A working definition of a native speaker is “a person who has spoken a certain language since early childhood” . Davis (1996) and Cook (1999) deconstructed this into attributes such as: subconscious knowledge of rules, intuitive grasp of meanings, ability to communicate within social settings, range of language skills, creativity of language use, identification with a language community, the ability to produce fluent discourse, and knowing differences between their own speech and that of the standard form of the language. Until the 1990s it was tacitly assumed that the only owners of a language were its native speakers. The objective of L2 learning was, therefore, to become like a native speaker as possible; any differences counted as a failure. The native speaker construct has however become increasingly problematic in SLA research.
On the one hand, it is a highly idealized abstraction. Native speakers of any language vary from each other in many aspects of grammar, pronunciation, and vocabulary for dialectal, social and regional reasons. So which native speaker should be used as a model? For French is it an inhabitant of Paris, Marseilles, Geneva, Quebec City, Parimaribo or Ouagadougou? The choice of a native speaker is related more to the status of particular varieties than to any properties of language. Additionally, any native speaker commands different genres of language rather than possessing a single monolithic form; speech varies from one moment to the next, accommodating to the speaker, situation, topic, and other factors. Hence the model of the native speaker in SLA research needs to allow for substantial variation.
On the other hand, this seemed to be one group exercising power over another (Phillipson). Since Boas, linguistics has refrained from value judgments about different groups of speakers. Treating the native speaker as the model for SLA is falling into the same trap of subordinating the group of L2 users to the group of native speakers, to which they could never belong by definition. The alternative is to treat successful L2 users as the model against which L2 users are measured. This led to the description of L2 user speech in its own terms (Jenkins, 2000) and to research on L2 grammar and vocabulary centered around the VOICE project (Vienna Oslo International Corpus of English) into English as Lingua Franca (Seidlhofer, 2004).
SLA research has then been questioning its faith in the native speaker as the only true possessor of language, leading some to reinforce the importance of an idealised speaker of a standard form of a language, others either to a more flexible version of the native speaker or to the establishment of a non-native model. The issue has been closely related to language teaching both in terms of the model that students should be offered in the classroom and of the relative merits of native speaker and non-native speaker teachers.