Printer Friendly

Evidentiality in Livonian.

Evidentiality is a grammatical category specifying the source of the information described by the proposition. The present paper will provide a preliminary but comprehensive picture of the Livonian evidentiality system, taking as a starting point the semantic distinction between direct and indirect evidentiality (see Willett 1988). In the case of direct evidentiality, the information comes from a primary source, that is, the speaker himself is a witness of the situation described by the proposition. In the case of indirect evidentiality, on the other hand, the information comes from a secondary source--it is inferred or assumed on the basis of some independent evidence or heard from someone else. Direct evidentiality may be further specified into visual, auditory, or other, and indirect evidentiality into inferrential, assumptive, reported, or other; these specifications are usually called evidential values (see e.g. Plungian 2001).

From a functional point of view a distinction is usually drawn between grammatical means, whose primary function is to express evidentiality, and grammatical means, whose primary function is to express other (non-evidential) categories but which convey evidentiality as a secondary conventionalized meaning (e.g. the perfect or the pluperfect). Following Aikhenvald (2004 : 105), we regard the former as evidentials and the latter as evidential strategies. In several languages of the Circum-Baltic area indirect evidentiality takes the value of the reported evidential. Thus, in languages like Estonian or Latvian, evidentials are used to show that the information is acquired from another person.

Another premise we adopt comes from the Estonian descriptive grammatical tradition and relates to the mood system of this language. In the early seventies, the Estonian linguist Huno Ratsep tried to accommodate a special category into Estonian grammar. This category, called m o d e o f reporting, consisted of two opposite values--direct and indirect mode of reporting. According to Ratsep, Estonian moods should be classified on the basis of these two modes of reporting. The unmarked indicative and the imperative belong to the direct mode of reporting. The reported (the so-called oblique mood) and the jussive, on the other hand, constitute, respectively, the indicative and the imperative of the indirect mode of reporting (Ratsep 1971). The same model pertains to Livonian where a distinction could be drawn between the reported indicative (the oblique mood) and the reported imperative (the jussive) (cf. Erelt, Metslang, Pajusalu 2006 : 126). The Livonian evidentials are formally and semantically close to Estonian and Latvian evidentials; nevertheless, they reveal some specific features, which make them unique among the neighbouring languages (cf. e.g. Walchli 2001; Erelt, Metslang, Pajusalu 2006 : 125-132; Krautmane 2006; 2010 : 13-18; 65-68).

The present paper will discuss the Livonian evidential moods and evidential strategies, their forms, functions, and use. Livonian will be compared first and foremost with Estonian, which is a genetically close language, as well as a contact language. The analysis is based on the Livonian texts, which were written down during fieldwork in the second half of the 19th and in the 20th century by J. A. Sjogren, E. N. Setala, L. Kettunen, J. Magiste, and S. Suhonen. The texts include mainly folk narratives describing reported information. In this paper language data of these manifold sources are presented in Standard Livonian orthography.

1. Reported indicative

1.1. The form of the oblique mood

In both main varieties of Livonian--Salaca and Courland Livonian--the present tense form of the reported indicative is marked by the suffix -(j)i, e.g. Example 1 (KET 67).

(1) Tammon ka voll-i viga, tam um kiton, ku ta

he:DAT also be-REP.IND.SG problem he.NOM be.PRS.3SG say:APP if he.NOM

so-ji tija makkoks kilmo vieta juodo, get-REP.IND.SG empty.GEN stomach:INSTR cold:PART water:PART drink:INF

ta e-ji id reitkoks uloks

he.NOM go-REP.IND.SG only time:INSTR crazy:INSTR

'He is also reported to have a problem; he has said that if he is said to be able to drink water on an empty stomach, he is reported to go mad at once (i.e. to fall ill)'

In the case of negation the negation verb ab 'not', which is identical with the indicative negation marker, is placed in front of the affirmative form both in the singular (Example 2 (KET 37)) and the plural (Example 3 (KET 28)).

