Upload
voquynh
View
218
Download
0
Embed Size (px)
Citation preview
The Perception of Discontinuous Dependencies by
18-Month-Olds: On the Process of Acquiring Verbal Passives
João Claudio de Lima Júnior and Letícia Maria Sicuro Corrêa
1. Introduction
Studies concerned with the acquisition of passives have traditionally aimed
to answer whether children at a certain age have, or not, the ability to derive
(comprehend/produce) verbal passives. In this respect, two major currents can be
considered: one that regards the acquisition of passives as a late achievement,
dependent on maturation (Borer & Wexler, 1987; Snyder & Hyams, 2015, among
many others); and another one that considers it an early achievement (Crain,
Thornton & Murasugi, 1987/2009; O’Brien, Grolla & Lillo-Martin, 2006; Bencini
& Valian, 2008). Little attention has been devoted, however, to the procedures
whereby children acquire these structures.
The present paper does not engage with any of the cited currents and changes
the focus of attention to early processing abilities that are assumed to be required
in order for the grammatical knowledge pertaining to passives to be achieved.
These abilities involve detecting patterns that can be relevant for the identification
of the discontinuous dependency that characterize the form of these structures in
languages such as Portuguese.
In Portuguese, verbal passives display a specific morphophonological
pattern, the Aux-foi (eventive be in its simple past form in Portuguese) plus the
participial form of the main verb (see (1)). The present work investigates whether
young children acquiring Brazilian Portuguese (who do not necessarily produce
more than one-word utterances) can track this non-adjacent dependency. In order
to do so, a preferential attention test (analogous to the Headturn Preference
Procedure (cf. Kemler-Nelson et al. (1995)), inspired in Santelmann & Jusczyk
(1998), was designed. It contrasts the mean listening time of stories containing
the regular pattern of verbal passives, foi_V-do (see (1)), with an ungrammatical
pattern involving the auxiliary foi and the Portuguese imperfective past morpheme
–va, forming the complex foi_V-va (see (2)). The main hypothesis is that children
can track such a dependency.
* João Claudio de Lima Júnior, Pontifical Catholic University of Rio de Janeiro (PUC-
Rio), [email protected]. Research supported by the CNPq grant 152661/2016-6).
Letícia Maria Sicuro Corrêa, Pontifical Catholic University of Rio de Janeiro (PUC-Rio),
[email protected]. Research supported by the grant CNPq-PQ 309499/2016-0.
© 2018 João Claudio de Lima Júnior and Letícia Maria Sicuro Corrêa. Proceedings of the 42nd annual Boston University Conference on Language Development, ed. Anne B. Bertolini and Maxwell J. Kaplan, 465-478. Somerville, MA: Cascadilla Press.
(1) O cachorro foi amarrado.
The_Masc._Sing. dog beeventive_Past_Sing. tie_Part._Masc._Sing.
‘the dog was tied’
(2) *O cachorro foi amarrava.
The_Masc._Sing. dog beeventive_Past_Sing. tie_Impefective Past.
‘*The dog was ties’
The paper is organized as follows: in the next section, an experimental study
investigating children’s ability to track the morphophonological dependency of
verbal passives is reported. In section 3, a brief characterization of the acquisition
of passives mainly concerned with languages such as Portuguese is proposed in
the light of a procedural view of language acquisition (Corrêa, 2009; 2014).
Section 4 brings our final remarks.
2. Experimental Study
Discontinuous dependencies between functional items have been a matter of
great interest in studies concerned with the interface language-cognition
(Golinkoff, Hirsh-Pasek & Schweisguth, 1998; Santelmann & Jusczyk, 1998;
Tincoff, Santelmann & Jusczyk, 2000; Höhle et al, 2006). In the classic study of
Santelmann & Jusczyk (1998) children’s ability to track the morphosyntactic non-
-adjacent dependency between the verb be and the gerund morpheme –ing was
tested. These functional items are codependent and together form progressive
tenses in English.
Santelmann & Jusczyk (1998) conducted 5 experiments1 on the HPP
(Headturn Preference Procedure) (cf. Kemler-Nelson et al., 1995) aiming at
investigating the effect of limitation in the tracking of the cited dependency.
In Santelmann & Jusczyk (1998), infants between 15 and 18 months were
exposed to stories containing grammatical sentences as in (3). These sentences
were contrasted with ungrammatical sentences, as in (4). Notice that in (4) the
auxiliary verb be was replaced by the modal can.
