Professional Documents
Culture Documents
research-article2021
SLR0010.1177/02676583211017604Second Language ResearchGupton and Sánchez Calderón
second
language
Original Article research
reconsidered
Timothy Gupton
University of Georgia, USA
Abstract
We examine the second language (L2) acquisition of variable Spanish word order by first language
(L1) speakers of English via the acquisition of unaccusative and transitive predicates in various
focus-related contexts. We employ two bimodal linguistic tasks: (1) acceptability judgment task
(B-AJT) and (2) appropriateness preference task (B-APT). Both present contextualized prompts
similar to previous studies, followed by response options with accompanying audio to control for
intonation and pauses. Results suggest a number of key findings: (1) by the high intermediate level,
L2ers acquire the relevant syntactic and syntax information structure interface competencies
for both predicate types; (2) native speakers and L2 groups exhibit optionality, and only differ in
nuanced ways; and (3) advanced learners show signs of acquiring syntactic and syntax–information
structure competencies in numerous contexts, but display minor differences regarding optionality
with corrective focus, an interface incorporating multiple interfaces (syntax–prosody pragmatics).
Unlike the predictions of the Interface Hypothesis (IH), this subtle, non-native-like divergence is
characterized by divergent knowledge of optionality similar to that found among native speakers.
Attempting to understand more completely the development of native-speaker optionality, we
also conduct a corpus study of child-directed Spanish from CHILDES and find that, although
syntactic theory explains much of the data, it cannot account for all of the variability in the data.
Results suggest that children are exposed to apparent optionality from the earliest stages.
Keywords
discourse, English, information structure, Spanish, syntax–focus interface, unaccusativity
Corresponding author:
Timothy Gupton, University of Georgia, 210 Herty Drive, Athens, GA 30602-0002, USA.
Email: gupton1@uga.edu
2 Second Language Research 00(0)
I Introduction
The issue of optionality is an obstinate one in generative linguistic research and
theorizing. Within a Minimalist model of the grammar driven by feature-checking
(e.g. Chomsky, 1995, et sequens), different derivational outcomes should be moti-
vated by a different makeup of lexical and functional features within the Numeration.
Viewed through this lens, all syntactic operations should have an effect on the even-
tual derivational outcome and the associated interpretation of a given utterance.
Within a system like this, optionality is undesirable. Nonetheless, the empirical data
suggest otherwise: optionality has proven to be a stubborn problem in the analysis of
scrambling phenomena (e.g. Fukui, 1993). It has also been attested among aphasics
and in first-language acquisition, although transitorily in the case of the latter (for
discussion, see, for example, Avrutin, 2004). In recent years, researchers in multilin-
gualism have taken a particular interest in optionality, noting its presence among
child second-language (L2) acquirers (Pladevall Ballester, 2010), adult L2 acquirers
(Sorace, 1999, 2000a; White, 1992), and among bilingual Heritage Speakers in the
United States (de Prada Pérez and Pascual y Cabo, 2012; Silva-Corvalán, 1994). It
also forms a crucial part of the Interface Hypothesis (IH, e.g. Sorace, 2011; Sorace
and Filiaci, 2006), which seeks to make predictions about outcomes among adult L2
acquirers, in particular among extremely successful adult L2 acquirers. This theo-
retical perspective assumes continuity with respect to access to Universal Grammar
consistent with Full-Transfer/Full Access (FT/FA) approaches (e.g. Schwartz and
Sprouse, 1996) such that advanced acquirers of an L2 will exhibit signs of grammar
restructuring at the level of the narrow syntax. For the IH, optionality is seen as
residual, and suggestive of instability at the point of interface between the syntax and
other linguistic modules. This is because the first language (L1) is understood to
instantiate an additional alternative that is not obviously unavailable in the L2. On
this view, optionality is something that should be eliminated from the emerging sys-
tem, but is not. The examination of numerous syntactic competencies brings to the
foreground leading syntactic accounts, such as those related to the syntax of different
predicate types, such as Perlmutter’s (1978) Unaccusative Hypothesis (UH) and
Baker’s (1988) Uniformity of Theta Assignment Hypothesis (UTAH). Despite
learner successes, however, residual optionality has been found at the interface of
syntax with other external modules of the grammar, such as discourse, pragmatics,
and information structure, and recent formulations of the IH (e.g. Sorace, 2011) have
sought to explain why. In the case of Spanish, Zubizarreta (1998), a syntactic pro-
posal that accounts for information structure – in particular, focus – has received a
great amount of attention owing to the predictions it makes for the syntax as well as
the syntax–information structure interface. In the current article, we seek to answer
the following initial research questions:
as the basic sentence-generating mechanism. The most basic, default word orders in
Spanish are found in sentences lacking previous context. For Zubizarreta (1998),
these are spoken ‘out of the blue’ or in response to a ‘Guess what?’ question. This is
similar to Kuroda’s (1972) thetic sentences, in which no constituent is privileged via
previous mention within a given discourse context. The default word order for
Spanish with regular, monotransitive predicates is SVO, the same as in English.1
Intransitive predicates exhibit numerous asymmetries, which has become known as
‘split intransitivity’. According to Perlmutter’s (1978) Unaccusative Hypothesis,
there are two types of intransitive predicates: unaccusative and unergative.2
Unaccusative predicates select a DP complement that is generated in the same syn-
tactic position as theme objects of transitive predicates, while unergative predicates
select a DP that is generated in the typical (preverbal) subject position. Viewed
through the lens of Baker’s (1988) Uniformity of Theta-Assignment Hypothesis, by
which theta roles are uniformly assigned to dedicated structural positions, we have
the structures in (3):
It may be this similarity that is responsible for the tendency noted by Oshita (2001), who
proposed ‘the unaccusative trap’, by which L2 acquirers initially classify intransitive
Gupton and Sánchez Calderón 5
The Spanish L2 acquirer has to realize that, in addition to preverbal subjects, Spanish
allows for postverbal subjects, and therefore, the edge-feature or [D] feature in TP
need not require movement to Spec, TP. This may be relevant for word order related
to the narrow syntax because these are thetic sentences that trigger the unmarked,
default word order. In narrow-focus contexts, however, there are more constituents at
play and more possible word orders. Therefore, from the complexity perspective, we
predict that L2 acquirers of Spanish should acquire native-like competence with
intransitive (unaccusative) predicates in narrow-focus contexts prior to transitive nar-
row-focus contexts. This motivates our third research question:
• Research question 3: Given their relatively greater complexity, are transitive pred-
icates more difficult to acquire than intransitive predicates?
With the exception of the detail related to the position of subjects of unergative
predicates illustrated in (4), the analyses above are crucial in that they are explicitly
assumed by the studies that we review in Section III. We summarize these word order
6 Second Language Research 00(0)
Thetic
context
Unaccusative predicate VS
Unergative predicate SV (VS)
Transitive predicate SVO
predictions associated with the predicate types and discourse contexts examined in
Table 1. Despite the predictions of the narrow-syntactic approach, following
Zubizarreta’s (1998) account of Spanish, not all word orders are felicitous; rather,
different word orders are available according to the discursive context at hand.
Assuming the structures in (4a, b) and (5) above then, the subject DP may remain in situ
in its postverbal position and is not required to undergo leftward, prosodic scrambling
(p-movement) in order to ensure proper alignment between the FPR and the R-NSR.4
Rather, movement of the F-marked subject to the preverbal position is claimed to trigger
a corrective/contrastive reading on the subject. To account for these possibilities, we add
narrow-focus word order predictions to Table 1.
Corrective/contrastive focus fronting is of particular interest to the current study
because it also represents a focus-related property.
Gupton and Sánchez Calderón 7
As we can see in (9), immediately following the waiter’s erroneous assertion, the cor-
rective focus reply simultaneously opens a variable (‘I ordered X’) and resolves it with
the correct information (‘X = a macchiato’).5 For Zubizarreta (1998), fronted focus
may or may not involve syntactic movement to the left edge, but information focus
(i.e. narrow focus) must target the right edge. An alternative possibility following
Titov’s (2012) account of Russian would posit that focus at the rightmost syntactic
edge is for information focus encoding, while movement to a left peripheral position
(e.g. FocP as in Rizzi, 1997) is the result of a disambiguating operation to encode con-
trast/correction, which may be motivated by criterial feature checking.6 Here, we
remain agnostic regarding the precise details of corrective focus fronting, but highlight
that more than one syntactic position is available for the corrective constituent: the in
situ position and the ex situ (i.e. fronted, left-edge) position. If variation among native
speakers for narrow-focus encoding allows in situ information focus, i.e. at the left or
right edge, then both focus-related phenomena – corrective focus and narrow-focus –
may be interpreted as potentially invoking optionality.
Sorace and colleagues proposed the Interface Hypothesis (IH) in order to account for
subtle non-target (i.e. non-native) performance among extremely advanced L2 acquirers.
