots before they become nouns, they should also be able to modify √Roots that are later categorized as verbs or adjectives. This prediction is borne out. The data in (22) and (23) show that diminutive prefixes in Halkomelem can also modify verbs and adjectives.  (22)  a.  lhí:m b.  lhi-lhi:m              pick                                 DIM-pick              ‘to pick’                          ‘to pick a little bit’  (23)  a.  p’eq’ b.  p’í-p’eq             white                                DIM-white             ‘white’                             ‘a little bit white’  157 To summarize, Halkomelem diminutives are syntactic modifiers that merge with category- free √Roots (Wiltschko 2008), as illustrated in (24).  (24) MODIFIER                      √Root                                      2                        Halkomelem DIM        √Root  Thus, both types of expressive morphology missing in Russian are found cross-linguistically: (i) in German, diminutive heads can only merge with a noun category, and (ii) in Halkomelem, diminutive modifiers merge with category-free √Roots (Table 6.5).   HEADS MODIFIERS Merge with √Roots Russian Halkomelem Merge with nouns German, Russian Russian Table 6.5: Expressive morphology in Russian, German, and Halkomelem  6.3. EXPRESSIVE MORPHOLOGY IN THE HIGHER NOMINAL DOMAIN Here I discuss the third question: Are expressives restricted to the lower nominal domain (√Roots and nouns)? So far we have been dealing with expressive morphology merging either with √Roots or with nouns (25).  (25)                                 D                                 3                                 D                     #                                            3                                          #                  n                                                      3                    EXPR                           n                √Root   However, nothing in the analysis forces the expressives to be low; rather, I have shown that there is no unified syntax of expressives. This finding leads us to expect that expressives might also be found in the higher nominal domain: merging with functional categories in the nominal extended projection such as Number (#) and Determiner (D) (Abney 1987; 158 Grimshaw 1991; Ritter 1995, among others), as illustrated in (26). Here I show that this is indeed the case. Empirical evidence will be drawn from the following languages: Brazilian Portuguese (Romance), Southern Barasano (an Eastern Tucanoan language of Colombia), Welsh (Celtic), and Tongan (Polynesian).  (26)                                   D                                   3                                 D                  #                                             3                                            #                   n EXPR                                                3                                                       n                √Root    6.3.1. Expressive morphology merging with Number Consider first expressive morphology in Brazilian Portuguese. It has been argued that the diminutive /ziɲ/ is adjoined to the category Number in Brazilian Portuguese (Bachrach & Wagner 2007). The evidence comes from the fact that /ziɲ/ has the same distribution as other linguistic objects that modify Number, namely coordinate compounds. To illustrate this, I first describe coordinate compounds and then I will compare them to the diminutive /ziɲ/.  In coordinate compounds, the second conjunct always agrees with the first one (Bachrach & Wagner 2007). For example, in (27) the second conjunct agrees with the first one in gender. In (27a), the first conjunct professor-a ‘teacher’ is feminine, and the second one vampire-a ‘vampire’ is also feminine. In (27b), professor ‘teacher’ is masculine, and vampire-o ‘vampire’ is also masculine (compare with (28), where there is no agreement in gender and the examples are ill-formed).  (27) Brazilian Portuguese coordinate compounds      a.  profesor-a       vampir-a                  b.  profesor            vampir-o          teacher-FEM   vampire-FEM                  teacher.MASC    vampire-MASC            ‘vampire teacher (FEM)’                        ‘vampire teacher (MASC)’ (Bachrach & Wagner 2007:3) 159 (28) Brazilian Portuguese (no agreement in gender)18      a. * profesor-a       vampir-o               b.* profesor           vampir-a           teacher-FEM    vampire-MASC                   teacher.MASC   vampire-FEM             ‘vampire teacher’                                 ‘vampire teacher’  The second conjunct also agrees with the first one in number. For example, in (29a), jornaw ‘newspaper’ is singular and livro ‘book’ is also singular. In (29b), jornaj-s ‘newspapers’ is plural and livro-s ‘books’ is also plural (compare with (30), where there is no agreement in number and the examples are ill-formed).  (29) Brazilian Portuguese coordinate compounds        a.  jornaw             livro                    b.  jornaj-s              livro-s           newspaper.SG  book.SG                         newspaper-PL    book-PL           ‘newspaper-booklet (SG)’                 ‘newspaper-booklets (PL)’ (Bachrach & Wagner 2007:9)  (30) Brazilian Portuguese (no agreement in number)         a. * jornaw              livro-s                 b.* jornaj-s              livro            newspaper.SG   book-PL                           newspaper-PL     book.SG           ‘newspaper-booklet(s)’                       ‘newspaper-booklet(s)’  Based on the observation that plural is marked on both conjuncts, Bachrach & Wagner (2007) propose that coordinate compounds are a case of adjunction to Number, as illustrated in (31). The first Number node is the host and the second one is an adjunct. This representation of coordinate compounds accounts for the descriptive generalization that each compound receives separate but agreeing number marking.19   18  Thanks very much to Maria Amelia Reis Silva for help with the data in (28) and (30). 19  In this structure, I ignore the representation of theme vowels. See Bachrach & Wagner (2007:3) for an analysis of theme vowels in Brazilian Portuguese. 160 (31)                                     #1                                    3                             #1                                       #2                     3           3                    n                  #1       n                   #2                  4                 g     4                 g                  jornaw               s     livro                  s  Now consider the diminutive marker /ziɲ/. According to Bachrach & Wagner (2007), /ziɲ/ behaves just like the second conjunct in a coordinate compound: it agrees with the first conjunct in gender and number. For example, in (32a), the first conjunct amig-a ‘friend’ is feminine and the diminutive zĩɲ-a ‘small’ is also feminine. In (32b), amig-o ‘friend’ is masculine and the diminutive zĩɲ-o ‘small’ is also masculine.  (32) Brazilian Portuguese diminutive /ziɲ/        a. amig-a          zĩɲ-a                         b.  amig-o             zĩɲ-o           friend-FEM     small-FEM                     friend-MASC    small-MASC             ‘small friend (FEM)’                           ‘small friend (MASC)’ (Bachrach & Wagner 2007:4)  Example (33) shows that /ziɲ/ also agrees with the first conjunct in Number. In (33a), jornaw ‘newspaper’ is singular and the diminutive is singular. In (33b), jornaj-s ‘newspapers’ is plural and the diminutive is also plural.20  (33) Brazilian Portuguese diminutive /ziɲ/        a.  jornaw-zĩɲ-o                                   b.  jornaj-s-ĩɲ-o-s           newspaper.SG-book-MASC.SG                  newspaper-PL-small-MASC-PL           ‘small newspaper (SG)’                       ‘small newspapers (PL)’ (Bachrach & Wagner 2007:9)  Since the distribution of /ziɲ/ is identical to the distribution of coordinate compounds, Bachrach & Wagner argue that the diminutive /ziɲ/ has the same structure as coordinate  20  See Bachrach & Wagner (2007:10) on the phonological alternations in this example.  161 compounds. In the structure in (34), the first conjunct is the host and the diminutive is an adjunct. Thus, the diminutive /ziɲ/ merges with Number in Brazilian Portuguese.  (34)                                    #1                                    3                               #1                                     #2                       3          3                        n                  #1       n                  #2                    4                g      4                g                    jornaj                s     /ziɲ/-o              s  Another example of a diminutive merging with Number is found in Stump (1993). In Southern Barasano, the diminutive suffix -aka attaches outside the plural morphology. In (35b), -ri is a plural suffix (cotɨ-ri ‘pot-PL’). In (35c), the diminutive attaches outside this plural suffix (cotɨ-ri-aka ‘pot-PL-DIM’). This means that the plural is formed first, and only after that is the diminutive formed.  (35) Southern Barasano diminutive -aka        a.  cotɨ                       b. cotɨ-ri                 c.  cotɨ-ri-aka           pot.SG                                   pot-PL                             pot-PL-DIM           ‘pot’                             ‘pots’                         ‘small pots’ (Stump 1993:7)  A similar distribution of diminutives is found in Welsh (Celtic). Like in Southern Barasano, Welsh diminutive suffixes -os (endearment) and -ach (contempt) attach after the plural is formed (Stump 1993). For example, in (36b), the plural is formed with the plural suffix -ed (merch-ed ‘girl-PL’). In (36c), the diminutive -os attaches outside the plural suffix (merch-ed- os ‘girl-PL-DIM’). This means that diminutivization is introduced after pluralization. An example with the diminutive -ach, which shows the same distribution as the diminutive -os, is given in (37).  (36) Welsh diminutive -os        a.  merch                       b.  merch-ed               c.  merch-et-os           girl.SG                            girl-PL                        girl-PL-DIM           ‘girl’                             ‘girls’                         ‘girls (with endearment)’  162 (37) Welsh diminutive -ach        a.  pryf                           b.  pryf-ed                   c.  pryf-et-ach           worm.SG                        worm-PL                     worm-PL-DIM           ‘worm’                          ‘worms’                      ‘worms (with contempt)’          (Stump 1993:7)  To summarize, in three unrelated languages––Brazilian Portuguese, Southern Barasano, and Welsh––expressive morphology merges with Number.  6.3.2. Expressive morphology merging with Determiner Finally, we also find expressive morphology merging with Determiners. In Tongan, there are two sets of determiners: ordinary and emotional (Churchward 1953; Hendrick 2005). Ordinary determiners are definite (he, e) or indefinite (ha). Emotional determiners, in addition to being definite (si‘i) or indefinite (si‘a), express affection, friendship, pity, or some other positive emotion (Churchward 1953:23) (Table 6.6).   