(3) Gradma is singing.
(4) *Grandma can singing.
Their prediction was that infants would listen differently to the two types of
stories (see (5-6)) if they could track the morphosyntactic dependency at stake.
The results show that 18-month-olds tend to listen longer to stories with the
grammatical complex Aux-be+V-ing than to stories with the ungrammatical
complex. The same results have not been attested with 15-month-olds.
1 The details of each experiment are omitted due to the fact that they are irrelevant to the
purposes of the present paper.
466
(5) Natural version (grammatical sentences)
At the bakery, everybody is baking bread. One person is mixing the flour
and water together. Someone else is adding salt and yeast. In the next
room, a big machine is kneading the dough. Another is shaping the
loaves for the oven. The whole place is starting to smell great!
(6) Non-natural version (ungrammatical sentences)
At the bakery, everybody can baking bread. One person can mixing the
flour and water together. Someone else can adding salt and yeast. In the
next room, a big machine can kneading the dough. Another can shaping
the loaves for the oven. The whole place can starting to smell great!
[see Santelmann & Jusczyk (1998, p. 111)]
The same experimental procedure is entertained here regarding passive
sentences. The working hypothesis is that 18-month-olds can track the
discontinuous dependency of verbal passives involving the auxiliary ser and the
participial morpheme –do. Adaptations had to be made, though, concerning the
form of the contrast. In English, the focus of the ungrammaticality of the modified
version was the auxiliary. In Portuguese, the participial morpheme (see (7)) was
replaced by the morpheme -va of the indicative imperfective past (see (8)).
(7) O chão foi molhado.
The_Masc._Sing. Floor Be_eventive_Past_Sing. wet_part._masc._sing.
‘The floor was wettened’
(8) *O chão foi molhava.
‘*The floor was wets’
The independent variables were Age (younger; older infants) and Type of
complex (normal; modified). Mean listening time was the dependent variable. It
is predicted that if children can track the morphosyntactic dependency of passives,
a significant difference between normal and modified conditions will be obtained.
Although the results reported in Santelmann & Jusczyk (1998) have shown that
infants, on average, listen longer to stories with grammatical sentences, no
prediction has been made concerning the direction of the results.
2.1. Method
Participants. 27 infants participated in the test, 5 of which had their data
excluded from analysis for not completing the task. The remaining participants
(13 girls and 9 boys) were equally divided into two age groups (13-15; mean age
14;8 months VS 17-20; mean age 18;14 months).
Material. 8 experimental stories were created.2 Each of them had 6
experimental sentences. Two versions of these stories were created: one
2 The test is not directly compromised with meaning, but with the (morpho)phonological
patterns. The decision to make real stories instead of disconnected sentences is to make the
material sound natural.
467
containing normal sentences with the complex Aux-foi+V-do (see (9)) and one
containing exactly the same elements, but with a specific modification in the
complex, the participial morpheme –do was substituted by the morpheme –va,
forming the complex Aux-foi+V-va (see (10)).
(9) Normal version
Quando entrou em casa, a formiguinha falou: [O chão do meu quarto
foi molhado]. Na cozinha, [o garfo foi babado pelo meu amiguinho]. Em
cima da pia, [o queijo foi aguado] e [o melão foi cortado]. Não dá para
acreditar! [O fogão foi quebrado]. No final, [meu carro foi roubado da
garagem]. Que será que aconteceu aqui?
Translation: As soon as the ant got into its house, it said: the floor of my
room was wetted. In the kitchen, the fork was slobbered by my friend.
On the sink, the cheese was watered and the melon was cut. I can't
believe it! The stove was broken! In the end, my car was stolen from the
garage. What might have happened here?
(10) Modified version
Quando entrou em casa, a formiguinha falou: [O chão do meu quarto
foi molhava]. Na cozinha, [o garfo foi babava2 pelo meu amiguinho].
Em cima da pia, [o queijo foi aguava] e [o melão foi cortava]. Não dá
para acreditar! [O fogão foi quebrava]. No final, [meu carro foi roubava
da garagem]. Que será que aconteceu aqui?
On average, each experimental story lasted 27 seconds (never less than 26
and never more than 28). Stories were all controlled to have the same number of
words (57 words). The experimental sentences, in particular, also had the same
number of words. Only dissyllable verbs of the first conjugation (ending in–ar)
were used in the complexes aux+part to keep the number of syllables controlled.
