Abstract This paper focuses on Spanish grammatical mood variation in Comment Clauses (e.g., es mejor que no vayas (subjunctive) / vas (indicative) ‘it’s better you not go’) in the speech of two generations in New York City. The data come from 36 participants, 18 from each of two generational cohorts. Carried out within the variationist-sociolinguistic research paradigm, we test grammatical mood against eight variables, four external (generation, region, speaker sex, language skill) and four internal (grammatical tense, clause type, lexical identity, negation). Statistical findings reveal that generation significantly conditions subjunctive use (the first generation has a significantly higher rate of use of subjunctive forms than does the second generation); English skill conditions first-generation subjunctive use (those with ‘good or excellent’ English skills have a higher subjunctive rate than those with ‘fair or poor’ English skills); clause type conditions both generations’ subjunctive use (impersonal constructions yield a higher subjunctive rate than personal constructions); lexical identity and negation in the matrix clause both condition first-generation use of mood (gustar ‘to like’ favors the indicative; importar ‘to be important’ and ser + impersonal expression ‘to be’ both favor the subjunctive). Generational differences are thus observed with respect to both social and linguistic conditioning factors.

中文翻译：

---

纽约市西班牙语中的评论条款和情绪选择

摘要 本文重点研究了纽约市两代人的演讲中评论从句（例如es mejor que no vayas（虚拟语气）/vas（指示性）'it's better you not go'）中的西班牙语语法语气变化。数据来自 36 名参与者，两代人各有 18 人。在变异主义-社会语言学研究范式内进行，我们针对八个变量测试语法情绪，四个外部变量（世代、地区、说话者性别、语言技能）和四个内部变量（语法时态、从句类型、词汇同一性、否定）。统计结果显示，这一代显着地影响了虚拟语气的使用（第一代虚拟语气的使用率明显高于第二代）；英语技能条件第一代虚拟语气使用（英语技能“好或极好”者虚拟语气率高于“一般或差”英语技能者）；子句类型条件两代人的虚拟语气使用（非人称结构产生比人称结构更高的虚拟语气率）；矩阵从句中的词汇同一性和否定都限制了第一代语气的使用（gustar 'to like' 倾向于指示性；importar 'to be important' 和 ser + 非人称表达 'to be' 都支持虚拟语气）。因此，在社会和语言条件因素方面都可以观察到代际差异。子句类型条件两代人的虚拟语气使用（非人称结构产生比人称结构更高的虚拟语气率）；矩阵从句中的词汇同一性和否定都限制了第一代语气的使用（gustar 'to like' 倾向于指示性；importar 'to be important' 和 ser + 非人称表达 'to be' 都支持虚拟语气）。因此，在社会和语言条件因素方面都可以观察到代际差异。子句类型条件两代人的虚拟语气使用（非人称结构产生比人称结构更高的虚拟语气率）；矩阵从句中的词汇同一性和否定都限制了第一代语气的使用（gustar 'to like' 倾向于指示性；importar 'to be important' 和 ser + 非人称表达 'to be' 都支持虚拟语气）。因此，在社会和语言条件因素方面都可以观察到代际差异。