SLABAKOVA (2000) - L1 - Transfer - Revisited - The - L2 - Acquisition

L1 transfer revisited: the L2 acquisition of
telicity marking in English by Spanish and

Bulgarian native speakers*
ROUMYANA SLABAKOVA
Abstract
This paper investigates the claim that the native grammar of the learners
is the initial state of second-language acquisition, as far as the acquisition
of universal grammar parameters is concerned. Two opposing views on L1
transfer are discussed: the first hypothesis maintains that learners start out
with the L1 parameter value (Schwartz and Sprouse’s 1994, 1996 full-
transfer/full-access hypothesis), while the second hypothesis argues that
L1 transfer plays a minimal role in the acquisition process (Epstein et al.
1996’s no-transfer/full-access hypothesis). The parameter under investiga-
tion is the aspect parameter, postulating two different ways in which lan-
guages mark telicity in the verbal phrase. In order to distinguish between
the two views of transfer with experimental means, the study examines the
competence of two groups of low-intermediate learners of English, native
speakers of Spanish, a language sharing the same parameter value with
English, and of Bulgarian, a language exhibiting the opposite parametric
value. Results indicate that the differences in the performance of learners
from the two language groups are directly traceable to their native language.
Thus the full-transfer/full-access hypothesis receives experimental support.
1. Introduction
This study addresses the nature of the initial grammar in second-language

(L2) acquisition. There is currently a lively debate in the L2 literature
within the universal grammar ( UG) framework ( Eubank 1993–1994,
1996; Schwartz and Sprouse 1994, 1996; Vainikka and Young-Scholten
1994, 1996; Epstein et al. 1996; White 1996a) as to the role of the native
language (L1). Although research on second-language acquisition has
been concerned with L1 transfer from the very beginning of its existence
(see Gass 1996 for an overview and discussion), the productive discussion
Linguistics 38–4 (2000), 739–770 0024–3949/00/0038–0739

© Walter de Gruyter
Brought to you by | University of Iowa Libraries

Authenticated | 128.255.6.125
Download Date | 5/10/13 4:24 PM
740 R. Slabakova
of this phenomenon is far from over. What is more, within the principles-
and-parameters framework the issue of language transfer has taken
another significance. Language transfer can be productively studied in
the context of UG parameters. Parameters have more than one value, or
setting, and a certain parameter’s values may or may not coincide in the
L1 and the L2. In studying the acquisition of a parameter the question
of L1 transfer can be effectively tackled. If learners are demonstrated to
initially adopt the L1 value of the parameter, then their hypothetical
search space is effectively constrained by their first grammar.
Most researchers agree with the fact that there are nontrivial differences
between L1 and L2 acquisition. L2 acquirers already have a grammar in
place and this grammar influences the L2 acquisition process in some way.
However, whether or not L1 transfer exists and what exactly constitutes
transfer is still hotly debated. The L1 grammar as the L2 initial state
position is argued for in the early work of White ( White 1985, 1986,
1989) and more recently by Schwartz and Sprouse (1994, 1996), Brown
(1998), and Montrul (1997), among others. White (1989) argues that
learners initially adopt the L1 value of a parameter. Schwartz and
Sprouse’s full-transfer/full-access hypothesis extends White’s claims to
the whole initial state of the L2 grammar. They argue that ‘‘all the
principles and parameter values as instantiated in the L1 grammar imme-
diately carry over as the initial state of a new grammatical system on
first exposure to input from the target language’’ (Schwartz and Sprouse
1996: 41). UG allows access to all the values of parameters, so that
parameter resetting can occur in most cases. At all times, the learners’
grammar is a natural-language grammar, that is, it is constrained by UG.
But not all researchers agree that L1 transfer is necessarily implicated
in adult L2 acquisition. Epstein et al. (1996, 1998) propose that UG is
directly accessed and ‘‘transfer is not part of the acquisition model itself ’’
(Gair 1998: 80). Following White (1996b) I will call the latter position
the no-transfer/full-access-to-UG hypothesis. The no-transfer/full-access
and full-transfer/full-access approaches share the assumption that UG is
actively implicated in adult SLA. They differ in their positions as to the
initial state of the L2 acquisition. According to direct-access-to-UG pro-
ponents ( Epstein et al. 1996; Flynn and Martohardjono 1994; Flynn
1987, 1996; Martohardjono 1993), all the ‘‘hypothesis space of UG’’ is
the hypothesis space of the L2. Crucially, the parameters already set to
the learners’ L1 values do not influence their initial analyses of the L2
input. Thus, these researchers do not speak of ‘‘parameter resetting’’ but
of ‘‘parameter setting.’’ If the L1 plays no role in the acquisition process,
then all possible settings of a parameter provided by UG will be open
and available, and speakers from different native languages are predicted

L1 transfer revisited 741
to demonstrate precisely the same acquisition sequences of a given L2

property. In order to support this position, then, one needs to provide
evidence that at the early stages of acquisition, learners do not entertain
the L1 values of UG parameters. Instead, they home in on the L2
values directly.
In this paper I will be looking at the role of L1 parameter settings in
an area that is relatively new in generative L2 acquisition research. While
the L2 acquisition of aspect has a long tradition of functionalist and
descriptive approaches, studying verbal phrase ( VP) aspect in terms of
a parameter within the generative approach offers a new perspective for
describing language variation and its acquisition. Recently, an important
domain of extensive study in L2 research has been the aspect hypothesis,
investigating the development of temporal-aspectual systems in interlan-
guage (see Andersen and Shirai 1996; Bardovi-Harlig 1999 for recent
surveys of the approaches and findings of this research). Andersen and
Shirai (1994: 133) state the aspect hypothesis as follows: ‘‘First and
second language learners will initially be influenced by the inherent
semantic aspect of verbs and predicates in the acquisition of tense and
aspect markers associated with/or affixed to these verbs.’’ As the citation
attests, the main goal of this program of research is to study the acquisi-
tion of tense–aspect (grammatical aspect) morphemes and its interaction
with verbal aspect (see below for a description of the principled distinction
between these two types of aspect). It is also important to note that this
line of inquiry has relied predominantly on spontaneous and elicited
production data (see Table 1 and Table 2 in Bardovi-Harlig 1999 for a
survey of designs). However, the actual interpretation that learners attri-
bute to telic and atelic VPs and their knowledge of the significance of
objects for marking telicity have not been studied directly. Furthermore,
the experimental study reported on here employs a comprehension test
of aspectual interpretation to address a competence issue at the syntax–
semantics interface. Thus, the present study represents a new piece of the
aspectual knowledge development puzzle.
2. The parametric distinction
Comparative theoretical studies of aspect within generative grammar

(Smith 1997 [1991]; Slabakova 1997a, 1997b, 1997c) propose that there
is a structural parametric distinction between English and Slavic with
respect to telicity marking. In Slabakova (1997c) I briefly discuss the
Spanish facts and propose that Spanish patterns with English as far as
VP-level aspect (or situation aspect, as defined below) is concerned.

742 R. Slabakova
Smith (1997 [1991]) makes the important distinction between view-

point aspect, reflected in grammatical aspectual tense morphemes (e.g.
English simple and progressive tenses, Bulgarian imperfect and aorist
tenses, Spanish preterite and imperfect tenses; see examples below), and
situation aspect, the inherent lexical class of verbal predicates ( Vendler’s
1967 states, activities, accomplishments, and achievements). We turn to
situation aspect first.
2.1. Situation aspect
The most widely accepted ontology-based classification of verbs into

lexical classes is that of Vendler (1967), who divided verbal phrases into
four classes as in (1):
(1) States Events
Activities Accomplishments Achievements
know run run a mile die

be sick travel travel from X to Y arrive
remain burn burn out find a wallet
be tall read read the book recognize
All the verbs in human languages can be viewed as reflecting either a
state or an event. A state is defined as a stable condition of some entity
for a period of time, where no change appears from time 1 to time 2.
Events, on the other hand, are dynamic situations where some change
or changes obtain from time 1 to time 2. I use the term ‘‘event’’ to enclose
all nonstative classes. The first type of event is an activity. To take an
example from the list in (3) above, the verb run in the sentence John is
running in the park denotes an homogeneous process going on in time
with no inherent goal. If we are observing John from an imaginary
vantage point in the air, we would see John in different locations at
time 1 and at a subsequent time 2. A second type of event is an accom-
plishment, a situation that involves a process going on in time and an
inherent culmination point, after which the event can no longer continue,
as in John ran a mile. We can imagine an accomplishment as a complex
event containing an activity and a culmination of that activity. Finally,
an achievement is similar to an accomplishment in that it also has an
inherent endpoint, but in this class the culmination, or change of state,
is instantaneous. Such a momentary event is represented by the sentence
John found a wallet.

