在有多位角色的文本中，作者常以不同的呈現形式介紹角色，如：人名、代名詞或角色描述名，且角色呈現的順序也不同。當文本中以代名詞指稱前文提及的角色時，讀者需要推論代名詞指稱何者。唯有正確的代名詞推論，才能促成文章的連貫，因此，角色提及順位與呈現形式等因素都會影響讀者代名詞推論歷程。本研究旨在探討此兩因素對中文讀者在代名詞推論的影響，以代名詞推論句的反應時間為依變項探究其推論歷程。本實驗以24位成人與26位六年級學童為研究對象，實驗設計為三因子混合設計：年級×提及順位×呈現形式，採用自我控制閱讀速度作業，紀錄逐句的閱讀時間。結果顯示三因子交互作用未達顯著，進一步分別探討提及順位（第一順位／第二順位）與呈現形式（人名／角色描述名）二因子對兩群年齡層讀者在代名詞推論之影響。結果發現，成人讀者在代名詞推論時，不受提及順位與呈現形式兩因素的影響；六年級學童在代名詞指涉第一順位角色時，閱讀時間不受角色呈現形式的影響，但在代名詞指涉對象為第二順位角色時，以人名呈現角色的方式會有助於代名詞推論。最後，從讀者在角色提及順位（第一順位／第二順位）與呈現形式（人名／角色描述名）所形成的四種情況之閱讀時間來看，六年級學童在代名詞推論的閱讀模式和成人的閱讀模式相近，此結果對於學童代名詞推論的閱讀發展歷程有更進一步的瞭解。

Narrative texts often contain several characters, and these characters have different levels of importance in the text. For example, some characters play primary protagonists, whereas others play supporting or minor roles. These characters are introduced in different orders and presented through different types of character expressions, such as proper nouns, pronouns, or role descriptions. When an author uses pronouns to refer to the characters mentioned in earlier sentences, readers must figure out which characters the pronouns are referring to. Therefore, pronouns can be integrated in earlier sentences. As a result, in later sentences, referential processing becomes essential for building discourse coherence. The order of mention and the types of character expression affect readers' pronoun resolution; and developmental data in the literature have shown that the ages of readers affect their pronoun resolution, too. The present study focused on how three factors affected pronoun resolution, namely the order of mention, the type of character expression, and the reader's age. We recruited 24 college students and 26 sixth grade students as the study participants. A three-way mixed design (age × order of mention × type of character expression) was adopted to examine the interaction patterns of these factors. The participants were asked to read texts sentence-by-sentence at their normal speed, and the reading time was then recorded. The results showed that no three-way interaction occurred. Two-way interactions were further examined in participants from separate age groups. The adult participants were unaffected by the order of mention and type of character expression. The sixth grade students were unaffected by the type of character expression when a pronoun referred to the first-mentioned character; however, in terms of the second-mentioned character, the proper name expression of the characters enhanced pronoun resolution in readers. This implies that the name of the character is more important than the role description for readers. Moreover, the reading time patterns were observed to be similar across these four two-way conditions for both age groups. Finally, some applications of pronoun resolution instruction and practical examplaes are serve as suggestions for teachers.