Mismatches in Russian Nominal Ellipsis - - PDF document

9/4/20 1

Mismatches in Russian Nominal Ellipsis

Maria Polinsky

1

Мы давно называемся взрослыми И не платим мальчишеству дань, И за кладом на сказочном острове Не стремимся мы в дальнюю даль. Ни в пустыню, ни к полюсу холода, Ни на катере ...к этакой матери. Но поскольку молчание - золото, То и мы, безусловно, старатели. Промолчи - попадешь в богачи! Промолчи, промолчи, промолчи! (Александр Галич)

2

2

Outline

• Setting the stage: main moving parts
• Russian NPE and PNE: feature matches and

mismatches

• Analysis of the data
• Conclusions and outstanding questions

3

3

Setting the stage

4

4

9/4/20 2

Two main characters

• Ellipsis
• phi-features in the noun phrase

5

5

Ellipsis

• Basic assumptions

§ the ellipsis site has structure § ellipsis involves PF-deletion triggered by a syntactic feature (Merchant 2001)

6

6

Ellipsis

• Ellipsis is regulated by an identity condition of

some sort:

• Pat left but I don’t know when Pat left.
• They bought a foreign car, but I don’t know which

foreign car they bought.

• Kim is a birdwatcher and her siblings are

birdwatchers too.

7

7

Identity condition on ellipsis

• Strict syntactic identity is too strong of a

requirement (Merchant 2001; Kroll 2019; Rudin 2019; Ranero 2020)

• Either the Board grants the license by December 15
• r it explains why it di

didn dn’t grant the license by December 15.

• No student finished the exam except Kim di

did finish the exam.

8

8

9/4/20 3

Identity condition on ellipsis

• Semantic identity is too weak
• synonymy is insufficient to license ellipsis

*Jamie is no longer a bachelor and Peter did get married too (Omer Preminger’s example)

• voice and argument structure mismatches are rarely

available (Chung 2013, Merchant 2013)

9

9

Identity Condition

• Syntactic condition is too strong
• Semantic condition is too weak
• Q1: What should be included in the Identity

Condition?

10

10

Identity Condition

• Syntactic condition is too strong
• Semantic condition is too weak
• Q1: What should be included in the Identity

Condition?

• Preview of the answer: the identity condition

should be formulated in reference to features

11

11

Ellipsis contexts to consider today

• Noun phrase ellipsis (NPE)

Mary’s daughter and Jane’s daughter are friends.

• Predicate-nominal ellipsis (PNE)

Kim is a linguist and Pat is a linguist too.

12

12

9/4/20 4

Features inside a noun phrase

• Decompositional approach to noun phrase

(independently motivated):

• Roots are acategorial (Harley 2014; Merchant 2019,

a.o.)

• Gender is on the categorizing n which combines with

√ROOT and carries formal gender features (Kramer 2015)

• Status of NumP is less clear (Ritter 1998; Picallo

2019)

13

13

Decomposing noun phrases

14

14

Noun phrase decomposition and ellipsis

15

15

Decomposing noun phrases: Outstanding questions

Do all derivational affixes have the same status?

• Yes, they are all functional heads (Marantz 2001;

Marvin 2003)

• Yes, they are all roots (Lowenstamm 2015)

16

16

9/4/20 5

Test case: Russian gender

• Little n’s in Russian (possibly elsewhere in Slavic)

n[+F]: feminine gender feature, triggers feminine concord n[-F]: masculine gender feature, triggers masculine concord nØ1: no gender feature, triggers masculine concord nØ2: no gender feature, triggers neuter concord

17

17

Test case: Russian gender

• Q2: Russian gender
• a. what is the decompositional structure of the

Russian noun phrase (with the emphasis on gender and number)?

• b. what is the status of Russian affixes used to

derive gendered nouns?

18

18

Test case: Russian gender

• Q2a: what is the decompositional structure of

the Russian noun phrase (with the emphasis on gender and number)?

• Preview of the answer:
• It is simpler than you think
• It is more articulated than you think

19

19

Test case: Russian gender

Q2b: what is the status of Russian affixes used to derive gendered nouns?

• Preview of the answer:
• Not all derivational affixes are created equal

20

20

9/4/20 6

Test case: Russian gender

Q2b: what is the status of Russian affixes used to derive gendered nouns?

