Relationships have been hypothesized between various areas of language learning success and FD/I at both ends of the continuum. It has been suggested (see, for example, Chapelle and Roberts, 1986) that in foreign language learning, FI students will be better at learning the L2 through the study of grammar rules and vocabulary, whereas FD students will be better at learning through interaction with native speakers. However, most of the studies of FD/I in foreign language learning to date have focused on the relationship between FD/I and foreign language aptitude or success. Many of these studies have found FI to be a predictor of success in foreign language learning, especially when that success has been measured by discrete tests such as structure tests, cloze tests and dictations.

For example, Naiman et al (1978) found that FI is an important predictor of success in the later stages of language learning, but not the early stages. FI students have also been found to be better than FD students at imposing structure on their work (Genesee and Hamayan, 1980), performing well in discrete language multiple-choice tests, cloze tests and dictations (Chapelle, 1988) and achievement-oriented final exams (Fuller-Carter, 1988).

Although the correlation between FI and success at foreign language learning has been well documented, links between the FD cognitive style and any kind of language learning success, though postulated, have rarely been demonstrated. It is often suggested that FD might relate to communicative competence because of the alleged superior interpersonal skills of the FD individual, but this has rarely been proven to be the case.

Chapelle and Roberts (1986) found communicative competence to be strongly and significantly correlated not with FD but with FI. Johnson and Rosano (1993) did find that FD correlated with tutors' ratings of students' communicative competence. However, they admit that out of a population of 45 students, data was only available for five of the students, as most tutors felt that they had had insufficient opportunity to observe their students' proficiency. This fact significantly weakens the strength of their findings. Abraham (1985) looked at the language learning preferences of students with FI and FD cognitive styles. She found that when learning L2 grammar, FI students performed better with a deductive lesson and FD students performed better with an inductive lesson. Thus it seems that, as one might predict, some relationships have been found between FI and language learning strengths and preferences, but very few links have been found between FD and language learning strengths or preferences.

Findings with respect to the holistic/analytic cognitive style dimension do seem to be more balanced, from the point of view of students at the holistic end of the continuum. Holistic students have been found to benefit from the presence of 'advance organizers', such as subtitles (Douglas and Riding, 1993), and to prefer group work and closed tasks (Riding and Read, 1996). The holistic/analytic cognitive style has been found to be a significant predictor of the choice of language made by undergraduates when selecting a language subsidiary. For example, in a study carried out at a British university, Littlemore (in press) found that, when choosing a subsidiary language, analytic students were more likely to opt for Japanese, whereas holistic students were more likely to choose to study Spanish. This suggests that analytic students may be drawn towards more apparently 'complex' languages than holistic students. The holistic/analytic style dimension has also been found to affect the choice of communication strategies made by language students (Littlemore, 2001b) when faced with gaps in their target language lexis. For example, if he or she does not know the word 'acorn' a holistic student is more likely to compare it to a similar object (eg 'a small pipe'), whereas an analytic student is more likely to break it down into parts (eg 'small and brown, with one part inside another'). Holistic students have also been found to be better at interpreting and producing novel metaphors in the target language (Littlemore, 2001c).