The distance between the auxiliary foi and the termination (–do or –va) was
always the same throughout the trials of the test; that is, two syllables.
An extra story, shorter than the experimental ones (17 seconds), was also
created. No structure containing aux+part was used in it. This story opened the
activity to give the participants some time to get used to the experimental
situation.
A female native speaker of Portuguese recorded all stories by means of the
Praat software. She was instructed beforehand to read the experimental sentences
without emphasizing any of the areas of interest, particularly the modified ones.
After some practice, she was asked to read the stories in a soft manner, as if she
were really telling a story to a baby. Once it was recorded, no alteration in the
material was necessary.3
Device. The experiments were run in a baby-lab soundproof cabin with a
chair for the caregiver to sit with his/her baby on the lap. The chair is opposite a
3 Native speakers were asked to judge the reading. They found both types of stories equally
amiable, fluent, and natural.
468
SONY TV set. A SONY camera was hidden between the parts of a drape. This
drape has the same color of the wall of the cabin and is situated right behind the
TV set. The camera is strategically positioned opposite the chair so as to capture
the images of the baby face.
In the control cabin, outside the soundproof cabin, the experimenter
controlled the test by means of a Macbook computer in which the Habit 2000
software was installed. This software is usually used in preferential listening tests
to control semi-automatically the presentation of stimuli and time the attention
children give to these stimuli. A Pioneer sound amplifier was used to remotely
control the volume of the stimuli. A Phillips TV set received the images of the
child from the inside of the baby-lab cabin. The caregiver and the experimenter
also used two mp3 players with earphones to avoid any sort of interference during
the test.
Procedure. All infants were tested individually. Toys were available and
strategically positioned on a rubber mat right in front of the door to the baby lab
cabin. Children were invited to play on this mat with the caregivers and
experimenter. The aim was to make children at ease in the lab.
The caregivers were previously oriented about how to proceed inside the
cabin. For instance, they were asked to calm down their children before the
beginning of the test; they were told to use the mp3 players and to put the
earphones on; they were also informed not to give any further instruction to their
children after the test began.
As soon as the door of the experiment cabin was closed, the test started. From
the outside of the cabin, the experimenter, who also had earphones, could observe
the babies’ behavior inside the experiment cabin.
The test started with the exhibition of a humanized (anthropomorphic
characteristics) light bulb on the TV screen in the format of an animated gif. The
gif blinked and its exhibition was synchronized with the emission of a sharp and
intermittent sound. The image and sound were used as the attention getter
throughout the test; that is, a mechanism with the aim of attracting children’s
attention to the experimental stimulus to be subsequently presented.
As soon as the child reacted to the exhibition of the attention getter, fixating
his/her eyes on the TV screen, the first stimulus was released. While the auditory
stimulus was being played, another gif, now the face of a girl moving her mouth
as if she were speaking, was simultaneously exhibited. The same image was used
in all trials.
After listening to the shorter story, created to make infants familiarized with
the test, the child had to listen to 8 different stories (4 of the normal version e 4 of
the modified version). The stories were semi-randomized in a way that babies
never listened to two normal and two modified stories in sequence.
Child’s attention to the TV monitor was controlled from the outside of the cabin.
The attention time is here characterized as listening time. The Habit software
registered the listening times automatically.
When the baby deviated his/her attention from the TV screen for more than
2 seconds, the attention getter was immediately released. As soon as the child’s
469
attention was retaken, a new stimulus was let off. This procedure successively
occurred until all the stories were released. The whole time inside the baby-lab
cabin was very short, usually 5 to 8 minutes. At any child’s sign of clear
discomfort, the procedure was immediately halted.
2.2. Results
The maximum mean listening time registered was 16.8 seconds (in the
normal condition). The total mean listening time of the test considering both
conditions (normal and modified) was 9.42 seconds. All the means of the test are
presented in figure 1.
Age Condition Mean listening Standard StandardGroups time (sec.) Deviation Error
13-15
18-20
Normal 8.57 1.74 0.52
Modified
Normal
Modified
7.85 2 0.62
0.5
0.923.06
1.668.4
12.85
Figure 1: all means obtained in the test
A 2 (Age) X 2 (Type of Story) ANOVA was conducted. Type of story was a
repeated measure. Two significant main effects were obtained Age and Type of
story.
