RT Journal Article SR Electronic(1) A1 Huiskens, Lucie YR 1986 T1 Zicom, een Computergestuurd Oefenprogramma Voor Zinsontleding JF Toegepaste Taalwetenschap in Artikelen VO 25 IS 1 SP 91 OP 100 DO https://doi.org/10.1075/ttwia.25.10hui PB John Benjamins SN 0169-7420, AB In this article I try to show that the relationship between grammar teaching and grammar learning is not as close as one often assumes on intuitional grounds. My arguments are based on research at the University of Amsterdam. They are of a psychological, didactic and psycholinguistic nature.In the first place, teachers assume too easily that learners will integrate into their knowledge system the rules presented to them during the instruction in more or less the same way in which these rules have been presented. Instead, learners have their own individual way of internalizing knowledge. They not only build up an incomplete network of rules, but they also invent idiosyncratic rules that cannot be explained by L1, L2, or the instruction (Van der Linden 1985).Secondly, it seems that these networks of rules resist restruc-turing. That is, new rules are incorporated into the system with some difficulty; the learner seems to prefer to stick to what he already knows.In the third place, grammar tests do not always seem to test precisely the things that were taught during instruction. From the more or less complete set of rules taught, the "simple" rules are often absent in the tests, because it is supposed that these are known by all students. Therefore, it is mainly the "difficult" rules, i.e. "exceptions" that figure in the test. As a consequence, a student's results do not reflect his language proficiency in an everyday sense which is why native speakers have often great problems in making these tests). Obtaining a 80% score (a com-monly used norm to pass the exam) means a more than 80% mastery of the rules concerned.Finally, there is a discrepancy between the knowledge presented to the student in the instruction and the knowledge acquired in other learning situations. This difference concerns the register of the language. Whereas the student is confronted with different registers in his learning process (poetic, familiar, "argot",...), the grammar instruction (and test) limits itself to the norm of the standard language ("français standard"), and considers expressions of another register incorrect.For all these reasons, there are discrepanties between the teaching and the learning of grammar. Careful research and reflection on this problem could lead to a closer relationship between both activities., UL https://www.jbe-platform.com/content/journals/10.1075/ttwia.25.10hui