——__.

 

 

LIBRARY
Michigan State
\ University

090;)“;

 

 

 

This is to certify that the
thesis entitled

THE ACQUISITION OF THE DISTINCTION BETWEEN
UNACCUSATIVE AND UNERGATIVE VERBS BY
ENGLISH SPEAKERS AT THE LOWER
INTERMEDIATE LEVEL OF L2 RUSSIAN

presented by

Elena Yuryevna Selezneva

has been accepted towards fulfillment
of the requirements for the

Department of Linguistics and
MA. degree in Germanic, Slavic, Asian and
African Languages

Major Professofi Signature
09 022‘ «0 5"”

Date

 

 

MSU is an Affirmative Action/Equal Opportunity Institution

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K‘IProj/Acc&Pres/ClRC/DaleDue.indd

THE ACQUISITION OF THE DISTINCTION BETWEEN UNACCUSATIVE AND
UNERGATIVE VERBS BY ENGLISH SPEAKERS AT THE LOWER
INTERMEDIATE LEVEL OF L2 RUSSIAN

By

Elena Yuryevna Selezneva

A THESIS
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS
Department of Linguistics and Germanic, Slavic, Asian and African Languages

2005

ABSTRACT
THE ACQUISITION OF THE DISTINCTION BETWEEN UNACCUSATIVE AND
UNERGATIVE VERBS BY ENGLISH SPEAKERS AT THE LOWER
INTERMEDIATE LEVEL OF L2 RUSSIAN
By

Elena Yuryevna Selezneva

This study aims to explore the acquisition of split intransitivity in L2 Russian by English
learners in relation to the Split Intransitivity Hierarchy (SIH) (Sorace 1996) and the
Unaccusative Trap Hypothesis (UTH) (Oshita 2001). The study involved two groups of
subjects: one group of L1 English speakers studying Russian for 3- 4 years at MSU and a
control group of Russian natives: graduate students at MSU. Subjects were tested on
their knowledge of the behavior of the intransitive verbs with respect to Genitive of
Negation construction (Brown, 1999). The magnitude estimation technique was used for
the elicitation of acceptability judgments. The results show that both groups are subject to
SIH in their judgments on unaccusatives; however, their judgments on unergatives do not
pattern according to the predictions. The results did not support the SIH and instead

supported UTH.

ACNOWLEDGMENTS

I would like to express my sincere gratitude to University Distinguished
Professor Susan Mary Gass, my academic adviser, for her encouragement and significant
help.

I wish to acknowledge Professor David K. Prestel and Assistant Professor
Christina Schmitt, the members of my advisory committee, for their interest in my work
and suggestions. I appreciate the time, feedback and support you provided me throughout
the process.

I would like to express my deepest appreciation to Dr. Thomas L. Maleck for
providing financial support that made it possible for me to live, study and finish this
work, as well as for his friendly attitude and guardianship over all these years.

I wish to thank Ellen Maleck, who together with Dr. Thomas L. Maleck,
actually made it possible for me to come to MSU to broaden my horizons.

I want to express sincere gratitude to the rector of Volgograd State University
of Architecture and Civil Engineering, Vladimir Aleksandrovich Ignatiev, and the
Director of International programs Of the same university, Vera Vladimirovna
Galishnikova, for their support, understanding, and sympathy.

I would like to thank all my friends at MSU: international, American and
Russian students for their participation in the research experiments as well as for their
friendly support and inspiration.

Last, but not least, I want to thank my family for their understanding and faith.

iii

TABLE OF CONTENTS

ACKNOWLEDGMENTS ........................................................................ iii
TABLE OF CONTENT ........................................................................... iv
INTRODUCTION ................................................................................... 1
CHAPTER 1

THE PHENOMENON OF UNACCUSATIVITY
1.1 Split Intransitivity ................................................................. 6

1.2 Research on acquisition of
unaccusative — unergative distinction ..................................... 12

1.3 The Unaccusative Trap Hypothesis (Oshita, 2001)

and Split Intransitivity Hypothesis (Sorace,1995) ..... l ................ 16
CHAPTER 2
PHENOMENON OF UNACCUSATIVITY IN RUSSIAN ................................. 23
2.1 Diagnostics for unaccusativity in Russian .................................. 23
2.2 Genitive of Negation ............................................................. 27
2.3 L1 acquisition of GM ........................................................... 30
CHAPTER 3
METHODOLOGY AND PROCEDURE ......................................................... 33
3.1 Research questions and Hypotheses ...................................... 33
3.2 Method ............................................................................. 35
3.2.1. Participants .......................................................... 35
3.2.2 Materials .............................................................. 36
3.3 Procedure ........................................................................... 39

iv

3.4 Analysis .............................................................................. 40

CHAPTER 4
RESULTS OF EXPERIMENT .................................................................... 42
4.1 Results and Discussions ..................................................... 42
4.2 Research Questions and Answers .......................................... 50
4.4 Heritage Learners ............................................................... 56
CHAPTER 5
CONCLUSIONS AND RECOMMENDATIONS ........................................... 63
APPENDICES ..................................................................................... 69
APPENDIX A
QUESTIONAIRE FOR ENGLISH NATIVE SPEAKERS ....................... 70
APPENDIX B
QUESTIONAIRE FOR RUSSIAN NATIVE SPEAKERS ...................... 71
APPENDIX C .
VOCABULARY TEST ................................................................. 72
APPENDIX D
INSTRUCTION GIVEN TO SUBJECTS .......................................... 74
APPENDIX E
GRAMMATICALITY JUDGMENT TEST ......................................... 76
REFERENCES .................................................................................. 80

INTRODUCTION

The field of language acquisition, in general, and second language acquisition
(SLA), in particular, has become one of the most fascinating and exciting fields in
language studies in recent years. Though people have always valued the knowledge of
foreign languages, until the middle of last century most researches focused on language
pedagogy and not language acquisition.

In the 195 Os the focus shifted from the improvement of teaching methods to
learning, from the teacher and the way he should present the language phenomena to the
learner and the way he learns the target language structures (Mitchell & Myles 2004).

Language acquisition can be defined as a process through which the learner
creates a mental representation of the language to which he/she is exposed (Culicover &
Nowak 2003, 24).

Current theories Of language learning and language acquisition are based on
research in a wide variety of fields, including psychology, neurolinguistics, sociology,
anthropology, and linguistics.

The linguistic approach to language learning has become one of the most
influential since Chomsky presented his fundamental theory of the Language Faculty,
which can explain simultaneously why human languages have certain properties and why
children are so efficient in learning languages.

According to Chomsky’s theory words are combined into sentences according to
a set of principles, which constitute Universal Grammar (UG), an innate biologically

endowed language faculty (Chomsky 1986). UG determines in advance what grammars

can or cannot be. It places requirements on the form Of grammars, providing an inventory
of possible grammatical categories and features; UG also constrains the functioning of
grammar by determining the nature of the computational system. The computational
system includes the kind of Operation that take place, as well as principles that grammars
are subject to.

It is widely accepted that UG constrains native language (L1) acquisition. The
evidence that native language acquisition is possible only because children are born with
an innately —— determined language faculty (UG) is considerable. The evidence that the
same innate ability is involved in Second Language Acquisition by older learners is less
clear.

In general the term Second Language Acquisition (SLA) refers to the process of
learning another language after the native language has been learned, both in a classroom
situation as well as in more “natural” exposure situation (Gass & Selincer 2001, 5).

Hence, the acquisition of L2 in adulthood is different fi'om the acquisition of L1
both in the context of acquisition and the nature of development. The adult learners
already know (at least) one other language: we can assume that the initial state of the
child and the adult is not the same. For adults other components of mind have already
maturated. For children, language acquisition and the development of other cognitive
abilities occur simultaneously. Input can be different too, for adults it may involve
written language along with spoken language. The final state is also different: many
adults display L2 performance that differs from that of a native speaker.

All these differences lead to questioning the role of UG in second language

acquisition. Though, as correctly mentioned by Mitchell & Myles (2004), Universal

Grammar is a theory of natural languages, Claiming that it plays no part in second
language acquisition would be problematic if we take “natural language” as a language
spoken by humans.

First and second language acquisition is similar in many ways. Second language
learners also go through fairly rigid stages when acquiring certain constructions in the
second language. For example, in English, both L1 and L2 learners will acquire
morphemes such as, -ing, plural, irregular past, regular past, -s plural, ‘s possessive in
relatively the same order.

Mitchell & Myles (2004) point out that the question of whether or not UG is
available to the second language learners was crucial fifteen years ago. Now it has been
replaced by more focused questions: how UG interacts with other modules involved in
language learning and what the role of the first language settings is.

The strongest argument for the accessibility Of UG to second language learners is
the fact that they do not produce interlanguages that violate UG principals. That is why
most of the work is concentrated on the availability of the parameters.

White (2003) points out, that L2 learners develop IL (interlanguage) grammars
that are different from the grammars of N85 but are still constrained by UG, and that this
is due to the L2 input interaction with UG and the L1 grammar.

According to Hawkins (2001a) and Herschensohn (2000) the evidence of access
to UG in L2 acquisition reveals itself as follows:

0 Learners can acquire functional categories which do not exist in the L1
(for example English L2 learners acquire clitics in French)

0 There are no wild interlanguage grammars

o Learners exhibit knowledge that goes beyond input
0 In some cases steady state is native-like.

To demonstrate convincingly that interlanguage grammars are constrained by
principals of UG, the phenomenon under investigation must be underdetermined by the
L2 input. That is, it must not be something that could be acquired by Observation Of the
L2 input, including statistical inferencing based on frequency of occurrence, on the basis
Of analogy, or on the basis of instruction; and the phenomenon should work differently in
the L1 and L2. That is, it must be underdetermined by the L1 grammar as well. In this
way, transfer of surface properties can be ruled out as an explanation of any knowledge
that L2 learners attain (White 2003).

The issue Of grammatical analysis of IL system has attracted the attention Of many
researchers. Special attention has been devoted to L2 acquisition of grammatical
properties of specific lexico-syntactic classes of verbs - intransitives (Sorace 1993a,b,
1995, 1996, 2000; Sorace & Shomura 2001; Oshita 2000, 2001; Hirakawa 1995, 1999,
2001; Montrul 2004). There are two types of intransitive verbs: unaccusatives and
unergatives. They possess distinctive semantic and syntactic properties. A number of
studies have found that L2 learners are aware Of these distinctions but they have
difficulties defining the appropriate syntactic realization of the distinction, and this can
lead to non- native grammars even among learners at the advanced level,

The phenomenon Of unaccusativity has been extensively investigated in variety of
languages. Several hypotheses have been proposed to account for the acquisition of this
phenomenon by L2 learners: the Unaccusative Trap Hypothesis (Oshita 2001) and Split

Intransitivity Hypothesis (Sorace 1995). Among them there are studies supported either

the former (Oshita 2001) or the latter (Sorace 1993a, b, 1995, 1996, 2000; Sorace &
Shomura 2001; Montrul 2004). All the studies were devoted to Western European or
Non-European languages and none of the studies examine L2 acquisition of this
phenomenon in Slavic languages. The evidence from Slavic languages will give more
validity to either Of these hypotheses as it is another language group.

The present study attempts to find evidence of these hypotheses in L2 Russian.
An interesting question in the context is how L2 learners establish the language specific
realization of the distinctions between verb types. The diagnostics for unaccusativity in
Russian differs from the diagnostics in English. Hence, the transfer of surface prOperties
is ruled out. English has a subtle evidence for unaccusativity, so the phenomenon can be
considered to be underdetermined by the L1 grammar.

The phenomenon under investigation, the unaccusatives — unergatives
distinction in L2 Russian can provide good evidence for access to UG as well. It
satisfies all the criteria mentioned above to demonstrate that interlanguage grammars are
constrained by principals of UG.

The thesis is organized as follows: in Chapter 1, I will introduce the phenomenon
of split intransitivity and talk about recent research on the acquisition of this distinction
by L2 learners; in Chapter 2, I will touch upon the diagnostics for unaccusativity in
Russian and explain in more detail the Genitive of Negation which is my primary
interest; in Chapter 3, I will discuss the methodology of the experiment and in Chapter 4,
will present its results with some observations. Finally, in Chapter 5, there is a
conclusion, which will be focused mostly on the results of the experiment with NNS and

NS, and a discussion of the possibilities for future research.

CHAPTER 1

THE PHENOMENON OF UNACCUSATIVITY

In section 1.1 of this Chapter I will introduce the phenomenon of split
intransitivity: the unaccusatives —— unergatives distinction between intransitive verbs. In
1.2 I will discuss recent research on the acquisition of this distinction by L1 and L2
learners. In 1.3 I will explain the two Hypotheses proposed to account for the acquisition

of the unaccusative — unergative distinction.

1.1 Split Intransitivity

The Unaccusative Hypothesis (UH) was first introduced by Perlmutter (1978).
UH stipulates that there are two types of intransitive verbs across languages: unaccusative
and unergative, which have distinct syntactic and semantic properties. The original
formulation of UH states that the difference between unergatives and unaccusatives is
syntactically represented and semantically encoded. Later, two alternative approaches to
unaccusativity, syntactic and semantic, were developed in response to UH.

The supporters of the semantic approach consider the difference between the two
verb types reducible to the notion of agentivity and telicity (Dowty 1991, Van Valin

1990). Unergative verbs (i.e. run, swim, work) are usually agentive and denote a process

without an end point, while unaccusative verbs (i.e. die, arrive, exist) are nonagentive and
telic and denote an action with a particular end point.