(2) Sis kure kitiz algo ta rokandogo. Tama

then devil.NOM say:PST.3SG proh:imp he.NOM speak:IMP.SG he.NOM

kitiz ta ab rokandi-ji

say:PST.3SG he.NOM neg speak-REP.iND.SG

'The devil then said that he should not speak. He said that he is reported not to speak

(3) (ma) kulis set kord, ku min izaiza

I.NOM hear:PST.1SG several time when my grandfather.NOM

nedi ridliz un kitiz, algo ne ents

they:PART scold:PST.3SG and say:PST.3SG proh:imp they REFL

jelaidon sieda-toda sizzol andagod, ku ne

animal:PL:DAT this_or_that:PART therein give:IMP:PL that they

ab iks ap ttuoi burad ab voll-i-d, un ku

no one no other witch:PL neg be-REP.IND-PL and that

ne ents jelaidi set iz id jara

they REFL animal.PART only themselves once PRF.PTCL

rikkijid, abjoudi aindi sizzol andos,

spoil:REP.iND:PL evil:PL.PART plant:PL.PART to_them give:INF:INE

motlos siekoks mingis tija buran vasto jello

think:iNF:iNE this:iNSTR some.GEN trivial:GEN witch:DAT against act:INF

'(I) heard several times that my grandfather was scolding them and said that they should not feed this or that to their animals, that no one of them is reported to be witch, and that they themselves spoiled their animals by feeding evil plants to them, thinking that they could act against some insignificant witch with it'

Interestingly, although the morphological distinction between the active (personal) and the impersonal voice is extremely pervasive in the Finnic languages (extending to both finite and non-finite verb forms), this distinction is neutralized in the present tense forms of the Livonian reported indicative. Consider Example 4 (SUH 26) where the unmarked personal form is used in the context of the impersonal.

(4) Ma um kulon ku oksakaldi Rigos sal Venas

I.NOM be.PRS.1SG hear:APP that stickleback:PL.PART Riga:iNE there Daugava:INE veijji-ji-d

fish-REP.IND-PL

'I have heard that sticklebacks have been reportedly caught there in Riga in Daugava River'

The (j)i-marked form coincides with the form of the agent noun although historically it could be also regarded as a form of the present participle (see Audova 2003). (1) Nowadays the forms with this suffix still function as present participles in the Karelian, Veps, Mordvinian, and Saami languages. Thus, originally the (j)i-form could have been the nominative form of the present participle (or the form of the total object--modifier of the utterance verb: the genitive in the singular, the nominative in the plural). The grammaticalization of a nominative participle into a marker of evidentiality is not an exceptional development in Finnic--in the Voru and Setu dialects of South Estonian, for instance, the nominative form ending in -v of the present participle has become a marker of reported evidentiality. Notably, the Voru-Setu v-form differs from the vat-marked reported evidential in Standard Estonian, which originated in the partitive case form of the present participle. The nominative source of the Livonian reported indicative marker makes it especially close to the reported evidentials in the eastern dialects of South Estonian despite the fact that agreement with the subject in number is a specific feature of Courland Livonian (cf. ta soji in Example 1 and ne rikkijid in Example 3).

The grammaticalization path of the reported indicative in Livonian is probably similar to the path assumed for Estonian. The evidential meaning was foregrounded when the non-finite verb form, which functioned as a complement of an utterance verb, became an independent clause, whereby the non-finite form acquired the function of a finite form (cf. Ikola 1953 : 48-49). This kind of reanalysis of the complement clause is considered to be a possible source of Latvian and Lithuanian reported evidentials (see Walchli 2000 : 194-195).

What makes the Livonian reported indicative special is, however, the fact that the non-finite form with the suffix -(j)i is used also as a nomen agentis. Unlike participles, which typologically constitute a common source of grammatical evidentials (especially in the languages of Eastern Europe; see Kehayov 2008), agent nouns lend themselves only rarely to the expression of evidentiality. The only cases of syncretism between agent nouns and evidentials, that we are aware of, come from the Qashqadarya dialect of Arabic (spoken in Uzbekistan) and Khowar (Indo-European, northern Pakistan). Qashqadarya Arabic is especially close to Livonian because the syncretism there is also threefold: active participle = agent noun = indirect evidentiality (Isaksson 2000 : 394). In other words, both Livonian and Qashqadarya Arabic employ a single verb form as an adjective-attributive participle (e.g. running horse) as a nomen agentis (e.g. runner) and as an indirect evidential (e.g. x is reportedly running). On the other hand, Khowar differs both formally and functionally from Livonian and Qashqadarya Arabic in that an agent noun is combined with the auxiliary 'be; become' in this language to produce inferential evidentiality (Bashir 2006 : 5).

The past tense forms of the Livonian reported indicative coincide with the past participles. Differently from Estonian, Courland Livonian generally has different forms in the singular and the plural, as is the case in Finnish. On the other hand, in Salaca Livonian, similarly to Estonian, the past reported indicative forms are not marked for number. Past participles without an auxiliary verb are by far the most frequent markers of the past tense of the reported indicative (Examples (5) and (6))--the active forms exhibit the ending -(o)n in the singular (in the case of a monosyllabic base in Courland Livonian also -nd) and -(o)nd in the plural; the passive forms exhibit the ending -dod/-tod. All these forms correspond to the nominative forms of the past participle. Compound past tense forms of the reported indicative are formed by the (j)i-form of the auxiliary verb volli 'be' and the past participle of the lexical verb; these forms are much less common (see Example 7 (SET 382)).