For Age F(1,20) = 9.31 p<.006 SS=64.49 MSe=6.93 the mean listening time
of older children was 10.63 sec. and of younger children was 8.21 sec.. An
increase of listening time is thus obtained as infants get older.
For Type of story F(1,20) = 26,9 p<.00005 SS=73.43 MSe=2.73) the mean
listening time of normal stories was 10.71 sec. and of modified stories was 8.13
sec.. Such an effect shows that infants, in general, prefer the normal stories to the
modified ones. An interaction between the independent variables Age*Type of
story F(1,20) = 14 p<.001 SS=38.32 MSe=2.73 was also obtained.
At last, a pairwise-comparison reveals that longer listening times were
obtained in the normal condition ([Normal]vs[Modified] t(10)=5,97 p<.0001) in
the older group only (see graph 1). It thus suggests that, as predicted, 18-month-
470
olds are able to detect the morphophonological pattern of passives while 15-
month-olds are not.
Mea
n lis
teni
ng t
ime
(sec
.)
16
14
12
10
6
0
2
4
8
n.s.
*
13-15 17-20Normal Modified Normal Modified
8,57 7,85 12,85 8,4
Graph 1: Mean listening time in the two groups tested and standard deviation bars;
the asterisk indicate the significant effect
2.3. Discussion
It has been demonstrated here that Brazilian babies identify the complex
Aux_foi+V-do as American babies identify the complex Aux_is+V-ing in English
(cf. Santelmann e Jusczyk, 1998). These results replicate previous ones
concerning the ability of 18-month-olds to detect discontinuous morphosyntactic
dependencies (Cf. Golinkoff, Hirsh-Pasek & Schweisguth, 1998; Santelmann &
Jusczyk, 1998; Höhle et al, 2006). As in these previous studies, children younger
than 18 months were not able to track such a dependency here.
It is relevant to mention that all previous investigations conducted in
Portuguese, although few and incipient, indicate that the occurrence of passives
is very low in spoken language, particularly in the caregiver-children interaction
(Cf. Perotino, 1995; Pesirani, 2009; Estrela, 2016). Lots of studies have been
suggesting that non-adjacent dependencies would be acquired via statistical
learning mechanisms (see references in Sandoval & Gómez (2013)). However,
the apparent low frequency of passives in Portuguese did not seem to be
problematic for the 18-month-olds.
471
The novelty of these results is that, for the first time, the ability of very young
children to detect the morphophonological pattern that characterizes verbal
passives has been shown experimentally. The next section inserts the
interpretation of these results in the context of a procedural theory of language
acquisition.
3. A procedural theory of language acquisition and the case of passives
The procedural view of language acquisition in Corrêa (2009; 2014) aims to
reconcile the early identification of morphophonological patterns in speech
processing with the notion of interfaces between an internal language and
processing systems - as expressed in the Minimalist Program of generative
linguistics (Cf. Chomsky, 1995; subsequent work). Under this view, the idea of
innately guided learning (Cf. Jusczyk & Bertoncini, 1988) and the phonological
bootstrapping hypothesis (Cf. Morgan & Demuth, 1996) are considered to be
compatible with an initial state of language that constrains the identification of
the information that can be grammatically relevant in the input data4.
It has been assumed that the grammatical information for the identification
of a particular grammar is available at the interfaces between (internal) language
and the so-called performance systems, particularly as morphophonological
patterns. The sort of information that the linguistic theory assumes to be necessary
for the functioning of a universal computational system - formal features - is
expressed in these patterns. Once formal features are represented in the lexicon,
they provide information for the computational system to build syntactic
structures. These features are initially represented in an underspecified manner
and will be progressively specified in the course of the development based on
referential/semantic information.
Results interpreted in the light of the phonological bootstrapping hypothesis
have shown that infants are able to detect morphophonological distributional
4 It is important to clarify that in the first place the procedural view of language acquisition
addresses the issue of how the computational system of human languages (Cf. Chomsky,
1995; subsequent work) is bootstrapped (put to work) (Cf. Corrêa, 2009). The approach
can be considered a first-merge theory in the sense that it speculates about the time the first
merge operation - as stated by Chomsky - occurs. In a more concrete way, suppose that an
infant is innately guided to detect regularities in the linguistic input, as assumed by Jusczyk
and Bertoncini (1988). In this sense, an infant may detect that a certain sound
corresponding to a determiner ‘the’ is usually uttered before sounds corresponding to
words uttered to name different entities in the world, nouns. An infant acquiring English,
for instance, notices that the former, determiners, always precedes the latter, nouns. Based
on these two pieces of information represented in the lexicon - (i) the distinction between
words pertaining to closed (determiners) and open classes (nouns) and (ii) order - the
computational system may proceed to derive its first syntactic phrase and send it to the
logical form in order to seek for an interpretation (cf. Corrêa, 2009; 2014). Corrêa claims
that the representation of this rudimentary information would suffice for the computational
system to be initialized.
