Skip to main content

Advertisement

Log in

Force shift: a case study of Cantonese ho2 particle clusters

  • Published:
Natural Language Semantics Aims and scope Submit manuscript

Abstract

This paper investigates force shift, a phenomenon in which the canonical discourse conventions, or force, associated with a clause type can be overridden to yield polar questions with the help of additional force-indicating devices. Previous studies attribute force shift to the presence of a complex question force component operating on semantic content. Based on utterance particles and particle clusters in Cantonese, we analyze force shift as resulting from compositional operations on force-bearing expressions. We propose that a simplex force, such as assertion or question, denotes unanchored sentence acts, while a force-shifting particle like Cantonese ho2 is an anchoring function anchoring a sentence act to the speaker while querying whether or not the addressee can perform the sentence act. The proposed semantics makes predictions about ho2’s interactions with addressee-changing operations and imperatives, as well as about a larger family of force shift phenomena.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7

Similar content being viewed by others

Notes

  1. This working definition excludes indirect speech acts, such as rhetorical questions (e.g., Is the Pope Catholic?) and questions taken to be commands (e.g., Can you pass the salt?).

  2. All Cantonese sentences in this paper are given in Jyutping, a romanization system developed by the Linguistic Society of Hong Kong. Particles under investigation are set in boldface and are glossed using force/act-level terminology (like assertion and question) rather than sentence-type terminology (like declarative and interrogative) because there are many more particles in Cantonese than sentence types traditionally recognized in the literature. The convention used for glosses is as follows: asp = aspect marker, asrt = assertive particle, bpq = polar question with a negative bias, polq = polar question particle, up = utterance particle, whq = wh-question particle.

  3. A reviewer points out that a brief pause is allowed, but not required, between ho2 and the preceding particle. Similar observations are also made in Tang (2020). There are a few possible reasons for the optional pause. First, it may be due to the fact that ho2 begins with a glottal consonant, which does not have any constriction in the oral cavity. If ho2 is said quickly, the consonant [h] is very weak. The lower particle and ho2 almost sound like one particle. If one wants to clearly indicate that there are two particles, then it is natural to insert a pause before ho2, much like inserting a glottal stop between two vowels. Second, the pause may also be semantically/pragmatically motivated. As is argued later in this paper, ho2 anchors a sentence act to both the speaker and the addressee, which is more complex than simple anchoring. The optional pause may be used to draw the hearer’s attention to this complexity.

  4. There are other varieties of Cantonese spoken in other regions with slightly different inventories of utterance particles and their clusters. We reserve potential dialectal variations for future research.

  5. The four speakers all grew up and lived in Hong Kong and were aged between 28 and 40 at the time of data collection, between 2018 and 2019. They were friends and relatives of the first author and were recruited through her personal network. Their participation was voluntary and not compensated. The judgements were collected in separate elicitation sessions, each with the first author as the interviewer and one of the informants as the interviewee. In each session, the interviewer presented the contexts and target sentences, group by group, in written Cantonese and spoken Cantonese. She then asked the interviewee to offer a felicity judgment for the target sentences in each group. A sentence was reported as ‘grammatical’, ‘marginal’ (prefixed with ‘?’), or ‘ungrammatical’ (prefixed with ‘*’) if there was consensus from more than half of the five involved parties. A sentence was reported as ‘infelicitous’ (prefixed with ‘#’) if it was judged as infelicitous in the given context but otherwise grammatical. .

  6. It is unclear to us why gaa3 and aa3 differ in their compatibility with aspectual classes. Since both particles can form particle clusters with ho2, we assume that their aspectual selection properties do not interact with force shift. Another particle closely related to gaa3 is ge3, which is also regarded as a declarative particle. It is often conjectured that gaa3 is itself a cluster involving ge3 and aa3. However, unlike gaa3-ho2, which is accepted by all informants we have consulted, a reviewer pointed out that some speakers find ge3-ho2 degraded. We searched Hong Kong Cantonese Corpus, credited to Luke and Wong (2015), and confirmed that the cluster ge3-ho2 is indeed non-existent. However, the cluster is reported to be acceptable in Matthews and Yip (2011). We acknowledge the inter-speaker variation but leave it open in this paper.

  7. Some have expressed doubts towards treating ne1 as a question marker. For example, Law (1990) analyzes it as a marker of tentativeness.

  8. A reviewer suggests that maa3 is less common in Hong Kong Cantonese than Guangzhou Cantonese. Sybesma and Li (2007) speculate that maa3 may be borrowed from Mandarin.

  9. Maa3 and me1 are the primary polar question particles in Cantonese. The general purpose particle aa3 can also be used to mark polar questions and is discussed in Sect. 4.2.

  10. What is not possible, we think, is an analysis in which simplex particles are treated as content-level complementizers, such as treating ne1 as an interrogative complementizer, à la Karttunen (1977), that turns a proposition into a set of propositions. Our refutation of this analysis is based on the observation that these particles resist embedding environments.

  11. This question can be felicitous if Ziming tried to be sarcastic and pretend that he is committed to the incorrect answer. This does not challenge the generalization that gaa3-ho2 preserves speaker commitment. Rather, it shows that the assertive force associated with gaa3-ho2 is very similar to the assertive force of ordinary assertions in the sense that both can give rise to sarcasm.

  12. Maa3-polar questions and aa3-questions are generally not acceptable with ho2. However, the acceptability can be remedied by a mechanism known as addressee shift, which we discuss in Sect. 4.2.

