Does Estonian have the jussive?/[TEXT NOT REPRODUCIBLE IN ASCII].

A peculiar feature of Estonian is that the third person imperative marker has extended to all the persons, as well as the impersonal, e.g.

Ma / sa / ta / me / te / nad kirjutagu kiri 'I / you / he / we / you / they should write a letter'; Kirjutatagu kiri 'A letter should be written'.

It is not a recent phenomenon. Such forms can be found already in folk songs and in a 17th-century grammar by J. Hornung. These forms can also be found in the grammars by E. Ahrens and F. J. Wiedemann, e.g.

Saagu sa soossa surema, / Kannu otsa kanguma 'May you die in the swamp / wither on a stump'; Sago minna, piddago minna / sinna / temma 'May I / you / he go' (Hornung 1693 : 76); ehk ma jaagu ilma; ehk ma wajugu maa sisse 'may I be left without it; may I sink into the ground' (Ahrens 1853 : 12); sagu ma enne aegu surema / kadugu sa nenda ma pealt kui kaste rohu pealt; ehk meie jagu ilma, teie olgu kidetud; nahku mina sind, kus ma naen (Wiedemann (1875 : 467-468) 'may my death be premature / may you disappear from the face of earth like dew from the grass; or may we leave without it, may you be praised; may I see you where I can see'.

Karl-August Hermann found a suitable name for these forms in Estonian grammar; he called this paradigm the optative (Hermann 1884 : 99):

"Koneviiside sekka on ka veel arvata

5. Sooviv koneviis ehk luhedalt soovkone (modus optativus) naitab, et midagi soovitakse, ihaldatakse, naituseks: mina palugu, sina palugu, tema palugu, meie palugu, nemad palugu." *

Actually, Mihkel Veske was the first linguist who treated the gu-/ku-marked mood as an independent mood and called it the optative (1). However, he thought that the optative was only used in the first person (Veske 1879 : 71, 74). Harald Pold (1923 : 64) may have discussed the optative after the example of these men. However, he did not provide any examples about the first and second persons of this mood.

The 1960s witnessed a new interest in the existence and peculiarity of this paradigm. Fanny de Sivers, who lived outside Estonia, wrote about the so-called indirect imperative (1969 : 60-61). In Estonia, Mati Hint wrote the following words in his review of the grammar handbook by Johannes Valgma and Nikolai Remmel:

"Juhitagu veel tahelepanu kaskiva koneviisi kolmanda poorde vormi omaparasele uldistumistendentsile, mida ei ole voimalik ka kirjakeeles enam pidurdada. Kolmanda isiku vorm on kindlalt kasutusel ainsuse esimese isiku (mina) puhul, kuid ta levib mujalegi; vrd. lauseid nagu: "Mina tehku tood ja sina muudkui passid"; "Mina muudkui istugu siin paev labi"; "Meie tehku kogu see too ja nemad vahivad pealt"; "Meie mingu koju, nemad voivad peole edasi jaada"; "Teie (sina) tehku tood, nemad saavad palga". Selle uldistatud vormiga valjendatav kask on niisama kaudne voi veel kaudsem kui kolmanda isiku puhul. Kirjakeele grammatika ei voi ignoreerida sellist taiesti asendamatut valjendusvoimalust, mis lisab kaskiva koneviisi vormistikule paralleelse ja taieliku paradigma." (Hint 1969 : 335). **

If M. Hint claims that we are dealing with a new tendency towards generalization, then he is wrong. However, that is not important. What is important is that the observation by M. Hint drew the attention of grammarians to this phenomenon. As noted, M. Hint was unable to specify the role of this paradigm in grammar. Huno Ratsep also took an interest in this phenomenon. His well-known article "Kas kaudne koneviis on koneviis?" (1971) introduced a new category into the Estonian grammatical description--the category of the mode of reporting (Estonian teatelaad). This category expresses the relation between the speaker and the source of the message and has two members: the direct mode of reporting or directal (Estonian direktaal) and the mediated mode of reporting or indirectal (Estonian indirektaal). In the first case the speaker acts at the same time as the source of the message; in the second case the source of the message is someone else, the speaker only mediates the message. H. Ratsep claimed that the gu-/ku-marked imperative, which has extended to all the persons, is the imperative of the mediated mode of reporting. On the other hand, the oblique mood that had hitherto been regarded as an independent mood (ma/sa/ta/me/te/nad kirjutavat kirja 'I/ you/he/we/ you/ they am/is/are said to be writing a letter' is the indicative of the mediated mode of reporting. Thus, according to H. Ratsep, Estonian has only three moods: the indicative, the imperative, and the conditional, whereas the first two occur in two modes of reporting--in the direct and the mediated mode of reporting. According to this treatment, the imperative (leaving aside the past tense and the negative) looks like this:

