Treffer: Computer Processing of Swedish Syntactic Data. Some Preliminaries and Tentative Results.

The encoding of Swedish syntactic information in a study of linguistic data processing is described in detail. The choice of syntactic variables and classification criteria, illustrated with examples, is also discussed. In addition, uses of the coding for contrasting languages, as in this study's comparison of Swedish, Finnish, and English word order, are explored. The coding key consists of 64 variables corresponding to syntactic or semantic properties of a clause. Each clause in the corpus is classified in a subcategory of one of the variables. The variables are divided into four groups: those identifying the clauses, those specifying some syntactic properties of the entire clause or its context, those classifying the clause constituents (verb, subject, object, complement, initial adverbial, adverbial in the central field, non- final adverbial in the end field, and final adverbial in the end field), and those specifying what transformations have been applied to the clause and what constituents they have affected. (MSE)