  13. A reviewer suggested that questions operated by ho2 may be related to conjectural questions. Conjectural questions, as defined in the literature (see Eckardt 2020 for an overview) have a very distinct profile. The asker of a conjectural question does not expect the addressee to know the answer and thus does not request an answer. Consequently, the addressee can remain silent without violating the rules of discourse. Ho2-questions, on the other hand, have an opposite effect on the discourse. They mandatorily require a response from the addressee, who is expected to know the answer and also respond. Another point of difference between conjectural questions and ho2-questions is that the former is anchored to the propositional content, while the latter is a higher level question not about a propositional content but about a sentence act.

  14. The use of a ne1-ho2 question in this context would be felicitous if Ada were trying to get Beth to explain the content of the talk without admitting that she thought Beth was more knowledgeable than her, which would be a discourse effect of using a ne1-question. .

  15. This is assuming Ada did not have the obnoxious intention of embarrassing Beth with her inability to answer the ne1-question. A reviewer also points out that some speakers would accept (43-a) if ne1 were replaced by aa3. This is likely because the question with aa3 can be used as a way to complain that the talk wasn’t delivered clearly. We leave open the question of why ne1 does not have the same complaint use.

  16. As expected, a negative answer to (48) is also possible when accompanied by an appropriate final particle, as shown below:

    1. (i)
      figure aq
  17. Note that Krifka’s system does not make use of tuples, as we do in this paper, to distinguish between different discourse participants’ commitment sets.

  18. This is likely too strong. A weaker alternative requirement is that p has not been recently added to the speaker’s discourse commitment set.

  19. We define \(\mathsf{Dox}^{c}_{x}\) and ¬p as follows: \(\mathsf{Dox}_{x} := \{p \mid x\text{ believes }p\text{ in }c\}\); ¬p: = {wwp}.

  20. Despite the connection, our definition of questions is different from the commitment space approach of Krifka (2015) and related studies. For Krifka (2015), a question gives rise to a new commitment space with an original root. For us, a question gives rise to an issue, which is a set of commitment spaces.

  21. Not all studies model polar questions as a set of two possibilities. For example, Bolinger (1978), Gawron (2001), van Rooy and Safarova (2003), Biezma (2009), Biezma and Rawlins (2012), Roelofsen and Farkas (2015), and Bhadra (2020) treat polar questions as consisting of a singleton answer, while earlier studies like Hamblin (1973) treat them as consisting of both positive and negative answers. We are not committed to a particular treatment of polar questions in this study. If polar questions turn out to be more amendable to a singleton analysis, the analysis proposed here can be recast along the lines of the singleton approach.

  22. Following Krifka’s (2013) suggestion about English response particles like right or okay, we assume that the affirmative morpheme hai in Cantonese can refer to speech acts.

  23. In exceptional circumstances such as one in which the addressee suffers from amnesia, this question would be deemed acceptable.

  24. Like ne1, aa3 may also be used in A-not-A questions and alternative questions.

  25. Ne1-questions can be used as self-directed questions, which as a speech act are felicitous when the answer is not known to the speaker (see Garrett 2001 for Tibetan; Murray 2010 for Cheyenne). Based on these cross-linguistic facts, Bhadra (2020) analyzes self-directed questions and rhetorical questions as being speech acts that do not raise issues, unlike true information-seeking questions. Ne1 questions are thus compatible with being both information seeking and non-information seeking questions.

  26. The definition in (80) needs to be slightly modified for maa3 so that it only yields a polar question.

  27. This solution is not without problems. For one thing, it is well known that rhetorical questions may admit only one answer. However, they are acceptable in many languages, including Cantonese.

  28. It is worth pointing out that declarative aa3 also cannot be used when there is no addressee. This suggests that it may also specify the addressee like the interrogative aa3. However, declarative aa3 is nonetheless compatible with ho2. We take this to indicate that it is not merely the addressee specification that is problematic for ho2. Rather, it is the interaction of the addressee specification and the definedness condition of an unanchored sentence act that causes issues for ho2. We reserve the role of the addressee specification in assertions for future research.

  29. Imperative clauses may admit a range of markers, include aa3 and laa1. As noted earlier, aa3 may appear in different types of clauses. We gloss it as imp based on its environment.

  30. Alternatively, the speaker parameter may still be present but is just not used in anchoring the sentence act associated with the lower force. It may still be involved in signaling that the higher polar question act is performed by the speaker.

  31. That said, there is some initial evidence, from Jeong (2018), that inquisitive and non-inquisitive rising declaratives may have slightly different intonational patterns.

References

Download references

Acknowledgements

We would like to thank Pranav Anand, Justin Bledin, Simon Charlow, Gennaro Chierchia, Kate Davidson, Veneeta Dayal, Jim Huang, Manfred Krifka, Kyle Rawlins, Paul Portner, Jessica Rett, Deniz Rudin, Sze-Wing Tang, two anonymous reviewers of Natural Language Semantics, and the editor Florian Schwarz for their valuable comments and suggestions. We also thank the audiences at SuB-22, NELS-48, Harvard Meaning and Modality Group, UCLA Linguistics Colloquium, and Georgetown Semantics Reading Group. We extend our gratitude to Lisa Chan and Wesley Law for data judgments. This research received financial support from the Start-Up Grant of Nanyang Technological University.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Haoze Li.

Ethics declarations

Competing Interests

The authors have no conflicts of interest.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Law, J.HK., Li, H. & Bhadra, D. Force shift: a case study of Cantonese ho2 particle clusters. Nat Lang Semantics (2024). https://doi.org/10.1007/s11050-023-09219-8

Download citation

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s11050-023-09219-8

Keywords

Navigation