Title: The acquisition of writing in a bilingual setting: mutual interferences of Portuguese and German 

 

 

Authors:

Beatriz Dias; Biologist; Edith Menke, Speech Therapist, Clara Loureiro, Psychologist; Martin Lauterbach, Neurologist and Psychiatrist,

LEL (Language Research Laboratory), UNIC (Neurological Clinical Investigation Unit), Faculty of Medicine of Lisbon

 

E-mail: labling@fm.ul.pt

 

Postal address: Laboratório de Estudos de Linguagem, Centro de Estudos Egas Moniz, Hospital Santa Maria, 1649-028 Lisboa

 

 

Introduction:

The ease of children to acquire two, or more, languages is strongly related with the social setting in which the languages are learned. The best language proficiency is obtained in a setting called “one person – one language”.

Children that attend a bilingual school learn simultaneously two different phoneme-grapheme correspondence-systems (PGC-S). The PGC-S of each language does not give unequivocal rules how to write all words of this language. In order to solve these ambiguities the children must learn context rules and building up a word form memory. Beside those intra-language ambiguities children that attend a bilingual school have to deal with the fact that in the other language certain phonemes are written differently.  This additional demand can lead up to inter-language ambiguities and consequently to specific interference errors.

The aim of this study is to investigate the interference of the two PGC-S and its course along the progression of education. For this purpose a longitudinal study was designed. The follow-up will last at least three years, in order to describe the general and the intra-individual developmental pattern.

 

Methods:

 

Participants: After informed consent of the parents and oral assent of the children, 77 pupils of the German School in Lisbon e 119 truly monolingual pupils of two private schools of Lisbon area, attending second to fourth grade, were included in the study. The pupils of the German School formed the study group and the truly monolingual Portuguese children the control group. Children with learning disabilities, according to the information given by the teacher, were excluded.

 

Material: The material consists of three lists of words: German and Portuguese mono- and disyllabic nouns and a list of pseudo-words. Words were collected from the text books of the first grade of basic school, in order to be sure that these words are known by all of the children. For the sampling of the pseudo-words we chose only phonemes that make part of the phonetic repertoire of both languages and obeyed to the phonotactic rules of both. Thus, the phonetic form obtained, is the same for the two languages, while the written form is different. Each list is 25 words long, preceded by 5 training items.

 

In order to obtain information about the linguistic context of each child, we developed a questionnaire for the parents. We asked for: the mother tongue; the knowledge of German and Portuguese language of each of the parents; the number of siblings and their language ability to speak both languages; the time that the child has already contact with the languages; the time that the child spends with German or Portuguese peers; if the child has contact, and to which extent, to other languages beside the two investigated ones; finally we asked for an estimation of the linguistic abilities of the child to speak both languages and if applicable in other languages.

The teachers of the children were also requested to make the same estimation about the language abilities of their pupils to speak, to write and to understand Portuguese and German.

 

 

Procedures:

The items were recorded digitally and played back via PC with external loudspeakers using Windows Media Player. The word lists were recorded with native speakers of German or Portuguese. The recording of the pseudo-words was done by a journalist who has English as mother tongue, but also speaks Portuguese and German fluently. Thus we intended to avoid a bias to one of the investigated languages. The items were played back from a digital recording to groups of five to nine children in a dictation task. Altogether the children wrote four lists. The pseudo-word list were presented twice, in a Portuguese and in a German setting and the children were asked to write them according to the phoneme-grapheme correspondence rules of each language. The sequence, if Portuguese or German was presented first, as well as if pseudo-word or words came first, was alternated at random. The language, in which the instructions were given, was always congruent to the language that the children were asked for to write in the following dictation. According to this we obtained four sub-groups. Because of the difficulty of the task and to be not too demanding with the pupils of the second grade, we shortened the lists to 15 words each. The pupils of the second grade wrote 80 items and the pupils of the third and fourth grade wrote 120 items. After dictation of the half of the items the children paused for at least 10 minutes before going forward to the other lists of words and pseudo-words, dictated in the other language.

The Portuguese monolingual group performed only the dictation of words and pseudo-words written according to the phoneme-grapheme rules of Portuguese.

With the help of the information obtained by the parental and teacher’s questionnaire, we classified the children of the study group in three sub-groups: truly bilingual, German and Portuguese mother tongue.

 

Data analysis:

The number of correct answers and the type of errors are the dependent variables in this study. Orthographic errors were analyzed on five different levels:

1. We counted the overall error rate for each list;

2. We calculated an error index that consists of the quotient of the number of errors divided through the number of phonemes of the word. The value of this index varies from 0 to1. Whereas 0 is a correct answer, 1 means that for every segment of the word an error was found. By definition we considered all forms that exceeded an index of 0,75 as not classifiable.

