Hostname: page-component-745bb68f8f-hvd4g Total loading time: 0 Render date: 2025-02-06T13:13:00.789Z Has data issue: false hasContentIssue false

Estonian conditional clauses: The degree of hypotheticality and the link to temporal and concessive clauses

Published online by Cambridge University Press:  12 June 2013

Helen Plado*
Affiliation:
Helen Plado, Jakobi 2–430, 51014 Tartu, Estonia. helen.plado@ut.ee

Abstract

Estonian conditional clauses have previously been divided into two clear-cut groups: real and unreal, with indicative and conditional main verbs of conditional clauses, respectively. This article defends the view that it is a question of the degree of hypotheticality that a sentence conveys, and it treats hypotheticality as a continuum that includes groups of linguistic forms, which have a relatively clear core and are separated by fuzzy transition areas. Secondly, the article concentrates on the relationship between Estonian conditional clauses and temporal clauses. As these clause types have the same marker (kui), the article discusses whether it is always possible to distinguish between these two clauses and which factors are relevant for determining whether the clause is a temporal or conditional one. Thirdly, the relationship between Estonian conditional and concessive clauses is under consideration, focusing particularly on Estonian scalar concessive conditional clauses.

Type
Research Article
Copyright
Copyright © Nordic Association of Linguistics 2013 

1. INTRODUCTION

In Estonian, conditional sentences are one of five sub-types of complex sentences with adverbial clauses that express cause-and-effect relations.Footnote 1 Unlike the other types (e.g. causal, purpose, concessive and consecutive sentence), the sub-type of conditional sentences conveys hypothetical causality. A conditional clause sets the scene for the event referred to in the main clause. If the condition expressed in the subordinate clause is realized, the state of affairs or the event described in the main clause will take place. In Estonian, in addition to conditional sentences proper, there are also other types of constructions that can express conditional relations, as well as conditional clauses that do not set conditions but are, rather, used to express a wish (see (1a) below, compared to conditional sentence proper in (1b)), a suggestion, or a question.Footnote 2 Usually, these clauses are not embedded in an (overt) main clause. Conditional clauses that express questions seem to be common across languages, as demonstrated among others by Vallauri (Reference Vallauri2004) primarily with respect to Italian, but in passing also for other languages, such as German, Swedish, Finnish, and Japanese.

  1. (1)

In this article, I will focus on Estonian conditional clauses proper. I will examine the issue of their varying degrees of hypotheticality as well as the links between Estonian temporal and conditional clauses, which are marked by the same subordinator on the one hand, and between conditional and concessive clauses, on the other hand. The article consists of five sections: Section 2 introduces the form of Estonian conditional clauses, Section 3 describes the position of conditional mood in the Estonian mood system, Section 4 demonstrates how conditional clauses have been classified, Section 5 discusses the relationship between real and unreal conditional clauses in Estonian, and Section 6 discusses the connections between conditional clauses and temporal and concessive clauses in Estonian.

The data analyzed originates primarily from the 1990s sub-corpus of the Corpus of Written Estonian (http://www.cl.ut.ee/korpused/kasutajaliides) from the University of Tartu. This sub-corpus was chosen because at the time of collecting the material for this article, it was the most recent body of textual material (fiction and newspapers texts) that formed a sufficiently large sample. The sub-corpus of the 1980s, which contains far fewer texts but whose material falls into three distinct categories of formality (fiction, news, and scientific texts) was used when I needed additional data to investigate the use of conditional markers in different language registers. Since the two sub-corpora held relatively few occurrences of scalar conditional concessive clauses, I expanded my research to the sub-corpus that at the time contained texts from the end of the 1990s and the beginning of the 2000s. That sub-corpus consists of news texts from the nation-wide daily newspaper Postimees and the weekly Eesti Ekspress.

Firstly, I introduce the form of Estonian conditional sentences (markers and possible main verb forms of conditional as well as main clauses) and also the Estonian conditional mood, as it plays a crucial role in the interpretation of Estonian conditional sentences.

2. THE FORM OF ESTONIAN CONDITIONAL SENTENCES

2.1 Markers of Estonian conditional sentences

In the conditional constructions in most languages, either the protasis (the condition) or the apodosis (the consequent) is marked, or sometimes both are marked. Although in some languages (such as Mandarin Chinese) the marking is not overt, it is the overt marking of the protasis that characterizes most languages (see Comrie Reference Comrie1986:87).

The protasis is also marked obligatorily in Estonian conditional sentences, while the marking of the apodosis is optional. Estonian conditional clauses either begin with the conjunction kui ‘if’, as in (2a), or bring the verb up to the initial position, as in (2b) (see also e.g. Erelt et al. Reference Erelt, Kasik, Metslang, Rajandi, Ross, Saari, Tael and Vare1993:308).Footnote 3 Corpus data show that the use of the conjunction kui is much more frequent than fronting the verb of the clause. Additionally, a verb may only be placed in the initial position if the conditional clause precedes the main clause.

  1. (2)

The Estonian conditional conjunction kui (as well as the Finnish kun, see e.g. Ikola et al. Reference Ikola, Palomäki and Koitto1989:93–94; Herlin Reference Herlin1998:160–163, which shares its origin with the former) is a polysemous word. In addition to conditional use, it can appear as a temporal (see Section 5.1), concessive (see Section 5.2), comparative, identifying or copulative conjunction (EKSS). Historically, kui is the instructive form of the pronominal stem ku- and is related to other conjunctions that have evolved from the same stem (the stem has been retained in the pronoun kumb ‘which (of the two)’; see Karelson Reference Karelson1959:187; Mägiste Reference Mägiste2000:1039), cf. the Estonian kuna ‘while, because, whilst’, kuigi ‘although’, kuni ‘until’, and kuid ‘but’ (see Mägiste Reference Mägiste2000:1019–1020, 1045–1047).

In addition to the protasis, the apodosis can also be marked in Estonian, although in that case the use of a marker is optional (in my data, the apodosis was marked in about 41% of conditional sentences). The apodosis may be marked either by siis ‘then’, as in (3a) below, or the adessive or translative case of the word juht ‘case’ (juhul and juhuks, respectively), as in (3b). The marker siis is used most frequently, and it is the only possible marker of the apodosis in the case of conditional clauses whose verb occupies the clause-initial position.

  1. (3)

Footnote 4 The choice of the marker for the main clause depends on the degree of formality of the text – the more formal the register is, the more often some form of the word juht is used. While in fiction the main clauses marked by juhul or juhuks constitute only 2.4% of all marked main clauses, the corresponding figure for news texts is 6.4% and 20% for research texts.

The next sub-section demonstrates which verb forms can be used in Estonian conditional clauses. Remarks are also made about the form of the verb of the main clause.

2.2 The main verb of Estonian conditional clauses

The main verb of the Estonian conditional clause can be in the indicative mood, as in (4a), the da-infinitive form, as in (4b), or in the conditional mood, as in (4c).

  1. (4)

According to the Estonian academic grammar (i.e. Erelt et al. Reference Erelt, Kasik, Metslang, Rajandi, Ross, Saari, Tael and Vare1993:308), conditional clauses with the verb in indicative mood or the da-infinitive form express real conditions, and clauses with the verb in the conditional mood express unreal conditions. One aim of the present paper is to find out whether this statement holds.

