Gradable modality in Brazilian Portuguese - Presentation - PLUS 2013 (original) (raw)

Gradable modality in Brazilian Portuguese In this paper I investigate the semantics of gradable epistemic modality in Brazilian Portuguese (BrP) by comparing the modal verbs ‘dever’ and ‘ter que’. It is common assumption that both express necessity and are sociolinguistic variants. However, a close look in the uses of these verbs shows they are not synonymous. I argue they give different semantic contributions because (i) they mobilize different conversational backgrounds: ‘ter que’ is not compatible with epistemic modal bases while ‘dever’ is; and (ii) ‘dever’ works like a probability operator, expressing a grade of comparative possibility. I base my analysis on the relative and gradual kratzerian approach for modality in natural language, where the meaning of modals depends on contextually given parameters, namely, the modal base and the ordering source. Apart from the sociolinguistic and morphosyntactic differences (‘dever’ conveys formality and is a defective verb), ‘dever’ and ‘ter que’ express different meanings because ‘dever’ is evidential, while ‘ter que’ is not. In other words, the use of ‘dever-p’ (p = embedded proposition) requires the availability and evaluation of (indirect) evidences towards p, from which an inference about p can be drown (example 1). A ‘deve-p’ sentence can also be complemented with a sentence warning to the possibility of p not to be the case, which is an evidence that ‘dever’ does not express necessity (example 2). On the other hand, ‘ter que’ does not require the availability and evaluation of evidences, which blocks it to convey an epistemic inference, like ‘dever’ conveys (example 3). If ‘dever-p’ means that, according to the evidences available, p is a better possibility than non-p (without excluding non-p), ‘ter que-p’, on the other hand, does not allow this comparison and intuitively excludes non-p. Thus ‘ter que-p’ cannot be complemented with sentences warning to the possibility of p not to be the case (example 4). Also, ‘ter que’ is adequate to be uttered when there is no evidence towards p. One can use ‘ter que-p’ to express her wish, or a way to achieve a goal (example 5). I conclude that ‘dever’ and ‘ter que’ differ in force and in the types of conversational background they are compatible with. While ‘ter que-p’ expresses necessity (in all worlds that come closer to the ideal determined by the ordering source, p is the case), ‘dever-p’ expresses that p is a better possibility than non-p: in the worlds that are closer to the ideal, the non-p worlds are not better ordered than the p worlds. Moreover, different from ‘dever’, ‘ter que’ is not evidential (not compatible with epistemic modal base), but conveys bouletic, deontic and teleological meanings. This proposal points towards a new line of investigation of BrP modals and indicates a specialization in this language modal system: ‘dever’ is preferably epistemic; ‘ter que’ covers the other kinds of modalities. EXAMPLES: Example (1): Context: Ana, Sam and Ria are planning to go to the restaurant X. Sam tells Ana he wants to drink whiskey. This is the first time they are going to the restaurant X, and don’t know the menu. Ana asks Ria if she knows if they serve whiskey in the restaurant. Ria answers: (1) Deve ter (whiskey). Deve to.have (whiskey). ‘There must/might be (whiskey)’. Sentence (1) is naturally paraphrased as “É provável que tenha” (it is probable that have). Example (2): consider the same context in (1). (2) Deve ter (whisky), mas pode ser que não tenha (precisamos checar pra ter certeza). Deve to-have (whisky), but may to.be that not have.SUBJ (we need to check to be sure). The derivation of the meaning follows along this lines: we’re going to a restaurant; as far as we know about restaurants, they usually serve whiskey; so, since we’re going to a restaurant, there is a probability of there to be whiskey; but we’ve never been there before, so we cannot be sure. A ‘ter que’ sentence would not convey an inference in this situation, but the speaker’s position: If Ria utters ‘Tem que ter’, she conveys that she strongly wants there is whiskey there. A ‘ter que’ sentence is adequate to be uttered by Sam, since he is the one who declaredly wants whiskey. Example (3): ‘Ter que’ is odd in a diagnosis scenario. Assume we are observing the weather conditions, which are indicating rain (dark clouds, sultriness, etc). According to our knowledge about weather conditions, we utter: (3a) Deve chover. (‘Deve to.rain.’) (3b) # Tem que chover. (‘Tem que to.rain.’) Example (4): (4) ?? Tem que ter (whisky), mas pode ser que não tenha (precisamos checar pra ter certeza). ??Tem que to.have (whisky), but may to.be that not have-SUBJ (we need to check to be sure). Example (5): Context: Ria meets the new neighbor, talks to him for few minutes, and finds him interesting (for a date, for instance). He gives no clue about his marital status. Ria, planning to flirt with him, may utter: (5) Ele tem que ser solteiro! (He tem que to be single, so I can ask him out, etc.) Sentence (5) is not an inference about him being single (because Ria has no information from which she can infer that). She just hopes him to be single, so she can ask him out, etc., which yields a teleological reading. REFERENCES: Von FINTEL, K & GILLIES, A. 2007. An opinionated guide to epistemic modality. Oxford Studies in Epistemology 2: 32-62. KRATZER, Angelika. 2010. Modals and Conditionals: New and revised perspectives. Oxford Studies in Theoretical Linguistics. Oxford University Press. LASSITER, Daniel. 2010. Gradable epistemic modals, probability, and scale structure. Proceedings of SALT 20. P 197-215. PIRES DE OLIVEIRA, R.; Scarduelli, J. 2008. Explicando as diferenças semânticas entre ‘ter que’ e ‘deve’: uma proposta em semântica de mundos possíveis. Alfa Revista de linguística (UNESP. São José do Rio Preto. Online), v. 52, p. 215-236. PORTNER, Paul. 2009. Modality. Oxford University Press. YALCIN, Seth. 2010. Probability operators. Philosophy Compass. P 916-937.