3. We adopted the classification schema of Bishop and Clarkson, 2003, for orthographic errors. In this schema all forms are classified weather they are orthographically or phonetically acceptable or not: orthographically acceptable (OA); orthographically unacceptable (OU); phonetically acceptable (PhA); phonetically unacceptable (PhU). These results in the following four categories: OA/PhA, OA/PhU, OU/PhA, OU/PhU. They describe differentially the severity of an orthographic error. OA/PhA represents the less severe error, as the auditory discrimination of the stimulus was well succeeded, but the child chose the wrong form among the other homophonic forms. OI-errors indicate insufficient knowledge of the graphotactic rules and PhU-errors indicate difficulties in the discrimination of the phonetic sequence. In this contexto the OU/PhU-errors represent the most severe errors, since both steps of the correspondence process are disturbed..

4. We describe the error type in a so called SIOA-analysis. The letter “S” designates substitution errors. These are further subdivided into errors of generalization (a homophonic phoneme is wrongly written with the dominant form), interference (a phoneme is written following the rules of the other language), minimal pairs (the phoneme is substituted by another which distinguishes only in one articulatory aspect), context rule (the item is written wrongly because of a lack of knowledge of a context rule) and simple substitutions (none of the above mentioned categories is observed). The letter “I” designates inversions, where all graphemes are present but the sequence of graphemes is inverted. The letter “O” designates omission of graphemes and finally the letter “A” corresponds to additions of graphemes.

5. Accents wrongly used in Portuguese and nouns written without capitals in German were also registered. If a stimulus word was substituted by another more frequent word, we classified this as a “verbal paragraphia”. A pseudo-word transformed into a word was classified as a “lexicalization”.

Whereas the first three levels of analyses refer to the whole word, the SIOA-schema is a sub-segmental analysis.

 

 

Statistical approach and hypothesis:

Parametric and non-parametric statistical tests will be applied to compare the performance for the Portuguese lists of the pupils of the German and Portuguese School, stratified by years of schooling. We expect a higher error rate for the pupils of the German school for the second and possibly the third grade, as they have to deal with two phoneme-grapheme correspondence systems simultaneously. For fourth grade we expect an equal proficiency compared to the truly monolingual Portuguese pupils.

In order to establish the developmental pattern, comparisons of the above mentioned dependent variables between the grades of each school will also be performed. The overall error rate as well as the number of errors in each sub-category is expected to decrease with the progression of education.

In the study group correlation between the parental and teacher’s questionnaires about linguistic proficiency will be calculated and the results will determine the allocation to language proficiency groups. The severity and the diversity of the errors are expected to decrease with the progression of education in the way that the rate of unacceptable errors should be minimal. For the not truly bilingual children the phonetic unacceptable errors will preponderate in the second language, because the auditory discrimination ability is less accurate and the mother tongue can interfere in the writing. We expect to observe interference errors especially for the pupils that are not truly bilingual. As there is no word form memory for pseudo-word, we expect a higher incidence of interference errors for the pseudo-words list, because its writing obliges the strict use of the phoneme-grapheme correspondence rules.

The sequence of languages (German/Portuguese or Portuguese/German) and stimulus type (word/pseudo-word or pseudo-word/word) is not expected to have an effect on test performance.

 

 

Perspectives

 

By a preliminary analysis of the data we realized that pupils of the fourth grade unexpectedly made still a considerable number of errors, especially in the pseudo-words. Thus we plan to extend the study to the fifth and sixth grades.

We pretend to introduce into the study a second control group of truly monolingual German children, in order to have normative data for the German tasks.

In collaboration with the teachers of the German School in Lisbon we will use the results as a framework to detect specific difficulties in bilingual learning and to develop alternative approaches of teaching. 

 

 

References:

 

Bishop, D., & Clarkson, B. (2003). Written language as a window into residual language deficits: a study of children with persistent and residual speech and language impairments. Cortex, 39, 215-237.

 

Goldstein, B., Fabiano, L., & Washington, P. (2005). Phonological skills in predominantly English-speaking, predominantly Spanish-speaking, and Spanish-English Bilingual children. Language, Speech, and Hearing Services in schools, 36, 201-218.

 

McBride-Chang C, Bialystok E, Chong KK, Li Y. Levels of phonological awareness in three cultures. J Exp Child Psychol. 2004 Oct;89(2):93-111.

 

Snow, C. (1998) Bilingualism and second language acquisition. In Gleason, J. & Ratner, N. (Eds.), Psycholinguistics (pp. 453-481). Orlando: Hartcourt Brace Publishers

 

Tallal, P.; Miller, S.L.; Bedi, G.; Byma, G.; Wang, X.; Nagarajan, S.S.; Schreiner, C. Jenkins, W.M. & Maerzenich, M.M.. Language comprehension in language-learning impaired children improved with acoustically modified speech. Science 1996; 271: 91-84