The followers of the syntactic approach (Burzio 1986) consider the difference to
be solely structural. Unergatives have an external argument, agent in the subject position;
Unaccusatives have an internal argument, a theme that moved to the subject position and

left a trace. Consider Figure 1:

Figure 1. Deep structure (D-structurel) of two types of intransitive verbs

 

 

 

1 a. Unergative 1. b. Unaccusative
D1 -Syntax: S S
/\ l
Subj \i’ VP
/\ ,
V V Obj
Lexicon: V (Ext Arg) V (Int Arg)

 

 

In D- Syntax, the argument of the unergative verb is outside of verbal phrase (VP)
(Figure 1a) and the argument of an unaccusative verb (Figure 1b) is inside VP. Thus, the
single argument of an unaccusative verb is syntactically equivalent to the direct Object of
a transitive verb, whereas the single argument of an unergative verb is syntactically
equivalent to the subject Of a transitive verb.

If we look at the surface structure (S-structurez), they seem to be the same,

because the arguments are in the subject position in the sentences. Consider Figure 2:

 

l D-Structure: a level where the underlying or deep structure of the sentence is represented. (White, 2003,
275)

2 S-Structure: a level where the surface structure of the sentence is represented. (White, 2003, 228)

Figure 2. S-structure of two types of intransitive verbs

 

 

 

a. Unergative b. Unflusative
S-Syntax: /S\ /S\
Subj V’ Subj V’
Np | Np /\
V V Obj

 

In current Minimalist theories, the single argument of both unaccusatives and
unergatives is generated in the VP domain and moves to the functional domain. The
distinction between the subject of the unaccusative verb and the subject of the unergative
verb is explained through the asymmetry in terms of light — v, a semi-functional head
(Chomsky 1995). The subject of the unergative verb is introduced by v, and the subject of

the unaccusative verb belongs to the lexical verb. Consider Figure 3:

Figure 3. Syntactic structure of two types Of intransitive verbs

 

 

 

3 a. Unergative 3. b. Unaccusative
vP vP
/\ l
NP v ’ v’

 

 

 

The movement of constituents is performed through the operation attract. In this
Operation, a chain forms between the target of movement and the goal. A trace is an
unpronounced copy of the moved element, which is an antecedent of the trace
(Chomsky 1995).

Levin & Rappaport (1995) support the original idea Of Perlmutter (1978) and
bring both approaches together. They argue that unaccusativity is syntactically a unified
phenomenon, but emphasize that the distinction between unaccusatives and unergatives is
semantically encoded. As an example, they use the case of resultative phrases in English.

Consider examples la-c:

1. a. The floor had been swept clean of debris. Passive
b. The river froze solid. Unaccusative
C. *The officers laugh helpless Unergative (Levin & Rappaport, 1995).
S-Structure subjects of passive (1a) and unaccusative verbs (lb) can predicate
resultatives. This situation is impossible with unergative verbs (1C). Levin and Rappaport
point out that the explanation for this can be found in the structural difference between
the S-structure subjects of these verbs. S-structure subjects of unaccusatives and passives
are logical Objects. S-structure subjects of unergatives are logical subjects. This would
not be possible if the unaccusative — unergative distinction was only a semantic
property. At the same time, the projection of arguments is regulated by specific linking
rules which are sensitive to semantic notions, such as change of location, and state.
Resultative constructions are incompatible with all types of unaccusative verbs. Stative

verbs such as remain and verbs of inherently directed motion such as come, go, and

arrive, are not involved in the resultative construction. Consider examples 2 a-b:

2. a. Carla remained in the country bored.

b. Willa arrived breathless (L&RH 1995256).

In (2a), the sentence does not mean that Carla became bored by remaining in the
country, and (2b) it does not mean that Willa became breathless as a result of arriving.
These facts illustrate that the syntactic structure of unaccusatives is not the only factor
effecting their performance with respect to different syntactic tests for unaccusativity;
semantic characteristics must also influence their behavior.

The difference between these two classes of verbs is considered universal. Two
distinct types Of intransitive verbs can be found in all languages with some language
specific syntactic and morphological features. For example, in English, unaccusative
verbs can appear both with existential subjects and in resultative constructions, whereas
unergatives cannot (Perlmutter 1978). In Italian and Dutch unaccusative verbs can appear
with the perfective auxiliary essere, zijn (be) whereas unergative and transitive verbs take
the auxiliary avere, hebbe (have) (Burzio 1986, Sorace 1993a, b). In Spanish transitive
and unaccusative verbs allow bare plurals in postverbal position and absolutive
constructions3 while unergative verbs do not behave in this way (Montrul 2004). In
Japanese, the adverb takusan (a lot) has two distinct interpretations. With unaccusative
verbs it quantifies the argument of the verb, and with the unergatives, it quantifies the
activity described by the verb and not the argument (Hirakawa 2001).

From the literature review that follows it is evident that unaccusative verbs pose

a learning problem, as are identical to unergative verbs in their surface syntactic

 

3 Clauses with a past participle that modify a postposed noun phrase and agree with it in number and
gender. (Montrul, 2004)

10

structure. The question that arises is how language learners acquire the distinction of
these two very similar types of verbs.

A substantial body of research in SLA is devoted to the acquisition Of this
distinction. For most linguists, the acquisition of unaccusativity represents a “poverty of
stimulus” problem. It is not Clear how learners identify verbs belonging to different types
of intransitives. Both of them have one argument but they do not behave identically in
regards to different syntactic constructions. This distinction poses even more serious
problems for L2 learners. According to Juffs (1998), this type of verbs is
underrepresented in second language teaching materials and, as a result, the distinction is
not taught in the language classroom.

In general, very few errors have been reported in L1 acquisition Of this
distinction. Children begin to distinguish these two types of verbs very early
(Babyonyshev, et a1, 2001; Borer and Kenneth Wexler 1992; Von Hout 1996). In contrast
to L1 acquisition, a lot of mistakes have been found in L2 acquisition even among
advanced learners. The difficulties with unaccusatives were found in L2 English (Oshita,
2001, Hirakawa, 1995, Zobl, 1989) as well as in other then English L25: Italian and
French (Sorace, 1993), Japanese (Hirakawa, 1999), I will discuss the research in more
details in 1.2.

Two hypotheses were proposed to account for these difficulties. Oshita (2001)
proposes the Unaccusative Trap Hypothesis. He argues that at the first stage of
acquisition, learners employ one single argument linking rule for both types of

intransitive verbs. The rule is not sensitive to any semantic notion. Later, learners begin

ll

to distinguish unaccusatives and unergatives and reconstruct interlanguage grammar. At
this stage of acquisition they become sensitive to semantic notion as well.

Sorace (1995) proposes the Split Intransitivity Hypothesis. It states that
unaccusative and unergative verbs display various behaviors across and within languages.
This behavior depends on their position in the hierarchy, which distinguish core
unaccusative and unergative verbs from peripheral verbs. The hierarchy identifies the
notion Of telicity a core notion for unaccusatives and agentivity for unergatives.
Peripheral verbs can vary in their behavior across languages. I will talk in more detail
about these two hypotheses in 1.3.

Having briefly introduced the phenomenon, I move to a discussion of the
previously conducted research on the acquisition of the unaccusative - unergative
distinction in L2 and the hypotheses proposed to account for the acquisition of this

distinction.

1.2 Research on the acquisition of unaccusatives and unergatives

Much research has been done in the past few years on the acquisition of the
unaccusative —— unergative distinction by L2 learners. Most research has been devoted to
L2 English or other European languages. However, a number of studies were done on
non-Indo-European languages (Sorace & Shomura 2001, Hirakawa 1999, 2001). The
existence of this phenomenon in non - European languages confirms the universality and

the validity of UH for different language types.

12

A great deal of research is devoted to the analysis of incorrect morphology.
Recent findings show that L2 learners are aware of the argument structure distinction
between unaccusative and unergative verbs, though the syntactic tests for unaccusativity
vary from language to language.

The research on the acquisition of the unaccusative — unergative distinction
began with English, arguably, the most well described language in the field of linguistics.
Zobl (1989) was one of the first scholars who described these types of mistakes in an L2
and attracted attention to the problem.

English does not have the difference in the auxiliary selection as a diagnostic of
unaccusative and unergative, unlike Italian and Dutch (See Sorace 1993a, b and Burzio
1986). Nevertheless, L2 English learners regardless Of L1 treat it as if observing this
distinction. The unaccusative verbs sometimes appear with auxiliary be and a past
participle in interlanguage English. This is a common L2 production error in speech and

writing, as well as an error in grammaticality judgment tasks. Consider examples in 3:

3. a. *An accident was happened. Montrul (2004, 242)
b. *Most of people are fallen in love and marry with somebody
(Zobl 1989, 204).
C. *My mother was died when I was just a baby (Zobl 1989, 204).

In (3), learners (3a L1 - Spanish, 3 b,c - Chinese) moved the Theme argument into
the subject position, but ungrammatically have added passive morphology. Zobl (1989)
proposes an explanation for this type of error with unaccusative verbs by pointing out the
similarities between unaccusatives and passives. Both of them have an internal argument
(logical objector theme) but no external argument (logical subject or agent). In addition,

the internal argument moves to the subject position. Zobl assumes, following Burzio

l3

(1981) that across languages unaccusatives have the argument structure, shown in (4).
ESL learners adopt it as the null hypothesis.

4. [VP [v Open NP the door]]

Learners of English note the parallel with the passive. The theme originates as
an internal argument of the verb and raises to a subject position for reasons of case. They
take the passive morphology as the means to indicate the NP raising. Once they acquire
the passive, unaccusatives are subject to this syntactic rule of NP movement because the
passive in English indicates a connection between the internal argument and the subject
position.

Zobl concludes that the syntactic rule never fully replaces the lexical rule,
which explains why learners produce both native-like and ungrammatical unaccusatives
in English.

Production and acceptance of causative errors with unaccusative and unergative

verbs is another common error for L2 English learners. Consider examples in 5:

5. a. *The man disappeared the ball (Montrul 2004, 242).
b. *I was just patient until dried my clothes (Zobl 1989, 204).
c. *Sometimes comes a good regular wave (Zobl 1989, 204).

Speakers fail to move the Theme argument in (5) to the specifier of IP in order to get
Nom Case.

Montrul (2004) mentions another very common mistake in L2 English:
incorrect rejection Of unaccusative verbs in grammatical noun phrases - verb sentences
(NP-V) in judgment tasks. However, she also notes the correct use of this sentence

pattern in all other cases (6):

14

6. John left

Oshita (2000) presented data from the database of Written English, which he
searched for ten unaccusatives and ten unergatives verbs. Out of 941 token of
unaccusatives, there were 38 sentences involving passivized unaccusatives. Such errors
totaled only 4% of all errors with unaccusatives. Out of 640 token of unergative verbs,
there was only one error of this type, which total 0.15%. Errors of this kind practically do
not exist. The findings show that unergatives and unaccusatives are represented as two
distinct classes in interlanguage grammars; otherwise the same type of errors would
appear with both types of verbs. Therefore, learners are aware that an unaccusative verb
poses only the internal theme — argument (White 2003).

White (2003), discussing Oshita’s findings, concludes that L2 learners correctly
represent the argument structure of unaccusative verbs as having an internal theme
argument. Learners map this argument to the correct position within VP. Usually, L2
learners experience problems relating argument movement in the syntax, to the
morphology associated with NP movement.

However, it is not clear how learners distinguish which particular verb belongs
to which class. Moreover, the reason for the classification is unclear, because they receive
only positive evidence. To account for the difficulties of acquisition of the distinction two

different hypotheses were proposed. Now, I will move to their discussion.

15

1.3 The Unaccusative Trap Hypothesis (Oshita 2001) and Split Intransitivity

Hypothesis (Sorace 1995)

A number Of studies have shown that unaccusatives cause problems for L2
learners of English with various L1 backgrounds even at high intermediate and advanced
levels of interlanguage development (Oshita 2000, Hirakawa 1995).

Several hypotheses were proposed to account for the difficulty of
unaccusative —- unergative acquisition. Two of these hypotheses are relevant to the
present study: the Unaccusative Trap Hypothesis (Oshita 2001) and the Split
Intransitivity Hypothesis —SIH- (Sorace 1995).

Oshita (2001) stipulates that the distinction of two types Of intransitive verbs,
unergatives (with underlying subjects) and unaccusatives (with underlying objects), may
not exist at early stages of L2 acquisition. According to him, both types are syntactically
represented as unergatives. This idea is referred to as the Unaccusative Trap Hypothesis
(UTH). It offers general predictions about the grammatical development Of IL systems.
Learners at low and even intermediate levels of proficiency will use one-argument verbs
(unaccusatives and unergatives) in the same syntactic environments. If a syntactic error
arises, it is more likely to affect unaccusative than unergative. This happens because
unaccusative is initially misanalysed as unergative and undergoes reanalysis later.
Syntactic errors exclusively observed with unaccusatives should become apparent only
after this verb class is correctly distinguished from the unergative class in the IL lexicon.

Similarly, the target syntactic structure is possible only with unaccusatives. For example,

16

the there-insertion structure can be fully acquired only after the two Classes of
intransitives are correctly differentiated in the IL lexicon.

Oshita (2001) proposes a three-stage process for the acquisition of
unaccusativity. The first stage is characterized by a Single-Argument Linking Rule. This
rule causes learners to treat all intransitive verbs as unergatives. Although this does not
pose a problem when the verb is indeed unergative, the rule leads to non-native-like forms
when the verb is unaccusative. It is strictly a rule that is found only in L2 grammars.

In the second stage, the learner becomes aware of the intransitivity split and
produces forms that deviate from those found in the L1 such as the passivization of
unaccusatives, reluctance to accept NP-V word order, and the production of V-NP
structures.

In the third stage of the process, learners “achieve a native grammar” (Oshita
2001, 289). This stipulation contradicts the evidence shown by other researchers (see
Zobl 1989, Hirakawa 1999). One of the characteristic properties of the acquisition of
unaccusativity in L2 is the persistence of errors, even at very advances stages of learning.
This means that the acquisition of near-native grammars is possible to be achieved. It
may require a longer period of time in contact with the target language grammar.

