Formal Approaches to South Asian Languages (E-Journal)
Not a member yet
116 research outputs found
Sort by
Naãn as a Tag Question and a Discourse Marker in Hindi-Urdu
No full textThis study investigates the Hindi-Urdu particle naãn, which is ubiquitous in casual speech, but whose exact contribution has not been satisfactorily described. While some instances of naãn contribute an interrogative meaning, others are incompatible with one - raising the question of whether naãn is even a single lexical item. Following the diagnostic tests for biased questions developed in Farkas and Roelofson (2017) and Goodhue (2018), as well as the investigation into German discourse markers in Kaufmann and Kaufmann (2012), this paper proposes that naãn should, in fact, be analyzed as two separate lexical items, naãn1 and naãn2. This investigation reveals that naãn1 appears exclusively in clause-final position, and contributes a meaning very similar to English reversed-polarity tag questions (questions like you\u27re going, aren\u27t you?), where a declarative sentence anchor is followed by a tag with the opposite polarity of the anchor), while naãn2 appears clause-medially or after imperative verbs, is incompatible with an interrogative interpretation, and behaves similarly to the unstressed form of the German discourse marker doch, with the additional ability to contribute contrastive topicalization
Mora stress in Shina as Contrastive Foot Structure
Full text linkI argue that so-called ‘mora stress’ in Kohistani Shina (spoken in Northern Pakistan, realized as falling vs. rising accent) is best analyzed as a difference in the alignment of a LH* pitch accent to two types of feet (moraic vs. syllabic trochees). This paper offers a formalization of the mapping of (intonational) tones to foot structure and argues that independent evidence for a foot-based approach comes from a process of stress advancement, where stem stress predictably shifts to a following suffix when the final mora of the stem is accented; yet it remains on the stem when the stem accent is on a non-final mora. I end by briefly discussing typological and theoretical implications of our analysis, also drawing on a comparison with a similar advancement pattern in Lithuanian (known as de Saussure’s Law)
Deriving ignorance in questions: evidence from Sinhala
Full text linkThe epistemic component of a question has always been a crucial aspect in the analysis of the meaning of a question. However, while the meaning of questions has mostly been analyzed in terms of indirect questions embedded under the factive know, the ignorance component associated with a direct question has received very little attention in the literature. This paper highlights the significance of the ignorance component of a matrix question in the analysis of its meaning and proposes to account for it in terms of a presupposition in association with the indefinite component within a question. It develops and tests the hypothesis with some crucial data from Sinhala, an Indo-Aryan language spoken in Sri Lanka
Bengali Verb-stranding VP Ellipsis and Ellipsis Identity Conditions
Full text linkThis paper first presents additional evidence, on top of those in Simpson, Choudhury, and Menon (2013), for Verb-stranding VP Ellipsis (VVPE) in Bengali. It then demonstrates that the language violates the Verbal Identity Requirement (VIR) of Goldberg (2005), and then goes on to argue that the violation of the VIR is what should be expected, while the adherence of a language to it is what shouldn’t be. The paper concludes itself by suggesting that the VIR might very well be an artifact of more underlying interactions between syntax and phonology. 
