Available online at www.sciencedirect.com ScienceDirect Journal of Pragmatics 70 (2014) 130--151 www.elsevier.com/locate/pragma Meaning construction in verbomusical environments: Conceptual disintegration and metonymy Paula Pérez-Sobrino * Department of Modern Languages, University of La Rioja (Logroñ o, Spain), c/ San José de Calasanz 33, 26004 Logroñ o, La Rioja, Spain Received 6 October 2013; received in revised form 18 June 2014; accepted 22 June 2014 Abstract In this paper I explore the workings of meaning (re)construction strategies in programmatic musical works, where the music stands for a broader extra-musical narration. The analysis of ten fragments of classical and contemporary music involving text and music reveals that conceptual disintegration in connection to metonymy emerges as a crucial tool for meaning (re)construction in programmatic music. This research presents four major contributions to the field. First, this paper holds for the complementariness of networks of conceptual disintegration and metonymic mappings in order to convincingly account for conceptual disintegration as a product (i.e., the multimodal expression) and as a process (i.e., the conceptual operation). Second, concerning the product, this paper provides a theoretical categorization of conceptual disintegration in terms of the ‘‘degree of disintegration’’ and ‘‘degree of subsidiarity’’ between the represented part and the whole conceptual package. Third, concerning the process, this work claims that metonymy arises as powerful analytical tool because it counts on a higher degree of constraint than blends. A view from Conceptual Metonymy Theory allows us to expand the inventory of possible meaning reconstruction processes in multimodal use: metonymic echoing, metaphtonymy, metonymic cueing, source-in-target metonymies and multiple source-in-target metonymies. Fourth, this paper deals with musical and verbomusical examples, largely unexplored in cognitive-linguistic studies. © 2014 Elsevier B.V. All rights reserved. Keywords: Conceptual disintegration; Inference; Meaning construction; Music; Multimodal metonymy; Patterns of interaction 1. Introduction This paper aims at delving deeper into the different levels of multimodal meaning rendered in verbo-musical environments. More specifically, it focuses on a particular strategy of musical meaning (re)construction, 1 i.e., the way in which different patterns of conceptual disintegration that structure the multimodal expression trigger the activation of metonymic reasoning at the conceptual level, in ten examples of program classical and contemporary music involving music and text. Program or programmatic music refers to a type of art music that attempts to render an extra-musical narrative. While the vast majority of music scholars readily acknowledge the existence of programmatic * Tel.: +34 941 299416; fax: +34 941 299419. E-mail addresses: paula.perezs@unirioja.es, paula.perez.sobrino@gmail.com. 1 The term ‘‘meaning (re) construction’’ is preferred in this paper over ‘‘meaning construction’’ because it refers to both the production and interpretation of meaning. http://dx.doi.org/10.1016/j.pragma.2014.06.008 0378-2166/© 2014 Elsevier B.V. All rights reserved. P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 131 pieces, the idea of reconstructing a composer’s intention per se is actually somewhat controversial2 (cf.; Barthes, 1965; Bloom, 1973; Wimsatt, 1954). The interpretation of the message that program music aims to convey is always dependent on the inferential capabilities of the speakers to a higher or lower extent. In this paper, we contend that this process of meaning reconstruction is aided by the combined workings of conceptual disintegration and metonymy. As shall be argued, these theoretical constructs guide and constrain the meaning reconstruction process, so that the inferential load is largely reduced, the cognitive cost lessened, and the intended message more straightforwardly recovered. In this paper, conceptual disintegration relates to a selective projection of a multimodal manifestation, whereas metonymy refers to the cognitive process based on granting access from a subdomain (in this case, the partially represented multimodal representation) to a larger and more-encompassing matrix domain. Disintegration and metonymy emerge as crucial tools for meaning (re)construction in programmatic musical works, where the music evokes a broader extra-musical narration. Disintegration as a tool for meaning (re)construction has not received very much scholar attention. Although Fauconnier and Turner (2002:119) have acknowledged its concomitancy to conceptual integration,3 nothing has been published in this regard since Hougaard’s (2005) and Bache (2005) interesting insights in Journal of Pragmatics. Framed within Conceptual Blending Theory (CBT; Fauconnier and Turner, 2002), Hougaard (2002:3) defines disintegration as ‘‘a process by which one unified and discrete structural element in mental space gets to receive multiple counterpart relations and is projected to (an)other mental space(s) as two, or more, separate structural elements’’. Notwithstanding the attention that CBT has drawn to the issue of conceptual disintegration, the paradigm cannot sufficiently account for the whole phenomenon of meaning (re)construction. Among other shortages (for a critical review of CBT, see Gibbs, 2000; Ritchie, 2004; Ruiz de Mendoza, 1998), the conceptual blend cannot provide us with a convincing separation between disintegration as a product (the linguistic or multimodal expression) and as a process (i.e., the conceptual operation). As Gibbs emphasizes (2000:531), ‘‘the processes of linguistic understanding are different from the products we consciously think about when we read or hear verbal expressions’’ (let ‘‘linguistic’’ and ‘‘verbal’’ stand for ‘‘multimodal’’ for the purposes of this paper). In order to comply with this theoretical distinction, this research delves deeper into the nature and scope of conceptual disintegration and offers an analysis that differs from that of Hougaard (2005) and Bache (2005) in at least four ways. First, it holds for the complementariness of networks of conceptual disintegration and metonymic mappings in order to convincingly account for disintegration as a product (the linguistic or multimodal expression) and as a process (i.e., the conceptual operation). Meaning (re)construction is here understood as a two-step process: the first involves the configuration of certain cues to structure the multimodal manifestation ( product), and the second relates to the cognitive operations triggered at the conceptual level by those multimodal cues ( process). Second, as regards the partial representation of a multimodal expression, or product, this paper provides a theoretical categorization in terms of the ‘‘degree of disintegration’’ (which yields a distinction between substitution and fragmentation) and ‘‘kind of subsidiarity’’ between the represented part and the whole conceptual package (on the basis of inherent dependency or ad hoc dependency). This two-variable taxonomy, here labeled as the Multimodal Conceptual Integration Model (MCIM, Pérez-Sobrino, 2014), goes a step further than CBT in offering at least two criteria to distinguish different types of disintegrated conceptual networks and the principles motivating them. Third, concerning disintegration as a process, this work claims that metonymy arises as a powerful analytical tool in connection to conceptual disintegration because it counts on a higher degree of constraint than blends and it features a greater interactional dimension. As will be shown, the patterns of interaction between metonymies propounded in Ruiz de Mendoza (2000), i.e., metonymic expansion and metonymic reduction, can account in finer grain for all the inferential activity arising from Hougaard’s (2005) conceptual disintegration processes, i.e., splitting and partitioning selection. Furthermore, a view from Conceptual Metonymy Theory (henceforth, CMyT) allows us to expand the inventory of possible meaning reconstruction processes: evidence is shown to prove the productive working of multimodal metonymic echoing, multimodal metaphtonymy, multimodal metonymic cueing, multimodal source-in-target metonymy and multimodal 2 For instance, Mailman (2012) shows tight connections between the intention of the composer and the program structuring the music (e.g., the prelude to Wagner’s Das Rhinegold, where the inevitable flowing of a river is depicted as a gradual increase in pitch range, rate of repetition, number of ostinato layers, orchestral thickness, and proportion of upward motion; Mailman, 2012: chapter 6) and equally meaningful correspondences between a program and a musical work that was not meant to refer to any extra-musical reality (as it is the case of Lucier’s Crossings, a minimalistic piece which unpurposefully yet readily ‘‘evokes the revving of an airplane engine, ominous sirens or a nuclear explosion in slow motion’’; Mailman, 2012: chapter 6). Mailman (2009) thus argues in favor of the feasibility of a strong coordination between the musical work and the structuring program, while also stressing the legitimate role of discourse in creating meaning out the musical work, given that ‘‘what seems more narrative in kind depends a lot on convention and context’’ (Mailman, 2009:406). 3 In words of Fauconnier and Turner (2002:119),‘‘integration and compression are one side of the coin; disintegration and decompression are the other’’. 