The verb of the main clause can be in different moods. Therefore, the mood of the verb of the main clause does not depend on the mood of the verb of the conditional clause (Erelt et al. Reference Erelt, Kasik, Metslang, Rajandi, Ross, Saari, Tael and Vare1993:308).

In the case of the degree of hypotheticality of conditional clauses, the distinction between the indicative and the conditional mood of the main verb plays a crucial role. Therefore, a short overview of the Estonian mood system, the functions of the conditional mood, and the verbal markers of conditional mood is provided.

3. CONDITIONAL MOOD IN ESTONIAN

There are five moods in Estonian: indicative, imperative, conditional, quotative and jussive. Indicative is an unmarked mood (both in form and meaning) that forms an opposition with the quotative on the basis of indirectness, with the imperative on the basis of directivity, and with the conditional on the basis of unreality (Metslang & Sepper Reference Metslang, Sepper, Rothstein and Thieroff2010:533). Jussive is a peripheral mood that expresses an indirect order (Erelt et al. Reference Erelt, Kasik, Metslang, Rajandi, Ross, Saari, Tael and Vare1993:37).

The Estonian conditional mood marker is -ks(i), which is followed by the personal ending in the first and the second person singular and all persons in the plural. However, it is possible to leave the personal ending out; in that case all the forms are identical; for example, ma, sa, ta, me, te, nad söö-ks ‘I, you, (s)he, we, you, they would eat’. The Estonian conditional mood has two tenses: present and past, the latter can be formed either analytically or synthetically. The analytical form consists of the verb olema ‘to be’ in a conditional mood and the past participle form of the main verb; for example, ole-ks(i-n) söö-nud ‘I would have eaten’. The synthetic past form is an artificial compound form introduced by Loorits (Reference Loorits1922) (söö-nu-ks(i-n) ‘I would have eaten’).

According to Erelt et al. (Reference Erelt, Kasik, Metslang, Rajandi, Ross, Saari, Tael and Vare1993:34), the main function of the conditional mood is to mark the unreality of the event or the situation. In addition to unreality, the conditional mood can also (mostly at the same time) express desirability, as in (5a), consideration, as in (5b) (Metslang Reference Metslang and Erelt1999:109,112; Metslang & Sepper Reference Metslang, Sepper, Rothstein and Thieroff2010:540), proposing, and guessing, as in (5c); these additional functions are often served by modal verbs in the conditional mood (see Metslang & Sepper Reference Metslang, Sepper, Rothstein and Thieroff2010:540).

  1. (5)

The pragmatic use of the condition to express politeness is also widespread (particularly with peaks as the directive ‘should’, and with tahaks ‘would like’ to soften a statement or wish) (see Metslang & Sepper Reference Metslang, Sepper, Rothstein and Thieroff2010:540). In such cases, the conditional mood softens the speech act (see Pajusalu & Pajusalu Reference Pajusalu and Pajusalu2004:264).

4. TYPES OF CONDITIONAL CLAUSES

Conditional clauses have been divided mainly on the basis of either the degree of hypotheticality they convey or the cognitive domain in which the events or state of affairs expressed by the conditional and the main clause are considered. Firstly, I will introduce conditional clause types based on how probable the realization of the events or state of affairs is regarded to be, and secondly, the classification of conditional clauses made on the basis of the cognitive domain within which the assumptions expressed by the conditional and the main clause fall.

4.1 Types of conditional clauses according to the degree of probability they convey

When classifying conditional clauses in the basis of how probable the realization of the event or state of affairs described in the conditional clause is considered to be, a distinction is commonly made between two main types: real conditionals and unreal conditionals. Real conditionals refer to real present, as in (6a), habitual or generic situations, as in (6b), or to past situations, as in (6c) (Thompson, Longacre & Hwang Reference Thompson, Longacre, Hwang and Shopen2007:255).

  1. (6)
    1. a. If it's raining out there, my car is getting wet.

    2. b. If you step on the brake, the car slows down.

    3. c. If you were at the party, then you know about Sue and Fred.

Unreal conditionals refer to unreal imaginary situations. Two types of unreal conditionals are often distinguished: hypothetical, as in (7a), and counterfactual, as in (7b). Hypothetical conditionals refer to situations that might arise, while counterfactual conditionals are used to denote events that did not or even could not happen (Thompson et al. Reference Thompson, Longacre, Hwang and Shopen2007:256).

  1. (7)
    1. a. If I were you, I'd talk once again with him.

    2. b. If you had been at the office yesterday, you would have met our new college.

Similar to the classification above, Estonian linguistics (Erelt et al. Reference Erelt, Kasik, Metslang, Rajandi, Ross, Saari, Tael and Vare1993:308; Külmoja Reference Храковский1998) divides conditional clauses into two groups: real and unreal. A slightly different approach has been suggested by Rannut (Reference Rannut1981), who distinguishes between three types of conditional clauses: real, possible, and unreal. However, she also points out that possible and unreal conditional clauses take an identical form in Estonian and should, for this reason, be treated together, which leaves us with two main types of conditional clauses: the real and the hypothetical.Footnote 5

A different classification of conditional clauses, based on hypotheticality as a continuum, is proposed by Comrie (Reference Comrie1986:88–93). According to Comrie, there are no clear-cut divisions, and different languages distinguish different degrees of hypotheticality along this continuum. By the term ‘hypotheticality’, Comrie means the degree of probability of the realization of the situations referred to in the conditional clause. In the continuum of hypotheticality, he distinguishes greater and lower hypotheticality – greater hypotheticality means lower probability and lower hypotheticality greater probability. For Comrie (Reference Comrie1986), a sentence stating a fact would represent the lowest degree of hypotheticality and a counterfactual sentence the highest degree. In many European languages (e.g. German, Russian, Latvian, and English – languages Estonian has had or still has close ties to), the degree of hypotheticality is indicated by the choice of the verb form. Hypotheticality is also treated as a continuum in the Finnish academic grammar (i.e. Hakulinen et al. Reference Hakulinen, Vilkuna, Korhonen, Koivisto, Heinonen and Alho2004:1081), in whose authors’ view expressions containing a verb in the conditional mood evoke a pronounced degree of hypotheticality and leave the realization of the situation open.

In Section 5, I discuss whether Estonian conditional clauses form clear-cut groups, as described so far in the Estonian linguistics literature, or whether we should instead consider a continuum of hypotheticality (as suggested by Comrie (Reference Comrie1986) from the typological point of view). For this reason, I will not take clear-cut groups (that are formed according to the main verb form of the conditional clause) of conditional clauses as a premise, but rather concentrate on the degree of hypotheticality (in Comrie's sense) and examine corpus data to determine whether or not they fall into clear-cut groups.

4.2 Conditional clauses in different cognitive domains

Most causal and conditional clauses can be divided according to the cognitive domain within which the assumptions expressed by the subordinate and the main clauses fall. Although Rutherford (Reference Rutheford1970) discussed some adverbial clauses and conjunctions that have different functions, and Lowe (Reference Lowe and Monaghan1987) had already divided causal clauses on similar bases in 1987, this approach gained popularity after the publication of Sweetser (Reference Sweetser1990). According to Sweetser (Reference Sweetser1990), there are three types of conditional clauses:

(i) Conditional clauses that act in the content domain – these are ‘prototypical’ conditional clauses – those that show the sufficient conditions for the realization of the event or state of affairs described in the main clause; for example, in sentence (8) the precondition for my coming is his coming.