Sorace (1995, 2000) has proposed SIH to account for the systematic differences
within the syntactic classes of unaccusative and unergative verbs. The hierarchy
distinguishes core unaccusative and unergative verbs from more peripheral verbs. The
proposed hierarchy places crosslinguistically unaccusative verbs on one end, and
invariably unergative verbs on the other. Languages differ in the point along the

hierarchy that separates unaccusatives from unergatives, and these two categories of

17

intransitive verbs are distinguished in each language by differences in syntactic behavior.
Sorace and Shomura (2001) argue that the difficulty in acquiring the split between
unaccusatives and unergatives is caused by the problem of systematical linking “a
multicategoriai lexical-semantic level to a necessarily binary syntactic level...” (Sorace
& Shomura 2001 , 249). This hierarchy, which is based on aspectual parameters, identifies
the notion of “telic dynamic Change” (Sorace & Shomura 2001, 249) at the core of
unaccusativity, and “agentive nonmotional activity” (Sorace & Shomura 2001, 249) at
the core of unergativity. The extremes of the hierarchy thus, consist of maximally distinct
core verbs ——verbs of change of location (i.e., arrive) and verbs Of agentive nonmotional
activity (i.e., work) — which consistently display unaccusative or unergative
characteristics, respectively. In contrast, peripheral verb types between the extremes are
prone to variable syntactic behavior. The overall hierarchy of split intransitivity is

represented in Figure 4 (Sorace,1995):

figure 4.
Split Intransitivity Hierarchy

 

 

Unaccusative (least variation)
l

 

1. Change of Location

 

2. Change of Condition

 

3. Appearance

 

4. Continuation of preexisting condition

 

5. Existence

 

6. Uncontrolled process

 

a. emission

 

b. involuntary reaction

 

7. Controlled motional process

 

 

8. Controlled nonmotional process I)

 

 

 

 

Unergative (least variation)

 

18

Peripheral verb types include (arranged in order of closeness to the core): (1).
verbs denoting indefinite change in a particular direction (i.e., rise), (2) change Of
condition (i.e., wilt), (3) appearance (i.e., appear), (4) continuation of a preexisting
condition (i.e., stay), and (5) states (i.e., exist, suflice). Peripheral verbs which are closer
to the unergative core include verbs denoting (7) motional processes (i.e., swim), and
various kinds of uncontrolled processes (such as (6) body functions [i.e., sweat]),
(6a) involuntary reaction (i.e., tremble), and (6b) emission (i.e., rattle). The hierarchy
does not include dyadic verbs alternating with transitive variants (i.e., break, increase),
which are weakly unaccusative and display unergative behavior in some languages (see
Sorace 2000).

The problem with hierarchies of this sort, as Montrul (2004, 243) notes, is that it
is difficult to find counter examples and to make sharp and clear distinctions among
categories of grammatical elements. She stipulates that peripheral verbs may be subject to
event-type shifts though sometimes they can be either telic or atelic, as in 7:

7. a. I ran for an hour.

b. I ran to the store.

Montrul (2004, 243) points out that they can receive multiple argument realization
depending on other elements in the sentence like arguments, adverbials, clitics, or
prepositional phrases. In turn, this grammatical variability is reflected in how individual
speakers may conceptualize these verbs. By contrast, core verbs are not Open to such
variation, and this is the strongest part of the prediction. Montrul (2004) interprets the
hypothesis as if there should be observed a trace with core unaccusatives and no trace

with core unergatives. However, with the more peripheral verbs individual speakers’

19

production and acceptance may vary greatly. Learners’ judgments will be less sure with
peripheral verbs than with core verbs (see also Sorace 2000 and Sorace & Shomura
2001)

Montrul (2004) presents data that provide the psycholinguistic evidence
supporting SIH proposed by Sorace (2000). She examined 35 native Spanish speakers
and 44 English speaking learners of Spanish in regards to on-line processing and the
acquisition of unaccusative—unergative distinction. The results of the use of Off-line
measures such as absolute—scale grammaticality judgment task suggest that native and
non-native speakers have more determinate judgments with core verbs and display
inconsistent judgments with less core and peripheral verbs.

To summarize, the unaccusative—unergative distinction is cross-linguistic and can
be found in all languages with languages implementing different syntactic and
morphological behavior to show the difference. There is a discussion whether the
distinction should be purely syntactic (Burzio 1986), purely semantic (Dowty 1991, Van
Valin 1990) or if it should be syntactic and semantic simultaneously (Sorace 2000 and
Levin & Rappaport 1995).

The L1 and L2 learners can learn the difference between these two types of
intransitive verbs from the linguistic data. The question arises from how learners discover
this difference if given only positive evidence, 1. e., which particular verb belongs to
which class, and the reason for the classification. The unaccusative—unergative distinction
is not taught in the classroom and it is not clear how L2 learners distinguish these types
of verbs. The evidence from the L1 cannot always help as the diagnostic since the

unaccusative—unergative distinction varies from language to language. Two hypotheses

20

were proposed to account for the fact that L2 learners distinguish unaccusative vs
unergative verbs UTH (Oshita 2001) and SIH (Sorace 1995).

UTH stipulates that both types of verbs are syntactically represented as
unergatives. The acquisition of the distinction occurs in three stages. In the first stage,
Single-Argument Linking Rule causes Ieamers to treat all intransitive verbs as
unergatives. In the second stage, the learner is aware of the intransitivity split. In the third
stage of the process, learners “achieve a native grammar” (Oshita 2001, 289). What
contradicts the evidence from other researchers as one Of the characteristic properties of
the acquisition of unaccusativity in L2 is the persistence of errors, even among advanced
learners.

SIH distinguishes core unaccusative and unergatives verbs from peripheral verbs.
These two categories of intransitive verbs are distinguished in each language by
differences in syntactic behavior. This hierarchy, which is based on aspectual parameters,
identifies the notion of “telic dynamic change” at the core of unaccusativity and that of
“agentive nonmotional activity” at the core of unergativity.

The present study aims to investigates if there is evidence that supports either of
these hypotheses in Russian. UTH states that both types of intransitive verbs are
represented as unergatives in IL grammars on the early stage. To support the hypothesis
we need to show that L2 learners use both types of verbs in the same syntactic
environment. The syntactic structures, possible only with unaccusatives, are acquired
only if two classes of verbs are correctly distinguished. The present study will address all

these issues.

21

In order to investigate if L2 learners follow the hierarchy proposed by Sorace,
core and peripheral unaccusative and unergative verbs will be used for the experiment. If
L2 learners follow the proposed hierarchy, they will be more confident in their judgments
with core unaccusative and unergative verbs and less sure with peripheral. They will
consider ungrammatical examples less ungrammatical with peripheral verbs and
grammatical example less grammatical.

The purpose of this study is to investigate acquisition of the unaccusative-
unergative distinction by native speakers of English in L2 Russian. English is not the
easiest language in which to study these matters, since evidence for unaccusativity in
English is subtle. Russian, on the other hand, provides a strong test for unaccusativity
with the genitive-of-negation (GN) construction.

In the next chapter, I will talk about the syntactical peculiarities of GN

constructions and other diagnostics for unaccusativity.

22

CHAPTER 2

PHENOMENON OF UNACCUSATIVITY IN RUSSIAN

In this chapter I will talk about the unaccusative—unergative distinction in
Russian. There are four traditional diagnostics for unaccusativity in Russian: 1) po-
phrases, 2) na- phrases, 3) Locative Inversion, 4) Genitive of Negation (GN). In what
follows I will present information on all four (2.1) although, only the fourth one (GN)
will be used in the study and presented in more details in 2.2. The chapter will also

include partial conclusions.

2.1 Diagnostics for Unaccusativity in Russian

Several syntactic and morphological diagnostics for unaccusativity have been
proposed for Russian. However, these diagnostics are quite distinct from those proposed
for other European and Non-European languages. If you recall from the previous chapter,
different languages have different syntactic and morphological features. For example, in
English, unaccusative verbs can appear both with existential subjects and in the
resultative constructions whereas unergatives cannot (Perlmutter 1978). In Italian and
Dutch unaccusative verbs can appear with the perfective auxiliary essere, zijn (be)
whereas unergative and transitive verbs take the auxiliary avere, hebbe (have) (Burzio
1986, Sorace 1993a, b). In Spanish transitive and unaccusative verbs allow bare plurals in
postverbal position and absolutive constructions while unergative verbs do not behave in

this way (Montrul 2004). In Japanese, the adverb takusan (a lot) has two distinct

23

interpretations. With unaccusative verbs it quantifies the argument of the verb, and with
the unergatives, it quantifies the activity described by the verb and not the argument
(Hirakawa 2001).

We can not use auxiliary selection as a diagnostic for unaccusativity in Russian
since there are no auxiliaries in the past tense. The past participle agreement will not
provide relevant information either since all verbs in the past tense agree with their
subjects, regardless of their argument structure. Consider examples in 8:

8. a. Unaccustive
Otvet prishel
answer-Mask Sg arrived-Masc Sg
‘The answer came.’
b. Unergative
Anna zvonila.
Anna-Fem Sg called-rem Sg
‘Anna called.’

There are four traditional diagnostics for unaccusativity in Russian: l)po-phrases,
2) na- phrases, 3) Locative Inversion, 4) Genitive of Negation (GN).

Babby (1980), Schoorlemmer (1995) and Pesetsky (1982) proposed distributive

po-phrases in Russian as a diagnostics of unaccusativity. Pesetsky argues that distributive

po-phrases are limited to direct Objects and subjects of unaccusative predicates, as shown

in (9):
9. a. Transitive
J a dal mal’chikam po jabloku.
I gave boys-Dar ( po) apple-Dar
‘I gave the boys an apple each’ (Pesetsky 1982, 69).
b. Unaccusative

Po jabloku upalo s kazhdogo dereva.
apple-Dar fell from each tree-Gen
‘A (different) apple fell from each tree’ (Babby 1980, 45).

24

c. Unergative
*V kazhdoy kvartire smeyalos’ po-mal’chiku
in each apartment laughed (po) boy-Dat
‘A (different) boy laughed in each apartment’ (Schoorlemmer 1995, 33).

P0 assigns Dat. Case when it appears with the subjects of a transitive verb, as shown
in (9a) above. When subjects of unaccusative predicates appear in po-phrases, as in (9b),
the verb displays a lack of subject—verb agreement, which is marked by the impersonal
morphology—o ending in the past tense. Po phrases are ungrammatical with unergative
verbs.

A second diagnostics discussed by Borik (1995) and Schoorlemmer (1995)
analyses the behavior of verbs with the prefixes na- (V a lot) and pere-/po- (V all).
Unlike standard verbal prefixes in Russian, these prefixes are quantifiable and affect the
interpretation of the internal argument of the verbs. Borik and Schoorlemmer show that a

quantification of this type applies only to internal arguments: NPs in the direct object

position with transitive verbs and subject position of unaccusative verbs as in (10):

10. a. Transitive (Borik 1995, 32-33)
Deti nakupili (mnogo) knig.
Children-Nom bought a-lot books-Gen
‘The children bought a lot of books.’

b. Unaccusative

Mnogo travy naroslo v parke.
a-lot grass-Gen grew in park

‘A lot of grass grew in the park.’

c. Unergative

*Mnogo detej naigralo v parke.
many children-Gen played in park
‘Many children played in the park.’

25

A third diagnostics is Locative Inversion. This diagnostics has been discussed by
Babyonyshev (1996). Babyonyshev argues that Locative Inversion is only possible with
unaccusative predicates, as shown below in 11-12 (examples from Babyonyshev 1996):

1 1. Unaccusatives

a. V uglu valjalas’ kurtka.

in corner lay jaCket-Nom

‘In the corner lay a jacket.’

b. V sadu rosli tri rozy.

in garden grew three I’OSCS-Nom

‘In the garden grew three roses.’

12. Unergatives

a. *Sebe pod nos napeval Petja.

self under nose sang Petja-Nom

‘To himself sang Petya.’

b. *V kvartire svistit Vanja.

in apartment whistles Vanya-Nom

‘Vanya is whistling in the apartment.’
Babyonyshev (1996) notes that Russian only allows inversion with unaccusative
predicates to maintain a neutral interpretation in the discourse, despite relatively free
word order.

The evidence for unaccusativity in Russian is extensive and diverse. However,
there are several reasons to eliminate mentioned above diagnostics from the current
experiment. First, the structures discussed above are difficult to acquire as evidenced by
continued errors even in advanced stages of learning. Another reason for not including
these structures for the present study is that they are less frequent in speech and thus, it
will provide less evidence for L2 learners.

The fourth diagnostic for unaccusativity is the Genitive of negation (GN) and this

is my primary interest. In contrast to the po- and na- construction, it is widely used in

26

every day language and can give positive evidence for its identification and acquisition
by a language learner.

Pesetsky (1982) was the first to argue that the GN construction could be used
productively as a diagnostic for unaccusativity in Russian. Like direct objects, subjects of
unaccusative verbs can receive Genitive Case under negation. Brown (1999) and Bailyn
(1997) discussed the fact that GN is only licensed on NPs within the scope of sentential

negation and generally receives an existential interpretation.

2.2 Genitive of Negation

A significant topic in Russian syntax which got a lot of attention in recent years is
the problem of genitive of negation. Genitive of negation (GN) in Russian is a
phenomenon in which an argument that bears structural Case in an affirmative sentence
can surface bearing Genitive Case (Gen) in the corresponding negative sentence.
Consider examples in 13:
13. Otvet /*Otveta prishel/ *prishlo
AnswerNom/Gen came Mask! Neut Sg
The answer came.
Otvet-a/ Otvet ne prishlo/ prishel
AnswerOcn lNom neg came Neut/ Mask Sg
The answer didn’t come.
The phenomenon has received considerable attention in the literature both in

linguistics and language acquisition (see Brown 1999, Bailyn 1997, Pesetsky 1982,

Babyonyshev et al 2001).