132 P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 multiple-source-in target metonymies as additional means to reconstruct absent information. This proposal proves also favorable to the issue of predicting the cognitive impact of conceptual disintegration, since the scope of possible inferences arising from a metonymic mapping is much more narrowed down than that of blends. The fourth novelty of this paper resides in the corpus of study. Traditionally, research in CBT and CMyT has offered linguistic evidence to support their claims on the conceptual working of everyday thought. However, further evidence in non-verbal modes is necessary for a theory about thought to be exhaustive. In this regard, musical understanding remains a complex and largely unexplored issue in Cognitive Linguistics (CL). Only musical metaphor has drawn scholarly attention (Brower, 2000; Cook, 1992; Cox, 2011; Forceville, 2009a; Larson, 2012; Larson and Johnson, 2004; PérezSobrino, 2014; Spitzer, 2008; Zbikowski, 2002). In compliance with this theoretical necessity, this work provides evidence of the productive workings of musical metonymy in a selected bulk of multimodal examples including text and music. As will be shown, the interaction of metonymy with other cognitive operations guides and steers the conceptualization of certain musical elements as a person, a person in motion or even an event. The remainder of this paper is structured as follows. In section 2, I address some basic theoretical notions that will ground the subsequent analysis. I deal with the concept of conceptual disintegration (2.1) which, for the purposes of this paper, is subdivided in cases of substitution and fragmentation. I briefly introduce some basic notions of metonymy (2.2) and sketch the basic workings of meaning reconstruction operations (2.3), in which patterns of disintegration creatively combine with metonymy. In section 3 I turn to analyze several patterns of disintegration in combination with metonymy (which I illustrate with ten examples of classical and contemporary music with text). I retrieve the main proposals of this paper in section 4, and suggest further lines of potential research. 2. Theoretical framework 2.1. Conceptual disintegration Conceptual disintegration has been traditionally associated to selective projection in CBT studies4 (Turner, 2001; Fauconnier and Turner, 2002). Within this paradigm, disintegration serves to fragment or partition conceptual wholes into elements that may be recruited for individual projection to blended spaces, in a process that usually remains unnoticed. However, Hougaard (2005:1673) notices that selective projection requires an extra effort when the selected projected elements stem from the same tightly integrated source. These cases, labeled as partitioning selection (to make a distinction to regular selection, Hougaard, 2005:1673), thus require a more laborious inferential activity. This observation sets an interesting starting point to survey the unattended facet of conceptual disintegration as a necessary counterbalance to traditional analysis of conceptual integration in form of blends. Subsequently, Bache (2005) has continued in this line by further acknowledging the vital role of disintegration within CBT. In spite of this interesting insight, the pervasiveness and ubiquity of CBT to all kind of mental activity posits several problems, mainly concerned with a satisfactory insight on specific phenomena. I share Gibbs’ (2000) skepticism in relation to the excessive explanatory power of blends, which makes them vulnerable to criticism. From CBT, it still remains to be explained how the blend as a process influences the product of conceptual disintegration. In the remainder of this section, I offer a tentative taxonomy to frame the interplay of networks of conceptual disintegration with metonymy (as will be shown in section 4). In this paper, conceptual disintegration is considered part of a broader descriptive framework designed to describe patterns of conceptual integration in a greater degree of detail than CBT: the Multimodal Conceptual Integration Model (MCIM; Pérez-Sobrino, 2014, wherein the model is applied to the study of verbomusical networks of conceptual integration). This new typology incorporates three basic novelties with respect to CBT: (a) it establishes different degrees of (dis)integration between conceptual inputs; (b) it contemplates different degrees of subsidiarity between the integrated inputs; and (c) it applies to multimodal hybrids, not only to linguistic examples.5 4 Conceptual blending is a cognitive operation whose uniform and dynamic structure underlies probably all cognitive phenomena, including categorization, metaphor, metonymy, metaphor, and analogy among others (analyses of conceptual integration in linguistic examples from a CBT perspective are given in Coulson, 1996; Fauconnier and Turner, 1998, 2002; Oakley, 1998). The main exponent of research on conceptual integration in music and verbomusical instances from a CBT perspective is Zbikowski (2002--present). 