472
regularities in the speech sound since very early - far before the first year of age.
As previously pointed out, children are assumed to be guided (Cf. Jusczyk &
Bertoncini, 1988) to take these patterns as grammatically relevant. Therefore, in
the particular acquisition of passives, the Aux-foi+V-do might be the sort of
information that will lead young children to acquire the structures at stake.
In a language such as Portuguese, the complex Aux-foi+V-do seems to
function as an index for the acquisition of passives. This index may be represented
in the lexicon as a rudimentary formal feature (call it Π5). At first, the detection
of this complex solely means that the dependency between the discontinuous
elements was identified. Π would be thus a formal feature that promotes a
syntactic binding between the auxiliary foi and the participle -do. In minimalist
terms, these functional items might be merged together. Having this information
represented in the lexicon, a merge operation between the auxiliary foi and the
participial morpheme may be driven. The result of this rudimentary computation
may be formally represented as a hierarchical structure6 (see (11)).
(11) [foiP foi[Π] [partP –do[μΠ] [VP V]]]
It is well known that the process of acquiring a language is based only upon
positive evidence. In this regard, a particular morphosyntactic index (Aux_foi+V-
do) from which an abstract formal relation is extracted (see (11)) will certainly
help infants in their task of mapping syntactic information into the semantic
interface. How children would then proceed to acquire the semantic information
of Portuguese verbal passives via syntax?
Once children have a feature represented in their lexicon informing that this
merge is possible, their subsequent task would be to differentiate the semantic
properties of the Aux-component of the complex at the semantic inteface. This
differentiation needs to be done in relation to analogous complexes - the
complexes of adjectival passives (see (12-13)), respectively resultative and stative
passives.
(12) O cachorro ficou amarrado.
The_Masc._Sing. dog become_Past_Sing. tie_Part._Masc._Sing.
‘The dog got tied’
(13) O cachorro está amarrado.
The_Masc._Sing. dog bestative_Past_Sing. tie_Part._Masc._Sing.
‘The dog is tied’
5 The name Π is just a formalization, a way to account for the distinctive nature of this
dependency in relation to other morphosyntactic dependencies. 6 It must be clarified that, in this formal representation, the maximum projection foiP is
used instead of auxP. This choice is justifiable due to the fact that it is possible that the
child identifies dependencies, at least initially, on the basis of an item-to-item relation (Cf.
Tincoff, Santelmann & Jusczyk, 2000). In this respect, the infant could not initially
associate foi with the flexional category of the verb ser (beeventive).
473
Each passive predicate is thus semantically distinct in Portuguese. According
to Embick (2004), the semantics of these complexes may be formally
characterized by means of two semantic features: [agentivity] and [eventivity]. In
this respect, a resultative predicate, as in (12), is positively marked only for
eventivity [−agentive; +eventive]; stative predicates, as in (13), are not positively
marked for any of the two features [−agentive; −eventive]. An eventive predicate,
as the ones in verbal passives, is positively marked for both features [+agentive;
+ eventive]. The identification of these semantic properties will lead children to
syntactically specify the properties of Π.
Notice that the distinctive semantic property of eventive passives is
agentivity. Roughly, the presence of an auxiliary ser (beeventive) and the participial
form of the main verb must indicate to the parser/computational system the need
for an agentive layer to be operational while deriving eventive passives. When it
occurs, it is possible to affirm that Π has been syntactically specified, meaning
that a voice/passive feature (cf. Collins, 2005; Lima Júnior & Augusto, 2015) has
been eventually represented in the lexicon and is available for syntactic
computation.
An experiment run with Brazilian children showed that the most fundamental
distinction regarding verbal (eventive) and adjectival passives (resultative and
stative) has been made by children as young as 3 years of age (Lima Júnior,
Augusto & Corrêa, 2016). The results of a truth-value judgment task revealed that
3 year-olds master resultative and stative passives as adults do. When
verbal/eventive passives are concerned, although these children hesitated to
answer as adults do, they demonstrated to have an understanding of the eventive
nature of verbal passives.