ORDINARY EMOTIONAL DEFINITE  he, e si‘i INDEFINITE ha si‘a Table 6.6: Tongan determiners (from Hendrick 2005:908)  The emotional determiners are always speaker-oriented and express the speaker’s assessment (Hendrick, personal communication). For example, in (38), the definite emotional determiner si‘i expresses the speaker’s pity toward the horse. In (39), the indefinite emotional determiner si‘a expresses the speaker’s sympathy toward the child.21  (38)    ‘Oku    hela   ‘a         si‘i                        hōsí.            PRES   tired   PREP    DEFIN.EMOT.DET  horse           ‘The (poor) horse is tired (pity)’ (Churchward 1953:23)  21  When si‘i is used in a postposition, it behaves like an adjective and means ‘small’:  (i) ha               konga   si‘i          INDEF.DET  part       small         ‘A small part’  (Churchward 1953:24)   163  (39)    Kuo    lavea     si‘a                       tamasi‘i?            PERF   be.hurt  INDEF.EMOT.DET  child            ‘Has a child been hurt (sympathy)?’ (Churchward 1953:24)  Hendrick (2005) suggests that the emotional determiners express the speaker’s emotions toward the nouns they head. The proposed structures for si‘i hōsí ‘the (poor) horse (pity)’ and si‘a tamasi‘i ‘a child (sympathy)’ are given in (40) and (41), respectively. In (40), the definite emotional determiner si‘i heads the noun hōsí ‘horse’ and expresses pity toward the horse. In (41), the indefinite emotional determiner si‘a heads the noun tamasi‘i ‘child’ and expresses sympathy toward the child.  (40)                                     D                                     3                                      D                     #                                     si‘i          3                         DEFIN.EMOT       #                  n                                                                5                                                              hōsí  ‘horse’  (41)                                     D                                     3                                  D                     #                              si‘a            3                      INDEFIN.EMOT      #                  n                                                              5                                                      tamasi‘i  ‘child’  Thus, in Tongan, expressive morphology targets the determiner domain. The expressive markers are the emotional determiners si‘i (definite) and si‘a (indefinite) that express the positive emotion of the speaker toward the nouns headed by the determiners.  6.3.3. Summary To summarize, expressive morphology can be found not only in the low nominal domain (√Root and n), but also in the higher nominal domain (# and D), as illustrated in (42).  164 (42)                                        D                                        3              Tongan          D                       #                                                   3              BP22, Barasano, Welsh   #                 n EXPR                                                    3              Russian, German                    n                 √Root               Russian, Halkomelem  6.4. EVIDENTIALS So far we have seen expressive morphology in the nominal domain. The question arises: can it also be found in the verbal domain? In Russian, there is no expressive morphology in the verbal domain. However, evidential morphology has been widely attested across languages in the verbal domain. Under some views, evidentials and expressives are considered closely related as markers of epistemology (see Willett 1988 for a literature review). For example, Chafe (1986) views evidentials in the broad sense as epistemology markers that code the speaker’s attitude toward his/her knowledge of a situation, and in the narrow sense as marking the source of such knowledge. Compare with expressives which code the speaker’s attitude towards the referent (§2.2.3.1).  Evidentials “put in perspective or evaluate the truth value of a sentence both with respect to the source of the information … and with respect to the degree to which this truth can be verified or justified” (Rooryck 2001:125). Under this view, evidentials indicate both the source and reliability of the information23. Here we can understand reliability as evaluation of the information source by the speaker. In this respect, expressives and evidentials are viewed as similar, as they both involve evaluation. They are, however, different with respect to what is being evaluated: evidentials evaluate the source of information, while expressives evaluate the referent. Examples of evidentials are given below. In (43), seems indicates that the source of information is sensory, while looks and sounds indicate that the source of information is visual and auditory, respectively.  22  BP = Brazilian Portuguese (abbreviation from Bachrach & Wagner 2007). 23  See Blain & Déchaine (2006) on a different view: evidentials only code the source of information. 165  (43)  a.  Sally seems happy.                Source: sensory          b.  Sally looks happy.                 Source: visual          c.  Sally sounds happy.               Source: auditory (Blain & Déchaine 2006:12)  Under the Evidential Domain Hypothesis (Blain & Déchaine 2006), evidential markers are a heterogeneous class and can merge in different syntactic positions. They can merge with the proposition (C), tense (T), aspect (Asp), and the predicate (v) (44).  (44)                                     C                                     3                                    C                 T                                               3                                            T                   Asp EVID                                                   3                                                        Asp                v                                                                     3                                                                   v                   D   Empirical evidence for the Evidential Domain Hypothesis is drawn from the following languages: Quechua, Cherokee, Turkish, Lillooet, Cree, and Tariana. Let us discuss evidential markers in some of these languages.  In Quechua, some evidential markers (-mi, -si, -chá) are part of the proposition-marking system, while others (-rqa, -sqa) are part of the tense-marking system (Blain & Déchaine 2006). The proposition-marking evidential markers are part of the focus-marking system and are in complementary distribution with other focus markers. The evidential marker -mi codes direct evidence, while the makers -si and -chá code indirect evidence. For example, in (45a), -mi indicates that the speaker has direct visual evidence of the rain. In (45b), -si indicates that the speaker has indirect evidence of the rain.  166 (45) Quechua evidentials  (C-domain)    a.  DIRECT    Para-sha-n-mi    rain-PROG-3-DIRC.EVID     ‘It is raining’              (Speaker has visual evidence)   b.  INDIRECT   Para-sha-n-si    rain-PROG-3-REPORT    ‘It is raining’    (Speaker has been told) (Faller 2002:25, 27)  The tense-marking evidential markers have a past tense orientation. The evidential marker -rqa codes direct evidence, while -sqa codes indirect evidence. For example, in (46a), -rqa indicates that the speaker had visual evidence of the rain. In (46b), -sqa indicates that the speaker did not experience the event.  (46) Quechua evidentials  (T-domain)   a.  DIRECT    Para-sha-rqa-n                         rain-PROG-PAST-3                        ‘It was raining’               (Speaker had visual evidence)   b.  INDIRECT   Para-sha-sqa     rain-PROG-NONEXP.PAST     ‘It was raining’     (Speaker did not experience the event)  (Faller 2003:20)  Let us now look at evidential markers in Cherokee. In Cherokee, there are two evidential markers: the suffix -ʌ?i, which marks direct evidence experienced first-hand by the speaker, and the suffix -e?i, which marks indirect evidence acquired indirectly by the speaker (Aikhenvald 2004). These evidential markers are part of the Tense system and, like the tense- marking evidential markers in Quechua, they have a past tense orientation (Blain & Déchaine 2006). For example, in (47a), -ʌ?i indicates that the speaker has visual evidence that a cat ran. In (47b), -e?i indicates that that the speaker had indirect evidence (i.e., was told) that he spoke.  167 (47) Cherokee evidentials (T-domain)   a.  DIRECT    wesa u-tlis-ʌ?i                       cat    3-run-1STHAND.PAST                       ‘A cat ran’ (Speaker saw it running)   b.  INDIRECT   u-wonis-e?i    3-speak-NON1STHAND.PAST    ‘He spoke’ (Speaker was told) (Blain & Déchaine 2006:15)  In Turkish, the so-called ‘perfect of evidentiality’ (Aikhenvald 2004) is part of the aspectual domain and is used to assert the necessary truth of the proposition with respect to the speaker’s knowledge state. For example, in (48a), the evidential marker -mus indicates that the speaker has been listening to someone play. In (48b), the evidential marker -ymIs indicates that the speaker has been told that the minister is sick.  (48) Turkish evidentials (Asp-domain)   a.  PERCEPTION iyi    cal-iyor-mus    good play-ASP-1STHAND.COP      ‘She is, as I hear, playing well’ (Speaker has been listening)   b.  REPORTATIVE  bakan      hasta-ymIs                           minister   sick-NON1STHAND.COP     ‘The minister is reportedly sick’ (Speaker has been told) (Blain & Déchaine 2006:16)  According to Blain & Déchaine (2006), in Lillooet (Salish), the ‘out-of-control’ morpheme ka-…-a is a direct evidence marker that is part of the predicate domain. In an alternative analysis, the morpheme ka-…-a is a modal (see Davis, Matthewson, & Rullmann, 2007). Blain & Déchaine (2006) suggest that the morpheme ka-…-a has a non-volitional ‘by- accident’ reading and corresponds to inaccessibility to consciousness. Consider the examples in (49). In (49a), the predicate sek-en-ás ‘hit-dir-3erg’ is volitional and means that the boy hit the ball on purpose (accessible to consciousness). In (49b), the predicate ka-sék-s-as-a with the ‘out-of-control’ marker ka-…-a has a non-volitional reading: the boy hit the ball accidentally (inaccessible to consciousness).  168 (49) Lillooet evidentials: direct evidence (v-domain)   a.  ACCESSIBLE TO CONSCIOUSNESS           sek-en-ás           ti      sq’úm’ts-a   ti     twéw’wet-a         hit-DIRC-3ERG    DET   ball-DET        DET  boy-DET           ‘The boy hit the ball’       b.  INACCESSIBLE TO CONSCIOUSNESS             ka-sék-s-as-a                   ti      sq’úm’ts-a    ti     twéw’wet-a          OOC-hit-CAUS-ERG-OOC   DET   ball-DET         DET   boy-DET           ‘The boy hit the ball accidentally’ (Demirdache 1997)  To summarize, evidential markers, like the closely related expressive morphemes, are a heterogeneous class and occupy different positions in a syntactic tree (50). This means that the behaviour of expressives is not special to expressives. Evidentials display the same behaviour, but in the verbal domain.  (50)                                        C                                        3              Quechua          C                    T                                                  3              Cherokee, Quechua     T                Asp EVID                                                    3              Turkish                                  Asp                 v                                                                     3                                                                   v                    D              Lillooet  6.5. CONCLUSIONS As expected, the typological gaps identified in the Russian expressive morphology are found cross-linguistically. Expressive heads that can only merge with nouns are found in German, while expressive modifiers that merge with category-free √Roots are found in Halkomelem.  The current analysis of expressive morphemes makes the following prediction. If expressive morphemes can merge in different syntactic positions (with √Roots and with n), we should also be able to find expressive morphology in the higher syntactic domain (merging with # and with D). This prediction is borne out cross-linguistically. 169 (51)                                        D                                        3              Tongan            D                     #                                                  3              BP, Barasano, Welsh    #                  n EXPR                                                   3              Russian, German                   n               √Root               Russian, Halkomelem  Finally, the proposed system of expressive morphemes is parallel to the system for evidential morphemes proposed in Blain & Déchaine (2006). In both systems, expressive and evidential morphemes merge not in one, but in different syntactic positions. The fact that evidentials are found in the clausal/verbal domain while expressives are found in the nominal domain might indicate that evidentials and expressives are not two different systems, but instead are parts of the same system, as illustrated in (52).  (52)                                      C                                      3              Quechua          C                  T                                                3              Cherokee, Quechua  T                  Asp EVID                                                   3              Turkish                                 Asp                v                                                                    3                                                                   v                  D              Lillooet                                                          3              Tongan                                                       D                     #                                                                                         3              BP, Barasano, Welsh                                               #                 n EXPR                                                                                          3              Russian, German                                                               n               √Root               Russian, Halkomelem   Another way to approach these findings is that evidentials and expressives are two distinct systems which are syntactically parallel. In this case, we expect to find evidentials in the nominal domain and expressives in the clausal/verbal domain. We might also expect to find 170 morphemes that are both markers of evidentiality and expressivity. The question whether evidentials and expressives are parts of the same system or parts of two different but parallel systems remains for further research.             171 Chapter 7: Conclusions and topics for further research  7.1. CONCLUSIONS I have argued that the syntax of Russian expressive suffixes varies across two dimensions: (i) how the suffix is merged (as a syntactic head or as an adjoined modifier), and (ii) where the suffix is merged (with category-free √Roots or with categories). I showed that attitude suffixes are noun heads that can either merge with √Roots (1a) or with syntactic categories n/a/v (1b). In contrast, size suffixes are noun modifiers that can only adjoin to a noun category (2).  (1) HEADS        a.           n                               b.             n                                  2                                     2                                n          √Root                               n            n/a/v                             EXPRattitude                               EXPRattitude  2                                                                                             n/a/v          √Root  (2)  MODIFIER                n                                  2                   EXPRsize         n                                          2                                            n           √Root  I investigated the functional and formal properties of Russian expressive suffixes in a systematic way, which, to the best of my knowledge, has not been done before. In doing so, I analyzed how expressive suffixes pattern along several kinds of criteria (gender/class change, category change, subcategorization). A byproduct of this analysis is that I showed how grammatical gender of an expressive form can be predicted from its inflectional class (combined with animacy and natural gender of the base).  One implication of this analysis is that expressives are not associated with special formal properties as opposed to non-expressives (descriptive linguistic objects). They have the same syntax as non-expressives that are distinguished on the basis of their syntactic types: head vs. modifier, and attachment to √Roots vs. categories. Another implication is that the formal 172 criteria developed in this thesis can be extended to set up a cross-linguisitc typology of expressives.  These findings have important implications for the form/function mapping in the realm of categorization. The problem of the diversity of grammatical categories within the Principles and Parameters framework is among the core issues of modern linguistic theory. How can we explain the tension between language diversity and language universals? Is the same semantic “concept” universally mapped onto the same syntactic category? I showed that even within a single language (Russian), the same function “expressive” does not map onto the same form. On the other hand, the two different semantic types (attitude vs. size) map directly onto the two different syntactic types (head vs. modifier). In view of this, we can raise the following question: what determines whether the form of a linguistic object is the same of different: its function or its semantic type?  7.2. TOPICS FOR FURTHER RESEARCH 7.2.1. Complex expressive suffixes In this thesis, I have not investigated complex expressive suffixes. Complex expressive suffixes look like sequences of simplex suffixes, with the diminutive -k as the last suffix of a sequence (Stankiewicz 1968:102).24  Examples of such suffixes are -ušk, -onk,-on’k/-en’k, and -išk. However, these suffixes are not exactly sequences of simplex suffixes, because what looks like the first suffix of a sequence cannot be used independently (without the diminutive -k) in the same word. For example, in (3), the complex suffix -ušk does not consist of two simplex suffixes -uš and -k, because -uš cannot be used independently in (3c). In contrast, the diminutive suffix -k can be used independently, but in this case, the meaning of the resulting word is different. The complex suffix -ušk has an affectionate meaning, while the simplex suffix -k has a diminutive meaning (3d). The data in (4) show the same behaviour with the complex suffix -onk.   24  Stankiewicz (1968) uses the term “compound” suffixes.  173 (3)  a.  golov-á                             b.  golóv-ušk-a           head-N.SG (FEM)                     head-EXPR-N.SG (FEM)           ‘head’                                    ‘head (affect)’        c. * golov-uš-a                      d.  golóv-k-a           head-EXPR-N.SG (FEM)          head-EXPR-N.SG (FEM)           ‘head (expr)’                        ‘head (dim)’  (4)  a.  ruk-á                                 b.  ruč’-ónk-a           hand-N.SG (FEM)                    hand-EXPR-N.SG (FEM)           ‘hand’                                   ‘hand (contempt)’        c. * ruč’-on-a                        d.  rúč’-k-a           hand-EXPR-N.SG (FEM)          hand-EXPR-N.SG (FEM)           ‘hand (expr)’                        ‘hand (dim)’  What makes complex expressive suffixes an interesting research topic is that they have different distributional properties from those of simplex expressive suffixes. To illustrate this, let us discuss the complex suffixes -on’k/-en’k and -išk. 7.2.1.1. The suffix -en’k/-on’k -en’k/-on’k is an affectionate suffix, where the distinction “e” vs. “o” is purely orthographic, just like for -ek/-ok (Shvedova et al. 1982:214). This is a complex suffix because what looks like the first suffix -en’/-on’ of the sequence -en’k/-on’k cannot be used independently, as illustrated in (5c). What looks like the second suffix -k of the sequence can be used independently, but with a different meaning: -en’k has an affectionate meaning, while -k has a diminutive meaning, as illustrated in (5d).  (5)  a.  ruk-á                                b.  rúč’-en’k-a           hand-N.SG (FEM)                    hand-EXPR-N.SG (FEM)           ‘hand’                                   ‘hand (affect)’       c. * ruč’-en’/on’-a                d.  rúč’-k-a           hand-EXPR-N.SG (FEM)          hand-EXPR-N.SG (FEM)           ‘hand (expr)’                        ‘hand (dim)’  The affectionate suffix -en’k/-on’k behaves differently compared to the affectionate simplex suffixes that have been described in this thesis: it can change the inflectional class of a noun, 174 but it can never change the syntactic category. This suffix can attach to nouns of all inflectional classes, always producing a Class II noun, as illustrated in (6)–(8).  (6)  a.  bóg                                              b.  bóž-en’k-a           God.N.SG (MASC; CLASS I)               God-EXPR-N.SG (MASC; CLASS II)             ‘God’                                              ‘God (affect)’  (7)  a.  dóč’                                              b.  dóč'-en’k-a           daughter.N.SG (FEM; CLASS III)        daughter-EXPR-N.SG (FEM; CLASS II)             ‘daughter’                                         ‘daughter (affect)’  (8)  a.  d’ád’-a                                          b.  d’ád’-en’k-a            uncle-N.SG (MASC; CLASS II)            uncle-EXPR-N.SG (MASC; CLASS II)              ‘uncle’                                              ‘uncle (affect)’  The fact that -en’k can change inflectional class for Class II can be accounted for if we assume that it is a syntactic head specified for Class II in its lexical entry (just as we did with simplex affectionate suffixes). However, simplex affectionate suffixes can change syntactic category: they attach to nouns, adjective, or verbs, every time forming a noun (Chapter 5). This is not the case with the complex affectionate suffix. As the data in (9)–(10) illustrate, -en’k cannot change syntactic category and does not form a noun from adjectives or adverbs (this suffix does not attach to verbs).25  (9)  a.  