5 For an alternative model to deal with interrelationships between various component media, and potential sources of potential meaning(s), see Cook’s (1992) ‘‘Models of Multimedia’’. Briefly put, the essence of the model is to determine whether component media are communicating consistent, coherent or contrastive information via different perceptual modalities. In spite of its proved validity, the emphasis put on the interrelation of modalities does not make it a suitable model to study conceptual disintegration networks. By contrast, the interest of the MCIM in the degree of disintegration and the kind of dependency between the represented input and the domain to which it belongs lends further support to the choice of this model to explore how the multimodal expression motivates different cognitive patterns at the conceptual level. [(Fig._1)TD$IG] P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 133 Fig. 1. Schematic representation of conceptual disintegration by substitution and fragmentation (black for partial present structure; white and discontinuous line for absent structure which needs to be reconstructed or inferred). Regarding the degree of integration, the MCIM places conceptual disintegration and conceptual integration at opposite extremes of the same continuum. Conceptual disintegration designates the highlight of partial selection from a certain conceptual package, whereas conceptual integration involves the incorporation of several elements or partial projections of several elements. The productivity of the patterns of interaction between multimodal elements makes it impossible to build an exhaustive inventory of the unlimited possibilities of disintegration. For the sake of clarity in the explanation, I hitherto restrict myself to set up identifiable cases at the extremes of a continuum of disintegration.6 Substitution is a subcase of conceptual disintegration whereby pre-existent material (A) is fully replaced by related material (A’). Take for example a situation where one person is yelling with anger to another, and the latter says ‘‘Don’t bark at me!’’. Here, ‘‘bark’’ has substituted human speech verbs such as ‘‘yell’’ or ‘‘shout’’. Due to the lack of linguistic traces pointing out that the addressee is a human being, the hearer has to rely on the context to infer it. This task is relatively easy to accomplish, since humans and animals are related by the Great Chain of Beings (Lakoff and Turner, 1989), a cultural folk model shared across cultures. Another possibility of disintegration is fragmentation. This subcase consists in the exclusive focus on partial information (a) over the rest of the broader conceptual domain (A). For instance, ‘‘I need a hand here’’ highlights ‘‘hand’’ over the whole domain ‘‘worker’’ or ‘‘person who can help’’. See Fig. 1 for an overview of the two processes. The introduction of the second variable kind of dependency increases the complexity of the issue, but at the same time offers a feasible explanation of the different principles governing conceptual disintegration. From the strongest sort of dependency to the weakest, this model contemplates two possibilities of relationship between the conceptual structure and its partial representation: inherent enrichment and ad hoc enrichment (in an application to multimodal environments of Ruiz de Mendoza, 2011:108). This distinction sheds some light on the expected cognitive effort invested in the meaning reconstruction task. The more entrenched the relationship is between the musical representation and the extra-musical narration it refers to, the stronger their dependency is, and thus the less necessary it is to refer to contextual elements to recover the absent conceptual structure necessary to interpret the musical narrative. By contrast, the weaker the dependency between the musical representation and the external structure to which it relates, the more difficult it is to interpret the musical piece, and therefore, the greater the need to refer to contextual information. As will be claimed, the violin trills in example 1 resemble the chirp of birds (in terms of melody, rhythm, dynamics, etc.), so that the hearer could do without external cues to make that connection (for instance, the title Spring reinforces this supposition, but it is not an essential part of the interpretive task). By contrast, in example 3, the ad hoc connection between