  1. (8) If he comes, I will also come.

In this type of conditional sentence, there is a cause-and-effect relationship between the event or state of affairs of the conditional and main clauses; in sentence (8), his coming is the cause of my coming, which is the effect. This is also the type of conditional clause that can convey different degrees of hypotheticality.

(ii) Conditional clauses that act in the epistemic domain – this type of conditional clause describes the premise(s) on the basis of which one can conclude the event or state of affairs described in the main clause; in other words, there is an inferential relationship between the event or state of affairs described in the conditional and main clause. In sentence (9), one can conclude from the situation described in the conditional clause that the situation expressed in the main clause holds.

  1. (9) If the lights are on, he must be at home.

The cause-and-effect relationship between events or the state of affairs described in the protasis and the apodosis is mostly reversed (compared to content conditional sentences): the cause is expressed in the apodosis and the effect in the protasis (Dancygier Reference Dancygier1998:86; Itkonen Reference Itkonen2001:334). In sentence (9), his being at home is the reason for the lights, not vice versa. However, there can be an inferential relationship between the event or state of affairs described in the conditional and the main clause also in sentences where the protasis expresses the cause and the apodosis the effect, as shown in example (11) below.

(iii) Conditional clauses that act in the speech-act domain – these conditional clauses comment upon the speech acts performed in the main clause and make the speech act relevant in the communicative situation. In sentence (10), being thirsty is not a sufficient condition for the juice being in the fridge; rather, the speaker makes the offering of juice (via the main clause) relevant in the situation – the offering is relevant if the partner is thirsty.

  1. (10) If you are thirsty, there is juice in the fridge.

Sometimes (e.g. in Dancygier Reference Dancygier1998; Dancygier & Sweetser Reference Dancygier, Sweetse, Couper-Kuhlen and Kortmann2000, Reference Dancygier and Sweetser2005) metalinguistic conditional clauses have also been regarded as a distinct type of conditional clause. However, neither the content of the speech act conditional clauses nor metalinguistic conditional clauses deliver the condition of the event or state of affairs in the main clause, but rather comment on the speech act or some word expressed by/in the main clause. For that reason Plado (Reference Plado and Erelt2008, Reference Plado2010) has discussed these types of clauses together, and referred to this group as ‘conversational conditional clauses’.

It is debatable whether we can talk about hypotheticality in the case of epistemic and speech act conditional clauses. As epistemic conditionals seem to be non-predictive, they are not used with hypothetical forms (Dancygier Reference Dancygier1998:87–88). For example, if we add greater hypotheticality to sentence (11a), the epistemic reading gives way to a reading of content space in (11b). However, the conditional mood does not rule out an epistemic reading in all cases, as demonstrated in example (11c). The sentence in (11c) can be used, for example, in the situation where the speaker and the hearer disagree about whether the person in question has taken entrance examinations, but they both know that the person did not go to the university in 1995.

  1. (11)

In the case of speech act conditionals, hypotheticality is not relevant. However, Sweetser (Reference Sweetser1990:118–119) notes that although the speech act is always accomplished, the speech act conditional clause adds some politeness and allows the hearer a little more room to maneuver.

Dancygier (Reference Dancygier1998) connects the two approaches discussed in the previous and this sub-section and differentiates two types of conditional clauses: predictive (which includes content conditional clauses, both open and hypothetical) and non-predictive (which includes mostly epistemic and speech-act conditionals).

In the following section, I will discuss types of Estonian conditional sentences according to how probable the realization of the events or states of affairs described in the conditional and the main clause is regarded as being. As this is relevant in the case of content conditional sentences, I will concentrate on these.

5. DEGREE OF HYPOTHETICALITY IN ESTONIAN CONDITIONALS

All earlier works on Estonian conditional clauses cited in Section 4.1 (Rannut Reference Rannut1981, Erelt et al. Reference Erelt, Kasik, Metslang, Rajandi, Ross, Saari, Tael and Vare1993:308, Külmoja Reference Храковский1998) are based on the premise that there is a direct link between the form of the main verb and the meaning of the conditional clause – conditionals whose main verb is in the indicative mood or in the da-infinitive express a real conditional meaning, while conditionals whose main verb is in the conditional mood express an unreal meaning.Footnote 6 According to Erelt et al. (Reference Erelt, Kasik, Metslang, Rajandi, Ross, Saari, Tael and Vare1993:308), in Estonian, the verb of real conditional clauses can be either in the indicative or in the form of the da-infinitive.Footnote 7 There is no requirement that the verbs of the conditional and main clauses be in the same mood (see (12)). As for unreal conditional clauses, their verbs are always in the conditional mood (Erelt et al. Reference Erelt, Kasik, Metslang, Rajandi, Ross, Saari, Tael and Vare1993:308).

  1. (12)

Although Erelt et al. (Reference Erelt, Kasik, Metslang, Rajandi, Ross, Saari, Tael and Vare1993) do not make a clear distinction between the categories of realis and irrealis, the premise that there is a clear correspondence between the mood of the main verb and the meaning of the conditional clause, and also the claim that the main function of the conditional clause is to mark irreality, hint at the existence of the grammatical categories of realis and irrealis.

On the other hand, Monticelli (Reference Monticelli2003:354) demonstrates that in Estonian, there is no one-to-one correspondence between the modality of the sentence and the form of the main verb of the conditional clause. He shows that two sentences in which the main verbs of the protases are in different moods (indicative and conditional) can be closer with respect to modality than two sentences whose respective main verbs of the protasis are in the conditional. In this section of the paper, I will test whether or not the above-mentioned correspondence between form and meaning holds. I will thus explore whether, in the case of Estonian, it is justified to talk about grammatical categories of realis and irrealis that are marked by the indicative and conditional mood, respectively. I will also examine the question whether we should treat hypotheticality as a binary choice or as a scale with clear groups at either end but a diffuse transition area between them.

As conditional clauses set a framework for the event or state of affairs in the main clause, the question is whether the modal meaning of a conditional sentence is determined exclusively by its conditional clause or if it also depends on the form of the main verb and the meaning of the main clause.

Conditional sentences analyzed here come from the sub-corpus of the 1990s, which includes fiction and newspaper texts.Footnote 8 The following analysis is based on 3021 conditional sentences (for general characteristics of the sample, see Table 1).

Table 1. Conditional sentences analyzed.

5.1 Conditional clauses with the verb in the indicative mood

Conditional sentences where the verb of the conditional clause is in the indicative mood may have varying degrees of hypotheticality. Sentences that express past situations, such as (13a), recurring situations or habitual/generic situations, such as (13b, c), or which refer to typical sequences of events in the natural world, such as (13d), usually exhibit a low degree of hypotheticality. Conditional clauses that express past situations will often refer to habitual and/or recurring events in the past. They usually express a low degree of hypotheticality because from the present point of view, it is already known that the event or state of affairs has occurred. As the conjunction kui marks both temporal and conditional clauses, clauses with the verb in the past tense are often regarded as temporal (see also Section 6.1 below). However, the clause with the verb in the past tense can also refer to a non-recurring event or state of affairs, as in sentence (13e). In such cases, the sentence expresses a higher degree of hypotheticality and it is considered a conditional sentence. In cases like (13b–d), there is background knowledge that the sequence of the content of the conditional and main clauses is somehow generic: it has happened before and it will probably happen likewise in the future, or even every time the content of the conditional clause is realized. This knowledge is often also supported by the adverbs that express the recurring nature of the situation, for example, tihtipeale ‘often’ in (13b) and alati ‘always’ in (13c).

