A description only looks at features that are observable. In this step of the analysis you tell what is the case (Kohn, 1974). Here errors are not yet being explained, but rather they are categorized into different linguistic areas. An error might occur due to phonetic reasons, it might also occur as a result of morphological reasons as well as for syntactic ones. Also errors can appear in the areas of pragmatics or semantics. Error classification tries to classify an error to one of those linguistic fields. It specifies how the forms formulated by the learner deviate from the forms a native speaker would have produced in the respective context (Barkhuizen & Ellis, 2005). It tries to find an answer to the question of “what the error consists of” (Ellis, 1994, p. 57). Apart from developing categories this step of an error analysis also consists of “recording the frequency of the errors in each category” (Barkhuizen & Ellis, 2005, p. 60).
The French learn to value and practice eloquence from a young age. Almost from day one, students are taught to produce plans for their compositions, and are graded on them. The structures change with fashions. Youngersters were once taught to express a progression of ideas. Now they follow a dialectic model of thesis-antithesis-synthesis. If you listen carefully to the French arguing about any topic they all follow this model closely: they present an idea, explain possible objections to it, and then sum up their conclusions. [...] This analytical mode of reasoning is integrated into the entire school corpus.