• Preview of the answer:
• Not all derivational affixes are created equal, so both

Marantz and Lowenstamm are partially right

21

21

Outline

• Setting the stage: ellipsis, phi-features, Russian
• Russian NPE and PNE: feature matches and

mismatches

• Analysis of the data
• Conclusions and outstanding questions

22

22

Russian NPE and PNE: Feature matches and mismatches

23

23

Section outline

• Number matches and mismatches
• Gender: three main classes
• More gender: focus on morphology

24

24

9/4/20 7

Number under ellipsis

Moi druz’ja kartežniki, i moj muž tože my friends card players and my husband too ‘My friends are card players, and my husband too.’ Moj muž kartežnik, i moi druz’ja tože my husband card player and my friends too ‘My husband is a card player, and my friends too.’

25

25

Number under ellipsis

[NUMBER] mismatches are available (and predicted, e.g., Saab 2019)

26

26

Number under ellipsis

*Oni bliznecy/ trojnjaški i moj muž tože they twins triplets and my husband too (‘They are twins/triplets, and my husband too.’) *Moj muž bliznec/trojnjaška i oni tože my husband twin/triplet and they too (‘My husband is a twin/triplet, and they too.’)

27

What is going on here?

27

Section outline

• Number matches and mismatches
• Gender: three main classes
• More gender: focus on morphology

28

28

9/4/20 8

Gender mismatches under ellipsis

• Greek

(Merchant 2014, Alexiadou 2015, Sudo & Spathas 2016)

• Spanish

(Depiante & Masullo 2001, Donatelli 2019, Merchant 2014,

Ranero 2019, 2020, Saab 2010)

• Portuguese

(Bobaljik & Zocca 2011)

29

29

Gender mismatches under ellipsis

• Greek (Merchant 2014, Alexiadou 2015, Sudo &

Spathas 2016)

• Spanish (Depiante & Masullo 2001, Merchant

2014, Donatelli 2019, Ranero 2019, 2020, Saab 2010)

• Portuguese (Bobaljik & Zocca 2011)
• To my knowledge, Slavic languages have not

been systematically investigated with respect to gender matches under ellipsis

30

30

Background

Masculine/feminine pairs of animate nouns fall into three distinct classes under NPE and PNE

• nouns that never license ellipsis of their

counterpart, regardless of gender

• nouns that license ellipsis both ways (M > F, F >

M)

• nouns in which the masculine noun of the pair

licenses ellipsis of the feminine version, but not vice versa (M > F, *F > M) (Merchant 2014)

31

31

Examples

• Class I: ellipsis impossible, common in kinship

terms (*John is a good uncle and Mary is a good aunt too) John [PredP is a good uncle] and Mary [PredP is a good aunt] too

32

32

9/4/20 9

Examples

• Class I: ellipsis impossible, common in kinship

terms (*John is a good uncle and Mary is a good aunt too)

• Class II: either element can antecede the other

(John is a good lawyer and Mary is a good lawyer too; Mary is a good lawyer and John is a good lawyer too)

33

33

Examples

• Class I: ellipsis impossible, common in kinship terms

(*John is a good uncle and Mary is a good aunt too)

• Class II: either form can antecede the other

(John is a good lawyer and Mary is a good lawyer too; Mary is a good lawyer and John is a good lawyer too)

• Class III: the masculine can antecede the feminine

but not the other way around (John is a good actor and Mary is a good actress too; *Mary is a good actress and John is a good actor too)

34

34

Let’s add Russian

Three main classes as in other languages

• Class I: car’/carica ‘tsar/tsarina’, deduška/babuška

‘grandfather/grandmother’, baran/ovca ‘ram/ewe’

• Class II: nominal advokat ‘lawyer’, doktor ‘doctor’,

avtor ‘author’, etc.; deadjectival dežurnyj/ dežurnaja ‘person on call’; epicene tupica ‘dunce’

• Class III: princ/princessa ‘princ(ess)’, ljotčik/ljotčica

‘pilot’, sportsmen/sportsmenka ’athlete’, idiot/idiotka ‘idiot’, krasavec/krasavica ‘looker, beauty’; žiraf/žirafa ‘giraffe’; most names of nationalities

35

35

Class I: Kinship terms

*Vera babuška, i Kostya tože V grandmother and K too (‘Vera is a grandma and Kostya too.’) *Kostya deduška, i Vera tože K grandfather and V too (‘Kostya is a grandpa and Vera too.’)

