An important factor impeding the development of explicit theories of translation has been the centrality of coincidence. Skilled translating consists not of following rules or algorithms of equivalence, but of generating coincidences between the materials of the source language and those of the target language. Conventional aspirations of linguistic theory emphasize degrees of generality, uniformity and formality, which such an activity does not readily seem to fit. Also, language science and linguistics have consistently rated form over meaning and language system over communicative context, while translation is an activity in which meaning dominates over form, and context immediately controls and influences how the language system is used. Some approaches to text and discourse have been striving to revise traditional theoretical aspirations in order to attain better models of language use, and may thus provide a basis for unifying theory with practice in translation.