(5) ta kito-n mis sa minnon maksad (SET 156)

s/he say-APP.SG what you.NOM I:dat pay:PRS.2SG

's/he was reported to say what you pay me'

(6) Un nei nemat salola-tot un jello-nd (KET 101)

and so they.NOM marry-PPP and live-APP:PL

And so they are reported to have been married and lived

(7) ma um kulon, se Kuolka nana voll-i sieda

I:nom be.PRS.1SG hear:APP.SG this Kolka cape be-REP.iND.SG this:PART

vito suggo-n

water:ILL arise-APP.SG

'I've heard that Cape Kolka was arisen like that'

An even less common construction is the so-called double perfect (see Krautmane 2010 : 56), which consists of the (j)i-form of the auxiliary verb, the past participle of the auxiliary verb, and the past participle of a lexical verb, as in Example 8 (KET 29):

(8) Klous sellist rokko iz usk, ta andiz ama

Klous such.PART story:PART neg:PST believe.coNNEG he give:PST.3SG all:PART

viga Didrikon kis tanda voll-i vond jara

blame:PART Didrik:DAT who he:PART be-REP.iND.SG be:APP.SG prf.ptcl

burro-n

bewitch-APP.SG

'Klous did not believe this story; he put all the blame on Didrik, who was reported to have put a spell on him'

Such forms containing two past participles occur marginally also in Estonian (Kehayov 2004 : 820) and usually convey what Aikhenvald (2004 : 157-158) calls 'conceptual distance'. They are motivated by the unwillingness of the speaker to make commitments as to the truth of the reported facts.

The reported indicative occurs also in constructions referring to the future (cf. Metslang 1994 : 536-537, 615; Norvik 2012); for example, with the future copula limo 'will be' (Example 9 (SET 210)) and in a resultative construction with a future meaning formed with verb somo 'get' in (Example 10 (KET 138)).

(9) sid um keratod,[ku se kenig unso neiz] ku sina

here be.PRS.3SG write:PPP that this king.NOM dream:INE see:PST.3SG that you

li-j i sie kenig tidaron pa miekoks

will_be-REP.IND.SG this.GEN king:GEN daughter:DAT as husband:INST

un laz se kenig jomand votag, andag

and IMP.PTCL this queen:NOM capture:IMP.SG give:IMP.SG

sinda benda kaddo un laz tiego sinnon tutkam

you:PART hangman:GEN hands:ILL and IMP.PTCL do:IMP.SG you:DAT end:GEN

'It is written here [that this king saw in his dreams] that you are reported to get married to this princess, and this queen should capture you and hand you over to the hangman and kill you'

(10) Ta ito kabalo minnon naktis ku

he.NOM one:iLL piece:iLL I:dat show:PST.3SG that

ma so-ji ils pug-dod

I.NOM get-REP.iND.SG up hang-PPP

'He continuously showed me that I was reported to be hanged'

While evidential forms of the 'get'-verb expressing future can be found in Estonian, the construction with the copula limo is absent from Standard Estonian. Notably, such an explicit marking of the future tense in the reported indicative is characteristic of Latvian (Holvoet 2001 : 114-118).

The reported indicative forms occur also in the presumably Latvian-inspired debitive construction, which consists of a (j)i-form of the verb 'be' and the action noun of the lexical verb in the partitive. Example 11 (SET 222) contains a reported debitive future form, and Example 12 (KET 140) shows a present tense form.

(11) kure um kiton, laz lakko idog juro un

devil.NOM be.PRS.3SG say:APP imp.ptcl go:iMP.SG by_the_evening there and

laz satellog valmoks, mup uondzol

IMP.PTCL place:iMP.SG ready:iNSTR tomorrow morning:ADE

li-ji le-mos-t suoda pal tegiz

will_be-REP.iND.SG go-ACTNM-PART war.GEN to again

'Devil has said that one should go by the evening and place an order, it is reported that tomorrow morning one has to go to war again'

(12) Ta kitiz iks ta tauz, tammon voll-i

he.NOM say:PST.3SG all_the_same he:DAT be-REP.iND.SG

sie loda palkerato-mis-t

this.GEN desk:GEN on write-ACTNM-PART

'He said that it didn't matter, he was reported to have to write on this desk'

The forms discussed so far belong to the paradigm of the Livonian evidential proper. In addition, perfect past tense forms of the direct mode of reporting are sometimes employed as an evidential strategy. In particular, these forms may function as the past tense form of the reported indicative; see um kiton in Example 13 (= Example 1) (KET 67).