                        direct imperative
                  (= Estonian otsene imperatiiv)

                        singular                 plural
personal
   1st person      --                         me kirjutagem
   2nd person      sa kirjuta 'write'         te kirjutage
   3rd person      ta kirjutagu               nad kirjutagu
impersonal                         kirjutatagu

                        mediated imperative
               (= Estonian vahendatud imperatiiv)

                        singular                 plural
personal
   1st person      ma kirjutagu 'I            me kirjutagu
                   shuold write'
   2nd person      sa kirjutagu               te kirjutagu
   3rd person      ta kirjutagu               nad kirjutagu

H. Ratsep's theory, however, did not find its way into grammars. Tiit-Rein Viitso's article about the system of Estonian inflectional patterns (1976) mentions briefly that

"Kuigi teatamislaadi analuus sellisena on moeldav, pole see vastavate muuteparadigmade defineerimiseks vajalik. Teatamissituatsiooni voib soovi korral vaadelda parast koneviiside tuvastamist tavalise koneviisi maaratluse --koneviis osutab raakija suhtumisviisi verbi abil valjendatud tegevusse --jargi." (Viitso 1976 : 158). ***

T.-R. Viitso was the third linguist after M. Veske and K. A. Hermann who regarded the gu-/ku-marked imperative an independent mood, calling it the concessive mood (Estonian moonev koneviis) or the jussive. T.-R. Viitso's laconic statement does not explain the background of his claim. It is unclear whether he regards the 'speakers mode of attitude' as the meaning of reportedness (the above quotation could be interpreted this way) or the concessive meaning (as one might conclude judging by the choice of the Estonian term). These two meanings, however, are far from identical. Be that as it may, T.-R. Viitso's system of moods was adopted.

The authors of the academic grammar of Estonian (EKG) adopted the mood system, as suggested by T.-R. Viitso, and also the names for the moods. At the same time, they clearly interpreted the mediacy of a command as a meaning that characterizes the mood and regarded concession and doubt as secondary meanings of these moods. Upon choosing this path the naming of the gu-/ku-marked mood as the concessive mood or the jussive was not fully justified. These terms were adopted because there were no better ones. However, the authors of the academic grammar took a somewhat broader view of mediacy than H. Ratsep, who had in mind only the difference between the speaker and the source of the message. The authors of the academic grammar treated under the mediated message also those cases where the recipient of the message is not the listener but someone else. This gave rise to the need to change the paradigm of the direct imperative. If mediacy is understood so broadly, then the third person imperative is always mediated.

Therefore, it was omitted from the paradigm of the imperative. According to the academic grammar, in the singular the imperative has only the second person, but in the plural it has the first and the second person:

                singular       plural

1st person        --         kirjutagem
2nd person      kirjuta      kirjutage
3rd person        --             --

In addition to the above-mentioned treatment of the imperative, Aarand Roos (1982) has also studied this topic. One should mention, however, that, being an Estonian living abroad, A. Roos was not acquainted with all the research that had been published in Estonia. This would explain the fact that he refers to M. Hint but not to H. Ratsep or T.-R. Viitso. Moreover, A. Roos neglected the papers by M. Veske and K. A. Hermann. Similarly to M. Veske, K. A. Hermann, T.-R. Viitso, and the academic grammar (EKG), A. Roos, too, regards the gu-/ku-marked imperative as an independent mood and calls it the optative like M. Veske and K. A. Hermann. A. Roos, however, understands the essence of the optative as indirectness of the source of the command rather than optativity (Roos 1982 : 12.13).

