The cognitive style dimension that has received the most attention, particularly in language learning, is that of field dependence/independence (FD/I) which was proposed by Herman Witkin (see, for example, Witkin et al, 1977). This is a measure of a person's ability to separate or disembed a single item from a larger context. While a more field-independent person will find it easier to overcome the influence of the larger context, a more field-dependent subject will have greater difficulty in doing so. The definition of FD/I has been extended to include a social dimension (Witkin and Goodenough, 1977). FD subjects are alleged to perform better in social contexts given their superior interpersonal orientation. The technique generally used for measuring FD/I is the Group Embedded Figures Test (GEFT). This test consists of approximately fifty items, in which the participants must locate a simple figure within a much more complex one. Those participants who are able to complete the task quickly and successfully are labelled Field Independent (FI) and those who are less successful are labelled Field Dependent (FD).

A tendency towards FI is thought to be beneficial to the development of cognitive restructuring abilities. This type of cognitive analysis involves changing or using the 'field' of available information rather than accepting it and learning it 'as it is'. Restructuring can take various forms, such as breaking up an organized field so that its parts can be viewed as separate from the background, providing structure in a field that lacks it, or imposing new organization on a field. FI has been found to correlate with success on tasks in the subjects' native language. For example, Berent and Silverman (1973: 1328) found FI students outperformed FD subjects on the verbal part of a paired-associate learning task. Kagan (1980) found a correlation between FI and greater syntactic complexity of language used by children speaking their native language.

In recent years, a number of aspects of FD/I have been heavily criticized (see Chapelle, 1992; Chapelle and Green, 1992; Skehan, 1998; Littlemore, 2001b). Firstly, the Group Embedded Figures Test (GEFT) does not deal adequately with the cognitive attributes of the field dependent subjects. In this test, the field independent subject is able to find a simple figure embedded in a more complex one and the field dependent subject is simply unable to complete the task. FD individuals are simply those who are unable to organize and impose structure upon ambiguous stimuli. This implies a mental deficiency rather than a style of thinking. Secondly, the GEFT does not offer the subject a choice to see how he or she prefers to deal with a situation. Rather, it requires the individual to perform in a particular way and assess how able he or she is to meet the task requirements. Again, what is being most directly assessed by the test of FD/I is not so much a style of functioning as an ability to function well in a certain type of task.

In the light of the above criticisms, it has been suggested that FD/I should be replaced by a 'holistic/analytic' cognitive style dimension, (see, for example, Riding and Cheema, 1991). This style dimension, which can be measured by means of the computer-based Cognitive Styles Analysis (CSA) (Riding, 1991) is very similar to FD/I, except for one thing: what the holistic/analytic distinction offers that is not offered by the FD/I distinction, is a cognitive element to the equivalent of the 'FD' end of the continuum. Holistic processing consists of drawing together pieces of information and treating them as a whole, or of perceiving similarity and togetherness, whereas analytic processing emphasizes the perception of difference and separateness. Thus it can be said to be a true bi-polar measure of cognitive style and is not simply a test of intelligence. The fact that there are cognitive skills associated with the holistic end of the scale might give researchers clues as to the kinds of cognitive tasks these individuals might be capable of, whereas before researchers had only looked for social skills. In the following section, we discuss the impact of both FD/I and the holistic/analytic cognitive styles on foreign language learning.