36

36

9/4/20 10

Class I: Nobility names

*Оrlov byl graf, O was count i Rostova tože and R too (‘Orlov was a count and Rostova was a countess.’) *Rostova byla grafinja, i R was countess & Lenin tože L too (‘Rostova was a countess, and Lenin was a count.’)

37

37

Class I: Animal names

*Tupogub byk, i Burёnka tože T bull & B too % Burёnka korova, i Tupogub tože B cow & T too

38

38

Class II: epicenes, professions, deadjectivals

F > M Maša tupica/advokat/rvan’/dežurnaja M dunce/lawyer/raggedy/on duty i Petja tože & P too M > F Petja tupica/advokat/rvan’/dežurnyj P dunce/lawyer/raggedy/on duty i Maša tože & M too

39

39

Class III

*Maša advokat-essa/učitel’-nica, M lawyer-FEM/teacher-FEM i Petja tože. & P too (‘Masha is a lawyer/teacher and Petya too.’) *Emma krol’č-ixa, i Freddi tože. E bunny-FEM & F too (‘Emma is a she-bunny, and Freddie too.’)

40

40

9/4/20 11

Gender under ellipsis: Main generalizations

41

Cl Class I Cl Class II Cl Class III M/ M/F F pairs Yes Yes Yes El Ellipsis Symmetrically bad Symmetrically good Asymmetrical; %M > F, * F > M

41

Comparing S, P , G, and R

• Recurring pattern of class asymmetries
• Class content varies across languages despite

some overlap

• Class content also varies across speakers (stay

tuned)

42

42

Section outline

• Number matches and mismatches
• Gender: three main classes
• More gender: focus on morphology

43

43

Adding morphology

44

Cl Class I Cl Class II Cl Class III M/ M/F F pairs Yes Yes Yes Mo Morpho hology re relating M an and F Suppletion; de derivational af affixes es Zero affixes; gender inflection on deadjectivals De Deriva vation

• nal

af affixes es El Ellipsis Symmetrically bad Symmetrically good Asymmetrical; %M > F, * F > M

44

9/4/20 12

Assumptions

• Defective paradigms can participate in ellipsis

(Abels 2018; Mendes 2020; Merchant 2014; Pertsova 2016); repair under ellipsis circumvents morphophonological gaps:

45

45

Assumptions

• Defective paradigms can participate in ellipsis;

repair under ellipsis circumvents morpho- phonological gaps: e.g., Russ. kočerga ‘firepoker’ *gen.pl

u Peti

• dna kočerga, a u nas pjat’ *gen.pl

by P

• ne firepoker.NOM

but by us five ‘Peter has one firepoker, and we have five.’

46

46

Gender-sensitive morphology

• Inflectional flavor: gender concord ending on

deadjectival nouns zavedujušč-ij zavedujušč-aja head-M head-F

• Derivational flavor

deriving F from M: -ka, -essa, -ša, -ica, -ixa, -(in)ja deriving M from F: -ak; suppression of –a, deletion

• f -ka

47

47

Some examples

• F from M

deputatka, deputatša, deputessa (<deputat) advokatka, advokatša, advokatessa, advokatixa, %advokatis(s)a, %advocatica (<advokat)

• M from F (non-productive)

ved’m-ak ‘witcher’ (<ved’ma) dojar ‘milkman’ (<dojarka) njan’ ‘babysitter’ (<njanja)

48

48

9/4/20 13

Morphological irregularities

M but no (regular) F: mudak ‘asshole’, bolvan ‘moron, dumbass’, pentjux ‘loser’, balbes ‘dumbhead’,

• lux ‘dullard, sap’, indejec ‘Native American’,

korejec ’Korean’, turok ‘Turk’, kovboj ‘cowboy’, lešij ‘goblin’, paxar’ ‘plowman’, rycar’ ‘knight’ F but no (regular) M: mašinistka ’typist’, manikjurša ‘manicurist’, njanja ‘babysitter’, povituxa ‘midwife’, mokrica ‘douchebag’, mymra ‘old stick’, bajaderka ‘temple dancer’

