First LING3201 Language Acquisition Term Project paper 2011
Natalie Yu Tsz Sum (1155000186) Fall
Usage and Development of Cantonese Sentence-final Particles wo3, wo4, and wo5 in native Cantonese-speaking Children
1. Introduction Sentence-final Particles
Sentence-final particles (SFPs) are minimal lexemes occurring at the end of a sentence with no referential meanings, but many of them are related to modality and pragmatics. They can appear in the form of discourse particles and act as qualifiers (or grammatical modifiers) of the sentence. Linguist Chao Yuen-Ren has described sentence-final particles as "phrase suffixes" just as word suffixes are in construction with the word preceding it, a sentence-final particle or phrase suffix is "in construction with a preceding phrase or sentence, though phonetically closely attached to the syllable immediately preceding it" (Chao 1968). Sentence-final particles are found mainly, but not exclusively, in many East Asian languages, including Mandarin and Cantonese.
2. Central Research Issues 2.1 Literature Review
The Classification of Cantonese SFPs by A. Law (2002)
1
"Cantonese sentence-final particles (SFPs) are bound forms attached to the end of sentences and constitute an important grammatical category in the language. The number of SFPs in Cantonese ranges from 30 (Kwok 1984) to 95 (Leung 1992), depending on how one counts them. Functionally, especially in the early studies, they are often said to be similar to intonation in nontonal languages (e.g. Chan 1998, Cheung 1972, Kwok 1984), mainly because many SFPs carry emotive meanings. Some studies (e.g. Luke 1990) prefer the term "utterance particles", suggesting that these particles are attached to utterances rather than sentences. Numerous studies have found that SFPs express a wide range of meanings such as aspect, focus, modality, speech acts and temporal order (cf. Chan 1998, Cheung 1972, Fung 2000, Kwok 1984, Law 1990, Luke 1990, Leung 1992, Matthews and Yip 1994, Lee and Yiu 1998a, 1998b, 1999 and Lee and Law 2000, 2001)." (Law 2000)
1
Excerpt of table on Cantonese SPFs by Leung Chung-sum
(Leung 2005)
2.2 Research Focus
It is significant how an additional word aka an SFP at the end of a sentence will change the meaning of the whole sentence or the focused constituent it covers. The Cantonese SFP system is much richer and complicated than the Mandarin one. The number of SFPs in Mandarin is also much smaller than in Cantonese. Unlike English, verbs do not carry tense-markers in themselves in Chinese. See these sentences in colloquial Cantonese. (1) ? ? ? ? Siu-ming eat rice "Siu-ming has a meal." (?) Note that it is difficult to determine what the speaker is trying to express through this sentence because it is ambiguous in terms of semantics due to unknown illocutionarity. (2) ? ? ? ? ? aa1 Siu-ming eat rice SFP "Siu-ming will eat (a meal)."/ "Siu-ming eats (a meal)!" (3) ? ? ? ? aa3 1 Note that Leung did not mark the SPFs with corresponding Jyutping tones, those little numbers are only indicating different meanings that he treats as separate entries, like those in dictionary. I have analyzed and marked the respective meanings of wo3, wo4, and wo5 in the next paragraph. 2
Siu-ming eat rice SFP "Is Siu-ming eating (a meal)?" (4) ? ? ? ? laa1 Siu-ming eat rice SFP "Siu-ming (you) should eat." (5) ? ? ? ? laa3 Siu-ming eat rice SFP "Siu-ming is going to start eating (a meal)." (6) ? ? ? ? laa4 Siu-ming eat rice SFP "Is Siu-ming going to/already eat (a meal)?" (7) ? ? ? ? zaa1 Siu-ming eat rice SFP "Siu-ming only eats (a meal)" 8. ? ? ? ? zaa4 Siu-ming eat rice SFP "Is Siu-ming only eating rice/a meal?"2 Most SFPs in Cantonese switch sentences between interrogatives and non-interrogatives with tonal change, but notice this is not the case for the three wo's wo3, wo4, wo5. Interestingly, none of sentences (9) to (11) are interrogatives. (9) ? ? ? ? ? ( wo3) (10) ? ? ? ? ? ( wo4) (11) ? ? ? ? ? ( wo5) Combining the observations from both Law (2000) and Leung (2005), wo3 in different contexts expresses reminder, negotiation, unfairness, blame, and introduction of new information, wo4 expressed surprised or unexpectedness; while wo5 is a quotative marker. Wo4 and wo5 also appear to be incompatible with question, whereas wh-phrase has to be interpreted as an indefinite pronoun in a sentence with wo4 or wo5. E.G. ? ? wo3? ? ? ? ? ? ? ? ? ? ? wo4? ? ? ? ? ? wo5? In this paper, I would like to find out if the actual usage of the three wo's insert to sentences match with those from the table showing all sentence-final particles in Cantonese according to Ann Law 2 Unlike other sentence, where "? ? " is interpreted as having a meal, because the zaa4 SFP indicates the
element of "only", therefore the sentence can be either asking if Siu-ming is eating only white rice, or if Siuming is only eating a meal (and not participating in other activities).