(13) Tammon ka voll-i viga, tam um kiton,

he:DAT also be-REP.iND.SG problem he.NOM be.PRS.3SG say:APP

ku ta so-ji tija makkoks kilmo vieta

if he.NOM get-REP.iND.SG empty.GEN stomach:iNSTR cold:PART water:PART

juodo, ta e-ji id reitkoks uloks

drink:iNF he.NOM go-REP.iND.SG only time:iNSTR crazy:iNSTR

'He is also reported to have a problem; he has said that if he is said to be able to drink water on an empty stomach, he is reported to go mad at once (i.e. to fall ill)'

1.2. The use of the reported indicative

The reported indicative (the oblique mood) is similar to the Estonian reported evidential with regard to both its development and its semantics and syntax. However, the overall impression from Livonian texts is that in Livonian the reported indicative is more common than in Estonian. Figure 1 shows the linguistic devices used to express reported statements in the decreasing order of specialized grammaticality.
Figure 1. Scale of marking devices of reported evidentiality.

--marked with a special evidential mood ('oblique mood')

  --marked with an evidential strategy
(other verb forms, such as the perfect or the pluperfect)

    --lexically marked or left unmarked


Livonian tends to mark reported evidentiality on a regular basis by means of its oblique mood. Estonian, on the other hand, tends to occupy a position towards the right-hand end of the scale by marking evidentiality by means of evidential strategies, lexically or leaving it unmarked. As a rule, Estonian avoids repeated use of the reported evidential marker. For example, in the Estonian equivalent of Example 13 (see Example 14) one would normally use a form of the oblique mood (olevat 'is said to be') at the beginning of the sentence, followed by an evidential strategy (indicative pluperfect oli utelnud 'had said'), and leave the predicates of the conditional sentence unmarked in relation to evidentiality (cf. saab 'gets', jaab 'remains'). (3)

(14) Temal ka ole-vat viga, ta ol-i utel-nud,

he:ADE also be-REP.IND problem he.NOM be-PST.3SG say-APP

kui ta saa-b tuhja maoga vett juua,

if he.NOM get-PRS.3SG empty.GEN stomach:iNSTR water:PART drink:INF

jaa-b ta uhe korraga haigeks

fall-PRS.3SG he.NOM at_once ill:iNSTR

'He is also reported to have a problem; he has said that if he is said to be able to drink water on an empty stomach, he is reported to go mad at once (i.e. to fall ill)'

In Livonian, the reported indicative can occur in various types of subordinate clauses, in independent clauses, or in main clauses. A certain type of subordinate clauses is represented by complement clauses of perception or utterance verbs, such as kito 'say', rokando 'speak', kulo 'hear', nagto 'show' or kizzo ask where the hearsay meaning (reported evidentiality) is already expressed by the complement taking verb. After such verbs Estonian usually has forms of the unmarked indicative, as the reported indicative is considered redundant (Sepper 2005 : 59). In the Estonian Dialect Corpus, which includes dialect texts of the 19th and 20th c., examples of complement clauses in the oblique mood as Example 15 (EDC, Kihnu dialect) are very rare and seem to be restricted to the southern part of the Estonian language area.

(15) jah sos uteldas et sos saa-vat 'leiba moni

yes then say:PRS.iMPRS that then get-REP.iND bread.PART some.NOM

utell oo moni+gord

say:PRS.3SG (hesitation) sometimes

yeah, it is said then that one reportedly gets bread, at least sometimes, people say

On the other hand, the Livonian material shows that double marking of hearsay is highly common in this language. The majority of the examples above include a complement clause of a perception or utterance main verb, in which the complement verb takes the form of the reported indicative. Example 5, where the complement verb is in the (unmarked) indicative, is an exception.

The reported indicative occurs in Livonian both in indirect (Example 16 (KET 140)) and in direct questions (Example 17). Whereas its occurrence in indirect questions is self-evident--considering its common use in complements of utterance verbs, its occurrence in direct questions deserves attention as it is very rare in Estonian (Kehayov 2004 : 827; Sepper 2005 : 16; EDC). Example 17 from Salaca Livonian (SjW 328) is a question about whether there is some second- or third-hand evidence about the deeds of the workers. Here the use of the reported form refers to the presupposition that the information expected in the response comes from a third party.