Aspectual literature often makes use of the term ‘‘telicity.’’ A clause is

defined as telic if the situation it describes has a natural (inherent)
endpoint, which has to be reached, and after which the situation cannot
conceivably continue. A clause is defined as atelic if the situation it
describes has no such endpoint. Thus, states and activities are atelic while
accomplishments and achievements are telic. Here are some examples:
(2) Atelic
a. John ran.
b. John ran laps.
Telic
c. John ran the marathon.
Consider the event described in (2c). After all the 28 miles and 853
yards of the marathon distance have been covered by John, there is not
a yard more that he can run that will be described by the sentence in
(2c). The situation has reached its inherent endpoint, measured out by
the length of the marathon distance. In the case of the events encoded
in (2a) and (2b), however, the running events can potentially continue
indefinitely and are described as atelic.
Another useful distinction has to do with the cardinality of determiner
phrases (DP)s. A DP is of specified cardinality if its denotation can be
exhaustively counted or measured. A DP is of unspecified cardinality if
its denotation cannot be exhaustively counted or measured. Cardinality
is orthogonal to definiteness, as the following examples demonstrate.
(3) Specified cardinality

an apple, three cakes, the cake
(4) Unspecified cardinality
apples, cake
There are an overt definite and an optional indefinite article in

Bulgarian. Thus the marking of cardinality parallels that in English (see
[3]–[4] above):
(5) a. Specified cardinality

(edna) jabelk-a, tri jabelk-i, torta-ta
an apple three apple-s cake-DET
b. Unspecified cardinality
jabelk-i, torta
apple-s cake

744 R. Slabakova
The same is true of Spanish:

(6) a. Specified cardinality
una manzana, diez manzana-s, la torta
an apple ten apple-s the cake
b. Unspecified cardinality
manzana-s, torta
apple-s cake
Although all three languages under consideration mark DP cardinality
in a similar fashion, they do not use it uniformly to affect the telicity of
the verbal phrase. As we will see below (in sections 2.3, 2.4, and 2.5),
specified cardinality of the DP object signals telicity in English and
Spanish; specified cardinality of the DP object is irrelevant for telicity
calculation in Bulgarian.
2.2. Viewpoint versus situation aspect
The second aspectual distinction described by Smith (1997 [1991]) as

viewpoint aspect (also known as sentential or grammatical aspect) has
to do with the choice of the speaker with respect to how to present the
event: from the outside as a complete whole, or from the inside as it is
unfolding. As Comrie notes, aspectual tenses reflect ‘‘different ways of
viewing the internal temporal constituency of a situation’’ (Comrie 1976:
3). For example, the English sentences in (7) differ in viewpoint aspect:
they present the event as a bounded whole, (7a), or as an event whose
progress is viewed from within, with no indication of initial or final
boundary, (7b).1
(7) a. Bounded
Cordelia ate an apple.
b. Unbounded
Cordelia was eating an apple.
Every sentence in natural languages has to encode both viewpoint and
situation aspect. The two aspectual distinctions obviously interact. Both
sentences in (7) present a telic event with an inherent endpoint, although
(7a) views this telic event as a bounded whole while (7b) views it as an
unbounded process.2
As examples in (8) below show, Spanish also distinguishes viewpoint
aspect in the aspectual tenses imperfect and preterite. The Spanish
viewpoint-aspect distinction is not exactly parallel to the English one;
however, for the purposes of this study, we will leave these differences

aside.3 I will come back to this contrast in aspectual tenses in the

discussion section.
(8) a. Bounded
Julieta practic-ó tenis.
Juliet practice-3sS/PRET tennis
‘Juliet did her tennis practice’ (... this morning, and is no longer
playing).
b. Unbounded
Julieta practic-aba tenis.
Juliet practice-3sS-IMP tennis
‘Juliet practiced tennis.’ (She used to do it habitually.)
‘Juliet was practicing tennis.’ (... when I saw her, and may still
be playing).
It is well known that Spanish and English aspect are different, and
that, for example, English learners have considerable difficulties in acquir-
ing the Spanish system of aspectual tenses (see note 3). However, the
Spanish–English differences are only in the encoding of viewpoint aspect,
not of situation aspect. We can think of at least two layers of aspect
marking in a universal clause structure, situation aspect being checked
at the VP level, and viewpoint aspect at the TP level (Giorgi and Pianesi
1997). I will argue below that although Spanish and English have different
viewpoint aspect, they share the marking of situation aspect. In this
study, I am trying to tease the two aspectual distinctions apart and study
the acquisition of the way telicity is marked in English, separately from
the aspectual tenses. The test sentences employ exclusively the simple
past tense in accomplishment and activity verbal phrases, where the telic
or atelic interpretation is due to the specified or unspecified cardinality
of the object DP. Thus, telicity is a variable while viewpoint aspect is
kept constant. However, since the two aspectual distinctions work
together (see [7] above), I will consider some possible interactions in the
discussion section.
2.3. The English value of the VP-aspect parameter
In English, the verbal form itself does not indicate whether the event is
telic or atelic.4 Verkuyl (1972, 1993) has argued convincingly that it is
the cardinality of the nominal arguments that determines the interpreta-
tion. An object of specified cardinality brings forward a telic reading as
in (9a). An object of unspecified cardinality indicates an atelic reading
as in (9b).

746 R. Slabakova
(9) a. Telic
Claire ate an apple/the apple/three apples/a bag of popcorn.
b. Atelic
Claire ate apples/popcorn.
Some particles are considered to be overt markers of telicity in English
(Brinton 1988). For example, (10a) is grammatical with the particular
up while (10b) is not.
(10) a. Claire ate her apple up/Claire ate up her apple.
b. */?Claire ate up apples/Claire ate apples up.
2.4. The Spanish value
In Spanish, marking telicity also depends on the cardinality of the object

(see Nishida 1994; Bonneau et al. 1994, 1995). The following Spanish
examples demonstrate the importance of the object’s cardinality for the
aspectual interpretation.
(11) a. Telic
Juan comi-ó diez manzanas.
John eat-3sS/PRET ten apples
‘John ate ten apples.’
b. Atelic
Juan comi-ó manzanas
John eat-3sS/PRET apples
‘John ate apples.’
Like English (where some particles are overt telicity markers), Spanish
has an optional telicity marker beyond the cardinality of the object
determiner phrase (DP): the ‘‘reflexive’’ clitic se (Nishida 1994; Bonneau
et al. 1994, 1995; Sanz 1996). The effect of this telic clitic is such that its
presence is ungrammatical in a sentence with an eventive verb and an
unspecified-cardinality object, (12b), or without an object, (12c), that is,
an atelic sentence.
(12) a. Juan se comió un pollo.
John CL eat-3sS/PRET a chicken
‘John ate a chicken.’
b. *Juan se comió manzanas.
John CL eat-3sS/PREL apples

c. *Juan se comió.
John CL eat-3sS/PRET
‘John ate.’
Thus the Spanish facts closely mirror the English facts reflected in (10a)
and (10b): the two optional markers of telicity (up/se) can only cooccur
with a specified cardinality object. In other words, a specified-cardinality
object is the minimum requirement for a telic interpretation. An
additional, but optional, marker of telicity may be added to reinforce it.
2.5. The Bulgarian value
Bulgarian signals telicity through perfective prefixes, or preverbs (PV ),

which are lexically selected morphemes. One might want to equate perfec-
tive preverbs in Bulgarian with English particles signalling a telic event,
such as up in drink up, the only difference being that English particles
are optional while Bulgarian preverbs are not. Brinton (1988) argues
that such particles are overt telicity markers in English. However, particles
and preverbs have different scope effects over the object’s cardinality.
When perfective verbs (bearing a preverb like na-) combine with DPs
of specified cardinality in Bulgarian, the sentence is interpreted as telic.
English and Bulgarian are parallel in this respect, as the translations
attest.
(13) a. Telic
Toj na-pis-a tri pisma.
he PV-write-3sS/AORIST three letters
‘He wrote three letters.’
b. Telic
Xudozniket na-risuva edna kartina.
artist-DET PV-paint3sS/AOR a picture
‘The artist painted a picture.’
However, when perfective verbs combine with bare plural or mass NPs
in Bulgarian, which should be equivalent in effect to English unspecified-
cardinality NPs, the event is still interpreted as telic. English and
Bulgarian work differently here (see [15]).
(14) a. Telic
Toj na-pis-a pisma.
he PV-write-3sS/AORIST letters
‘He wrote some letters.’