3
(Law 2002) and Leung Chung-Sum (Leung 2005), in other words, to see if children are using them with the adult "standards". And non-interrogative neutral SFPs are said to be the first to be acquired in according to Lee (Lee 1995), that means wo3 (expressing reminder) is supposed to be acquired earlier than wo4 and wo5. I aim to find evidence for (or against, if there is any) both claims, and alongside anything interesting.
3. Methodology 3.1 Source of Data
In order to observe children's development on their usage of the three wo's, naturalistic data is mandatory. Experimental approach might not be sufficient in obtaining desired and conclusive results, as it requires many different kinds of scenarios and contexts for utilization of the three wo's from children. Therefore, I turn to the Hong Kong Cantonese Child Language Corpus (CANCORP) [2011 version]. The CANCORP is a one-million-character child language corpus constructed to study the early grammatical development of Cantonese-speaking children in Hong Kong. It is a set of audiorecordings of the longitudinal language development of eight Cantonese-speaking children who were each observed for around 12 months. The beginning age of observation was between 1;07 and 1;11 for four of the children, and between 2;02 and 2;08 for the remaining four children; the age at which observation ended was between 2;07 and 3;08. The mean number of observation sessions for each child, with each session lasting approximately one hour, was 21. Four of the child subjects were male, and the other four female. The CANCORP consists 171 transcripts, were coded according to the CHAT format (Codes for the Human Analysis of Transcripts), and tagged with 33 parts-of-speech labels. The transcripts recorded conversational exchanges between the children and various adults, mostly the investigators, and often caretakers and other members of the family as well.
Table showing coded names, genders, recorded ages and numbers of files of the eight subject children
3.3 Procedures
4
First, I looked at the appearance frequency of the three wo's in all 171 transcripts of the eight children. Including adult input (uttered by interviewers, caretakers, and children), there are 104 entries containing wo3, 89 entries containing wo4, and 23 entries containing wo5. Using sampling method3, among entries including only child input, 64% containing wo3, 32% containing wo4, and exactly 1 production containing wo5 among all 171 transcripts. Illustrations cited are mainly of Chunyat (CCC), and the examples do not include utterances considered to be imitation preceding adult speech, meaning within a considerable time in the conversation, there is no occurrence of the same SFPs as the target SFPs uttered by the child. Chosen excerpts that specifically illustrate my findings are presented in the following sections. Before jumping into the data analysis section, I must introduce the transcript coding method. The transcript code is composed of two parts: child subject's code name and his/her age when the interview took place. E.G. ccc020307 is the file of child Chunyat (coded ccc) when 2 years, three months, and seven days old. (ccc020307) *MOT: ? ? ? ? ? ? ? ? ? ? ? ? . *MOT: ? ? ? ? ? ? ? ? . *CHI: ? ? ? . *MOT: ? ? *CHI: ? . *MOT: ? ? ? ? *MOT: ? ? ? ? *CHI: ? ? &wo3. *INV: ? ? &lo1. 3.4 Data Analysis 3.4.1 Data Analysis on wo3 (ccc020410) *INV: ? ? ? &a3 ? *CHI: ? &a3. *MOT: ? . *INV: ? ? ? ? &a3 ? *CHI: ? ? ? &wo3. *CHI: ? ? ? . *INV: ? ? ? ? ? &gaak3 ? *CHI: ? ? ? &a3 . *INV: ? ? ? ? &a3 ?
First let us look at transcripts coded ccc020307, ccc020410, and ccc020624.