49

49

Morphological irregularities

M but no (regular) F: mudak ‘asshole’, bolvan ‘moron, dumbass’, pentjux ‘loser’, balbes ‘dumbhead’,

• lux ‘dullard, sap’, indejec ‘Native American’,

korejec ’Korean’, turok ‘Turk’, kovboj ‘cowboy’, lešij ‘goblin’, paxar’ ‘plowman’, rycar’ ‘knight’ F but no (regular) M: mašinistka ’typist’, manikjurša ‘manicurist’, njanja ‘babysitter’, povituxa ‘midwife’, mokrica ‘douchebag’, mymra ‘old stick’, bajaderka ‘temple dancer’

50

50

Morphological irregularities: Some examples

kovboj ‘cowboy’ ≠ kovbojka ‘plaid shirt’ korejec ‘Korean’ ≠ korejka ‘porkloin’ mašinistka ‘typist’ ≠ mašinist ‘engine driver’

51

51

PNE in native speaker Russian: Regular predicate nominals

1 2 3 4 5 6 7 8 Class I Suppl Class I Der i ved Class II deADJ Class II nominal Class III M > F F > M

nephew-niece

52

Likert scale, 1-7; 112 respondents ram--ewe

52

9/4/20 14

PNE in native speaker Russian: Regular predicate nominals

1 2 3 4 5 6 7 8 Class I Suppl Class I Der i ved Class II deADJ Class II nominal Class III M > F F > M

nephew-niece

53

Likert scale, 1-7; 112 respondents ram--ewe

• 53

PNE in native speaker Russian: Regular morphology

• The results support the division of gender pairs

into basic classes

• Minor asymmetry in Class II, with M being

preferred antecedent (we may need better measures…), especially with deadjectival forms

54

54

PNE in native speaker Russian: Irregular morphology on predicate nominal (cowboy ~ plaid shirt)

1 2 3 4 5 6 7

M > F F > M

Likert scale, 1-7, 120 respondents

55

55

PNE in native speaker Russian:

• nly one gender of predicate nominal

available (‘dumbass’ type)

56 Likert scale, 1-7, 120 respondents 1 1.5 2 2.5 3 3.5 4 4.5 5

F missing M missing

56

9/4/20 15

Summary of the empirical data

• Number mismatches are tolerated but not with

pluralia tantum

• Gender mismatches vary depending on noun

class and morphology

57

57

Outline

• Setting the stage: ellipsis, phi-features, Russian
• Russian NPE and PNE: feature matches and

mismatches

• Analysis of the empirical data
• Conclusions and outstanding questions

58

58

Toward an analysis

59

59

Main points

• Lexical vs functional categories in ellipsis
• Constraints on roots in ellipsis
• Constraints on features in ellipsis
• Different types of derivational affixes with respect

to ellipsis

60

60

9/4/20 16

Lexical constraint on ellipsis

• the sluice cannot contain any lexical items that

are not present in the antecedent (Ross 1969, Merchant 2001, Chung 2006, Saab 2008)

61

61

Lexical constraint on ellipsis

• the sluice cannot contain any lexical items that

are not present in the antecedent

• Strict Root Identity (Saab 2008; Ranero 2020):

syntactic roots in the antecedent and in the sluice must be strictly identical

62

62

Root Identity at work

• Genuine suppletive forms: no ellipsis

(ram-ewe, many kinship terms)

• One of the forms is absent (null root): no ellipsis

Although ellipsis often allows speakers to circumvent the material that is absent, that does not happen in the roots

• A pluralia tantum noun has a root different from

that of the singular form: no ellipsis

63

63

Explaining number mismatches

Two possible solutions:

• pluralia tantum and collective nouns include an

additional NUM projection, next to the root (Picallo 2018) [NumP [nP [NumP [√P

• there is only one NumP

, but pluralia tantum nouns have a dedicated root, and the singular form represents a different root; no root identity bliznec1 <PL> bliznec2 <SG>

64

64

9/4/20 17

Consequences of Root Identity

Phe Pheno nomeno non Fo Form 1 Fo Form 2 Genuine suppletion baran ‘ram’, deduška ‘grandpa’

• vca ‘ewe’,

babuška ‘grandma’ One of the roots is null (absent)

• lux ‘dumbass’

(M only) no corresponding feminine form Accidental root homophony bliznecy ‘twins’ koreec ‘Korean’ pilot ’pilot’ bliznec ‘twin (sg)’ korejka ‘porkloin’ pilotka ‘triangle- shaped cap’

65

65

What’s in the root?