(16) Ta kiziz voi min lus ti-ji ka molto bildidi

he ask:PST.3SG whether I:DAT appeal-REP.IND.SG also picture:PART.PL paint:INF

'He asked whether I'm reported to like painting also pictures'

(17) Mis nant tum tuonikad sal pate-ji?

what those ten worker:NOM.PL there do-REP.IND

'What are those ten workers said to do there?'

Thus, the Livonian reported indicative is functionally more neutral than its Estonian equivalent. For this reason, it does not generally imply uncertainty or doubt on behalf of the speaker towards the propositional content. An epistemic appraisal is communicated either by the so-called double perfect (see Example 8) or by lexical means (Krautmane 2010 : 67).

In independent sentences of popular narratives the most common past tense form of the predicate is the past participle (as in Examples 5 and 6) closely followed by the indicative perfect. One might assume that the function of the latter is to foreground an event in the narrative, as um ouron 'is said to have screamed' in Example 18 (KET 31).

(18) Se perimiez, se vo-nt se suz, se irgo-n sieda

this.NOM farmer.NOM it be-APP this wolf.NOM it start-APP this:PARt

poisso kisko. Sis se poiss um ouro-n kuolm kord

boy:PART tear:INF then this.NOM boy.NOM be.PRS.3SG scream-APP three times

'This farmer, it is said to have been this wolf, it is said to have started to tear this boy. Then this boy is said to have screamed three times'

2. Reported imperative (jussive)

Livonian forms of the reported imperative include the imperative particle las and the imperative form of the verb. The reported imperative is used to indicate that the source of the command or request is not the speaker but someone else. The addressee of the command can be any person, but similarly to the reported indicative, here, too, the predicate agrees with the subject in number but not in person. The affirmative form of the jussive singular consists of the particle las ~ laz and the k(o)/g(o)-marked form of the lexical verb (Example 19 (KET 146)). The negative singular is formed by the prohibitive verb form algo 'may not' and the k(o)/g(o)-marked form of the verb (Example 20 (KET 128)). The affirmative plural consists of las ~ laz and the kot/got-marked form of the verb (Example 21 (KET 60)), and the negative plural is formed by algo/algod and the kot/got-marked form of the verb (Example 22 (KET 49)).

(19) Ama kitiz, la ma vota-k punda

mother.NOM tell:PST.3SG IMP.PTCL I.NOM take-iMP.SG pound.NOM

zepo, la-k piezzom

soap:PART go-IMP.SG wash:INF

Mother told me to take a pound of soap and to go and wash myself

(20) Un ta irgon kutsort pallo ku a l-go broutso-g

and he start:APP coachman:PART beg:INF that PROH-IMP drive-IMP.SG

nanton taga pits kilgiz-rekidi

they:DAT behind along sideroad:PL.PART

And he is said to have started to beg the coachman not to drive after them using minor roads

(21) Siz iks risting maddon kitiz [--] laz meg lak-ko-t

then one person.NOM we:DAT tell:PST.3SG IMP.PTCL we.NOM go-IMP-PL

sino somo

there receive:INF

'Then someone told us [--] to go there to receive (fishnets)'

(22) Ma kitiz al-go ni lak-ko-t tul jello

I.NOM tell:PST.1SG PROH-IMP now go-IMP-PL fire:GEN house:ILL

'I told them now not to enter the house with a fire'

The origin of the particle las is probably the same as that of the Estonian particle las--the Finnic verb stem developed into a modal and directive particle similarly to the German lass (Metslang 2000). A similar particle with similar functions (lai) is found also in Latvian. The verb form with the marker -k(o)/ -g(o) may have developed either from the imperative third person singular or from the earlier optative (cf. Erelt, Metslang 2004). Estonian, too, has a jussive form marked by -ku/-gu and the modal (imperative) particle las, which has the same origin and is used in the same function. Unlike Livonian, however, in Estonian las is incompatible with the morphological jussive marker -ku/-gu; it occurs either with the inflected indicative form of the verb or with the da-infinitive.

While the reported indicative occurs both in main and subordinate clauses, the jussive occurs mostly in subordinate clauses, usually in complement clauses expressing the report.

Examples like 4, 7, 9 and 19-22 show that reporting is consistently marked with a special mood--the reported indicative or the reported imperative. In texts when reporting several messages, including those that are mutually inclusive or originate from different sources, reported clauses containing predicates in both evidential moods may be intertwined, see also Examples 20 and 23 (KET 141).

