Evidentiality in Estonian and some other languages. Introductory remarks[TEXT NOT REPRODUCIBLE IN ASCII].

Estonian belongs to those languages, where evidentiality, that is, reference to the source of information is expressed morphologically (Standard Estonian, present: ma / sa / ta / me / te / nad lugevat 'I / you / he / we / you / they are said to read'; past: ma / sa / ta / me / te / nad (olevat) lugenud) 'I / you / he / we / you / they are said to have read'. In Estonian linguistics the morphological paradigm of evidentiality has been called the oblique mood (modus obliquus). However, because we are dealing with reported evidentiality, more recent grammars use the term 'quotative' (EKG 1993; 1995). The same phenomenon occurs in Livonian, for example, ta lu'ggiji 'he is said to read', and in the Baltic languages, for example, Standard Latvian: vins lasot 'he is said to read'; Standard Lithuanian: Jis daug dirbas 'He is said to work a lot'. In addition to those Standard Estonian forms that are treated as a mood, Common Estonian, especially the Estonian dialects, knows some other possibilities to express evidentiality in declarative sentences, both morphological and periphrastic. One of the peculiarities of Estonian is that in addition to the quotative, the third person imperative forms, which have become generalized into all the persons, have also evidential implications (Standard Estonian, present: ma / sa / ta / me / te / nad lugegu 'I / you / he / we / you / they should read'). Estonian linguists have regarded this paradigm as an independent mood as well--the jussive, which differs from the imperative. However, the essence of this mood has remained unclear. A mood that is similar to the Estonian jussive can be found in Livonian as well, for example, (laz) ma lu'ggog 'I should read'.

The evidentiality systems of Estonian, as well as Livonian and the Baltic languages, were discussed at the seminar "Indirect Mode of Reporting--a Specific Feature of the Baltic Areal?" held at Puhajarve in South Estonia on November 16, 2001. The seminar was organized in the framework of the conference "Language Contacts of Small Languages" (November 15-17). Efforts were made to find typological parallels in more distant languages as well. Petar Kehayov compared the Estonian evidential with that of Bulgarian, and Ago Kunnap provided an overview of a symposium on evidentiality held at the University of La Trobe (Melbourne). The present issue publishes four presentations made at Puhajarve and the presentation about the Enets evidential by Ago Kunnap at the University of La Trobe. The other presentations made at Puhajarve will be published in the conference proceedings "Vaikeste keelte kontaktid" (Voru 2002; in press).

Both in Estonian and the Baltic languages the oblique moods developed from the participles. In Standard Estonian the present form of indirect mood is marked by -vat (historically the partitive form of the present participle). The standard language adopted it from South Estonian dialects. At the same time the Voru dialect uses as a quotative marker in addition to the partitive form of the present participle also the v-final nominative of the present participle, for example, ta ei tundov teda arq 'he is said not to recognize him'. At present efforts are being made to establish this older but now rare form in the Voru literary language. Sulev Iva showed in his report "Expression of Indirectness in the Voru Language" (to be published in the conference proceedings) that in certain word types the language tries to keep the v-marked oblique mood also formally separated from the present participle. Namely, in verbal monosyllabic stems that end in consonants the present participle ends in -va/-va, for example, votva; the oblique form, however, still ends in -v, for example, vottov. The oblique mood was distinguished from the present participle also in those verbal stems that end in a long vowel or a diphthong, where the oblique marker was formed from two present markers that where connected with a schwa: -bev/-bov, for example, pernaso tutar kudabov kangast 'the landlady's daughter is said to be weaving a fabric' (see also Kask 1984 : 251-270; Keem 1997 : 49).

One of the peculiarities of Estonian in comparison with the other circum-Baltic languages is that in North and Central Estonian the infinitives are used as evidential markers. In North Estonian the da-infinitive, which is to a certain extent used in common Estonian, is the main evidential form, for example, Ta olla haige 'He is reported to be ill'. In the Central Estonian the present evidential is expressed by means of the ma-infinitive, for example, Ta elama hasti 'He is said to live well'. However, in both areas one can come across other evidential forms. The report "About the Reportive Mode in the Western Estonian Dialects" by Ellen Niit (to be published in the conference proceedings) focused on these forms in western dialects. The report showed that in the western dialect the da-infinitive marker has become regrammaticalized to such an extent that it is attached to the impersonal stem as well, for example, Muhu: joanibe tuldade meitele kulase 'it is said that we are going to have quests on midsummer day'.

In addition to the oblique forms that are clearly based on the participle and the infinitive, the northwestern part of the Tartu dialect (Rannu, Noo, Puhja) reveals the na-marked form, for example, Rannu: karu olna serande luum, et ...' the bear is reported to be such an animal that ...', and in Saaremaa one can find the na-/ne-marked oblique forms, for example, Juhan kiitna, et ta noor obu joosna tukkis nobemini kut Sandri oma; Ta pidane andama aja, milla ta naitama akab 'Juhan is reported to brag that his young horse can run much faster than Sander's horse; He should provide a time when he will begin to show it'. The origin of these forms is somewhat unclear. Because these forms have mostly a present meaning, one may assume that they proceeded from the ne-marked potential. In fact, this assumption has been made concerning the insular dialect (Grunthal 1910; Kask 1984 : 279-281); in the case of the Tartu dialect, however, it has been suggested that these forms may have proceeded also from the nud-participle (Kask 1984 : 280). The article "Evidentiality in South Estonian" by Helle Metslang and Karl Pajusalu compares these two development paths from the point of view of the contemporary theory of grammaticalization and draws a conclusion that the hypothesis concerning the potential is somewhat better grounded.

It would have been worthwhile to discuss one of the main means of expressing (present) evidentiality in North Estonian dialects in the contemporary standard language, namely, the construction with the modal verb pidama 'must, have to'. In the standard language the imperfect form of the pidama-verb expresses reported evidentiality in the present, for example, Ta pidi haige olema (= Ta olevat haige) 'He is said to be ill'. On the other hand, the present form of the pidama-verb has the meaning of inferred evidentiality, for example, See peab kull karu olema, kes siin kainud 'A bear must have visited this place' (cf. Erelt 2001). In the North Estonian dialects, however, the pidama-verb clearly implies reported evidentiality, for example, Ansekula: Ta peab outu juht olema 'He is said to be a driver'.

A number of seminar reports discussed the relationship between evidentiality and commands. The article "Does Estonian Have the Jussive?" by Mati Erelt is based on his seminar report. It attempts to prove that the Estonian jussive, which developed by generalizing the third person imperative form into the other persons (ma / sa / ta / me / te / nad lugegu 'I / you / he / we / you / they should read'), is actually an optative paradigm that has only some evidential implications. Therefore, its treatment as an independent mood is justified.

This view finds indirect support in the article "Reported Commands in Lithuanian Compared to Estonian" by Birute Klaas, which provides an overview of the treatment of the possibilities to express commands that are directed at the third person in Lithuanian grammars. B. Klaas shows that although the third person forms are usually regarded as belonging to the imperative paradigm, they clearly have an optative meaning. In fact, some forms that have become archaic have developed from the old optative. On the other hand, the more recent forms have also developed from the present indicative by attaching a modal particle or a suffix that has developed from the latter. However, in Lithuanian the optative meaning has not extended to the first and the second persons, as is the case in Estonian, and unlike Estonian, it has no evidential implications. In Lithuanian evidentiality is related to only statements. A command that is directed at the speaker is expressed only by means of the complex sentence, whereas the verb in the subordinate clause stands in the conditional. This alternative is possible in Estonian as well, for example, Ta utles, et ma teeksin = tehku see too ara 'He said that I should do this job'.

In Latvian the third person imperative form has not become generalized to the other persons either. Liena Muiczniece showed in her report "Concerning the Expression of the Indirect Command in Latvian" (to be published in the conference proceedings) that in Latvian indirect commands are now expressed by means of the particle lai 'let'. This particle is associated with the present indicative or the oblique mood, for example, lai tu skatoties 'you should look'. The corresponding Estonian particle las has no evidential meaning; the construction expresses a command that is directed at the listener, for example, las ma teen 'let me do it., las ta teeb 'tell him that he should do it; allow him to do it'.

Livonian reveals an unusual situation as well. Similarly to Estonian, the third person imperative has become generalized into the first and the second persons, but almost always the particle laz is added to this form. We are dealing with a clear case of reported imperative: (laz) ma lu'ggog--(laz) meg lu'ggogod 'I should read--We should read'. Livonian also shows a peculiarity in that evidentiality is expressed in the declarative sentence by an agent noun (performing various functions of the present participle), which agrees with the subject in number (e.g. ta lu'ggiji--ne lu'ggijid 'he read; they read'. Erika Krautmane discussed evidentiality in Livonian in her report "Concerning the Expression of Indirectness in Livonian".

Two articles discuss languages other than the circum-Baltic languages. The article by Petar Kehayov "Typology of Grammaticalized Evidentiality in Bulgarian and Estonian" focuses on the comparison of Estonian and Bulgarian systems of evidentiality. The article shows that the systems are rather different both functionally and formally. Functionally Bulgarian seems to be closer to Lithuanian than to Estonian. For example, both in Bulgarian and in Lithuanian the evidential forms also perform the function of inferentiality and mirativity (Ambrazas 1997 : 263-265).

The article "On the Enets Evidential Suffixes" by Ago Kunnap presents an evidential system where in addition to indirect evidentiality also direct auditive evidentiality is represented.

Evidentiality has been a topical problem in typological linguistics for over a decade. This was proved by Ago Kunnap's account of the international symposium at the University of La Trobe (Melbourne). We hope that the Puhajarve seminar also made a contribution to the description of this phenomenon.

Abbreviations

EKG--M. Erelt, R. Kasik, H. Metslang, H. Rajandi, K. Ross, H. Saari, K. Tael, S. Vare, Eesti keele grammatika I. Morfoloogia. Sonamoodustus; II. Suntaks. Lisa: Kiri, Tallinn 1995; 1993.

REFERENCES

Ambrazas, V. 1997, Lithuanian Grammar, Vilnius.

Erelt, M. 2001, Some Notes on the Grammaticalization of the Verb pidama in Estonian--Estonian: Typological Studies V, Tartu (Tartu Ulikooli eesti keele oppetooli toimetised 18), 7-25.

Grunthal, V. 1941, Itamerensuomalaisten kielten yksikon nominatiivi objektin edustajana aktiivin yhteydessa, Helsinki.

Kask, A. 1984, Eesti murded ja kirjakeel, Tallinn (Eesti NSV Teaduste Akadeemia Emakeele Seltsi toimetised 16).

Keem, H. 1997, Voru keel, Tallinn.

MATI ERELT (Tartu)

[TEXT NOT REPRODUCIBLE IN ASCII].