  1. (13)

Sometimes sentences whose main verbs have the same form still exhibit differing degrees of hypotheticality. For instance, there are sentences that express the probability of a situation, as in (14a), and/or imaginative sentences in which someone imagines what could be, while the actual realization of the imagined event/situation is left open, as in (14b), or the event/situation is regarded as (rather) unlikely, as in (14c).

  1. (14)

In the case of (14a), the adverb nüüd ‘now’ supports the interpretation that the realization of the event of the conditional clause is regarded to be possible or even probable. If that adverb was not there, the sentence would have an imaginary reading with a open reference for realization. In sentence (14b), the hearer actually does not get any clues how probable the realization of the situation described in the sentences is regarded to be. In sentence (14c), however, the previous context (the steersman's doubt and also irony in his voice) and especially the adverb isegi ‘even’ in the main clause refer to a higher degree of hypotheticality.

In some examples illustrated above, the adverb of the main clause has an effect on the interpretation of the conditional sentence. I will now describe whether the conditional mood of the verb in the main clause affects the interpretation of the degree of hypotheticality of the conditional sentence. In the data, there are only 120 conditional sentences in which the verb of the subordinate clause is in the indicative mood and the verb of the main clause in the conditional mood. Although the data sample is too small for definitive conclusions, we can nevertheless observe some tendencies.

The use of the conditional mood in the main clause can have different effects on the interpretation of the conditional clause and the whole sentence. Although the use of the conditional in the main clause in most cases hardly affects the degree of hypotheticality in the sentence, in some cases conditional verbs can significantly boost the hypotheticality. For example, in (15a) the use of a conditional verb in the main clause affects the interpretation that it is remarkably unlikely for the conversation partner to be in a similar situation (also the adverbial ühel päeval ‘one day’, which refers to an unknown time in the future, supports the interpretation of an imaginary situation). By contrast, if an indicative verb had been used in the main clause (15b), the probability would have been rather high and correspondingly, the degree of hypotheticality would have been low. For example, if it is known that there are significant problems at the company and all the workers are treated in the same manner, and so the conversation partner would probably be in the same position at some time, an indicative verb could be used.

  1. (15)

However, as also Metslang & Sepper (Reference Metslang, Sepper, Rothstein and Thieroff2010:540) point out, conditional mood with modal verbs are mostly used to express desirability, consideration, proposals, guessing, and politeness. Accordingly, the use of the conditional mood in the main clause usually co-occurs with modal verbs and so softens the utterance and expresses a polite proposal, as in (16a), or adds a shadow of doubt, as in (16b), rather than remarkably increasing the degree of hypotheticality of the sentence.

  1. (16)

Thus, it can be concluded that the construct kui + verb in the indicative mood is rather unmarked in the sense of hypotheticality. Although in the majority of cases it is used to express rather low degrees of hypotheticality (i.e. the events or state of affairs whose realization is regarded as probable), higher degrees of hypotheticality may also be expressed by this construct. When the subordinate clause is in the indicative mood, the conditional mood is rarely used in the main clause; however, it can be used (i) with modal verbs, where its effect on the degree of hypotheticality of the sentence is negligible, or (ii) with non-modal verbs, where it substantially affects the degree of hypotheticality. The next sub-section concentrates on conditional sentences where the main verb of the conditional clause is in da-infinitive form.

5.2 Conditional clauses with the verb in da-infinitive form

As infinitives are not marked by the categories person, number, tense, and mood, it is interesting to identify which factors become important for interpreting the degree of hypotheticality in the conditional sentence.

Sentences in which the verb of the conditional clause appears as a da-infinitive and the verb of the main clause is in the indicative mood mainly have a low degree of hypotheticality. Such conditional sentences express mostly habitual, generic or recurring events or states of affairs, as (17) illustrates.

  1. (17)

Sentences where the conditional clause contains a da-infinitive form and the main clause is in the conditional mood are quite rare; the data analyzed included only 35 such sentences. This type of sentence can express a lower as well as a higher degree of hypotheticality. Consider (18).

  1. (18)

In (18a) there is a conditional modal verb, which does not increase the degree of hypotheticality, but adds a note of caution and politeness and thereby softens the utterance, whereas in (18b) and (18c) the use of the conditional mood in the main clause increases the degree of hypotheticality and invites the reader to imagine a situation in which the conditions are realized. In sentence (18b), in addition to the conditional mood in the main clause, contextual cues also play a role in the imaginative interpretation: the question of which concert venue one of Estonia's biggest towns would need is under discussion, and hence the participants in the communicative situation imagine what the positive and negative aspects of different concert halls are. Sentence (18c) also does not provide any additional lexical cues for the imaginary interpretation but again, there are contextual cues for the imaginary interpretation in this example. A really fat man is under discussion, and the participants of the situation described by the sentence (boys) know that the person would not actually deflate.

Despite the small sample size, we can conclude that in the case of conditional sentences with a da-infinitive verb in the protasis, the degree of hypotheticality depends primarily on the verb form of the main clause: if it is in the indicative mood, the degree of hypotheticality of the sentence is low, and if it is in the conditional mood, the degree of hypotheticality of the sentence is high. (This does not seem to hold in the case of modal verbs in the main clause.)

Finally, I will discuss the conditional sentences with the main verb of the conditional clause in the conditional mood.

5.3 Conditional clauses with the verb in the conditional mood

In terms of their degree of hypotheticality, conditional sentences with the verb of the conditional clause in the conditional mood are more homogeneous than sentences with a conditional clause whose verb is in the indicative mood or the da-infinitive form. Conditional clauses in which the verb is in the present conditional express imaginative conditions: what would happen if the event or state of affairs described in the subordinate clause became true. Although all these sentences have a high degree of hypotheticality, this degree also varies to a certain extent. A lower degree of hypotheticality in this case means that someone imagines a situation that does not exist at the moment, but whose realization in the future is left open, as in (19a), whereas the sentences with a higher degree of hypotheticality express a situation that is at odds with the current state of affairs, as in (19b), and this is rather unlikely, (19c), or even impossible, (19d).

  1. (19)

Example (19c) describes an imaginary situation that does not hold at the moment; however, the context, where the hearer has already learned the Bible by heart, demonstrates that the situation is not regarded as totally impossible (as, for example, in (19d)).

Conditional clauses in which the verb is in the past conditional are counterfactual – they express conditions that did not or could not have come true. This type of conditional sentence is used to speculate what would have happened or what could be different at the moment of speaking if the events expressed in the conditional clause had happened (20a, b).

  1. (20)

As can be seen in (20b), sentences with a past verb in the main clause and a present conditional verb in the conditional clause still remain counterfactual.