(23) Sis ta kitiz, u ma voll-i vigali, aga ma ab

then he.NOM say:PST.3SG either I.NOM be-REP.IND.SG guilty or I.NOM NEG

voll-i vigali, la ma kita-k tuota, aga

be-REP.IND.SG guilty IMP.PTCL I.NOM say-IMP.SG that:PRT whether

ma voll-i ka saksa spiek

I.NOM be-REP.IND.SG also german:GEN spy.NOM

'He then said that I was either said to be guilty, or I was said not to be guilty, I should say whether I was said to be also a German spy

In addition to reporting someone's command, the jussive is used to express necessity, both participant-internal necessity (where the source of necessity is the same as the logical subject of the report, such as Example 24 (KET 62)) and necessity that is conditioned by external circumstances (see Example 25 = Example 9 (SET 210)).

(24) Minnon rosko rodo um, se min um

I:DAT a_little money:PART be.PRS.3SG it.NOM I:GEN be.PRS.3SG

armakstomost ku ma voib leibo entson sodo roz

spare:ACTNM.PART that I.NOM can:PRS.1SG bread:PART self:DAT get:INF a_little

vosto, al-go ma nalgo lak-ko, al-go ma

buy:INF PROH-IMP I.NOM hunger:ILL go-IMP.SG PROH-IMP I.NOM

tallo nalgo kuolo-g, tola tulab jeddo

in_winter hunger:ILL die-IMP.SG winter.NOM come:PRS.3Sg forth

'I have a little money; I have to keep it in order to buy a little bread, so that I won't be hungry, may I not die from hunger in winter, winter will arrive'

(25) sid um keratod,[ku se kenig unso neiz] ku sina

here be.PRS.3SG write:PPP that this king.NOM dream:INE see:PST.3SG that you

li-ji sie kenig tidaron pa miekoks

will_be-REP.iND.SG this.GEN king:GEN daughter:DAT as husband:iNSTR

un laz se kenig jomand votag, andag sinda

and IMP.PTCL this queen:NOM capture:IMP.SG give:IMP.SG you:PART

benda kaddo un laz tiego sinnon tutkam

hangman:GEN hands:ILL and IMP.PTCL do:IMP.SG you:DAT end:GEN

'It is written here [that this king saw in his dreams] that you are reported to get married to this princess, and this queen should capture you and hand you over to the hangman and kill you

By comparison with Estonian, the use of the jussive is more common in the Livonian material. The Estonian jussive forms (marked by -ku/-gu or by the particle las) are very rarely used (Sepper 2005 : 61; EDC). It is more common in Estonian to express a mediated command by means of the conditional; for example, the Estonian equivalent of the jussive in Example 21 would be et me laheksime 'so that we should go'. Both participant-internal and participant-external necessity (dynamic and deontic modality) are usually expressed by means of modal verbs in Estonian; for example, the Estonian translation of Example 25 would be kuninganna peab votma ja andma sind timukale ja tegema sulle lopu 'the queen must capture you and hand you over to the hangman and to kill you. However, in Example 24 the form of the Estonian clause of purpose would require the conditional despite modality: (et) ma nalga ei jaaks, talvel nalga ei sureks 'so that I would not be hungry, not die of hunger in winter'.

Furthermore, it has been observed that the Livonian jussive occurs in deliberative questions (such as How could I do this?) and that such occurrences cannot be found in Standard Estonian (Kehayov, Lindstrom, Niit 2011).

Conclusion: the Livonian reported evidential by comparison with its Estonian correlate(s) and other cross-linguistic evidence

As noted by Walchli (2000), Kehayov (2008), and Koptjevskaja-Tamm and Walchli (2001 : 715-723), the Livonian evidential shares some common features with Estonian and Latvian. The concluding section will list the similarities and the differences between Livonian and Estonian, the closest relative of Livonian among the Finnic languages. The most striking evidentiality-related features shared by the two languages are as follows:

* Both Livonian and Estonian distinguish between the reported indicative (the oblique mood) and the reported imperative (the jussive).

* Both languages make use of bound morphemes (suffixes) to encode the reported indicative and the imperative.

* In both languages the reported indicative seems to have developed as a result of reanalysis of the complement clause of an utterance verb.

* In both languages forms of the reported indicative developed from present participles.

* Both languages use past participles to express the past tense of the reported indicative.

* In both languages the forms of the reported indicative neutralize past tense distinctions in the unmarked indicative.

* Similarly to Estonian, past perfect tense forms are employed in Livonian as an evidential strategy. Furthermore, in both languages the conceptual distance between the speaker s knowledge and the reported propositional content is increased by means of the so-called double perfect.