27

Brown (1999) points out that Gen can only occur on a non-oblique VP-internal
argument of a negated verb: subjects of unaccusative verbs and objects of transitive verbs
(see 14 a—b). However, it is disallowed on the subject of both transitive and intransitive
verbs and direct objects of verbs when assigning some oblique case (Dat, Instr), (see 14

c-d) (all examples are taken from Brown 1999):

14. a. Internal argument of Transitive
Ya ne pishu stixov
I neg write poems Gen.
I don’t write poems.

b. Internal argument of Unaccusative
Otvet-a ne prishlo
AnswerOen neg came Neut Sg
The answer didn’t come.

0. External argument of Transitive
*Studenta ne *pisal /*pisalo
Student Masc Gcn/ neg wrote Masc Sg lNeut Sg

Student ne pisal /*pisalo
Student Masc./Nom neg wrote Masc. Sg /Neut Sg
Student didn’t write

d. External argument of Intransitive (unergative)
Studenty/* Studentov ne spyat’
Students Masc./Nom/Pl /students Masc/Gen/Pl neg sleepspPI
Students don ’t sleep.
e. Oblique Direct Object
Vladimirov ne upravljaet fabrikoj /*fabriki
Vladimirov Neg direct factory Fem Instr/*factory Fem Gen
Vladimirov doesn’t direct the factory.
The GN is accompanied by the impersonal agreement morphology. As illustrated

in (14b), when the subject of a negated sentence takes Gen case, the predicate bears the

third person singular and neuter agreement morphology.

28

However, as shown in (15), when the same subject takes Nominative (Nom) case,
the agreement is kept intact.
15. Otvet ne prishel
Answer Nom neg came Mask Sg
The answer didn ’t come
As shown in examples (14b) and (15), Gen seems to apply optionally when the

sentence has an unaccusative verb. Moreover, Gen can alternate with other cases.

Consider table 1 (Brown. 1999):

Table l. Altemation of Gen with other cases

 

 

 

 

 

Verb Type Gen alternates with
1. transitive Acc
2. unaccusative Nom
3. unergative ~

 

 

In constructions with transitive verbs, Gen can alternate only with ACC and never
with Nom. In constructions with unaccusatives verbs Gen can alternate only with Nom
and never with ACC. Gen is never used with unergative verbs.

However, as both Brown (1999) and Bailyn (1997) point out, and the translations
to examples in (14b, 15) here repeated as (16a,b) respectively, indicate, the two
alternatives in such cases have distinct interpretations. In particular, the GN variants have
an existential or indefinite interpretation, whereas the Nominative counterparts have a
definite interpretation.

16. a. Otvet-a ne prishlo
Answer Masc Gen neg came Neut Sg
No answer came
b. Otvet ne prishel

Answer Masc Nom neg came Masc Sg
The answer didn’t come

29

 

In the existential copula construction in (17), the subject of a negative existential
sentence cannot bear Nom-case and Gen case seems to apply Obligatorily.
17. a. Na stole est’ zurnaly.
on table be magazines PI. Nom
‘ There are magazines on the table’
b. Na stole net zurnalov/*zurnaly.
on table neg-be magazines PI Gen/ *Nom
There are no magazines on the table’ (Bailyn 1997, 86)

The phenomenon under investigation provides good evidence for testing the
acquisition of the unaccusative — unergative distinction by L2 learners. Though GN has
received considerable attention from syntactic and semantic fields, very few researchers
have looked at the acquisition of the distinction by L1 learners (Babyonyshev et a1 2001),
in the manner presented in 2.3 and no research has been devoted to the L2 acquisition of

it. Thus, the present study initiates the promising investigation in the second language

acquisition of the unaccusative — unergative distinction.

2.3. Research on L1 acquisition of GM.

Recent research has shown that children begin to distinguish the unaccusative —
unergative types of verbs very early (Babyonyshev, et a1, 2001; Borer and Kenneth
Wexler 1992). Von Hout (1996) stipulates that children already have all the necessary
knowledge about the syntactic distinction but they need to learn the semantic
classification by distributional evidence. Babyonyshev, et a1, (2001), Borer and Kenneth
Wexler (1992) point out that certain syntactic structures are still not acquired and this fact

poses a problem for the production of adult - like structures with unaccusative verbs.

30

Babyonyshev, et al (2001) investigate the acquisition of the split intransitivity
phenomenon in L1 Russian by Children ranging in age from 3 to 6. The results show that
they master GN with the direct Object of transitive verbs earlier than they master it with
intransitive verbs. By the age of three they are able to distinguish these two types of
verbs.

The present study tests the hypothesis of Borer and Wexler (1992), which
stipulates that young children lack the ability to represent A-Chains (argument chain) that
link thematic subjects and object positions. According to Baker (1988), semantic role
labels are associated with syntactic positions. The idea is referred to as a Uniformity of
Theta —role assignment Hypothesis (UTAH). Consider example 18:

18. a. Was Opened the door.

b. The dOOI'i was opened (Babyonyshev, et a1 2001, 8)

Children consider 18b ungrammatical, though the structure is grammatical in
adult language. Children seem to be unable to assign any O-role to “door” which is in the
subject position and its O-role belongs to the object position.

Results of an experiment by Babyonyshev, et al (2001) have shown that Russian
children at age three understand the use of GN. They can use the subtle semantic
distinction, such as wide vs. narrow scope of negation, or nonspecific vs a specific object
(Babyonyshev, et al 2001, 25). The results of the children’s performance with
unaccusative verbs are the most interesting for us. Statistical analysis reveals the
differences in the response to two types of unaccusative verbs. The mean number of GN

responses for regular unaccusatives is 45% and 47% with bleached" unaccusatives. The

 

’ Bleached unaccusative: unaccusatives ( including existential be) that require Gen case on their argument
regardless of the specificity of it (Babyonyshev, et a1, 2001 2001,13)

31

significance of the difference between means reveals that the children treat unaccusative
verbs as unergatives. There is an age difference in the responses. The means for in the
group of children older that 4 is higher than those of 4 and younger. .50 vs .40 the
difference is 10%. For bleached unaccusatives the difference is even higher, 31% and
highly significant t (28) = 3.21, p < .001 Children’s performance on the unaccusatives
improves with age.

The study agrees with the results of other studies in L1 acquisition of this
phenomenon. Most of them indicate that children acquire the syntactic distinction
between unaccusative and unergative verbs before age 4 (Snyder, Hyarns & Crisma 1995;
von Hout 1996). In the next chapter I will talk about research questions, design and

methodology of the study.

32

CHAPTER 3

METHODOLOGY OF THE EXPERIMENT

In this chapter, I will discuss the research questions and hypotheses as well as the

methodology and procedure used in the research.

3.1.

Research questions and Hypotheses

Several research questions were proposed for the present study.

1.

Do L2 learners of Russian differentiate the unaccusative and unergative verbs by
allowing GN on unaccusatives and disallowing it on unergatives?

Do L2 learners of Russian differentiate the unaccusatives and unergatives
depending on the position of monadic verbs on Sorace’s SIH?

DO L2 learners distinguish better the unergative syntactic behavior of the verbs
denoting motional processes and worse the behavior of the verbs denoting non-
motional processes?

DO L2 Ieamers distinguish better the unaccusative syntactic behavior of the verbs
denoting change of location and worse the behavior of the stative verbs?

Do L2 learners show a stronger preference for grammatical sentences over
ungrammatical sentences with core unergative and unaccusative verbs, and a
weaker preference with peripheral verbs if they recognize the difference at all?

Do grammaticality judgments of L2 learners differ from those of L1 speakers?

33

7. Do L2 learners analyze unaccusative verbs as unergatives, following Single
Argument Linking Rule, as predicted by Oshita’s UTH?

8. Have L2 learners achieved near-native proficiency in distinguishing unaccusatives
and unergatives?

9. Will L2 Ieamers allow the Genitive Case in all the structures since it is Optional?

Several specific predictions were made regarding the use of the genitive of
negation by L2 Ieamers in accordance with the general predictions made by UTH and

SIH hypotheses.

1. Learners will differentiate the unaccusative and unergative verbs by allowing
GN on unaccusative and disallowing it on unergatives.

2. Learners will differentiate the unaccusatives and unergatives depending on the
position of monadic verbs on Sorace’s SIH.

3. Learners will distinguish better the unergative syntactic behavior of the verbs
denoting motional processes and worse the behavior of the verbs denoting non-
motional processes.

4. Learners will distinguish better the unaccusative syntactic behavior of the verbs
denoting change of location and worse the behavior of the stative verbs.

5. Learners will show a stronger preference for grammatical sentences over
ungrammatical sentences with core unergative and unaccusative verbs, and a
weaker preference with peripheral verbs if they recognize the difference at all.

6. The grammaticality judgments of L2 Ieamers will differ from those of L1

speakers.

34

7. L2 Ieamers will analyze unaccusative verbs as unergatives, following the Single
Argument Linking Rule, as predicted by Oshita’s UTH.

8. L2 Ieamers have not achieved near-native proficiency in distinguishing
unaccusatives and unergatives.

9. Learners will not allow the Genitive Case in all the structures since it is

Optional.

3.2. Method

3. 2. 1. Participants

The first group of participants consists of college students: 20 native speakers of
English ranging in age from 18 to 29 years (mean 23.5); 8 males, 12 females. All of them
had learned Russian in a classroom setting for at least 2.5 years but with little exposure to
Russian outside the university environment. The second group of participants consists of
6 heritage Ieamers. The data from them will be discussed separately. The third group of
participants consists of 11 Russian native speakers, age 23 to 32 (mean 27.5) who are
enrolled in various graduate academic programs at MSU. The length of stay in the US.
ranges from 4 months to 7 years (mean 3.9 years). The subjects have strong connections
to their native language and country. They speak Russian every day and visit their native
country every year. One of the subjects had spent only four months in the US. and still

has strong connections with Russia and the Russian language.

35

3. 1.2. Materials

Questionnaire: All subjects were asked to fill out background questionnaires
that were slightly different for native speakers of English and Russian (Appendix A, B).
The Questionnaire helped to eliminate subjects with L1 other than English and provide
some background information about the possible interfering factors such as living abroad
for a long time (for both L1 and L2 speakers), and lack of connection with the Russian
language and Russia (for L1 speakers).

Vocabulary test: One week before the experiment, all subjects in Group 1
completed a vocabulary test containing 44 lexical items, 22 verbs and 22 nouns
(Appendix C), where among the target words there were distractors. Among verbs there
were 11 target (unaccusative and unergative verbs) and 11 distracters. All nouns can be
considered distracters as the primary goal of this study is unaccusative and unergative
verbs.

Subjects were asked to provide not only the translation of the words but to
indicate if (a) they knew the word, (b) they had seen the word before but were not entirely
sure of its meaning, (c) they definitely didn ’t know the word. The self report procedure is
taken from Smith (2004). The test determined their familiarity with the lexical items in

the task and helped to eliminate the possible lexical difficulties.

Magnitude Estimation (ME) .' magnitude estimation is a technique that can be used
for the elicitation of acceptability judgments. This method was originally developed in
psychophysics and recently applied to the measurement of linguistic acceptability. ME

was used for the first time on acceptability judgments by Sorace (1992) to compare the

36

results with those obtained using more familiar techniques. The major reason to employ
ME to linguistic data elicitation is that it gives the opportunity to obtain more finely
grained data as there is only one restriction on the informant from the nature of the scale
itself.

Standard procedures of linguistic acceptability judgments usually form category
scales (acceptable, *(unacceptable)) or limited ordinal (acceptable, ?(questionable),
*(unacceptable), ** (highly unacceptable)) scales. These scales require absolute rating
judgments, rather than relative ranking judgments. Ordinal scales do not provide
information about the relative distance between adjacent points on the scale. When ME
technique is used for the elicitation of acceptability judgments it provides data with more
accuracy.

The procedure is conducted as follows: participants are asked to assign any
number to the first sentence and then to assign proportional numbers to successive
sentences, so as to reflect the perceived degree of acceptability of each sentence with
respect to the first one. For example, if a sentence appears to be 10 times as acceptable as
the previous one, it should be assigned a number 10 times as large. Higher numbers
correspond to more acceptable sentences.

ME gives subjects the freedom to set up their own range and categories of
judgments, thus enabling them to make finer distinctions in their judgments. Also, all
judgments are proportional: subjects are asked to state not only if one sentence is better
or worse than another, but also how many times better or worse it is. The scale on which
judgments are made is open-ended. Subjects can always add a higher score or an

additional intermediate rating. The net result is that subjects are able to produce

37

judgments which distinguish only the differences they observe. The mean responses
obtained with the help of ME technique are not merely averages of independent samples
but the reflection of robust trends across subjects.

Featherston (2002) and Keller (2003) point out that the results Obtained with ME
technique exhibit more differentiation than standard judgments are assumed to contain
and ME makes it possible to treat linguistic acceptability as a continuum and directly
measures acceptability differences between stimuli.

To confirm the experiment one must be sure that the participants understand how
to perform magnitude estimations. Many ME experiments use a control condition in
which participants are asked to perform magnitude estimations of the length of a line.
These have been shown to be proportional to the actual length of the line. Unlike other
dimensions, linguistic acceptability has no Obvious “physical” scale to locate the
informants’ impressions.

Keller (2003) has recently argued that a power law of the same kind as that
Obtained in psychophysics can be derived by plotting estimated linguistic acceptability

against the number of linguistic constraints violated in the stimuli.

38

3. 3. Procedure

One week before the experiment the subjects were asked to complete a
vocabulary test.

The following week the participants completed a short practice session (limited
to 7—10 min) where they farniliarized themselves with the ME process. During the
session they were asked to judge line lengths (Appendix D).