žád-n-ij                                     b.   žad-n’-en’k-ij           stingy-ADJ-MASC.N.SG                         stingy-ADJ-EXPR- MASC.N.SG           ‘stingy’                                           ‘stingy (affect)’         c. * žad-n’-en’k-a                          d.* žad-en’k-a             stingy-ADJ-EXPR-N.SG                         stingy-EXPR-N.SG             ‘stingy animate (affect)’               ‘stingy animate (affect)’  (10)  a.  bístr-o                                       b.  bístr’-en’k-o             quick-ADV.SUFF                                       quick -EXPR-ADV.SUFF             ‘quickly’                                        ‘quickly (affect)’   25  The diminutive suffix -k behaves the same way when added to the adverb n’e-mnóg-o ‘a little bit’: n’e-mnóž- k-o ‘a little bit (dim)’. To the best of my knowledge, this is the only case where the diminutive -k merges with an adverb and does not change syntactic category (see also Efremova 2006). This is a puzzling case that deserves further attention: perhaps investigation from a historical perspective can shed some light on this phenomenon. 175          c. * bistr’-en’k-a               quick-EXPR-N.SG               ‘quick animate (affect)’  With respect to the data above, the following questions arise: What are the morphosyntactic properties of the complex suffix -en’k/-on’k? Is it a syntactic head specified for Class II in its lexical entry? And if it is, why is it unable to change syntactic category? 7.2.1.2. The suffix -išk The affectionate suffix -išk is a complex suffix because what looks like the first suffix of the sequence, -iš, cannot be used independently (11c). Like in the examples above that contain the affectionate suffix -en’k/-on’k, what looks like the second suffix of the sequence, namely -k, can be used independently, but with a different meaning (11d).  (11)  a.  sín                                              b.  sin’-íšk-a             son.N.SG (MASC)                               son-EXPR-N.SG (MASC)               ‘son’                                                            ‘son (affect)’          c. * sin-iš-a                                      d.  sin-ók             son-EXPR-N.SG (MASC)                   son-EXPR.N.SG (MASC)             ‘son (expr)’                                   ‘son (dim)’  What makes this suffix interesting to investigate is that, unlike any simplex affectionate suffix discussed in this work, it produces a change in the inflectional class of a noun depending on the animacy of the base. For example, when it attaches to an animate noun, it changes its inflectional class to Class II and consequently, it forms a noun of common gender (MASC/FEM), as illustrated in (12). However, when it attaches to an inanimate noun, it does not change either inflectional class, or gender (13)–(14). When attached to an inanimate masculine noun, the resulting noun acquires the neuter ending -o (13b). This is similar to the distribution of the augmentative suffix -išč’ (Chapter 3). However, unlike the augmentative -išč’ which does not “care” whether a base is animate or inanimate, the suffix -išk is sensitive to the animacy of the base.  176 (12)  a.  vór                                               b.  vor’-íšk-a              thief.N.SG (MASC; CLASS I)                thief-EXPR-N.SG (MASC/FEM; CLASS II)                ‘thief’                                                            ‘thief (affect)’  (13)  a.  dóm                                             b.  dom’-íšk-o              house.N.SG (MASC; CLASS I)                house-EXPR-N.SG (MASC; CLASS I)                ‘house’                                                         ‘house (affect)’  (14)  a.  oxót-a                                           b.  oxót’-išk-a              hunt-N.SG (FEM; CLASS II)                    hunt-EXPR-N.SG (FEM; CLASS II)                ‘hunt’                                                            ‘hunt (affect)’  The data above raise two questions: What are the morphosyntactic properties of the complex suffix -išk? And what accounts for its sensitivity to animacy of the base? 7.2.2. Delimitative verbs and delimitative prefix po- The delimitative prefix po- is interesting to investigate in its relation to diminutives, as it always indicates short events. The so-called delimitative verbs describe events limited in time. These verbs are formed by means of a productive prefix po- (traditionally called ‘delimitative prefix’: Zalizniak & Shmelev 2000). For example, in (15a), the verb gul’-á-t’ ‘walk’ describes an event of walking. In (15b), the delimitative prefix po- is used and the resulting verb describes an event of walking that is limited in time. The same behaviour of the delimitative po- is shown in (16).  (15)  a.  gul’-á-t’                                      b.  po-gul’-á-t’             walk-TH-INF                                              DELIM.PREF-walk-TH-INF             ‘to walk’                                          ‘to walk for a while’  (16)  a.  p’-í-t’                                          b.  po-p’-í-t’             drink-TH-INF                                              DELIM.PREF-drink-TH-INF             ‘to drink’                                         ‘to drink for a while’  What is relevant to this work and interesting about delimitative verbs is that they always describe events that are short. Evidence for this stems from the fact that delimitative verbs can be used with the adjective n’emnógo ‘a little bit’, and cannot be used with the adjective mnógo ‘a lot’ (17).  177 (17)  a.  Iván   po-p’-í-l                                 n’emnógo  čáj-u             Ivan   DELIM.PREF-drink-TH-PAST     a little bit   tea-GEN            ‘Ivan drank a little bit of tea’          b.* Iván   po-p’-í-l                                 mnógo    čáj-u             Ivan    DELIM.PREF-drink-TH-PAST   a lot         tea-GEN            ‘Ivan drank a lot of tea’  The fact that the delimitative prefix po- indicates a short event makes it similar in meaning to the diminutive suffixes described in this work that indicate the small size of a referent. In both cases, it is smallness of event/referent that is being indicated. In this respect, the following question arises: since the delimitative prefix po- and diminutive suffixes are similar semantically, does it mean that they are also similar syntactically? I have argued that diminutive suffixes are syntactic modifiers, but what is the syntax of the delimitative po-? Similar to the diminutive modifiers, the delimitative po- is restricted to only one category: verbs (it cannot be used with nouns or adjectives, as shown in (18)–(19)). Diminutives, on the other hand, are restricted to nouns.  (18)  a.  sín                                              b.* po-sin             son.N.SG (MASC)                                DELIM.PREF-son.N.SG (MASC)               ‘son’                                                              ‘small son’  (19)  a.  žád-n-ij                                     b.* po-žad-n-ij             stingy-ADJ-MASC.N.SG                            DELIM.PREF-stingy-ADJ-MASC.N.SG             ‘stingy’                                             ‘a little bit stingy’  However, unlike the diminutive modifiers, the delimitative prefix po- does not allow repetition of the same morpheme, as illustrated in (20)–(21). Thus, the delimitative po- has one property of a syntactic modifier, but it does not have another property. The question of the syntactic structure for the delimitative prefix po- remains for further research.  (20)  a.  po-gul’-á-t’                                 b.* po-po-gul’-á-t’             DELIM.PREF-walk-TH-INF                        DELIM.PREF-DELIM.PREF-walk-TH-INF             ‘to walk for a while’                         ‘to walk for a while’  (21)  a.  po-p’-í-t’                                     b.* po-po-p’-í-t’             DELIM.PREF-drink-TH-INF                       DELIM.PREF-DELIM.PREF-drink-TH-INF             ‘to drink for a while’                        ‘to drink for a while’ 178  7.2.3. Further issues From a phonological perspective, the following research topics arise. First, how do expressive suffixes in Russian interact with stress patterns? Some expressive suffixes seem to carry inherent stress (the suffix vowel is always stressed), e.g., -án’ (22) and -úg (23). Other expressive suffixes, such as -in, do not carry inherent stress (compare (24) and (25)). Why some expressives suffixes are inherently stressed, while others are not, remains an open question. Stress does not match up with the syntactic differences identified in this thesis; otherwise I could have used it as a diagnostic for the syntactic differences.  (22)  a.  mám-a                                        b.  mam-án’-a             mother-N.SG (FEM)                                mother-EXPR-N.SG (FEM)             ‘mother’                                           ‘mother (affect)’  (23)  a.  zv’ér’                                           b.  zv’er’-úg-a             animal.N.SG (MASC)                               animal-EXPR-N.SG (MASC/FEM)             ‘animal’                                            ‘animal (vulg)’  (24)  a.  skót                                             b.  skot’-ín-a            (expressive suffix stressed)             cattle.N.SG (MASC)                                  cattle-EXPR-N.SG (MASC/FEM)            ‘cattle/swine’                                   ‘cattle/swine (vulg)’  (25)  a. uród                                             b.  uród’-in-a    (expressive suffix not stressed)             ugly.N.SG (MASC)                                      ugly-EXPR-N.SG (MASC/FEM)             ‘ugly animate’                                   ‘ugly animate (vulg)’  Second, some expressive suffixes in Russian trigger palatalization in the base-final consonant, such as -úg (26). Other expressive suffixes do not trigger palatalization, such as -úx (27). Why expressive suffixes act differently with respect to palatalization remains a topic for further research. Palatalization does not line up with the syntactic differences identified in this thesis.  (26)  a.  žád-n-ij                                       b.  