In its original formulation (Sorace and Serratrice, 2009), the interface between syntax
and other modules of the grammar were predicted to result in instability and prolonged
optionality.7 In more recent research on the IH has found that not all syntactic interfaces
are equally problematic thus motivating a proposed divide between internal and external
interfaces (for a review, see, for example, Sorace, 2011; White, 2011a). Internal inter-
faces, such as those involving the syntax and modules such as the lexicon or semantics,
both of which may be implicated in the acquisition of split intransitivity, are predicted to
be unproblematic in second language acquisition (SLA) following a period of initial L2
transfer. The acquisition of transitive predicate word order is similarly predicted to be
unproblematic when involving an internal interface. External interfaces, however, are
predicted to result in greater instability and residual optionality because they involve an
interface between the syntax and an external module of the grammar, such as discourse
pragmatics. It is for this reason that the syntax–information structure interface then is a
prime candidate for exploration related to the IH. Of the examples reviewed above, those
involving sentences uttered ‘out of the blue’ or in response to a ‘What happened?’ sort of
8 Second Language Research 00(0)
question should involve default syntactic orders and thus not pose a serious problem for
acquisition since they do not invoke discourse-pragmatics. Replies to subject narrow-
focus questions (information structure) and corrective focus (pragmatics), however,
involve a syntax–discourse interface and are therefore predicted to result in residual
optionality owing to the greater cognitive load implied by the task of reconciling the
syntax with the demands of the interface, which as we have seen above, also incorporates
prosody. An empirical complication for this prediction lies in recent studies of Spanish
(e.g. Hoot, 2016; Jiménez-Fernández, 2015; Leal et al., 2018; Muntendam, 2013) that
have found that narrow focus can target constituents in non-rightmost positions. In short,
native speakers also accept and produce in situ replies to narrow-focus questions. To our
knowledge, whether this additional possibility invokes optionality is an unexplored
question. Nevertheless, it stands to reason that native speakers may exhibit syntactic
optionality, at least with respect to focus-related phenomena. If this is the case, the notion
of residual optionality in advanced L2 acquirers must be re-examined. Therefore, we
include two research questions to further guide our exploration of optionality.
The IH does not comment on native-speaker optionality, but its existence may go some
length in explaining it among advanced L2 acquirers. Given that this study examines two
focus-related phenomena bearing potential for optionality among native speakers, we
expect the answer to this question to be positive. Nevertheless, the mere presence of
optionality among native speakers does not guarantee that optionality among advanced
L2 acquirers will necessarily be similar. Therefore, we also propose the following
research question:
The IH predicts that very advanced speakers will exhibit residual optionality with inter-
face-related phenomena at external interfaces.
The IH as formulated in Sorace (2011) is not intended to predict or describe the devel-
opment of interface-related competencies among emerging bilinguals. However, we fol-
low suggestions by other researchers (Lardiere, 2009; Leal-Méndez and Slabakova,
2011; White, 2011b) that an examination of developmental trends among learners is
fundamental to learning more about how residual optionality develops and/or results.
Therefore, we add one more exploratory research question:
(10) You are at a party with your friend Magda. Magda leaves the room to get a drink, and
while she is gone a man whom you don’t know leaves. When she returns, she notices that
someone has left, but she is not sure who. Magda asks you: ¿Quién salió? (‘Who left?’) You
answer:
For (10), participants rated the acceptability of each sentence on a 7-point Likert scale
(−3 to 3) and then answered a question asking which sentence they preferred (A, B, or
No preference). The second task involved a similar contextualized prompt as (10), but
was followed by the question ‘What do you answer?’ There were four Spanish L2er
groups and a native-speaker control group from an unspecified selection of countries.
Linguistic history questionnaire results were used to exclude natives who did not other-
wise fit the desired native-speaker profile. At the time of data gathering, all participants
were living and studying or working in the USA and were reported to have advanced
English proficiency. The tasks included seven unaccusative predicates (desaparecer
‘disappear’, entrar ‘enter’, escapar ‘escape’, llegar ‘arrive’, nacer ‘be born’, salir
‘leave’, venir ‘come’) and seven unergative predicates (bailar ‘dance’, cantar ‘sing’,
dormir ‘sleep’, estornudar ‘sneeze’, gritar ‘shout’, llorar ‘cry’, nadar ‘swim’).
The results of the written task found that early-stage L2ers produced very little VS
order, and this continued even at the advanced level. Focusing on the advanced L2ers,
post-hoc statistical measures revealed only one significant statistical difference in com-
parison with the native-speaker group, and this was in the statistically higher percentage
of VS word order produced with unergative predicates in thetic sentences. The results of
the second task revealed that, overall, participants preferred VS order more than they
actually produced it, which is suggestive of emerging changes in the mental grammar of
the L2: as L2 level increased, there was a growing preference for VS. Advanced L2ers
actually preferred VS at statistically higher rates than the native speaker group, with the
exception of unaccusative predicates in narrow-focus contexts. Hertel (2000, 2003) do
10 Second Language Research 00(0)
not explicitly discuss these results within the context of the IH. Notwithstanding,
advanced learner results run contrary to the predictions of the IH.
What is of particular interest in Hertel (2000) is the data on the ‘no preference’
option. These results, which are suggestive of optionality, were not discussed and
only reported on in the Appendix (Hertel, 2000: 199). Although ‘no preference’
optionality was present in all groups and all contexts, we do not interpret this to mean
that all such optionality is equal. One might assume that optionality at a lower com-
petence level is representative of indeterminacy and less experience with the L2,
while at higher levels, it may be representative of a deeper knowledge of multiple
possibilities and variation in language.
Lozano (2006) examined two groups of advanced L2 acquirers of Spanish: a group
of British English L1 speakers (n = 17) studying Spanish at the University of Essex
and the other a group of Greek L1 speakers (n = 18) studying Spanish in Greece.
Controls were 14 Spanish native-speakers who were pursuing post-graduate studies in
the UK at the time of testing. Most of these native speakers were from Spain, but par-
ticipants were also from Argentina, Mexico, and Venezuela. Although no comments
are made regarding their English proficiency, it is fair to assume that they all knew
English at a rather high level since they had been admitted to a British university.
Lozano’s participants completed a contextualized AJT with paired SV/VS replies simi-
lar to Hertel’s (2000) AJT methodology, but with two important modifications: (1) a
5-point Likert scale (from 2 to −2) was used, and (2) there was no question about
response preference, meaning that preference was extrapolated from acceptability rat-
ings. The task included many of the same predicates as Hertel’s study, with the excep-
tion of the unergative predicate reír ‘laugh’ and the unaccusative predicate volver
‘comeback/return’, which did not appear in Hertel’s methodology.8 There were 24 tar-
get stimuli involving predicate type and focus type, with six unergative and six unac-
cusative predicate stimuli for narrow-focus contexts as well as six of each for thetic
contexts. All of the speaker groups preferred VS to SV in thetic sentences with unac-
cusative predicates. Additionally, there were no statistical differences among groups
for ratings of SV or VS. The three speaker groups preferred SV to VS word order in
thetic sentences with unergative predicates. However, there was a significant differ-
ence with respect to the acceptability of VS word order, with both L2 groups accepting
it significantly more than the Spanish natives. In narrow focus contexts, Spanish native
speakers preferred VS to SV, rating VS significantly higher than SV. The learner
groups, however, did not distinguish between SV and VS word order. Lozano inter-
prets equal acceptability combined with a higher SV rating than natives as optionality
among the L2 groups. A similar situation applies to ratings of unergative predicate
word order replies to narrow-focus contexts: native Spanish speakers rated VS signifi-
cantly higher than SV, but the L2 groups each rated SV slightly higher than VS, not
significantly higher. The fact that the learner groups both exhibit optionality with dis-
course-related stimuli leads him to conclude that the optionality present is consistent
with the IH.
Domínguez and Arche (2008, 2014) employed a contextualized preference task based
on Hertel (2003). Their study differed with respect to the participants, with 60 English-
L1–Spanish-L2 acquirers in the UK between 13 and 22 years of age. The native speaker
Gupton and Sánchez Calderón 11
controls (n = 20) were 17–18-year-old high school students in Spain. The authors sought
to determine if the L2 acquisition of VS word order was more dependent on purely syn-
tactic phenomena or on pragmatic and discursive information. Participants were pre-
sented with sentence-pair continuations of contexts like (10) and were instructed to
choose one of the paired sentences or the option ‘Both’. Domínguez and Arche interpret
the results as showing that VS word order takes some time for L2ers to acquire. Although
the advanced-level learner group in their study shows evidence of having acquired VS,
their results are not native-like and show a high level of optionality in their ‘Both’
response percentages for narrow focus and thetic contexts, the latter of which is not pre-
dicted by the IH. The authors suggest, following Papp (2000), that this optionality is due
to insufficiently transparent input resulting in a syntactic deficit related to the licensing
of pro in intransitive sentences and incomplete acquisition of split intransitivity. They
suggest that future studies examine the input that L2 acquirers are exposed to.