748 R. Slabakova
b. Telic
Xudozniket na-risuva kartini.
artist-DET PV-paint3sS/AOR pictures
‘The artist painted some pictures.’
Note that the presence of the perfective preverb imparts the meaning of
a specified number to the bare plural nouns prisma/kartini ‘letters/
pictures’, reflected in the English translations by adding the quantifier
some (see Filip 1993, 1994 for the same effect in other Slavic languages).
Unlike the one in (14), the sentence in (15) can only be interpreted as
habitual activities.
(15) a. Atelic
He wrote letters.
b. Atelic
The artist painted pictures.
Let us apply the usual telicity test of modifying the event with temporal
adverbs. Open-interval-denoting adverbs of the type ‘‘for X time’’ are
compatible with atelic predicates; closed-interval adverbs of the type ‘‘in
X time’’ are compatible with telic predicates. Example (16) shows that
the telic–atelic interpretation depends solely on the presence or absence
of the perfective preverb on the verb, since the object is a mass noun in
both (16a) and (16b).
(16) a. Atelic
Tja gotvi jadene 3 časa/*za 3 časa
she cook-3sS/AORIST food for 3 hours/*in 3 hours
‘She cooked food for three hours.’
b. Telic
Tja z-gotvi jadene *3 časa/za 3 časa
she PV-cook-3sS/AORIST food *for 3 hours/in 3 hours
‘She cooked some food in three hours.’
In English, bounded eventive verbs combined with bare plural or mass
DPs result in an atelic interpretation, as discussed above (see [9b] and
[11b] as well as the tests below).
(17) a. Atelic
He wrote notes for three hours/*in three hours.
b. Atelic
The artist painted pictures for three hours/*in three hours.
c. Atelic
She cooked food for three hours/*in three hours.
It is important to reiterate that the objects of the sentences in (16) and

(17) are all bare plural and mass nouns in form. However, in the
Bulgarian sentence in (16b), for example, the object is interpreted as a
specified quantity (of food), while in the English (17c) it is interpreted
as an unspecified quantity of food. This is due to the scope effect of the
perfective preverb. To summarize, the telic interpretation in Bulgarian is
due to the presence of the preverb, a lexically selected morpheme on the
verb signalling telicity. Since the verb form is telic, DPs that are mass
nouns or bare plurals in form take on an interpretation of specified
quantity or number.
One question remains to be discussed. The literature on Slavic aspect
is divided on the issue of whether Slavic perfective preverbs fall in the
domain of viewpoint or situation aspect (see discussion above). Most
researchers (Comrie 1976; Dahl 1985; Kučera 1983, among others) agree
that Slavic aspectual preverbs mark specific ways of presenting the situa-
tion as a process, a telic event, or a state. But it is also true that the vast
majority of research on Slavic aspect does not necessarily refer to the
two levels of aspect marking. Thus we can only conjecture on how most
researchers would solve the viewpoint versus situation aspect issue.
Among the ones who do have a clear position, Smith (1997 [1991]; see
chapter 10, written with Gilbert Rappaport) claims that perfective pre-
verbs encode viewpoint aspect. Brecht (1984), Filip (1993, 1994), Piñon
(1993), and Verkuyl (1999), however, convincingly argue that Slavic
preverbs’ contribution to the overall aspectual makeup of the sentence is
at the VP (or situation aspect) level. Brecht (1984: 12) explicitly relates
preverbs to telicity marking. In this article, I follow Brecht, Filip, Piñon,
and Verkuyl and refer the reader to the original literature for more
arguments.
2.6. The analysis
In studying various marking of situation aspect in different languages,

the principles-and-parameters framework seeks to reduce cross-linguistic
variation to independently needed variation in how languages encode
quantificational elements and how phrase-structure properties affect case
marking. The more general goal is to explain the language-acquisition
task as acquisition of the feature strength of functional projections and
the presence or absence of aspectual features in those functional projec-
tions in the universal phrase structure provided by UG. In other words,
the child does not need to learn what particular aspectual features her
language has activated or not. The acquisition burden is reduced to

750 R. Slabakova
mechanisms independently needed for case checking and predicate

formation.
In capturing the Bulgarian versus English/Spanish distinction in phrase
structure, I adopt the syntactic decomposition of eventive verbs approach,
following Larson (1988), Pustejovsky (1991), Hale and Keyser (1993),
and Travis (1992). The trees in (18a) and (18b) illustrate the proposed
phrase markers for English and Bulgarian respectively.
(18) a. English/Spanish b. Slavic
VP VP
tsubj Vf tsubj Vf
V AspP V PerfP
CAUSE CAUSE
DPobj Aspf Perf AspP
Preverb=[+telic]
Asp VP DPobj Aspf
[+telic]
(up, se) tobj Vf Asp VP
tobj V
V
Let us concentrate on the English tree in (18a) first. The double VP

( lower VP and upper VP) structure reflects the semantic fact that events
may be viewed as having at least two subevents (Dowty 1979): a causative
subevent and a resultant state. The upper VP denotes the causative
subevent and the lower VP denotes the resultant-state subevent of the
eventive classes. This decomposition is reflected by postulating a null
CAUSE morpheme in the head of the upper VP in a Larsonian VP shell
structure (Hale and Keyser 1993; Pesetsky 1995; Chomsky 1995).5
Event participants (arguments) take part in the aspectual composition
through case checking in AspP (accusative case) and TP (nominative
case). AspP is an important functional category for aspect construal. The
object moves to the spec of AspP to check accusative case and the verb
moves to the head Asp (Borer 1994; van Hout 1998 [1996 ]; Schmitt
1996; Travis 1992). It is at this point, in a spec–head relationship with
the verb, that the verb imparts its temporal properties to the object DP.
Depending on a verbal feature (or type of predicate) and on a nominal
feature (specified or unspecified cardinality), the aspect of the whole
VP is calculated ( Verkuyl 1993). Whenever the object is of specified

cardinality the interpretation is one of a telic event. Whenever the object

is of unspecified cardinality, being a mass or bare plural noun, the
interpretation is atelic.6 Thus the independently needed mechanism of
accusative case checking is also used for aspectual feature checking at
the syntax–semantics interface. The same analysis would apply to Spanish
at the level of VP aspect.7
It has been argued (Slabakova 1997a, 1997b, 1997c) that differences
between English and Bulgarian with respect to aspect can be described
by a parameter, following Smith (1997 [1991]). In Bulgarian, the telic
morpheme is overt; it is a lexical morpheme, usually a preverb,8 on the
verb. It occupies the head of a functional-projection perfectivity phrase
(PerfP), a position higher than the one in English. If a preverb is in the
Perf°, a position from which it c-commands the object, the interpretation
is telic. If there is no preverb in the Perf°, then the interpretation is atelic.
Consequently, the cardinality of the object in Slavic does not matter for
aspectual interpretation; it is only the presence or absence of preverb
that signals aspectual class. Notice that the lack of preverb is meaningful
in Slavic, as it signals atelicity. In English, of course, speakers do not
know whether a verb is telic or atelic if they see the form in isolation
(e.g. ate) ( Verkuyl 1972).9 The VP-aspectual distinctions can be
summarized as below:
English/Spanish
– [+telic] morpheme on the verb is null
– telicity depends on cardinality of object
Bulgarian
– [+telic] morpheme on the verb is overt
– telicity does not depend on cardinality of object
3. Rationale for the present study
In previous research (Slabakova 1997c, 1999) I studied the acquisition

of the aspectual parameter by Bulgarian native speakers learning English.
I will summarize some of the findings here. The three learner groups in
that study had to demonstrate that they had acquired the way English
marks (a)telicity by recognizing the telic–atelic contrast. While the two
more advanced groups (high intermediate and advanced) had already
reset their parameter value to the L2 target, the performance of the low-
intermediate participants exhibited an interesting dissociation. These
learners were as accurate as native speakers on the atelic condition, while
their performance on the telic condition was significantly less accurate.

