A Balkan View on the Left Periphery: Modal and Discourse Particles

Round 1

Reviewer 1 Report

The paper discusses the future and the subjunctive markers and the discourse marker haide in the Balkan languages with respect to their syntactic properties. It is well-structured, analytical and supported by the relevant references. There are, however, some minor inaccuracies that should be addressed:

p. 4

…in Greek the future reading arises given the -past +perfective specification of the verb…

In fact, the -past / -perfective (tha doulevo) is a future reading too. The example (9b) in Bulgarian (shte rabotja) is imperfective. In both languages the future interpretation is +/-perfective. In Albanian and Romanian there is no morphological marking for aspect, so the generalization this holds for the other Balkan languages is not correct.

p. 5

It should be stressed that the dissimilarities shown in examples (10) are a result of the diachronic development and the stage it has achieved in each of the mentioned languages and dialects, for example the omission or the preservation of the subjunctive marker.

p. 5

It should be noted that Serbian/Croatian is not a part of the Balkan Sprachbund and subsequently remains peripheral in the Balkan convergence processes. The morpho-syntactic properties of the future are another proof of its peripheral status.

p. 8

these are  verbal elements, either inflected (Albanian do is 3rd person for example)...

Albanian do is inflected when used as a full verb:

unë dua, ti do, ai do, ne duam...

As a grammaticalized future particle it is not inflected:

unë do të shkoj, ti do të shkosh, ai do të shkojë...

To be compared also with Bulgarian:

az ne shta, ti ne shtesh, toy ne shte...

az shte otida, ti shte otidesh, toy shte otide...

In section 3, Bulgarian is poorly represented, while Albanian is not even mentioned.

https://sq.wiktionary.org/wiki/hajde

In my opinion, the paper will improve if the Bulgarian and the Albanian discourse markers are covered by the analysis.

Author Response

Please see the attached

Author Response File: Author Response.docx

Reviewer 2 Report

The paper is dedicated to an old-standing problem of Balkan linguistics, but the author tries to enrich the theoretical reasoning by relating data from two apparently unconnected domains - future particles and discourse markers, arguing for the existence of a left peripheral projection related to Speech Act. The paper brings up various examples (some well-known, others new) from the secondary literature on Balkan languages. 

I raise here a number of points that the author may want to consider to strengthen his/her article. The issue of whether the future particles are instantiations of Mood or part of the left periphery should perhaps be argued for in more detail.  Also, some motivation could be given against treatments that relate these particles to clause-internal syntax: the literature contains arguments against a left peripheral accommodation of these particles, such as the relative positions of the subject, of various types of adverbs, and clitics, preceding the future periphrasis.

Concerning the mono-clausal vs. bi-clausal analysis of these clauses, the author says that the subjunctive particle is also part of the left periphery and that a restructuring context is created with the subjunctive particle acting as some sort of linker. In what sense the particle acts as a linker? More generally, what elements count as linkers in Balkan grammars? For Manzini and Savoia (2018) they count as a D element in Albanian, but does this reflect the general situation in all Balkan languages?

Also, which positions do linkers occupy in the left periphery? On p. 7, the author states that "the subjunctive particles are not inherently modal but derive their modality from their presence in the left periphery". Again, this statement is left without a hard and fast motivation and while it is true that subjunctive particle may also correspond to an infinitive in languages with infinitives, their presumable position within the CP area does not really make it clear to the reader why they are what they are.  As to the organization of the left periphery itself, specific reference to a particular theoretical framework (Rizzi’s 1997, or Giorgi’s 2010 About the Speaker Oxford Univ. Press), should be made clear, apart from the Speech Act projection which is supposedly higher than the rest of the CP. This assumption, as well as the works it is based on, seems to conceive the left periphery as a more fine-grained space, yet on p. 7, the author rejects the idea that Force and Fin correspond to Linker and M. The explanation that follows raises the following questions: complementizers are said to be "on top of the left periphery" (?) and to occupy a distinct position with respect to subjunctive particles, but what kind of positions are these? 

Interesting is the part dedicated to haide (and variants) and the relation between discourse markers and the left periphery. While it is true that this discourse particle has been borrowed from Turkish, it is not really clear what leads the author to conclude, on p. 12, that Balkan languages "have not only borrowed a lexical item from Turkish but have also adopted and adapted its syntactic properties". No real discussion follows about the syntactic properties of this marker in Turkish so the conclusion that it has been adapted to the specific requirements of the Balkan left periphery is left hanging in the air.  Also, it is not entirely clear that haide is an imperative verb, as also pointed out by Hill (2013b), cited in the paper. Even more dubious is that it occupies the SpeechAct position dedicated to the Hearer (Addressee?), given that unlike the highest verb in other "serial verb" constructions cited in the paper, it does not act as a host for clitics. In general, what does it mean for a projection to be on top of the left periphery? Is it still part of sentence grammar or does it belong to discourse grammar? Since other imperative verbs in a serial verb construction may act as clitic hosts, the conclusion on p. 13 that "whatever analysis accounts for (29) could perhaps extend to these elements [haide/ande] as well" remains dubious. A more serious discussion of these issues from a theoretical point of view would be highly useful for the reader. 

As regards section 4, although it summarizes some general statements known from previous works on Balkan linguistics, it is not clear to me how the E-language vs. I-language distinction bears on the issue of Balkanisms, as well as on the issue of surface similarities between the Balkan languages. The latter are well-known to share a "deep grammar" that has been shaped in language contact (contact-induced grammaticalization) but the way the author presents the issue opens more questions than those it tries to solve. In fact, it is not clear to me how language contact might have worked in reference to the precise phenomena under consideration. 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

The review is attached

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The paper has been significantly improved. I think it needs a final reading as a whole text.

Author Response

I would like to thank the reviewer for the positive review.

I have addressed the points raised below by providing the relevant clarifications.

Also following Reviewer’s 3 comments, I have modified the structure in (33a-b) and further elaborated on the structure suggested in (32a-b) in terms of Chomsky’s (2020) mechanism of Pair-Merge.

Main changes are highlighted (yellow) in the revised version.

Reviewer 2 Report

The new version of the article is considerably improved with respect to the original version. The framework has also been elaborated and the main contributions of the paper stand out more clearly now especially those regarding 'haide'.  

I have only one last question for the author relative to the structure in 21: What are the arguments that the modal particles of Bulgarian (da) and Romanian (să) are suitable for being merged under D like their Albanian and Greek counterparts? I understand from the author's response that D is being used as a cover term synonymous to 'linker' but still I find the use of the D  label problematic because subjunctive complements do not have the distribution of DPs (some of them cannot even select a nominal complement indicating the presence of a D projection). The author might want to add a footnote to clarify this issue. 

I also found two new references (Giorgi 2010, Sonnenhauser and Wiemer 2019) missing from the References list.      

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

attached

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.docx