The experiment itself consisted of a total of 90 sentences presented in random
order: 20 sentences with unaccusatives, 20 with unergatives, 10 with existential verbs and
40 with distractors. The sentences were organized in blocks of 10 to make the judgment
task easier for the subjects.

Primary attention was devoted to four types of verbs from Sorace’s SIH (see table
1): two unaccusative verbs: verbs of change of location npuxodumb (to arrive) and state
cymecmeoeamb (to exist); two unergative verbs: nonmotional activity pa6omamb (to
work) and motional activity nnaeamb (to swim). Subjects were asked to give their
grammaticality judgment of the sentences.

Schiitze (1996) indicates the possibility of the additional effect of neighboring
stimulus sentences on perception of target sentences. The marginal sentences can appear
more acceptable when preceded by much worse examples and less acceptable when
preceded by much better ones (Schtitze 1996, 155). One’s “absolute” judgments of a
stimulus can be exaggerated by difference between the stimulus and its context. This
influence by contrast can occur in the “intuitions” about grammaticality. In order to
remove this effect, different orders of sentences were used with both groups. The

sentences were the same for all the participants in this experiment.

39

There was an equal number of grammatical and ungrammatical sentences with
target structures and fillers. The block was started randomly with a grammatical or
ungrammatical sentence in both task variants. The order of blocks was randomized as
well (see Appendix E).

In my experiment subjects were asked to provide purely comparative judgments:
these are relative both to a reference item and the participant’s own previous judgments.
They can fix the value of the reference item themselves relative to which subsequent
judgments are made. They can add a higher score if they need to account for better
examples or a lower score if they need to account for a less grammatical example. The

results of the experiment will be discussed in Chapter 4.

3.4. Analysis

In order to examine the performance of L2 Ieamers regarding GN construction,
the present study employs four types of analysis: descriptive statistics, t-test and Two-
Factor Anova without Replication and post-hoe (Scheffé test). Only significant
differences are reported. p < .05 for all tests.

Two-Factor Anova without Replication will be performed on the results of each
group, on both unergative and unaccusative verbs. Verb types and NS or NNS groups are
variables in the analysis. If ANOVA shows a significant effect or interaction, the null
hypothesis is rejected and we need to find where the differences lie, not just that there is a
difference. A post-hoc (Scheffe') test will be performed to determine the location of the
difference. The Scheffe' test is customarily used with unequal sample sizes, thus it could

be used in the analysis of my data. Comparisons are performed both within categories

40

and across categories for each group. Once again, only significant differences at a minimum
significance level ofp .05 will be reported.

All statistics were performed using SPSS.

In Chapter 4 I will talk about the results of the experiment and the statistic

analysis of the data.

41

CHAPTER 4

THE RESULTS OF THE EXPERIMENT

In this chapter I will present the results of the experiment and the subsequent

discussion.

4.1. Results and Discussion

Questionnaire, L2 learners: The data showed that there are no interfering factors
to influence the results: none of the L2 learners had stayed abroad more than 5 weeks,
two subjects indicated that they speak with Russian natives on a regular basis. They did
not specify how regularly, however.

All subjects had been studying Russian not less than 2.5 years (mean 3.2). Eight
participants indicated that they speak Russian outside the classroom. Five subjects
indicated that they usually speak Russian with their friends; one — with Russians; one —
with her Russian boyfriend and one of the subjects indicated other.

There is a test to evaluate the level of proficiency of speakers of Russian as a
foreign language. None of the subjects took this test. Four subjects misunderstood the
question and interpreted it as if they were being asked about a test for their regular
Russian course.

All of the subjects had studied another foreign language: 14 of the participants

took Spanish, two German and four French.

42

Questionnaire, NS of Russian: The data of NS of Russian showed that there is a
factor that can influence the results of the survey. One of the participants had been in the
US. more than 7 years. This prolonged stay in the US. may influence the grammaticality
judgments of this participant. Overall, the length of stay of the participants from the
control group varied from 4 months to 7 years (mean 3.9). All subjects, except one that
had spent four months in the US, mentioned that they visit their home country every
year. These periodical visits to Russia extended from two weeks to two months. All
subjects indicated that they speak Russian every day with their friends, family, or

roommates.

Vocabulary test: The analysis of the data from the test helped to chose the lexical
items for the grammaticality judgment task. The selection of the vocabulary was based on
the frequency of verbs and nouns in the language and the high percentage (not less than
70%) of correct translations, provided by participants.

The data collected and shown (see next page Table 2), indicates that none of them
had problems with the vocabulary used in the grammaticality judgment test. However,
half of the participants were not confident about the meaning of the verb cyutecmeoeamb
“exist” and Opaxon “dragon” as they marked the column: “I have seen the word but I am
not sure of its meaning”. Nevertheless, they provided the correct translation of the words.
The reason for this may be simple guessing, as the phonological form of English word is
close to the Russian one. Another reason could be that they came across the word in the
text they read and remembered the meaning but they are not sure, as the word is not in
their active vocabulary. As the percentage of correct answers is very high, it is likely that

they do not experience lexical difficulties regarding these words.

43

For the reasons mentioned above, only four types Of verbs were chosen for the
actual test since the rest of verb types were not very well known by L2 learners. As a
consequence, unaccusative verbs proposed for the tests were very close to each other in
terms of SIH. This may lead to a lack of evidence for SIH support, because learners may
unifome treat these two types of verbs as core unaccusatives.

Table 2 represents the results of the vocabulary test and indicates the percentage
of the correct responses that motivated the choice of the vocabulary for the experiment.
As it was mentioned above, the selection of the vocabulary was based on the high
percentage (not less than 70%) of correct translations, provided by participants.

Even when a correct translation was given, participants were asked if they knew
the word or if they had seen the word but were not sure of the meaning. These results are

reflected in columns 4 and 5

44

Table 2. Translate from Russian into English

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Translate from Correct I know I have seen I don't
Russian into . translation the word but know
E . Translation the .
ngllsh word not sure of Its the word
meaning
1. MaJIb‘IHK (boy) 100% 100%
2. IIPICBMO (letter) 100% l 00%
3. mew (answer) 95% 70% 25% 5%
4. erarcOH (dragon) 79% 16% 63%*
5.01am (window) 100% 100%
6. cymecrnoearb (to exist) 68% 15 % 53%* 32%
7. morconarr (chocolate) 100% 100%
8. CTleCHT (student) 100% 100%
9. pafiorarb (to work) 100% 100%
10. Bpaq (doctor) 1 00% 84% 1 6%
11. npui’rrn ( to arrive) 84% 68% 16%" 16%
12.renerpaM1va (telegram) 100% 100%
13. mm. (to swim) 95% 69% 26%” 5%

 

 

 

 

 

 

 

*50% out of 63% wrote correct translation.
** The percentage of correct translation is around 70%.

Results of Grammaticality judgment task:
Descriptive Statistics

The analysis of the data revealed no significant differences among two categories
of unaccusative verbs: verbs of change of location and change of state for any group.
Similarly, no significant differences were obtained among two categories of unergatives
verbs: verbs of nonmotional process and verbs of motional process. Table 3 contains the

results of Descriptive statistics.

45

Table 3. Mean acceptability judgment on different verb types.

 

 

 

 

 

 

 

 

 

 

NS NNS
Mean SD Mean SD
Nom 10 O 9.73 1.03
Nonmotional process work
Gen 3.28 3 4.42 4.1
. Nom 10 O 9.56 1.3
Motronal process swim
Gen 3,28 3 4.82 4.2
Nom 9.9 1.5 5.74 4.2
Change of Location arrive
Gen 10 0 6.79 4.21
Nom 9.6 1.16 7.52 3.67
State exist
Gen 9.18 1.01 7.9 3.71

 

 

 

 

 

 

 

 

A graphical representation of the judgments of the two groups on unergative and
unaccusatives verbs is shown in Figures 4 and 5.

The GN construction is ungrammatical with unergatives. The graphs indicate that
the judgments on the two types of unergative verbs were very close in both groups. The
controls (NS group) clearly differentiate between grammatical Nom case and
ungrammatical Gen case, accepting the former and rejecting the latter (Figure 4), though
we have a rather high mean for acceptability of ungrammatical Gen. The reason for that
may be the judgments of grammaticality of one of the participants, who has stayed in the
US. for 7 years. Overall, his responses differ from the rest of the group by to 3 points.
We can stipulate that the length of stay in the foreign country can influence the
grammaticality judgments: the ungrammatical sentences are more accepted by him and
the grammatical ones are rated less when compared with the results of the other

participants.

46

Figure 4. Mean acceptability judgments on unergatives.

 

 

 

 

 

 

 

 

 

a: -__-7- #___ _ _- v,
I
s

10 if .-_ g

8 $1

i -". .«2:

Mean 5 . j ;,
acceptability , j "4Q: l
- i. ‘i‘ ‘ r3343 Ii
judgments 4 g; 1% l Nom case1
2 . - Rig, IGen case I
i - 3m? T";" 'i
0 ‘ 1

NS NNS NS NNS i

Non-motional motional

 

The NNS group followed the trend; though they are less sure about the verbs of
motional process, the difference is not significant (see Figure 4, above). They tend to
accept more ungrammatical sentences.

The graphical representation of the judgments of the groups on unaccusative
verbs is shown in figure 5 (next page). It reveals that the group of NS would prefer Nom
case to Gen in the constructions with state verbs. This result was not expected. The
participants tended to consider Gen “more grammatical” with location verbs and less with
state verbs. They indicated this behavior through their personal interaction with the
researcher. The opposite can be observed for Nom. However, the difference is not
significant. One of the possible explanations is the fact that these verbs are far from each
other on SIH and a state verb is associated with a more definite reading than change of
location verb (Recall: Gen argument of the verb has indefinite reading and Nom has

definite). Figure 5 shows the mean acceptability judgments on unaccusatives verbs.

47

Figure 5. Mean acceptability judgments on unaccusatives

 

 

Unaccusative verbs

 

l
|
I
l
I

 

 

 

 

 

 

10 .A

8 7:.

Mean 6
acceptability , - , [I
1 judgments 4 if; , I Nom case;
I 2 I Gen case I,

o ,
NS NNS NS NNS
Location change State

h _, fl _ r- _-_J

 

 

 

There is no significant difference in the judgments of the NNS group. They prefer

Gen more than Nom.

t-teSlS

In order to compare the verb types between the groups t-tests were performed.
The results showed a significant difference on one type of unaccusative verbs: change of
location (see Figure 6). t = 2.495, df=28, p < .05. The comparison of the sentences with
change of location verbs between the groups produces a significant difference.

There is no significant difference regarding the second type of unaccusatives
verbs: verb of state. t = 1.44, df=28, p < .05, ns. The fact that NSs prefer Gen over Nom
with verbs of Change of location, and Nom over Gen with the verbs of state might

attribute to the appearance of this significant difference. NNSs showed the same pattern

48

for both types of unaccusatives. This leaning towards Gen can be explained by means of
SIH: the core unaccusative verbs identify the notion of “telic dynamic change” and that

attracts Gen. State verbs express this notion less and N85 tend tO use Nom.

Two-F actor Anova and Scheffe'

One of the focuses of the research was to find out whether native speakers and L2
Ieamers respond differently to different classes of verbs: unaccusative and unergative and
different types of verbs depending on their position in the unaccusativity hierarchy
proposed by Sorace (2000). It was predicted that L2 Ieamers should be more sure in their
judgments with core unaccusative and core unergative verbs and less sure with peripheral
verbs as less core verbs are supposed to display variable syntactic behavior within and
across languages.

The statistical analysis applied to all the verbs confirmed these trends for the two
groups. A repeated measures ANOVA shows that the difference in the verb type is
significant F (3, 28) = 6.157, p<.001. There are significant differences between groups,
F (3, 28) = 6.295, p<.014 The interaction between the verb type and the group is also
significant, F (3, 28) = 8.151, p<.000. To find where the significant difference exists
post hoc Scheffé (p<.05) was conducted.

Differences in performance on the unaccusative verb type, change of state exist
versus change of location arrive, were not significant 1.76, p < .257. The differences in
performance on unergative type of verbs were not significant as well. Scheffé has not

confirmed the results of ANOVA.

49

ANOVA has shown significant differences between groups, F (3, 28) = 6.295,
p<.014. Scheffé has borne out this result. There is significant difference in judgment
between the groups. 1.62, p < .047

ANOVA has identified the significant difference in interaction between the verb
type and the group. Scheffé indicated the significant difference between NNS’s
judgments versus NS’s judgments regarding change of location verb: 5.23, p < .029

Thus, ANOVA and Scheffé confirm that two groups are different in their
judgments and the difference is located in the judgments of core unaccusative verbs. This
result validates the hierarchy proposed by Sorace (1995) for intransitive verbs. The
judgments regarding unergative verbs were not significantly different. The finding can be
explained by the fact, that Gen is ungrammatical with this type of verb. L2 learners

distinguish the two types of verbs in Russian.

4.2 Research questions and answers

I. Do L2 learners of Russian differentiate the unaccusative and unergative verbs by
allowing GN on unaccusatives and disallowing it on unergatives?
The original prediction was that Ieamers would differentiate the unaccusative and
unergative verbs by allowing GN on unaccusative and disallowing it on unergatives.
The experiment has shown that the Ieamers do allow Gen case on the
arguments of unaccusative verbs but at the same time they allow it on the arguments of
unergative verbs. Several reasons may account for this result: the optionality of case

assignment and, as a result, the lack of unambiguous evidence for the distinction of

50

unaccusatives and unergatives in L2; the subtle evidence for unaccusativity in English.
The prediction is not confirmed.

The data suggest that the learners are in the process of forming a TL grammar;
otherwise they would not allow Gen case on the arguments of unergative verbs. Gen is
ungrammatical with unergatives. If the participants were not in the process of forming a

TL grammar the data collected would be comparable to that of NSs.