The construct kui + verb in the conditional mood clearly refers to a high degree of hypotheticality. However, there is also some variation in this group in the sense of the degree of hypotheticality.

The degree of hypotheticality is primarily indicated by the mood of the main verb of the conditional clause. However, the connection between the indicative and conditional mood of the verb in the protasis and the degree of hypotheticality is not absolute. Additionally, the mood used in the main clause, and the lexical (mostly adverbs) and contextual factors affect the hypotheticality of the conditional sentence.

The next section concentrates on the connections that conditional clauses have with temporal and concessive clauses, and also investigates which role hypotheticality plays in these clause types.

6. THE RELATIONSHIP OF ESTONIAN CONDITIONAL CLAUSES TO OTHER ADVERBIAL CLAUSES

In this section I will describe the relationship of the Estonian conditional clause and closely related adverbial clauses – temporal and concessive clauses. Firstly, the links between Estonian temporal and conditional clauses are discussed, followed by those between conditional and concessive clauses, and concluding with a discussion of concessive conditional clauses.

6.1 The relationship between temporal and conditional clauses in Estonian

There are numerous examples in the world's languages of conditional markers evolving from temporal markers (see Heine & Kuteva Reference Heine and Kuteva2002:293, 326). Both Traugott (Reference Traugott and Haiman1985:290–292) and Kortmann (Reference Kortmann1997:197) claim that the sources of conditional conjunctions are mainly temporal conjunctions and other time expressions. Since, in many cases, such linguistic units retain the historical meaning that they possessed at an earlier stage, there are many languages in which the same conjunction or expression is used to mark both temporal and conditional clauses (see Thompson et al. Reference Thompson, Longacre, Hwang and Shopen2007:257) and/or in which conditional markers also express temporality to a certain degree (Tabakowska Reference Tabakowska, Athanasiadou and Dirven1997:278–280). However, it is also important to note that the conditional relationship between two events or states of affairs always involves temporal relationships: the effect described by the main clause cannot take place earlier than the condition (or cause) described by the conditional clause.

Linguists have long argued over whether it is possible to distinguish between temporal and conditional clauses in languages which mark both types of clauses using the same means, that is, whether it is always possible to tell if a certain clause is a temporal or conditional one. To some extent, this seems to be a language-specific issue. However, Thompson et al. (Reference Thompson, Longacre, Hwang and Shopen2007:257) claim that there are many languages in which only conditional clauses that predict future events and temporal clauses that express the future are indistinguishable.

Estonian is one of the languages in which the conditional conjunction (kui) and the temporal conjunction have the same form.Footnote 9 Moreover, the optional apodosis marker siis of the main clause (see Section 2.1 above) can occur in conjunction with both conditional and temporal subordinate clauses. However, based on the context, in most cases it is possible to interpret a subordinate clause as temporal or conditional, depending on the presence or absence of causal relations between the event or state of affairs described in the subordinate and main clause, and the degree of hypotheticality that characterizes the meaning expressed in the subordinate clause. Still, this ambiguity is relevant only in content conditional clauses – clauses that express the direct condition of the event or state of affairs described in the main clause and that convey the interpretation of some degree of hypotheticality. For example, sentence (21) receives a temporal reading if the interpretation is that the speaker might be sure that the people referred to will at some point have a car. It receives a conditional reading if the interpretation is that the speaker only imagines how these people would behave if the state of affairs expressed in the subordinate clause were (ever) to come true.

  1. (21)

  1. (i) ‘How we cussed them, swearing that if we got cars we would certainly not be like them.’

  2. (ii) ‘How we cussed them, swearing that when we got cars we would certainly not be like them.’

Hence, the hypotheticality becomes one of the distinctive factors – if the sentence contains at least some degree of hypotheticality, it gets a conditional reading. The second important factor for distinguishing between temporal and conditional sentences is the presence or absence of the cause-and-effect relationship between the events or states of affairs expressed by the subordinate and main clause: there is no (explicit) cause-and-effect relationship between situations expressed by the temporal and main clause, but the relationship is present in the case of situations delivered by the conditional and its main clause.

However, this ambiguity in the meaning of the sentence can also be intentional (e.g. when the speaker does not want to make a straightforward promise). At the same time, in conversation, it is possible to lose the ambiguity by emphasizing the conjunction kui by extra word stress. In this case, the subordinate clause acquires strong conditional meaning. (The same effect is sometimes achieved in written texts by writing the conjunction in capital letters.) Also the marker juhul can be used in the main clause to show a straight conditional relationship between the events or states of affairs described in the subordinate and the main clauses.

One important factor for determining the degree of hypotheticality, and therefore for determining whether a clause is temporal or conditional, is tense. As there is no grammatical future in Estonian (the future is expressed using the present tense), the relevant distinction is between the present and the past. If the event took place in the past, it is generally (but not always) characterized by a low degree of hypotheticality because the event is known to have taken place. Such clauses are interpreted as temporal, as in (22a). However, sentence (22a) would acquire a conditional reading if it expressed a recurring event or state of affairs in the past; in other words, if the same situation took place more than once and the reference to a prisoner was non-specific. Furthermore, it could also acquire a conditional reading if the verb of the conditional and main clause was in the present tense and the sentence referred to the future, as in (22b).

  1. (22)

A special case is the use of the perfect tense in subordinate clauses because on the one hand, the perfect tense implies that the event took place in the past but on the other hand, it establishes a reference to the present. In certain situations, the perfect tense in Estonian may even refer to the future (Erelt et al. Reference Erelt, Kasik, Metslang, Rajandi, Ross, Saari, Tael and Vare1995:77–78). For example, in (23) the use of the past tense in the main clause does not affect the interpretation of the probability of the event expressed in the subordinate clause; in the main clause the vantage point is in the past, and from this perspective the event expressed in the subordinate clause will take place in the future. It is not clear whether the operation in Mexico is completed but the implication is that it will probably come to an end.

  1. (23)

To summarize, there are three factors that are relevant for determining whether a clause is temporal or conditional: the presence of a cause-and-effect relationship between the event or state of affairs described in the subordinate and the main clause, the presence of hypotheticality, and the tense. The first two factors indicate the conditional relationship between the situations expressed by the two clauses. The factor of tense is not equally absolute: although in most cases the past tense refers to temporal relationships between the situations, it does not hold in all cases. As Section 5.1 shows, clauses with the main verb in the past can also express situations with conditional relationships.

I will next discuss the fuzzy transition area between temporal and conditional clauses. Sentences (24a–f) demonstrate this most clearly.

  1. (24)

The subordinate clause in (24a) clearly expresses time only; there is no cause–effect relationship between the state of affairs described in the protasis and the apodosis, and the situations of the sentence are not marked by hypotheticality as these have already happened. In (24b) the subordinate clause also possesses shades of a conditional meaning – it acquires a temporal reading but it also determines the conditions necessary for the (recurring) event in the main clause to take place (drawing may take place only when everybody has fallen asleep). In (24c) one can no longer tell whether its primary interpretation should be one of temporality or of conditionality, since both are possible. In (24d) the subordinate clause would probably acquire a conditional reading, yet the use of the adverb alles ‘only/yet, not earlier than’ also allows a temporal reading, as alles in that position typically specifies the temporal relationship and brings the temporal adverbial into focus. In (24e) the subordinate clause is clearly conditional – it sets the conditions necessary for the event in the main clause to occur. At the same time it is unclear whether the event will actually take place. Finally, sentence (24f) is a step further: the verb in both the subordinate and the main clause is in the conditional mood, which is explicitly more hypothetical.