Parafita Couto et al. (2015) are innovative in examining Sorace’s (1995, 2000b) hier-
archy for intransitive predicates (Figure 1), with a desire to see if it plays a role in the
acquisition and processing of split intransitivity among English-L1–Spanish-L2 acquir-
ers. According to this hierarchy, top and bottom extremes are the most prototypical for the
corresponding predicate type, and exhibit the least syntactic variation. The authors
employed a contextualized AJT (5-point Likert with ‘I don’t know’ option) similar to
Lozano (2006), but with an expanded set of 76 distinct verbs in three subdivisions (low,
medium, high) of prototypicality. Stimuli consisted of focus (narrow-focus) and non-
focus (thetic) contexts with unergative and unaccusative predicates for each context type.
Experimental participants were 20 English-L1–Spanish-L2 acquirers aged 29–72 (M =
46.6) who had lived in Madrid for an average of 20.7 years. These were educated Spanish
L2ers who used Spanish on a daily basis and rated their own Spanish abilities as near ceil-
ing (M = 3.74/4.0). Native speaker controls were 20 Spanish monolinguals aged 19–31
(M = 23) living in Madrid.
They found that SV is favored for unergative predicates and unaccusative predi-
cates in thetic contexts, but that VS was rated slightly higher for high- and medium-
level prototypical predicates. In narrow-focus contexts, however, VS order was
12 Second Language Research 00(0)
preferred for both predicate types. Against the predictions of Zubizarreta (1998), SV
word order was also highly accepted for a large number of predicates – even in narrow-
focus contexts – by the bilinguals as well as the monolinguals. Importantly, the
advanced L2 group treated unaccusative and unergative predicates identically, sug-
gesting that they have not converged on native-like competence. The results did not
support Sorace’s Hierarchy, with no clear distinction between intransitive verb types,
similar to de Prada Pérez and Pascual y Cabo (2012). Additionally, given that there was
no differential outcome or distinction between focused and non-focused contexts, they
find that the results do not support the IH.
Despite differences related to methodology and results, the studies above have much
in common. They all find that early L1 English acquirers of L2 Spanish depend on
English SV(O) word order early on, but these L1 transfer effects are eventually over-
come and L2ers begin to accept VS order. At higher levels, optionality is present, either
explicitly, via a ‘no preference’ or ‘both’ reply in preference tasks, or implicitly, via simi-
lar ratings of SV and VS word orders. The definition of ‘advanced-level’ acquirers varies
and does not always involve near-native L2 Spanish speakers. These studies also part
ways on the root cause of the non-native-like results, with some attributing them to a
syntactic deficit not predicted by the IH, while others claim that they are due to interface
instability, in line with the IH. The origin of the persistent optionality found remains
unclear. It is unknown if it is a product of residual interference from the L1 system, or
reflective of the target L2. Many of the previous studies reviewed in this section have
neglected native-speaker results that do not adhere to predictions, leading us to suspect
that a proper examination of native-speaker optionality is in order to better understand
optionality involving multiple syntactic positions.9
IV Methodology
Participants in this study consisted of 33 undergraduate students (28 female, M = 20.3
years) at a public state university in the southeastern United States and were enrolled in
third- and fourth-year coursework in Spanish, meaning that all of these English L1 par-
ticipants had chosen a minor or major specialty in Spanish. There was also a very
advanced group (AGG, n = 13, 10 female, M = 37.7 years) of English-L1–Spanish-L2
acquirers consisting of Spanish teaching assistants, instructors, and professors. The
native-speaker (NS) control group consisted of participants who were from Argentina,
Peru, Spain and other Latin American countries (NS, n = 20, 13 female, mean age =
38.6). Eleven of these participants were residing in the US at the time of data collection,
while the other nine were residing in Valencia, Spain.10 Those of whom were living in
Spain had various connections to a study abroad program organized by that same
American university in Valencia. Participants were not paid for their participation.
The methodology consisted of informed consent followed by an initial linguistic history
questionnaire consisting of nine questions, which served to disqualify potential participants
who did not have the desired profile. Next, participants completed a series of untimed,
Internet-based questionnaires in three blocks with 91 questions total. This total included a
20-question bimodal acceptability judgment task (B-AJT) and a 20-question bimodal appro-
priateness preference task (B-APT), which were presented in counterbalanced blocks
Gupton and Sánchez Calderón 13
pseudo-randomly assigned as part of the informed consent process. Each was preceded by
three training questions. The B-AJT and the B-APT followed the model employed in Hertel
(2003) and Lozano (2006), by which participants read a paragraph-length context followed
by pair-sentence replies similar to (10), but with accompanying audio recordings of each
response option to avoid participants accommodating responses with unintended pauses or
intonation, following recommendations in Kitagawa and Fodor (2006). As we saw in (6–8),
this is especially important given the syntactic flexibility of Spanish and the differences of
meaning that may be encoded via prosodic stress on a particular constituent. We provide
examples of these in the subsections below. After completing these questionnaires, partici-
pants completed a shortened version (36 questions) of the University of Wisconsin Spanish
placement test. Each questionnaire took 15–20 minutes, following the suggestions of market
research (e.g. la Bruna and Rathod, 2005), which suggest that the ideal questionnaire length
for avoiding fatigue is 17–20 minutes. Additionally, participants were invited to rest between
blocks if necessary. The total participation time for most participants was 45–60 minutes.
(11) Context: You are at a café with a group of friends. Everybody orders drinks and coffees.
While the waiter is preparing the order, your colleague Beatriz goes outside to make a call. The
waitress arrives at the table and, seeing that there is a coffee extra, asks you: ‘What happened?
Weren’t there more people here?’ You answer:
Participants evaluated each possible word order response using a five-point radio button
Likert scale (12). These definitions always appeared on screen.
In sum, each participant rated the acceptability of 40 sentences as part of the B-AJT. All
response option recordings were made by a male and female native speaker of Peninsular
Spanish and were counterbalanced.
Given this state of affairs, it is possible that learners have been exposed to such variation.
Consider the following sample context from the B-APT (13):
(13) Context: You are in a restaurant with your family. After eating, you order a café
macchiato, but the waiter brings you a latte. You tell him that there has been a mistake. He says
to you: ‘But you ordered a latte, right?’ You correct him:
Gupton and Sánchez Calderón 15
As we can see in (13), the corrective focus reply simultaneously opens a variable (‘I
ordered X’) and resolves it with the correct information (‘X = a macchiato’).12 For
Zubizarreta (1998), fronted (corrective) focus may or may not involve syntactic move-
ment to the syntactic left edge, but information focus (narrow focus) must target the right
edge. An alternative possibility following Titov’s (2012) account of Russian would posit
that focus at the right syntactic edge is for information focus encoding, while movement
to a left peripheral position (e.g. FocP as in Rizzi, 1997) is the result of a disambiguating
operation to encode contrast/correction, which may be motivated by criterial feature
checking. If variation among native speakers allows in situ information focus, i.e. at the
left or right edge, then both focus-related phenomena may be interpreted as invoking
optionality. To gain insight on such optionality, the following preference scale (14)
appeared after the APT contexts:
There were five tokens each for subjects and objects in narrow-focus contexts and five
tokens of each in corrective-focus contexts, netting a total of 20 target stimuli with 40
possible replies. All of these appear in Appendix 2. As with the B-AJT, participants were
also asked to listen to audio files of each paired-sentence reply to ensure that they cor-
rectly understood the intended meaning.
Table 2. Bimodal acceptability judgment task (B-AJT) mean appropriateness ratings, standard
deviation, and ranges of word orders in thetic contexts by predicate type and speaker group.
Interface instability among the AGG group will involve optionality that is divergent
from the NS group.
V Results
Examination of the placement test results confirmed that the very advanced group (M =
34.9 years, range 31–36 years) and the native speaker groups (M = 34.0 years, range 30–36
years) were at or near ceiling, but the undergraduate learner group exhibited differences, and
was divided into two sub-groups using the lowest native-speaker score (30 points) as the
cut-off. This netted a low-intermediate group (LG1, n = 15, M = 25.7 years, range 17–29
years) and a high intermediate group (LG2, n = 18, M = 32.4 years, range 30–36 years).13
1 B-AJT results
The data gathered in the B-AJT was analysed using R Studio (RStudio Team, 2020)
using the dplyr package. Descriptive statistics, including mean acceptability ratings,
standard deviation, and high/low rating ranges are provided for thetic contexts in Table 2
and for narrow-focus contexts in Table 3.