3 Using sampling method, I first counted the numbers of the three wo's in every transcript of the following four children: CCC, LTF, CGK, and MHZ. Then I calculated the percentages by dividing the child entries (of CCC, LTF, CGK, MHZ) that contain the target SFPs by the total numbers of entries from the four children. The sampled results reflect the characteristics of the subjects from which they are drawn. Note that I chose the four children randomly, with the only criteria of two females and two 5
In males. ccc020307, ccc020410 and ccc020624, the child subject seemingly did not use wo3 accurately in the adult sense wo3 should not be applied in this context where lacks meaning of reminder, negotiation, unfairness, blame, or new information. Neutral SFPs such as aa3 should be used instead in adult context. The child might yet have learned the correct usage of wo3. (ccc020800) *INV: ? ? ? ? ? *INV: ? ? ? ? ? ? ? *CHI: ? ? ? ? &a3. *CHI: ? ? ? ? ? ? &wo3. *INV: ? ? ? ? . *INV: ? ? ? ? ? ? ? *CHI: ? ? ? ? ? &lo1. Here in ccc0208000, the child subject uses wo3 and correctly expressed the meaning of reminder and negotiation. In respond to the interviewer's question "? ? ? ? ? ? ?", the child replied with "? ? ? ? ? ? &wo3", intending to tell and ask the mother not to hide a certain object. At this point the child has learned to correctly used wo3 as a SFP, and he proceeded to produce more correct usage of wo3 after this transcript. See more correctly used examples in transcripts ccc020907, ccc020923, ccc021013, (ccc020624) and ccc021027. *INV: ? ? ? ? &a3. *INV: ? ? ? ? *CHI: ? ? . *INV: ? ? ? ? ? ? &gaak3 ? *CHI: ? &wo3. *INV: ? ? *CHI: ? &a3.
3.4.2 Data Analysis on wo4
6
(ccc020507) *CHI: ? ? ? ? &a3. *INV: ? ? ? ? &a3. *INV: ? &a3. *CHI: ? ? ? ? ? &wo4 ? *INV: ? ? ? ? ? ? . *INV: ? ? .
(ccc020624) *INV: ? +... *INV: ? ? ? ? ? . *CHI: ? ? ? ? ? ? &a3. *CHI: ? ? ? ? &wo4. *INV: ? ? ? ? &a4 ? *CHI: ? .
In ccc020507, the child says "? ? ? ? ? &wo4" by expressing the unexpected fact that the light is not turned on. And in ccc020624, the child says "? ? ? ? ? ? &a3" first, then "? ? ? ? &wo4". Here, it is more than unexpectedness that the child is expressing. Rather, there is a sense of contrast by comparing sister going to play but not him 4. However, when the interviewer precedes to ask the child if he is going to play, the child reply with a yes. There is a contradiction in meaning here. It is possible that the child misinterprets the interviewer's question of "? ? ? ? &a4 ?" as asking if he wants to play. (ccc020817) *INV: ? ? ? ? &a3 ? *CHI: 0. %com: The telephone rings. The child's grandmother goes to answer the phone. *INV: ? ? ? ? ? ? &ga3 ? *CHI: ? ? ? ? ? &wo4. *INV: ? &wo3 . *INV: < ? > [//] ? ? ? ? ? &wo3 . Ccc020817 illustrates a correct usage of wo4. "? ? ? ? ? &wo4" is an answer replying the interviewer's question of "? ? ? ? ? ? &ga3", stating a sudden realization of not remembering something. That counts as a marker expressing surprise. See more correct utterances with wo4 in ccc020907 and ccc021013.
3.4.3 Data Analysis on wo5
4 "? ? " is the nickname of child subject coded ccc his Chinese name is Chan Chun-Yat. 7
There is only 1 child utterance of wo5 in the CANCORP and it is from Tinfaan (coded ltf).
(ltf030218) *SIS: ? ? &a3 ? ? . *INV: ? ? &lu3 ##. *INV: ? &a3. *CHI: ? ? ? ? ? ? ? &wo5. *INV: ? &a3. *INV: ? ? ? ? &la3. *INV: ? ? ? +/. *CHI: ? ? # ? ? ? ? . At the beginning of the excerpt above, the child's sister says "? ? &a3 ? ? ", and the child replies with "? ? ? ? ? ? ? &wo5". It is remarkable that the child uses wo5 for quoting what his sister has said earlier. Also note that this is the only utterance among eight children recorded, and it happened when the child subject was three years, two months, and eighteen days old. It is relatively rare and late comparing to other SFPs.
3.4.4 Supporting Data
From the table showing the distribution of SFPs in another child subject (coded MHZ) according to types of preceding utterance produced by adults (see appendix entry A), it provides affirmation towards the initial assumption of wo3 appearing before wo4 and wo5 aside from the data listed above.
4. Findings 4.1 Findings on wo3
8
Wo3 has the earliest occurrence among the three wo's, but it is also with lower accuracy. Misuse of wo3 happens more frequently than that of wo4 and wo5. It could be due to (i) probability as wo3 is used more, the chances of getting it wrong is also higher; or (ii) wo3's early emergence children are not fully capable in applying SFPs even the neutral non-interrogative ones at early stages. As a side note, I realized that except reminder, negotiation, unfairness, blame, and introduction of new information, wo3 could also express the surprised, unexpected meaning. (Note that I did not get this observation from the CANCORP, and there is no relevant example from the children.) E.G. ? ? wo3, ? ? ? ? ? ? ? Down rain SFP, you PL still go walk hill "It's raining, how come you guys are still going hiking?"
