What do different methods of data collection reveal about evidentiality? 1 SEPPO KITTILÄ, ERIKA SANDMAN & LOTTA JALAVA (UNIVERSITY OF HELSINKI) Outline 2 1. 2. 3. 4. 5. 6. 7. Background Grammars Corpora Field methods Questionnaires Participant observation Conclusions Background 3 Evidentiality is currently one of the most studied topics in linguistics (see e.g. Aikhenvald & Dixon 2003; Aikhenvald 2004; Peterson & Sauerland 2010). There are, for example studies dealing with evidentiality systems and the use of evidentials in discourse. Background 4 Our study is a part of the project ’Interactional, cross- linguistic, theoretical and areal perspectives on evidentiality and egophoricity’. The project comprises three parts: 1. Theoretical and cross-linguistic perspectives on evidentials 2. Egophoric marking in Amdo Sprachbund: interactional, typological and areal perspectives 3. Evidential functions in the Uralic languages and Siberian contact area Visit: https://blogs.helsinki.fi/evid-ego/ Background 5 Studying many different facets naturally means that we must use various data collection methods. These methods constitute the topic of this talk. Both pros and cons of the employed methods will be examined. We will also discuss how the different methods complement each other. This will be done in light of hearsay evidentials. Case study 6 Our presentation is based on hearsay evidentials, as in: Finnish (1) Aino on kuulemma söpö Aino COP.PRES.3SG hearsay cute ’They say that Aino is cute’ Ngiyambaa (Donaldson 1980: 276) (2) bura:y-dja=lu ga:-y-aga child.ABS-EV=3ABS bring-CONJ.M-IRR ’It’s said that she’s going to bring the children’ Case study 7 In what follows, we will discuss the different methods we are using in light of what they reveal about hearsay evidentials. As will be shown, the different methods are, of course, not contradictory, but they rather enable us to approach evidentiality from various perspectives. For example, data gathered in the field is directly relevant to typology. All methods also have their limitations. Grammars 8 General: Hearsay evidentials are common across languages (quotatives and hearsay evidentials). Grammars also provide us with information on some secondary uses (distancing) and the genres where hearsay evidentials are common may be discussed. There are languages, in which hearsay evidential is the only evidential category available (e.g., Estonian and Nuosu). Grammars reveal what kind of hearsay evidentials are common across languages, but they do not necessarily tell what is common within a single language. Grammars 9 Applicability: Grammars constitute the starting point for any cross-linguistic study on the topic. They are also relevant to studies of individual languages; for example does cross-linguistic frequency correlate with frequency within a language? Grammars aid field linguists in recognizing certain evidential categories in the language(s) they are studying. Grammars 10 Problems: grammars usually do not discuss evidentiality in languages where the category is not obligatory. Moreover, usually only the primary uses of evidentials are accounted for, and the use of evidentials for secondary functions (e.g. surprise, irony) is not necessarily discussed. Corpora 11 General: How reported evidentials are used in actual language; what are their functions, what is common/rare within languages? Are reported evidentials typical of certain genres? The use of lexical expressions of evidentiality (languages for which large corpora are available usually lack evidentiality as an obligatory category). Corpora 12 Applicability: Usage-based studies on evidentiality: how are evidentials actually used in a single language? The study of lesser known languages (e.g. Wutun, Nenets) can be informed by studies on evidential expressions in more extensively studied languages (e.g. Finnish, for which corpora are available). The study of evidentiality strategies (e.g. tense in many Uralic languages). Corpora 13 Problems: corpora may be available for mainly certain genres that do not necessarily give us a complete picture of the use of evidentials. For example, the reported/hearsay evidential of Finnish would not be attested in a newspaper corpus. Moreover, large corpora are mainly available for languages that lack obligatory evidentiality. Potential problem/question: is Internet a reliable corpus? What can we say about evidentials in light of their uses in the Internet? Example 14 Secondary uses of Finnish hearsay evidential kuulemma in data obtained from Internet: (3) Puhutko unissasi? Minä kuulemma puhun. Do you speak in your dreams? They say that I do. http://www.napsu.fi/vastaukset/kysymys/28152 (4) Sinä kuulemma luet omia kirjoja, piirtelet omaan vihkoon ja istuskelet. They say that you read your own books, and make some drawings in your notebook and sit around. http://lamppu.tumblr.com/post/15356683334/joo joskus-ennen-joulua-sanoin etta-sain Field linguistics 15 General: Information about less studied languages can be gathered only via field linguistics. Information on evidentiality in different extralinguistic contexts. Information on secondary uses. Applicability: Field linguistics provides invaluable data for typology and linguistic theory. Field linguistics includes several different methods (elicitiation, stimuli-based techniques, text collection, participant observation) that complement each other. Field linguistics 16 Problems: Data may be biased towards certain genres (e.g. folktale narratives) of a language under study. It is important to study not only monologues but also dialogues! In the case of translation equivalents, interference from the source language may cause some problems. Gathering data is very time-consuming, informants may be very hard to get in contact with. Example 17 Hearsay evidential in multiple perspective constructions (see Evans 2007) in Wutun (Sinitic, Erika Sandman’s field notes): (5) hai mi-mai-de re sho-li still NEG-sell-NMLZ FACT REP-SEN.INF ’(He) still didn't sell (the thangka) (for sure), they say.’ (6) gu-jhege qhi-zhe sho-di-li 3P-PAUC go-PROSP say-PROGR-SEN.INF sho-li REP-SEN.INF ‘I have heard from somebody that they say they will go’. Questionnaires 18 General: Information on exactly those features we are interested in, we can create exactly the scenes that should provide us with the relevant information. For example, we can create scenarios where the speaker has more than one kind of evidence available (which are rarely discussed in grammars). Study of variation (for example, between different age groups). Questionnaires 19 Applicability: Questionnaires complement other data collection methods. For example, corpora may provide us with some ideas about the language use that can be checked by using questionnaires. Questionnaires can also be used in field linguistics. One of the goals of our project is to develop questionnaires and stimuli-based techniques for studying evidentials in the field. Questionnaires 20 The limits of use of hearsay evidentials. For example, where hearsay evidentials are no longer possible. Are they possible with facts, questions etc. Questionnaires may be filled by a very high number of informants, which make them very helpful in any study of evidentiality. Example 21 For example, the visual/inferential evidential of Finnish is used with more direct sources of information, while the hearsay evidential is used whenever the access to the information source is less direct. See the examples on the next slides. Examples 22 You read in a paper that Muammar Gaddafi has been arrested. You tell a friend about this right away: Muammar Gaddafi on______pidätetty Muammar Gaddafi has________been arrested Näköjään Kuulemma Both 97 36 1 Examples 23 You read in a paper that Muammar Gaddafi has been arrested. You tell a friend about this in two hours: Muammar Gaddafi on______pidätetty Muammar Gaddafi has________been arrested Näköjään Kuulemma 7 127 Participant observation 24 General: Allows studying the use of evidentials in real-life situations. The use of evidentials in dialogues involving two or more speakers. Important in studying the secondary uses of evidentials (e.g. surpirse, irony, sarcasm, telling lies). Applicability: Complements other methods of data- gathering (corpora, questionnaires). Is an essential method in field linguistics. Example 25 Hearsay dubitative evidential in Tundra Nenets (Samoyedic branch of the Uralic language family, Lotta Jalava’s field notes): (7) xə-ńaəd=ći ťeńewə-wna-w? (ENS/NenTay2011) what-ABL=CL know-DUB-1SG>SG ‘How would I have known that?' (as I clearly didn’t know) • The hearsay dubitative evidential encodes indirect knowledge of a situation to which the speaker does not believe, or, retelling something that the speaker does not commit herself to (evidential +, epistemic -). • It is often used for encoding doubt or pretence, commonly in interaction and dialogues but never in newspaper texts and only rarely in written texts. Concluding remarks 26 Evidentiality comprises many different facets and it is therefore important to use many data collection methods for studying it. Different methods reveal different features of evidentiality, all of which are central if our goal is to arrive at a comprehensive understanding of evidentiality. Whenever evidentiality is discussed, the pros and cons of different data collection methods should be taken seriously. References 27 Aikhenvald, Alexandra Y. & R.M.W. Dixon 2003 (eds.). Studies in Evidentiality. Typological Studies in Language 54. Amsterdam: John Benjamins. Aikhenvald, Alexandra Y. 2004. Evidentiality. New York: Oxford University press. Donaldson, T. 1980. Ngiyambaa. The language of the Wangaaybuwan. Cambridge: Cambridge University Press. Evans, Nicholas 2007. View with a View: towards a Typology of Multiple Perspective. Proceedings from Berkeley Linguistics Society 1: 93-120. Berkeley: University of California. Peterson, Tyler & Uli Sauerland 2010 (eds.). Evidence from Evidentials. University of British Columbia. Working Papers in Linguistics, vol. 28.