To examine the effects of the independent variables of speaker group (NS, AGG, LG2,
LG1), predicate type (transitive (TR) vs. unaccusative (UNACC)), word order (SV vs. VS),
and context type (thetic (TH) vs. subject narrow-focus (SNF)) on the dependent variable
(acceptability rating), linear mixed-effects regressions were run using the LMER package
(Bates et al., 2015) in RStudio. We performed four regressions in total to determine the pre-
cise interactions between predicate type and context type, and their influence on the accept-
ability of VS word order. A random effect was entered for stimulus item and individual
participant number.14 The results for the linear regression models appear in Tables 4–7 in
Appendix 3. Under fixed effects, looking at the p-values for inter-group (Group) compari-
sons, we can see that the only group that exhibited significant differences from the native
Gupton and Sánchez Calderón 17
Table 3. Bimodal acceptability judgment task (B-AJT) mean appropriateness ratings, standard
deviation, and ranges of word orders in subject narrow-focus contexts by predicate type and
speaker group.
controls was the lowest-level intermediate group, LG1 (p < 0.001). We can see in Tables 4–7
that the effects of individual variables and paired-variable interactions are often conflicting
and contradictory. Therefore, we focus on the p-values for the interactions between the three
variables: predicate type (TR/UNACC), context (TH/SNF), and (VS) word order.
In Table 4 we see that the p-value for transitive predicates in subject narrow-focus con-
texts with VS word order favor is significant (t = 3.559, p < 0.001). The directionality of
the t-value indicates that the interaction between these factors favors a high acceptability
rating. In Table 5, the interaction of transitive predicate, thetic context, and VS word order
favors a low rating (t = −3.559, p < 0.001). In Table 6, the interaction of unaccusative
predicate with subject narrow-focus context and VS word order favors a significantly low
level of acceptability (t = −3.559, p < 0.001). In Table 7, the interaction of thetic predicate,
thetic sentence context, and VS word order (t = 3.559, p < 0.001) favors a high acceptabil-
ity rating. We discuss the implications of this data further in Section VI.
To address research questions 2, 4, 5, and 6, we compared ratings of SV and VS word
orders for each response dyad rated in the B-AJT. Optionality was coded as ‘1’ where
participants gave SV and VS orders identical acceptability ratings. We ran linear mixed-
effects regressions to provide insight on the factors favoring optionality. These regres-
sion models appear in Tables 8–11 in Appendix 3.
In each table, we can see inter-group comparison results in t-values and p-values, finding
no statistical differences between learner groups and the NS controls. In Table 8, we see that
the interaction between subject narrow-focus (SNF) and a transitive predicate (TR) strongly
favors optionality (t = 3.956, p < 0.001). In Table 9, the interaction between transitive
predicate (TR) and thetic sentence (TH) context does not favor optionality (t = −3.956, p <
0.001). The interaction between unaccusative predicate (UNACC) and subject narrow-focus
(SNF) context, seen in Table 10, also does not favor optionality (t = −3.956, p < 0.001).
Table 11 shows that the interaction between an unaccusative predicate (UNACC) and a
18 Second Language Research 00(0)
Figure 2. Results for bimodal appropriateness preference task (B-APT): Object narrow-focus
context, VO proportions by group.
thetic sentence (TH) context strongly favors optionality (t = 3.956, p < 0.001). We discuss
the theoretical implications of these outcomes in Section VI.
2 B-APT results
Recall that, for the appropriateness preference task, participants only rated replies for
contexts involving narrow focus or corrective focus. Here, we report preference propor-
tions for responses A-C from the forced-choice response scale.15 Upper and lower limits
of Wilson 95% confidence intervals are represented by error bars. Overlapping propor-
tions among groups (visualized in the error bars) indicate an absence of statistical differ-
ences. Non-overlapping proportions indicate statistically significant differences. We start
with object narrow-focus contexts, for which VO was the preferred reply across groups
(Figure 2). Here, there were no significant inter-group differences.
Gupton and Sánchez Calderón 19
Figure 3. Results for bimodal appropriateness preference task (B-APT): Subject narrow-focus
context, VS proportions by group.
For subject narrow-focus contexts, VS was the most common reply among the
advanced L2ers (Figure 3). There was no significant difference between the AGG group
and the native speaker group, but both of those groups differed from the lower level L2er
groups. The most common reply was SV among the LG1 group (0.5818) and the LG2
group (0.4421).
As we can see in Figure 4, the native speaker group preferred the response indicat-
ing that SV and VS are equally acceptable. Curiously, all of the learner groups recog-
nized the possibility of both in a manner similar to the native speakers.
For contexts that elicited corrective focus on an object constituent, all of the speaker
groups preferred VO word order (Figure 5). Although this preference was slightly lower
among native speakers, this was not a statistically significant difference.
20 Second Language Research 00(0)
Figure 4. Results for bimodal appropriateness preference task (B-APT): Subject narrow-focus
context, VS = SV reply proportions by group.
Native speaker ratings were lower in part because of higher response proportions
indicating optionality of VO and OV forms, as in Figure 6. The L2 groups appear to
recognize this possibility. Their replies did not differ significantly from the NS
group.
For replies eliciting a contrast involving the subject, VS was the preferred word order
among the advanced L2 (AGG) group (Figure 7). This difference compared to native
speakers was not statistically significant. The AGG and the NS group rating proportions
significantly differed from the lower-level L2 (LG1, LG2) groups. The highest-rated
reply among native speakers was the one recognizing the possibility of optionality: a SV
or a VS reply. Figure 8 shows that this preference differed statistically from all of the L2
groups. SV word order was the preferred reply for the LG1 group (0.6364) and the LG2
group (0.5474). Figure 9 shows that all groups exhibit similar levels of optionality
(VO = OV) for object narrow focus contexts.
Gupton and Sánchez Calderón 21
Figure 5. Results for bimodal appropriateness preference task (B-APT): Object corrective
focus context, VO reply proportions by group.
Figure 6. Results for bimodal appropriateness preference task (B-APT): Object corrective
focus context, VO = OV reply proportions by group.
Figure 7. Results for bimodal appropriateness preference task (B-APT): Subject corrective
focus context, VS reply proportions by group.
the native speakers chose the answer permitting optionality. This view of optionality sug-
gests that interface optionality may not, in fact, be a symptom of non-target L2 acquisi-
tion, but rather an inherent part of native-like competence, be it monolingual or
multi-lingual. In generative syntactic models, optionality is generally considered unde-
sirable, yet a descriptively adequate account seeking to improve on Zubizarreta (1998)
may eventually need to explain it.
Regarding research question 3, regressions performed on the B-AJT data collected
and reported on here suggest no reason to suspect that transitive predicates are more dif-
ficult to acquire than intransitive predicates. If such an asymmetry exists, it is possible
that these particular L2 learners have already progressed beyond such a stage.
In response to Question 4 and whether native speakers exhibit syntactic optionality,
we have seen that their performance on a number of experimental tasks suggests an
affirmative answer. Regarding research question 5, we have seen that the advanced
24 Second Language Research 00(0)
Figure 8. Results for bimodal appropriateness preference task (B-APT): Subject corrective
focus context, SV = VS reply proportions by group.
learners (AGG) exhibited subtle optionality in comparison with the native speaker con-
trols in this study. We interpret these results as being consonant with the Interface
Hypothesis (IH); however, we would like to emphasize that the results of this and other
studies strongly suggest that optionality may also be found among native speakers.
Therefore, optionality itself is not inherently residual or a sign of instability. In fact, in
answering research question 6, the regression data in Tables 4 —11 suggest that learners
successfully acquire native-like optionality related to the interface of narrow syntax and
the syntax–information structure by the high intermediate level. It appears that the
acquisition of corrective focus, which incorporates the interface of narrow syntax,
information structure, and prosody, does not reach the same native-like levels with
respect to subtle optionality. The results from the B-APT examined in Figure 8 capture
the one point of difference of the advanced (AGG) group. Treating all interfaces equally,
the data examined in this study might seem to be suggestive of early acquisition of
Gupton and Sánchez Calderón 25
Figure 9. Results for bimodal appropriateness preference task (B-APT): Object narrow-focus
context, VO = OV reply proportions by group.
1 Examining input
Numerous paradigms in SLA research agree on the importance of linguistic input. In
recent years, SLA researchers working in the variationist paradigm have examined
learner-directed input (e.g. Gurzynski-Weiss et al., 2018; Long and Geeslin, 2020) to
examine the variables at play in learner-directed input produced by native and non-native
speaker language instructors. Given the unexpectedly high ratings of SV with unaccusa-
tive predicates by native speakers in our study and in others, we examined and analysed
child-directed Spanish input from the CHILDES database (MacWhinney, 2000) to deter-
mine what factors influence subject position in the emerging L1 grammar.18 We extracted
3,739 Spanish clauses (identified by the presence of a finite verb/predicate) produced by
23 different individuals in the corpus.19 Given that constituent type influences whether
subject–verb inversion takes place in Spanish questions (see, for example, Torrego,
1984), all interrogative sentences were eliminated. We also eliminated all null-subject
sentences since those would not be instructive with respect to factors that influence sub-
ject position. We coded the remaining declarative clauses as follows: the dependent vari-
able, subject position (±preverbal), and two independent variables: intransitive predicate
type (±unaccusative), and whether the sentence was a narrow focus response (±focus).
