Evidentiality in Uzbek and Kazakh
|
|
P ST P RF CP ST -di/-DI -gan/-GAn -(i)b/-(I)p [+C ONFIRMATIVE ] [Ø C ONFIRMATIVE ] [-C ONFIRMATIVE ] Both the perfect -gan/-GAn and the converbial past -(i)b/-(I)p may express non-firsthand information source, but this meaning is not primary and must be contextually specified. It is only because they are not marked as confirmative that these meanings may surface. The copular forms ekan/eken and emish/-mIs fall in a second class. These morphemes, when not used to express admirativity or emotivity express non-firsthand information source. The use of these forms to express non-firsthand information source is equivalent to what occurs in Turkish, Azerbaijani, and other languages of the Eurasian evidentiality belt, in which “the combination of taxis (marking for anteriority) and resultativity, i.e. the double marking of a past auxiliary and a past participial form, can result in the pluperfect’s functioning as a genuinely pure reported form” (Friedman 1979, 345). Copular ekan/eken is not, of course, limited in distribution to perfects, but may occur with a wide variety of forms. Like the double marking of the perfect in the languages discussed by Friedman, the use of ekan/eken is morphosyntactically marked, so the surfacing of the more semantically marked non-firsthand meaning is not surprising. 122 What remains to be decided here is whether the non-firsthand meanings of ekan/eken (as well as emish/-mIs) place them into a category of EVIDENTIALITY , or whether these evidential meanings are secondary to some other meaning. Although the label EVIDENTIAL is often conferred upon -gan/-GAn and -(i)b/-(I)p , these morphemes may only be linked to evidentiality by virtue of their primary non-confirmative meanings. In discussions of evidentiality, it is STATUS or MODALITY that poses the greatest challenge to the existence of an EVIDENTIAL category. In his reanalysis of Jakobson’s (1957/1971) verbal categories, Aronson eliminates the category of EVIDENTIALITY and suggests that it “should be regarded as closely related to, or, better, a subvariety of, STATUS (E n /P s )” (1991, 116). Aronson considers confirmativity to be a sub-variety of STATUS , that is, the “subjective evalutation of the narrated event by the speaker, i.e., E n /P s ” (1991). Scholars who propose that EVIDENTIALITY and MODALITY / STATUS are distinct categories do not deny that a link exists (see Aikhenvald 2003; de Haan 1999; Cornillie 2009), yet attempt to account for the differences between what they see as two different categories in a variety of ways. DeHaan (1999) proposes that evidentiality and modality differ with regard to the scope of negation; modals always fall under the scope of negation and (fully grammaticalized) evidentials scope over negation. In Maricopa, for example, a visual evidential cannot have the meaning I didn’t see; this meaning must be expressed by an independent verb (Gordon 1986, 85, as cited in de Haan 1999, 12-13): (180) a. Waly-marsh-ma-?-yuu NEG -win. DUAL - NEG - VIS ‘(I saw) They didn’t win.’ b. Marsh-m waly-?-yuu-ma-k win. DUAL - DS NEG -1 SG -see- NEG - ASP ‘I didn’t see them win.’ 123 In the simpler non-firsthand sort of evidential meaning found in Uzbek, Kazakh, and many other languages of the Eurasian evidentiality belt, the scope of negation is much more difficult to determine. This is likely due to the fact that ekan/eken and similar morphemes in other languages cannot be associated with any single verb of perception, as these forms merely state the existence of non-firsthand evidence without stating their source: (181) U ket-ma-gan ekan. (Uz) he go- NEG - PRF EVID ‘There is evidence that he didn’t go.’ ~ ‘There is no evidence that he went.’ Faller (2006) proposes that two types of evidentiality, one that operates on the illocutionary level (evidentiality proper) and one that operates on the propositional level (modal evidentiality). To the first class she assigns the evidential enclitics of Cuzco Quechua, and to the modal evidential class she assigns the German modal sollen, which typically indicates reportatitvity. Under this compromise approach, evidential meaning is expressed in two locations in the grammar, and languages may differ as to where this meaning is expressed. Under DeHaan or Faller’s approaches, there exists a variety of modality that expresses of evidential meaning. Assigning a given form in some language MODAL status does not mean that evidential meaning cannot be expressed. For the purposes of this work, I will not make any claims as to the existence or non- existence of EVIDENTIALITY as a distinct category. Rather, I focus on the most likely candidates in Uzbek and Kazakh (ekan/eken, emish/-mIs) and determine whether they are, indeed, evidentials. In support of these