žad-n’-úg-a        (triggers palatalization)             stingy-ADJ-MASC.N.SG                            stingy-ADJ-EXPR-N.SG (MASC/FEM)            ‘stingy’                                              ‘stingy animate (vulg)’  179 (27)  a.  gr’áz-n-ij                                     b.  gr’áz-n-úx-a   (does not trigger palatalization)              dirty-ADJ-MASC.N.SG                             dirty-ADJ-EXPR-N.SG (MASC/FEM)             ‘dirty’                                               ‘dirty animate (vulg)’  From a typological perspective, an interesting research question arises: What is the morphosyntax of expressives suffixes in other Slavic languages? All Slavic languages are known to be rich in expressive morphology. With this respect, it would be interesting to investigate the morphosyntactic types of expressive suffixes in other Slavic languages and compare them to those of Russian. 180 REFERENCES Abney, S. (1987). The English noun phrase in its sentential aspects. Unpublished Ph.D. thesis, Massachusetts Institute of Technology. Aikhenvald, A. (2004). Evidentiality. Oxford: Oxford University Press. Apres'an, U. (1995). Integral'noe opisanie jazyka i sistemnaya leksikografia. Moskva. Arad, M. (2003). Locality constraints on the interpretation of roots: The case of Hebrew denominal verbs. Natural Language and Linguistic Theory, 21, 737-778. Aronoff, M. (1976). Word formation in generative grammar. Cambridge: MIT Press. Babko-Malaya, O. (2003). Perfectivity and prefixation in Russian. Journal of Slavic Linguistics, 11, 5-36. Bachrach, A., & Wagner, M. (2007). Syntactically driven cyclicity vs. output-output correspondence: The case of adjunction in diminutive morphology. Proceedings of the 30th Penn Linguistics Colloquium. Available at: http://ling.auf.net/lingbuzz/@GQgxYONSUpFoqyOV/pwrJlqTA?6 Beard, R. (1966). The affixation of adjectives in contemporary literary Serbo-Croatian. Unpublished Ph.D., University of Michigan. Beard, R. (1995). Morpheme-lexeme base morphology. Albany: SUNY Press. Beard, R., & Szymanek, B. (1988). Bibliography of morphology 1960-1985. Amsterdam, Philadelphia: Benjamins. Bierwisch, M. (2003). In E. Lang, C. Maienborn, & C. Fabricius-Hansen (Eds.), Modifying Adjuncts (pp. 113-159). Berlin, New York: Mounton de Gruyter. Blain. E. & R.-M. Déchaine. (2006). The evidential domain hypothesis. Proceedings of WSCLA 11 (Workshop on the Structure and Constituency of the Languages of the Americas 11). University of British Columbia. Vancouver, BC, Canada, pp. 12–25. Bobaljik, J. (2002). Syncretism without paradigms: Remarks on Williams 1981, 1994. Unpublished manuscript. Bonet, E. (1991). Morphology after syntax. Unpublished Ph.D. thesis, Massachusetts Institute of Technology. Bratus, B. V. (1969). The formation and expressive use of diminutives. Cambridge: Cambridge University Press. Bühler, K. (1934). Sprachtheorie. Jena. 181 Chafe, W. (1986). Evidentiality in English conversation and academic writing. In W. Chafe, & J. Nichols (Eds.), Evidentiality: The linguistic coding of epistemology (pp. 261- 272). Norwood, N.J.: Ablex Publishing. Chomsky, N. (1995). The minimalist program. Cambridge, MA: The MIT Press. Chomsky, N. (2001). Derivation by phase. In M.J. Kenstowicz (Ed.), Ken Hale: A life in language (pp. 1-52). Cambridge, MA: The MIT Press. Churchward, C. (1953). Tongan grammar. Oxford: Oxford University Press. Clements, G.N. & Hume, E. (1995). The Internal Organization of Segments. In J. Goldsmith (ed.) The Handbook of Phonological Theory. Blackwell Handbooks in Linguistics. Blackwell Publishers, 245-306. Corbett, G. (1980). Animacy in Russian and other Slavonic languages: Where syntax and semantics fail to match. In C. Chvany, & R. Brecht (Eds.), Morphosyntax in Slavic (pp. 43-61). Columbus: Slavica. Corbett, G. (1982). Gender in Russian: An account of gender specification and its relationship to declension. Russian Linguistics, 6(2), 197–232. Corbett, G. (1991). Gender. Cambridge: Cambridge University Press. Corbett, G., & Fraser, N. (2000). Default genders. In B. Unterbeck, & M. Rissanen (Eds.), Gender in grammar and cognition (pp. 55-98). Berlin; New York: Mouton de Gruyter. Davis, H., Matthewson L., & Rullmann, H. (2007) ‘Out of control’ marking as circumstantial modality in St’át’imcets. Unpublished manuscript. Available at: http://semanticsarchive.net/Archive/GVhMzFjM/DMR%20circumstantial.pdf Dement'ev, A. (1953). Umen'shitel'nye slova v russkom jazyke. Russkii iazyk v shkole, 5. Demirdache, H. (1997). Out of Control in St'át'imcets. In A. Mendikoetxea & M. Uribe- Etxebarria (eds.), Theoretical Issues in the Morphology-Syntax Interface, supplement to the International Journal of Basque Linguistics and Philology, 97-143. Derkach, T. (2005). Formal and semantic aspects of denominal augmentatives in Polish, Ukrainian and Russian. Unpublished manuscript. Dimitrova-Vulchanova, M. (1999). Verb semantics, diathesis and aspect. Lincom Europa, München. 182 Doleschal, U., & Schmid, S. (2001). Doing gender in Russian: Structure and perspective. In M. Hellinger, & H. Bussmann (Eds.), Gender across languages (pp. 253-282). Amsterdam/Philadelphia: John Benjamins Publishing Company. Dressler, W. (1997). Studies in pre- and proto morphology. Vienna: Verlag der Österreichischen Akademie der Wissenschaften. Dressler, W., & Merlini Barbaresi, L. (1994). Morphopragmatics: Diminutives and intensifiers in Italian, German, and other languages. Berlin: Mouton de Gruyter. Durnovo, N. (1922). De la declinaison en grand-russe litteraire moderne. Revue des Etudes Slaves, 2, 235-255. Durovich, L. (1964). Paradigmatika spisovnej rushtiny. Slovenske pedagogicke nakladatel'stvo. Bratislava. Efremova, T. (2006). Sovremennyj tolkovyj slovar' russkogo jazyka. Moskva: Astrlel'. Embick, D. (1997). Voice morphology and its interfaces with syntax. Unpublished Ph.D. thesis, University of Pennsylvania. Embick, D. (1998). Voice systems and the syntax/morphology interface. In H. Harley (Ed.), MITWPL 32: Papers from the UPenn/MIT roundtable on argument structure and aspect (pp. 41-72) MITWPL. Embick, D., & Noyer, R. (2005). Distributed morphology and the syntax/morphology interface. Unpublished manuscript. Available at: http://www.ling.upenn.edu/~embick/interface.pdf Ettinger, S. (1974). Form und Function in der Wortbilding: Die Diminutiv- und Augmentativmodification im Lateinischen, Deutschen und Romanischen. Ein kritischer Forschungsbericht 1900-1975 (1st ed.). Tübingen: Narr. Ettinger, S. (1980). Form und Function in der Wortbilding: Die Diminutiv- und Augmentativmodification im Lateinischen, Deutschen und Romanischen. Ein kritischer Forschungsbericht 1900-1975 (2nd ed.). Tübingen: Narr. Faller, M.T. (2002). Semantics and pragmatics of evidentials in Cuzco Quechua. Unpublished Ph.D. thesis, Stanford University. Faller, M.T. (2003). Propositional- and illocutionary-level evidentiality in Cuzco Quechua. In J. Anderssen, P. Menendez-Benito & A. Werle (eds.) SULA 2. Fentslova, M. (1985). Rol' znacheniia iskhodnogo slova pri obrazovanii znacheniia deminutive, issledovannaia na materiale razgovornoi rechi russkogo i cheshskogo iazykov. Ceskoslovenska Rusistika, 30(5), 221-227. 183 Fowler, G. (1996). An articulated theory of aspect and prefixation in Slavic. In J. Toman (Ed.), Annual Workshop on Formal Approaches to Slavic Linguistics: The College Park Meeting 1994 (pp. 97-122) Michigan Slavic Publications. Gillis, S. (Ed.). (1998). Studies in the acquisition of number and diminutive marking. Antwerpen: University Antwerpen. Gorshkova, K., & Xaburgaev, G. (1981). Istoricheskaia grammatika russkogo iazyka. Moskva: Vysshaia Shkola. Grimshaw, J. B. (1991). Extended projections. Unpublished manuscript. Halle, M. (1994). The Russian declension: An illustration of the theory of distributed morphology. In J. Cole, & C. Kisseberth (Eds.), Perspectives in phonology (pp. 29- 60). Stanford: CSLI Publications. Halle, M. (1997). Distributed morphology: Impoverishment and fission. In B. Brüning, Y. Kang & M. McGinnis (Eds.), Papers at the interface (pp. 425-449) MITWPL. Halle, M., & Marantz, A. (1993). Distributed morphology and the pieces of inflection. In K. Hale, & S. Keyser (Eds.), The view from building 20: Essays in linguistics in honor of Sylvain Bromberger (pp. 111-176). Cambridge, Mass.: MIT Press. Halle, M., & Matushansky, O. (2006). The morphophonology of Russian adjectival inflection. Linguistic Inquiry, 37(3), 351-404. Hansson, G. (1993). Padenie reducirovannyx i upotreblenie predlogov vo, so, ko v sovremennom russkom jazyke. Unpublished B.A. Thesis, University of Iceland. Harley, H., & Noyer, R. (1999). State-of-the-article: Distributed morphology, 4(4) 3-9. Harley, H., & Noyer, R. (2003). Distributed morphology. In L. Cheng, & R. Sybesma (Eds.), The second GLOT international state-of-the-article book (pp. 463-496). Berlin: Mouton de Gruyter. Helbig, G. (1974). Geschichte der neueren Sprachwissenschaft. Reinbek: Rowohlt. Hendrick, R. (2005). Tongan determiners and semantic composition. Language, 81(4), 907- 926. Hermans, B. (2002). Overapplication of yer vocalization in Russian. In H. Broekhuis, & P. Fikkert (Eds.), (pp. 85-95). Amsterdam: John Benjamins. Available at: http://www.benjamins.nl/jbp/series/AVT/19/art/0010a.pdf Isachenko, A. (1962). Die Russische Sprache der Gegenwart I: Formenlehre. Halle: VEB Max Niemeyer Verlag. 184 Ivanova, N. (1965). Narechija s suffiksom emocional'noj ocenki -en'k (-on'k) v sovremennom russkom jazyke. Russkii iazyk v shkole, 1, 83-85. Jakobson, R. (1958). Morfologicheskie nabludenija nad slavanskim skloneniem. American contributions to the fourth international congress of Slavists (pp. 127-156). The Hague: Mouton. Jakobson, R. (1960). Linguistics and poetics. In T. Sebeok (Ed.), Style in language (pp. 350- 377). Cambridge, Mass: The MIT Press. Kalasniemi, M. (1992). Ocenochnyj suffiks -#k v russkom jazyke. University of Jyvaskyla: Jyvaskyla. Karcevskij, S. (1932). Sur la structure du substantif russe. Charisteria guilelmo mathesio quinquagenario (pp. 65-73). Prague: Cercle linguistique de Prague. Karcevskij, S. (1948). Notes de morphologie russe (2nd ed.). Geneve: Ecole d'interpretes de l'Universite de Geneve. Kikuchi, Y. (1994). Keigo [honorifics]. Tokyo: Kadokawa Shoten. Kolomiets, V. G. (1988). Slovoobrazovatel'nye formy prilagatel'nyx s umen'shitel'no- laskatel'nym znacheniem v slavyanskix yazykax, Izvestiya Akademii Nauk SSSR, Seriya Literatury i Yazyka, 47(6), 550-558. Kosmeda, T. (1999). Vyrazhenie kategorii laskatel'nosti i unichizhitelnosti. Slowotworstwo, semantika i skladnia jezykow slowianskich, 1, 16-23. Kratzer, A. (1999). Beyond ouch and oops: How descriptive and expressive meaning interact. A comment on David Kaplan's paper. Unpublished manuscript. From http://semanticsarchive.net/Archive/WEwNGUyO/ Leech, G. (1974). Semantics. Penguin Books. Lightner, T. (1972). Problems in the theory of phonology. Vol. I: Russian phonology and Turkish phonology. Edmonton: Linguistic Research, Inc. Mandel'štam. (1903). Russkie umen'shitel'nye suffiksy so storony iz znachenija. Zhurnal Ministerstva Narodnogo Prosveschenija, 7-8. Manova, S. (2004). Derivation versus inflection in three inflecting languages. In W. U. Dressler, D. Kastovsky, O. E. Pfeiffer, F. Rainer, F. Gardani & M. A. Pöchtrager (Eds.), Morphology and its demarcations (pp. 233-252). Amsterdam, Netherlands: Benjamins.  