"Kaesoleva artikli pohjal tahaksin liigituse aluseks seada lause toelise "peremehe", predikaadis peituva motte algataja voi inspireerija. Seega saaks definitsioon jargmise sonastuse: Kui tegelik kaskija-soovija voi tegelike kaskijate-soovijate hulgas on ka koneleja, on imperatiiv otsene ja me nimetamegi koneviisi imperatiiviks. Kui aga kaskijaks-soovijaks on keegi kolmas, on imperatiiv kaudne ja me voiksime seda nimetada kas voi optatiiviks. Selle liigituse pohjal langeksid optatiivi koik juhud, kus kask esineb korvallauses ja pea- ja korvallausel on eri alus, nait. "Opetaja utles, ta kirjutagu kirjatoo valmis."" ****

Actually, A. Roos treats the gu-/ku-marked mood in the same vein as the academic grammar. However, it is only a label because on both occasions it is the indirectness of the source of the message that serves as the distinctive feature of the corresponding mood. The treatment of the imperative by A. Roos differs from all the other treatments in that, according to A. Roos, the Estonian direct imperative has also the first person, which is expressed by the same gu-/ku-marker. For example, in the following sentence by Tammsaare Armastagu ma sind vanaisa, ukskoik kui palju, teda armastan ma ikka rohkem 'No matter how much I love you, grandpa, I love him more anyway' armastagu ma expresses an unfulfilled wish of the speaker. Thus, this form does not belong to the paradigm of the optative (Roos 1982 : 12). However, there is no need to supplement the imperative paradigm in such a way because the speaker may act as the source of the command also for the other persons of the optative, for example, in curses: Kadugu sa maa pealt! 'May you disappear from the face of the earth'. Thus, it appears that the mediacy of the command (wish) is not the best criterion for making a distinction between the imperative and the optative.

This is how our grammarians have treated the gu-/ku- marked paradigm. We are faced with the problem how we should re-write Estonian grammar in the light of current research. At first sight it seems that it is not difficult to harmonize the treatment by H. Ratsep with the contemporary treatments of evidentiality. It would be tempting to place the mediacy of the source of message both in the indicative and the imperative under the category of the referential (reported) evidential (Willett 1988). We might claim that the morphological evidential, which is associated with a command and is very rare in other languages (see Aikhenvald 2000 : 63), does actually exist in Estonian. The other meanings besides mediacy could be treated as secondary.

Below, however, I would like to explain once again the treatment according to which the gu-/ku-marked paradigm is an independent mood and that the imperative has no third person.

About eighty years ago Karl Leetberg (1921 : 50) wrote that "Kaskimine voib sundida ainult sellele, kellega raagitakse, s.o. teises tegijas, aga soovimise mottes ka sellele, kellest raagitakse (kolmandale tegijale): laulgu tema, laulgu nemad." *****