752 R. Slabakova
These findings suggest that low-proficiency learners considered telic

English VPs as atelic, because the telic verbal form in English does
not have an overt preverb and thus resembles the atelic form in Slavic.
I proposed that this discrepancy in accuracy bears on the learners’ initial
hypothesis of the L2 input.
In my previous research I argued that transfer can be demonstrated in
a cross-sectional study by comparing learners at different levels of profic-
iency. The present study attempts to confirm L1 transfer by comparing
two groups of participants at a similarly low level of proficiency: native
speakers of Bulgarian and native speakers of Spanish. This research design
is intended to reveal overt differences in the competence of the learners
directly traceable to their native language. Recall that Spanish purport-
edly has the same value of the aspectual parameter as English, while
Bulgarian has a different value. If the Spanish group reveals a different
pattern of recognizing telicity marking from the Bulgarian group, and
especially if the Spanish group shows a similar pattern to the English
control group, the hypothesis of L1 transfer will receive substantial
support.
4. Hypothesis for SLA
It was hypothesized that L2 learners will start out with the L1 value of
the proposed parameter (see White 1985, 1989; Schwartz and Sprouse
1994). This means that beginning and low-intermediate Bulgarian learn-
ers of English will consider the verb as crucial in determining the aspectual
interpretation of the sentence and will not be aware of the fact that in
English it is the cardinality of the object that is crucial in determining
telicity. Since in Bulgarian the presence/absence of a preverb signals
telicity, their learning task is to notice the absence of a telicity marker
on the verb and to subsequently reinterpret the significance of objects.
More specifically, they will perform as in hypothesis 1 below:
Hypothesis 1. Bulgarian learners will perform more accurately in recog-

nizing the atelicity of a dynamic verb and an unspecified-cardinality
object (e.g. make cakes) than in recognizing the telicity of a dynamic verb
with a specified-cardinality object (e.g. make a cake). On the assumption
of transfer, the make a cake–type telic sentence, lacking a preverb in
English, would straightforwardly be interpreted as atelic by Bulgarian-
speaking learners. Since in Slavic a verbal form with no preverb is atelic,
a similar form of the English verb (make in the examples above) will
indicate atelicity to the learner. In other words, where test sentences elicit

atelic responses from native speakers of English, the Bulgarian learners

will also produce atelic responses; but where test sentences elicit telic
responses from native speakers of English, the Bulgarian learners will,
again, respond with atelic. So learner responses will appear to match
those of English natives only in the case of English atelic sentences; but
the learner responses should be completely different from the native-
speaker responses in the case of English telic sentences.
On the other hand, Spanish beginning and low-intermediate learners
are predicted to perform as in hypothesis 2 below:
Hypothesis 2. Spanish native speakers will judge both the telic and atelic
sentences accurately. Where test sentences elicit atelic responses from
native speakers of English, the Spanish learners will produce atelic
responses; where test sentences elicit telic responses from native speakers
of English, the Spanish learners will, correctly again, respond with telic.
Their L1 Spanish grammar does not allow them to treat telic and atelic
sentences as if they are the same and facilitates their acquisition of English
telicity marking.
5. Method
5.1. Participants
Twenty-two native speakers (NS) of Bulgarian and 21 NSs of Spanish,

all adult learners of English, participated in the experiment. They were
tested in Bulgaria and Argentina, respectively. Most of the subjects in
both groups were high school and university students. They participated
voluntarily and were given the written tests in their classrooms. Their
mean age was 18.3 years for the Bulgarian group and 17.1 years for the
Spanish group. Thirty-two NSs of English served as controls, 16 speakers
of British English (BrE) and 16 speakers of North American English
(NAmE ). BrE and NAmE controls were necessary because at the start
it was hypothesized that there might be some dialectal differences between
BrE and NAmE in the domain of aspect. It turned out that there
were none.
5.2. Tasks and materials
5.2.1. Independent measure of proficiency. Participants in the Bulgarian

and Spanish groups were judged as low-intermediate learners of English

754 R. Slabakova
on the basis of an independent measure of proficiency: a cloze test (see

Slabakova 1997c for details). Apart from the first sentence, given whole
for establishing context, every seventh word was omitted throughout the
whole passage, giving 40 blanks altogether (see Appendix 1). Subjects
were asked to provide a word that would fit meaningfully in that space.
The exact-word method of scoring was used, that is, if a blank was filled
with the exact match of the word in the original text, one point was
given. If no word was supplied or if the word supplied was meaningful
but not the exact match of the original word, no point was given. Thus,
the maximum score was 40.
The cloze procedure has been demonstrated in the L2 literature to
have substantial concurrent validity as an integrative test of overall
proficiency in English as a second language (Hinofotis 1980; Irvine et al.
1974; Jonz 1990). The cloze requires the student to draw upon several
language skills simultaneously and involves complex language processing,
while the focus is on the content. In particular, it has been argued by
Hanania and Shikhani (1986) that cloze tests are significantly correlated
with standardized ESL tests in predicting the learner’s level of proficiency.
In this study, the cloze was chosen because the comprehensive versions
of standardized tests take a long time to fill in, and the attention span
of these low-intermediate learners would have been stretched unneces-
sarily. It was also felt that on formal tasks that focus on one item at a
time, the classroom learner can bring to bear a conscious knowledge of
rules that have been formally learned but are not yet part of his/her
linguistic competence (Dulay et al. 1982).
5.2.2. Aspectual interpretation task. Participants were asked to assess

on a scale from −3 to +3 how well two clauses in complex sentences
combine with each other. The seven-point scale was used in order to give
subjects sufficient space for encoding nuanced judgments between the
two extremes, ‘‘a perfectly natural combination’’ and ‘‘a very unnatural
combination.’’ There were 28 sentences in all, 12 test sentences in two
conditions (see Appendix 2) and 16 fillers.
(19) Characteristic and telic (C+T)
Antonia worked in a bakery and made a cake.
(20) Characteristic of atelic (C+A)
Sharon worked in a bakery and made cakes.
Since the first clause in test sentences as in (19) established a habitual
situation, it was expected that the telic clause that followed, if indeed it
was interpreted as telic by the participants, would be considered a worse
than perfect match. The assumption here is that conjunction joins like

constituents, compatible not only syntactically but semantically as well.

A habitual clause and a clause denoting a one-time complete event in
the past differ as types of events. On the other hand, the same habitual
clause (see [20]) in combination with an atelic second clause was expected
to be judged as a better match than the one in (19). The conjunction
joins two semantically equivalent clauses. Thus the prediction is that the
Spanish and English NS groups will judge sentences like those in (19) as
sufficiently different from sentences like those in (20), while the Bulgarian
NSs will fail to demonstrate this contrast.
In my previous work (see Slabakova 1997c, 1999) two more conditions
were tested, where the combination of clauses was atelic+atelic (a good
match) and telic+unfinished (an impossible match). The latter combina-
tion of clauses was easily detected as impossible, so all the Bulgarian
learners of English performed at ceiling. Thus, it was judged that the
more subtle contrast of good match–worse match, rather than the clear-
cut contrast between a good match and a bad match, would better reveal
the finer-grained knowledge of the beginning learners.
6. Results
6.1. Independent measure of proficiency
One of the conditions for establishing transfer effects is to use learner

groups at a comparable level of proficiency. If this condition is not met,
one can never infer that differences in performance are due to transfer
and not to lower linguistic competence instead.
The cloze test results were used in order to ascertain that the two
groups of participants were at the same level of proficiency (see Table 1).
The cutoff point was 16 out of 40 possible points. Cloze scores were
compared in a two-sample t-test. It was found that they did not differ
significantly (t(1, 43)=1.56, p=0.125), and consequently, the partici-
pants can be assumed to be at the same level of proficiency in English.
Table 1. Mean and standard deviation of participant groups in cloze task
Participant groups M SD Score range
NAmE controls (n=16) 25.9 2.64 21–31

BrE controls (n=16) 26.1 1.76 22–38
Bulgarian learners (n=22) 7.02 4.14 6–14
Spanish learners (n=21) 7.98 5.1 6–16

756 R. Slabakova
6.2. Aspectual interpretation task
6.2.1. Group results. Figure 1 illustrates the participants’ judgment of

how well the telic and atelic clauses combine with an atelic habitual
context. The maximum score is 3. In this task, we are looking for a
significant difference between C+T (telic) and C+A (atelic) condition
means.
Native speakers and Spanish learners recognize the distinction between
sentences like those in (19), Antonia worked in a bakery and made a cake,
versus (20), Sharon worked in a bakery and made cakes. Repeated-
measures ANOVA was performed on the actual (unconverted ) data
points for each participant group. The relevant statistics are given in
Table 2. For example, the first line of Table 2 indicates that the repeated-
measures ANOVA for the North American control group demonstrates
a significant effect of sentence type (the telic–atelic contrast) at F (1,30)=
8.08, p=0.0001. Both for American English and for British English
controls the contrast is highly significant. The same is true of Spanish
low-proficiency learners of English. But Bulgarian native speakers behave
differently. They do not show a statistically significant contrast between
telic and atelic sentences ( p=0.085).
Figure 1. Aspectual judgment means on telic and atelic sentences
Table 2. Statistic effect of telicity by participant groups (repeated-measure ANOVA)
Participant groups F df p
NAmE controls 38.08 1,30 0.0001

BrE controls 42.95 1,30 0.0001
Bulgarian learners 3.56 1,42 0.085
Spanish learners 17.69 1,40 0.0001