185 Marantz, A. (1997). No escape from syntax: Don't try morphological analysis in the privacy of your own lexicon. Proceedings of the 21st Annual Penn Linguistics Colloquium. Available at: http://dingo.sbs.arizona.edu/~hharley/courses/Oxford/Marantz.pdf Marantz, A. (2001). Words. Unpublished manuscript. Marvin, T. (2002). Topics in the stress and syntax of words. Massachusetts Institute of Technology. Matushansky, O. (2002). On formal identity of Russian prefixes and prepositions. MIT Working Papers in Linguistics 42: 217-253. Müller, G. (2005). A distributed morphology approach to syncretism in Russian noun inflection. Unpublished manuscript. Available at: http://www.uni- leipzig.de/~muellerg/mu52.pdf Nesset, T. (1994). A feature-based approach to Russian noun inflection. Journal of Slavic Linguistics, 2(2), 214-237. Nesset, T. (2003). Gender assignment in Ukrainian: Language specific rules and universal principles. Poljarnyj vestnik, 6, 1-15. Nieuwenhuis, P. (1985). Diminutives. Unpublished Ph.D. thesis, University of Edinburgh. Noyer, R. (1998). Impoverishment theory and morphosyntactic markedness. In D. Lapointe, D. Brentari & P. Farrell (Eds.), Morphology and its relation to phonology and syntax (pp. 264-285). Palo Alto: CSLI. Ogol’cev, V. (1960). Emocional'nye i ekspressivnye znachenija suffiksa im'en prilagatel'nyx -en'k/-on'k. Russkii iazyk v shkole, 2, 8-13. Ooishi, S. (1975). Keigo [honorifics]. Tokyo: Chikuma Shoboo. Pesetsky, D. (1995). Zero syntax: Experiencers and cascades. Cambridge, MA.: MIT Press. Plyamovataya, S. (1955). O grammaticheskoj prirode i klassifikacii im'en suschestvitel'nyx s umen'shitel'no-ekspressivnymi suffiksami v sovremennom russkom jazyke. Russkii iazyk v shkole, 6, 4-11. Plyamovataya, S. (1961). Razmerno-ocenochnye imena suschestvitel'nye v sovremennom russkom jazyke. Moskva. Polterauer, I. (1981). Die Deminutiva in der modernen russischen Schriftsprache. Vienna: VWGÖ. 186 Popoff-Böcker, E. (1973). A semantic approach to the diminutive suffixes in Russian. VII miedzynarodowy kongres slawistow w warszawie 1973: Streszczenia referatow i komunikatow. (pp. 408-409). Warsaw: PAN. Popov, E. A. (1967). The semantic structure of the Russian diminutives. Unpublished Ph.D. thesis, Stanford University. Potts, C. (2003). The performative nature of expressive content. Unpublished manuscript. Potts, C. (to appear in 2007). The expressive dimension. Theoretical Linguistics 33(2): 165- 197. Availabel at: http://people.umass.edu/potts/papers/potts-expressives06.pdf Potts, C., & Kawahara, S. (2004). Japanese honorifics as emotive definite descriptions. Proceedings of Semantics and Linguistic Theory, 14, 235-254. Available as a manuscript at: http://people.umass.edu/potts/papers/potts-kawahara-salt14-paper.pdf Protasova, E. (2001). Rol' diminutivov v sovremennom russkom jazyke. Trudy po russkoj i slavjanskoj filologii - Lingvistika, 5, 72-88. Rakušan, J. (1981). On the expressivity of nominal suffixes in modern Russian. Russian Linguistics, 6, 41-55. Ritter, E. (1995). On the syntactic category of pronouns and agreement. Natural Language and Linguistic Theory, 13, 405-443. Rojina, N. (2004). English particles, Russian prefixes, and prepositional phrases. Unpublished Master’s thesis, University of Tromsø. Rooryck, J. (2001). Evidentiality. Glot International, 5(4), 125-133. Rowicka, G. (1999). On ghost vowels. A strict CV approach. Unpublised Ph.D. Thesis, University of Leiden. Rubach, J. (1986). Abstract vowels in three-dimensional phonology: The yers. Linguistic Review, 5, 247-280. Scalise, S. (1984). Generative morphology. Dordrecht: Foris Publications. Scalise, S. (1988). The notion of 'head' in morphology. Yearbook of Morphology, 1, 229-245. Schlenker, P. (2003). A plea for monsters. Linguistics and Philosophy, 26(1), 29-120. Schneider, K. P. (2003). Diminutives in English. Max Niemeyer Verlag Gmbh. Schütze, C. (1995). PP attachment and argumenthood. Papers on Language             Processing and Acquisition. MITWPL Working Papers in Linguistics, 26, 95-151. 187 Shanskaja, T. (1961). O rodovoj prinadlezhnosti slov s suffiksami sub'ektivnoj ocenki. Russkii iazyk v shkole, 6, 13-17. Shvedova, N. (Ed.). (1970). Grammatika sovremennogo russkogo literaturnogo jazyka. Moskva: Nauka. Shvedova, N., Arut'unova, N., Bondarko, A., Ivanov, V., Lopatin, V., Uluxanov, I., et al (Eds.). (1982). Russkaya grammatika. Moskva: Nauka. Spencer, A. (1985). A non-linear analysis of vowel-zero alternations in Polish. Journal of             Linguistics 22: 249-280. Spiridonova, N. F. (1999). Russkie diminutivy: Problemy obrazovaniia i znacheniia. Izvestiia Akademii Nauk, Seriia Literatury i Iazyka, 58(2), 13-22. Stankiewicz, E. (1954). Declension and gradation of substantives in contemporary standard Russian. Unpublished Ph.D. thesis, Harvard University. Stankiewicz, E. (1968). Declension and gradation of Russian substantives. The Hague: Mouton. Stankiewicz, E. (1978). The inflection of Serbo-Croatian substantives and their genitive plural endings. In H. Birnbaum (Ed.), American contributions to the eighth international congress of Slavists I: Linguistics and poetics (pp. 666-681). Columbus: Slavica. Steriopolo, O. (2007). Jer vowels in Russian prepositions. Formal Approaches to Slavic Linguistics. The Toronto Meeting 2006. Michigan Slavic Publications, 365-385. Available at: http://www.indiana.edu/~sls2006/Handouts/SteriopoloSLS.pdf Stump, G. (1993). How peculiar is evaluative morphology. Journal of Linguistics, (29), 1-36. Svenonius, P. (2004). Slavic prefixes inside and outside of VP. Nordlyd. Special issue on Slavic prefixes 32(2), 323-361. Szpyra, J. (1992). Ghost segments in nonlinear phonology: Polish yers. Language 48: 277- 312. Thelin, N. (1975). Notes on general and Russian morphology. Studia Slavica Upsaliensia, 15, 30-32. Timberlake, A. (2004). A reference grammar of Russian. Cambridge: Cambridge University Press. Tokieda, M. (1940). Kokugogaku genron [principles of Japanese linguistics]. Tokyo: Iwanami. 188 Trager, G. (1940). The Russian gender categories. Language, 16, 300-307. Tsujimura, T. (1978). Keigo to hikeigo [honorifics and anti-honorifics]. Keigo: Ronshuu Nihongo Kenkyuu [Honorifics: Papers on Research in Japanese], 9 218-231. Unbegaun, B. (1957). Russian grammar. Oxford: Clarendon. Vaseva, I. (1977). Narechiia s suffiksom sub''ektivnoi otsenki v russkom i bolgarskom iazykakh. Sapostavitelno Ezikoznanie/Sopostavitel'noe Jazykoznanie/Contrastive Linguistics, 2(6), 9-22. Vinogradov, V. (1972). Russkij jazyk (2nd ed.). Moskva: Uchpedgiz. Vinogradov, V., Istrina, E., & Barxudarov, S. (Eds.). (1952). Grammatika russkogo jazyka I: Fonetika i morphologija. Moskva: AN SSSR. Voeikova, M., & Dressler, W. (2002). Pre- and proto-morphology. München: Lincom Europa. Volek, B. (1987). Emotive signs in language and semantic functioning of derived nouns in Russian. Amsterdam: Benjamins. Wade, T. (2000). A comprehensive Russian grammar. Blackwell Publishing. Wierzbicka, A. (1984). Diminutives and depreciatives: Semantic representation for derivational categories. Quaderni di Semantica, 5, 123-130. Willett, T. (1988). A cross-linguistic survey of the grammaticization of evidentiality. Studies in Language, 12(1), 51-97. Wiltschko, M. (2006). Why should diminutives count? In H. Broekhuis, N. Corver, R. Huijbregts, U. Kleinhenz & J. Koster (Eds.), Organizing grammar. Linguistic studies in honor of Henk van Riemsdijk. Berlin: Walter de Gryter. Wiltschko, M. (2008). The syntax of non-inflectional plural marking. NLLT. Wiltschko, M., & Steriopolo, O. (2007). Parameters of variation in the syntax of diminutives. Unpublished manuscript. Yearley, J. (1995). Jer vowels in Russian. Occasional Papers in Linguistics, 18, 533-571. Zalizniak, A., & Shmelev, A. (2000). Vvedenie v russkuju aspektologiju. Moskva: Jazyki Russkoj Kul’tury. Zalizniak, A. (1967). Russkoe imennoe slovoizmenenie. Moskva: Nauka. 189 Zoll, C. (1998). Parsing below the segment in a constraint-based framework. Stanford:            CSLI Publications.    190 APPENDIX   Examples of data showing change in the formal properties of attitude suffixes   1. VULGAR SUFFIXES  1.1.      The suffix -ág  Diagnostics The suffix -ág Data Does it change syntactic category? a (1)-(2) Does it change grammatical gender? a (3) Does it change inflectional class? a (3)-(4) Table A.1: The suffix -ág  i. Change in category  (1)    a.  x’ítr-ij                                       b.   x’it’-ág-a             sly-ADJ.MASC.N.SG                                sly-EXPR-N.SG (MASC/FEM)             ‘sly                                                 ‘sly animate (vulg)’  (2)    a.  zdoróv-ij                                   b.   zdorov’-ág-a            big-ADJ.MASC.N.SG                                big-EXPR-N.SG (MASC/FEM)             ‘big’                                                ‘big animate (vulg)’  ii. Change in gender/inflectional class  (3)  a.  kón’                                             b.  kon’-ág-a            horse.N.SG (MASC; CLASS I)               horse-EXPR-N.SG (MASC/FEM; CLASS II)           ‘horse’                                              ‘horse (vulg)’  (4)  a.  pár’en’                                        b.  parn’-ág-a           guy.N.SG (MASC; CLASS I)                   guy-EXPR-N.SG (MASC; CLASS II)           ‘guy’                                                ‘guy (vulg)’       191 1.2.      The suffix -ák  Diagnostics The suffix -ák Data Does it change syntactic category? a (5)-(6) Does it change grammatical gender? a (7)-(8) Does it change inflectional class? a (7)-(8) Table A.2: The suffix -ák  i. Change in category  (5)  a.  gul’-á-t’                                         b.  gul’-ák-a            walk/play-TH-INF                                       walk/play-EXPR-N.SG (MASC/FEM)           ‘to walk/play’                                    ‘animate who likes to walk/play (vulg)’  (6)  a.  p’is-á-t’                                          b.  p’is-ák-a           write-TH-INF                                                 write-EXPR-N.SG (MASC/FEM)           ‘to write’                                            ‘writer (vulg)’  ii. Change in gender/inflectional class  (7)  a.  čórt                                                b.  č’ert’-ák-a            devil.N.SG (MASC; CLASS I)                  devil-EXPR-N.SG (MASC/FEM; CLASS II)           ‘devil’                                                ‘devious animate (vulg)’  (8)  a.  kón’                                                b.  kon’-ák-a            horse.N.SG (MASC; CLASS I)                  horse-EXPR-N.SG (MASC/FEM; CLASS II)           ‘horse’                                                ‘horse (vulg)’  1.3.      