In addition to temporal sentences, conditional sentences are also rather closely related to concessive sentences. The next sub-section describes the relationship between these two types of adverbial sentences.

6.2 The relationship between conditional and concessive clauses in Estonian: Estonian concessive conditional clauses

Estonian has several means of marking concessive clauses. The most typical is the conjunction kuigi ‘although’ (Karu 2006:69; Plado & Lindström Reference Plado, Lindström and Erelt2012:141), illustrated in (25). In its form, it combines the temporal and/or conditional clause marker kui and the emphasis particle -gi/-ki.

  1. (25)

Concessive conditional clauses have the characteristics of both a conditional and a concessive clause. A conditional clause expresses a condition that facilitates the occurrence of the event of the main clause, while a concessive conditional clause expresses an unfavourable condition that should prevent the event in the main clause from taking place but does not. This links concessive conditional clauses to concessive clauses. Concessive clauses differ from concessive conditional clauses in that the former do not express conditions necessary for the event in the main clause to take place. Instead, in concessive clauses it is understood that events referred to in the main and the subordinate clause will all take place. Thus, on the one hand, while the subordinate clause of a concessive conditional sentence resembles a conditional clause in that it delivers conditions and contains hypotheticality, one does not know for sure whether or not the situation expressed by this clause will take place. On the other hand, the main clause of the sentence resembles the main clause of a concessive sentence as the event or state of affairs described in it will take place in any case, regardless of the unfavourable factors described in the subordinate clause.

Haspelmath & König (Reference Haspelmath, König, van der Auwera and O'Baoill1998:563) distinguish three types of concessive conditional clauses: scalar, alternative and universal. I have provided Estonian examples for these types: (26a) includes a scalar, (26b) includes an alternative, and (26c) includes a universal concessive conditional clause.

  1. (26)

Since scalar concessive conditional clauses are closely related to conditional clauses in Estonian, it is this type of clause – which Erelt et al. (Reference Erelt, Kasik, Metslang, Rajandi, Ross, Saari, Tael and Vare1993:308) classifies as conditional – that will form the exclusive focus of this section.

Concessive conditional clauses tend to evolve into concessive clauses and are thus among the historical sources for the development of concessive clauses (Haspelmath & König Reference Haspelmath, König, van der Auwera and O'Baoill1998:567–568). According to König (Reference König1986:240–241), concessive conjunctive words have often evolved from temporal and/or conditional conjunctive words that mean ‘even’, ‘also’, or ‘and’. Examples include the English even though, even so; the German obwohl, obgleich, obschon, wenngleich, wenn auch; the French quand même; the Latin et si; the Finnish joskin, etc. They were probably first used in conditional and concessive conditional clauses, which were distinguished on the basis of their context. The prototypical Estonian concessive conjunction kuigi has also previously been used as a marker of concessive conditional sentences (Plado & Lindström Reference Plado, Lindström and Erelt2012:141–143).

In Estonian, scalar concessive conditionals are constructed by introducing an additional element into the conditional. There are five ways in Estonian to use a conditional clause as a scalar concessive conditional clause as illustrated in (27).

  1. (27)

The first type involves adding the adverb isegi ‘even’ to the subordinating conjunction kui thereby forming the subordinating conjunctive phrase isegi (siis/juhul,) kui as in (27a). It is probably not the case that the Estonian concessive conditional marker isegi kui is perceived as a single unit, because in one-third of the cases the correlate siis or juhul as well as a comma are inserted between isegi and kui. The second type consists of adding the particle ka ‘also; and’, as in (27b), the third adds the emphatic particle -gi/-ki (which can sometimes mean ‘also; and’) to the verb in the subordinate clause, as in (27c). The fourth inserts the adverb veel ‘still/more’ into the clause, as in (27d) and is very rare. The fifth simply uses the conditional form to convey a concessive conditional meaning, as in (27e).

The adverb isegi and the particles -gi/-ki and ka hint that the realization of the situation described in the clause is regarded as rather improbable, i.e. these markers increase the degree of the hypotheticality of the subordinate clause, yet at the same time they also reinforce the reality of the events described in the main clause (Koks Reference Koks2004:500). Therefore, in the case of concessive conditional clauses it is important to keep separate the degree of hypotheticality in the subordinate and the main clause.

The most frequent strategy for marking concessive conditional clauses in Estonian is the addition of the particle -gi/-ki to the subordinate clause (33% of the instances in the 1990s sub-corpus) and using the conjunctive phrase isegi (siis/juhul,) kui (30%). The other strategies were the use of the conditional clause (19%) and the addition of the particle ka (17%). It should be noted that conditional clauses are often used to express concessive conditional meanings in fiction but seldom in media texts. The use of the adverb veel is marginal (1%).

The adverb ikka ‘still’ (plus the emphatic particle -gi) may be added to the main clause to emphasize its concessive conditional meaning, as is shown in (28) below. If one wishes to use the form of a conditional clause to convey a concessive conditional meaning, then one mostly adds a particle to the main clause to emphasize that meaning (27e).

  1. (28)

As seen above, conditional clauses may be used to express concessive conditional meanings. At the same time, the subordinate clause containing the conjunction kui may even receive concessive interpretations, as in (29), yet this rare use suggests that it has not become fully conventional yet.

  1. (29)

Kuigi – the prototypical concessive conjunction in current Estonian – has previously also been used as a marker of concessive conditional sentences. The first instances of its use can be found in 19th century texts.Footnote 10 During the period 1890–1910, kuigi appeared both in concessive conditional clauses, shown in (30a), and concessive clauses, shown in (30b).

  1. (30)

The material analyzed from the corpus of Written Estonian shows that the development of the modern use of the conjunction kuigi was essentially complete by the 1950s. Thus, it could already then be used to mark a concessive conditional, provided the verb of the subordinate clause was in the conditional mood (31) (see also Karu Reference Karu, Klaas and Tenjes2004:56).

  1. (31)

7. CONCLUSION

Estonian conditional clauses either begin with the conjunction kui ‘if’, or bring the verb up to the initial position. The use of the conjunction kui is much more frequent than fronting the verb of the clause. The marking of the apodosis is optional.

The conditional sentences analyzed in this study show that in Estonian, hypotheticality appears to be a continuum that includes groups of linguistic forms, which have a relatively clear core and are separated by fuzzy transition areas. Although the degree of hypotheticality varies both between groups and within them, in most cases, conditional sentences in which the main verb of the protasis is in the indicative mood or takes the form of the da-infinitive have a low degree of hypotheticality; however, sometimes this kind of conditional sentence conveys a higher degree of hypotheticality. Conditional sentences in which the verb of the protasis is in the conditional mood are more homogeneous. Conditional clauses whose verb is in the present conditional express imaginary conditions, whereas conditional clauses in which the verb is in the past conditional are counterfactual – they express conditions that did not or could not have come true.

