Content uploaded by Anna Gavarró
Author content
All content in this area was uploaded by Anna Gavarró on Jun 23, 2015
Content may be subject to copyright.
CHAPTER NINE
THE BY-PHRASE AND THE ACQUISITION OF
MAJORCAN CATALAN PASSIVES
ANNA GAVARRÓ, XAVIER PARRAMON AND
LUCRÈCIA RALLO
The literature on language acquisition has shown over the years that passive
sentences (1a) are poorly understood by children when compared to active
sentences (1b): see Maratsos et al. (1985) for English, Verrips (1996) for Dutch,
Pierce (1992) for Spanish, Babyonyshev & Brun (2003) for Russian, Terzi &
Wexler (2002) for Greek, Sugisaki (1999) for Japanese, amongst many others.
Moreover, Maratsos et al. (1985) showed that 4-year-old children understand
comparatively well passive sentences that involve “actional” verbs (1), whereas
they fail in comprehension tasks that involve “nonactional” passives (2).
(1) a. The boy was kissed/kicked/held by the girl.
b. The girl kissed/kicked/held the boy.
(2) The boy was loved/seen/remembered by the girl.
To account for these phenomena, Fox and Grodzinsky (1998) proposed that the
difficulty with passives follows from difficulties in theta-transmission to the by
phrase. In this paper we evaluate their hypothesis in view of some new results from
Majorcan Catalan.
The paper proceeds as follows. In section 1 we provide the theoretical
background to our study. In section 2 we detail the features of passives in
CHAPTER NINE
Majorcan Catalan and present the experiment. In section 3 the results are provided,
and conclusions are drawn.1
1. Passives and the by phrase
In their 1998 paper Fox and Grodzinsky, in an attempt to replace Borer and
Wexler’s (1987) Maturation Hypothesis, formulated the Relaxed Maturation
Hypothesis which stated that
Children’s difficulty with passive constructions is related to properties of the by-
phrase. Specifically, we argue that children are in full control of all aspects of the
passive construction except for the ability to transmit the external theta-role of the
predicate to the by-phrase. (Fox & Grodzinsky 1998: 311)
Fox and Grodzinsky showed that, given the variety of argumental structures a
predicate can have, the by-phrase may be assigned various theta-roles, as shown in
(3) for English:
(3) a. Bill was killed by Mary. (AGENT)
b. The package was sent by John. (SOURCE)
c. The letter was received by Bill. (GOAL)
d. The professor was feared by all students. (EXP)
From this, they concluded that (i) theta-transmission is thematically unlimited,
and (ii) theta-transmission involves the transmission of a theta-role that is
otherwise realized by an implicit argument (as in a short passive, where the
external theta-role is implicit, e.g. The package was sent).
In contrast, in nominals by is limited to assigning an AFFECTOR theta-role
(meaning: an AGENT, POSSESSOR, CAUSE), as in (4a,b,c,), and assignment of other
theta-roles results in ill-formedness, as in (4d,e).
(4) a. the imprisonment of refugees by the government
b. the destruction of the city by lightning
c. a book/article by John
d. the receipt of the package (*by John) (GOAL)
1 The authors are grateful to Maria del Mar Vanrell for the Majorcan Catalan version of the
experiment, and for recording it, and to Cèlia Rosselló Galmés for running the experiment
with a subset of the children. We are indebted to the teachers and children of the Col.legi
Pius XII in Palma de Mallorca, for their kind willingness to take part in this project, and to
project HUM-2006-13295-C02-01 for financial support.
2
THE BY-PHRASE IN THE ACQUISITION OF MAJORCAN CATALAN
PASSIVES
e. the fear of Harry (*by John) (EXP)
Therefore, there appears to be no theta-transmission with nominals. By acts as
independent theta-role assigner, and assigns an AFFECTOR theta-role. If, by
hypothesis, theta-transmission fails in early development the complement of by in a
passive will fail to get a theta-role. There is the possibility that by assigns to its
complement position the role that it assigns in nominals, namely an AFFECTOR
theta-role. This non-adult-like behaviour would result in seemingly adult-like
behaviour with (i) truncated passives, (ii) non-truncated passives of actional verbs,
in which the by phrase receives an AFFECTOR theta-role (w/o theta-transmission);
non-adult-like behaviour would be predicted with nontruncated passives of
nonactional verbs, in which assignment of an AFFECTOR theta-role to the
complement of by would be at odds with the argumental structure of the verb.