2. Do L2 learners of Russian differentiate the unaccusatives and unergatives depending
on the position of monadic verbs on Sorace ’s SIH?

The results of a repeated measures ANOVA shows that the difference in the
verb type is significant F (3, 28) = 6.157, p<.001. However, Scheffé test did not reveal
significant difference between the grammaticality judgments of unaccusative and
unergative verbs depending on their position on Sorace’s SIH.

We can not make a generalization about the general behavior of L2 Russian in
regards to SIH due to the small sample size of the participants and the impossibility to

include all the verb types in the study.

3. Do L2 learners distinguish better the unergative syntactic behavior of the verbs
denoting motional processes and worse the behavior of the verbs denoting non-
motional processes?

The original prediction was that learners will distinguish better the unergative syntactic

behavior of the verbs denoting motional processes and worse the behavior of the verbs

denoting non-motional processes.

51

t-test did not show any significant difference in distinguishing the unergative
verbs denoting motional processes vs. non-motional processes. r=1.76, df= 36, ns.

This result can be explained by the fact that the verbs are very close on SIH.
However, the results of descriptive statistics have revealed the general trend. Both groups
tend to be surer in their judgments about verbs denoting nonmotional process. The fact that
the verbs are very close on SHI could influence the results. The prediction that Ieamers have

followed the SIH is considered to be partially confirmed.

4. Do L2 learners distinguish better the unaccusative syntactic behavior of the verbs
denoting change of location and worse the behavior of the stative verbs?

The original prediction was that learners will distinguish better the unaccusative syntactic
behavior of the verbs denoting change of location and worse the behavior of the stative
verbs.

t-test did not show significant difference in distinguishing the unaccusative
verbs denoting change of location vs. the state verbs. t=0.083, df= 36, ns.

Scheffé analysis did not reveal the significant differences as well. The
hypothesis that the Ieamers follow the SIH is not confirmed.

The results of descriptive statistics have shown the general trend in the
judgments though other statistical tests did not confirm the results. The hypothesis that

the learners follow the SIH is considered to be partially confirmed.

5. Do L2 learners show a stronger preference for grammatical sentences over
ungrammatical sentences with core unergative and unaccusative verbs, and a weaker

preference with peripheral verbs if they recognize the difference at all?

52

The original prediction was that learners will show a stronger preference for grammatical
sentences over ungrammatical sentences with core unergative and unaccusative verbs,
and a weaker preference with peripheral verbs if they recognize the difference at all.

The results of the Descriptive statistics confirm the prediction. NNS show a
preference for grammatical sentences over ungrammatical with core unergative verbs and

weaker preference with peripheral. Consider Table 4:

Table 4. Mean acceptability judgment on core and peripheral unergative verbs

 

 

 

 

 

 

NNS
Mean SD
Nom 9.73 1.03
Nonmotional process work
Gen 4.42 4.1
Nom 9.56 1.3
Motional process swim
Gen 4.82 4.2

 

 

 

 

 

 

The GN constructions are ungrammatical with unergative verbs. The result of descriptive
statistics show that L2 learners are surer about the grammatical Nom case with core
unergative verb work then with more peripheral verb swim. The descriptive statistics
showed that there is a tendency to allow more ungrammatical sentences with peripheral
unergative verbs. The means of grammaticality judgments were 4.82 for motional process
vs 4.42 for nonmotional. However, Scheffé analysis did not reveal the significant
differences. This result can be explained by the fact that the verbs are very close on SIH.
The results did not reveal the preference for unaccusative verbs as both

constructions are grammatical.

53

6. Do grammaticality judgments of L2 learners difler from those of L1 speakers?
The original prediction was that the grammaticality judgments of L2 Ieamers will differ
from those of L1 speakers.

This prediction is confirmed by the results of the Descriptive statistics analysis
(see Figures 4 and 5) and ANOVA (F (3, 28) = 6.295, p<.014). There is a significant
difference between the grammaticality judgments of L2 learners and NS in relation to

unaccusative verbs.

7. Do L2 learners analyze unaccusative verbs as unergatives, following Single
Argument Linking Rule, as predicted by Oshita ’s UT H ?
The original prediction was that L2 learners will analyze all unaccusative verbs as
unergatives, following Single Argument Linking Rule, as predicted by Oshita’s UTH.
The results of the experiment should be treated with caution as the study is not
longitudinal and it shows only the present state of the learners. This is one of the

limitations of the study as UTH is a developmental hypothesis.

8. Have L2 learners achieved near-native proficiency?

The original prediction was that L2 learners have not achieved near-native proficiency.
The results of the experiment have shown that L2 Ieamers have not achieved

near-native proficiency, as they consider ungrammatical forms to be grammatical and

their judgments differ from those of NS.

54

9. Will L2 learners allow the Genitive Case in all the structures since it is optional?
The original prediction was that learners will not allow the Genitive Case in all the
structures since it is Optional

The results of the Descriptive statistics did not confirm this prediction. NNS do
allow Gen case though it is optional.

In section 4.3 I will talk about the acquisition of the phenomenon of split
intransitivity by heritage learners of Russian. The grammaticality judgments presented by
them differ from both native speakers and second language learners. To account for this
fact I want to hypothesize that there is a phenomenon of incomplete acquisition of the
native language. The fact is understandable as they did not get any formal instruction in

Russian before coming to the US.

55

4.3 Heritage Learners

In this section I will touch upon an observation that was made while conducting
the research. Among L2 learners there were several heritage Ieamers (HL), who have
immigrated from the former Soviet Union to the US. when they were from 4 to 6 years
Old. This factor is relevant to the data collected in the study. Below, I will refer briefly to
several definitions of the term heritage learner. Then, I will talk about the distinction in
language acquisition between NS and HL, NNS and HL, and talk about the recent
research on this question. Finally, I will present evidence from the acquisition of the
unaccusative - unergative distinction in Russian by heritage learners.

Valdés (2000) refers to heritage speakers as those who are raised in a home where
a non-English language is spoken, who speak or merely understand the heritage language,
and who are to some degree bilingual in English and the heritage language.

According to Kagan (2001), heritage learners are neither typical students of
foreign language nor are they native language speakers. Native speakers learn a language
in childhood and continue using it throughout their life in multiple interactions in a full
language community. Students acquire a foreign language when their own native
language is fully developed. Their full language community will continue to be in the
native language. Heritage Ieamers had age appropriate native language skills when they
began learning the new language that was to become their primary language because of

immigration.

56

There is a defining distinction between heritage language and foreign language
acquisition. Heritage language acquisition begins at home, as opposed to foreign
language acquisition which is usually begun in a classroom.

All the participants of this study left Russia or the former Soviet Union. They did
not get any formal instruction in Russian language as they did not go to a Russian school.
All of them were lsolely or primarily educated in English. In the United States they
continue to speak Russian only with their farnily members but English has been the
dominant language in their university and professional life.

Six participants of the experiment got the same task as native speakers of Russian
and foreign language learners. First, they were asked to translate forty words from
Russian into English. The data showed that none of them has problems with the
vocabulary used in the grammaticality judgment test. However, half of the participants
were not certain about the meaning of the verb cyutecmeoeamr, “exist”, as they marked
the column “I have seen the word but not sure of its meaning”. But as they gave a correct
translation of the word it is considered that they do not experience any lexical difficulties
regarding this semantic item. One of the participants marked the word dparcon “dragon”
as unknown at the same time he marked “exist” as unknown, and that may have
influenced the grammaticality judgments of the subject. Consider table 5 (on the next

page) where the target lexical items and the participants’ responses are shown.

57

Table 5. Translate from Russian into English (heritage learners)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Translate from . gglziion. I know hibidebietegtthe I don't
Russ1an into Translatron the . know
English word 21:23:? the word
1. manbtmrc (boy) 100% 100%

2. HHCBMO (letter) 100% 100%

3. orae'r (answer) 100% 100%

4.z1pa1<orr (dragon) 100% 100%

5. OKHO (window) 100% 100%

6. cymecrsoaarb (to exist) 100 % 50 % 50%*

7. moxonan (chocolate) 100% 1 00%

8. cryrrerrr (student) 100% l 00%

9. pa60rarb (to work) 100% 100%

10.13paq (doctor) 100% 100%

11. npni’rrn “03:33:; ‘0 100% 80% 20%"
12.1enerpa11ma (telegram) 100% 100%

13. 11.11be (to swim) 100% 100%

 

 

*One participant did not indicate the translation of the word.
** One participant did not indicate the translation of the word.

Results of Grammaticality judgment task:
Descriptive Statistics

The analysis of the data revealed that grammaticality judgments of heritage
learners differ from both native speakers and foreign language learners. There is

significant difference in judgments between two categories of unaccusative verbs: verbs

of change of location and change of state for NS and HL groups.

58

However, no significant differences were Obtained among two categories of

unergatives verbs: verbs of nonmotional process and verbs of motional process. Table 6

contains the results of Descriptive statistics.

Table 6. Mean acceptability judgment on different verb types

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NS NNS HL
Mean SD Mean SD Mean SD
Nonmotional process Nam 10 0 9.73 1.03 10 0
work Gen 3.28 3 4.42 4.1 2.7 2.6
Motional process Nom 10 0 9.56 l .3 9.17 1 .6
swim Gen 3,28 3 4.82 4.2 1.8 1.9
Change of Location Nom 9.9 1.5 5.74 4.2 8.3 4.0
arrive Gen 10 O 6.79 4.21 5 3.69
State Nom 9.6 1.16 7.52 3.67 9.6 0.52
exist Gen 9.18 1.01 7.9 3.71 4.83 3.4

 

The graphical representation of the data is given in figures 9 and 10

59

 

Figure 9. Mean acceptability judgments on unergatives.

 

Unergative verbs

 

 

 

 

 

The GN construction is ungrammatical with unergatives. Unexpectedly, HLs
allow Gen with unergative verbs, and what is more surprising, they allow it with a core
unergative verb: the verb of nonmotional process work. Another surprising result was that
they less frequently allow Gen with the verb of motional process swim, though this is a
more peripheral unergative verb. Their judgment is even better than that of NS. They are
surer with a peripheral verb than with a core verb. We can stipulate that HL do not
acquire these types of verbs as a pattern. They learn each verb separately.

It is difficult to say whether this is a general trend among all HLs of Russian due
to the modest sample size.

Figure 10 (next page) represents the grammaticality judgments of the participants
with respect to unaccusative verbs.

It was not expected that HLs will prefer Nom case to Gen in the constructions
with change of location verbs. They tend to consider Gen ungrammatical with change of

location verbs and change of state verbs, and with both types of verbs they prefer Nom

60

Figure 10. Mean acceptability judgments on unaccusatives.

 

Unaccusative verbs

s. x, I
'. h ‘
r..‘:-'Jp‘.'
%
%
"L

10

8 $122: '-

7 fil; ‘1 . '

Mean 6 gI

acceptability 5 i u
judgments 4%I IQ?

3

1 I

O

 

new

N arrive G arrive N exist G exist

 

 

 

We got two statistically significant results: on the one hand, NS-HL with the
unaccusatives verb arrive+ GEN and on the other, NS-HL with the verb exist + GEN.
Heritage learners do not allow Gen case on the argument of unaccusatives verbs. They
treat unaccusative verbs as unergatives. The reasons may be the optionality of the case
assignment in Russian.

SO we can stipulate that IL grammar of heritage learners differs from that of NS.
Montrul proposes the explanation of such phenomenon, which she called incomplete

language acquisition (2004b). HLs do not acquire the language to the same extent as NS.

t-test results
The comparison of verb types between the groups of NS and HL showed a
significant difference on one type of unaccusative verbs: Change Of location. t = 2.686,

df=15, p < .05.

61

There is also a significant difference regarding the second type of unaccusatives
verbs: verb of state as well. t = 3.843, df=15, p < .05. One of the explanations can be the

fact that HLs are subject to incomplete acquisition as it is discussed by Montrul (2004).

ANOVA and Scheffé

A repeated measures ANOVA shows that the difference in the verb type is
significant F (3, 34) = 5.611, p<.001. There are significant differences between groups,
F (3, 34) = 5.325, p<.006. The interaction between the verb type and the group is also
significant, F (3, 34) = 4.274, p<.001. To find where the significant difference exist post
hoc (Scheffé) test was conducted. p<.05

The post hoc test shows significant difference between the judgments of HL and
NS 2.69, p<.011, NNS and NS 1.62, p<.047 and there is no significant difference
between the judgments of HL and NNS.

The post hoc test does not reveal the significant difference between two verb
types; unaccusatives and unergatives.

The next Chapter presents a conclusion.

62

CHAPTER 5

CONCLUSION AND DISCUSSIONS

The phenomenon of intransitivity has attracted a lot of attention from researches
in Second Language acquisition theory. As L2 learners experience problems in acquiring
the distinction between two types of intransitive verbs several hypotheses were proposed
to account for it. The Split Intransitivity Hierarchy by Sorace was found to explain the
variations regarding the acquisition of unaccusative and unergative verbs in Western
European languages. The Unaccusative Trap Hypothesis by Oshita accounted for
different facts in Japanese and Chinese. This study aims to find evidence of SIH and
UTH in Russian, though Russian contrary to Western European languages does not have
unambiguous evidence for split intransitivity: the phenomenon under investigation, which
distinguishes unaccusative and unergative verbs, is optional Gen case marking allowed
by unaccusative verbs. The semantic difference (definite or indefinite interpretation of the
argument of the verb), is the only source of differentiating to the optionality in the input.

