帶施事題元的漢語述賓式合成複合詞有兩種：OVS型和VOS型。前者是純粹的向心型詞結構，而後者卻不是如此。在VOS型當中，如果V和O分別是單音節的，那麽VO是VP經模式聯體記憶而轉化過來的詞根，跟S再組成複合詞。如果V和O分別是雙音節的，那麽VO是VP直接進入複合詞，即所謂“短語入詞”現象。模式聯體記憶和運用規則之間的相互補充反映了語法運作的經濟原則。漢、英兩種語言中的合成複合詞都為這一原則提供了證據。不同的是，漢語合成複合詞中的施事題元是一個粘著詞根，而英語合成複合詞中的施事題元卻可以是一個動詞後綴。這樣，漢語合成複合詞在成分結構中直接就滿足了“題元指派統一論”的要求，而英語合成複合詞（如果施事題元是一個動詞後綴的話）就需要在邏輯形式中去完成相同的要求。

Chinese OVS compounds are generated by morphological rules, whereas VOS compounds are far less homogenous. The VO part is a root word if the V and O are monosyllabic, or a VP if the V and/or O are disyllabic. In the former case, the syllable count may have conditioned a pattern-associated memory that has transformed a VP into a root. In the latter case a lexico-syntactic compound has been created. These findings are in agreement with the principle of memory vs. symbol-processing proposed by Pinker (1999) and provide proof that the principle works in more areas than inflectional morphology and may indeed be universal.