Fox and Grodzinsky (1998) ran a truth-value judgement task with thirteen
children, native speakers of English, and observed three response patterns. Group 1
children (including two children, aged 4;1 and 4;9) performed like adults. Group 2
children (including eight children, age range 3;6–5;5) displayed poor performance
on the nonactional nontruncated passives; when the by-phrase is eliminated from
the nonactional passives, these children’s performance is perfect – and their
behaviour is as predicted by Fox and Grodzinsky’s hypothesis. Group 3 children
(three children, age range 4;3–4;9) had poor performance on nonactional
nontruncated passives and also nonactional truncated passives – contrary to the
predictions of Fox & Grodzisnky’s hypothesis.
Fox and Grodzinsky’s (1998) results were not replicated in other studies of
passive in English (see for example Gordon and Chafetz 1990, Hirsch and Wexler
2006a). They were also subject to a number of objections by the fact that children
were shown to ignore the AFFECTOR theta-role assigned by by in adult English
(Hirsch & Wexler 2006b): in a study of spontaneous production no AGENT by
phrases was ever produced by children, none either in child-directed speech;
according to their experimental results, children systematically misinterpreted the
story by M as the story about M.
In this study we explore the predictions of Fox and Grodzinsky’s hypothesis for
Majorcan Catalan, a variety of Catalan in which passives are analogous to English
passives in many respects. In particular, we test the prediction (ii) of Fox and
Grodzinsky’s hypothesis, namely, that children appear to perform like adults with
non-truncated passives of actional verbs, in which the by phrase receives an
AFFECTOR theta-role (w/o theta-transmission). We will not consider short passives
here, but refer the reader to Parramon (2009) for results on short passives in
Catalan.
3
CHAPTER NINE
2. An experiment on Majorcan Catalan
In Majorcan Catalan, just as in other varieties of Catalan, passive sentences
involve the auxiliary ésser ‘be’ followed by the verb in the participle. The
preposition that introduces the external argument is de ‘of’. The sentences tested in
our experiment were thus active and long passive sentences, exemplified in (5).
(5) a. Es nin gran pega an es nin petit.
the child older hits P the child little
‘The older child hits the younger child.’
b. Es nin petit és abraçat des padrí.
The child little is hugged P-the grandfather
‘The younger child is hugged by the grandfather.’
Interestingly, the preposition de is the same that is found in nominals to
introduce an AFFECTOR theta-role as shown in (6).
(6) a. es llibre de Tolstoi (AGENT)
the book of Tolstoy ‘the book by Tolstoy’
b. sa destrucció de s’huracà (CAUSE)
the destruction of the hurricane ‘the destruction by the
hurricane’
c. es llibre de Na Maria (POSS)
de book of D Maria ‘Mary’s book’
This situation sharply contrasts with that in the other of varieties of Catalan,
which present the preposition per, as in (7):
(7) El nen petit és abraçat per l’avi.
the child little is hugged by the grandfather
With nominals, per introduces a GOAL or a PATH, never an AFFECTOR:
(8) a. un llibre per la Maria (GOAL)
a book for D Maria ‘a book for Maria’
b. un camí per la selva (PATH)
4
THE BY-PHRASE IN THE ACQUISITION OF MAJORCAN CATALAN
PASSIVES
a way through the jungle ‘a path through the jungle’
c. *la novel.la per Tolstoi (AGENT)
the novel by Tolstoy
The similarity between the Majorcan Catalan preposition de and English by
motivated our study, although it should be noted that de can also introduce a
THEME theta-role with nominals (it is in fact very unrestricted with respect to the
interpretation of its complement), as in (9), unlike English by.2
(9) es llibre de lingüística ’the book on linguistics’ (THEME)
Our experiment, designed under the auspices of COST [European Cooperation
in Science and Technology] action A33 (Armon-Lotem et al. submitted), is a
picture selection task. Testing was preceded by a warm up in which the child was
introduced to the characters in the pictures and familiarised with the actions
depicted. The child had to select the correct picture out of a set of four; in all cases,
the picture matching the test sentence appeared amongst a picture of the reverse
action, a picture of the same action carried out by different characters and a picture
without any action (to control for the understanding of the verb). All sentences
were semantically reversible and contained actional verbs with animate characters.