morphemes being evidentials is the fact that they express evidential meaning. That is, ekan/eken express non-firsthand information source, and emish/-mIs express reported information. This fact alone, however, is insufficient to claim evidential status for these 124 morphemes, as morphemes in many other languages (e.g. English must, German sollen) may express evidential meaning, but are not universally considered evidentials. There are three issues with considering ekan/eken and emish/-mIs evidentials. The first is the fact that these forms are never obligatory, a criterion for evidentiality that has been claimed by some authors. The second is the strong connection between these morphemes and (non-)confirmativity, a variety of meaning that quite clearly belongs to the category of STATUS or MODALITY . The third issue with categorizing these morphemes as evidentials is the fact that they also express admirativity, which is the expression of surprise at a newly discovered fact and is often considered a sub-type of non-confirmativity A number of authors have proposed that a language with true evidentiality will obligatorily express it, at least within paradigms where evidentiality is semantically compatible with mood. According to Aikhenvald (2004), in a language with true evidentiality (as opposed to an evidential strategy) declarative sentences will specify the type of information the speaker has for the truth of the sentence. Some authors have denied that this criterion is valid (de Haan 1999; McCready and Ogata 2007) and even Aikhenvald (2004) states that languages in which there exists an unmarked term pose problems for this criterion. In Uzbek and Kazakh, clauses without ekan/eken or emish/ -mIs are both formally unmarked (in that they lack a certain morpheme) and functionally unmarked (in that no evidential meaning - such as firsthand information source - is expressed). So, if a speaker chooses to employ the simple past in either Uzbek or Kazakh, no claim has been made that the speaker witnessed or otherwise participated in the event described. Under a strict interpretation of the obligatory marking criterion, Uzbek and Kazakh do not exhibit true evidentiality, as the expression of information source is never required. 125 The fact that ekan/ekan and emish/-mIs need not be present indicates that when they are present, the utterance containing them is formally marked; this corresponds to functional markedness for non-confirmativity. Due to the markedness of these morphemes, speakers who employ them must have a motive for employing them, and this motive is an unwillingness to confirm the contents of the utterance. The non-confirmativity of these morphemes is further supported by the alternation between ekan/eken and -(i)b/-(I)p. When a speaker wishes to express non-firsthand information source and simple past tense, the speaker has the option of employing -(i)b/-(I)p, which is an acceptable substitute because both morphemes express non- confirmativity. While ekan/ekan and emish/-mIs do, indeed, express non-firsthand information source, they also express non-confirmativity, and this is manifested as doubt or uncertainty. The final issue with definitively categorizing ekan/eken and emish/-mIs as evidentials is the fact that these forms also bear what is known as ADMIRATIVE meaning. Admirativity is defined as the expression of “surprise at a newly discovered and previously unsuspected fact,” and is a variety of non-confirmativity (Friedman 1988). In Uzbek and Kazakh, the admirative (or, more broadely, emotive) usage of these morphemes has expanded to interrogative forms, resulting in the formation of rhetorical questions. The properties of these interrogatives will be fully described in the Chapter 5; for now the ability of these forms to express these meanings should be viewed as further evidence against a strict evidential categorization. Although ekan/ekan and emish/-mIs express evidential meaning, they do not belong to a category of evidentiality. Rather, because their primary purpose is to express non- confirmativity, they belong to the category of STATUS or MODALITY . Although many authors consider EVIDENTIALITY to be a valid category, there are a number of analyses under which evidential meaning may be expressed by members of the STATUS / MODALITY category. Under the 126 analysis here, in which non-confirmativity is a specific category (to borrow a term from Whorf 1938/56) under the generic category of STATUS / MODALITY , non-firsthand information source is merely an extension of non-confirmative meaning. While non-confirmativity expresses a lack of speaker confidence in the truth of an utterance, the evidential meanings apparent in ekan/eken provide a reason for that lack of speaker confidence. Yüklə Dostları ilə paylaş:
|