The suffix -ál  Diagnostics The suffix -ál Data Does it change syntactic category? a (9)-(11) Does it change grammatical gender? * Does it change inflectional class? * Table A.3: The suffix -ál     192 Change in category  (9)  a.  pr’i-l’ip-á-t’                               b.  pr’i-l’ip-ál-a           VERB.PREF-cling-TH-INF                       VERB.PREF-cling-EXPR-N.SG (MASC/FEM)           ‘to cling’                                         ‘clinging animate (vulg)’  (10)  a.  pod-p’ev-á-t’                            b.  pod-p’ev-ál-a            VERB.PREF-sing-TH-INF                        VERB.PREF-sing-EXPR-N.SG (MASC/FEM)           ‘to join in singing’                           ‘yes-man (vulg)’  (11)  a.  xník-a-t’                                    b.  xník-al-a             complain-TH-INF                                    complain-EXPR-N.SG (MASC/FEM)            ‘to complain’                                  ‘complaining person (vulg)’  1.4.      The suffix -án  Diagnostics The suffix -án Data Does it change syntactic category? * Does it change grammatical gender? a (12)-(13) Does it change inflectional class? a (13)-(14) Table A.4: The suffix -án  Change in gender/inflectional class  (12)  a.  púz-o                                          b.  puz-án             belly-N.SG (NEUT; CLASS I)                belly-EXPR.N.SG (MASC; CLASS I)             ‘belly’                                             ‘animate with distinct belly (vulg)’  (13)  a.  gub-á                                          b.  gub-án             lip-N.SG (FEM; CLASS II)                     lip-EXPR.N.SG (MASC; CLASS I)             ‘lip’                                                 ‘animate with distinct lips (vulg)’  (14)  a.  Kól’-a                                         b.  Kol’-án             Kolia-N.SG (MASC; CLASS II)             Kolia-N.SG (MASC; CLASS I)               ‘Kolia’                                              ‘Kolia (vulg)’         193 1.5.      The suffix -ár  Diagnostics The suffix -ár Data Does it change syntactic category? * Does it change grammatical gender? a (15) Does it change inflectional class? a (16) Table A.5: The suffix -ár  Change in inflectional class  (15)  a.  sobák-a                                      b.  sobač’-ár-a             dog-N.SG (FEM; CLASS II)                  dog-EXPR-N.SG (MASC/FEM; CLASS II)             ‘dog’                                               ‘dog (vulg)’  (16)  a.  kót                                              b.  kot’-ár-a             cat.N.SG (MASC; CLASS I)                    cat-EXPR-N.SG (MASC; CLASS II)            ‘cat’                                                 ‘cat (vulg)’  1.6.      The suffix -áx  Diagnostics The suffix -áx Data Does it change syntactic category? a (17)-(19) Does it change grammatical gender? * Does it change inflectional class? * Table A.6: The suffix -áx  Change in category  (17)  a.  za-mar-á-t’                                 b.  ras-mar-áx-a             VERB.PREF-make.dirty-TH-INF           VERB.PREF-make.dirty-EXPR-N.SG (MASC/FEM)               ‘to make dirty’                                 ‘animate who makes itself dirty (vulg)’  (18)  a.  ras-t’er’-á-t’                               b.  ras-t’er’-áx-a             VERB.PREF-lose-TH-INF                         VERB.PREF-loose-EXPR-N.SG (MASC/FEM)               ‘to lose’                                            ‘animate who loses things (vulg)’  (19)  a.  po-b’ir-á-t’-s’a                           b.  po-b’ir-áx-a             VERB.PREF-take-TH-INF-self                VERB.PREF-take-EXPR-N.SG (MASC/FEM)               ‘to live by begging’                          ‘beggar (vulg)’ 194  1.7.      The suffix -íl  Diagnostics The suffix -íl Data Does it change syntactic category? a (20)-(22) Does it change grammatical gender? * Does it change inflectional class? * Table A.7: The suffix -íl   i. Change in category  (20)  a.  stráš-n-ij                                     b.   straš-íl-a              ugly-ADJ-MASC.N.SG                             ugly-EXPR-N.SG (MASC/FEM)              ‘ugly’                                               ‘ugly animate (vulg)’  (21)  a.  maz-á-t’                                       b.  maz’-íl-a             miss-TH-INF                                                miss-EXPR-N.SG (MASC/FEM)               ‘to miss the target’                            ‘animate who misses the target (vulg)’  (22)  a. za-vod’-í-t’                                 b.  za-vod’-íl-a             VERB.PREF-lead-TH-INF                          VERB.PREF-take-EXPR-N.SG (MASC/FEM)                ‘to live lead’                                      ‘ringleader (vulg)’  1.8.      The suffix -in   Diagnostics The suffix -in Data Does it change syntactic category? a (23) Does it change grammatical gender? a (24)-(25) Does it change inflectional class? a (24)-(25) Table A.8: The suffix -in   i. Change in category  (23)  a.  koso-lap-ij                                   b.   koso-láp’-in-a             crooked-paw-ADJ.MASC.N.SG               crooked-paw-EXPR-N.SG (MASC/FEM)             ‘awkward’                                         ‘awkward animate (vulg)’ 195   ii. Change in gender/inflectional class  (24)  a.  bolót-o                                         b.  bolót’-in-a             swamp-N.SG (NEUT; CLASS I)              swamp-EXPR-N.SG (FEM; CLASS II)             ‘swamp’                                            ‘swamp (vulg)’  (25)  a. uród                                              b.  uród’-in-a             ugly-N.SG (MASC; CLASS I)                  ugly-EXPR-N.SG (MASC/FEM; CLASS II)            ‘ugly animate’                                   ‘ugly animate (vulg)’  1.9.      The suffix -ób  Diagnostics The suffix -ób Data Does it change syntactic category? a (26) Does it change grammatical gender? a (27) Does it change inflectional class? a (27) Table 9: The suffix -ób  i. Change in category  (26)  a.  žád-n-ij                                        b.   žad-ób-a             stingy-ADJ-MASC.N.SG                            stingy-EXPR-N.SG (MASC/FEM)             ‘stingy’                                              ‘stingy animate (vulg)’  ii. Change in gender/inflectional class  (27)  a.  st’id                                             b.  stid-ób-a             shame-N.SG (MASC; CLASS I)               shame-EXPR-N.SG (FEM; CLASS II)            ‘shame’                                             ‘shame (vulg)’  1.10. The suffix -ot  Diagnostics The suffix -ot Data Does it change syntactic category? a (28) Does it change grammatical gender? a (29)-(30) Does it change inflectional class? a (29)-(30) Table A.10: The suffix -ot 196   i. Change in category  (28)  a.  m’él-k-ij                                      b.   m’el-k-ot-á             small-ADJ-MASC.N.SG                              small-ADJ-EXPR-N.SG (FEM)             ‘small/shallow’                                 ‘smallness/shallowness (vulg)’  ii. Change in gender/inflectional class  (29)  a.  sm’éx                                           b.  sm’ex-ot-á             laughter.N.SG (MASC; CLASS I)           laughter-EXPR-N.SG (FEM; CLASS II)            ‘laughter’                                          ‘laughter (vulg)’  (30)  a.  srám                                             b.  sram-ot-á             shame.N.SG (MASC; CLASS I)               shame-EXPR-N.SG (FEM; CLASS II)            ‘shame’                                             ‘shame (vulg)’  1.11. The suffix -óx   Diagnostics The suffix -óx Data Does it change syntactic category? a (31)-(32) Does it change grammatical gender? a (33) Does it change inflectional class? a (34) Table A.11: The suffix -óx   i. Change in category  (31)  a.  vi-p’iv-á-t’                                   b.  vi-p’iv-óx-a             VERB.PREF-drink-TH-INF                        VERB.PREF-drink-EXPR-N.SG (MASC/FEM)               ‘to drink up’                                      ‘boozer (vulg)’  (32)   a.  o-b’ir-á-t’                                    b.  o-b’ir-óx-a             VERB.PREF-take-TH-INF                          VERB.PREF-take-EXPR-N.SG (MASC/FEM)             ‘rob/fleece’                                        ‘robber (vulg)’  ii. Change in gender/inflectional class  (33)   a.  právd-a                                       b. pravd-óx-a               truth-N.SG (FEM; CLASS II)                 truth-EXPR-N.SG (MASC/FEM; CLASS II)               ‘truth’                                               ‘truth telling person (vulg)’ 197  (34)  a.  Igor’                                              b.  Igor’-óx-a             Igor’.N.SG (MASC; CLASS I)                  Igor’-EXPR-N.SG (MASC; CLASS II)            ‘Igor’’                                                ‘Igor’ (vulg)’  1.12. The suffix -úg  Diagnostics The suffix -úg Data Does it change syntactic category? a (35)-(36) Does it change grammatical gender? a (37)-(38) Does it change inflectional class? a (37)-(38) Table A.12: The suffix -úg  i. Change in category  (35)  a.  xvat-á-t’                                       b.  xvat’-úg-a             grab-TH-INF                                                grab-EXPR-N.SG (MASC/FEM)             ‘to grab’                                           ‘grabber (vulg)’  (36)  a.  žád-n-ij                                         b.  žad-n’-úg-a             stingy-ADJ-MASC.N.SG                            stingy-ADJ-EXPR-N.SG (MASC/FEM)             ‘stingy’                                              ‘stingy animate (vulg)’  ii. Change in gender/inflectional class  (37) a.  vór                                                b.  vor’-úg-a             thief.N.SG (MASC; CLASS I)                   thief-EXPR-N.SG (MASC/FEM; CLASS II)            ‘thief’                                                ‘thief (vulg)’  (38)  a.  zv’ér’                                            b.  zv’er’-úg-a              animal.N.SG (MASC; CLASS I)             animal-EXPR-N.SG (MASC/FEM; CLASS II)             ‘animal’                                            ‘animal (vulg)’            198 1.13. The suffix -úk   Diagnostics The suffix -úk Data Does it change syntactic category? a (39)-(40) Does it change grammatical gender? a (41)-(42) Does it change inflectional class? a (41)-(42) Table 13: The suffix -úk   i. Change in category  (39)  a.  pódl-ij                                          b.   podl’-úk-a             mean-ADJ.MASC.N.SG                              mean-EXPR-N.SG (MASC/FEM)            ‘mean’                                                ‘mean animate (vulg)’  (40)  a.  zl-ój                                              b.  zl’-úk-a             angry-ADJ.MASC.N.SG                               angry-EXPR-N.SG (MASC/FEM)            ‘angry’                                                ‘angry animate (vulg)’  ii. Change in gender/inflectional class  (41)  a.  kón’                                              b.  kon’-uk-a             horse.N.SG (MASC; CLASS I)                  horse-EXPR-N.SG (MASC/FEM; CLASS II)            ‘horse’                                                ‘horse (vulg)’  (42)  a.  tvár’                                             b.  tvar’-úk-a              animal.N.SG (FEM; CLASS III)           animal-EXPR-N.SG (MASC/FEM; CLASS II)             ‘animal’                                             ‘animal (vulg)’  1.14. The suffix -úx  Diagnostics The suffix -úx Data Does it change syntactic category? a (43)-(44) Does it change grammatical gender? a (45)-(46) Does it change inflectional class? a (45)-(46) Table A.14: The suffix -úx   199 i. Change in category  (43)  a.  za-v’ir-á-t’                                    b.  za-v’ir-úx-a             VERB.PREF-lie-TH-INF                              VERB.PREF-lie-EXPR-N.SG (MASC/FEM)                ‘to lie’                                                ‘liar (vulg)’  (44)  a.  gr’áz-n-ij                                       b.  gr’az-n-úx-a             dirty-ADJ-MASC.N.SG                                dirty-ADJ-EXPR-N.SG (MASC/FEM)             ‘dirty’                                                ‘dirty animate (vulg)’  ii. Change in gender/inflectional class  (45)  a.  gólod                                            b.  golod-úx-a           hunger.N.SG (MASC; CLASS I)               hunger-EXPR-N.SG (FEM; CLASS II)                ‘hunger’                                            ‘hunger (vulg)’  (46)  a.  stíd                                                b.  stid-úx-a             shame.N.SG (MASC; CLASS I)               shame-EXPR-N.SG (FEM; CLASS II)             ‘shame’                                             ‘shame (vulg)’   2. AFFECTIONATE SUFFIXES  2.1.  