Figure 2. Aspectual interpretation means across condition
In order to confirm that Bulgarian learners do not recognize the

contrast, let us compare mean judgments across condition (Figure 2).
The important thing to be noticed in this figure is that in judging atelic
sentences, all control and learner groups, including the Bulgarian learners,
demonstrate more or less the same acceptance mean, around 2 out of
maximum 3. On the other hand, in judging telic sentences, the Bulgarian
learner group stands out by giving them a high mean score of 1.45, while
the rest of the participants judge them around the 0.5 mark ( Table 3).
Two one-factor repeated-measures ANOVAs were performed on telic
and atelic sentences separately, looking for an effect of group. On telic
sentences, there was a highly significant effect of group. Post-hoc Scheffé
analysis showed that this group effect was due to the performance of
Bulgarian learners. In particular, the Bulgarian group mean differed
significantly from all the other group means. Spanish learners and the
two control groups’ performance did not differ. Crucially, on atelic
sentences all learner groups and controls performed without any differ-
ence. Thus, the performance of the Bulgarian learners can be described
more adequately now. It is not only the case that they do not exhibit
any contrast between telic and atelic sentences. In addition, they perform
accurately on atelic sentences and inaccurately on telic sentences.
6.2.2. Individual results. In discussing learners’ competence, individual

results are potentially indicative of what is really going on in their
Table 3. Statistic effect of group for telic and atelic sentences in aspectual interpretation
task (repeated-measure ANOVA)
Type of sentence F df p
Telic 5.8 3,73 0.0001

Atelic 1.36 3,73 0.27

758 R. Slabakova
grammars and may uncover contrasts that the group results are obscur-
ing. In order to inspect individual learner contrasts, data points were
transformed in the following way. The mean over six tokens of each
type, telic and atelic sentences, was calculated. Next, the mean of telic
sentence judgments was subtracted from the mean of the atelic sentences
for each individual learner. The resulting difference was indicative of how
well the contrast between telic and atelic sentences is represented in the
particular learner’s grammar. Next, the relative contrast between telic
and atelic sentences was plotted in Figures 3 to 5 for each individual.
What we expect to see in a group of participants who have a well-
established contrast of telicity in their grammar is consistent high positive
scores. We will consider high positive scores to be indicative of a L2
target-like hypothesis space.
Let us turn to the contrasts illustrated in Figure 3 first. The large
majority of the controls (with one outlier) rate the conjunction of charac-
teristic and atelic clauses consistently higher than the conjunction of
characteristic and telic clauses, thereby demonstrating the validity of the
test for studying knowledge of telicity marking.
Unlike English native speaker controls, Bulgarian beginning learners
of English show variable behaviour. Their hypothesis space is largely
distributed on both sides of the zero plotted on the X axis (see Figure 4),
indicative of the fact that a number of learners do not have the telic–
atelic contrast in their grammar. The columns on the positive side indicate
that there are individual learners in the Bulgarian group who already
distinguish between telicity and atelicity. On the other hand, the columns
on the negative side indicate that there is also a number of learners who
have not established the telicity contrast in their interlanguage grammar.
Figure 3. Atelic–telic mean difference in controls

Figure 4. Atelic–telic mean difference in Bulgarian learners
A comparison of the Spanish group’s individual performance

( Figure 5) to that of the control group and the Bulgarian group reveals
that Spanish beginners’ hypothesis space resembles much more closely
that of the native speakers than that of the Bulgarian learners. In particu-
lar, there are no individuals who show a negative contrast.
Table 4 shows the number of individual participants distributed along
a continuum of relative telic–atelic mean difference. Note that 15 of the
controls and 11 of the Spanish participants demonstrate a clear-cut
contrast with more than 2 points between telic and atelic means. Only
one individual in the Bulgarian group shows a contrast in that range.
On the other hand, there are six individual Bulgarian learners who
demonstrate a negative contrast, that is, they judge combinations of
characteristic and telic clauses to be more natural than combinations of
characteristic and atelic clauses. Compare that number to only one outlier
control and no Spanish learners with negative contrast. The distribution
reflected in Table 4 confirms that the Bulgarian learner group is not
Figure 5. Atelic–telic mean differences in Spanish learners

760 R. Slabakova
Table 4. Number of participants distributed in relation to their relative atelic–telic mean

difference
Participant Number of Number of Number of Number of

group participants participants participants participants
whose contrast whose contrast whose contrast whose contrast
is higher than 2 is between 1.99 is between 0.99 is below 0
and 1 and 0
Control group 16 10 5 1
Spanish group 11 2 8 0
Bulgarian 1 8 7 6
homogeneous with respect to telicity marking. There are some individuals

who are still operating with the Bulgarian value of the parameter (that
is, they are in the full-transfer stage). Others have already reset the
parameter with native speakers and Spanish learners (that is, they are in
the full-access stage). These findings support our hypothesis of L1 transfer
and at the same time argue for successful UG access in adulthood.
7. Discussion
To summarize the results of the experimental study, native-speaker con-

trols demonstrate a significant contrast between telic and atelic sentence
means and they successfully match telic and atelic sentences with the
relevant context, thereby confirming the efficiency of the test. The Spanish
group patterns with the controls, while the Bulgarian learner group
does not.
Let us now take up the results for the Spanish group in more detail.
Spanish low-intermediate learners of English have acquired the distinction
between telic and atelic sentences in the target language in the sense that
they prefer atelic sentences in combination with a habitual context to
telic sentences in a habitual context, as native speakers do. Crucially,
none of the Spanish individual learners demonstrates a negative contrast
between telic and atelic means (see Table 4). These findings can be
explained only if we assume that they are aided in their telicity acquisition
by their native language. Crucially, note that the Spanish learners are as
accurate in judging telic as they are in judging atelic sentences.
These experimental results also constitute support for our initial claim
that English and Spanish exhibit the same value of the aspectual param-
eter (contra Sanz 1996; see note 7). If English and Spanish had conflicting

ways of marking telicity, the high accuracy of Spanish native speakers in

acquiring English telicity marking would be unaccounted for.
It is worth pointing out that Spanish low-proficiency learners are so
accurate in the marking of telicity given that Spanish marks aspect some-
what differently from English, if we take the whole clause structure into
consideration. As I pointed out earlier, Spanish has another aspectual
opposition in the system of tenses, preterite vs. imperfect, which could
have had an impact on the subjects’ judgments of telicity in English. The
most natural, meaning-based association of the L1 aspectual tenses with
the L2 aspectual tenses would be to map Spanish preterite on English
simple past tense; Spanish imperfect onto English past progressive. At
least as far as eventive verbs are concerned, this mapping would actually
facilitate viewpoint-aspect comprehension. The following sentences are
translations into Spanish of the test sentences exemplified in (19) and
(20), assuming the mapping delineated above. Subjects will have been
judging sentences like those in (21) and (22).
(21) Bounded telic
Antonia trabajó en una pastelerı́a e
Antonia work-3sS/PRET in a bakery and
hizo una torta.
make-3sS/PRET a cake
‘Antonia worked in a bakery and made a cake.’
(22) Bounded telic
*/?Antonia trabajó en una pastelerı́a e
Antonia work-3sS/PRET in a bakery and
hizo tortas.
make-3sS/PRET cakes
‘Antonia worked in a bakery and made cakes.’
The two sentences differ in grammaticality, with the sentence combining
preterite tense with an unspecified-cardinality object as in (22) being
strongly dispreferred to ungrammatical. According to native speakers’
judgments, there is a clash between the atelicity and the boundedness of
the test clause, especially in the context of another bounded clause. If
the learners were mapping English simple past and Spanish preterite
uniformly, then one would expect them to be much less accurate on atelic
than on telic sentences, which is not supported by the results of the
experiment. These possibilities leave us with the question of whether the
Spanish low-proficiency learners are actually mapping their native preter-
ite onto the English simple past tense, or whether they are interpreting
utterances directly with their interlanguage grammars. The results seem
to indicate that no mapping of Spanish aspectual tenses onto English