While distinguishing passive predicates and even after having the passive
feature represented in the lexicon, children still need to be able to recognize which
verbs are, or are not, passivizable. This may change from language to language.
Notice that the verb have and its counterpart in Portuguese ter do not admit
passivization (see (14)). As for walk, in English, the verb is perfectly passivizable
while in Portuguese caminhar is not (see (15)). In this regard, overgeneralizations
have been experimentally attested in English over the age of 4 (Cf. Pinker,
Lebeaux & Frost, 1987). Aspectual and semantic features, such as affectedness,
also seem to help British and Brazilian children to have a better performance on
verbal passives (see Messenger et al., 2012; Lima Júnior & Augusto, 2014).
(14) a. *The dog was had by the boy.
b. * O cachorro foi tido pelo menino.
(15) a. The dog was was walked by the boy.
b. *O cachorro foi caminhado pelo menino.
It is important to notice that passive sentences are pragmatically marked in
Portuguese as they are in English. Additionally, these sentences have been
reported to be more costly than adjectival and active sentences even for adults
474
(Griffin & Bock, 2000; Ferreira, 2003; Lima Júnior & Corrêa, 2015). Reversibility
of theta roles and pragmatic factors also have been shown to add to the burden of
processing passives (Ferreira, 2003; Lima Júnior & Corrêa, 2016; Lima Júnior,
Corrêa & Augusto, 2016). Therefore, even if the syntactic information was
properly represented and is at work, the use of passives at certain circumstances
may be unfavored.
At last, language external factors such as inhibitory control can contribute to
the development of the ability to cope with these structures (cf. Lima Júnior,
Augusto & Corrêa, 2016). It is expected then that children take longer to find out
when it is more advantageous to use a passive sentence instead of alternative
structures, such as an active or a topical sentence, for instance (cf. Lima Júnior,
Corrêa & Augusto, 2016). Thus, having difficulties to parse passives at a certain
age does not necessarily mean that the competence to parse it is not existent. Being
able to cope with these apparently cognitive difficulties is also necessary for an
adult-like performance with passives to be attested experimentally.
To summarize, a procedure view of the acquisition of passives in languages
such as Portuguese would include:
(a) Detecting the phonetic and distributional pattern corresponding to
AUX+Part at the phonetic interface;
(b) differentiating the Aux-foi+V-do complex from other analogous
complexes (adjectival passives, for instance) at the semantic interface
(see Crawford, 2012; Koring, Sangers & Wexler, 2015).
(c) identifying semantic restrictions associated with main verbs in passive
structures (Cf. Pinker, Lebeaux & Frost, 1987; Lima Júnior & Augusto,
2014);
dealing with the cost of computing/processing passives in different
conditions (Bencini & Valian, 2008; Crain, Thornton & Murasugi, 2009;
Lima Júnior, Corrêa & Augusto; 2016).
4. Final Remarks
In this study the non-adjacent dependency of passives has been investigated
in the light of a procedural theory of acquisition. A preferential attention test (a
similar testing ground to the Headturn Preference Procedure) was designed. In
the task, the complex involving the auxiliary of passives and the participial
morpheme was contrasted with an ungrammatical complex, in which the
participle was replaced by the imperfective past morpheme –va. The results show
that children at the age of 18 months tend to listen longer to stories involving the
grammatical complex that provides morphophonological information concerning
the voice/passive feature in Portuguese. The same was not observed with 15-
month-olds, though. This difference has been interpreted here as an evidence
that children at the age of 18 have the syntactic information required to proceed
to the acquisition of the semantic specifications of verbal passives.
475
References
Bencini, Giulia & Valian, Virginia (2008) Abstract sentence representations in 3-year-olds:
Evidence from language production and comprehension. Journal of Memory and
Language, 59, 97–113.
Borer, Hagit & Wexler, Kenneth (1987) The Maturation of Syntax. In Roeper, Tom &
Williams, Edwin (eds.) Parameter setting (pp. 123-72), Dordrecht: Reidel.
Chomsky, Noam (1995) The Minimalist Program. Cambridge, Mass.: MIT Press.
Collins, Chris (2005) A Smuggling Approach to Raising in English. Linguistic Inquiry, 36
(2), 289-298.