Given that the majority of the stimuli on our AJT were animate and human, we also
sought to determine if these factors might lead to unaccusative predicates being analysed
as agentive/causative predicates, similar to unergatives, which might result in greater
levels of SV word order. Therefore, we also coded two additional independent variables:
animacy (±animate) and whether subjects were human (±human). Given that all varia-
bles were binary, we used R Studio (RStudio Team, 2020) using the plm package to run
a random-effect logistic regression on the remaining 478 clauses produced by 16
Gupton and Sánchez Calderón 27
different individuals (9 female, 7 male) from Spain (n = 14) and Mexico (n = 2), with
participant as the random factor. The very little sociolinguistic information available for
the subjects in this corpus was limited to gender (±male) and country of origin
(±Peninsular). We offer a summary of the different independent variables examined as
well as the dependent variable (subject position) in the cross-tabulations in Table 12.
The only factors that were found to be statistically significant were predicate type and
focus type, with focus favoring a postverbal position (Estimate coefficient −0.25, stand-
ard error 0.06, p < 0.001) and unaccusative predicates also favoring a postverbal posi-
tion for subjects (Estimate coefficient −0.19, standard error 0.05, p < 0.001). Whether a
subject constituent was human (Estimate coefficient −0.0003, standard error 0.06, p =
0.996) or animate (Estimate coefficient −0.02, standard error 0.06, p = 0.708) had no
significant influence on subject position. The sociolinguistic variables were admittedly
limited in scope: male (Estimate coefficient −0.14, standard error 0.10, p = 0.165), and
Peninsular (Estimate coefficient −0.04, standard error 0.17, p = 0.833). These were not
significant, either.
Acknowledgements
We would like to thank Gary Baker, Stacey Casado, Nicole DeCraene, Inma Garnes, Jim Jones, Araceli
Peris, Leonor Sales, and Julia Sturm for logistical and technical support. We would also like to thank
Tania Leal, Kim Love-Myers, Jamie Monogan, Melissa Whatley, and Wenbo Wu for discussion of the
statistical analysis and results. We would also like to thank our anonymous reviewers and the editors of
Second Language Research for their helpful discussion. All remaining errors are our own.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this
article.
ORCID iD
Timothy Gupton https://orcid.org/0000-0003-4970-823X
Notes
1. This is not the case for a subset of psychological predicates similar to gustar ‘to please’,
which select a dative-marked experiencer argument in subject position. We do not consider
such predicates in this study, but they clearly form part of the input that the L2 acquirer is
exposed to, especially in the formal foreign language classroom. For more on the Spanish L2
acquisition of this by English L1 speakers, see, for example, VanPatten, 2004.
2. For an opposing critical view of the Unaccusative Hypothesis, see Mackenzie, 2006.
3. Here, we reduce unaccusative predicates to VP alone for the sake of simplicity. If unaccusa-
tive predicates are understood to also project a little v level, this must be a defective v that is
unable to assign an agent/cause theta role. Despite the fact that they only select a DP, assign-
ing it an agent/cause theta role in Spec, vP, unergative predicates may be understood to have
30 Second Language Research 00(0)
a big V level (VP) at which they merge with an abstract, phonologically null object comple-
ment. This has crosslinguistic validity given that in many languages, unergative predicates
alternate with a transitive variant such as in English dream (a dream) and Spanish gritar (un
grito) ‘scream (a scream)’.
4. An alternate possibility is explored in Lozano (2006), who assumes a TP-internal focus posi-
tion as in Belletti (2004) to which clause-final subjects move to check a [+Focus] feature.
5. For more on the details of this approach to focus fronting in Spanish, see López, 2009.
6. Haegeman and Zanuttini (1991) first proposed criterial features for negation. Notably, crite-
rial features are not lexical features of the type typically involved in the feature-checking
model used in Minimalist syntax (for further discussion, see Chomsky 1995). For further
discussion of criterial features and freezing, see Rizzi, 2010. For him, movement motivated
by such feature checking cannot be optional.
7. The Interface Hypothesis is hardly uncontroversial, and a number of studies have reported
findings that contradict its predictions (e.g. Cuza, 2013; Gómez Soler and Pascual y Cabo,
2018; Rothman, 2009).
8. The unergative predicates cantar and nadar, and the unaccusative predicates desaparecer and
nacer used in Hertel (2000, 2003) were not included in Lozano.
9. The optionality examined in the current article does not involve the dichotomy ‘expression
vs. non-expression’ of the type examined in Sorace and Filiaci (2006). Rather, it always
involves expression of a constituent in more than one possible syntactic position. For slightly
different understandings of optionality, see also Boyd, 2007; McGregor, 2013; Miyagawa,
2011; Müller, 2003.
10. These two native-speaker sub-groups were compared in Gupton (2017) and no statistically
significant differences were found between them. A third native speaker sub-group, Galician–
Spanish bilinguals, is also examined there. These 13 participants were not included in this
study because (1) their group results differed from the Catalan–Spanish bilingual group from
Valencia in a number of B-APT comparisons, and (2) because their profile did not represent
the sort of speaker that the intermediate-level learner groups would typically have contact
with.
11. VclS results from the direct object having appeared in the preceding context and being
expressed as an object clitic pronoun. In pilot testing, participants found that VclS was found
to be more natural than VOS.
12. For more on the details of this approach to focus fronting in Spanish, see López, 2009.
13. Both intermediate groups are labeled as such owing to the level of university Spanish class
that students were enrolled in (third-year) at the time of data collection.
14. To determine whether the regression model with the variables examined made significantly
better predictions than the model with the random effects alone, an ANOVA was run. The
results (X2 = 426.84, df = 10, p < 0.001) confirm a significant difference between the two.
15. Of the 1,280 total preference replies, choices of ‘Neither’ (n = 11) and ‘I don’t know’ (n = 5)
were extremely scarce.
16. Such cases are referred to as ‘unaccusative mismatches’ in Levin and Rapaport, 1989; Levin
and Rapaport Hovav, 1992.
17. A reviewer asks whether exposure to English could be to blame for optionality among the NS
group. However possible, this would not explain the fact that the same optionality also exists
among the native speakers from Valencia, who do not know English beyond what they have
been exposed to in the Spanish education system. See also note 10.
18. A reviewer rightly questions whether child-directed speech is a good indication of adult L1
speech. According to e.g. Genovese et al. (2020), infant-directed speech exhibits, among other
characteristics, syntactic simplification. Clark (2016: 25–26), however, disagrees with this
view, claiming that ‘child-directed speech is often singularly well tailored to its addressees,
Gupton and Sánchez Calderón 31
highly grammatical in form, and virtually free of errors.’ Even if the former position is on the
right track, we seek to determine what this sort of simplification might look like and what sort
of clues that might provide for the child.
19. Although the individuals in the CHILDES corpus are not the same as those who participated
in the AJT and APT, we feel that the child-directed language is comparable to what these
speakers would have received as children, at the very least for the participants who grew up
in Spain.
References
Alexiadou A and Anagnostopoulou E (1998) Parametrizing AGR: Word order, V-movement and
EPP-checking. Natural Language and Linguistic Theory 16: 491–539.
Avrutin S (2004) Optionality in child and aphasic speech. Lingue e Linguaggio 1: 65–96.
Baker M (1988) Incorporation: A theory of grammatical function changing. Chicago, IL: The
University of Chicago Press.
Bates D, Mächler M, Bolker B, and Walker S (2015) Fitting linear mixed-effects models using
lme4. Journal of Statistical Software 67: 1–48.
Belletti A (2004) Aspects of the low IP area. In: Rizzi L (ed.) The structure of CP and IP: The
cartography of syntactic structures: Volume 2. Oxford: Oxford University Press, pp. 16–51.
Birdsong D, Gertken LM, and Amengual M (2012) Bilingual language profile: An easy-to-use
instrument to assess bilingualism. Austin, TX: COERLL, University of Texas at Austin.
Available at: https://sites.la.utexas.edu/bilingual (accessed May 2021).
Bley-Vroman R (1983) The comparative fallacy in interlanguage studies: The case of systematic-
ity. Language Learning 33: 1–17.
Boyd JK (2007) Comparatively speaking: A psycholinguistic study of optionality in grammar.
Unpublished PhD thesis, University of California, San Diego, CA, USA.
Chomsky N (1995) The minimalist program. Cambridge, MA: MIT Press.
Clark E (2016) First language acquisition. 3rd edition. Cambridge: Cambridge University Press.
Cuza A (2013) Crosslinguistic influence at the syntax proper: Interrogative subject–verb inversion
in heritage Spanish. International Journal of Bilingualism 17: 71–96.