The results of the study show that learners successfully apply the rule of GN with
unaccusatives verbs though they do not have the same pattern as N85. N85 distinguish
core vs. peripheral unaccusatives verbs, but they treat core and peripheral unergatives
verbs alike as no significant differences were revealed in their judgments. It seems likely
that NSs consider both types of verbs as core as they are very close at the hierarchy. This
is one of the limitations of the present study. The appearance of verbs denoting various

uncontrolled processes (such as body functions, involuntary reaction and emission),

63

might attribute to different results as they are more distinct from the verbs denoting
nonmotional process according to SIH. The choice of verbs was motivated by the
frequency of the verbs in the language and NNSs’ awareness of their meaning determined
by the vocabulary test.

NS showed a definite preference of Gen vs. Nom case regarding unaccusatives
verbs. Though the statistic result did not reach the level of significant difference it can be
treated as the possible conformation of SIH.

NNSs do not display the same gradation in their judgments as native speakers,
but they seem to develop in the direction of the native pattern. The results of this study
should be treated with caution due to the small sample size of the participants and the
incomplete coverage of verb types proposed for SIH.

One of the characteristics for Unaccusatives verbs in Russian, GN, is syntactically
optional whereas unergatives verbs do not demonstrate any optionality. The fact might
lead to the potential difficulties in differentiating these two types of verbs. Learners may
prefer one of the choices and disallow another one, or they may apply the rule of
optionality to both types of verbs. Research with advanced and near-native learners
should be conducted to provide an answer to this question.

L2 learners of Russian differentiate the unaccusative and unergative verbs by
allowing GN on unaccusatives but at the same time they allow it on the arguments of
unergative verbs. Several reasons may account for this result: the optionality of case
assignment and, as the result, lack of unambiguous evidence for the distinction of

unaccusatives and unergatives in L2; the subtle evidence for unaccusativity in English.

64

L2 learners of Russian differentiate the unaccusatives and unergatives
depending on the position of monadic verbs on Sorace’s SIH. The results of a repeated
measures ANOVA shows that the difference in the verb type is significant F (3, 28) =
6.157, p<.001. However, the Scheffé test did not reveal significant differences between
the grammaticality judgments on unaccusative and unergative verbs depending on their
position on Sorace’s SIH.

The original prediction was that Ieamers will better distinguish the unergative
syntactic behavior of the verbs denoting motional processes and worse the behavior of
the verbs denoting non-motional processes. The t-test as well as Scheffé did not show any
significant difference in distinguishing the unergative verbs denoting motional processes
vs. non-motional processes. However, the results of descriptive statistics have revealed the
general trend. Both groups tend to be surer in their judgments about verbs denoting
nonmotional process. The fact that the verbs are very close on SIH could influence the
results. The prediction that learners have followed the SIH is considered to be partially
confmned.

L2 learners distinguish better the unaccusative syntactic behavior of the verbs
denoting change of location and worse the behavior of the stative verb. The t-test as well
as Scheffé did not show significant differences in distinguishing the unaccusative verbs
denoting change of location vs. the Change of state verbs. The NS’s group considers Gen
case more grammatical with verbs of change Of location. The results of descriptive
statistics have shown the general trend in the judgments, though other statistical tests did
not show the significant differences in the results of two groups. The hypothesis that the

Ieamers follow the SIH is considered to be partially confirmed.

65

 

The original prediction was that Ieamers will show a stronger preference for
grammatical sentences over ungrammatical sentences with core unergative and
unaccusative verbs, and a weaker preference with peripheral verbs if they recognize the
difference at all. The results of the Descriptive statistics confirm the prediction. NNS
show a preference for grammatical sentences over ungrammatical with core unergative
verbs and weaker preference with peripheral. The GN constructions are ungrammatical
with unergative verbs. The result of descriptive statistics show that L2 Ieamers are surer
about the grammatical Nom case with core unergative verb work then with more
peripheral verb swim. One of the reasons may be that they are less familiar with the verb
swim as it is shown by the results of the vocabulary test: only 69% of students marked the
column “I know the word”. For the verb work the result is 100%. The descriptive
statistics showed that there is a tendency to allow more ungrammatical sentences with
peripheral unergative verbs. However, Scheffé analysis did not reveal the significant
differences. The results did not reveal the preference for unaccusative verbs as both
constructions are grammatical.

The original prediction that the grammaticality judgments of L2 Ieamers will
differ from those of L1 speakers was confirmed by the results of the Descriptive
statistics analysis and ANOVA. There is a significant difference between the
grammaticality judgments of L2 Ieamers and NS.

The results of the experiment should be treated with caution as the study is not
longitudinal and it shows only the present state of the Ieamers. This is one of the

limitations of the study as UTH is a developmental hypothesis.

66

The original prediction that L2 Ieamers will analyze all unaccusative verbs as
unergatives, following Single Argument Linking Rule, as predicted by Oshita’s UTH was
not confirmed. The results of the experiment have shown that L2 learners do not analyze
all unaccusative verbs as unergatives as they do allow Gen case on the argument of the
verb.

L2 learners allow the Genitive Case in all the structures though it is optional.
Some interesting observations were made during the experiment with regard to heritage
learners of Russian. The grammaticality judgments presented by them differ from both
native speakers and second language learners. To account for this fact I have
hypothesized that there is a phenomenon of incomplete acquisition of the native
language. The participants did not get any formal instruction in Russian before coming to
the US. and the only source of language input was their family.

Though, the GN construction is ungrammatical with unergatives, HLs allow Gen
with unergative verbs, and what more is surprising, they allow it with a core unergative
verb: the verb of nonmotional process. Another surprising result was that they less
frequently allow Gen with the verb of motional process swim, though this is a more
peripheral unergative verb. It is difficult to say whether this is a general trend among the
HLs of Russian due to the modest sample size.

HLs prefer Nom case to Gen in the constructions with change of location verbs.
They tend to consider Gen ungrammatical with location verbs and change of state verbs,

and with both types of verbs they prefer Nom

67

So we can stipulate that IL grammar of heritage learners differs from that of NS.
Montrul proposes the explanation of such phenomenon, which she called incomplete
language acquisition (2004b). HLs do not acquire the language to the same extent as NS.

Their judgments are not significantly different from those of NNSs. The reason
can be the same formal instructions at the university level, comparable to that of NNSs’

There are a number of possibilities for future research on this topic. It would be
interesting to investigate the grammaticality judgments of advanced Ieamers of Russian
by including in the test other diagnostics of unaccusativity, as well as all types of verbs
according to the SIH proposed by Sorace (1995). It would definitely be a good topic for
doing further research to investigate if HL are really so different from NS and NNS. For
that purpose interesting findings can be produced when comparing them with the

advanced L2 Ieamers and NS.

68

APPENDIXES

69

APPENDIX A

QUESTIONNARI FOR ENGLISH NATIVE SPEAKERS

. L1 if other then English

 

. How long have you been studying Russian?

 

. Have you been to Russia? How long did you stay there?

 

. DO you speak Russian outside the classroom?

 

If yes, specify: where and with whom

 

 

. Did you take any test of Russian as a foreign language?

 

If yes, what was the score?

 

. Have you studied any other foreign language?

 

If yes, specify

 

7O

APPENDIX B

QUESTIONNAIRE FOR RUSSIAN NATIVE SPEAKERS

1. L1 if other then Russian

 

2. How long have you been abroad?

 

3. How long have you been in the USA?

 

4. How often do you go to Russia?

 

5. When was the last time you were there?

 

6. How often do you speak Russian?

 

7. Where and with whom do you speak Russian?

 

8. Have you studied any other foreign language (except English)?

 

If yes, specify

 

71

APPENDIX C

VOCABULARY TEST

Translate from Russian into English

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Translate from Translation I know I have seen the I don't
Russian into the word word but not sure know
English of its meaning the
word
1. MaJIb‘IHK (boy)
2. HHCBMO (letter)
3. Borra (water)
4. Hafirn (to find)
5 t1y13cr13013arracsr (to feel)
6. noaanrsca (to appear)
7. orBeT (answer)
8. Barry'rb (to wither)
9. nparcon (dragon)
10. yrposa (threat)
11. nuuo (face)
12. pasrmrbca (to overflow)
13. TOHyTB (to sink)
14. OKHO (window)
15. paOOTaTB (to work)
16. uBeTOK (flower)
17. cymechOBarL (to exist)
18. moxonarr (chocolate)
19. mars (to sneeze)
(to fear, to be
20. Goarbca afraid)
21 . JIPICTOK (leaf)
22. ynerarbca (to smile)
23. ytmrenb (teacher)
24. crrarb (to sleep)
25. Bcrpeqa (meeting)
26. crynenr (student)
27. ocraBarbcsr (to stay)
28. renerpaMMa (telegram)
29. MOJIOKO (milk)
30. MOp03 (frost)

 

 

 

 

 

72

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

31 . Berep (wind)
32MH8JI (magazine)
33. normen (must, have to)
34. HOT]: (to sing)
35. rearp (theatre)
36. 11an (to swim)
37. CMCSITBCH (to laugh)
38. 61,111; (to be)
39 Bpaq (doctor)
(to come,
40. npm‘irn to arrive)
41. t{mar}, (to read)
42. qaii (tea)
43. nonHHMarbca (to arise)
44. CHIICTI: (to fit)

 

 

 

73

 

 

 

APPENDIX D

INSTRUCTION GIVEN TO SUBJECTS

The purpose of this exercise is to get you to judge the acceptability of some
Russian sentences. These sentences are all different. Some will seem better to you, but
others will not. What we are after is not what you think of the meaning of the sentence,
but what you think of the way it is constructed. Your task is to judge how good or bad
each sentence is by assigning a number to it. First we will do a small exercise: you need

to judge the length Of the lines, which I show you and give the number for it.

1.

 

You don’t need to give the exact length. Only approximate. 8. The next line I show you:

2.

 

is twice as short as then the first one, and you will assign it 4. The next line is twice
as large as the first one:
3. you
will assign it 16.
Now we will try to do the same with Russian sentences: for each sentence, assign a
number to show how good or bad that sentence is in proportion to the first sentence in a set.
For example, the first sentence is:

(1) the cat the mouse ate.

74

If it seems bad, you would give it a small number, 10 for example and if the next
example:
(2) the cat ate the mouse
is 10 times better for you then you should give it 100, and if the next example:
(3) cat the mouse ate the.

Seems to be 2 times worse you will assign it 5.

Use any number you like for the first sentence. Judge each sentence in proportion
to the reference sentence. Use any positive numbers you think appropriate. There are no
'correct' answers, so whatever seems right to you is a valid response. Only your first

impression is important, so don't spend too much time thinking about your judgment.

75

APPENDIX E
GRAMMATICALITY JUDGMENT TEST
Below is one of the variants of the test. The second variant consisted of an equal number
of the same sentences but organized in different order. The symbol “*” indicates the
ungrammaticality of the sentence. The target sentences and the distractors are separated

here. In the original test they were presented in the mixed order.

Grammaticality judgment test (target sentences)

fl

CryzreHra He npmmro. *Student (Gen) didn 't arrive (Neut)
Cryrrenra He npnnren. *Student (Gen) didn ’t arrive (Masc)
CryrreHr Irpmrren. Student (Nom) arrived(Masc.)

CryneHra npnrrmo. *Student (Gen) arrived (Neut)

Cryzrenr He npmrrno. *Student (Nom) didn’t arrive (Neut.).)
TenerpaMMbr He npurrmo. Telegram (Gen) didn ’t arrive (Neut)
TenerpaMMa He npmmra. Telegram (Nom) didn ’t arrive (F em. )

TenerpaMMbr He npmnna. *Telegram (Gen) didn’t arrive (F em. )

.‘OP‘IHP‘S‘PPE"

TenerpaMMbr npnrnno. *Telegram (Gen) arrived (Neat)

fl
O

. TenerpaMMa npmnna. Telegram (Nom) arrived(F em. )

11. Bpaqa He pa6oraer. *Doctor (Gen) does not work
12. Bpaq He pa60raer. Doctor (Nom) does not work
13. Bpaqa pa60raer. *Doctor (Gen) works

14. Bpaq pa6oraer. Doctor (Nom) works

15. Marrquca He IIJILIBCT. *Boy (Gen) does not swim.
16. MaJIb‘IIIK He IIJIBIBCT. Boy (Nom) does not swim.

l7. Maanmca HJIBIBBT. *Boy (Gen) swims.

76

18.
19.
20.

21.
22.
23.
24.

25.
26.
27.
28.
29.

30.
31.
32.
33.
34.
35.
36.
37.
38.
39.

40.

41
42

43.
44.

MaJIb‘IHK IIJIBIBCT. Boy (Nom) swims.
He IIJIbIBCT MaJIb‘IHK. Boy (Nom) does not swim

He pa60rae'r Bpaq. Doctor (Nom) does not work

IIpaKOHLI He cymeCTBonr. Dragons (Nam) don ’t exist
leaKOHOB He cymecrsyer. Dragons (Gen) don ’t exist
IIpaKOHLr cymecrayror. Dragons (Nom) exist

,IIpaKOHOB cymechyer. *Dragons (Gen) exist

Bpaqa He pafioraer. Doctor (Gen) does not work
IIpaKOHOB He cymecrayer. Dragons (Gen) don ’t exist.
Maanmca He IIJIBIBCT. *Boy (Gen) does not swim.
OrBeTa He npuumo. Answer (Gen) didn ’t arrive (Neut).

Cryirerrra He anero. *Student (Gen) didn’t arrive (Neut).