Each verb was presented twice: once in an active sentence and once in a passive
sentence. The verbs tested appear in (10).
(10) Verbs tested: sempentejar ‘push’, examinar ‘examine’, besar
‘kiss’, alimentar ‘feed’, rentar ‘wash’, rapinyar ‘scratch’,
pentinar ‘comb’, dibuixar ‘draw’, embrutar ‘dirty’, acariciar
‘stroke’, abraçar ‘hug’, tapar ‘cover’, dur ‘carry’, estirar ‘pull’,
afaitar ‘shave’, fotografiar ‘photograph’, pintar ‘paint’,
pessigollejar ‘tickle’, pegar ‘hit’, mossegar ‘bite’, perseguir
‘chase’, eixugar ‘dry’
2 It is standard to analyse de as a weak preposition, inserted only for the licensing of the
complement, as Case marker, and therefore to consider it semantically vacuous (even in
partitive contexts, see Martí 2010 for Catalan). It could be argued, then, that Majorcan
Catalan de and English by are dissimilar in that respect. However, by has also been
considered a semantically empty Case marker in the English passive construction (see
Bresnan 1982, Zubizarreta 1985).
5
CHAPTER NINE
Each child was tested on 22 active sentences and 22 long passives. Sentences
were presented in random order. A total number of 51 children, aged 4, 5 and 6, all
native speakers of Majorcan Catalan, were tested in Palma de Mallorca; their
details appear in (11).
(11) Subjects
Age group # Mean age Age range
4-year-olds 5 4;6 4;2–4;11
5-year-olds 17 5;6 5;1–5;11
6-year-olds 29 6;7 6;1–6;11
total 51 5;6 4;2–6;11
3. Results: Passive delay in Majorcan Catalan
The results we obtained for 4, 5 and 6-year-olds appear in (12). Given that the
child had four pictures to choose from, the results with actives are clearly
indicative of above-chance performance; long passives, in contrast, were poorly
understood (and most of the time the child’s error consisted in choosing the reverse
reading of the sentence; for reasons of space, we do not quantify different error
types here).
(12) Percentage of correct answers, actives vs. passives
4-year- olds 5-year-olds 6-year-olds
active 63.6% 79.7% 90.4%
long passive 26.5% 25.4% 58%
The statistical analysis shows that performance is a function of the sentence
type: active vs. passive (OR=8.857); by age, there are no statistically significant
differences between 4- and 5-year olds, but there are differences between these two
groups and 6-year-olds (for 4-year olds OR=0.181, for 5-year olds OR=0.271).
There are statistical differences in the interpretation of actives between all age
groups, and in the interpretation of (long) passive sentences there are no
differences between 4- and 5-year olds, but there are statistically significant
differences between 4-year olds and 6-year olds (OR=0.235) and between 5-year
olds and 6-year olds (OR=0.219).
6
THE BY-PHRASE IN THE ACQUISITION OF MAJORCAN CATALAN
PASSIVES
Let us now compare these results for Majorcan Catalan with those available for
peninsular Catalan, obtained from the same experiment (see Gavarró and Parramon
2011 for details).
(13) Percentage of correct answers, actives vs. passives
3-year- olds 4-year-olds 5-year-olds 6-year-olds
active 83.9% 94% 97.5% 98.8 %
long passive 20% 37.8% 31% 95.4%
Again, actives are well understood, and long passives are poorly understood
(here only as far as age 5). The statistical analysis indicates statistical differences
between the two groups (Majorcan Catalan vs. peninsular Catalan), with the
peninsular Catalan-speaking children performing better on both actives
(OR=5.776) and passives (OR=1.902). There is no explanation at this point for this
difference, which may be due to a combination of factors, for example: the way in
which the test was administered, which was less effective in Majorcan Catalan,
and/or the size of the sample, which was also smaller in Majorcan Catalan.
