Whys and Wherefores: The Aetiology of the Left Periphery (With Reference to Vietnamese)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. (Speaker-Oriented) Adverb Placement: The left-peripheral adverbials discussed by the author are mostly speaker-oriented and, should, in theory, take scope over the entire proposition. This implies that they should be located within the left periphery so as to scope over the proposition. The data in (55–57) demonstrate that these adverbials can either precede or follow the subject (the latter order is preferred). To explain this, the author adopts a movement approach. However, since the topic marker in Vietnamese is optional, an alternative analysis may be possible. Specifically, could these adverbs be base-generated before the subject, while their post-subject position is due to the topicalization of the subject? Note that the examples in (55-57) are root clauses, not embedded ones.
2. Anchoring the subject: Ko (2005) demonstrates that certain Chinese Scope-Bearing Elements (SBEs) (e.g. meiyouren ‘nobody’, henshao ren ‘few people’, zhiyou NP ‘only NP’, zuiduo liang-ge ren ‘at most two people’) cannot undergo topicalization as shown in (42) below. Because these SBEs are obligatorily anchored in the subject position, they serve as diagnostic tools for determining the surface position of in-situ why in Chinese. Specifically, in her (36) below, the ungrammaticality of (36a) is attributed to the fact that weishenme ‘why’ must be base-generated above the SBEs.

(From Ko, 2005)

(42) {*Meiyouren₁/*henshaoren₁/*zhiyou lisi₁/*zuiduo liang-ge ren₁},

      {Nobody/few people/only Lisi/at most two people}

     Zhangsan shuo [(ta₁/tamen₁) hen congming]

     Zhangsan said (she/they) very smart

     ‘{Nobody1/few people1/only Lisi1/at most two people1} Zhangsan said that {she1/they1} {is/are} very smart.’

(36) a. *{Meiyouren/zhiyou Lisi/henshaoren} weishenme cizhi?

     b. Weishenme {meiyouren/zhiyou Lisi/henshaoren} cizhi?

     Why {nobody/only Lisi/few people} resign

     ‘Why did {nobody/only Lisi/few people} resign?’

If a similar phenomenon is observed in Vietnamese, could it be used to determine the syntactic position of speaker-oriented adverbs? Specifically, is the SBE–Adverb or Adverb–SBE order acceptable in Vietnamese, considering that SBEs occupy the subject position? If SBEs remain fixed in their base-generated positions, their interaction with speaker-oriented adverbs may provide a useful diagnostic for examining whether these adverbs undergo movement or are base-generated in the left periphery.

Ko, Heejeong. 2005. Syntax of Why-in-situ: Merge into [Spec,CP] in the overt syntax, Natural Language & Linguistic Theory 23 (4): 867-916.

3. The Status of ‘People’ in (81): The noun ‘people’ in (81) appears to be a generic noun. A crucial question is whether such generic nouns can undergo topicalization in Vietnamese. If they cannot, it follows that tại sao ‘why’ can only precede ‘people’, as topicalization of ‘people’ would not be an option. This restriction could provide additional support for the syntactic positioning of tại sao which must be structurally higher than generic noun subjects.
4. (64b): The ungrammaticality of (64b) may be attributed to a semantic requirement in topic constructions, where a larger whole is introduced as an Aboutness Topic (AT), followed by a reference to a smaller part of that whole. If reversed, such an order may violate the expected part-whole discourse relation, leading to an unnatural interpretation.
5. {...TOP…SUBJ…QP∀…} and the Position of QP∀: The author states that “The acceptability of this word order is obviously inconsistent with the idea that QP∀-movement in Vietnamese involves raising to a dedicated Foc position, …” However, suppose we adopt Belletti’s (2004) proposal of a low periphery within the IP domain (above vP and below IP). In that case, the post-subject QP∀ may occupy an IP-internal FocP rather than a CP-internal one.

Belletti, A. 2004. Aspects of the low IP area. In L. Rizzi (ed.), The Structure of CP and IP. New York and Oxford: Oxford University Press, 16-51.

6. The postulation of ReasonP: The author follows Shlonsky & Soare (2011) in assuming a ReasonP to host tại sao ‘why’. However, a recent paper by Yang (2021) challenges the existence of ReasonP in embedded contexts in Chinese. Specifically, he demonstrates that Chinese ‘why’ cannot be embedded under matrix control verbs which select nonfinite complements, as shown in (18) (Yang’s example):

(18) Ni dasuan   a. song shei?

      you plan         give who

                           b. song shenme?

                           give what

                           c. shenmeshihou song?

                           when give

                           d. zenme song?

                           how give

                           e. *weishenme song?

                           why give

          ‘Who/what/when/how/*why do you plan to give?’

Note, in particular, that the above examples differ from S&S’s paradigm in that these sentences are direct questions. Under S&S’s framework, weishenme ‘why’ in Chinese should first merge in an embedded ReasonP before undergoing LF movement to the matrix IntP to derive matrix scope. However, the ungrammaticality of (18e) suggests that such a derivation is impossible. Yang (2021) argues that the notion of ReasonP should be dispensed with in Chinese and that Rizzi’s IntP alone should be sufficient enough for deriving the scope of ‘why.’ Given this observation, I wonder whether a similar reasoning can be applied to Vietnamese.