4.2 Findings on wo4
Only one instance of wo4 was found in one child before 2;0; while some spontaneous uses were found in the other two children at 2;5. Three-year-old children used this particle sparsely. wo4 is more productive after 3;0 some children did not acquire wo4 until three-years-old. It is possible that children acquiring wo4 later because (i) the usage of wo4 requires the complex task of assessing hearer's knowledge and (ii) it often co-occurs with other surprise interjections and degree intensifiers EG. ? ? ? , ? ? ? ? ? wo4. Note that adult input should not be contributing to the late emergence of wo4. In sum, wo4 is used by children to a limited extent From the analysis in 3.4.2 on wo4, other than an unexpected expression, wo4 also occurred to mark contrasts. However, I cannot find or come up with any examples in which wo4 is used to indicate contrasts, and I am uncertain to define the usage cited in 3.4.2 as a wrong usage as a native Cantonese speaker, it does not sound bad or wrong to me. This is left for discussion.
4.3 Findings on wo5 Among the three wo's, wo5 is the least used in the CANCORP, and it is not just in the children's speech, but also in the adult's. There are several possibilities why wo5 is not productive: (i) due to children's representational complexity, (ii) contexts that require quotative speech are more restricted (unlike wo3 which expresses various meanings) and rarer, and (iii) insufficient positive adult input. Note that children do possess the concept of reported speech as they will say something like "? ? ? ? ? ? ? ".
9
5. Conclusion
In respond to the issues raised in 2.2 research focus, even with occasional errors, the data shows really clearly of children's early sensitivity to functional categories in this case wo3, wo4, and wo5. And from analyzing the CANCORP, undoubtedly SFPs are one of the first functional categories to emerge in Cantonese children. Neutral SFPs do appear in child language earlier than interrogative and non-neutral SFPs as described in Lee, et al (1995). Also, through the process of researching, I see there might be more that the three wo's could represent other than meanings listed by Leung (2005) and Law (2000). It is amazing to discover how intricate young children's speech is contrary to most people think; in this case that early child has the knowledge of inventory of functional syntactic category. Observational results are in favor of continuity hypothesis that the principles that the child uses to fix her/his grammar are constant over the course of development in child first language acquisition.
6. Appendix A.
10
11
6. References 1. Lee, Thomas H.T., Colleen H Wong, Samuel Leung, Patricia Man, Alice Cheung, Kitty Szeto, and Cathy S P Wong. 1996. The Development of Grammatical Competence in Cantonesespeaking Children. Report of RGC earmarked grant 199194. 12
2. CANCORP [2011 version]. Chinese TAG version. http://www.cuhk.edu.hk/lin/langacq/CANCORP2011_Big5.rar 3. Language Acquisition Laboratory, CUHK http://www.arts.cuhk.edu.hk/~lal/index.php?id=9&lang=1#CANCORP 4. Chao, Yuen Ren (1968). A Grammar of Spoken Chinese. Berkeley: University of California Press. p. 149. 5. Fox, Barbara (2007), Principles shaping grammatical practices: an exploration. Discourse Studies 9, p. 303 6. Lee, Thomas H.T., Colleen H Wong (1996). Cahiers de Linguistique - Asie Orientale 27 (2): 211-228. 7. MacWhinney, B. (2000) The CHILDES project: Tools for analyzing talk. Third Edition. Mahwah, NJ: Lawrence Erlbaum Associates. 8. Chao, Y. R. (1947). Cantonese Primer. Cambridge, Massachusetts: Harvard University Press. 9. Fung, Roxana S.-Y. (2000). Final particles in standard Cantonese: semantic extension and pragmatic inference. Ph.D. Dissertation, the Ohio State University. 10. Luke, Kang Kwong. 2000. Sentence particles in Cantonese. Workshop on Cantonese particles. Hong Kong Polytechnic University. 11. Cheung, Hung-Nin. (1972). Cantonese as spoken in Hong Kong. The Chinese University of Hong Kong. 12. Maddieson, Ian. (1978). Universals of tone. In JH Greenberg (Ed.), Universals of human language: Phonology (Vol. 2). Stanford: Stanford University Press. 13. (2005). . , . 14. Pinker, S. (1987) The bootstrapping problem in language acquisition. In B. MacWhinney (Ed.), Mechanisms of language acquisition. Hillsdale, NJ: Erlbaum.
13