CHILDES Portuguese Florianópolís Corpus

Leonor Scliar-Cabral



Participants: 1
Type of Study: longitudinal
Location: Brazil
Media type: audio
DOI: doi:10.21415/T5F597

Browsable transcripts

Download transcripts

Link to media folder

Citation information

Project Description

These data were collected, transcribed, and examined during Leonor Scliar-Cabral’s doctoral thesis work in 1974. The records were reviewed during the adaptation at the University of Florianópolís to CHILDES in 1993. They consist of one Brazilian Portuguese child’s 5530 utterances in broad phonetic transcription (including intonational patterns) collected in three sessions:

  1. Age 1;8.21, MLU 1.45 with 5 hours of recording for 1320 utterances.
  2. Age 1;10.20, MLU 2.22 with 6 hours of recording for 2245 utterances.
  3. Age 2;2.8, MLU 2.40 with 6 hours of recording for 1966 utterances.
  4. there are also 9 audio files after 2;2.8 that are not transcribed

The phonetic transcription was made by Scliar-Cabral, who has phonological training. Three speech therapy students checked the transcriptions. The intonational patterns that were transcribed in the original version were not adapted to CHILDES, but this information will be added in the future. Notes were taken for describing the situational context and/or any relevant information. The child’s mother and father were helpful in translating his lexical creations. No videotapes were made. The following measures were computed: MLU, utterance types, utterance tokens, type-token ratio, upper bound, number of imitations, percentage of imitations, lexicon size, number of nouns, verbs, adjectives, locatives, and pronouns, functors required, functors present, and percentage of functors present. Other classes computed were: copulas; modals; pivotal operators; discourse operators; tags; interjections; onomatopoeias; stereotypes. Pronouns were subclassified as possessive, demonstrative and interrogative. MLU was computed using the criteria suggested by Brown (1973) and Bowerman (1973). Utterances were coded only as imitations if they had an identical intonation to the previous utterance, as well as a complete lexical overlap. Omitting imitations and onomatopoeias, the MLU computed by CLAN matched that computed by hand.

The child was a Brazilian Portuguese native speaker, living in an upper-middle class suburb of São Paulo. He attended a part-day upper-middle-class nursery where Brazilian Portuguese was spoken. His parents were of upper class Jewish background: although the father was a fluent speaker of French, the parents spoke only Brazilian Portuguese with their child. The father was a professor of linguistics at the University of São Paulo; the mother was a psychologist. There were frequent and tight contacts with other relatives, like uncles, aunts, and grandparents who also used Brazilian Portuguese. As is common among Brazilian upper-middle-class families, there was a housekeeper and a nanny who used a lower-class sociolinguistic variety, sometimes from different regions of Brazil. Excluding these inputs, all the other adults used the Brazilian Portuguese variety of São Paulo city and its surroundings.

Thanks to the following members of Brazilian National Research Council (CNPq): Andréia Veríssimo Agostini, Cloves Cardoso Carrera, Flávia Maria do Nascimento, Iris Marjorie Boing Imhof, Mara Gonzales Bezerra, Mirella Nunes Giracca, Richard Fernando de Souza Costa, Santo Gabriel Vaccaro, TamírisTeodoro Vieira, and Tiago Kroich