Third person accusative pronouns in Brazilian Portuguese in a Socio- and Psycholinguistic perspective

Description

The study of linguistic variation, which traditionally relied mostly on spontaneous production data, has recently been focusing on other psycholinguistic methodologies, such as elicited productions, reaction and reading times, and priming effects (cf. Chevrot, Drager, & Foulkes, 2018). For instance, Kaschak & Glenberg (2004) and Squires (2014) found faster reading times when an individual was more familiar with a specific variant; moreover, those times can become faster as individuals have more contact with previously unfamiliar variants, what Clopper (2014) considers a result of an expansion of mental representations in order to accommodate them.
In this regard, schooling may be important for children to have contact with different variants, especially those that are not acquired during early childhood. For example, in Brazilian Portuguese (BP), third person accusative pronouns can be realized either as strong pronouns ele(s)/ela(s), normally acquired by children, or as clitic pronouns o(s)/a(s), acquired via schooling (Kato, Cyrino, & Corrêa, 2009). Production data has shown that clitic pronouns are seldom used orally (Corrêa, 1991; Zanellato et al., 2021), contrasting with their higher use in the written modality (Othero & Schwanke, 2018; Othero et al., 2018; Zanellato, 2017). As argued by Lira (2021), this difference is connected to a higher level of monitoring in written productions, which might be linked to the perception of formality attributed to clitic pronouns in BP (cf. Schwenter et al., 2022). In fact, for Labov ([1972] 2008), the linguistic repertoire must be controlled by a sociolinguistic monitor, which involves more conscious cognitive processes that lead to the selection of the most appropriate variant depending on the social context and the speakers’ social profile.
Therefore, two productions tasks which allowed different degrees of speech monitoring (spontaneous and elicited) and a sentence-matching task (whose rational says that the reaction time to decide if two given sentences are the same or not is slower when they are agrammatical or unfamiliar) were conducted with 24 schooled BP native speakers in order to establish whether the level of monitoring affects the type of pronoun produced, whether there are differences at the processing level when participants read sentences containing each pronoun, and whether
any social factors influence their performance. Results showed that, as expected, strong pronouns are prioritized in spontaneous speech (91.8%) and clitic pronouns are prioritized in more monitored situations (88.3%). No difference in reaction time was found during the sentence-matching task (strong pronouns: 1650ms; clitic pronouns: 1638ms; p = 0,604), which might be due to their own morphological nature, a by-product caused by the higher use of clitic pronouns during the elicited production task (suggesting either a priming effect of such pronouns or inhibition of the strong pronouns), or that they are equally accessible (and their use depends on specific social contexts). Gender and education were the relevant social factors detected, in the sense that women and more educated speakers tended to prioritize the more formal variant (clitic pronouns), as anticipated by Labov ([1972] 2008).

Period	26 Jun 2024
Event title	6th conference on Variation and Language Processing
Event type	Conference
Location	Vigo, SpainShow on map
Degree of Recognition	International