Davies M (2006) A frequency dictionary of Spanish. New York: Routledge.
de Prada Pérez A, and Pascual y Cabo D (2012) Interface heritage speech across proficiencies:
unaccusativity, focus, and subject position in Spanish. In: Geeslin K and Díaz-Campos M
(eds) Selected proceedings of the 14th Hispanic Linguistics Symposium. Somerville, MA:
Cascadilla Proceedings Project, pp. 308–18.
Domínguez L (2013) Understanding interfaces. Amsterdam: John Benjamins.
Domínguez L and Arche MJ (2008) Optionality in L2 grammars: the acquisition of SV/VS con-
trast in Spanish. In: Chan H, Jacob H, and Kapia E (eds) BUCLD 32 proceedings. Somerville,
MA: Cascadilla Press, pp. 96–107.
Domínguez L and Arche MJ (2014) Subject inversion in non-native Spanish. Lingua 145: 243–65.
Fukui N (1993) Parameters and optionality. Linguistic Inquiry 24: 399–420.
Geeslin K and Guijarro-Fuentes P (2006) A longitudinal study of copula choice: Following devel-
opment in variable structures. In: Sagarra N and Toribio AJ (eds) Selected proceedings of
the 9th Hispanic Linguistics Symposium. Somerville, MA: Cascadilla Proceedings Project,
pp. 144–56.
Gell-Mann M (1995) What is complexity? Remarks on simplicity and complexity by the Nobel
Prize-winning author of The Quark and the Jaguar. Complexity 1: 16–19.
Genovese G, Spinelli M, Romero Lauro L et al. (2020) Infant-directed speech as a simplified but
not simple register: A longitudinal study of lexical and syntactic features. Journal of Child
Language 47: 22–44.
32 Second Language Research 00(0)
Gómez Soler I, and Pascual y Cabo D (2018) On focus and weight in Spanish as a Heritage
Language. Revista Española de Lingüística Aplicada / Spanish Journal of Applied Linguistics
31: 437–66.
Gupton T (2017) Early minority language acquirers of Spanish exhibit focus-related interface
asymmetries: Word order alternation and optionality in Spanish–Catalan, Spanish–Galician,
and Spanish–English bilinguals. In: Lauchlan F and Parafita-Couto MC (eds) Bilingualism
and minority languages in Europe. Newcastle upon Tyne: Cambridge Scholars, pp. 214–41.
Gurzynski-Weiss L, Geeslin K, Daidone D et al. (2018) Examining multifaceted sources of input:
Variationist and usage-based approaches to understanding the L2 classroom. In: Tyler A,
Ortega L, Uno M, and Park H-I (eds) Usage-inspired L2 instruction. Amsterdam: John
Benjamins, pp. 291–311.
Haegeman L and Zanuttini R (1991) Negative heads and the NEG criterion. The Linguistic Review
8: 233–51.
Henry A (2005) Non-standard dialects and linguistic data. Lingua 115: 1599–1617.
Hertel T (2000) The second language acquisition of Spanish word order: Lexical and discourse
factors. Unpublished PhD thesis, Pennsylvania State University, State College, PA, USA.
Hertel T (2003) Lexical and discourse factors in the second language acquisition of Spanish word
order. Second Language Research 19: 273–304.
Hoot B (2016) Narrow presentational focus in Mexican Spanish: Experimental evidence. Probus
28: 335–65.
Jiménez-Fernández AL (2015) Towards a typology of focus: Subject position and microvariation
at the discourse–syntax interface. Ampersand 2: 49–60.
Jung D and McKinnon S (2019) A variationist analysis of overt subject placement in bilingual
Arizonan and monolingual Mexican Spanish. Unpublished paper presented at the 23rd
Hispanic Linguistics Symposium, El Paso, TX, USA.
Kitagawa Y and Fodor JD (2006) Prosodic influence on syntactic judgments. In: Fanselow G,
Féry C, Schlesewsky M, and Vogel R (eds) Gradience in grammar: Generative perspectives.
Oxford: Oxford University Press, pp. 336–58.
Kuroda S-Y (1972) The categorical and the thetic judgment: evidence from Japanese syntax.
Foundations of Language 9: 153–85.
la Bruna A and Rathod S (2005) Questionnaire length and fatigue effects. Bloomerce White Paper
5: 2–17.
Lardiere D (2009) Some thoughts on the contrastive analysis of features in second language acqui-
sition. Second Language Research 25: 173–227.
Leal T, Destruel E, and Hoot B (2018) The realization of information focus in monolingual and
bilingual native Spanish. Linguistic Approaches to Bilingualism 8: 217–51.
Leal-Méndez T and Slabakova R (2011) Pragmatic consequences of p-movement and focus front-
ing in L2 Spanish: Unraveling the syntax–discourse interface. In: Herschensohn J and Tanner
D (eds) Proceedings of the 11th Generative Approaches to Second Language Acquisition
Conference (GASLA 2011). Somerville, MA: Cascadilla Proceedings Project, pp. 63–75.
Levin B and Rapaport M (1989) An approach to unaccusative mismatches. Proceedings of NELS
19: 314–28.
Levin B and Rapaport Hovav M (1992) The lexical semantics of verbs of motion: The perspec-
tive from unaccusativity. In: Roca I (ed.) Thematic structure: Its role in grammar. Berlin:
DeGruyter, pp. 247–69.
Long A and Geeslin K (2020) Examining the role of instructor first language in classroom-based
oral input. In: Gurzynski-Weiss L (ed.) Cross-theoretical explorations of interlocutors and
their individual differences. Amsterdam: John Benjamins, pp. 161–87.
López L (2009) A derivational syntax for information structure. Oxford: Oxford Studies in
Theoretical Linguistics.
Gupton and Sánchez Calderón 33
Lozano C (2006) Focus and split-intransitivity: the acquisition of word order alternations in non-
native Spanish. Second Language Research 22: 145–87.
Mackenzie I (2006) Unaccusative verbs in Romance languages. New York: Palgrave.
MacWhinney B (2000) The CHILDES project: Tools for analyzing talk. 3rd edition. Hillsdale, NJ:
Erlbaum. Available at: http://childes.talkbank.org (accessed May 2021).
McGregor WB (2013) Optionality in grammar and language use. Linguistics 51: 1147–1204.
Miyagawa S (2011) Optionality. In: Boeckx C (ed.) The Oxford handbook of linguistic minimalism.
Oxford: Oxford University Press, pp. 354–376.
Montrul S (2002) Incomplete acquisition and attrition of Spanish tense/aspect distinctions in adult
bilinguals. Bilingualism: Language and Cognition 5: 39–68.
Müller G (2003) Optionality in optimality-theoretic syntax. In: Cheng L and Sybesma R (eds)
The second GLOT International state-of-the-article book. Berlin: Mouton de Gruyter,
pp. 289–321.
Muntendam AG (2013) On the nature of cross-linguistic transfer: A case study of Andean Spanish.
Bilingualism: Language and Cognition 16: 111–31.
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.
Oshita H (2001) The unaccusative trap in second language acquisition. Second Language
Acquisition 23: 279–304.
Pallotti G (2015) A simple view of linguistic complexity. Second Language Research 31: 117–34.
Papp S (2000) Stable and developmental optionality in native and non-native Hungarian gram-
mars. Second Language Research 16: 173–200.
Parafita Couto M, Mueller Gathercole V, and Stadthagen-González H (2015) Interface strate-
gies in monolingual and end-state L2 Spanish grammars are not that different. Frontiers in
Psychology 5: 1–17.
Perlmutter D (1978) Impersonal passives and the unaccusativity hypothesis. In: Proceedings of the
4th Annual Meeting of the Berkeley Linguistic Society, pp. 157–89. Berkeley, CA: Berkeley
Linguistic Society.
Pires A and Rothman J (2009) Disentangling sources of incomplete acquisition: An explanation
for competence divergence across heritage grammars. International Journal of Bilingualism
13: 211–38.
Pladevall Ballester E (2010) Child L2 development of syntactic and discourse properties of
Spanish subjects. Bilingualism: Language and Cognition 13: 185–216.
Polinsky M (1997) American Russian: Language loss meets language acquisition. In: Browne W
et al. (eds) Annual workshop on formal approaches to Slavic linguistics: The Cornell meeting
(1995). Ann Arbor, MI: Michigan Slavic Publishers, pp. 370–406.
Polinsky M (2006) Incomplete acquisition: American Russian. Journal of Slavic Linguistics 14:
191–262.
RStudio Team (2020) RStudio: Integrated Development for R. RStudio. Boston, MA: PBC.
Available at: http://www.rstudio.com (accessed May 2021).
Rizzi L (1997) On the fine structure of the left periphery. In: Belletti A and Rizzi L (eds) Elements
of grammar. Dordrecht: Kluwer, pp. 281–337.
Rizzi L (2010) On some properties of criterial freezing. In: Panagiotidis EP (ed) The complemen-
tizer phase: Subjects and operators. Oxford: Oxford University Press, pp. 17–32.
