Erros gráficos na transcrição de ditados de aprendizes brasileiros de ele

Abstract

Portuguese and Spanish are two languages very close to each other as they belong to the family of the Romance languages. However, their differences deserve to be highlighted as they are manifested in the Spanish writing activity, through various types of errors, among them, the spelling ones. In this sense, from the analysis of transcripts of dictations made in SFL (in High School and Undergraduate) classes, this work aims to (i) analyze the spelling errors resulting from poor understanding of Spanish phonemes and; (ii) to detect the occurrence of spelling errors resulting from the phonological transference from Portuguese to Spanish, as well as from the morphological transference from the one to another; plus the unfamiliarity or he forgetfulness of the orthographic rules of the new language. The object of research is the spelling of fifteen words extracted from the text ¿Sabías que los nidos de las abejas se llaman colmenas? and the relation of the phonemes /β/, /b/, /x/, /λ/, /θ/, /ɾ/, /r/ and /n/ with the graphemes found in these words. The corpus is composed by 38 dictations transcribed by Brazilian students, among adolescents and adults. The text brings 103 words in a basic vocabulary and comes from a comic book written in Spanish. Regarding theory, in this work we have the contribution of the Contrastive Analysis (LADO, 1957), the Analysis of Errors (CORDER, 1967; ELLIS, 1994) and the analysis of Interlanguage (SELINKER, 1972); as well as theory regarding Phonetics and Phonology, Spelling (in Language Teaching) and phonological-orthographic peculiarities between Spanish and Brazilian Portuguese (MASIP, 2001, 2010, FROST, 1992; QUILIS, 2010); besides the relation between the phonological and the orthographic systems themselves (CAGLIARI, 2002; LLOP, 1999); among other authors. In terms of methodology, we have the contribution of Variationist Sociolinguistics (LABOV, 1972, TARALLO, 2007, GUY & ZILLES, 2007) in the quantitative analysis of the social variables involved in the research (age, sex, school level and type of previous educational institution). Among some of the results obtained, the analyzes indicated, for example, (1) that the phoneme-grapheme relation between / β / and the letters <b> e <v> in abejas, árboles, arriba, trabajan, viven - was in the great majority successful, with a percentage of graphical errors between 0 and 23%; (2) other phonemes - / x /, / ɾ /, / r /, / λ / and / n / - in these same words, presented spelling errors and/or in greater quantity and/or a greater variation of misspelled words as in "abelhas", "aberras", "aberas", "árbores", "ariba", "trabaran", "trabalhoØ", "viben." One of the relations between phoneme and the expected spelling in Spanish was between / λ / e <ll> in the word llaman. In a total of eleven variations of misspellings of this word, four are related to the abovementioned relation - eg lhaman, llhaman, ilhaman, chaman - while the others have spelling errors in the second syllable, in the relation between / n / and <n >, having <n> the grammatical function of representing the simple present of the third person in the plural form in Spanish, - present in the words llaman, trabajan, utilizan, ven and viven - while in Portuguese – in trabalham [tɾa´baʎãw] – such an ending is represented by the grapheme <m> and phoneme /w/. Regarding the relationship between the spelling errors and the social data of the research participants, the results showed that the variable age - divided into three age groups - showed a higher error rate in the first range (18 to 20 years), with a total of 198 errors out of 359 (198/359), approximately 55.15% errors. The lowest index was located in the age group between 21 and 25 years old, with an error rate of 72/359 or 20.05%. The gender and level of education variables presented small differences when compared to the number of errors found in the samples, and were not considered relevant as extralinguistic factors. However, the school origin variable (whether private or public) prior to the IFRN caught our attention to the result that 55.43% of the errors were from students who came from public education, against 44.57% from private education.