14.1 Introduction
Dingemanse (Reference Dingemanse and Enfield2017: 195) defines ‘marginalia’ as “typologically unexceptional phenomena that many linguists think can be ignored without harm to linguistic inquiry.” He applies this concept to the analysis of ideophones (like kibikibi ‘energetic’ in Japanese) and interjections (like Ouch! in English), showing that both share a certain degree of syntactic independence (hence their common portrayal as marginal for grammatical theory) but arguing that these categories still shed light on central aspects of the design of human language:
Ideophones challenge us to take a fresh look at language and consider how it is that our communication system combines multiple modes of representation. Interjections challenge us to extend linguistic inquiry beyond sentence level and remind us that language is social-interactive at core.
The concept of marginalia can be easily related to one of the central tenets of Construction Grammar: to provide a full account of the grammar-lexicon of languages, thus challenging presumed distinctions between core and periphery in grammar. In this sense, construction grammarians have often given special attention to irregular and idiomatic constructions:
Our reasons for concerning ourselves with otherwise neglected domains of grammar are not so that we can be left alone, by claiming territory that nobody else wants, but specifically because we believe that insights into the mechanics of the grammar as a whole can be brought out most clearly by the work of factoring out the constituent elements of the most complex constructions.
One phenomenon that has been neglected until recently in descriptive and theoretical grammatical analyses is insubordination, the main clause use of subordination markers (Evans Reference Evans and Nikolaeva2007), such as an if-clause used as a polite request in English (If you could open the door, please). Insubordination can by no means be considered an exceptional phenomenon. Since the seminal work by Evans (Reference Evans and Nikolaeva2007), which includes thirty-seven languages representing twelve linguistic families, a large amount of literature has been devoted to describing and analyzing independent constructions with subordination markers in a wide array of languages belonging to unrelated linguistic families. Scholars have been very active in Romance (Debaisieux Reference Debaisieux2006; Lombardi Gras Reference Gras2011, Reference Gras, Jacob and Ploog2013, Reference Lombardi Vallauri, Evans and Watanabe2016; Patard Reference Patard2014; Sansiñena Reference Sansiñena2015; Sansiñena et al. Reference Sansiñena, De Smet and Cornillie2015a, Reference Sansiñena, De Smet and Cornillie2015b; Gras & Sansiñena Reference Sansiñena2015, Reference Gras and Sansiñena2017, Reference Gras, Sansiñena, Bouzouita, Enghels and Vanderscheuren2021; Vallauri Reference Vallauri, Evans and Watanabe2016; Hirata-Vale et al. Reference Hirata-Vale, Oliveira and Silva2017; Alves & Hirata-Vale Reference Alves and Hirata-Vale2021) and Germanic languages (Verstraete et al. Reference Verstraete, D’Hertefelt and Van Linden2012; Brinton Reference Brinton2014; Wide Reference Wide, Boogaart, Colleman and Rutten2014; D’Hertefelt Reference D’Hertefelt2018), but there are studies on insubordinate constructions in many non-Indo-European languages as well: Athabaskan and Eskimoan (Mithun Reference Mithun2008; Cable Reference Cable2011) and Altaic (Robbeets Reference Robbeets2009), among others.
Generally, studies tend to concentrate on specific constructions – complement or conditional clauses, non-finite verb forms or particles – whether in a single language or in a sample of related languages. Since independent insubordinate constructions tend to be highly polyfunctional, many of these studies try to disentangle whether the different meanings can be better represented as separate constructions or as instances of a single schematic construction (e.g., Verstraete et al. Reference Verstraete, D’Hertefelt and Van Linden2012; D’Hertefelt & Verstraete Reference D’Hertefelt and Verstraete2014; D’Hertefelt Reference D’Hertefelt2018 on independent complement and conditional constructions in Germanic languages; Gras Reference Gras, Jacob and Ploog2013, Reference Gras, Evans and Watanabe2016; Gras & Sansiñena Reference Sansiñena2015, Reference Gras and Sansiñena2017, Reference Gras, Sansiñena, Bouzouita, Enghels and Vanderscheuren2021; Sansiñena Reference Sansiñena2015 on independent complement constructions in Spanish).
This chapter tackles a different angle on insubordination. Since most insubordinate constructions tend to occur in informal conversations, they serve as an interesting case for analyzing the discourse–grammar–prosody interconnections, given that grammar in interaction has inherently a phonetic shape and needs to be related also to the turn structure of the conversational flow (see also Chapter 12). A relevant case study in this respect is the contrastive insubordinate conditional construction (CICC) (Montolío Reference Montolío1999; Schwenter Reference Schwenter2016). This construction, exemplified in (1), occurs typically as the dispreferred second part of an adjacency pair stating a reason why the speaker considers the previous turn to be inappropriate. In this case, by stating that the addressee is wearing their glasses, the previous question by the addressee is regarded as pragmatically inappropriate.Footnote 1
(1)
A: ¿Has visto mis gafas? ‘Have you seen my glasses?’ B: ¡Si las llevas puestas! if them wear.2sg.prs.ind put ‘But you are wearing them!’
The CICC clearly qualifies as a construction in that it pairs an idiosyncratic form (a self-standing conditional protasis) with a non-compositional meaning (contrast with previous context). This is an interesting case study for two main reasons. First, as already pointed out, the construction typically instantiates the dispreferred second part of an adjacency pair (Montolío Reference Montolío1999), but it remains unclear whether this is a formal restriction, equivalent to morphosyntactic constraints. And second, previous studies have pointed to the existence of prosodic restrictions (Montolío Reference Montolío1999; Schwenter Reference Schwenter2016) though no detailed prosodic research was conducted.
The integration of prosodic and discursive information in the description of grammatical constructions does not pose a theoretical challenge in Construction Grammar since most constructional approaches define grammatical constructions as pairings of any aspect of phonological, morphological, and syntactic form and any of semantic, pragmatic, and discursive meaning. However, the fact that a specific construct licensed by a construction must have all its features of form and meaning specified does not mean that these must be stated at the level of the construction. As a taxonomic theory, Construction Grammar allows information to be stored at different nodes of the constructional network, enabling some features to be inherited from more abstract constructions, on the one hand, and some features to be specified only in more concrete constructions, on the other. The goal of this chapter is to reflect on the integration of prosody and discourse in a constructional model that uses constructions and networks through the analysis of contrastive insubordinate conditional constructions in Spanish.