Corrêa, Letícia M.S. (2009) Bootstrapping language acquisition from a minimalist
standpoint: On the identification of phi-features in Brazilian Portuguese. In: Pires, A.
& Rothman, J. (Eds.). Minimalist Inquiries into Child and Adult Language
Acquisition: Case Studies across Portuguese. p. 35-621. Berlin: Mouton de Gruyter.
Corrêa, Letícia M.S. (2014) Interface Information and Computational Cost: An Integrated
Procedural Approach to Language Acquisition with Some Implications for SLI. In
Costa, João; Fiéis, Alexandra; Freitas, Maria J.; Lobo, Maria; Santos, Ana L. (Eds.)
New Directions in the Acquisition of Romance Languages. Cambridge publishers,
CA, 2014.
Crain, Stephen; Thornton, Rosalind, Murasugi, Keiko (1987) Capturing the evasive
passive. Paper presented at the 12th Boston University Conference on Language
Development.
Crain, Stephen; Thornton, Rosalind, Murasugi, Keiko. (2009) Capturing the evasive
passive. Language Acquisition, 16(2), 123-133.
Crawford, Joan. (2012) Developmental Perspectives on the Acquisition of the Passive.
Doctoral dissertation, University of Connecticut, Storrs, CT.
Embick, David. (2004) On the Structure of Resultative Participles in English. Linguistic
Inquiry, 35 (3), 355-392.
Estrela, Antónia P. (2016) A estrutura passiva num corpus de aquisição. Fórum
Linguístico, Florianópolis, 13 (1), 1009-1021. DOI: 10.5007/41039.
Ferreira, F. (1994) Choice of passive voice is affected by verb type and animacy. Journal
of Memory and Language, 33, 715-736.
Ferreira, Fernanda. (2003) The misinterpretation of noncanonical sentences. Cognitive
Psychology, 47, 164-203.
Golinkoff, Roberta M., Hirsh-Pasek, Kathy, Schweisguth, Melissa A. (2000) A reappraisal
of young children's knowledge of grammatical morphemes. In: Jürgen Weissenborn
& Barbara Höhle (Eds). Approaches to bootstrapping: phonological, lexical, syntactic
and neurophysiological aspects of early language acquisition.
Amsterdam/Philadelphia: John Benjamins, 2000.
Griffin, Zenzi M & Bock, Kathryn. (2000) What the eyes say about speaking.
Psychological Science, 11, 274–279
Höhle, Barbara; Schmitz, Michaela; Santelmann, Lynn M.; Weissenborn, Jürgen (2006)
The recognition of discontinuous verbal dependencies by German 19-month-olds:
Evidence for lexical and structural influences on children’s early processing
capacities. Language Learning and Development (2 (4), pp. 277– 300.
Jusczyk, Peter W. & Bertoncini, Josiane (1988) Viewing the development of speech
perception as an innately guided learning process. Language and Speech, 31, 217-238.
Kemler-Nelson, Deborah G.; Jusczyk, Peter W., Mandel, Denise R., Myers, James; Turk,
Alice E. (1995) The Head-turn Preference Procedure for testing auditory perception.
Infant Behavior and Development, 18, 111–116.
476
Koring, Loes; Sangers, Nina; Wexler, Kenneth. Who is the agent when it is left implicit?
Paper presented at the 12th Generative approaches to language acquisition (GALA)
congress, 2015.
Lima Júnior, João C. & Augusto, Marina R.A. (2014) The relevance of aspectual and
semantic features for the comprehension of short passive sentences involving agentive
and non-agentive verbs. In In Costa, J.; Fiéis, A.; Freitas, M.J.; Lobo, M.; Santos, A.L.
(Eds.) New Directions in the Acquisition of Romance Languages. Cambridge
publishers, CA, 2014.
Lima Júnior, João C. & Augusto, Marina R.A. (2015) Is 'smuggling' really necessary? The
most recent analyses of passive sentences reconsidered in terms of phasehood and
cyclic movement. ReVEL, special edition, n. 9, 2015. [www.revel.inf.br].
Lima Júnior, João C. & Augusto, Marina R.A. (2017) PassiveP and the distinction between
eventive, resultative, and stative passives. Diadorim, special edition, pp.: 27-52, Rio
de Janeiro.
