本研究利用轉換式，分析所有可能的三項系列題。三項系列題的推理時間可以分為三成份加以研究;個別前提的理解，較大消息結構（LSI）之形成，自LSI演繹所需消息單位。當解決三項系列題時，依靠前題的情況，受試者會傾向於採用形式處理或內容處理的方式。各種語言中，句子內的字排列次序並不相同，因而會決定受試著採用形式或內容處理的方式。若以符號表示前題，則受試者會傾向於採用形式處理的方式。使用中文及符號寫成的三項系列題所得結果，均支持理論上的預測。

　　　　　A concise analysis of various forms of the three-term series problem, each of which retains the original order among three terms, is presented in termsof transformations. The reasoning time is then decomposed into three interacting components: comprehension of individual premises, formation of a larger structure of information (LSI) by combining information from the two premises, and deducing from this LSI an item of information not explicitly stated in either premise. Depending on certain conditions of the premises, one tends toadopt either form-based or content-based processing. Certain languages having a particular word order within propositions predispose one to adopt form-based processing not only for the negative-equative but also for the comparative construction. Expressing the premises in a symbolic form also leads one to adopt form-based processing. As a consequence, information from linearly transitive premises can be combined more efficiently. As for the component process of information deduction, the resultant LSI, depend only on whether the premises are written in a symbolic or linguistic fom. The results obtained from an experiment with the three-term series problems expressed in a symbolic form and written in a language different from English are in accordance with the present model.