The chapter is organized as follows. Section 14.2 summarizes the main features of insubordination as a phenomenon, including its main formal and functional properties in Spanish, and offers a constructional analysis of the CICC in Spanish, providing evidence in favor of its constructional status. Section 14.3 describes the intonational analysis of the construction, while Section 14.4 presents the discursive analysis. Section 14.5 presents the conclusions.
14.2 Insubordination in Spanish
14.2.1 Insubordination
Insubordination “can be defined diachronically as the recruitment of main clause structures from subordinate structures, or synchronically as the independent use of constructions exhibiting prima facie characteristics of subordinate clauses” (Evans & Watanabe Reference Evans, Watanabe, Evans and Watanabe2016: 2). From a synchronic point of view, a construction is an instance of insubordination if it fulfills these two criteria: (i) that the construction shows features that are typical of subordinate clauses in the language at hand (non-finite verb forms, complementizers, subordinating particles, etc.), and (ii) that it is not a syntactic constituent of another syntactic structure uttered by the same speaker or another speaker, in the same or previous turns; that is, it is not a case or regular ellipsis. In this sense the CICC presented in example (1) qualifies as an instance of an insubordinate construction since (i) it bears subordinating marking (the conditional conjunction si ‘if’) and (ii) it cannot be parsed as a conditional protasis of the interrogative sentence uttered by the previous speaker (*¿Has visto mis gafas si las llevas puestas? lit. ‘Have you seen my glasses if you’re wearing them?’).
In terms of form, Spanish insubordinate constructions fall into two patterns: non-finite verb forms, either infinitives (2) or gerunds (3), and subordinating conjunctions followed by a verb in the indicative (4) or subjunctive mood (5).
(2)
Infinitive with retrospective imperative interpretation A: Tengo hambre. ‘I’m hungry.’ B: Pues haber comido antes. dp have eaten before ‘You should have eaten first.’
(3)
Gerund with imperative interpretation Andando, que es tarde walk.grd comp be.3sg.prs.ind late ‘Let’s walk, it’s late.’
(5)
<que ‘that’ + subjunctive> with optative interpretation ¡Buenas noches, que duerm-as bien! good night comp sleep-2sg.prs.sbjv well ‘Good night, sleep tight!’
Both formal patterns differ in terms of frequency. A study of insubordinate constructions in a corpus of informal conversations among adult speakers of Peninsular Spanish (Briz & Val.Es.Co. Reference Briz2002) showed that subordinating conjunctions are much more frequent than non-finite verb forms (Table 14.1): Non-finite verb forms constitute only 4.61 percent of the total amount of insubordinate constructions in the corpus, and no gerunds were found. As for the subordinating conjunctions, even though the literature identifies independent uses of several subordinating conjunctions, que ‘that’ and si ‘if’ represent most of the cases found in the corpus (83.52 percent).
Table 14.1 Distribution of insubordinate constructions in the Val.Es.Co. corpus (Reference Briz2002)
From a semantic-pragmatic perspective, Evans (Reference Evans and Nikolaeva2007) points out that insubordinate constructions tend to express the same meanings across languages. Considering data coming from twelve language families, Evans (Reference Evans and Nikolaeva2007) builds a functional cross-linguistic typology of insubordinate constructions, which consists of three macro functions: (i) They express indirectness and interpersonal control (especially directives, but also permissives, warnings, and threats), like the imperative infinitives (2) and gerunds (3); (ii) They express various kinds of modal framing (especially epistemic and deontic, but also evaluative), like the complement clauses with optative interpretation (5); (iii) They signal a high degree of presupposed material (focus, contrast, reiteration, among others), such as the complement clauses with quotative interpretation (4).
This typology has received different elaborations in further research in Romance and Germanic languages. On the one hand, it has been noted that there exists substantial overlap between functions (i) (indirectness) and (ii) (modal framing), since insubordinate directives could fall in either category (Gras Reference Gras2011, Reference Gras, Evans and Watanabe2016; D’Hertefelt Reference D’Hertefelt2018). Recent work has proposed that these uses of insubordination might be better explained as minor sentence type constructions (Siemund Reference Siemund2018; Gras Reference Gras2020; Pérez et al. Reference Pérez, Gras and Brisard2021): formally marked alternatives of major sentence types, which normally add an expressive value to their major counterparts.Footnote 2 On the other hand, function (iii) has been rephrased as ‘discourse insubordination’ (Gras Reference Gras2011, Reference Gras, Evans and Watanabe2016; Verstraete et al. Reference Verstraete, D’Hertefelt and Van Linden2012) because patterns marking this function tend to express relations between the insubordinate clause and some other parts of discourse. The CICC would be an instance of discourse insubordination, given that it expresses some sort of contrast with previous discourse.
The diversity of form, meaning, and function these constructions show leads to the question of whether we are dealing with a single unified phenomenon or with distinct phenomena that share a formal feature. Parts of the literature argue in favor of setting insubordination apart from the extension of dependency for discourse uses, being referred to as ‘extension of dependency’ (Mithun Reference Mithun2008, Reference Mithun, Beijering, Kaltenböck and Sansiñena2019) or ‘dependency shift’ (D’Hertefelt & Verstraete Reference D’Hertefelt and Verstraete2014; D’Hertefelt Reference D’Hertefelt2018).Footnote 3 An example of dependency shift is the use of complement clauses to elaborate on a previous turn, as in the following example:
(6)
Elaborative complement structure in Danish A: om vi skulle fråga våra eh förstaklassare här om dom vill ha betyg eller inte skulle dom inte fatta vad det handlade om vet inte hur vad betyg eller vad det e (…) så det ju nånting som / andra lägger på B: ja A: att det det kommer ju sen atomatist i comp it it come.prs part afterwards automatically in skolan att man får betyg school.def comp one get.prs grades å då kommer den här konkurrensen ännu mera in tror jag va A: ‘if we were to ask our first-graders here if they want to have a diploma or not they wouldn’t understand what it was about don’t know how what grades or what it is (…) so it’s something that / others impose’ B: ‘yes’ A: ‘that it it then comes automatically in school that one gets grades and then this competition starts even more I think right’ (D’Hertefelt & Verstraete Reference D’Hertefelt and Verstraete2014: 92)
These uses express a discursive meaning, are discursively dependent on a previous utterance or turn, and do not allow for the reconstruction of a main clause. According to the distinction between insubordination and dependency shift, both phenomena differ in terms of their meaning/function, their degree of (in)dependence, and their plausible diachronic development, as summarized in Table 14.2.