Lima Júnior, João C.; Augusto, Marina R.A.; Corrêa, Letícia M.S. (2016) Discrimination
of passive predicates by Brazilian Portuguese-speaking children. Paper presented at
the Romance Turn 8 (RT8), which was held in Universitat Autònoma de Barcelona,
Bellaterra, October.
Lima Júnior, João C. & Corrêa, Letícia M.S. (2015) A natureza do custo computacional na
compreensão de passivas: um estudo experimental com adultos. Letras de Hoje, Porto
Alegre, vol. 50, n.1, p. 91-101, jan-mar
Lima Júnior, João C. & Corrêa, Letícia M.S. (2016) Sentence-image mapping and the
relevance of temporal/aspectual information in the comprehension of passive
sentences by children. In Marianne Carvalho Bezerra Cavalcante et alli (Eds.).
Questões em aquisição da linguagem e psicolinguística: fonologia e prosódia, língua
de sinais, sintaxe e processamento (pp. p. 575-589). 1ed. João Pessoa: Mídia Gráfica
e Editora ltda., v. 1.
Lima Júnior, João C.; Corrêa, Letícia M.S.; Augusto, Marina R.A. (2016) Favorable
processing conditions in the production of passive sentences by Brazilian Portuguese
speaking children Paper presented at the Romance Turn VIII (RTVIII), Universitat
Autònoma de Barcelona (UAB), Bellaterra - Barcelona, Spain.
Messenger, Katherine; Branigan, Holy P & McLean, Janet F & Sorace, Antonella. (2012)
Are children’s early passives semantically constrained? Evidence from syntactic
priming. Journal of Memory and Language, 66, 568-587.
Morgan, James L & Demuth, Katherine. (1996) Signal to Syntax: an overview. In Morgan,
JL & Demuth, K. (eds.). Signal to Syntax: Bootstrapping from speech to grammar in
early acquisition (pp. 1-22), Mahwah, New Jersey: Lawrence Erlbaum Associates.
O´Brien, Karen, Grolla, Elaine & Lillo-Martin, Diane. (2006) Long passives are
understood by young children. BUCLD, 30 Proceedings, 441-451.
Perotino, Silvana. (1995) Mecanismos de indeterminação do agente: o fenômeno da
apassivação na aquisição da linguagem. MA Dissertation, Universidade de Campinas
(UNICAMP).
Pesirani, Mariana AA. (2009) Aquisição de 'ser' e 'estar' em Português Brasileiro, com
especial atenção as estruturas passivas. Paper presented at the 17 SIICUSP.
Pinker, Steven , Lebeaux, David & Frost, Loren A. (1987) Productivity and constraints in
the acquisition of the passive. Cognition, 26, 195-267.
Snyder, William & Hyams, Nina (2015) Minimality effect in children's passives. In Elisa
Di Domenico, Cornelia Hamann and Simona Matteini (eds.) Structures, Strategies and
Beyond: Studies in honour of Adriana Belletti, 343–368.
477
Sandoval, Michelle; Gomez, Rebecca L. (2013) The development of nonadjacent
dependency learning in natural and artificial languages. Cognitive Science.
Santelmann, Lynn M.; Jusczyk, Peter W. (1998) Sensitivity to discontinuous dependencies
in language learners: Evidence for limitations in processing. Cognition, 69, 105-134.
Tincoff, Ruth; Santelmann, Lynn M., Jusczyk, Peter W. (2000) Auxiliary verb learning
and 18-month-olds' acquisition of morphological relationships. In: Howell SC, Fish
SA, Keith-Lucas T (eds.) Proceedings of the 24th Annual Boston University
Conference on Language Development (pp. 726–737). Somerville, MA: Cascadilla
Press.
478
Proceedings of the 42nd annualBoston University Conference on Language Development
edited by Anne B. Bertolini and Maxwell J. Kaplan
Cascadilla Press Somerville, MA 2018
Copyright information
Proceedings of the 42nd annual Boston University Conference on Language Development© 2018 Cascadilla Press. All rights reserved
Copyright notices are located at the bottom of the first page of each paper.Reprints for course packs can be authorized by Cascadilla Press.
ISSN 1080-692XISBN 978-1-57473-086-9 (2 volume set, paperback)ISBN 978-1-57473-186-6 (2 volume set, library binding)
Ordering information
To order a copy of the proceedings or to place a standing order, contact:
Cascadilla Press, P.O. Box 440355, Somerville, MA 02144, USAphone: 1-617-776-2370, [email protected], www.cascadilla.com