• lack of root identity blocks ellipsis
• roots are not all the same with respect to

reference to mixed-gender groups in the plural

66

66

What’s in the root?

• lack of root identity blocks ellipsis, but roots are

not uniform with respect to gender mismatch in the plural

67

Coordination impossible Coordination possible */?Maria y Juan son mis padres. Maria y Juan son mis tios. (‘Maria & Juan are my parents.’) ‘Maria & Juan are my aunt and uncle.’

67

What’s in the root?

• lack of root identity blocks ellipsis, but roots are

not uniform terms of gender mismatch in the plural

68

Coordination impossible Coordination possible *Maša i Petja moi djad’ja/ djadi Maša i Petja moi plemjanniki. (‘M & P are my aunt & uncle.’) ‘M & P are my niece and nephew.’

68

9/4/20 18

djadja vs plemjannik

• Some roots that do not participate in ellipsis still

allow plural in reference to mixed-gender groups (plemjannik ‘nephew’)

• Such plurals are root-specific (not all nouns act

like this) but arbitrary (can vary arbitrarily across languages; cf. Russian plemjannik and English nephew; cf. Spanish tio and Russian djajdja)

69

69

What’s in the root?

plemjannk djadja Plural can apply to mixed- gender groups yes no Encyclopedic meaning component <MALE> of <FEMALE> included in root specification no yes Can combine with a null n yes no

70

70

What’s in the affix?

• Bottom line: nouns of different classes can

combine with different categorizing n heads

71

71

New approach to Identity Condition

• Identity Condition on Ellipsis:

Antecedent and material properly contained within the ellipsis site must be featurally non- distinct (Ranero 2019, 2020, building on Chomsky 1965)

72

72

9/4/20 19

Identity Condition via features

Antecedent feature Sluice feature Mismatch possible?

X[+Y] X[-Y]

NO

X[-Y] X[+Y]

NO

X[null] X[+Y]

YES

X[+Y] X[null]

YES

73

73

What’s in the affix?

• Class I (kinship, nobility, differentiated animals)
• Some roots include encyclopedic information

regarding the features <MALE> or <FEMALE> (as seen in the reference to mixed-gender groups)

• Gender mismatch is impossible both ways because

both the masculine and feminine counterpart are specified for [GENDER]

• in principle those roots that do not include

<MALE> or <FEMALE> can combine with null n

• but it is more informative to use the n with feature

[-F] or [+F]

74

74

Class I nouns decomposed

Identical root, ‘bunny’ krolik

nP[√ROOT32 n[-F]]

krol’čixa

nP[√ROOT32 n[+F]]

75

75

Class I nouns decomposed

Distinct roots: papa

nP[√ROOT64 n[-F]]

mama

nP[√ROOT59 n[+F]]

76

76

9/4/20 20

Ellipsis in class I (identical roots)

Antecedent feature Sluice feature Mismatch possible?

X[+Y] X[-Y]

NO

X[-Y] X[+Y]

NO

77

77

Class II nouns decomposed

• the form referring to a male includes the n head

that is not specified for gender, and the form referring to a female includes n[+F]

• “masculine” nouns can be used with female

referents:

• na naš

zavedujuščij kafedroj she our head department ‘She is our department head.’

78

78

Class II nouns decomposed

zavedujuščij

nP[√ROOT nØ]

zavedujuščaja

nP[√ROOT n[+F]]

79

79

Ellipsis in Class II

Antecedent feature Sluice feature Mismatch possible?

X[null] X[+Y]

YES

X[+Y] X[null]

YES

80

80

9/4/20 21

What about epicene nouns?

Anton tupica, i Anna tože tupica A dunce & A too ‘Anton Is a dunce and Anna is too.’ Anna tupica, i Anton tože tupica A dunce & A too ‘Anna is a dunce and Anton is too.’