Negation, Imperatives, and Agreement in Coorgi
Full text linkThis paper presents novel data from a severely understudied Dravidian language Coorgi showing that one of the negation morphemes in the language (-le negation) cannot occur with imperatives or with agreement. We show that another form of negation, -at, on the other hand, can occur with both. The paper explores possible reasons behind this pattern, and we propose that this distributional difference between -le and -at stems from a difference in their syntactic position. Namely, we argue that -le is head-negation (head of NegP), while -at is adjunct negation occupying an adjunct position (SpecVP). We show that this syntactic difference between the two types of negation in Coorgi places the observed patterns within a robust cross-linguistic generalization: negation is banned in imperative contexts in (some) languages where negation morphemes are syntactically heads, while negation is allowed in imperative contexts in languages where the negation morphemes are adjuncts (Zeijlstra 2004, Bos?kovic? 2004, 2012). We thus argue that the patterns of the distribution of negation in Coorgi is due to a cross-linguistically attested pattern of head Neg blocking Affix- Hopping. The proposed analysis provides an insight into another pattern of the distribution of negation in the language – namely, its cooccurrence or the lack thereof with agreement morphology
Semantic mismatch and microvariation in Telugu Psych-PC predicates: Event structure of \u27-ki\u27 and \u27-gaa\u27
Full text linkA well-known generalization in Dravidian languages is that Psych-Predicates need dative subjects in order to predicate. This paper discusses the data from Telangana Telugu (TT), which uses nominative subjects with psych-PCs and shows that datives are not obligatory in the presence of \u27-gaa\u27. It builds on Balusu\u27s (2016) idea of \u27-gaa\u27 as a [+eventive] Pred0 and explains its dynamic semantics via the subevental structure within Ramchand\u27s (2008) First Phase Syntax (FPS). Finally, it explains the experiencers in TT which use both \u27-ki\u27 and \u27-gaa\u27 and yet give dynamic semantics using dative incorporation
Modal-Aspect interactions in Bangla
Full text linkBhatt (1999) observed that ability modals crosslinguistically carry an actuality entailment when they appear with the perfective aspect (e.g., the English John was able to swim the English Channel (on the episodic reading of the sentence) entails that John did, indeed (manage to) swim the English Channel). The imperfective counterpart, however, lacks this entailment. Hacquard (2006, 2009) showed that Bhatt’s generalization is insufficiently general in that the actuality entailments appear with all root modals, not just ability modals, independently of whether they are possibility or necessity modals. This paper presents novel data from Bangla where modals can take the progressive as well as the perfective aspect morphology, and also carry an actuality entailment, that asserts the actual existence of a process rather than the culmination. I show that this can be explained on Hacquard’s general account of Modal-Aspect interactions if one adopts a perspective on the semantics of the Progressive that has a component of its meaning anchored in the world of evaluation over and above its modal component. The actuality entailment with the Perfect also follows on particular assumptions about its meaning
Singular tum is not plural: a Distributed Morphology analysis of Hindi verb agreement
Full text linkHindi has a three-way honorificity contrast in the second person: low tuu vs. mid tum vs. honorific aap, and a two-way contrast between non-honorific and honorific DPs in the third person. Honorific DPs are said to be formally plural as they always trigger plural agreement, regardless of semantic number. In this context, I consider the formal number features associated with the non-honorific pronoun tum. Prior work has claimed that like honorific DPs, tum always bears formal plural features. This is motivated by the fact that in many cases, tum takes apparent plural morphology, regardless of semantic number. However, Bhatt & Keine (2018) note a puzzling exception to this generalization: tum takes the feminine singular affix -ii when semantically singular, and the feminine plural affix -i?i? when semantically plural. I account for this puzzle by assuming that only DPs that are honorific or semantically plural bear the formal plural feature. Since tum is not honorific, it does not bear this feature when it is semantically singular. I show that apparent plural morphology associated with tum can be accounted for if we assume this morphology is actually underspecified for number. The analysis is couched within a Distributed Morphology framework
The Lawngtlang Zophei verbal complex
Full text linkKuki-Chin languages offer complex and innovative agreement systems within Tibeto-Burman which make use of pre- and post-verbal subject and object agreement markers. The choice of verb paradigm often varies based on valency, mood, and polarity, among other factors. Outside of current work at Indiana University, Zophei is undescribed. This research contributes to the literature on Kuki-Chin agreement by offering a description of the verbal complex of Lawngtlang Zophei based on the intuitions of (co-author) Zai Sung, a native speaker. The description includes subject and object agreement paradigms for contexts varying by valency (transitive, intransitive), polarity (negative, affirmative), and mood (declarative, interrogative, imperative). In addition, the plural subject/object marker, directional, future, negative, and question markers are discussed along with the order of morphemes within the verbal complex. 
Two types of pluractionality within Kannada verbal reduplication
Full text linkMotivated by certain distributional and interpretive contrasts between Kannada reduplicated verbs carrying perfective vs. imperfective aspect-marking, here we pursue a view of these two constructions as instantiating two types of cross-linguistically attested pluractionalities — namely, event-external and event-internal pluractionality respectively. Such a characterization of Kannada reduplicated verbs allows us to borrow into their analysis several aspects of existing proposals for event-external and event-internal pluractionality, which in turn enables natural explanations for (many of) their distributional idiosyncracies