The problem of variation in SLA theory and research Cristina Sanz Georgetown University Introduction. Within the concept of variation in interlanguage, we need to distinguish between inter-learner and intra-learner variation. Inter-learner variation is used to refer to differences in the production of x number of L2 learners considered to be at the same level of language proficiency. Intra-learner variation can be further subdivided into diachronic variation or changes in production found when comparing the L2 production of the same learner at different moments in her learning process; and synchronic variation, or differences found in the production of one single learner at one point in her learning process. In this paper I concentrate on synchronic intra-learner variation. The first part of the paper discusses ways in which different perspectives in Second Language Acquisition (SLA) theory have approached variation, from those who place it at the heart of the SLA process, to those who consider it problematic and avoid it altogether. A special section is devoted to discussion of the way in which the psycholinguistic construct of attention has been used. The second part deals with issues related to L2 variation in research methodology and discusses results from an experimental study that attempted to proceduralize attention. The problem of variation in SLA theory. Two terms are important in order to understand the way interlanguage variation has been approached: knowledge and access, or the mechanisms that put knowledge to use. In Chomskyan views of acquisition, there are two types of linguistic knowledge, explicit and implicit knowledge, and production varies according to the intensity of the interaction between them. In a variationist approach to language acquisition, variation is the manifestation of one variable source of knowledge. In processing approaches to SLA, variation is the effect not only of different types of memory, but also of different ways of accessing that knowledge. Eckman (1994) presents a summary of the discussion between Tarone (1990), a variationist, and Gregg (1990), a Chomskyan, on a theoretical problem that centers around knowledge. The variationist view. Variationistshave adapted Labov's framework (1972) to SLA and consider variation essential to understanding acquisition (Adamson