This quotation clearly states that a command that is directed at a third party, that is, an addressee outside the speech situation, is not a proper command. Its implication is rather to emphasize a wish or a need that something should occur in relation to a third party. It does not mean that a third party is additionally made to do something. For example, if the speaker tells the recipient about a person who has no personal relation to the recipient that he looks ill--Ta mingu arsti juurde 'He had better see a doctor', then it does not mean that the recipient has to visit this person and forward the command. It only emphasizes the need to see a doctor, that is Ta mingu arsti juurde has the same meaning as Ta peaks arsti juurde minema 'He had better see a doctor'. Thus, the form under discussion has only a modal meaning, to be more precise, an optative and/or a deontic meaning but not an additional directive meaning as in the second person imperative. In many languages there is no third person form for the imperative; the modality that is associated with the third person is expressed by other means. As in Estonian the modal meaning has extended from the situation-external person to the participants in the speech situation, that is, to the first and the second persons, we need not regard such a form of the third person as an imperative. We can accept the existence of an independent mood as we have actually done in our grammar. (2) We can explain the existence of this mood only by the presence of a modal meaning because the modal meaning may, in fact, be the only meaning of the gu-/ku- paradigm. The optative meaning occurs, for example, in curses of the type Vajugu ma maa sisse! 'May I rather sink in the ground', which constitute the basic examples of the gu-/ku- paradigm but fit in with the contemporary standard language as well. Deontic necessity is expressed, for example, in the following sentence by Arvo Magi, borrowed from a paper by F. de Sivers (1969 : 60): Jaab siga auku kinni, sina lased jalga. Mina muudkui istugu ja oodaku hommikut nagu tuhkur puuris 'A pig gets stuck in a pit, you make off. Why should I be sitting and waiting for the morning like a polecat in a cage'. As for the name of this, then we can keep on calling it the jussive. This term is more or less suitable for our purposes, and one can find it in the grammar of some other languages. We could alternatively call it the optative, like M. Veske and K. A. Hermann and later A. Roos, but perhaps not the concessive mood because concession (Olgu (pealegi)! 'That's all right', Olgu, mis on! 'Be that as it may', Tulgu, kes tahes! 'Anyone is welcome') is not the main meaning of this mood. However, deontic meaning is often accompanied by a reference to an indirect source of the message, that is, to evidential meaning. It is not impossible that the evidential meaning may have developed from the deontic meaning. Even in Estonian, for example, the modal verb pidama 'must' has given rise to a semi-auxiliary that expresses indirect evidentiality, e.g. Ta pidi homme siia tulema 'He is supposed to come here tomorrow' (Erelt 2001). Evidentiality may proceed from the modal meaning not only in statements but also in commands because someone's specific wish serves as one of the main sources of deontic modality. One can see it explicitly in the use of the jussive in indirect speech, e.g. Juku utles, et ma olgu vait = et ma pean vait olema 'Juku told me that I should keep quiet'.

One might claim that the distinctive feature of the Estonian jussive is modality rather than evidentiality. We cannot speak here about morphological evidentiality but about an additional modal meaning, an evidential strategy (Aikhenvald 2000). One should emphasize once again that evidentiality as manifested in the gu-/ku- paradigm has nothing to do with the indirectness of the recipient, as the academic grammar erroneously claims; it is only the indirectness of the source of the message. This erroneous view later served as a basis for attempts to show that the past participle in sentences of the type Tulnud ohtul oigel ajal koju! 'You shouldn't have come home at a proper time at night' is a past jussive form if it refers to the third person and a past imperative form if it refers to the second person (Muiczniece, Metslang, Pajusalu 1999 : 147-148). In previous treatments this form was regarded as a past imperative for all the persons (Saareste 1952 : 37; de Sivers 1969 : 76-77; Magiste 1976). I think that semantically in all persons, as well as in the impersonal (Tuldud ohtul oigel ajal koju 'One should have come home at a proper time at night.), it is the simple past of the jussive or the optative. This form does not express a command but rather a wish, whereas it implies an evidential result (someone's information leads to a conclusion that some action should have been performed in the past). The compound form of the type olgu tehtud 'get it done' could be regarded as a perfect form of the jussive or the optative. Actually, it expresses the future in the past (an action reaching a deadline in the future), e.g. Olgu see too homseks tehtud 'See to it that this job will be finished by tomorrow.. The imperative, however, has no past forms.

By way of conclusion, I will present the imperative and jussive paradigms (in the affirmative) in the way they should look like on the basis of the above treatment.

                          imperative                     jussive

personal
 present      --        kirjutagem    ma kirjutagu    me kirjutagu
            sa kirjuta  te kirjutage  sa kirjutagu    te kirjutagu
              --           --         ta kirjutagu    nad kirjutagu
preterite     --           --             --               --
                                      sa kirjutanud   te kirjutanud
                                      ta kirjutanud   nad kirjutanud
perfect       --           --         ma olgu         me olgu
                                      kirjutanud      kirjutanud
                                      sa olgu         te olgu
                                      kirjutanud      kirjutanud
                                      ta olgu         nad olgu
                                      kirjutanud      kirjutanud
impersonal
  present     --           --                  kirjutatagu
  perfect     --           --                  olgu kirjutatud

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

Ahrens, E. 1853, Grammatik der Ehstnischen Sprache Revalschen Dialektes, Reval.

Aikhenvald, A. Y. 2000, Evidentials. Melbourne.

de Sivers, F. 1969, Analyse grammaticale de l'estonien parle, Clermont-Ferrand.

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 19), 7-25.

Hermann, K. A. 1884, Eesti keele grammatik, Tartu.