Rothman J (2007) Heritage speaker competence differences, language change and input type:
Inflected infinitives in heritage Brazilian Portuguese. International Journal of Bilingualism
11: 359–389.
Rothman J (2009) Pragmatic deficits with syntactic consequences: L2 pronominal subjects and the
syntax–pragmatics interface. Journal of Pragmatics 41: 951–73.
34 Second Language Research 00(0)
Schwartz B and Sprouse R (1996) L2 cognitive states and the Full Transfer/Full Access model.
Second Language Research 12: 40–72.
Shirai Y (1990) U-shaped behavior in L2 acquisition. In: Burmeister H and Rounds PL (eds)
Variability in second language acquisition: Proceedings of the tenth meeting of the Second
Language Research Forum: Volume 2. Eugene, OR: Department of Linguistics, University
of Oregon, pp. 685–700.
Silva-Corvalán C (1994) Language contact and change: Spanish in Los Angeles. Oxford: Clarendon.
Snyder W (2000) An experimental investigation of syntactic satiation effects. Linguistic Inquiry
31: 575–82.
Sorace A (1993a) Incomplete vs. divergent representations of unaccusativity in non-native gram-
mars of Italian. Second Language Research 9: 22–47.
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 unac-
cusative/unergative distinction. In: Eubank L, Selinker L, and Sharwood MS (eds) The cur-
rent state of interlanguage. Amsterdam: John Benjamins, pp. 153–75.
Sorace A (1999) Initial states, end-states, and residual optionality in L2 acquisition. In: Greenhill
A, Littlefield H, and Tano C (eds) Proceedings of the 23rd Boston University Conference on
Language Development. Somerville, MA: Cascadilla Press, pp. 666–74.
Sorace A (2000a) Differential effects of attrition in the L1 syntax of near-native L2 speakers. In:
Howell SC, Fish SA, and Keith-Lucas T (eds) Proceedings of the 24th Boston University
Conference on Language Development. Somerville: Cascadilla Press, pp. 719–25.
Sorace A (2000b) Gradients in auxiliary selection with intransitive verbs. Language 76: 859–90.
Sorace A (2011) Pinning down the concept of ‘interface’ in bilingualism. Linguistic Approaches
to Bilingualism 1: 1–33.
Sorace A and Filiaci F (2006) Anaphora resolution in near-native speakers of Italian. Second
Language Research 22: 339–68.
Sorace A and Serratrice L (2009). Internal and external interfaces in bilingual language develop-
ment: Beyond structural overlap. International Journal of Bilingualism 13: 195–210.
Sorace A and Shomura Y (2001) Lexical constraints on the acquisition of split intransitivity: evi-
dence from L2 Japanese. Studies in Second Language Acquisition 23: 247–78.
Titov E (2012) Encoding focus and contrast in Russian. In: Neeleman A and Vermeulen R (eds)
The syntactic expression of information structure. Berlin: DeGruyter, pp. 119–56.
Torrego E (1984) On inversion in Spanish and some of its effects. Linguistic Inquiry 15: 103–29.
VanPatten B (2004) Input processing in adult second language acquisition. In: VanPatten B
(ed) Processing instruction: Theory, research, and commentary. Mahwah, NJ: Lawrence
Erlbaum, pp. 5–31.
White L (1992) Long and short verb movement in second language acquisition. Canadian Journal
of Linguistics 37: 273–86.
White L (2011a) Second language acquisition at the interfaces. Lingua 121: 577–90.
White L (2011b) The Interface Hypothesis: How far does it extend? Linguistic Approaches to
Bilingualism 1: 108–10.
Yuan B (1997) How are lexical–semantic differences represented in L2 syntax? A study of the unac-
cusative–unergative distinction in L2 Chinese. In: Kellerman E, Weltens B, and Bongaens T
(eds) The generative study of second language acquisition. Mahwah, NJ: Lawrence Erlbaum,
pp. 215–38.
Yuan B (1999) Acquiring the unaccusative–unergative distinction in a second language: Evidence
from English-speaking learners of L2 Chinese. Linguistics 37: 275–96.
Zubizarreta ML (1998) Prosody, focus, and word order. Cambridge, MA: MIT press.
Gupton and Sánchez Calderón 35
Appendix 1
Bimodal acceptability judgment task stimuli.
A. Unaccusative predicate, thetic context
1. Estás en un café con un grupo de amigos. Todo el mundo pide bebidas y cafés.
Mientras la camarera prepara el pedido, tu colega Beatriz sale a la calle para
hacer una llamada. La camarera llega a la mesa y, viendo que hay un café de
sobra, te pregunta: ‘¿Qué ha pasado? ¿No había más gente?’ Tú le contestas:
a. Ha salido mi colega un momento.
b. Mi colega ha salido un momento.
2. Has salido con un grupo de amigos, y todos estáis en frente de una discoteca.
Ramón está diciendo que quiere ir a otro sitio y en ese momento pasan unos ami-
gos suyos. Mientras él los saluda tu amigo Fran entra en la discoteca sin espera-
ros. Vuelve Ramón y te pregunta: ‘¿Qué pasa? Pensaba que íbamos a ir a otro
bar.’ Tú le contestas:
a. Ha entrado Fran. No lo sé.
b. Fran ha entrado. No lo sé.
3. Vas a reunirte con unos amigos en el centro. Caminas a la parada de autobús con
tu amigo Jaime para coger el último bus al centro. De repente, él decide parar en
un bar para comprar tabaco. Tú sigues hasta la parada, que está cerca, y lees el
horario. Vuelves al bar y le dices:
a. No ha venido el autobús aún. ¡Vamos!
b. El autobús no ha venido aún. ¡Vamos!
4. Estás en una fiesta con tu amigo Juanjo. Él se va al baño y mientras está allí, entra
Pili, el alma de toda fiesta. Cuando vuelve Juanjo nota que todo el mundo está
más animado y te pregunta: ‘¿Qué ha sucedido aquí?’ Tú le contestas:
a. Ha llegado Pilar.
b. Pilar ha llegado.
5. Tu hermana está chateando por Internet con su amiga, Fátima, que está haciendo
un Erasmus en Bélgica. De repente escuchas gritos y risas. Vas a su cuarto y le
preguntas: ‘¿Qué pasa?’ Ella te contesta:
a. ¡Viene Fátima el próximo fin de semana!
b. ¡Fátima viene el próximo fin de semana!
7. Llegas al bar y hay un grupito de compañeras sentadas en una mesa. Parecen muy
emocionadas. Saludas a tu amiga Elena, que está con el grupo, y le preguntas:
‘¿Qué sucede?’
a. El novio de Susana le ha regalado un anillo de compromiso.
b. Le ha regalado el novio de Susana un anillo de compromiso.
8. Ves a tu amigo Enrique en el pasillo. Parece bastante contento, así que le pregun-
tas: ‘¿Qué cuentas?’ Te contesta:
a. Mira. Mi padre me ha prestado el coche para el fin de semana. ¿Te apuntas a
un viaje a Barcelona?
b. Mira. Me ha prestado mi padre el coche para el fin de semana. ¿Te apuntas a
un viaje a Barcelona?
9. Tu amiga Felisa te viene con un regalo en las manos. Le preguntas con curiosi-
dad: ‘¿Qué tal?’ Te responde:
a. Tu madre te ha enviado un regalo para tus cumpleaños.
b. Te ha enviado tu madre un regalo para tus cumpleaños.
10. Estás estudiando con los auriculares puestos cuando de repente tres amigos se te
acercan. Te quitas los auriculares y les preguntas: ‘¿Qué pasa?’ Te contesta Juan
Carlos:
a. Paco ha pedido una pizza. ¿Tienes hambre?
b. Ha pedido Paco una pizza. ¿Tienes hambre?
14. Estás saliendo de la estación de tren con un grupo de amigos que acaban de llegar.
Tu amigo, Jordi, te llama al móvil y te pregunta: ‘¿Quiénes han llegado?’ Le
contestas:
a. Han llegado Jaume, Estela y Patri.
b. Jaume, Estela y Patri han llegado.
15. Estás en una excursión a Mallorca con unos amigos. Juntos habéis alquilado unos
cuartos en un albergue. Durante el día un grupo decide ir a la playa, mientras tú
vas al centro con otro grupo. Cuando vuelves por la noche, parece que falta gente.
Tú le preguntas a Jaime, que fue con el otro grupo: ‘¿Quiénes no han vuelto?’ Él
te contesta:
a. No han vuelto Patricio y Toni. Van a acampar en la playa.
b. Patricio y Toni no han vuelto. Van a acampar en la playa.
Appendix 2
Bimodal appropriateness preference task stimuli.
A. Subject narrow focus
1. Estás en una fiesta cuando de repente llaman a la puerta. Abres la puerta y hay un
repartidor de un restaurante chino. Tú te das la vuelta y preguntas: ‘¿Quiénes
habéis pedido comida china?’ Tu amigo José María te dice:
a. La han pedido Margarita y Paula.
b. Margarita y Paula la han pedido.