762 R. Slabakova
aspectual tenses is going on, although the matter certainly deserves further
research.
On the other hand, Bulgarian low-proficiency learners did not
demonstrate that they have acquired the contrast between telic and atelic
sentences in English. They patterned with native speakers on judging
atelic sentences but were significantly less accurate in judging telic senten-
ces. This dissociation between telic and atelic sentences in the interlan-
guage of Bulgarian learners is hard to explain without resorting to transfer
of their L1 value of the aspectual parameter. If they had homed in on
the target value as the direct-access hypothesis suggests, they would have
exhibited the pattern of responses of the Spanish low-proficiency learners,
which is not the case. On a transfer view, however, without a perfective
preverb on the verb, the test sentences in both the C+T (telic) and the
C+A (atelic) conditions look acceptable to the learners. They all seem
‘‘atelic’’ to them, which, in turn, they match up to the characteristic
nature of the first clause in the test items.
If what we observe in the Bulgarian learners’ group is some kind of
L1 interference or delay in acquisition, the differential treatment of telic
and atelic marking in English still remains incomprehensible. Proponents
of the delay-of-acquisition explanation will have to account for why this
delay affects one type of simple transitive sentence and not the other.
They will also have to explain why the performance of the Bulgarian
participants differs markedly from the performance of the Spanish partici-
pants, although learners in both groups are at a comparable level of
proficiency.
Two other principled proposals about the initial state of L2 acquisition
have been made in the literature recently: Vainikka and Young-Scholten’s
(1994, 1996) minimal trees hypothesis and Eubank’s (1993–1994, 1996)
valueless features hypothesis. The former theory suggests that only lexical
categories but not functional categories transfer from the L1 to the L2
grammar, while the latter argues that functional categories also transfer
but their features are underspecified for the ‘‘strong/weak’’ values. It has
been suggested that telicity may cross-linguistically be the ‘‘marked’’
option in the sense that it is associated with some sort of overt marking:
some morpheme-like articles ( English the, a), number ( English three),
particles ( English up), clitics (Spanish se), perfective perverbs (Bulgarian
pro-), and so forth. Most if not all of these morphemes may be associated
with functional projections: DP, NumP, AspP, etc. It can reasonably be
supposed that the learners in the early stages either don’t have those
functional projections (along the lines of Vainikka and Young-Scholten’s
[1994, 1996 ] minimal trees hypothesis view) or they have them, but the
features associated with these projections are underspecified in some

relevant respect (along the lines of the Eubank’s 1993–1994 valueless

features hypothesis view). Note that this notion of ‘‘markedness’’ would
essentially capture the distinction between functional and lexical
categories. Here I briefly offer two arguments against such a position.
One argument is conceptual. English and Spanish atelic sentences such
as those in (9b) and (11b), repeated here for ease of reference, involve
bare plurals.
(9) Atelic
b. Claire ate apples
(11) Atelic
b. Juan comi-ó manzanas
John eat-3sS/PRET apples
Plural morphemes have to be treated on a par with the other ‘‘functional’’
morphemes: particles, determiners, numerical quantifiers, preverbs. Thus
it is not the case that atelicity marking enjoys some privileged unmarked
status and that it does not need to be checked in a functional category.
On the contrary, as I have argued following Verkuyl (1993), the aspectual
construal is calculated in the functional category AspP based on a
property of the verb and a property of the DP object.
The other argument has to do with the results of the present study.
The underspecification theory will predict no difference between Spanish
and Bulgarian learners, since all learners of English are supposed to
transfer functional categories with the strength of features underspecified.
Spanish and Bulgarian beginning learners of English would be making
use of exactly the same underspecified functional categories when faced
with the English L2 input. Therefore, it is logical to expect them to
perform similarly. But this prediction is not supported by the results of
the present study: we found that Spanish learners perform as well as
native speakers on both telic and atelic sentences, while Bulgarian learners
are as accurate as the controls on atelic but crucially not on telic sentences.
Spanish and Bulgarian learners’ performance suggests different under-
lying competence, which can only be attributed to their first language.
Thus the hypothesis of L1 transfer is the only principled way of explaining
the experimental results.
8. Conclusion
In the spirit of recent proposals that the exact nature of interlanguage

linguistic representations should be further investigated (Gair 1998;

764 R. Slabakova
Schwartz 1998; White 1996b), I have explored the initial values of the
aspectual parameter in the interlanguage of Spanish- and Bulgarian-
speaking learners of English. It was shown that Spanish low-proficiency
learners are very accurate in interpreting telicity and atelicity in English
while Bulgarian learners are accurate only on atelic sentences. It was
argued that this differential accuracy is directly related to the L1 value
of the parameter instantiated in the learners’ grammar at this point. Thus
the results support L1 transfer and argue against initial direct access to
UG. The findings of the reported experiment suggest that in the area of
aspect we find the same L1 transfer effects as in the previously studied
areas of second-language acquisition, like null subject, verb raising, and
others (see Gass 1996 for a comprehensive discussion). The first half of
the full-transfer/full-access hypothesis (Schwartz and Sprouse 1994, 1996)
has received experimental support from a cross-linguistic perspective.
Received 6 December 1999 University of Iowa

Revised version received
19 April 2000
Appendix 1. Cloze test
Please fill in the blanks in the following passage. Each blank must have one and
only one word.
Joe came home from work on Friday. It was payday, but he wasn’t too
excited about it. He knew that when he sat down and paid his
bills and set aside money for groceries, some for the car and a
small amount in his savings account, there wasn’t too much
left over for a good time .
He thought about going out for dinner at his favourite restaurant,
but he just wasn’t in the mood. He wandered about his
apartment and ate a sandwich. For a while, he couldn’t stop himself
from worrying about the money situation. Finally, he got into
his car and started driving . He didn’t have a destination in
mind , but he knew that he wanted to be far away from the
city where he lived.
He drove onto a quiet country road . The country sights made him
feel good . His mind wandered as he drove along small farms
and he began to imagine living on his own piece of land and
becoming self-sufficient. It had always been a dream of his, but he
had never done anything to make it a reality. Even as he was
thinking, his logical side was scoffing at his wild imaginings.
He debated the advantages and disadvantages of living in the country

and growing his own food. He imagined his farmhouse equip-

ped with a solar energy panel on the roof to heat the house in
winter and power a water heater. He envisioned fields of vegetables for
canning and preserving to last through the winter. If the crops
had a good yield, then he could sell the surplus and buy some
farming equipment with the extra cash .
Suddenly, Joe stopped thinking and laughed out loud, ‘‘I’m really
going to go ahead with all this?’’
Appendix 2
Characteristic+telic condition (C+T)
Umberto was a good mechanic and he fixed a car.

Antonio worked in a bakery and made a cake.
Emily took very good care of her children and she packed them an elaborate
lunch for school.
Mr. Jones was a salesman and he sold a used car.
Julie was a dressmaker and she sewed a suit.
Amanda was a shop-assistant and she wrapped a purchase carefully.
Characteristic+atelic condition (C+A)
Jim was a good mechanic and he fixed cars.

Sharon worked in a bakery and made cakes.
Penny took very good care of her children and she packed them elaborate lunches
for school.
Mr. Brown was a salesman and he sold used cars.
Anne-Marie was a dressmaker and she sewed clothes.
Patricia was a shop-assistant and she wrapped purchases carefully.
Notes
* Bonnie Schwartz planted the idea for this study when I presented portions of my
dissertation at the GASLA III conference in Montreal — I am grateful to her. I also
wish to thank Lydia White, Silvina Montrul, and Lena Gavruseva for reading and
commenting on drafts of the article. Parts of the article were presented at GASLA IV in
Pittsburgh, and I thank the audience, especially Joyce Bruhn de Garavito, Silvina
Montrul, and Liliana Sanchez, for useful discussion. I am especially indebted to Silvina
Montrul for arranging the testing in Argentina. Thanks also go to the teachers and
participants in the experiment in Bulgaria and Montreal, Canada. I was partly supported
by a FCAR doctoral fellowship and a SSHRCC grant number 410-95-0720 to Lydia

766 R. Slabakova
White and Nigel Duffield. The field trip to Bulgaria was made possible by a McGill
Faculty of Graduate Studies and Research Thesis Research Grant and the Professional
Partnership Programme Travel Award by the Association of Universities and Colleges
of Canada, for which I am grateful. All errors are my own. Correspondence address:
Department of Linguistics, University of Iowa, 557 English Philosophy Building, Iowa
City, IA 52242, USA. E-mail: roumyana-slabakova@uiowa.edu.
1. The terminological problem arises of what kind of aspect is encoded at each of these
levels. The term ‘‘perfective-imperfective’’ is used to describe viewpoint aspect in the
literature on Romance and Germanic aspect, but the same term is used for situation
aspect in the Slavic aspectual literature. To avoid any confusion, I will not use the term
‘‘(im)perfective’’ at all. Following Depraetere (1995) I will reserve ‘‘boundedness’’ for
the aspectual property encoded in viewpoint (grammatical, TP-level ) aspect and
‘‘telicity’’ for the aspectual property encoded in situation ( lexical, VP-level ) aspect.
2. An example of atelic bounded and unbounded sentences in English would be (i) and
(ii), respectively:
(i) Cosmo ate Granny Smith apples but now he can’t stand them.
(ii) Cosmo was eating apples all afternoon yesterday; he did no reading whatsoever.
3. In fact, Spanish has a contrast between simple and progressive tenses as well as between
preterite and imperfect tenses, which makes the viewpoint aspect in Spanish a four-way
distinction. Apart from (8a) and (8b), the following sentences are also part of the
viewpoint aspect in Spanish.
(i) Julieta est-aba practicando tenis.