Hint, M. 1969, Eesti grammatikakirjanduse pohimottelised ja konkreetsed probleemid. --KK 1969, 327-341.

Hornung, J. 1693, Grammatica Esthonica, Riga.

Leetberg, K. 1921, Eesti keele grammatika, Tallinn.

Muiczniece, L., Metslang, H., Pajusalu, K. 1999, Past Participle Finitization in Estonian and Latvian.--Estonian: Typological Studies III, Tartu (Tartu Ulikooli eesti keele oppetooli toimetised 11), 128-157.

Magiste, J. 1976, Viron imperatiivin preteriti.--Vir. 1976, 48-51.

Pold, H. 1923, Eesti keeleopetus koolidele. I jagu. Kolmas, parandatud trukk, Tallinn.

Roos, A. 1982, Imperatiivi mina-vormist ja optatiivist.--Finsk-Ugriska smaskrifter 5, Lund, 3-14.

Ratsep, H. 1971, Kas kaudne koneviis on koneviis?--Keel ja struktuur 5, Tartu, 45.69.

Saareste, A. 1952, Kaunis emakeel I, Lund.

Veske, M. 1879, Eesti keele healte opetus ja kirjutuse viis, Tartus.

Viitso, T.-R. 1976, Eesti muutkondade susteemist.--KK 1976, 148-162.

Willett, T. 1988, A Cross-Linguistic Survey of the Grammaticalization of Evidentiality.--Studies in Language, vol. 12, no. 1, Amsterdam, 51-97.

Wiedemann, F. J. 1875, Grammatik der Ehstnischen Sprache, St.-Petersbourg.

* 'One should include among the moods

5. The optative mood (modus optativus), which shows that something is wished, desired, for example, I should beg, you should beg, he should beg, we should beg, they should beg.'

** 'Let me draw your attention to the peculiar agreement of the third person imperative that we will be unable to stop also in the written language. The third person form is widely used for the first person (mina 'I'), but it is spreading elsewhere, cf. such sentences as "Mina tehku tood ja sina muukui passid" 'Why should I do the work and you're just idling'; "Mina muukui istugu siin paev labi" 'Why should I be sitting here the whole day'; "Meie tehku kogu see too ja nemad vahivad pealt" 'Why should we do all the work and they're just watching'; "Meie mingu koju, nemad voivad peole edasi jaada" 'Why should we go home, but they can remain at the party'; "Teie (sina) tehku tood, nemad saavad palga" 'Why should you do the work and they will be paid'. A command expressed by this generalized form is as indirect or even more indirect than in the case of the third person. A grammar of the written language should not ignore this unique possibility of expression that serves as an additional alternative and full paradigm of the imperative mood'.

*** 'Although such an analysis of the mode of reporting is conceivable, it is not needed for defining the corresponding inflectional paradigms. If necessary, one can discuss the situation of reporting after the moods are established according to the usual definition of the mood--a mood refers to the speaker's mode of attitude to an action expressed by means of a verb.'

**** 'On the basis of this article the classification should be based on the real "master" of the sentence, the initiator or inspirer of the thought to be found in the predicate. Thus, the definition could be worded as follows: if the actual commander-wisher is the speaker or the speaker is among the actual commanders or wishers, the imperative is direct, and we call this mood the imperative. However, if the commander or wisher is a third party, then the imperative is indirect, and we could call it the optative, if we like. According to this classification, the optative would cover all those cases where the command occurs in a subordinate clause and the main and the subordinate clause have different subjects, e.g. Opetaja utles, ta kirjutagu kirjatoo valmis 'The teacher said that he should write up his work'.'

***** 'Giving orders concerns only the person that one is talking, that is, another person, but in the sense of wishing also the person under discussion (a third party): laulgu tema, laulgu nemad 'he should sing, they should sing'.'

(1) Most early grammars used the term 'optative' for the conditional.

(2) M. Veske may have proceeded from the same considerations when he separated the third person imperative from the imperative and regarded it as an independent mood. However, we cannot guess a possible reason why K. A. Hermann treated the gu-/ku- forms as optative forms but retained them also as third person imperative forms.

MATI ERELT (Tartu

[TEXT NOT REPRODUCIBLE IN ASCII].