ITL - International Journal of Applied Linguistics - Volume 74, Issue 1, 1986

The main problem in learning foreign language vocabulary is retention. There exists a considerable amount of evidence pointing to the superiority of mnemonic techniques over rote repetition in the learning of verbal material. An investigation was carried out comparing the learning of FL vocabulary in three different conditions : (1) making formal and semantic links with known words (2) making verbal links of a solemy semantic nature (3) using habitual methods. The first experimental group demonstrated superiority, although scores were not significant on an anamysis of variance. Indications were given by the data that training and encouragement in the use of formal and semantic verbal association will lead to better vocabulary retention.

Using continuous variables derived in information theory and from a Markovian model of phonemic chaining, we have anamysed some strings of jar-gonic speech of French speakers (glossolalics, schizophasics, aphasies, poets) and of natural French2. These anamyses give a quantitative meaning to the concept of phonological plane. Discriminant anamyses separate our groups in a clear-cut fashion. Elementary numerical cmustering provides a meaningful interpretation. This type of approach could bring, in its future developments, further insights into the phonological systems of languages.

In recent studies on the acquisition of a second or foreign language in a non target language environment, environmental factors such as lack of native peer input, a formal language learning situation and pressure on the learner to produce the required target language structures have been considered to have attributed to a high incidence of L1 induced or interlingual errors.

In order to investigate this problem, I collected spontaneous speech samples from some eighty francophone students enrolled at the Faculté de Sciences Economiques and the Ecole de Solvay who were learning English at the Institut de Phonétique at the Université Libre de Bruxelles. The data represented two levels of learning, the first degree level when students entered the course and the third degree level which marked the end of the course. My hypothesis was that given a difference in emphasis within the TL syllabus from that of TL forms (communicative English) to that of TL content (business English) and given also that the same methodological approach was adopted with native teacher input and optimal corrective feedback, students might progress beyond Ll-induced to somewhere approximating TL rule formation.

The error samples were broken down into categories representing syntax, word order, morphology, semantic concepts and lexis. They were examined and compared across levels for examples of interlingual (Ll-induced) and organisational (non Ll-induced) errors.

An anamysis of the data indicated a decrease in interlingual errors but an increase in organisa-tional errors at the third degree level, suggesting the existence of a continuum that accommodates possibilities of fossilization due to lack of native peer input. My data also suggested the role of L1 in its influence on the formation of semantic concepts in TL and in organising, comparing and hypothesizing lexical domains in the two languages.