* In both languages the past tense forms of the reported indicative are conventionalized as main forms of the predicate in popular and traditional narratives.

At the same time, the Livonian reported evidential reveals a number of unique features, which are absent in Estonian, or at least in the standard version of this language:

* Despite the close genetic relationship between the two languages, their reported indicatives differ in terms of category assignment. The present tense form of the Estonian reported indicative is specified for voice whereas its Livonian equivalent lacks voice specification. At the same time, the reported indicative in (Courland) Livonian attracts number specification, which is absent in Estonian.

* Similarly to Latvian but differently from Estonian, the Livonian reported evidential (oblique mood) has future constructions as well as special debitive forms.

* A rather peculiar and typologically rare feature of the Livonian reported evidential is that it coincides in form with the agent noun. Such syncretism is to our knowledge unique in the European languages.

* The Livonian forms of the reported indicative and the reported imperative are more grammaticalized than their Estonian correlates, which is manifested by their frequency, distributional freedom, and semantics:

** Unlike contemporary standard and colloquial Estonian, where reported evidentiality is predominantly expressed by evidential strategies or lexically, Livonian is characterized by predominance of the oblique mood as a marker of reports.

** The Livonian oblique mood is often found in various types of dependent clauses and in direct questions, which is not the case in Estonian.

** Livonian tends to use the reported imperative (the jussive) in indirect commands and wishes and for the expression of modal necessity whereas Estonian prefers to use the conditional mood for the expression of commands/ wishes and modal verbs for the expression of necessity.

* The reported indicative of Livonian resembles formally and distributionally more the respective category in South Estonian than the reported indicative in Standard Estonian: Livonian and South Estonian forms of the reported indicative originate in nominative or genitive participles and occur in dependent clauses whereas their Standard Estonian counterparts are based on a partitive participle and are, as a rule, restricted to independent clauses.

* Unlike Estonian, where the modal (imperative) particle las and the jussive suffix -ku/-gu are incompatible, Livonian often uses the particle and the suffix together in a unified construction.

* The Livonian reported indicative lacks the epistemic overtones characteristic of its Estonian counterpart.

doi: 10.3176/lu.2012.1.04

Abbreviations

ACTNM--action noun, ADE--adessive case, APP--active past participle, CONNEG--connegative verb form, DAT--dative case, GEN--genitive case, ILL--illative case, IMP--imperative, IMPRS--impersonal voice, IND--indicative, INE--inessive case, INF--infinitive, INSTR--instrumental case, NEG--negator, NOM--nominative case, PART--partitive case, PL--plural, PPP--passive past participle, PRF--perfective, PROH--prohibitive verb, PRS--present (tense), PST--past (tense), PTCL--particle, REFL--reflexive pronoun, REP--reported evidentiality marker, SG--singular.

Sources of the examples: EDC--Tartu University's Estonian Dialect Corpus. http://www.murre.ut.ee/estonian-dialect-corpus/; KET--L. Kettunen, Untersuchung uber die Livische Sprache I. Phonetische Einfuhrung. Sprachproben, Tartu 1925 (ACUT B VIII. 3); MAG--J. Magiste, Muistoja Liivinrannasta. Liivin kielta Ruotsista. Suomentanut ja julkaissut Anneli Honko, Helsinki 2006 (MSFOu 250); SET--E. N. Setala, Naytteita liivin kielesta. Suomentanut ja julkaissut Vaino Kyrola, Helsinki 1953 (MSFOu 106); SjW--Joh. Andreas Sjogren's Gesammelte Schriften. Band II. Teil 1. Livische Grammatik nebst Sprachproben. Im Auftrag der Kaiserlichen Akademie der Wissenschaften bearbeitet und mit einer historischethnographischen Einleitung versehen von Ferdinand Joh. Wiedemann, St. Petersburg 1861; SUH--S. Suhonen, Liivin kielen naytteita, Helsinki 1975 (Castrenianumin toimitteita 5).

REFERENCES

Aikhenvald, A. Y. 2004, Evidentiality, Oxford.

Audova, I. 2003, Die ostseefinnischen Deverbativa auf *-pa/-pa und *-ja/*-ja in der Funktion des adjektivischen Attributs.--LU XXXIX, 1-19.

Bashir, E. 2006, Evidentiality in South Asian Languages.--The Proceedings of the LFG-06 Conference. http://cslipublications.stanford.edu/LFG/11/lfg06.html, 30-50.