Table 14.2 Differences between insubordination and extension of dependency
| Insubordination | Extension of dependency | |
|---|---|---|
| Meaning/function | Modal, illocutionary | Discursive |
| Dependency | Syntactic and discursive independence | Syntactic independency, but discursive dependency |
| Plausible diachronic development | Ellipsis of a main clause that gives rise to the modal/interactional meaning of the insubordinate construction | Extension of dependency from a clausal to a discursive domain (increase of scope) |
Even though this distinction has been useful for explaining the differences between superficially similar constructions in several languages (see references above), the CICC challenges it in several ways. Regarding its meaning/function, CICC has a discursive meaning, which consists in signaling a contrast between the proposition and some background content. In example (1), asserting that the addressee is wearing their glasses questions the appropriateness of the question (¿Has visto mis gafas? ‘Have you seen my glasses?’). However, there is also a modal-illocutionary dimension of the construction since it carries a strong assertion (see Section 14.2.2).
As for its degree of independence, CICC in its prototypical use as a dispreferred response is syntactically independent but discursively bound to a previous turn by the addressee. However, in some cases, its instances can be uttered to express the speaker’s surprise at a situation that challenges their previous assumptions. Example (7) could be said by a speaker who, after looking for their glasses, realizes that they are wearing them. In this context, the construction expresses a mirative meaning and is not bound to previous discourse material but to the situational context. A partially equivalent pattern in English is the mostly substantive construction “If it isn’t X” (e.g., Well, if it isn’t my old friend Tom!), which is “used to express surprise about meeting someone when it is not expected” (Merriam-Webster def. ‘if it isn’t’).
(7)
¡Si las llev-o puestas! if them wear-1sg.prs.ind put ‘But I’m wearing them!’
As for its plausible diachronic development, it has been suggested that CICC derives from epistemic conditionals (‘If you are wearing your glasses, why are you asking for them?’). As Montolío (Reference Montolío, Bok-Bennema, de Jonge, Kampers-Manhe and Molendijk2001: 201) argues, “only the protasis appears in this construction, the apodosis having been systematically omitted or ‘silenced’, as it were, given its invariant nature: why have you said what you have just said?”. This explanation is compatible with Evans’ (Reference Evans and Nikolaeva2007) hypothesis about insubordination as a result of the conventionalization of a main clause ellipsis.
In sum, CICC challenges the distinction between insubordination and extension of dependency in several respects: (i) It combines modal-illocutionary and discursive meanings; (ii) It shows discourse-boundedness, while still able to be used discursively independently; (iii) It can be accounted for as the result of main clause ellipsis. This situation is not exclusive to this construction. In fact, many Spanish insubordinate patterns combine a modal-illocutionary component with discursive functions. On the one hand, several minor imperative sentence type constructions show discourse restrictions in addition to their illocutionary value. Consider, for instance, infinitives with retrospective imperative interpretation: They have the illocutionary force of a reproach (‘You should have done it’), while they serve as dispreferred responses in interaction, therefore fulfilling a discourse-structuring function (reaction). On the other hand, some discourse insubordinate patterns allow the reconstruction of a main clause. This is the case of complement clauses with quotative interpretation, which always allow the reconstruction of a verb of saying, as the modified version of example (4) shows in (4′). Therefore, the CICC will be treated in this chapter as an instance of discourse insubordination and not as a case of extension of dependency.
(4′)
<que ‘that’ + indicative> with quotative interpretation A: Voy a cenar mañana ‘I’m coming for dinner tomorrow.’ B: ¿Qué? ‘what?’ A: Te digo que voy a cenar mañana you say.1sg.prs.ind comp come.1sg.prs.ind to dinner tomorrow ‘I’m telling you that I’m coming for dinner tomorrow.’
14.2.2 Contrastive Insubordinate Conditionals in Spanish
Contrastive insubordinate conditionals differ from the most frequent use of insubordinate conditionals discussed in the literature (Kaltenböck Reference Kaltenböck, Kaltenböck, Keizer and Lohmann2016; D’Hertefelt Reference D’Hertefelt2018; Lastres-López Reference Lastres-López2020), which are polite requests (8) and wishes (9).Footnote 4 Even though insubordinate conditional structures can express polite requests and wishes in Spanish, they are much less frequent than CICC, at least in informal conversation. Indeed, all instances of independent insubordinate conditional in the Val.Es.Co. corpus have a contrastive interpretation (Gras Reference Gras2011). It should also be noted that CICC is not exclusive for Spanish. Equivalent constructions can be found in several Romance languages: Catalan (Salvador Reference Salvador and Solà2002), Portuguese (Alves & Hirata-Vale Reference Alves and Hirata-Vale2021), and Italian (Lombardi Vallauri Reference Vallauri, Evans and Watanabe2016).
(8)
Conditional with polite request interpretation in Spanish Si pudieras abrir la puerta, por favor. if can.2sg.pst.sbj open the door please ‘If you could open the door, please.’
(9)
Conditional with wish interpretation in Spanish ¡Si tuviera 10 años menos! if have.1sg.pst.sbj 10 years less ‘If I were ten years younger!’
CICC should not be confused with suspended conditional constructions (Schwenter Reference Schwenter2016), that is, conditional protases whose main clause must be inferentially reconstructed by the addressee, as in (10):
(10)
A: ¿Vamos a la playa? ‘Let’s go to the beach!’ B2: Si quieres … if want.2sg.pres.ind ‘If you want to …’
In example (10), si functions as a conditional marker that introduces a protasis, whose apodosis has been left open for reconstruction: si quieres, vamos a la playa ‘If you want, we go to the beach’. The difference between regular bi-clausal patterns and suspended patterns, thus, has to do with the explicit or implicit character of the apodosis: explicit in the former, implicit in the latter.
CICC has idiosyncratic formal features that set it apart from conditional constructions, whether regular bi-clausal or suspended ones (Montolío Reference Montolío1999; Schwenter Reference Schwenter1998). First, it rejects subjunctive verb forms, as version B1 from example (10′) shows. Second, it rejects coordination with other clauses, as the ungrammaticality of B2 shows. And finally, it does not combine with ‘continuation rise’ intonation; on the contrary, previous studies have suggested it has exclamative intonation (Montolío Reference Montolío1999; Schwenter Reference Schwenter2016).
(10′)
A: ¿Vamos a la playa? ‘Let’s go to the beach!’ B1: *¡Si esté lloviendo! if be.3 sg.prs.sbjv raining B2: *¡Si está lloviendo y si hace frío! if be.3sg.prs.ind raining and if make.3sg.prs.ind cold *‘It’s raining and it’s cold!’
As for the meaning, most authors tend to identify a contrastive meaning (Contreras Reference Contreras1960; Almela Reference Almela1985; Porroche Reference Porroche, Martín Zorraquino and Montolío1998; Schwenter Reference Schwenter1998; Montolío Reference Montolío, Bok-Bennema, de Jonge, Kampers-Manhe and Molendijk2001; Schwenter Reference Schwenter2016). In particular, Montolío (Reference Montolío1999, Reference Montolío, Bok-Bennema, de Jonge, Kampers-Manhe and Molendijk2001) identifies two features that capture the meaning side of the construction: (i) It expresses a basic meaning of contrast, which can affect various aspects of a previous turn (its propositional content or illocutionary force); (ii) It indicates the inappropriateness of some aspect of the addressee’s previous turn or their non-linguistic behavior, and leads to an implicit conclusion: p / if q → p is inappropriate (e.g., ‘If it’s raining and cold, then it is inappropriate to suggest going to the beach’). In addition, Montolío suggests that CICC is discourse-placed in the sense that their instances “encode information about the context in which they appropriately occur” (Evans Reference Evans and Foley1993: 325). In particular, they normally occur as the dispreferred second part of an adjacency pair, often as a turn in itself.
However, as Montolío herself acknowledges, the construction also occurs in discursive contexts other than dispreferred responses. In order to address this limitation, Schwenter (Reference Schwenter2016) has developed an epistemic analysis of the construction. According to him, it is convenient to distinguish between the ‘codified’ meaning of the construction and a series of context-dependent (pragmatic) interpretations: “The coded, non-truth-conditional meaning of si is argued to be epistemic in nature, marking the proposition that it accompanies as one which is obviously true to the speaker” (Schwenter Reference Schwenter2016: 22). According to Schwenter, this epistemic analysis makes it possible to explain apparently non-contrastive contexts of use, such as the introduction of a premise that justifies a previous statement, as exemplified by the rhetorical question in (11) or the agreement with a preceding turn, as occurs in (12).
(11)
Tweet about Carmen Electra, an American actress and model ¿Quién quiere conocer a @CarmenElectra? Si está más buena que unos taquitos if be.3sg.prs.ind more good than some taquitos al pastor a las 2am al pastor at the 2am ‘Who wants to meet @CarmenElectra? She’s yummier than some taquitos al pastor at 2am.’
(12)
Conversation between faculty members A: Juana la han aceptado en Stanford. ‘Juana has been accepted at Stanford.’ B: Claro, si es muy inteligente. sure if be.3sg.prs.ind very intelligent ‘Sure, she’s very intelligent.’
However, this analysis also has two interrelated limitations. Although the meaning of epistemic certainty is common to both the contrastive uses and the non-contrastive uses that have just been presented in (11)–(12), Schwenter’s analysis does not indicate which discursive contexts favor the presence of the insubordinate conditionals. Normally, presenting a proposition as obviously true for the speaker is compatible with any discursive context, and yet the construction tends to be used in a limited set of contexts. Also, in relation to this aspect, Schwenter’s analysis is based on the qualitative analysis of a set of examples and does not indicate how often the various interpretations/contexts occur.
To summarize, the constructional status of this pattern is clear. On the formal side, the construction exhibits properties that cannot be predicted from the knowledge of conditional constructions, especially mood selection and clause combining. On the interpretive side, the construction marks an interactional contrastive and assertive meaning that cannot be attributed to any lexical item. However, the status of the prosodic and discursive information of the construction remains unclear. From a discourse-structural perspective, a corpus-based analysis is needed to analyze the discourse constraints of the constructions in order to provide an empirically grounded constructional representation. From a prosodic perspective, it has been suggested that CICC has exclamative prosody. Nevertheless, more prosodic research is needed in order to decide whether the construction accepts other prosodic patterns and whether the attested prosody is idiosyncratic (i.e., does not occur outside the construction) or inherited (it exists outside the construction). To answer these questions, let us review two studies: an interactional corpus study (Section 14.3) and a prosodic study (Section 14.4).
14.3 Discourse Structure
The section presents a study whose goal is to analyze the discourse restrictions of the CICC in a corpus of informal conversations and, based on this analysis, to reflect on the most appropriate way to incorporate discursive information in a constructional approach.
14.3.1 Methodology
The interactional study is performed on data coming from the Val.Es.Co. corpus (Briz & Val.Es.Co. Reference Briz2002), which consist of informal conversations among adults from Valencia (Spain). To ensure accuracy, a total of seventy-five instances were carefully extracted using a semi-automated process, ensuring that only occurrences of the CICC were included. Each token has been analyzed according to two types of parameters: conversational and rhetorical. For the conversational analysis, three parameters have been considered:
(1) Type of unit: The construction constitutes (i) a turn in itself, (ii) a turn constructional unit (TCU) in a complex turn, or (iii) a turn extension.
(2) Type of intervention: The construction is used as (i) an initiation (e.g., question or command), (ii) a preferred response (e.g., affirmative reaction or agreement), or (iii) a dispreferred response (e.g., a negative reaction or disagreement).
(3) Position: The construction occurs in (i) an independent position (as a turn in itself), (ii) turn-initial, and (iii) turn non-initial (preceded by a TCU).
Regarding the rhetorical parameters, two have been considered:
(4) Target: The CICC refers to (i) a previous turn, (ii) a previous utterance of the speaker, or (iii) the extralinguistic situation.
(5) The argumentative orientation: (i) counter-orientation or (ii) co-orientation.
Taking into account the combination of the above parameters, five functions of the construction have been proposed. Four of them were already mentioned in the literature (rebuttal, controversial agreement, mirativity, and justification) and a fifth one has been identified (polyphonic rebuttal). They will be explained in detail in the next section.
14.3.2 Distribution of Functions in the Corpus
Rebuttal
Following Montolío (Reference Montolío1999), those occurrences in which the construction is used in a dispreferred response are called rebuttal. From a conversational point of view, the reply can be a turn or a TCU in a complex turn, which preferably occurs in turn-initial position, but also in non-initial position. In an initial position, it introduces a premise that leads to an implicit conclusion. As in (13), from a conversation between a boy (A) and his girlfriend (B), in which the speaker introduces a premise (Yo no te pido más tiempo ‘I don’t ask you for more time’) that leads to the implicit conclusion ‘You shouldn’t give me more time’, which contradicts what A said in his previous turn.Footnote 5
(13)
A is thinking about ending his relationship with B (ML.84.A.1.: 157–161) A: no yo SÉ que debería darte más tiempo↓ del que te doy B: pero si yo no te pido más tiempo ↓ but if I no you ask.1sg.prs.ind more time yo lo que te pido es que estés SEGURO A: ‘No I know I should give you more time than I give you’ B: ‘but I don’t ask you for more time, what I ask is that you be sure’
In a non-initial position, the use of CICC introduces a premise that justifies an explicit conclusion, as in (14), in which the propositional content (lo dijo por cachondeo ‘he was joking’) reinforces the disagreement expressed explicitly by ¡qué va! ‘no way’.
(14)
A found a watch in the street (RB.37.B.1: 48–68) C: ¿pero él– pero él entendía↑ dee– de reLOJES↑ oo? A: ¡QUÉ VA↓! si lo dijo por cachondeo no way if it say.1sg by fun C: ‘But did he know about watches or?’ A: ‘No way! He was joking’
To summarize, the construction can be used in contexts of explicit disagreement as reinforcement of a statement that codifies said disagreement (14), or else in contexts of implicit disagreement (13) in which the propositional content gives rise to a conventional implicature of a negative nature (‘something of what you have said or done is inappropriate’).
Polyphonic Rebuttal
The term polyphonic rebuttal is used to refer to utterances that contradict an assumption evoked in the turn, as in (15). Throughout her lengthy intervention, speaker A describes certain embarrassing situations she experiences when her roommates invite their boyfriends or partners to the apartment they share. The propositional content of the utterance (si yo creo que acepto de puta madre ‘I think that I absolutely accept it’) contradicts an inference that can be derived from the previous statements uttered by the speaker herself: that she does not accept the behavior of her roommates.
(15)
E talks about her roommates’ behaviors (L.15.A.2.: 933–947) porque es quee a mí me parece muy bien↑ que venga el novio de Olga y que se acueste con ella/// pero lo comprendo perfectamente si se queda la noche a dormir/ no va a dormir con él ¿no?/ lo que pasa que tú– te armen UUN CACAO to(d)a la noche que (( )) entonces/ […] SI YO CREO QUE LO ACEPTO MÁS DE PUTA MADRE ‘because the thing is that I think it’s fine that Olga’s boyfriend comes and sleeps with her, but I understand it perfectly if he sleeps over, it’s not that she’s not gonna sleep with him, no? The thing is that they make a big fuss all night that, then/ […] I think that I absolutely accept it’
From a conversational point of view, it is a TCU in non-initial position. From the rhetorical-argumentative point of view, the utterance introduces a statement that opposes a possible inference derivable from the previous discourse. Unlike regular cases of rebuttal, in which the counter-argumentation is established between statements issued by different speakers, in polyphonic rebuttals the counter-argumentation relationship is established between what the speaker affirms and what could be inferred from their previous utterances. It can be considered as a monological extension of a dialogical resource.
Controversial Agreement
The term controversial agreement is used to describe cases in which the CICC is used as a preferred response to polemic issues. This is what happens in example (16), in which a young man (J) tells his aunt (P) and his mother (C) that he has finally got his driver’s license. The CICC occurs as a response to P’s previous turn: P comments that she had told him that driving was easy and J agrees by saying si es una tontería conducir ‘driving is easy’. J’s intervention expresses his agreement with P’s previous turn and simultaneously contradicts J’s own earlier belief about driving being difficult.
(16)
J had a hard time getting his driving license (G.68.B.1: 365–374) P: ¿qué? ¿cómo va el coche ya Juan? J: muy bien/ que lo diga la mamá→ C: ¡ay!/ está hecho un artista J: que fuimos a la boda dee–/ bueno/ al bautizo C: al bautizo P: ¿yo qué te dije? verás cómo eso te vas a ir tú mismo↑ soltando↑ J: si es una tontería conducir if be.3sg a silly.thing driving C: es una tontería Q: ‘how are you doing with the car, Juan?’ J: ‘very good. Ask mom!’ C: ‘oh! He’s such an artist!’ J: ‘we went to the wedding, well, the baptism’ C: ‘to the baptism’ P: ‘what did I tell you? You’ll see that you’re going to let yourself go’ J: ‘driving is easy’ C: ‘it’s easy’
From the conversational point of view, the construction can be used as a turn in itself or as a TCU in initial position. From the rhetorical-argumentative point of view, it introduces a statement that confirms the previous turn, while canceling an assumption present in the context. Although it would seem paradoxical that the same construction is used to express both disagreement and agreement, it is necessary to clarify that it is a particular type of agreement: The CICC is used in situations in which the asserted content was challenged by previous context.
Mirativity
Mirativity refers to the speaker’s surprised attitude towards a state of affairs perceived at the time of speaking (De Lancey Reference DeLancey2001). This denomination describes situations in which a speaker expresses a counter-expectation reaction to an extralinguistic stimulus, such as in (17). This excerpt comes from a conversation in which speaker A talks about a watch she found on the street. In her turn, A says that she took it to a watchmaker to find out how much it would cost to repair it. The CICC reproduces in direct speech the reaction of the watchmaker when he sees the watch.
(17)
A found a watch in the street and brought it to a watchmaker to have it assessed (RB.37.B.1, p. 227: 117–122) y en seguida quitó la caja↑ y dice ¡vaya reloj! y dice pues si este reloj es buenísimo and say.3sg.prs.ind dp if this watch be.3sg.prs.ind good.sup ‘And right away he took out the box and he said what a watch and he said this is a very good watch.’
Mirativity contexts are conversationally characterized as turns that are initiations – since they are not responses to previous turns – but at the same time they are reactions to extralinguistic stimuli. From the rhetorical-argumentative point of view, they express a contrast between the speaker’s expectations and their perception at the time of speaking.Footnote 6
Justification
The last function departs from the above cases in that the notion of contrast is not salient. In justification uses, the construction is used as a turn extension, as in (18), coming from a conversation between young male adult friends. Speaker A is talking about his experience as an Erasmus student at the University of Ghent.
(18)
A talks about food during his Erasmus stay in Ghent (H.38.A.1., lines 459–480) B: de tapas ni de coña ¿no? A: tenía una– tapas tampoco↓ ¡qué va!// plato combinao me lo hago yo/ si tenía allí yoo una cocina/// if have.pst.ind there I a kitchen mis huevos y mis cosas (RISAS) B: ‘Eating tapas, no way, right?’ A: ‘Tapas neither, no way! A combo-dish I do it myself/ I had a kitchen there, my eggs and my stuff’
In this example, the CICC is used in a turn extension since it relies on the illocutionary force of the previous TCU. Through the use of CICC, the speaker introduces an argument that reinforces the immediately preceding TCU (‘I could make a combined dish because I had a kitchen there’).
Overview
The results of the corpus study allow us to propose an empirically based analysis of the discursive properties of the construction. As Table 14.3 shows, the construction displays a clear preference for the expression of rebuttal (63 percent) and, therefore, for dispreferred response position, thus confirming the intuitions of grammarians who tend to represent CICC in this type of context. However, a still significant number of cases (37 percent) are not dispreferred responses. Nevertheless, these data can be interpreted in another way if one takes into account that three of the remaining functions – polyphonic rebuttal, controversial agreement, and mirativity – constitute reactive interventions that also express some kind of contrast, although not between the content of two turns but between utterances in one speaker’s turn (polyphonic rebuttal), between what the speaker asserts and what could be inferred from previous turns (controversial agreement), or between what the speaker just realized and their previous expectations (mirativity).
Table 14.3 Distribution of meanings/functions of CICC in the Val.Es.Co. corpus (Reference Briz2002)
| Meaning/function | Tokens | Percentage |
|---|---|---|
| Rebuttal | 48 | 63% |
| Justification | 10 | 13% |
| Mirativity | 7 | 9.2% |
| Polyphonic rebuttal | 6 | 7.8% |
| Controversial agreement | 5 | 6.5% |
| Total | 76 | 100% |
14.3.3 Constructional Representation
The constructionist approach allows for discourse information to be represented as features of a grammatical construction: “A construction is a set of formal conditions on morphosyntax, semantic interpretation, pragmatic function, and phonology, that jointly characterize or license certain classes of linguistic objects” (Fillmore Reference Fillmore, Gras, Östman and Verschueren1999: 113). However, as the corpus study shows, the discursive position of the construction cannot be considered a formal restriction comparable to the morphological and syntactic restrictions described in Section 14.2. These are formal conditions which license a well-formed expression. For instance, the selection of a subjunctive verb form (in bold) would lead to an unacceptable result (*!Si las lleves puestas! ‘If you are wearing them!’). By contrast, the fact that CICC may not occur as a dispreferred response does not lead to ungrammaticality but to a different pragmatic function. At the same time, the fact that discourse position cannot be considered a formal restriction does not imply that it cannot be part of the representation of the construction. On the one hand, it is convenient to point out that the construction in question cannot occur in just any context: The analysis presented identifies five discursive contexts. On the other, not all contexts have the same impact: Reactive interventions constitute 63 percent of the occurrences in the corpus, which increases to 86.5 percent if we include all reactive interventions.
Construction Grammar offers the tools to incorporate discursive information in the representation of a construction. On the one hand, discursive information can be represented as a series of pragmatic attributes and values (e.g., Fried & Östman Reference Fried and Östman2005; Linell Reference Linell, Berg and Diewald2009; Nikiforidou et al. Reference Nikiforidou, Marmaridou and Mikros2014), like the conversational and rhetorical features used in the study reported above. On the other hand, as already mentioned, Construction Grammar allows information to be stored at different nodes of the ‘constructicon’, enabling some features to be specified in more concrete constructions. In order to capture the uneven distribution of contexts of the CICC, with rebuttal as the most frequent function, Langacker’s (Reference Langacker1987) approach, which distinguishes a ‘schema’ and a series of ‘instances’ and ‘extensions’, is especially relevant. This conception, which has been applied in Cognitive Construction Grammar (Goldberg Reference Goldberg2006), not only allows information to be represented at various levels (like all constructionist approaches) but also explains prototypical effects. The schema allows us to represent the information shared by all members of the category. The CICC constructional schema should include the following features:
Form: < si + indicative>
Meaning: assertion and contrast.
At the same time, the distinction of instances and extensions makes it possible to account for the functional versatility of the construction as well as the relative correspondence between functions and discourse positions. Likewise, this distinction makes it possible to capture the existence of a prototypical instance – the turn that occurs as a dispreferred response – as well as a series of extensions, which move away, to a greater or lesser extent, from the prototype and maintain relations of family resemblance among them. Both the prototypical instance and the extensions have their own discursive and rhetorical features:
(1) Rebuttal:
(2) Polyphonic rebuttal:
Function: contradict an assumption derivable from the previous content of the turn.
Discourse position: non-initial TCU.
Target: a previous TCU by the speaker.
Rhetorical relationship: counter-orientation.
(3) Controversial agreement:
Function: agree with the interlocutor while canceling an assumption derivable from previous turns.
Discourse position: initial TCU or turn in itself, preferred response.
Target: a previous turn by an interlocutor.
Rhetorical relationship: co-orientation with the previous turn and counter-orientation with previous turns or shared knowledge.
(4) Mirativity:
Function: express surprise about an extralinguistic situation.
Discourse position: initial TCU or turn in itself.
Target: the extralinguistic situation.
Rhetorical relationship: counter-orientation.
(5) Justification:
Function: agree with the interlocutor while canceling an assumption derivable from previous turns.
Discourse position: non-initial, turn extension.
Target: a previous turn by an interlocutor.
Rhetorical relationship: co-orientation.
The distinction between schemas and instances also allows us to account for geographic variation as well as polyfunctionality. As Levshina (Reference Levshina2012) argues, regional varieties of the same language can differ in terms of the organization of the schemas and instances/extensions (subschemas in her terminology) of a construction. In the case of CICC, it has been noted that the mirative function is not available in Argentinian Spanish (Rodríguez Ramalle Reference Rodríguez Ramalle2011) even though the construction can be used in its prototypical function (rebuttal). As for polyfunctionality, other discourse-sensitive patterns, such as discourse markers, are well known for their ability to occur in different contexts, expressing diverse meanings. A model which distinguishes schemas and instances/extensions can account for these phenomena, modeling the different functions/contexts either as separate constructions or as instances of an abstract schema (Fried & Östman Reference Fried and Östman2005; Nikiforidou et al. Reference Nikiforidou, Marmaridou and Mikros2014).
14.4 Prosody
The goal of this section is to illustrate an empirical prosodic analysis of CICC and to reflect on the most appropriate way to incorporate prosodic information in constructional representations.Footnote 7
14.4.1 Methodology
The prosodic analysis is based on elicited data, using a Discourse Completion Task in which the participants were given discourse contexts resembling those discussed in the previous section, and the lexical material they should use (Vanrell et al. Reference Vanrell, Feldhausen, Astruc, Feldhausen, Fliessbach and Vanrell2018). In order to keep the experiment controlled while still accounting for discourse-structural variation, the two most extreme contexts were included in the study: the prototypical dispreferred response expressing rebuttal (19) and the initiation with a mirative value (20).Footnote 8
(19) You are with a close friend, and you are talking about a third person that has put on weight after a pregnancy. Your friend tells you that it is obvious that she has put on weight because she eats chocolate every day, but you both have seen her eating vegetables and you say to your friend: ¡(Pero) si merienda verdura! ‘But she eats vegetables!’
(20) You are with a close friend that has recently had a child. You think that the child only drinks formula but you see that his mother is preparing vegetables. When you realize the vegetables are for the baby, you say: ¡(Anda) si merienda verdura! ‘(Wow) She eats vegetables!’
For each context, two sentences were included depending on the lexical stress of the last word (paroxytone and proparoxytone words).Footnote 9 The task was performed three times by each speaker. As for the speakers, all the recordings were made by fourteen native speakers of Peninsular Spanish. The average age of the participants was 24.44 (σ=2.10) and they came from four provinces of Spain (Madrid, Barcelona, Cantabria, Seville), thus representing a comprehensive overview of Peninsular Spanish. The first situation (19) was recorded by each participant (eighty-four responses), while the second was recorded only by one participant (six responses) in order to check if it had a different prosodic realization.
The annotation was done in PRAAT and consists of the first tier, in which the utterance was transcribed orthographically by words, the second tier with a syllable segmentation and phonetic transcription, and the third tier with Break Indices using Sp_ToBi, the Spanish version of ToBI (Tones and breaks indices) to transcribe prosody (Beckman et al. Reference Beckman, Díaz-Campos, McGory and Morgan2002; Estebas‐Vilaplana & Prieto Reference Estebas-Vilaplana and Prieto2008). ToBI systems are based on the Autosegmental Metrical model (Pierrehumbert Reference Pierrehumbert1980). According to this model, pragmatic meaning is best encoded by the combination of pitch movements that occur between the last stressed syllable and the end of the intonational phrase, called nuclear configurations or prosodic contours.
ToBI systems propose the existence of Break Indices (from 0 to 4), which are used to mark prosodic separation. They use a tone level-based approach, where intonation is understood as a series of high (H) and low (L) tones. Finally, they mark stressed syllables with a star (*), intermediate boundaries with a dash (-), and final boundaries with a percent symbol (%). An inventory of fourteen different nuclear configurations for Castilian Spanish can be found in Estebas-Vilaplana and Prieto (Reference Estebas-Vilaplana, Prieto, Prieto and Roseano2010). Each nuclear configuration consists of a pitch accent (the tones borne by the last stressed syllable in the utterance) and a boundary tone and is paired with a pragmatic meaning: basic speech act distinctions (assertions, questions, commands) and several pragmatic distinctions that combine with illocutionary forces (e.g., reiterated commands, contrastive focus, or obvious assertions). The goal of the analysis was to test whether the construction at hand (i) has an idiosyncratic nuclear configuration (not included in the inventory), (ii) is compatible with any of the nuclear configurations already identified for Spanish, or (iii) has a preference for one or more of these configurations.
14.4.2 Nuclear Configurations
Three nuclear configurations or prosodic contours were found in this study: L* HL%, L+H* L%, and ¡H*L%. The first two appeared in all four varieties, whereas the third appeared only in Seville. As for the contexts, all the contours can appear in both context types (rebuttal and mirative). As shown in Figure 14.1, the majority pattern is L+H* L% (59 percent), followed by L*HL% (39 percent), and a token presence of ¡H*L% (2 percent).
Figure 14.1 Attested nuclear configurations cases and percentages
The pattern L+H*L%, which consists of a rising stressed syllable and a low boundary tone (Figure 14.2), has been attested in almost all of Spanish as the typical contour of contrastive focus (Hualde & Prieto Reference Hualde, Prieto, Frota and Prieto2015).
L*HL% consists of a low last stressed syllable and a rising-falling boundary tone (Figure 14.3) and it has been described as the intonational pattern of narrow focus and contradiction statements (Prieto & Roseano Reference Prieto and Roseano2010).
Figure 14.3 Waveform, spectrogram, and pitch contour of the utterance ¡Pero si merienda verdura! ‘But she eats vegetables!’ produced by a female speaker from Barcelona with the nuclear configuration L* HL%
Finally, the ¡H* L% contour, which consists of a rising stressed syllable from a high target to an extra high one (Figure 14.4), has been described as a dialectal pattern attested in Canarian yes/no questions. However, it has also been attested in Seville as a dialectal solution for refutational statements.
Figure 14.4 Waveform, spectrogram, and pitch contour of the utterance ¡Pero si merienda médula! ‘But she eats marrow!’ produced by a male speaker from Seville with the nuclear configuration ¡H*L%
The summary of the attested prosodic patterns, their schematic contour, and pragmatic meaning can be observed in Figure 14.5. The three patterns are very similar, both in form and meaning. Formally, they consist of a rising stressed syllable and a low boundary tone, and it is unclear whether speakers can perceive the difference between them. Moreover, their meanings are also very similar, related to the expression of focus and contrast. Therefore, they can be considered formal variants of the same pattern.
14.4.3 Constructional Representation
Building on the analysis just presented, this section discusses the status of prosodic contours in a constructional model of grammar. It is possible to identify three potential scenarios regarding the relationship between a grammatical construction and its prosody:
Scenario 1. The construction is prosodically neutral; it can combine with any prosodic contour in the language.
Scenario 2. The construction is prosodically idiosyncratic; it has its own prosodic contour, which does not occur outside the construction (Sadat-Tehrani Reference Sadat-Tehrani2008).
Scenario 3. The construction inherits its prosody from independently existing prosodic constructions, which pair a prosodic contour with a pragmatic meaning (Ogden Reference Ogden, Barth-Weingarten, Reber and Selting2010; Ward Reference Ward2019).
As shown in the previous section, CICC combines only with focus prosodic contours, with a special regional variant in Seville. Therefore, the prosodic behavior of this construction is consistent with the third scenario: The construction inherits an already existing prosodic contour. A theoretical explanation of the relationship between prosodic patterns and grammatical constructions in a constructional model might consist of two interrelated aspects: (i) to treat prosodic patterns as constructions (with regional variants when relevant) and (ii) to represent prosody as a feature of grammatical constructions (see Chapter 13 for this distinction).
