Null Anaphoric Possessor Arguments of Kinship Nouns and

Null Anaphoric Possessor Arguments of Kinship Nouns and Long-Distance Binding in Mandarin Hezao 1, 2 Ke , Ya 2, 3 Zhao , Liqun 2 Gao , Shuying 2 Liu The University of Michigan, Ann Arbor 1, Beijing Language and Culture University 2, Northwest University for Nationalities 3. Background Experiment 1 Experiment 2 Focus of this study: kinship nouns in Chinese, which is a typical kind of relational nouns (RNs). Test sentence: 2) a. Zhangsan dai-le erzi qu Qingdao. Zhangsan take-ASP son to Qingdao. ‘Zhangsan took (his) son to Qingdao. ’ Lead-in sentence: Now, let me say something about this story. In this story, Mickey Mouse talk to Donald Duck about going to an island for a trip. Finally, er…… Definition: RNs, in narrow sense, are Nouns which have more than one argument in their lexical meaning. (Barker, 1995; Partee 1987/1997; etc). As a RN, a kinship noun such as father has a lexically inherent extra argument. [[father]]= [λxλy. father (x, y)] [[Mary’s father]]= [λxλy. father (x, y)](m) = λy. father (m, y) Father picks up Mary as one of its arguments. Research Questions Q 1: Is the implicit argument pronominal or anaphoric? (Experiment 1) Q 2: Is the implicit argument a theta assigned syntactic argument or only a semantic argument? A related question: is there any syntactic processing involved in the interpretation of kinship nouns? (Experiment 2) Intuitions are not sharp enough. . . Control condition: Used a similar story except all the occurrences of kinship nouns were replaced with other nouns such as schoolbag, dog, etc. At the end of the story, a control sentence was presented: 1) Zhangsan qizhe luotuo qu lvxing. Zhangsan ride camel to travel ‘Zhangsan travelled on a/his camel. ’ Are the “semi-relational” such as pets and other human-owned things the same with kinship nouns? Achimova, Deprez & Musolino (in press): Controlled experimentation should be underscored as an important tool for theory construction. Experiment 1 Method Subjects: 17 master’s students in linguistics were recruited from Beijing Language and Culture University (BLCU). Procedure: § A variant of Truth Value Judgment Task (Crain & Thornton, 1998) § An experimenter told a participant a story. § Then test/control sentences were presented by the experimenter as a statement concerning the story. § The subjects’ role was to determine the truth value of the sentences according to the story. Materials: 8 stories (4 test and 4 control sentences) each followed by a test statement. Typical test item story (in Chinese): Zhangsan and Lisi wanted to bring their sons to Qingdao for a trip. Eventually, Zhangsan took Lisi’s son but not his own son to Qingdao. After the story, a lead-in sentence and test sentence (2 a) were presented to the subjects. Lead-in sentence: Now, let me say something about this story. In this story, Zhangsan talked to Lisi about going to Qingdao for a trip. Finally, er…… POSTER TEMPLATE BY: www. Poster. Presentations. c om The NP Mickey Mouse c-commands the kinship noun son, but Donald Duck, which is a NP inside the PP, does not c-command son. Results The participants rejected the test sentences with a rate of 89. 7%, but accepted the control sentences with a rate of 91. 2%. 100. 0% 90. 0% 80. 0% 70. 0% 60. 0% 50. 0% 40. 0% 30. 0% 20. 0% 10. 0% NP K Mickey-Mouse VP VP c-command VP V take-ASP PP P at 1. 5% 0. 0% V’ NP son NP V’ VP V Donald-Duck hospitalize DE time PP to island doesn’t c-command acceptance uncertainty V travel Figure 1: Percentage of rejection, acceptance and uncertainty to test/control sentence. Discussion § (The implicit argument of) kinship nouns should be obligatorily bound by their antecedents, where they can pick up their referents, while nonrelational nouns do not have bound arguments. § Reasons for exceptions: Ø Reasons to accept test sentences: 1. In one out of six acceptances of the test sentence a subject forgot the content of the story. 2. One subject said the test sentences were true in all his 4 answers. Ø Reasons to reject control sentences: 1. There are six rejections of the control sentences. A subject consistently rejected the control sentences because he thought the expressions were ambiguous and not accurate. However, this indicates that the subject realized that non-relational nouns are not obligatorily bound. 