Erelt, M., Metslang, H. 2004, Grammar and Pragmatics: Changes in the Paradigm Of The Estonian Imperative.--LU XL, 161-178.

Erelt, M., Metslang, H., Pajusalu, K. 2006, Tense and Evidentiality in Estonian.--Topics in Subjectification and Modalization (Belgian Journal of Linguistics 20), 125-136.

Holvoet, A. 2001, Studies in the Latvian Verb, Krakow.

Ikola, O. 1953, Viron ja liivin modus obliquuksen historiaa, Helsinki (Suomi 106, 4).

Isaksson, B. 2000, Expressions of Evidentiality in Two Semitic Languages--Hebrew and Arabic.--Evidentials: Turkic, Iranian and Neighbouring Languages, Berlin--New York (Empirical Approaches to Language Typology 24), 383-399.

Kehayov, P. 2004, Eesti keele evidentsiaalsussusteem mone teise keele taustal. Morfosuntaks ja distributsioon.--KK, 812-829.--2008, An Areal-Typological Perspective to Evidentiality: the Cases of the Balkan and Baltic Linguistic Areas, Tartu.

Kehayov, P., Lindstrom, L., Niit, E. 2011, Imperative in Interrogatives in Estonian (Kihnu), Latvian and Livonian.--LU XLVII, 81-93.

Koptjevskaja-Tamm, M., Walchli, B. 2001, The Circum-Baltic Languages. An Areal-Typological Approach. -Circum-Baltic Languages 2. Grammar and Typology, Amsterdam (Studies in Language Companion Series 55), 615-750.

Krautmane, E. 2006, The Evidential in the Latvian, Estonian and Livonian Languages. An Introduction to the Topic.--Valodnieciba. Somugristika, Riga (Latvijas Universitates Raksti 708), 66-72.--2010, Kuramaa liivi keele koneviiside susteem eesti ja lati keele taustal, Tartu (Unpublished MA thesis. University of Tartu).

LaPo11a, R. J. 2003, Evidentiality in Qiang.--Evidentiality in Typological Perspective, Amsterdam, 63-78.

Metslang, H. 1994, Eesti ja soome--futuurumita keeled?--KK, 534-547, 603-616.

--2000, Reflections on Development of a Particle in Estonian.--Estonian: Typological Studies IV, Tartu (Tartu Ulikooli eesti keele oppetooli toimetised 14), 59-86.

Norvik, M. 2012, Perifrastic Future Devices in Salaca Livonian.--LU XLVIII, 27-40.

Plungian, V. 2001, The Place of Evidentiality within the Universal Grammatical Space.--Journal of Pragmatics 33, 349-357.

Ratsep, H. 1971, Kas kaudne koneviis on koneviis?--Keel ja struktuur 5, Tartu 45-69.

Sepper, M.-M. 2005, Indirektaal 19. sajandi lopu ja 20. sajandi eesti kirjakeeles, Tallinn (Unpublished MA thesis. Tallinn University).

Walchli, B. 2000, Infinite Predication as a Marker of Evidentiality and Modality in the Languages of the Baltic Region.--Sprachtypologie und Universalienforschung 53, 186-210.

Willett, T. 1988, A Cross-Linguistic Survey of the Grammaticalization of Evidentiality.--Studies in Language 12, 51-97.

Address

Petar Kehayov

University of Tartu

petar.kehayov@ut.ee

Helle Metslang

University of Tartu

metslang@ut.ee

Karl Pajusalu

University of Tartu

karl.pajusalu@ut.ee

* The study was supported by the Estonian Science Foundation (grants 7464 and 8409) and by the Estonian Ministry of Education and Research (target-financed research projects SF180084s08 and SF0180076s08).

(1) The lack of impersonal voice marking in the reported indicative forms is a sign that these forms are grammatically more similar to agent nouns (which essentially are always active) than to participles (which are usually specified for voice in the Finnic languages).

(2) Typologically, evidentiality is often left unmarked in conditional clauses; see Aikhenvald (2004 : 255-256,) on Tariana (Arawakan) and LaPolla (2003 : 74-75) on Qiang (Tibeto-Burman).

PETAR KEHAYOV, HELLE METSLANG, KARL PAJUSALU (Tartu)
COPYRIGHT 2012 Estonian Academy Publishers
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Kehayov, Petar; Metslang, Helle; Pajusalu, Karl
Publication:Linguistica Uralica
Article Type:Report
Date:Mar 1, 2012
Words:5943
Previous Article:Perifrastic future devices in Salaca Livonian.
Next Article:Development, research, and sources of written Livonian.

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters