2.1 Some previous studies

Bustos (Reference Bustos1986: 232) explains the occurrence of the indicative mood in si-clauses from a formal pragmatic perspective of linguistic analysis. The analysis departs from the notion that both the particle si and the subjunctive mood presuppose that the speaker is neutral regarding the veracity of the clause, and that natural languages tend to avoid redundant information. Within this line of reasoning, the subjunctive mood would duplicate the information provided by the si-clause, which is why the indicative mood occurs in si-clauses.Footnote ^[4] Bustos (Reference Bustos1986) focuses primarily on mood choice in si-clauses, whereas conditionals such as those displayed in examples (2) and (3) above are left outside the scope of analysis.

From a similar theoretical perspective, Haverkate (Reference Haverkate2002) makes a distinction between the realis, potential and irrealis interpretations of si-clauses, and comments that clauses with realis and potential interpretations trigger the indicative mood, whereas the subjunctive mood occurs in clauses with an irrealis interpretation. Haverkate (Reference Haverkate2002: 171–173) illustrates the difference between the realis and the potential interpretation of si-clauses with the following examples:

The difference between examples (4) and (5) relates to the description of causal links between events in the real world, as in (4), and the description of events that have the potential to occur in the real world, as in (5). Haverkate (Reference Haverkate2002: 171–173) argues that an example like (4) describes generic factuality, which from the perspective of formal logic has a ‘material implication’, i.e. it has the illocutionary force of a statement on general laws, rules and principles that apply whenever the described condition is met. By contrast, the causal link in (5) is purely ‘based on a virtual state of affairs’. That is, the causal link between the two events is not factual in character but hypothetical (see also RAE 2009: 3550–3551).

Haverkate (Reference Haverkate2002) observes that the distinction between realis and potential implies certain difficulties in finding an explanation for the indicative mood in cases such as (5) above.Footnote ^[5] In order to provide an explanation, he refers to Bustos (Reference Bustos1986) and to the idea that natural language tends to avoid redundant information (see above). He also argues that the particle si expresses ‘a condition, or a possibility, and natural language does not seem to allow redundancy by allowing the possibility to be expressed again in the verb’ (Haverkate Reference Haverkate2002: 173). That is, the subjunctive mood would duplicate the condition or the possibility expressed by the particle si. Following the explanation provided by Lunn (Reference Lunn1989), Haverkate (Reference Haverkate2002: 173) further hypothesizes that a high degree of relevance of both the protasis – the subordinate clause – and the apodosis – the main clause – could explain the indicative mood in si-clauses.

The occurrence of the subjunctive mood in other Spanish conditional clauses is explained in accordance with the semantic meaning of the conjunction that introduces the conditional clause. The conditional conjunction en caso de que ‘in the case that’ emphasizes the potential character of the proposition, which motivates the occurrence of the subjunctive mood. By contrast, the occurrences of the subjunctive mood in clauses that are introduced by conditional conjunctions such as con tal de que ‘provided that’, bajo/con la condición de que ‘under/with the condition that’ and siempre que ‘as long as’ is explained by their directive force (Haverkate Reference Haverkate2002: 177).Footnote ^[6]

In a mental spaces analysis, Mejías-Bikandi (Reference Mejías-Bikandi2009) argues that the particle si represents a space builder that sets up a domain parallel to the speaker’s view of reality (see Fauconnier Reference Fauconnier1994: 115; Mejías-Bikandi Reference Mejías-Bikandi2009: 164). The indicative mood is thus a natural consequence of the fact that it designates asserted information. In other conditional clauses, the truth of the proposition depends on the fulfillment of the condition, and this is why the subjunctive mood occurs. It designates non-asserted information. Mejías-Bikandi (Reference Mejías-Bikandi2009: 164) illustrates this difference with the following formulas:

According to Mejías-Bikandi (Reference Mejías-Bikandi2009), these formulas describe the semantic relations that hold between the clauses. A prototypical conditional conjunction creates a semantic relation between the two clauses in which the achievement of the main clause event depends on the condition expressed in the conditional clause. By contrast, the particle si does not establish any particular semantic relation between the two clauses: ‘Rather, si invites us to consider a situation in which the proposition expressed by the protasis holds. At the same time the assertion is made that the proposition expressed in the apodosis also holds in this situation’ (Mejías-Bikandi Reference Mejías-Bikandi2009: 169).

From the standpoint of the present study, the analyses offered by Bustos (Reference Bustos1986) and Haverkate (Reference Haverkate2002) imply some explanatory problems. The first one is related to the notion of redundancy of information. Although this concept is useful for explaining the indicative mood in si-clauses, it seems to conflict with the occurrence of the Spanish subjunctive mood in potential contexts, i.e. after expressions such as quizás ‘maybe’, tal vez ‘maybe’ and es possible que ‘it is possible that’. Haverkate (Reference Haverkate2002) explains the occurrence of the subjunctive mood in potential contexts by the potential meaning of the subjunctive mood, but this would further suggest that the subjunctive mood repeats the information given by the conditional conjunction. Another problem concerns the theoretical notions of realis and potential. The distinction implies a static view of causal relations. In fact, the notion of material implication suggests that a causal relation only occurs in cases of the type: Always X${>}$Always Y. As example (5) above clearly shows, however, a potential si-clause can be causal in nature.

If human reasoning about causes and effects is taken into account, the indicative mood in example (5) can be explained by the fact that the speaker considers the causal relation to be in accordance with his/her experience of causes and effects in the world. Even though we do not always leave if someone is late, it is true that there are times when we lose our patience and do just that. Accordingly, the indicative mood occurs in a context in which the speaker expresses what s/he considers to be a causal relation between two events: X late${>}$Y leave. That is, the indicative mood is consistent with the speaker’s way of conceptualizing the causal force of the si-clause.Footnote ^[7]

2.2 The present study

The explanation offered by Mejías-Bikandi (Reference Mejías-Bikandi2009) is in many ways compatible with the one offered in the present paper. However, unlike Mejías-Bikandi, who explains the indicative mood in si-clauses within the theoretical framework of mental spaces, the present paper will focus on the causal relation between the si-clause and the main clause. That is, the ‘relevant inference’ drawn by the conceptualizer is causal in character. Although the causal link may be present to different degrees, the analysis will show that it is always there.

Moreover, the mood alternation in conditional clauses will be explained by the notion of dominion. The claim is that the indicative mood in si-clauses designates an event that is located within the conceptualizer’s epistemic dominion, whereas the subjunctive mood in other conditional clauses designates an event that is located outside the dominion of effective control. More exactly, the analysis will show that the meaning of the particle si – designating inferences and reasoning about causal relations – is related to the conceptualizer’s epistemic dominion, whereas other conditional conjunctions, overtly expressing a condition for the occurrence of the main clause event, evoke the dominion of effective control. Subsequently, mood choice designates whether the described event is located within or outside the conceptualizer’s dominion: (i) the indicative mood designates an event that is located within the conceptualizer’s dominion, and (ii) the subjunctive mood designates that the event is located outside the conceptualizer’s dominion.

It goes without saying that the Cognitive Grammar analysis pursued in the present paper conflicts with previous approaches to the analysis of mood choice, conditional clauses and the concept of reality. For example, Giannakidou (Reference Giannakidou1998, Reference Giannakidou, Musan and Rathert2011) claims that a fundamental feature of if-clauses is that they are nonveridical in character. The reason for this assumption seems to be that that they designate a hypothetical relation and imply a projection into the future (see also Section 2.1 above). From the theoretical perspective of Cognitive Grammar, however, the concept of reality is conceived of in more dynamic terms. As Langacker (Reference Langacker1991) points out: ‘[the] evolutionary momentum is conceived of as being strong enough that the future course of reality can be projected with considerable confidence’ (Langacker Reference Langacker1991: 228). That is, given our experience of how the world has evolved until the present, we have the capacity to also integrate future events into our conception of reality – and into our epistemic dominion.Footnote ^[8] The concept of dominion will be discussed in the following section.

4.1 Content domain

This analysis will focus on the notion that si-clauses are less conditional in character and that they describe a plausible and/or possible cause for the occurrence of the main clause event. By contrast, prototypical conditionals, such as a condición de que ‘on the condition that’, con tal de que ‘provided that’, siempre y cuando ‘as long as’, etc., seem to describe a necessary condition for the fulfilment of the main clause event. This difference is illustrated in the following examples:

Example (17) is an accurate description of a context in which someone proposes physical exercise to a friend in order for him/her to improve the health, whereas (18)–(20) describe a situation in which physical exercise is a necessary condition for improved health. That is, the protasis in example (17) designates a causal force that may be used in order to achieve the event described in the main clause. By contrast, examples (18)–(20) seem to put emphasis on the necessary condition for the fulfilment of the main clause event.

The difference between examples (17) and (18)–(20) becomes more obvious if we try to add an alternative to the condition expressed in the protasis:

The reason why examples (22)–(24) sound unnatural is probably because they create a semantic conflict between a necessary condition and an alternative to this condition. This is not the case in example (21). The relation between the two events is in accordance with the speaker’s conception of causes and effects in the world and is not due to necessary conditions. Indeed, we know that physical exercise improves our health, but we also know that health can be improved by other means under normal circumstances. Therefore, the si-clause in (21) describes a causal force that leads to the effect of the main clause event. This causal force is located within the speaker’s epistemic dominion and that is why the indicative mood occurs.

Another difference is related to the character of the effect. On a purely intuitive basis, causal relations can have either a positive or a negative outcome. Prototypical conditions, by contrast, tend to focus on a desired outcome. A condition is set up for the occurrence of an event with positive features. This is why the causal relations described in (25) and (26) are perfectly natural, whereas examples (27) and (28) seem to be less appropriate:

The examples displayed above refer to the conception that some events lead to the occurrence of other events. If you eat too much, it may have a negative consequence for your physical condition, and if you lean out of an open window, you may fall out. These relations are causal in character. Thus, examples (25)–(26) describe the speaker’s conception of causal relations in the world and not a condition for the outcome of the event described in the main clause. By contrast, examples (27)–(28) illustrate a semantic conflict in which a condition is expressed in a context of causal relations.

Accordingly, the semantic difference between si-clauses, on the one hand, and other conditional clauses, on the other hand, refers to how overtly the condition for the occurrence of the main clause event is expressed. This difference can also explain why conditional conjunctions typically trigger the subjunctive mood. To put a condition on someone or something evokes the dominion of effective control. In prototypical cases, the fulfilment of the condition is located outside the conceptualizer’s dominion of effective control. Consider the following examples:

The common denominator of examples (29)–(31) is that the conceptualizer (equated with the speaker) expresses a condition for the fulfilment of the main clause event. However, the occurrence of the event described in the conditional clause is located outside the conceptualizer’s dominion of effective control. In fact, placing a condition on someone to do something does not entail that the event will be realized. It is the referent of the subject of the conditional clause who is the active participant in the described event, and as such, s/he may decide whether to execute the action or not. The conceptualizer may wish that the action will occur but has no effective control over whether it in fact does.

The relevant conceptualizer can also be equated with the main clause subject:

In examples (32)–(33), the participant referred to by the subject of the main clause puts a condition on the subject of the conditional clause. The conditional conjunctions a condición de que ‘on the condition that’ and con tal de que ‘provided that’ evoke the dominion of effective control. The completion of this event, however, depends on a willingness to cooperate with the main clause subject. This being so, the subjunctive mood expresses that the event is located outside the main clause subject’s dominion. Another participant, i.e. the subject of the conditional clause, represents the active participant of the described event. Figure 3 illustrates the difference between si-clauses and clauses introduced by other conditional conjunctions.

Figure 3 Si-clauses and other conditional conjunctions: (a) si-clauses and the epistemic dominion, (b) conditional clauses and the dominion of effective control.

As illustrated in Figure 3a, the causal relation between the si-clause and the main clause focuses on the conceptualizer’s epistemic dominion. In Figure 3a, the higher degree of salience of the epistemic dominion is illustrated with the bold line surrounding the conceptualizer’s epistemic dominion. This causal relation corresponds to the conceptualizer’s view of causal relations in the world, which is why the si-clause is located within the epistemic dominion. By contrast, prototypical conditional clauses designate the conceptualizer’s dominion of effective control. This is illustrated in Figure 3b. The fact that the event described in the conditional clause depends on the action of a participant other than the conceptualizer entails that is located outside the conceptualizer’s dominion of effective control.

To conclude, the mood alternation analyzed above can be explained by the concept of dominion. Si-clauses do not primarily focus on the conditional relation between two events, but rather they express the speaker’s conception of causal relations. Accordingly, the indicative mood designates an event that is located within the speaker’s epistemic dominion. By contrast, prototypical conditional conjunctions evoke the dominion of effective control. The subjunctive mood adds the information that the described condition is located outside the conceptualizer’s dominion.

4.2 Epistemic domain

The epistemic domain differs from the content domain in that in the former the causal relation is more indirect than in the latter. Instead of designating a causal relation between two events, the si-clause designates the basis for the conclusion expressed in the main clause. This phenomenon was illustrated in example (15), e.g. If she is divorced, (then) she’s been married, in which the knowledge of the protasis leads to the conclusion expressed in the apodosis. Knowledge of the proposition expressed by the si-clause causes the conceptualizer to conclude that the main clause event is true. Obviously, this feature rules out the use of prototypical conditional conjunctions:

The fact that the particle si, rather than a prototypical conditional conjunctions, is used in order to describe the speaker’s inferences provides further evidence for the claim that this particle introduces a causal force that motivates the conceptual content of the main clause. Thus, the occurrence of the particle si in the epistemic domain can be conceived of as a functional extension from the content domain (see Sweetser Reference Sweetser1990: 141). The following examples illustrate the inferential causal relation between the si-clause and the main clause:

In examples such as (35)–(37), there is no direct causal relation between the two clauses: the light does not cause the arrival of ‘the Pérez’; the message does not cause Pedro’s departure; and the buying of the car does not cause financial wealth. Nonetheless, there is a causal relation if we consider the basis on which the speaker reaches a conclusion. These examples present inferential chains of the type: lights on${>}$they are at home; left message${>}$gone; and afford to buy a car${>}$rich. The conceptual content of the si-clause causes the conceptualizer to draw the conclusion expressed in the main clause.

In this sense, the function of the particle si within the epistemic domain resembles the function of the causal conjunction porque ‘because’. It is a well-known fact that this conjunction not only designates ‘purely’ causal relations, but also the conceptualizer’s inferences. It is thus perfectly natural to use porque ‘because’ in the epistemic domain:

A common explanation for the meaning of porque in this context is that it has an explicative function (see Oversteegen Reference Oversteegen1997, Sanders Reference Sanders1997, Pander Maat & Sanders Reference Pander Maat, Sanders, Kortmann and Couper-Kuhlen2000). The porque-clause provides an explanation for the reason why the speaker reaches a conclusion. In other words, the causal porque is analogous to the particle si. Both can introduce a clause that explains the underlying cause for the conceptualizer’s inference. Accordingly, the indicative mood reflects the fact that it designates the basis on which the conceptualizer draws a conclusion. Needless to say, the described event is located within the conceptualizer’s epistemic dominion.

4.3 Speech-act domain

The occurrence of si-clauses in the speech-act domain represents a further extension of the content domain. If the epistemic domain mirrors a cause–effect relation in which the si-clause is the basis for a conclusion, si-clauses in the speech-act domain tend to correlate with the speaker’s reasoning – or inferences – regarding the situation described in the si-clause. Alternatively, the si-clause may represent known information (see Haiman Reference Haiman1978). These features cause the conceptualizer to express the conceptual content of the main clause. Examples (41)–(42) illustrate this phenomenon.

Although examples (41)–(42) differ in some aspects, they do evoke a causal relation between the si-clause and the main clause.Footnote ^[11] This causal relation is perhaps more overtly expressed in example (41). The act of informing the hearer that there is food in the refrigerator is only relevant if the speaker has some reason to believe that the hearer is hungry. On the one hand, the si-clause can represent known information; that is, the speech act has been preceded by the information that the hearer is hungry, e.g. ‘I am really hungry’. On the other hand, some contextual evidence may cause the conceptualizer to infer that the hearer is hungry. In both cases, the conceptualizer has a reason to believe that the conceptual content of the if-clause is true. Thus, the indicative mood describes an event that is located within the conceptualizer’s epistemic dominion. The si-clause in example (42) is more properly regarded as a discourse marker expressing politeness and, as such, it imposes a permission assumed by the speaker to make an assertion. The if-clause is in this sense the cause for expressing the main clause: if I am allowed to say it${>}$I say it. In sum, the indicative mood in examples (41)–(42) describes events that are located within the speaker’s epistemic dominion.

Finally, Mejías-Bikandi (Reference Mejías-Bikandi2009: 168) comments on what he refers to as incredulity conditionals, e.g. Si tú eres Bill Gates, yo soy un marciano ‘If you are Bill Gates, I am a Martian’. From the perspective of the present analysis, the most plausible interpretation of this example is that the conceptualizer (equated with the speaker) does not believe the protasis to be true, which causes him/her to expresses the conceptual content of the apodosis. This causal relation can be formulated as If you say X (which I doubt to be true), I say Y (which is not true). The causal relation per se is located within the conceptualizer’s epistemic dominion.