Yang, Barry C.-Y. 2021. On the position of ReasonP, International Journal on Chinese Linguistics 8(2): 291-313.

Author Response

Comment 1.  (Speaker-Oriented) Adverb Placement: The left-peripheral adverbials discussed by the author are mostly speaker-oriented and, should, in theory, take scope over the entire proposition. This implies that they should be located within the left periphery so as to scope over the proposition. The data in (55–57) demonstrate that these adverbials can either precede or follow the subject (the latter order is preferred). To explain this, the author adopts a movement approach. However, since the topic marker in Vietnamese is optional, an alternative analysis may be possible. Specifically, could these adverbs be base-generated before the subject, while their post-subject position is due to the topicalization of the subject? Note that the examples in (55-57) are root clauses, not embedded ones.

Response 1. I have tried to make sense of this comment, but I feel that one of us may have misunderstood something. The data presented throughout the paper demonstrates that Vietnamese does not allow double topicalization, neither in main nor embedded clauses. Given this, the examples in (57), which show marginal Top-Adv-Subject order and fully acceptable Top-Subj-Adv order, would seem to demonstrate (i) that Subj-Adv order does not arise through topicalization (ii) the (to the reviewer) expected Adv-Subject is marked relative to Subj-Adv order. This is consistent with the analysis in (58), (even if I am personally sceptical of the existence of a dedicated Mod position). Furthermore, if the topic marker were merely optional we would expect the double-marked equivalent of (57b) to be acceptable: to my knowledge, it is not.

*Những cuốn sách đó] thì cô ấy thì nhất định phẚi đọc.

Now, as I point out in the revised conclusion, this does not preclude the possibility that the canonical position of subjects is in a position higher than IP, though not as high as TopP. I agree with the reviewer's theoretical intuition that speaker-oriented adverbs should have scope over the underlying proposition (including the subject), hence the facts in (55)-(58) are puzzling. But facts they seem to be.

Comment 2.  Anchoring the subject: Ko (2005) demonstrates that certain Chinese Scope-Bearing Elements (SBEs) (e.g. meiyouren ‘nobody’, henshao ren ‘few people’, zhiyou NP ‘only NP’, zuiduo liang-ge ren ‘at most two people’) cannot undergo topicalization as shown in (42) below. Because these SBEs are obligatorily anchored in the subject position, they serve as diagnostic tools for determining the surface position of in-situ why in Chinese. Specifically, in her (36) below, the ungrammaticality of (36a) is attributed to the fact that weishenme ‘why’ must be base-generated above the SBEs...[cont]

Response #2. As I understand it, this is largely restatement of the same concern as in #1, namely, that both AdvPs and why-adverbials should take scope over "anchored" subjects, as they seem to do in Chinese. Whether Vietnamese behaves in the same way as Chinese in having "unmoveable" SBEs is an interesting empirical question, deserving investigation (and which should shed light on Comment #1). Unfortunately, time constraints at this stage preclude further progress beyond a footnote (fn. 52).

Comment #3: The Status of ‘People’ in (81): The noun ‘people’ in (81) appears to be a generic noun. A crucial question is whether such generic nouns can undergo topicalization in Vietnamese. If they cannot, it follows that tại sao ‘why’ can only precede ‘people’, as topicalization of ‘people’ would not be an option. This restriction could provide additional support for the syntactic positioning of tại sao which must be structurally higher than generic noun subjects...'
Response #3. As much as this is an interesting question, but I don't understand why it is in any way crucial. The examples in (86) show that tai sao preferentially occurs above the topic position, with regular topicalized arguments, and marginally between the Top and Subj; those in (89) confirm that Top-Subj-tại sao is excluded with thematic subjects. Thus, it is unclear how evidence of the non-topicalizability of generic nouns would cause me to change anything about the claims made here.

Comment #4. 64b): The ungrammaticality of (64b) may be attributed to a semantic requirement in topic constructions, where a larger whole is introduced as an Aboutness Topic (AT), followed by a reference to a smaller part of that whole. If reversed, such an order may violate the expected part-whole discourse relation, leading to an unnatural interpretation.
Response #4. I have added a footnote to this effect (fn. 56).

Comment # 5.  {...TOP…SUBJ…QP∀…} and the Position of QP∀: The author states that “The acceptability of this word order is obviously inconsistent with the idea that QP∀-movement in Vietnamese involves raising to a dedicated Foc position, …” However, suppose we adopt Belletti’s (2004) proposal of a low periphery within the IP domain (above vP and below IP). In that case, the post-subject QP∀ may occupy an IP-internal FocP rather than a CP-internal one.

Response #5. I feel that this is repeating the same general objection that I have tried to address in the earlier responses. Moreover, even if there were a lower Mod position in IP, after Belletti (2004), it would not, I would suggest, explain the variability in the position of the "QP" (all that matters is that this expression is above cung!)

Comment #6. The postulation of ReasonP: The author follows Shlonsky & Soare (2011) in assuming a ReasonP to host tại sao ‘why’. However, a recent paper by Yang (2021) challenges the existence of ReasonP in embedded contexts in Chinese. Specifically, he demonstrates that Chinese ‘why’ cannot be embedded under matrix control verbs which select nonfinite complements, as shown in (18) (Yang’s example):...
Response # 6. I do not fully understand the reviwer's point here, unless it is to suggest that Chinese is simply different in this respect (which is of course entirely possible). The reviewer suggests that in Chinese ReasonP can be eliminated in favour of obligatory movement to Int. Yet the point of the examples in this section (84)-(89) is precisely to demonstrate that Vn tại sao cannot raise as high as Int; hence the necessity of ReasonP or some other node independent of Int. (If I have missed something here, it will I hope provoke a response, and progress will be made). I have included reference to Yang's work as a footnote (fn. 72).

Reviewer 2 Report

Comments and Suggestions for Authors

Review of “Whys and Wherefores: the Aetiology of the Left Periphery (with reference to Vietnamese)”

 

Submitted to Languages special issue on Vietnamese

 

This is an interesting, data-rich, very well-written paper.  I strongly recommend its publication in Languages.  The paper provides a wealth of information on the organization of the higher structure of clauses (primarily) in Vietnamese and creates a very detailed baseline for future researchers working on the ‘left periphery’ in Vietnamese and other languages, a truly invaluable resource.

 

I have only a few suggestions for potential adjustments – some typos and some general questions that come to mind. The author should feel free to make use of these remarks entirely as they feel appropriate and is beneficial for the fine tuning of the paper.

 

p.7 It is suggested that Vietnamese has a richer inventory of pre-subject constituents than most other MSEA languages.  Are we really sure about this?  Or has the kind of work that the author has put into this study of Vietnamese just not been carried out in other MSEA languages, to the same depth?  Certainly there is variation in complementizers available in Thai, Lao and Burmese, and a certain amount of topicalization in all three (as well as in Malay), and Khmer has clause initial interrogative particles in main clauses.

 

p. 11 In the tree in (14), can overt topics be added into SpecTopP between rằng and liệu? Giving rise to the sequence ‘…rằng Topic liệu… If not, what might disallow this?

Similarly, in the structure in (54), can any elements such as adverbials occur between rằng, là and thì? 

 

p.15 Typo line 497: rare should be sembrare

 

p.29 Is there support for a movement analysis of the AdvP nhất định to ModP, rather than optional base-generation of nhất định in either a higher or a lower position? 

 

p.33 Is there a typo?  The text states that (65b) is fully acceptable, but example (65b) is presented with two question marks ??

 

p.33 Typo: Firstl → Firstly

 

p.40 Typo: ‘the position tại sao..’  → ‘the position of tại sao..’

 

p.42 Typo:  ‘..adverbials such as can appear..’  Something is missing between ‘as’ and ‘can’.

 

p.46 In the interesting discussion of differences between Italian and Vietnamese, the author might find the following paper useful.  It compares Italian, English and Chinese, and distinguishes between moved and non-moved topic types (the latter being multiple in theory, the former being unique due to intervention effects on movement).

 

Liao, Roger and Grant Hung-Ta Kao. 2023. Extraction asymmetries in topic structures: a

comparative analysis.  Journal of East Asian Linguistics 32:63-90.

 

Once again, I enjoyed reading this article and think it will be a great contribution to the Languages special issue on Vietnamese.

Author Response

I thank the reviewer for their very supportive comments and suggested corrections. Below, I briefly indicate how these have been addressed: 

1. p.7 It is suggested that Vietnamese has a richer inventory of pre-subject constituents than most other MSEA languages.  Are we really sure about this? ...
Response #1. Point taken. Please see fn. 16.

Comment #2. 

p. 11 In the tree in (14), can overt topics be added into SpecTopP between rằng and liệu? Giving rise to the sequence ‘…rằng Topic liệu… If not, what might disallow this?

Similarly, in the structure in (54), can any elements such as adverbials occur between rằng, là and thì? 

Response #2. Please see fn. 23. With respect to (54): as discussed below tai sao can intervene between là and thì (86a); however nothing can intervene between rằng and lieu.

Comments #3 Typos on pp. 33, 40 and 42.
Response #3: fixed (thank you)

Comment #4 ...p.29 Is there support for a movement analysis of the AdvP nhất định to ModP, rather than optional base-generation of nhất định in either a higher or a lower position? 
Response #4. No, there is no evidence that would force a movement analysis. The wording is deliberately vague here "might be concluded...consistent with Rizzi/Bocci..." If one were committed to the Rizzi/Bocci functional sequence, this analysis would model the facts appropriately. It should be clear by the end that I remain quite sceptical of dedicated positions.

Comment #5.  "p.46 In the interesting discussion of differences between Italian and Vietnamese, the author might find the following paper useful.  It compares Italian, English and Chinese, and distinguishes between moved and non-moved topic types (the latter being multiple in theory, the former being unique due to intervention effects on movement)...

Response #5. I have included the reviewer's remarks as fn. 46.