Julieta was-3sS/IMPERFECT practicing tennis
‘Juliette was practicing tennis’ (when I saw her, and she may still be playing).
(ii) Julieta est-uvo practicando tenis.
Julieta was-3sS/PRETERITE practicing tennis
‘Juliette was practicing tennis’ (when I saw her, but she is no longer playing).
4. This is true with the exception of some achievement verbs like die that are inherently
telic.
5. The upper VP in this tree is more or less similar to some recent proposals like Kratzer’s
(1996) VoiceP, Bowers’s (1993) PredP, and Chomsky’s (1995) vP.
6. As it stands, this formulation is too strong, of course. Stative verbs are atelic regardless
of their objects (e.g. John likes the girl next door, Mary hates this house). Push-type verbs
are also not affected by the cardinality of their objects (e.g. John pushed the cart, Mary
drove the red car are atelic). My analysis captures these facts by postulating that both
stative and push-type verbs are marked with an atelic feature in the lexicon. Hence, their
objects’ cardinality is irrelevant for the aspectual composition. The experiment described
in this article does not include such verbs.
7. In fact, another analysis has been proposed for the same data. Sanz (1996) argues that
Spanish and English are parametrically opposed as to the marking of telicity. Sanz
proposes that two functional categories are needed in order to capture telicity marking:
one is TrP (transitivity phrase), where the interpretable feature [measure] and the unin-
terpretable feature accusative case are checked. The second is AktP (Aktionsart phrase)
where the interpretable feature [telicity] is checked. Sanz claims that her analysis is
supported by the fact that the telic clitic se in Spanish overtly checks the strong telic
feature, while without this clitic transitive verbs and specified cardinality objects are
ambiguous between a telic and an atelic interpretation. This is, however, not the case,
and in fact Sanz later contradicts herself as to this point (‘‘Accomplishments are possible

without the clitic,’’ 1996: 32; ‘‘The verb can check telicity when it has been delimited by
an object, but it cannot otherwise,’’ 1996: 33). This account loses the central observation
that the properties of the object are crucial for licensing the telic clitic. If English and
Spanish differed parametrically, one would expect to see some syntactic consequences in
this respect. But in fact, the syntactic behavior of the telic clitic se and the English telic
particle (e.g. up) is perfectly parallel. Essentially, there is no need to posit two functional
categories when one functional category checking accusative case and telicity is the
elegant solution.
8. There is only one perfective morpheme that is a suffix: the semelfactive -n, as in čukam
‘knock repeatedly’, čuk-n-a ‘knock once’.
9. The analysis I present here has been slightly modified from an earlier version (see
Slabakova 1997a, 1997b, 1997c). The most important claim has not changed: Bulgarian
perfective morphemes have a higher, c-commanding position as compared to the
English/Spanish (optional ) telicity morphemes up/se. Strictly speaking, the experiment
does not test the scope effects that obtain from the difference in positions, but the overt
versus covert quality of the telicity morpheme. Therefore, I will not offer extensive
justification for the higher position of the Bulgarian morpheme and I refer the reader to
the original publications.
References
Andersen, R.; and Shirai, Y. (1994). Discourse motivations for some cognitive acquisition
principles. Studies in Second Language Acquisition 16, 133–156.
—; and Shirai, Y. (1996). The primacy of aspect in first and second language acquisition:
the pidgin-creole connection. In Handbook of Second Language Acquisition. W. C. Ritchie
and T. K. Bhatia (eds.), 527–570. San Diego, CA: Academic Press.
Bardovi-Harlig, K. (1999). From morpheme studies to temporal semantics. Studies in
Second Language Acquisition 21, 341–382.
Bonneau, J.; Bruhn-Garavito, J.; and Libert, A. (1994). Strengthening AgrO. Talk given at
the Annual Meeting of the LSA, Boston.
—; Bruhn-Garavito, J.; and Libert, A. (1995). Feature checking and English particles. In
Proceedings of the 1995 Annual Meeting of the Canadian Linguistic Association. Toronto
Working Papers in Linguistics. Toronto: University of Toronto.
Borer, H. (1994). The projection of arguments. In Functional Projections, E. Benedicto and
J. Runner (eds.), 19–47. UMass Occasional Papers in Linguistics 17. Amherst, MA:
GLSA.
Bowers, J. (1993). The syntax of predication. Linguistic Inquiry 24, 591–656.
Brecht, R. (1984). The form and function of aspect in Russian. In Issues in Slavic
Morphosyntax, Michael Flier and R. Brecht (eds.), 9–33. UCLA Slavic Studies 10.
Columbus, OH: Slavica.
Brinton, L. (1988). The Development of English Aspectual Systems. Cambridge: Cambridge
University Press.
Brown, C. (1998). The role of the L1 grammar in the L2 acquisition of segmental structure.
Second Language Research 14(2), 136–193.
Chomsky, N. (1995). The Minimalist Program. Cambridge, MA: MIT Press.
—(1998). Minimalist inquiries: the framework. Unpublished manuscript, MIT.
Comrie, B. (1976). Aspect. Cambridge: Cambridge University Press.
Dahl, O. (1985). Tense and Aspect Systems. Oxford: Blackwell.

768 R. Slabakova
Depraetere, I. (1995). On the necessity of distinguishing between (un)boundedness and

(a)telicity. Linguistics and Philosophy 18, 1–19.
Dowty, David (1979). Word Meaning and Montague Grammar. Dordrecht: Reidel.
Dulay, H.; Burt, M.; and Krashen, S. (1982). Language Two. New York: Oxford University
Press.
Epstein, S.; Flynn, S.; and Martohardjono, G. (1996). Second language acquisition: theo-
retical and experimental issues in contemporary research. Behavioral and Brain Sciences
19(4), 677–714.
—; Flynn, S.; and Martohardjono, G. (1998). The strong continuity hypothesis: some
evidence concerning functional categories in adult L2 acquisition. In The Generative Study
of Second Language Acquisition, S. Flynn, G. Martohardjono, and W. O’Neil (eds.),
61–77. Mahwah, NJ: Erlbaum.
Eubank, L. (1993–1994). On the transfer of parametric values in L2 development. Language
Acquisition 2, 183–208.
—(1996). Negation in early German-English interlanguage: more valueless features in the
L2 initial state. Second Language Research 12(1), 73–106.
Filip, H. (1993). Aspect and the interpretation of nominal arguments. Chicago Linguistic
Society 28.
—(1994). Aspect and the semantics of noun phrases. In Tense and Aspect in Discourse,
Co Vet and Carl Vetters (eds.), 227–255. Berlin: Mouton de Gruyter.
Flynn, S. (1987). A Parameter-Setting Model of L2 Acquisition. Dordrecht: Reidel.
—(1996). A parameter-setting approach to second language acquisition. In Handbook of
Second Language Acquisition, W. Ritchie and T. Bhatia (eds.), 121–158. San Diego:
Academic Press.
—; and Martohardjono, G. (1994). Mapping from the initial state to the final state: the
separation of universal principles and language specific properties. In Syntactic Theory
and First Language Acquisition: Crosslinguistic Perspectives, vol. 1, Heads, Projections and
Learnability, B. Lust, M. Suner, and J. Whitman (eds.), 319–336. Mahwah, NJ: Erlbaum.
—; Martohardjono, G.; and O’Neil, W. (eds.) (1998). The Generative Study of Second
Language Acquisition. Mahwah, NJ: Erlbaum.
Gair, J. W. (1998). Functional categories in L2 acquisition: homegrown or imported?
Commentary on part I. In The Generative Study of Second Language Acquisition, S. Flynn,
G. Martohardjono, and W. O’Neil (eds.), 79–86. Mahwah, NJ: Erlbaum.
Gass, S. (1996). Second language acquisition and linguistic theory: the role of language
transfer. In Handbook of Second Language Acquisition, W. Ritchie and T. Bhatia (eds.),
317–345. San Diego: Academic Press.
Giorgi, A.; and Pianesi, F. (1997). Tense and Aspect. Oxford: Oxford University Press.
Hale, K.; and Keyser, J. (1993). On argument structure and the lexical expression of
syntactic relations. In The View from Building 20, K. Hale and J. Keyser (eds.), 53–109.
Cambridge, MA: MIT Press.
Hanania, E.; and Shikhani, M. (1986). Interrelationships among three tests of language
proficiency: standardized ESL, cloze and writing. TESOL Quarterly 20(1), 97–109.
Hinofotis, F. B. (1980). Cloze as an alternative method of ESL placement and proficiency
testing. In Research in Language Testing, J. W. Ollrr, Jr., and K. Perkins (eds.), 121–133.
Rowley, MA: Newbury House.
Irvine, P.; Atai, P.; and Oller, J. W., Jr. (1974). Cloze, dictation, and the test of English as a
foreign language. Language Learning 24, 245–252.
Jonz, J. (1990). Another turn in the conversation: what does cloze measure? TESOL
Quarterly 24(1), 61–83.