2. Vas con tu amigo Marcos al pueblo de su padre. Cuando llegáis a la casa, hay un
tractor nuevo enfrente de la casa. Marcos le pregunta a su padre: ‘Papá, ¿quién ha
comprado ese tractor?’ Su padre le dice:
a. Lo compró Alberto. No está mal, ¿eh?
b. Alberto lo compró. No está mal, ¿eh?
3. Quedas con tu hermano para un café. Cuando le saludas, notas que lleva una
cadena de oro nueva. Sabiendo que él tiene más de una ‘amiga’, le preguntas:
‘¿Quién te ha dado el collar?’ Te contesta, mostrándotelo:
a. Me lo ha dado Cecilia. ¿Te gusta?
b. Cecilia me lo ha dado. ¿Te gusta?
4 Estás en la oficina trabajando cuando llega un repartidor de flores con una docena
de rosas para tu compañera Paula. Le preguntas: ‘¿Quién te ha enviado esas
rosas?’ Ella mira la tarjetita y te contesta, feliz:
a. Me las ha enviado Paco.
b. Paco me las ha enviado.
5. Quedas con los compañeros de universidad en el centro. Llega Samuel en un
Mercedes que nunca has visto antes. Sabiendo bien que no puede ser suyo, le
dices: ‘¡Qué cochazo! ¿Quién te lo ha prestado?’ Te contesta:
a. Me lo ha prestado mi cuñado. El mío se me ha estropeado.
b. Mi cuñado me lo ha prestado. El mío se me ha estropeado.
14. Miquel, uno de tus compañeros de piso, llega a casa. Entra en la cocina y después te
viene y te dice: ‘Por fin. Marc ha limpiado la cocina. ¿La has visto?’ Tú le corriges:
a. La ha limpiado SERGI.
b. SERGI la ha limpiado.
15. Vuelves del baño del bar y ves que han llegado cinco amigos de la facultad que
venían de una clase de flamenco. De repente el camarero os trae una jarra de
sangría a la mesa. Los amigos que acaban de llegar te dicen: ‘¿Nos has pedido la
sangría?’ Tú les corriges:
a. La ha pedido FERRÁN, pero ha salido a fumar.
b. FERRÁN la ha pedido, pero ha salido a fumar.
Appendix 3
Regression statistics.
Table 4. Mixed effect linear regression examining the interaction between transitive
predicates (TR), subject narrow-focus (SNF) contexts, and their influence on the acceptability
rating of VS word order.
Random effects:
Number (Intercept) 0.097911 0.31291
Item (Intercept) 0.005635 0.07506
Residual 0.814376 0.90243
Fixed effects:
(Intercept) 3.83564 0.09592 39.99
SNF −0.29394 0.08479 −3.467 < 0.001
TR 0.10909 0.08479 1.287 0.198
OrderVS −0.63939 0.07025 −9.101 < 0.001
Group1 (LG1) −0.47208 0.11747 −4.019 < 0.001
Group2 (LG2) −0.16986 0.11173 −1.52 0.129
Group3 (AGG) −0.0626 0.12252 −0.511 0.609
SNF:TR 0.04848 0.11991 0.404 0.686
SNF:OrderVS 0.58788 0.09935 5.917 < 0.001
TR:OrderVS −0.64545 0.09935 −6.497 < 0.001
SNF:TR:OrderVS 0.5 0.14051 3.559 < 0.001
Table 5. Mixed effect linear regression examining the interaction between transitive
predicates (TR), thetic sentence (TH) contexts, and their influence on the acceptability rating of
VS word order.
Random effects:
Number (Intercept) 0.097911 0.31291
Item (Intercept) 0.005635 0.07506
Residual 0.814376 0.90243
Fixed effects:
(Intercept) 3.5417 0.09592 36.925
TH 0.29394 0.08479 3.467 < 0.001
(Continued)
42 Second Language Research 00(0)
Table 5. (Continued)
Fixed effects:
TR 0.15758 0.08479 1.858 0.063
OrderVS −0.05152 0.07025 −0.733 0.464
Group1 (LG1) −0.47208 0.11747 −4.019 < 0.001
Group2 (LG2) −0.16986 0.11173 −1.52 0.129
Group3 (AGG) −0.0626 0.12252 −0.511 0.609
TH:TR −0.04848 0.11991 −0.404 0.686
TH:OrderVS −0.58788 0.09935 −5.917 < 0.001
TR:OrderVS −0.14545 0.09935 −1.464 0.143
TH:TR:OrderVS −0.5 0.14051 −3.559 < 0.001
Table 6. Mixed effect linear regression examining the interaction between unaccusative
predicates (UNACC), subject narrow-focus (SNF) contexts, and their influence on the
acceptability rating of VS word order.
Random effects:
Number (Intercept) 0.097911 0.31291
Item (Intercept) 0.005635 0.07506
Residual 0.814376 0.90243
Fixed effects:
(Intercept) 3.94473 0.09592 41.127
SNF −0.24545 0.08479 −2.895 0.004
UNACC −0.10909 0.08479 −1.287 0.198
OrderVS −1.28485 0.07025 −18.289 < 0.001
Group1 (LG1) −0.47208 0.11747 −4.019 < 0.001
Group2 (LG2) −0.16986 0.11173 −1.52 0.129
Group3 (AGG) −0.0626 0.12252 −0.511 0.609
SNF:UNACC −0.04848 0.11991 −0.404 0.686
SNF:OrderVS 1.08788 0.09935 10.95 < 0.001
UNACC:OrderVS 0.64545 0.09935 6.497 < 0.001
SNF:UNACC:OrderVS −0.5 0.14051 −3.559 < 0.001
Table 7. Mixed effect linear regression examining the interaction between unaccusative
predicates (UNACC), thetic sentence (TH) contexts, and their influence on the acceptability
rating of VS word order.
Table 8. Mixed effect linear regression examining the interaction between transitive
predicates (TR), subject narrow-focus (SNF) contexts, and their influence on optionality.
Random effects:
Number (Intercept) 0.059097 0.2431
Item (Intercept) 0.001523 0.03903
Residual 0.179014 0.4231
Fixed effects:
(Intercept) 0.51894 0.06414 8.091
SNF −0.07273 0.04116 −1.767 0.077
TR −0.1303 0.04116 −3.166 0.002
Group1 (LG1) −0.115 0.0891 −1.291 0.197
Group2 (LG2) −0.08611 0.08475 −1.016 0.31
Group3 (AGG) −0.10577 0.09293 −1.138 0.255
SNF:TR 0.2303 0.05821 3.956 < 0.001
Table 9. Mixed effect linear regression examining the interaction between transitive
predicates (TR), thetic sentence (TH) contexts, and their influence on optionality.
Random effects:
number (Intercept) 0.059097 0.2431
Item (Intercept) 0.001523 0.03903
Residual 0.179014 0.4231
Fixed effects:
(Intercept) 0.44621 0.06414 6.957
TH 0.07273 0.04116 1.767 0.077
TR 0.1 0.04116 2.43 0.015
Group1 (LG1) −0.115 0.0891 −1.291 0.197
Group2 (LG2) −0.08611 0.08475 −1.016 0.31
Group3(AGG) −0.10577 0.09293 −1.138 0.255
TH:TR −0.2303 0.05821 −3.956 < 0.001
Table 10. Mixed effect linear regression examining the interaction between subject narrow-
focus (SNF) and unaccusative predicates (UNACC) contexts, and their influence on optionality.
Random effects:
number (Intercept) 0.059097 0.2431
Item (Intercept) 0.001523 0.03903
Residual 0.179014 0.4231
Fixed effects:
(Intercept) 0.38864 0.06414 6.059
SNF 0.15758 0.04116 3.828 < 0.001
UNACC 0.1303 0.04116 3.166 0.002
Group1 (LG1) −0.115 0.0891 −1.291 0.197
Group2 (LG2) −0.08611 0.08475 −1.016 0.31
Group3(AGG) −0.10577 0.09293 −1.138 0.255
SNF:UNACC −0.2303 0.05821 −3.956 < 0.001
Table 11. Mixed effect linear regression examining the interaction between unaccusative
predicates (UNACC), thetic sentence (TH) contexts, and their influence on optionality.
Random effects:
Number (Intercept) 0.059097 0.2431
Item (Intercept) 0.001523 0.03903
Residual 0.179014 0.4231
Fixed effects:
(Intercept) 0.54621 0.06414 8.516
TH −0.15758 0.04116 −3.828 < 0.001
UNACC −0.1 0.04116 −2.43 0.015
Group1 (LG1) −0.115 0.0891 −1.291 0.197
Group2 (LG2) −0.08611 0.08475 −1.016 0.31
Group3 (AGG) −0.10577 0.09293 −1.138 0.255
TH:UNACC 0.2303 0.05821 3.956 < 0.001