2. In another rejection the subject forgot the content of the story. Control condition: A similar story was presented except all the occurrences of kinship nouns were replaced with other nouns such as fish, hamburger, etc. At the end of the story, a control sentence was presented: Control sentence: 3)b. Xiaoqingwa [PP zai xiaohema shuijiao de shihou] chidiao-le hanbao. Frog at Hippo sleep DE time eat-off. ASP hamburger. ‘Mr. Frog, during the time when Mr. Hippo was asleep, ate (his) hamburger. ’ Results The participants overwhelmingly rejected the test sentences and accepted the control sentences, with a rejection rate of 95. 0% and an acceptance rate of 88. 3%, respectively. Results of Experiment 2 95. 0% 90. 0% 88. 3% Procedure: The same as in Experiment 1. Materials: 8 stories (4 test and 4 control sentences) each followed by a test statement. Typical test item story (in Chinese): Mickey Mouse and Donald Duck wanted to bring their sons on a trip. However before their departure, Donald Duck got a toothache, and was sent to the hospital. Eventually, Mickey Mouse took Donald Duck’s son but not his own son to the island. § Other evidences follow. Evidence 1 Kinship nouns VS. body-part nouns: Long -distance and local binding. Compare (4) with (5): 4) Lisi. J xihuan Zhangsan. K meitian xi toufa*J/K. Lisi like Zhangsan everyday wash hair ‘Lisi likes that Zhangsan washes his hair everyday. ’ 5) Zhangsan. J renwei Lisi. K hen liaojie erzi. J/K. Zhangsan think Lisi very know son ‘Zhangsan thinks Lisi know his son very well. ’ § The implicit arguments of kinship nouns can be long-distance bound (5), whereas those of body-part nouns can only be locally bound (4). This distinction can be accounted for if we assume there are two types of anaphora in Mandarin Chinese (MC). Evidence 2 60. 0% test sentence control sentence 50. 0% 40. 0% Subjects: 15 master’s students in linguistics were recruited from BLCU. None of them had participated in Experiment 1. § Experiment 2 shows that a syntactic requirement, i. e. a c-command requirement, is involved in the interpretation of kinship nouns. This evidence underpins the idea that the extra argument is syntactic rather than only semantic or pragmatic. 80. 0% 70. 0% Method § The results of the first experiment can be readily explained by assuming an extra syntactic argument inside the kinship nouns. This argument is anaphoric since it is obligatorily bound by its antecedent. The results also show kinship nouns are relevant to inalienability. Supporting a Syntactic Approach 100. 0% Experiment 2 General Discussion Asp. P Asp’ V’ control sentence 8. 8% Kinship nouns are different from nonrelational nouns in that a kinship noun has an implicit anaphoric possessor argument, and this argument is obligatorily bound by some NP which c-commands it. T’ test sentence 8. 8% Conclusions TP 91. 2% 89. 7% rejection Example: “semi-relational” nouns Test sentence: 3) a. Milaoshu [PP zai Tanglaoya zhuyuan de shihou] dai-le erzi qu xiaodao lvyou. Mickey-Mouse at Donald-Duck hospitalize DE time take-ASP son to island travel. ‘Mickey Mouse, during the time when Donald Duck was hospitalized, took (his) son to the island for a trip. ’ Control sentence: 2)b. Zhangsan na-le shubao hui sushe. Zhangsan take-ASP schoolbag to dormitory. ‘Zhangsan took a schoolbag to the dormitory. ’ Results of Experiment 1 Conclusions and General Discussion 30. 0% 20. 0% 11. 7% 10. 0% 5. 0% 0. 0% rejection acceptance Figure 2: Percentage of rejection, acceptance to test/control Discussion § (The implicit argument of) a RN is bound by its antecedent, which ccommands it. But a non-relational is free from binding. § Reasons for rejection of control sentences: Five out of seven occurrences of rejection of control sentence were in the third control sentence, where there is a contrast between possessors of a “fish” in the lead-in sentence. § A possible correlation between the existence of long-distance bound anaphoric possessor arguments of kinship nouns and long-distance bound reflexives. § If a language L has null longdistance bound anaphor, then L will also have overt long-distance bound anaphor (a counterpart to ziji ‘self’ in MC). § Sinhala, Korean, Italian, Dutch, Norwegian, Japanese, Russian have both a long-distance bound reflexive and long-distance bound null anaphoric argument, while Spanish and German have neither. Hezao Ke, hezaoke@umich. edu