Kratzer, A. (1996). Severing the external argument from itd verb. In Phrase Structure and
the Lexicon, J. Rooryck and L. Zaring (eds.), 109–138. Dordrecht: Kluwer.
Kučera, H. (1983). A semantic model of verbal aspect. In American Contributions to the
Ninth International Congress of Slavists, vol. 1: Linguistics, M. Flier (ed.), 171–184.
Columbus, OH: Slavica.
Larson, R. (1988). On the double object construction. Linguistic Inquiry 19, 335–391.
Martohardjono, G. (1993). Wh-movement in the acquisition of a second language: a cross-
linguistic study of three languages with and without movement. Unpublished Ph.D. thesis,
Cornell University.
Montrul, S. (1997). The L2 acquisition of dative experiencer subjects. Second Language
Research 13(3), 264–298.
Nishida, C. (1994). The Spanish reflexive clitic se as an aspectual class marker. Linguistics
32, 425–458.
Pesetsky, D. (1995). Zero Syntax. Cambridge, MA: MIT Press.
Piñon, C. (1993). Nominal reference and the imperfective in Polish and English. Proceedings
of NELS, 383–397.
Pustejovsky, J. (1991). The syntax of event structure. Cognition 41, 47–81.
Sanz, M. (1996). Telicity, objects and the mapping onto predicate types. A cross-linguistic
study of the role of syntax in processing. Unpublished Ph.D. thesis, University of
Rochester.
Schmitt, C. (1996). Aspect and the syntax of noun phrases. Unpublished doctoral
dissertation, University of Maryland at College Park.
Schwartz, B. (1998). On two hypotheses of ‘‘transfer’’ in L2A: minimal trees and absolute
L1 influence. In The Generative Study of Second Language Acquisition, S. Flynn,
G. Martohardjono, and W. O’Neil (eds.), 35–59. Mahwah, NJ: Erlbaum.
—; and Sprouse, R. (1994). Word order and nominative case in non-native language acquisi-
tion: a longitudinal study of (L1 Turkish) German interlanguage. In Language Acquisition
Studies in Generative Grammar, T. Hoekstra and B. Schwartz (eds.), 317–368. Amsterdam:
Benjamins.
—; and Sprouse, R. (1996). L2 cognitive states and the full transfer/full access model.
Second Language Research 12(1), 40–72.
Slabakova, R. (1997a). How light is a light verb? In Proceedings of the Fifteenth West Coast
Conference on Formal Linguistics, Brian Agbayani and Sze-Wing Tang (eds.). Stanford:
CSLI.
—(1997b). Slavic preverbs: aspect in phrase structure. Linguistics 35(4), 673–704.
—(1997c). Zero acquisition: second language acquisition of the parameter of aspect.
Unpublished Ph.D. thesis, McGill University.
—(1999). The parameter of aspect in second language acquisition. Second Language
Research 15(3), 283–317.
Smith, C. (1997 [1991]). The Parameter of Aspect, 2nd ed. Dordrecht: Kluwer.
Travis, L. (1992). Inner aspect and the structure of VP. Cahiers Linguistiques de l’UQAM
1, 130–146.
Vainikka, A.; and Young-Scholten, M. (1994). Direct access to X’ theory: evidence from
Korean and Turkish adults learning German. In Language Acquisition Studies in
Generetive Grammar: Papers in Honour of Kenneth Wexler from the GLOW 1991
Workshops, T. Hoekstra and B. D. Schwartz (eds.), 265–316. Amsterdam: Benjamins.
—; and Young-Scholten, M. (1996). Gradual development of L2 phrase structure. Second
Language Research 12(1), 7–39.
van Hout, A. (1998 [1996 ]). Event Semantics of Verb Phrase Alternations: A Case Study of
Dutch and its Acquisition. New York: Garland.

770 R. Slabakova
Vandler, Zeno (1967). Linguistics in Philosophy. Ithaca, NY: Cornell University Press.
Verkuyl, H. (1972). On the Compositional Nature of the Aspects. Dordrecht: Reidel.
—(1993). A Theory of Aspectuality. Cambridge: Cambridge University Press.
—(1999). Tense, aspect, and aspectual composition. In Topics in South Slavic Syntax and
Semantics, Mila Dimitrova-Vulchanova and Lars Hellan (eds.), 125–162. Amsterdam:
Benjamins.
White, L. (1985). The pro-drop parameter in adult second language acquisition. Language
Learning 35, 47–62.
—(1986). Implications of parametric variation for adult second language acquisition: an
investigation of the ‘‘pro-drop’’ parameter. In Experimental Approaches to Second
Language Acquisition, V. Cook (ed.), 55–72. Oxford: Pergamon.
—(1989). Universal Grammar and Second Language Acquisition. Amsterdam: Benjamins.
—(1996a). Clitics in L2 French. In Generative Perspectives on Language Acquisition,
H. Clahsen (ed.), 335–368. Amsterdam: Benjamims.
—(1996b). Universal grammar and second language acquisition: current trends and new
directions. In Handbook of Second Language Acquisition, W. Ritchie and T. Bhatia (eds.),
85–120. San Diego: Academic Press.


SLABAKOVA (2000) - L1 - Transfer - Revisited - The - L2 - Acquisition

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SLABAKOVA (2000) - L1 - Transfer - Revisited - The - L2 - Acquisition

Uploaded by

Copyright:

Available Formats

L1 transfer revisited: the L2 acquisition of

telicity marking in English by Spanish and

This study addresses the nature of the initial grammar in second-language

Linguistics 38–4 (2000), 739–770 0024–3949/00/0038–0739

Brought to you by | University of Iowa Libraries

Brought to you by | University of Iowa Libraries

to demonstrate precisely the same acquisition sequences of a given L2

2. The parametric distinction

Comparative theoretical studies of aspect within generative grammar

Brought to you by | University of Iowa Libraries

Smith (1997 [1991]) makes the important distinction between view-

2.1. Situation aspect

The most widely accepted ontology-based classification of verbs into

Activities Accomplishments Achievements

know run run a mile die

Brought to you by | University of Iowa Libraries

Aspectual literature often makes use of the term ‘‘telicity.’’ A clause is

(3) Specified cardinality

There are an overt definite and an optional indefinite article in

(5) a. Specified cardinality

Brought to you by | University of Iowa Libraries

The same is true of Spanish:

2.2. Viewpoint versus situation aspect

The second aspectual distinction described by Smith (1997 [1991]) as

Brought to you by | University of Iowa Libraries

aside.3 I will come back to this contrast in aspectual tenses in the

2.3. The English value of the VP-aspect parameter

Brought to you by | University of Iowa Libraries

2.4. The Spanish value

In Spanish, marking telicity also depends on the cardinality of the object

Brought to you by | University of Iowa Libraries

2.5. The Bulgarian value

Bulgarian signals telicity through perfective prefixes, or preverbs (PV ),

Brought to you by | University of Iowa Libraries

Brought to you by | University of Iowa Libraries

2.6. The analysis

In studying various marking of situation aspect in different languages,

Brought to you by | University of Iowa Libraries

mechanisms independently needed for case checking and predicate

Let us concentrate on the English tree in (18a) first. The double VP

Brought to you by | University of Iowa Libraries

cardinality the interpretation is one of a telic event. Whenever the object

3. Rationale for the present study

In previous research (Slabakova 1997c, 1999) I studied the acquisition

Brought to you by | University of Iowa Libraries

These findings suggest that low-proficiency learners considered telic

4. Hypothesis for SLA

Hypothesis 1. Bulgarian learners will perform more accurately in recog-

Brought to you by | University of Iowa Libraries

atelic responses from native speakers of English, the Bulgarian learners

Twenty-two native speakers (NS) of Bulgarian and 21 NSs of Spanish,

5.2. Tasks and materials

5.2.1. Independent measure of proficiency. Participants in the Bulgarian

Brought to you by | University of Iowa Libraries

on the basis of an independent measure of proficiency: a cloze test (see

5.2.2. Aspectual interpretation task. Participants were asked to assess

Brought to you by | University of Iowa Libraries

constituents, compatible not only syntactically but semantically as well.

6.1. Independent measure of proficiency

One of the conditions for establishing transfer effects is to use learner

Table 1. Mean and standard deviation of participant groups in cloze task

Participant groups M SD Score range

NAmE controls (n=16) 25.9 2.64 21–31