The first aspect rests on modeling prosodic contours as constructions that pair a phonological form (a prosodic contour) and a pragmatic meaning (illocutionary force or information structure status), in line with current research on the prosodic analysis of conversational data (Ogden Reference Ogden, Barth-Weingarten, Reber and Selting2010; Ward Reference Ward2019). Evidence for this approach comes from the fact that the same prosodic contours occur in constructions with different formal marking but similar meaning-function. Take, for instance, other Spanish insubordinate constructions with a rebuttal function (dispreferred responses), like ni que ‘not even that’ or como si ‘as if’ followed by past subjunctive verb forms (21),Footnote 10 which also inherit focus prosodic contours, as represented in Figure 14.6 (left and right).
(21)
A: ¿Me preparas la cena? ‘Can you cook me dinner?’ B: ¡Ni que fuera tu madre! not.even that be.1sg.pst.sbjv your mother ‘As if I was your mother!’ B’: ¡Como si fuera tu madre! as if be.1sg.pst.sbjv your mother ‘As if I was your mother!’
Figure 14.6 Waveform, spectrogram, and pitch contour of the utterances ¡Ni que fuera tu madre! and ¡Como si fuera tu madre! ‘As if I were your mother!’ produced with a focus intonational pattern (L+H* L%)
These insubordinate constructions differ from CICC in terms of their form: They are introduced by different subordinating conjunctions (ni que ‘not even that’, como si ‘as if’) and they select subjunctive verb forms. However, they coincide in their pragmatic function: They tend to occur in dispreferred second parts of adjacency pairs to express some correction of a previous statement (usually uttered by the interlocutor) by describing an extreme or absurd situation in which the previous turn would be appropriate. Moreover, the same prosodic contour is also found in simple declarative clauses with no specific marking if they express the same pragmatic function, as in (22), prosodically analyzed in Figure 14.7. Therefore, prosody does not mainly depend on lexicogrammatical form but on meaning-function.
(22)
A: ¿Me preparas la cena? ‘Can you cook me dinner?’ B: ¡No soy tu madre! not be. 1sg.prs.ind your mother ‘¡I’m not your mother!’
Figure 14.7 Waveform, spectrogram, and pitch contour of the utterance No soy tu madre ‘I am not your mother’ produced with a focus intonational pattern (L+H* L%)
Prosodic constructions, that is, pairings of an intonational contour with a meaning-function, can also have ‘allostructions’: structural variants of constructions that do not encode different meanings (Capelle Reference Capelle2006). On the one hand, it is not clear whether the changes in peak alignment that the data display in the alternation of L+H* L% and L* HL% are phonological, since they do not convey a dramatic change in meaning. On the other, intonational constructions are subject to dialectal variation, as the specific pattern found in Seville suggests.
The second option is to treat prosody as a feature of grammatical constructions. Lexical and phrasal constructions that can license sentences must inherit a prosodic construction, depending on the pragmatic function of the construction and the prosodic constructions available in the language. The range of prosodic constructions compatible with a specific construction depends on the functional specialization of that construction. Some constructions, like CICC, are only compatible with a specific prosodic construction due to their highly specialized pragmatic function. In other cases, a sentence-level construction can inherit more than one prosodic pattern because they can yield different pragmatic interpretations. For instance, insubordinate complement quotative constructions allow for the repetition of an utterance by the same speaker (Gras et al. Reference Gras, Pérez, Brisard, Hennecke and Wiesinger2023), as in (23). This construction can inherit two prosodic constructions (Figure 14.8): (i) declarative prosodic construction if the speaker simply repeats their previous statement and (ii) focus prosodic construction if the speaker tries to convey that their interlocutor did not react in accordance with their previous statement.
(23)
A: Son las nueve. ‘It’s 9.’ B: ¿Qué? ‘What?’ A: Que son las nueve. that be. pl.prs.ind the nine ‘I said it’s 9.’
Finally, these findings are compatible with construction-based prosodic research on the polyfunctionality of insubordinate constructions. Gras and Cabedo (Reference Gras and Cabedo2022) analyze independent constructions introduced by a ver si (lit. ‘to see if’) in Spanish, which can convey different modal-illocutionary values, like questions, wishes, and expressions of fear, and they find that these correlate with different nuclear configurations. Similarly, Fried and Machač (Reference Fried and Machač2022) analyze insubordinate constructions introduced by jestli (lit. ‘if, whether’) in Czech, which can express different degrees of epistemic certainty, and they observe that explicative constructions (‘I think that maybe p’) have slightly rising melody, whereas argumentative constructions (‘I think that probably not p’) have sharply falling intonation. Differences in prosody offer formal support for differences of meaning and are taken as evidence of separate constructional status.
14.5 Conclusions
One of the most attractive aspects of constructional approaches is that the notion of grammatical construction is flexible enough to represent whatever linguistic information is needed to explain how a linguistic pattern is used, regardless of whether the relevant information belongs to phonology, morphology, syntax, semantics, pragmatics, or discourse. While the meaning potential of constructions has been fruitfully developed in constructional literature, the prosodic dimension of constructions has received comparatively less attention.
The study of a marginalia, like insubordination, helps shed light on the discursive and prosodic functioning of grammatical constructions. As for the discursive aspects, it has been shown that CICC has discourse restrictions, but these are not comparable to morphosyntactic constraints which need to be satisfied by each instance of the construction. An alternative to including discourse information is considering constructions as complex categories made up of a schema and a series of instances and extensions. This analysis has several advantages: It (i) accounts for the effects of prototypicality (rebuttal as the main function), (ii) incorporates discursive information at the level of instances and extensions (not at the schema level), (iii) allows describing peripheral meanings (e.g., justification) with greater precision, and (iv) reflects dialectal variation since not all functions are attested in each geographical variety (e.g., mirativity).
Regarding prosody, it has been shown that CICC combines with a focus prosodic pattern. However, the prosodic information is not idiosyncratic, since the pattern exists outside the construction. The theoretical possibility explored in this chapter is that the prosodic patterns of a language (or language variety) can be represented as schematic constructions that pair a prosodic contour (form) with a pragmatic function (meaning) that are inherited by sentence-level constructions if their meanings are compatible.
In sum, a constructional model which considers constructions as complex categories with inheritance relationships from abstract constructions can account also for the discursive and prosodic features of constructions: Discursive information is better represented at the level of instances, which show different degrees of prototypicality, while prosodic information is better represented as inherited from general prosodic patterns in the language.