How do our result fit in with the predictions of Fox and Grodzinsky’s
hypothesis? In Majorcan Catalan the prediction was that long passives of actional
verbs would be well understood by children because the preposition de, just like by
in English, would assign an AFFECTOR theta-role to its complement, in spite of lack
of theta-transmission in children. Our results do not fulfil this prediction: Majorcan
Catalan children fail with long passives, just like their peninsular Catalan peers.
The characteristics of the preposition do not seem to affect performance, certainly
not in the direction expected, given that peninsular Catalan children outperformed
Majorcan Catalan children. It is not the case that long passives of actional verbs are
interpreted in an adult-like manner by children when the preposition introduces an
AFFECTOR theta-role. The Relaxed Maturation Hypothesis put forward by Fox and
Grodzinsky (1998) is not supported by our findings on Majorcan Catalan.
References
Armon-Lotem, Sharon et al. Submitted. A large scale crosslinguistic investigation
of the acquisition of passive.
Babyonyshev, Maria and Dina Brun. 2003. The role of aspect in the acquisition of
passive constructions in Russian. Paper presented at Generative Approaches
to Language Acquisition (GALA) 2003, Utrecht.
7
CHAPTER NINE
Borer, Hagit, and Kenneth Wexler. 1987. The maturation of syntax. In Parameter
Setting, ed. T. Roeper and E. Williams, 123-172. Dordrecht: Reidel.
Bresnan, Joan 1982. Control and complementation. Linguistic Inquiry 13(3): 343–
434.
Fox, Danny and Yosef Grodzinsky. 1998. Children’s passive: a view from the by-
phrase. Linguistic Inquiry 29: 311–332.
Gavarró, Anna and Xavier Parramon. 2011. The acquisition of Catalan passives
and the adjectival interpretation. Paper presented at LSRL, Ottawa.
Gordon, Peter, and Chafetz, Jill. 1990. Verb-based versus class-based accounts of
actionality effects in children's comprehension of passives. Cognition, 36(3),
227-254.
Hirsch, Christopher and Kenneth Wexler. 2006a. Children’s passives and their
resulting interpretation. In The Proceedings of the Inaugural Conference on
Generative Approaches to Language Acquisition –North America, ed. K. U.
Deen, J. Nomura, B. Schultz and B. Schwartz, University of Connecticut
Occasional Papers in Linguistics, 125–136.
Hirsch, Christopher, and Kenneth Wexler. 2006b. By the way, children don’t know
by. In Proceedings of the 30th annual Boston University Conference on
Language Development, ed. D. Bamman, T. Magnitskaia, and C. Zaller, 249-
261. Somerville, MA: Cascadilla Press.
Maratsos, M. P., D. E. C. Fox, J. A. Becker, and M. A. Chalkley. 1985. Semantic
restrictions on children's passives. Cognition, 19(2), 167-191.
Martí Girbau, Núria. 2010. The Syntax of Partitives. Doctoral disseration,
Universitat Autònoma de Barcelona.
Parramon, Xavier. 2009. The acquisition of actional passives in Catalan. MA
dissertation, Universitat Autònoma de Barcelona.
Pierce, Amy. 1992. The acquisition of passives in Spanish and the question of A-
chain maturation. Language Acquisition 2: 55-81.
Sugisaki, Koji. 1999. Japanese passives in acquisition. University of Connecticut
Working Papers in Linguistics 10: 145–156.
Terzi, Arhonto and Kenneth Wexler. 2002. A-Chains and s-homophones in
children’s grammar: Evidence from Greek passives. Proceedings of NELS 32:
519–537.
Verrips, Maike. 1996. Potatoes must peel: The acquisition of Dutch passive. PhD
dissertation, Universiteit van Amsterdam.
Zubizarreta, M. Luisa. 1985. Morphophonology and morphosyntax: Romance
causatives. Linguistic Inquiry 16(2): 247–289.
8