To sum up, the construct kui + a verb in the indicative mood is unmarked with respect to hypotheticality – although in the majority of cases it is used to express rather low degrees of hypotheticality (i.e. the event or state of affairs whose realization is regarded to be probable), higher degrees of hypotheticality can also be expressed by this construct. This construct is even unmarked in the sense of conditionality, as it can also express a temporal relationship between the situations expressed by a sentence. The construct kui + a verb in the da-infinitive is also unmarked in the sense of hypotheticality; the degree of hypotheticality is primarily referred to by the verb form in the main clause. In contrast, the construct kui + a verb in the conditional mood is marked by a sense of hypotheticality – it refers to the interpretation of high degrees of hypotheticality.

It also appears that although the main role in determining the degree of hypotheticality in an Estonian conditional sentence belongs to the conditional clause, the main clause too has a part to play. That part is particularly explicit in cases where the main verb of the subordinate clause takes a non-finite form that cannot be marked for tense or mood. However, in conditional clauses with finite verb forms, the main clause can also increase or decrease the degree of hypotheticality in the whole sentence significantly.

The Estonian conditional conjunction (kui) has the same form as a temporal conjunction. There are three factors that are relevant for determining whether the clause is a temporal or a conditional one: the presence of a cause-and-effect relationship between the event or state of affairs described in the subordinate and main clause, the presence of hypotheticality, and the tense. The analysis of the corpus material has demonstrated that although it is, in most cases, possible to interpret a subordinate clause as a temporal or conditional one, there are clauses that simultaneously express time and conditionality and that cannot be identified as temporal or conditional even with the help of the context.

The Estonian conditional marker is also historically linked through grammaticalization to the markers of concessive conditional and concessive clauses. Concessive conditional and sometimes even concessive meaning can be conveyed with a sentence that is marked as conditional. The commonly used Estonian concessive conjunction kuigi ‘although’ is clearly formed from the temporal and/or conditional clause marker kui and the emphasis particle -gi/-ki.

ACKNOWLEDGEMENTS

I would like to thank professor Mati Erelt and the reviewers of the article for their highly useful remarks. I am also grateful to our working group at the Institute of Estonian and General Linguistics at the University of Tartu for reading the paper and discussing it with me. The work reported in the article was supported by Estonian Science Foundation grant no. 8595.

APPENDIX

Abbreviations used in example glosses:

1, 2, 3 = first, second, third person; abess = abessive; abl = ablative; adess = adessive; all = allative; com = comitative; comp = comparative; cond = conditional mood; el = elative; ess = essive; gen = genitive; ill = illative; imp = imperative; impers = impersonal voice; in = inessive; inf = infinitive; neg = negation; part = particle; pl = plural; pr = present; prt = partitive; prtc = participle; pst = past; q = polar question; sg = singular; term = terminative; trsl = translative

Footnotes

1. I use ‘conditional clause’ for protasis and ‘conditional sentence’ for a clause combination instantiating a conditional relation.

2. The example source annotations AJA and AJAE mark example sentences taken from news texts of the corresponding decade in the University of Tartu's Corpus of Written Estonian. Abbreviations used in example glosses are listed in the appendix.

3. The abbreviation ILU marks example sentences taken from fiction of the corresponding decade in the University of Tartu's Corpus of Written Estonian.

4. Juhul can also be placed into the protasis with the conjunction (Saad surematuks, juhul kui lähed vasakule). The position of juhul is a question about focusing: if juhul is in the main clause, then it (and the whole conditional clause) is focused; if it is in the conditional clause, then it is non-focused (Erelt Reference Erelt1986:33).

5. This distinction coincides with the distinction between realis and irrealis in the sense of Palmer (Reference Palmer2001:145, 187), according to whom realis and irrealis are grammatical categories. In the case of Estonian, realis and irrealis would be marked by the indicative and the conditional mood, respectively.

6. However, Külmoja (Reference Храковский1998) mentions that conditional clauses with the verb in the indicative mood can also be used to denote the future.

7. In Estonian, there are certain types of adverbial clauses in which the verb may appear in the infinitive. The conditional clause, however, is the only one in which the implicit actor of the main verb does not have to be identical to the actor of the main clause. Main verb in the da-infinitive form can only occur in conditional clauses which are marked by the conjunction kui. For a discussion of Estonian conditional clauses with the da-infinitive verb, see Plado (Reference Plado2010).

8. For technical reasons (unfortunately, it was not possible to search the sub-corpus of the 1990s for sentences or clauses that begin with a verb), only conditional clauses marked by the conjunction kui are taken into account.

9. This holds only for the marking of the subordinate clause by the conjunction kui. As for verb-initial subordinate clauses there is no such an ambiguity. To distinguish temporal and conditional kui-clauses, I used questioning and substitution. In the former case I tested whether the question ‘When?’ or ‘On what condition(s)?’ was appropriate; in the latter case I tried to substitute the conjunction either with sel tingimusel, et ‘on the condition that’, or with sel ajal kui ‘at that time when’.

10. However, the corpus of Early Written Estonian (which includes 409,722 word units; http://www.murre.ut.ee/vakkur/Korpused/korpused.htm) does not yield a single instance of the use of the conjunction kuigi from the 19th century. Unfortunately the 19th-century material in the corpus for the most part stems from the first half of the century, leaving the years 1850–1890 largely uncovered.

References

REFERENCES

Comrie, Bernard. 1986. Conditionals: A typology. In Traugott et al. (eds.), 77–99.Google Scholar
Dancygier, Barbara. 1998. Conditionals and Prediction: Time, Knowledge and Causation in Conditional Constructions. Cambridge: Cambridge University Press.Google Scholar
Dancygier, Barbara & Sweetse, Eve [E.]r. 2000. Constructions with if, since, and because: Causality, epistemic stance, and clause order. In Couper-Kuhlen, Elizabeth & Kortmann, Bernd (eds.), Cause – Condition – Concession – Contrast: Cognitive and Discourse Perspectives, 111142. New York: Mouton de Gruyter.Google Scholar
Dancygier, Barbara & Sweetser, Eve [E.]. 2005. Mental Spaces in Grammar: Conditional Constructions. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
Erelt, Mati. 1986. Liitlause eesti keeles. Eesti-soome kontrastiivseminar [Complex sentence in Estonian: Estonian–Finnish contrastive seminar] (Helsingi 1986). Tallinn: Eesti NSV Teaduste Akadeemia.Google Scholar
Erelt, Mati, Kasik, Reet, Metslang, Helle, Rajandi, Henno, Ross, Kristiina, Saari, Henn, Tael, Kaja & Vare, Silvi. 1993. Eesti keele grammatika II. Süntaks [The grammar of Estonian II: Syntax]. Tallinn: Eesti TA Keele ja Kirjanduse Instituut.Google Scholar
Erelt, Mati, Kasik, Reet, Metslang, Helle, Rajandi, Henno, Ross, Kristiina, Saari, Henn, Tael, Kaja & Vare, Silvi. 1995. Eesti keele grammatika I. Morfoloogia. Sõnamoodustus [The grammar of Estonian I. Morphology: Word formation]. Tallinn: Eesti TA Eesti Keele Instituut.Google Scholar
EKSS = Eesti keele seletav sõnaraamat [Explanatory dictionary of Estonian]. http://www.eki.ee/dict/ekss/index.cgi?Q=kui&F=M (accessed 10 March 2013)Google Scholar
Hakulinen, Auli, Vilkuna, Maria, Korhonen, Riitta, Koivisto, Vesa, Heinonen, Tarja Riitta & Alho, Irja. 2004. Iso suomen kielioppi [The big grammar of Finnish]. Helsinki: Suomalaisen Kirjallisuuden Seura.Google Scholar
Haspelmath, Martin & König, Ekkehard. 1998. Concessive conditionals in the languages of Europe. In van der Auwera, Johan & O'Baoill, Dónall P. (eds.), Adverbial Constructions in the Languages of Europe, 563640. Berlin & New York: Mouton de Gruyter.CrossRefGoogle Scholar
Heine, Bernd & Kuteva, Tania. 2002. World Lexicon of Grammaticalization. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
Herlin, Ilona. 1998. Suomen kun [Finnish kun]. Helsinki: Suomalaisen Kirjallisuuden Seura.Google Scholar
Ikola, Osmo, Palomäki, Ulla & Koitto, Anna-Kaisa. 1989. Suomen murteiden lauseoppia ja tekstikielioppia [The syntax and text grammar of Finnish dialects]. Helsinki: Suomalaisen Kirjallisuuden Seura.Google Scholar
Itkonen, Esa 2001. Maailman kielten erilaisuus ja samuus [The diversity and the unity of the world's languages]. Turku: Turun yliopisto.Google Scholar
Karelson, Rudolf. 1959. Sidesõnad läänemeresoome keeltes [Conjunctions in Finnic languages]. Dissertatsioon filoloogiakandidaadi teadusliku kraadi taotlemiseks, Tartu.Google Scholar
Karu, Katrin. 2004. Mitmetähenduslikud mööndlaused eesti ja vene keeles [Polysemous concessive clauses in Estonian and Russian]. In Klaas, Birute & Tenjes, Silvi (eds.), Emakeel ja teised keeled IV [Estonian and other languages IV], 5563. Tartu: Tartu Ülikooli Kirjastus.Google Scholar
Karu 2006 = Кару, Катрин 2006. Уступительные конструкции в эстонском и русском языках [Concessive constructions in Estonian and Russian]. Ph.D. dissertation, University of Tartu.Google Scholar
Koks, Helen. 2004. Eesti kirjakeele mööndtingimuslausest [The concessive conditional clauses of literary Estonian]. Keel ja Kirjandus 7, 496503.Google Scholar
König, Ekkehard. 1986. Conditionals, concessive conditionals and concessives: Areas of contrast, overlap and neutralization. In Traugott et al. (eds.), 229–246.Google Scholar
Kortmann, Bernd. 1997. Adverbial Subordination: A Typology and History of Adverbial Subordinators Based on European Languages. Berlin & New York: Mouton de Gruyter.CrossRefGoogle Scholar
Külmoja 1998 = Кюльмоя, Ирина. 1998. Условные конструкций в эстонском языке [Conditional constructions in Estonian]. In Храковский, В. С. (ed.), Типология условных конструкций [Typology of conditional constructions], 350370. Санкт-Петербург: Наука.Google ScholarPubMed
Loorits, Oskar 1922. Tingiva ja kaudse kõneviisi minevik [Past tense of conditional and quotative mood]. Eesti Keel 5–6, 170173.Google Scholar
Lowe, Ivan. 1987. Two ways of looking at causes and reasons. In Monaghan, James (ed.), Grammar in the Construction of Texts, 3746. London: Frances Pinter.Google Scholar
Mägiste, Julius. 2000. Estnisches etymologisches Wörterbuch IV. 2. Auflage kuhtuma – loom. Helsinki: Finnisch-Ugrische Gesellschaft.Google Scholar
Metslang, Helle. 1999. Is the Estonian and Finnish conditional actually a conditional? In Erelt, Mati (ed.), Estonian: Typological Studies III, 97127. Tartu: University of Tartu.Google Scholar
Metslang, Helle & Sepper, Maria-Maren. 2010. Mood in Estonian. In Rothstein, Björn & Thieroff, Rolf (eds.), Mood in the Language of Europe, 528550. Amsterdam & Philadelphia, PA: John Benjamins.CrossRefGoogle Scholar
Monticelli, Daniele. 2003. Eesti keele tingimuslausete liigitamise semantilised aspektid [Some semantic aspects of the classification of Estonian conditional sentences]. Keel ja Kirjandus 5, 353361.Google Scholar
Pajusalu, Renate & Pajusalu, Karl. 2004. The conditional in everyday Estonian: Its form and functions. Linguistica Uralica 4, 257269.CrossRefGoogle Scholar
Palmer, Frank Robert. 2001. Mood and Modality, 2nd edn.Cambridge: Cambridge University Press.CrossRefGoogle Scholar
Plado, Helen. 2008. Adverbiaallausete funktsioonipõhine liigitus eesti keele põhjus- ja tingimuslause näitel [Classification of adverbial clauses, based on Estonian conditional and causal clauses]. In Erelt, Mati (ed.), Emakeele Seltsi aastaraamat 53 (2007), 122145. Tallinn: Teaduste Akadeemia Kirjastus.Google Scholar
Plado, Helen. 2010. Eesti keeled da-infinitiivis öeldisverbiga tingimusalused [Conditional clauses with da-infinitive predicate in Estonian]. ESUKA – JEFUL 2, 255272.CrossRefGoogle Scholar
Plado, Helen & Lindström, Liina. 2012. Eesti keele mööndlause: markerid ning koht alistuse ja rinnastuse teljel [Estonian concessive clause: Markers and position in the chain of subordination and coordination]. In Erelt, Mati (ed.), Emakeele Seltsi aastaraamat 57 (2011), 131161. Tallinn: Teaduste Akadeemia Kirjastus.Google Scholar
Rannut, Lehte. 1981. Põimlause eesti keeles [Complex sentence in Estonian]. Tallinn: Valgus.Google Scholar
Rutheford, William E. 1970. Some observations concerning subordinate clauses in English. Language 46 (1), 97115.CrossRefGoogle Scholar
Sweetser, Eve E. 1990. From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
Tabakowska, Elżbieta. 1997. Conceptualization: Conditionals as an instance of figure/ground alignment. In Athanasiadou, Angeliki & Dirven, René (eds.), On Conditionals Again, 273288. Amsterdam: John Benjamins.CrossRefGoogle Scholar
Thompson, Sandra A., Longacre, Robert E. & Hwang, Shin Ja. 2007. Adverbial clauses. In Shopen, Timothy (ed.), Language Typology and Syntactic Description, vol. II: Complex Constructions, 237300. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
Traugott, Elizabeth Closs. 1985. Conditional markers. In Haiman, John (ed.), Iconicity in Syntax, 289307. Amsterdam: John Benjamins.CrossRefGoogle Scholar
Traugott, Elizabeth Closs, ter Meulen, Alice, Reilly, Judy Snitzer & Ferguson, Charles A. (eds.). 1986. On Conditionals. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
Vallauri, Edoardo Lombardi. 2004. Grammaticalization of syntactic incompleteness: Free conditionals in Italian and other languages. SKY Journal of Linguistics 17, 189215.Google Scholar
Figure 0

Table 1. Conditional sentences analyzed.