Grammaticality judgment test (distractors)

5I He runny aneM) . I don’t write letters (Gen)

51 He mmry HHCbMa. I don’t write letters(Acc)

51 many KapaHnamoru. I write with a pencil (Instr)
51 many IIHCCM. *1 write letters (Gen

51 nnmy nucsMa. I write letters(Acc)

lIeBoqrca runner. T he girl writes

IIHceM a runny. *Letters (Gen) I write

Huesma a many. letters(Acc)] write

lIeBOqKn runner. *The girl (Gen). writes

51 rmnry KapaHrrarna. I write with a pencil(Gen)

He quraer neBorrKa. Doesn’t read the girl
IIeBOIIKa quraer. The girl reads

lIeBqurr quraror. The girls read
lIeBqua He quraer. The girl doesn ’t read

IleBOIIKH He Irm‘aror. The girls don’t read

77

45. Hnraer neBoqrca. Reads the girl

46. IIrrraror neBqun. Read the girls

47. He anraror neBquH. Don ’t read the girl/girls
48. IIrrraror neBqua. Reads the girl

49. IIrrrae'r IICBO‘IKH. Reads the girls.

50. B nome He )KHBCT neBqua. In the house doesn ’t live a girl(Nom)
51. )KnBeT 110M neBoqrca. *Lives the house a girl (Nom)

52. lIeBoqrca y noma >KHBe'r. *T he girl lives near the house

53. LieBqua He xcnser B norm. The girl (Nom) doesn ’t live in the house.

54. Knnra B some newer. The book(Nom) in the house lives

 

55. )KnBeT B ,LIOMe rcrmra. Lives in the house a book(Nom

56. KHnra He )KHBCT B 110Me. The book Worn doesn ’t live in the house
57. Karma )KHBCT 11a nome. *The book (Nom lives on the house

58. KHnr amay'r B home. *The books (Ace) live in the house

59. Knnra >KHByr B nome. *The book(Nom) live in the house

60. B noure Be 651110 01010. *In the house there was no window(Nom)
61. B nor/re Ire 651110 010121. In the house there was no window(Acc)

62. OKHO B 110Me He 651110. "' There was no window(Nom) In the house
63. 01013 B 110Me He 651110. There was no window(Gen) In the house
64. 01013 B 110Me 651110 *There was a window(Gen) in the house

65. OKHO B ,uOMe 651110. *There was a window(Nom)) in the house

66. IIeBqua 11106111 moxonarr. The girl (Nom) likes chocolate (Acc).

67. IICBO‘IKH 11106111 moxonazr. *The girl (Gen) likes chocolate (Ace).

68. lIeBqua 11106111 moxorrarra. *The girl (Nom) likes chocolate (Gen).

69. IIeBqua He 11106111 moxonazr. The girl (Nom) does not like chocolate (Acc).

70. IIeBqua He 11106111 Inorconarra. The girl (Nom) does not like chocolate (Gen).
71. Emma moxonazra He 11106111“. The girl (Nom) chocolate (Gen) does not like.
72. HeBOtrrca moxonarr He 11106111: The girl (Nom) chocolate (Acc) does not like.

73. IIononan neBOtncy He 11106111. The chocolate (Nom) the girl (Acc) does not like.

78

74.
75.

76.
77.
78.
79.

80.

81.
82.
83.
84.
85.
86.
87.
88.
89.
90.

Illoxonarra He 11106111 neBqua. The chocolate (Gen) the girl (Nom) does not like
IIIOKOIIarr He 11106111 neBquy. The chocolate (Nom) does not like the girl (Ace).

IIIOKOJIaJr He 11106111 neBoqrcy. The chocolate (Nom) does not like the girl (A cc).
KHHI‘a )KHBCT B Home. The book (Nom) lives in the house.

lIeBqua He qH'raer HHCBMO. The girl (Nom) doesn ’t read the letter (Ace)
Illoxonana Her. lleBqua ero C'5e11a. There is no ehocolate(Aee). The girl has
eaten it.

He xopomaa rues anrrrJIa MHe B r0110By. Not a good idea comes into my mind

3e11eHaa H1165! anmJIa mm B r0110By. A green idea came into my mind.

MHe rrpmrura 3e11eHaa Hrrea B rOJIOBy. Came a green idea into my mind.

B r0110By MHe npmrma 3e11eHaa M651. Into my mind came a green idea.
Xopomaa H1165! npnuma mm B rOJIOBy. A good idea came into my mind.
Xopouraa MHe B rOJIOBy npmmra men. A good came into my mind idea.

He xoporrraa P111651 npnruna MHe B ronOBy. Not a good idea came into my mind.
He xopomaa npmrma MHe B r0110By M651. Not good came into my mind an idea.
He xoponrasr B rOJIOBy wrest anrrrna me. Not good into my mind idea came.
Xoporrraa Hnea npnmna mm B r0110By. A good idea came into my mind.

Xopomaa npmmra H1165! MHe B rOJIOBy. A good come idea into my mind.

79

REFERENCES

80

REFERENCES

Babyonyshev, M., Ganger, J ., Pesetsky, D., Wexler, K., (2001). The Maturation of
Grammatical Principles: Evidence from Russian Unaccusatives. Linguistic Inquiry,
Vol. 32, Number 1: 1-44

Babyonyshev, M. (1996). Structural conventions in syntax and processing: Ph.D.
Dissertation, MIT

Baker, M. (1988). Incorporation University of Chicago Press.

Bailyn J. (1997). Genitive of Negation is Obligatory. In Annual Workshop on Formal
Approaches to Slavic Linguistics: The Cornell Meeting 1 995. [= FASL 4], eds.
Wayles Browne, Ewa Domsich Natasha Kondrashova, and Draga Zec., 84-114. Ann
Arbor, Michigan: Michigan Slavic Publications.

Bard, E.G., Robertson, D. and Sorace, A. (1996). Magnitude estimation of linguistic
acceptability. Language 72: 32-68.

Borik, O. (1995). Sintakticheskij priznak neakkuzativnosti glagola (na materiale russkogo
jazyka). MA Thesis, Dept. of Theoretical and Applied Linguistics, MGU

Borer, H., Kenneth W. (1987). The maturation of syntax. In Thomas Roeper and
Edwin Williams, eds., Parameter Setting, 23-172. D. Reidel Publishing Company.

Borer, H. and Kenneth Wexler. (1992). Bi-unique relations and the maturation of
grammatical principles. Natural Language and Linguistic Theory 10: 147-189.

Brassil, D. (2002). On the distribution of the definiteness effect in Spanish:
Relativizing the Partitive Case Hypothesis. CLS papers

Brown, S. (1999). The Syntax of negation in Russian: A Minimalist Approach.
Stanford, CA: CSLI Publications

Brown, J .D & Rogers. Th. (2002). Doing second language research. Oxford: Oxford
University Press

Burzio, L. (1986). Italian syntax. Dordrecht: Reidel. Can, A

Chomsky, N (1986). Knowledge of language: its nature, origin, and use. New York:
Praeger

Chomsky, N. (1995). The Minimalist Program. Cambridge, MA: MIT Press

81

Chomsky, N. (2002). On nature and language. edited by Adriana Belletti and Luigi
Rizzi, Cambridge ; New York : Cambridge University Press

Clahsen, H and Muysken, P. (1996). How adult second language learning differs from
child first language learning. Behavioral and Brain Sciences 19: 721-723

Culicover, P. W. and Nowak, A. (2003). Dynamical grammar: minimalism,
acquisition, and change. Oxford: Oxford University Press.

F eatherston, N. (2002). Magnitude estimation and what it can do for your syntax:
Some wh-constraints in German. Ms, Tiibingen

Gass, S. and Selinker, L. (2001). Second language acquisition: an introductory course
Mahwah, NJ. : Lawrence Erlbaum Associates

Guiraud-Weber, M. (2003). Euro pas o pyCCKOM reHeTHBe orpnuaHHa: Barium co
cropOH51. Russian Linguistics, 27: 363-384
Once again about Russian genitive of negation: the look from outside

Gregg Kevin L.(1989). Second language acquisition theory: the case for a generative
perspective. In Gass, S. M. & Schachter, J. (Eds), Linguistic perspectives on second
language acquisition (1 5—41). New York: Cambridge University Press.

Haegeman, L. (1995). The syntax of negation. Cambridge [England] ; New York,
NY, USA: Cambridge University Press.

Hatch, E., & Lazaraton, A. (1991). The research manual: Design and Statistics for
Applied Linguistics. Heinle $ Heinle Publishers, Boston, Massachusetts

Hawkins, R. (2001a). Second Language Syntax. A Generative Introduction.
Blackwell Publishers, Oxford

Hawkins, R. (2001b). The theoretical significance of Universal Grammar in second
language acquisition. Second Language Research 17, 4: 345-367

Herschensohn J. (2000). The second time around Minimalism and L2 Acquisition.
Amsterdam ; Philadelphia, PA : J. Benjamins Publishing Company

Hirakawa, M. (1995). L2 acquisition of English Unaccusative constructions. In D.
MacLaughlin &S McEwen (Ed), Preceedings of the 19111 Boston University
Conference on Language Development (291-302), Sommerville, MA: Cascadilla
Press

Hirakawa, M. (1999). L2 acquisition of Japanese unaccusatives verbs by speakers of

English and Chinese. In Kenno (ed.), The acquisition of Japanese as a second
language (89-113). Amsterdam: John Benjamins

82

Hirakawa, M. (2001). L2 acquisition of Japanese Unaccusative verbs. SSLA, 23: 221-
245.

von Hout, A. (1996). Event semantics and verb frame alternations. A case study of
Duch and its Acquisition. Tilburg Dissertation in Language Studies.

Juffs, A., 1998 The acquisition of semantics — syntax correspondences and verb
frequencies in ESL Materials. Language Teaching Research, 2: 93-123

Kagan, O. (2001). A New Perspective on Teaching Russian: Focus on the Heritage
Learner Slavic and East European Journal 45.3: 507-18.

Keller, F. (2003). A psychophysical law for linguistic judgments. Proceedings of the
25th Annual Conference of the Cognitive Science Society. Mahwah: Lawrence
Erlbaum.

Lasnik, H. (2001). Minimalism. In The MIT Encyclopedia of the Cognitive Sciences
(MIT ECS) Edited by Robert A. Wilson and Frank C. Keil MIT Press

Levin, B. and Rappaport Hovav, M. (1995). Unaccusativity at the syntax semantics
interface. Cambridge, MA: MIT Press.

Levin, B. and Rappaport Hovav, M. (1998). Building verb meaning. In M. Butt & W.
Geuder (Eds) The projection of arguments. Lexical and compositional factors (97—
134). Stanford, CA: CSLI Publications

Levin, B. and Rappaport Hovav, M. (2002). An event structure account of English
resultatives. Language, 77: 766-797.

Levin, B. and Rappaport Hovav, M. (1986). From lexical semantics to argument
realization, Cambridge: Cambridge University Press.

Macaro, E. (2003). Teaching and learning a second language: a review of recent
research London; New York: Continuum

Montrul, S. (2004). Psycholinguistic evidence for split intransitivity in Spanish
second language acquisition. Applied Psycholinguistics 25: 239-26 7.

Mitchell, R. and Myles, F. (2004). Second language learning theories. London :
Arnold ; New York : Distributed in the United States by Oxford University Press,

Oshita, H. (2000). What is happened may not be what appears to be happening: a

corpus study of “passive” unaccusatives in L2 English. Second Language Research
16: 293-324.

83

Oshita, H. (2001).The Unaccusative trap in Second Language acquisition SSLA, 23:
279-304.

Pesetsky, (1982) Paths and Categories MIT dissertation.

Perlmutter, D. (1978). Impersonal passives and the Unaccusative Hypothesis.
Proceedings of the Berkeley Linguistic Society, 4: 157—189

Pust, C P. (2000). Linguistic theory and adult second language acquisition: on the
relation between the lexicon and the syntax. Frankfurt am Main ; New York : Peter
Lang

Rizzi, L. (2000). Comparative syntax and language acquisition. London; New York :
Routledge

Schoorlemmer, M. (1995). Participial Passive and Aspect in Russian. Ph.D.
Dissertation, Utrecht University

Smith, B. (2004). Computer-mediated negotiated interaction and lexical acquisition.
SSLA, 26: 365-398.

Snyder, W., Hyams, N. and Crisma, P. (1995). Romance auxiliary selection with
reflexive clitics: Evidence for early knowledge of unaccusativity. In Clark E., (Ed),
The Proceedings of the T wenty-sixth Annual Child Language Research Forum.
Stanford, CA: CSLI.127-1 36

Sorace, A. (l993a). Incomplete and divergent representation of unaccusativity in
non-native grammar of Italian. Second Language Research, 9: 22-48

Sorace, A. (1993b). Unaccusativity and auxiliary choice in non-native grammars of
Italian and French: asymmetries and predictable indeterminacy. Journal of French
Language studies 3: 71-93.

Sorace, A. (1995). Acquiring linking rules and argument structures in a second
language: The unaccusative / unergative distinction. In L. Eubank, L. Selinker, & M.
Sharwood Smith (Eds), The current state of interlanguage (153—175). Amsterdam:
Benjamins

Sorace, A. (1996). The use of acceptability judgments in second language research. In
V. T. Bhatia and W. Ritchie (eds.) Handbook of Second Language Acquisition. New
York: Academic Press, 375-409

Sorace, A. (2000). Gradients in auxiliary selection with intransitive verbs. Language,
76: 859-890.

Sorace, A. and Shomura, Y. (2001).Lexical constraints on the acquisition of split
intransitivity Evidence fiom L2 Japanese. Cambridge University Press

84

Stockburger David, W. Introductory Statistics: Concept, Models, and Applications
http://www.psvchstat.smsu.edu/introbOOk/sbl_(00.htm

Valdés, G. (2000). Introduction. Spanish for Native Speakers, Volume 1. AATSP
Professional Development Series Handbook for Teachers K-l6. New York, NY:

White, L. (2003). Second language acquisition and universal grammar. Cambridge,
UK ; New York : Cambridge University Press

Zobl, H. (1989). Canonical typological structures and ergativity in English L2

acquisition. In S. M. Gass, & J. Schachter (Eds.), Linguistic perspectives on second
language acquisition (203—221). New York: Cambridge University Press.

85