Schema-based textual analysis of domain-controlled authentic texts

This paper reports a textual analysis of 22 authentic and unmodified newspaper and magazine articles dealing with a current political issue within the period of nine months. The analysis was based on schema theory, which divides all the schemata, i.e., words comprising the articles, into three main categories: semantic, syntactic and parasyntactic. It was hypothesized that the very nature of semantic, syntactic and parasyntactic schemata would render the schemata different in terms of their occurrence within texts analyzed. Chi-square tests confirmed the hypothesis and showed that semantic, syntactic, and parasyntactic schemata differ from each other significantly not only in type (X2= 8371.993, df = 2, p < .0001), but also in their subcategorization (X2 = 5866.867, df = 12, p < .0001) and thus endow both schema types and their subsumptions with psychological reality. Further descriptive analyses indicated that semantic schemata make up 77% of the whole text whereas parasyntactic and syntactic schemata accounted for only 17% and 6%, respectively. The pedagogical implications of these findings within foreign language teaching and testing context are discussed.