81

81

What about epicene nouns?

• Epicene nouns belong to class II
• they always include nØ

(cf. Atkinson 2015 for similar conclusions on French)

• they can optionally have a higher n specified as [+F]
• r [-F] (discourse gender)
• this is reflected in their variable agreement and does

not affect ellipsis

82

82

What about epicene nouns?

• Epicene nouns belong to class II
• they always include nØ

(cf. Atkinson 2015 for similar conclusions on French)

• they can optionally have a higher n specified as [+F]
• r [-F] (discourse gender)
• this is reflected in their variable agreement and does

not affect ellipsis Petja gor’kaja/gor’kij p’janica, P bitter.FEM/bitter.M drunkard a Maša načinajuščaja p’janica but M beginner.FEM ‘Petya is a complete drunkard, and Masha is only starting as a drunkard.’

83

83

Class III nouns

Anton is an actor and Anna is too *Anna is an actress and Anton is too

84

84

9/4/20 22

Approaching Class III nouns

• l-forms vs f-forms (Harley & Noyer 1999, a.o.)
• l-form affixes can change stress pattern and

trigger phonological rules

85

XP 5 X √P 5 √21 √35

f-forms l-forms

85

Approaching Class III nouns

86

• l-forms vs f-forms (Harley & Noyer 1999, a.o.)
• Properties of l-affixes:
• Stress-sensitive
• Combine with bound stems (as opposed to words)
• Closest to the stem (in terms of relative position)

(Harley & Noyer 1999; Lowenstamm 2015; Cremers et al. 2018)

86

Class III nouns: A proposal

• “masculine” nouns include a null n head

As with Class II nouns, the ”masculine” noun can be used in reference to a female:

Ona advokat she lawyer ‘She is a lawyer.’

87

87

Class III nouns: A proposal

• “masculine” nouns include a null n head
• “feminine” nouns include an l-affix (ditto for a

subset of class I nouns, esp. nobility terms)

88

advokat essa/ka/ša ‘female lawyer’

√

povar

√ Ixa ‘female cook’

88

9/4/20 23

Ellipsis in Class III

• Root Identity is not maintained, ellipsis is

impossible [nP [√P √povar √ixa]] ‘female cook’

<cook> <FEMALE>

89

89

Ellipsis in Class III

• Root Identity is not maintained, ellipsis is

impossible [nP [√P √povar √ixa]] ‘female cook’

<cook> <FEMALE> *Maša ploxaja povarixa, M bad.FEM cook.FEM a Petja xorošij povar but P good.M cook.M

‘Masha is a bad cook but Petya is a good one.’

90

90

What about M > F ellipsis in Class III?

*Maša ploxaja povarixa, M bad.FEM cook.FEM a Petja xorošij povar but P good.M cook.M Petja ploxoj povar, P bad.FEM cook.FEM a Maša xorošij povar but M good.M cook.M

91

91

What about M > F ellipsis in Class III?

Ellipsis involves the removal of identical material (root with a null gender categorizer, as in Class II), but not the relationship between a masculine and a feminine noun

92

an anteceden ecedent sl sluice den denotat ation el ellips psis OK OK? [nP ROOT n∅] [nP ROOT n∅] Either F or M yes

92

9/4/20 24

From three classes to two

If this analysis is on the right track, there are only two classes of nouns in ellipsis:

• One of the nouns has a null categorizing head,

making ellipsis mismatches possible

• All nouns have specified categorizing heads;

ellipsis possible only if the features are the same

93

93

PNE in native speaker Russian: Regular predicate nominals

1 2 3 4 5 6 7 8 Class I Suppl Class I Der i ved Class II deADJ Class II nominal Class III M > F F > M

nephew-niece

94

Likert scale, 1-7; 112 respondents ram--ewe

• 94

Outline

• Setting the stage: ellipsis, gender, Russian
• Russian NPE and PNE: feature matches and

mismatches

• Analysis of the empirical data
• Conclusions and outstanding questions

95

95

Conclusions and outstanding questions

96

96

9/4/20 25

What I have discussed

• Identity conditions on ellipsis
• Evidence for l- and f-affixes in Russian
• Gender encoding in Russian

