Acrisio Pires and his co-author Jason Rotham (University of Iowa) presented their paper at the 38th Annual Linguistic Symposium on Romance Languages held April 4-6 at the University of Illionois.
Abstract

Given that various influential proposals about syntactic change have been based on acquisition and learnability (e.g. Lightfoot 1999, Clark&Roberts 1993, Roberts 2001, 2007 and refs therein), tests of children’s knowledge of properties that are undergoing change in different dialects are relevant (but still lacking), and can provide new evidence to refine such proposals. With this in mind, the acquisition of Brazilian Portuguese (BP) is especially relevant, since many aspects to the formal/standard grammar were argued to have been at least partially eliminated from colloquial dialects, differently from European Portuguese (e.g. 3P accusative clitics, enclitic pronouns, null referential subject pro, see Azevedo 1989, Galves 2001, Kato&Roberts 1993, Naro 1981, Salles 2005 and refs therein). However, adult BP speakers that are tested for grammaticality judgments show evidence of full knowledge of certain grammatical properties that are argued to have been lost. Such is the case with inflected infinitives, infinitival forms that are overtly inflected for person/number, independently of tense (1b) (2a): adult BP speakers give evidence of knowledge of the grammatical properties of these forms (e.g. Koike 1983, Perini 1974, Quicoli 1996, Rothman & Iverson 2007). However, such properties were argued to have been partially or entirely eliminated from the grammar of colloquial BP (see Botelho-Pereira&Roncaratti 1993, Lightfoot 1991, Painter 1991, Pires 2002). In addition, corpus studies of colloquial dialects of BP indicate that alternative structures are used to the exclusion of inflected infinitives. This paper focuses on questions such as: (i) To which extent is full/native knowledge of certain grammatical properties determined only by native/child language acquisition? (ii) How can knowledge of grammar that results from native acquisition (Chomsky 1986, 2005) be teased apart from grammatical knowledge that results from late learning, in case the empirical naturalistic data (or the absence thereof) and grammaticality judgments (Chomsky 1964 and much later work) lead to opposite conclusions regarding the knowledge attained by adult speakers? We address these questions by investigating whether the morphosyntax and semantics of inflected infinitives are actually acquired as part of native BP grammatical systems or are instead acquired (or learned) later.
This study tested two hypotheses: (i) whether inflected infinitives are indeed no longer acquired as part of native BP (colloquial) dialects and (ii) whether these forms are actually learned late by BP speakers (e.g. as an artifact of learning standard BP in school). We tested a cross section of BP child/teenage groups (ages range from 6-15) from upper and lower income classes to determine if and when BP children come to acquire inflected infinitives. We present below the results for the upper income class: the only one that showed full knowledge of inflected infinitives.
We conducted two experiments, a morphological recognition task (MRT) and a truth value judgment task (TVJ task, e.g. Crain & Thornton 1998). Test materials were presented in two different (but similar) versions, according to the age of the subjects. Children up to age 9 were tested with a picture matching/choice task (e.g. McDaniel, McKee & Cairns 1996) that involved the participation of a puppet, and subjects age 10 and older were presented with a written version of the same task. The MRT involved 12 stories (3 stories testing 3PL inflected infinitives, 3 stories testing 1PL inflected infinitives, 6 control stories with non-inflected infinitives or present tense). Each story had two test sentences, and correct/incorrect test sentences varied randomly across stories. An example of the MRT with inflected infinitives appears in (1). After each story the child was asked to correct the test sentence if they thought it was incorrect. For the younger children, the test sentence was uttered by the puppet, Tigger, given that children were told that Tigger was still learning Portuguese, and sometimes needed help when he made mistakes.
In the TVJ task the subjects were told stories such as (2). They were then presented with the test sentence (uttered by the puppet as the end of the story) and asked to pick the picture/description that corresponded to that story end (or to say what happened in case they did not accept either picture/description). There were 24 stories, with a total of 12 test sentences (6 each for inflected infinitives 3PL and 1PL forms) and 12 fillers/distractors (including non-inflected infinitives). Test sentences targeted three syntactic/semantic properties of non-inflected infinitives (vs. non-inflected infinitives): (i) non-obligatory control (2a,c), (ii) strict reading under ellipsis and (iii) the possibility of split antecedents.
The experimental data clearly show (confirmed statistically across both tests) that children under the age of ten (and from the youngest school age) do not have grammatical knowledge of the distinctive syntax and semantics of inflected infinitives (for both tests these groups did not differentiate between inflected and uninflected infinitives, for every counterbalanced property, i.e. p>.05 for all relevant categories). In addition, children and teenagers above age ten develop such knowledge incrementally: we show that only children/teenagers above age 10 show incremental evidence of knowledge of inflected infinitives: by the age of 13, subjects performed as expected of a native grammar with inflected infinitives (every counterbalanced comparison was p<.001 for all relevant categories). The late learning (arguably post critical period) of inflected infinitives confirms proposals of Portuguese dialectal variation and BP diachronic change. However, BP educated speakers still learn inflected infinitives as teenagers, which can explain the grammatical knowledge shown by adults, and gives relevance to the potential role of standardization and literacy in the acquisition of grammatical knowledge that is restricted to a standard, non-native dialect but has been lost from native vernacular dialects.

(1) (Story testing 3PL inflected infinitive) A Margarida e a Minie gostam de esportes. Elas semprem fazem natação, mas hoje o esporte delas é outro. Por que a escolha delas mudou?
Daisy and Minnie like sports. They always swim, but today they picked a different sport. Why did their choice change?
Test sentences (uttered by puppet, for younger subjects):
a. *É muito difícil elas nadar todo dia. (incorrect non-inflected infinitive form)
It is very difficult they swim every day.
b. Agora é mais importante a Daisy e a Minnie correrem mais vezes. (3PL infl. inf)
Now it is more important for Daisy and Minnie to run more times.
(2) Story: O Mickey estava em casa, com o Pato Donald e a Margarida. O carro estava muito sujo e precisava de uma limpeza.
Mickey was at home, with Donald Duck and Daisy. The car was very dirty and needed cleaning.
a. Test sentence: O Mickey ficou satisfeito de lavarem o carro.
Mickey was happy of wash-INF-3PL the car.
‘Mickey was happy that they washed the car.’
b. Wrong picture: shows Mickey washing the car alone (no one else is on the picture).
c. Correct picture: shows Donald and Daisy washing the car (Mickey is absent).

This entry was posted on Thursday, April 10th, 2008 at 1:54 pm and is filed under Conferences, Syntax.