The suffix -án’   Diagnostics The suffix -án’ Data Does it change syntactic category? * Does it change grammatical gender? * Does it change inflectional class? * Table A.15: The suffix -án’   The affectionate suffix -án’ is not productive in Russian. It only attaches to kinship terms (47), and first names (48). It has an affectionate meaning, which is the reason why I include this suffix in the affectionate sub-group of the attitude suffixes. However, there is no data that would show that it can produce a change in the formal properties of the base.  200 (47)  a.  páp-a                                          b.  pap-án’-a             dad-N.SG (MASC; CLASS II)                  dad-EXPR-N.SG (MASC; CLASS II)            ‘dad’                                                ‘dad (affect)’  (48)  a.  Vás’-a                                          b.  Vas’-án’-a             Vasia-N.SG (MASC; CLASS II)              Vasia-EXPR-N.SG (MASC; CLASS II)             ‘Vasia’                                              ‘Vasia (affect)’  2.2.  The suffix -áš  Diagnostics The suffix -áš Data Does it change syntactic category? a (49)-(51) Does it change grammatical gender? * Does it change inflectional class? * Table A.16: The suffix -áš   Change in category  (49)  a.  ras-t’er’-á-t’                                 b.  ras-t’er’-áš-a             VERB.PREF-lose-TH-INF                          VERB.PREF-loose-EXPR-N.SG (MASC/FEM)                ‘to lose’                                             ‘animate who loses things (affect)’  (50)  a.  kut’-í-t’                                         b.  kut’-áš-a             carouse-TH-INF                                          carouse-EXPR-N.SG (MASC/FEM)               ‘to carouse’                                       ‘carousing animate (affect)’  (51)  a. m’íl-ij                                           b. m’il-áš-a             cute-ADJ.MASC.N.SG                                cute-EXPR-N.SG (MASC/FEM)            ‘cute’                                                 ‘cutie (affect)’  2.3.  The suffix -ón   Diagnostics The suffix -ón Data Does it change syntactic category? a (52)-(53) Does it change grammatical gender? a (54)-(55) Does it change inflectional class? a (54)-(55) Table A.17: The suffix -ón 201   i. Change in category  (52)  a.  gul’-á-t’                                        b.  gul’-ón-a             walk/play-TH-INF                                      walk/play-EXPR-N.SG (MASC/FEM)             ‘to walk/play’                                   ‘animate who likes to walk/play (affect)’  (53)  a.  sm’ír-n-ij                                      b.  sm’ir’-ón-a              meek-ADJ-MASC.N.SG                             meek-EXPR-N.SG (MASC/FEM)             ‘meek’                                               ‘meek animate (affect)’  ii. Change in gender/inflectional class  (54)  a.  zv’ér’                                            b.  zv’er’-ón-a              animal.N.SG (MASC; CLASS I)              animal-EXPR-N.SG (MASC/FEM; CLASS II)              ‘animal’                                             ‘animal (affect)’  (55)  a.  slást’                                             b.  slast’-ón-a              sweet.N.SG (FEM; CLASS III)              sweet-EXPR-N.SG (MASC/FEM; CLASS II)             ‘sweet’                                              ‘animate with sweet tooth (affect)’  2.4. The suffix -úl’   Diagnostics The suffix -úl’ Data Does it change syntactic category? a (56)-(57) Does it change grammatical gender? a (58) Does it change inflectional class? a (59) Table A.18: The suffix -úl’   i. Change in category  (56)  a.  za-máz-a-t’                                   b.  za-maz-úl’-a             VERB.PREF-make.dirty-TH-INF             VERB.PREF-make.dirty-EXPR-N.SG (MASC/FEM)                ‘to make dirty’                                  ‘animate who makes itself dirty (affect)’  (57)  a.  rod-n-ój                                         b.  rod-n-úl’-a             dear-ADJ-MASC.SG                                    dear-ADJ-EXPR-N.SG (MASC/FEM)            ‘dear’                                                 ‘dear animate (affect)’   202 ii. Change in gender/inflectional class  (58)  a.  kras-ot-á                                       b.   kras-ot-úl’-a             pretty-NOM-N.SG (FEM; CLASS II)        pretty-NOM-EXPR-N.SG (MASC/FEM; CLASS II)            ‘prettiness/beauty’                             ‘pretty animate (affect)’  (59)  a.  sín                                                b.  sin-úl’-a              son.N.SG (MASC; CLASS I)                   son-EXPR-N.SG (MASC; CLASS II)              ‘son’                                                 ‘son (affect)’  2.5. The suffix -ún’   Diagnostics The suffix -ún’ Data Does it change syntactic category? * Does it change grammatical gender? * Does it change inflectional class? a (60)-(61) Table A.19: The suffix -ún’   Change in inflectional class  (60)  a.  d’éd                                               b.  d’ed-ún’-a              grandfather.N.SG (MASC; CLASS I)     grandfather-EXPR-N.SG (MASC; CLASS II)              ‘grandfather’                                     ‘grandfather (affect)’  (61)  a.  P’ótr                                               b.  P’etr-ún’-a              P’otr.N.SG (MASC; CLASS I)                  P’otr-EXPR-N.SG (MASC; CLASS II)              ‘P’otr’                                                ‘P’otr (affect)’  2.6. The suffix -úr   Diagnostics The suffix -úr Data Does it change syntactic category? * Does it change grammatical gender? a (62) Does it change inflectional class? a (62)-(63) Table A.20: The suffix -úr  203  Change in gender/inflectional class  (62)  a.  n’ém’-ec           German-NOM.N.SG (MASC; CLASS I)   ‘German’         b.  n’em-č-úr-a   German-NOM-EXPR-N.SG (MASC/FEM; CLASS II)              ‘German (contempt)26”  (63)  a.  dóč’                                                b.  doč’-úr-a             daughter.N.SG (FEM; CLASS III)            daughter-EXPR-N.SG (FEM; CLASS II)             ‘daughter’                                           ‘daughter (affect)’  2.7. The suffix -ús’   Diagnostics The suffix -ús’ Data Does it change syntactic category? * Does it change grammatical gender? * Does it change inflectional class? a (64)-(66) Table A.21: The suffix -ús’   Change in inflectional class  (64)  a.  d’éd                                                b.  d’ed-ús’-a              grandfather.N.SG (MASC; CLASS I)       grandfather-EXPR-N.SG (MASC; CLASS II)              ‘grandfather’                                      ‘grandfather (affect)’  (65)  a.  sín                                                   b.  sin-ús’-a              son.N.SG (MASC; CLASS I)                      son-EXPR-N.SG (MASC; CLASS II)              ‘son’                                                   ‘son (affect)’  (66)  a.  P’ótr                                                b.  P’etr-ús’-a             P’otr.N.SG (MASC; CLASS I)                     P’otr-EXPR-N.SG (MASC; CLASS II)             ‘P’otr’                                                  ‘P’otr (affect)’     26  In this word, the suffix -úr has a derogatory rather than affectionate meaning. 204 2.8. The suffix -úš   Diagnostics The suffix -úš Data Does it change syntactic category? a (67)-(68) Does it change grammatical gender? * Does it change inflectional class? a (69)-(70) Table A.22: The suffix -úš   i. Change in category  (67)  a.  vr-á-t’                                              b.  vr-úš-a              lie-TH-INF                                                       lie-EXPR-N.SG (MASC/FEM)             ‘to lie’                                                 ‘liar (affect)’  (68)  a.  rod-n-ój                                           b.  rod-n-úš-a             dear-ADJ-MASC.N.SG                                   dear-ADJ-EXPR-N.SG (MASC/FEM)             ‘dear’                                                   ‘dear animate (affect)’  ii. Change in inflectional class  (69)  a.  brát                                                  b.  brat-úš-a              brother.N.SG (MASC; CLASS I)                brother-EXPR-N.SG (MASC; CLASS II)              ‘brother’                                              ‘brother (affect)’  (70)  a.  Páv’el                                              b.  Pavl-úš-a              Pav’el.N.SG (MASC; CLASS I)                 Pav’el-EXPR-N.SG (MASC; CLASS II)              ‘Pav’el’                                               ‘Pav’el (affect)’   REFERENCES Efremova, T. (2006). Sovremennyj tolkovyj slovar' russkogo jazyka. Moskva: Astrlel'. Ozhegov, S., Shvedova, N. (1999). Tolkovyj slovar’ russkogo jazyka. Moskva: Rossiyskaya Akademiya Nauk. Shvedova, N., Arut'unova, N., Bondarko, A., Ivanov, V., Lopatin, V., Uluxanov, I., et al (Eds.). (1982). Russkaya grammatika. Moskva: Nauka. Zalizniak, A. (1977). Grammaticheskij slovar’ russkogo jazyka. Moskva: Russkij jazyk.

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
United States 26 0
Germany 21 1
Israel 4 0
Iran 2 0
Sweden 2 0
Canada 2 2
Hong Kong 2 0
Malaysia 1 0
China 1 0
France 1 0
City Views Downloads
Unknown 28 0
Washington 20 0
Ludwigsburg 3 1
Stockholm 2 0
Toronto 2 2
Cupertino 2 0
Martin 1 0
Seremban 1 0
Mountain View 1 0
Shenzhen 1 0
Ashburn 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}

Share

Share to:

Comment

Related Items