97

97

Taking stock

• Apparent mismatches in Russian ellipsis are not

uniform

• They can be comprehensively accounted for by

two separate conditions on roots and affixes: a lexical and a featural condition

98

98

Back to the original questions

• Q1: What should be included in the Identity

Condition? Condition on roots Condition on categorizing heads

99

99

Root Identity Condition on ellipsis

• Every ROOT in the ellipsis site must have an

identical ROOT in the relevant portion of the antecedent and vice versa (Saab 2008, Ranero 2020)

100

100

9/4/20 26

Identity Condition on ellipsis

• Antecedent and material properly contained

within the ellipsis site must be featurally non- distinct (Ranero 2019, 2020)

101

101

Identity conditions on ellipsis

• Mismatches in Russian ellipsis are not uniform

and are comprehensively accounted for by two separate conditions on roots and affixes

• Altogether, the Russian data provide novel

support for identity conditions on ellipsis

102

102

Identity conditions on ellipsis

• Mismatches in Russian ellipsis are not uniform

and are comprehensively accounted for by two separate conditions on roots and affixes

• Altogether, the Russian data provide novel

support for identity conditions on ellipsis; elements of the semantic and syntactic conditions are needed in tandem

103

103

Decomposing Russian noun phrases

Q2: Russian gender

• a. what is the decompositional structure of the

Russian noun phrase (with the emphasis on gender and number)?

• b. what is the status of Russian affixes used to

derive gendered nouns?

104

104

9/4/20 27

Decompositional structure of the Russian noun phrase

[NumP [nP [nP [√P Root ⇧ ⇧

discourse categorizing gender gender head head

Root homophony accounts for some restrictions

• n ellipsis

105

105

Russian “gender” affixes

• At least some derivational affixes are roots (l-

forms), supporting the notion that derivational affixes are not uniform in nature

(contra Lowenstamm 2015, Marantz 2001, in support of Creemers et al. 2018)

106

106

Russian “gender” affixes

• Outstanding questions:
• More precise diagnostics needed to determine which

Russian affixes are roots and which are functional head

• Variation across speakers, esp. with respect to the

suffixes –ka and –ica

• Ongoing change in spoken Russian as new feminine

derivations (feminitives) enter the language

107

107

Side effect: Homophony

• Root homophony (cf. pluralia tantum)
• Affix homophony:

a reasonable expectation is for the same form to be available as an l-form and f-form; variation across speakers should also be expected

108

108

9/4/20 28 Gender mismatches in other languages: Same as in Russian?

• With specified identity conditions and possible

homophony of bound forms, gender and number mismatches can be reduced to a minimum

• Three classes established across languages are

no longer needed

• Can gender and number mismatches reported in
• ther languages be similarly reduced?

109

109

THANK YOU!

• Special thanks to Ruth Kramer, Terje Lohndal,

Ora Matushansky, Omer Preminger, and Rodrigo Ranero

• Many thanks to the MultiGender Project at CAS,

Oslo for their support

110

110

Selected references

Chomsky, N. 1965. Aspects of the theory of syntax. Cambridge: MIT Press. Creemers, J., Don, J., & P . Fenger. 2018. Some affixes are roots, others are

• heads. NLLT 36: 45-84.

Lowenstamm, J. 2015. Derivational affixes as roots: Phasal spell-out meets English stress shift. In A. Alexiadou et al. (eds.)The syntax of roots and the roots of syntax, 230–259. Oxford: OUP . Marantz, A. 2001. Words and things. Ms., MIT. Marvin, T. 2003. Topics in the stress and syntax of words. PhD diss., MIT. Ranero, R. 2019. Voice mismatches in Kaqchikel sluicing. WSCLA proceedings. Ranero, R. 2020. The eventive core is not special in ellipsis. Ms., UMD. Ross, J. R. 1969. Guess who? CLS 5: 252-286.

111

111

Selected references

Saab, A. 2008. Hacía una teoría de la identidad parcial en la ellipsis. PhD Diss, University of Buenos Aires. Saab, A. 2010. (Im)possible deletions in the Spanish DP . Iberia 2.2: 45-83. Saab, A. 2019. Nominal Ellipsis. In J. van Craenenbroeck and T. Temmerman (eds)., The Oxford Handbook of Ellipsis, 526-561. Oxford: OUP .

112

112