the main theme and the character it represents, i.e., the beloved, makes it necessary for the hearer to rely on information external to the musical piece (the textual program provided by the composer) to interpret the Fantastic Symphony as an autobiography. In enrichment, conceptual disintegration takes place when the relationship between the inputs involved is of inherent subsidiarity or ad hoc subsidiarity. Take the example ‘‘The car industry is turning green’’; here ‘‘green’’ substitutes ‘‘naturefriendly product’’. This connection is possible because green is the color of the chlorophyll, a pervasive and intrinsic component in nature. Therefore, ‘‘green’’ and ‘‘nature-friendly’’ share a relation of inherent dependence in the context of natural environments. Additionally, the elements involved in a substitution or fragmentation network can become dependent on one another through ad hoc dependency. Consider the example ‘‘He has climbed his way to the top’’, where the PATH and the VERTICALITY schema, while not initially interdependent, become subsidiary to one another in order to express that the greatest achievements are at the top of an upward path. See Fig. 2 for a schematic representation. Table 1 provides a schematic overview of this paradigm for the study of multimodal conceptual disintegration. Example 1 (Vivaldi’s Four Seasons -- the Spring), 2 (Schubert’s Gretchen by the Spinning Wheel), 3 (Berlioz’s Fantastic Symphony), and 4 (Edith Piaf’s ‘‘Je ne regrette rien de rien’’ in Christopher Nolan’s movie Inception) render cases of 6 The reader is referred to Author (2014) for the explanation and examples of additional types of conceptual integration (incorporation and multiple incorporation) and degrees of dependency (combination) within this model. 134 P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 Table 1 Conceptual disintegration in the MCIM (with examples). Degree of disintegration Kind of dependency Enrichment Inherent subsidiarity Ad hoc subsidiarity Substitution Fragmentation 1) Vivaldi’s violin and viola trills in The Four Seasons - Spring 2) Schubert’s left hand in Gretchen by the Spinning Wheel* 3) Berlioz’s idée fixe in Fantastic Symphony 4) Nolan’s Inception adaption of Edith Piaf’s Non, je ne regretted rien 6) Grieg’s In the Hall of the Mountain King* 5) Papageno’s theme in Mozart’s The Magic Flute 7) Ravel’s use of the pentatonic scale in Laindenotte, Imperatrice des Pagodes 8) Lohegrin and Elsa’s leitmotifs in Wagner’s Lohengrin 9) Nokia tune 10) Sword leitmotif in Wagner’s The Ring of the Nibelungs [(Fig._2)TD$IG] Fig. 2. Conceptual disintegration by inherent enrichment in ‘‘The car industry is turning green’’ and by ad hoc enrichment in ‘‘He has made his way to the top’’ (white for absent domain; black for present subdomains that are or have become dependent to the more-encompassing domain). substitution where the metonymy is triggered by an absent cue. This is the case of instrumental pieces where the title or the textual program, not present in the music, play a more or less important role in the interpretation of the musical work. The manner in which the hearer recovers the missing information, determined by the type of dependency between the partially represented input and the whole conceptual material to which it belongs, is also variable: examples 1 and 2 show cases where the proper reconstruction of the musical narrative is based on the physical resemblance between the external physical event (the birds’ chirp and the motion of a spinning wheel, respectively) and the musical motive it motivates (either via the direct imitation of the external sound in the melody or the reproduction of the trajectory of the event in the melodic line). Since there is a manifest and recognizable analogy (in terms of melody, melodic line, dynamics, rhythm, etc.), the musical motive is inherently dependent on the event it resembles. However, in examples 3 and 4, the musical expression becomes dependent on an external linguistic cue provided by the surrounding verbomusical context for the extra-musical narration to be interpreted. The rest of the examples, 5 (Mozart’s The Magic Flute -- Papageno’s theme), 6 (Grieg’s Peer Gynt -- In the Hall of the Mountain King), 7 (Ravel’s Mother Goose -- Laindeonette, Empress of the Pagodas), 8 (Wagner’s Lohengrin -- Lohengrin and Elsa leitmotifs), 9 (Nokia tune), and 10 (Wagner’s The Ring of the Nibelungs -- Sword leitmotif) deliver cases of fragmentation. In these examples, the incomplete conceptual structure that triggers the metonymic operation is present in the multimodal expression. Once again, the interpretation of this multimodal input relies either on the hearer’s encyclopedic knowledge or on ad hoc contextual cues. For instance, the ensuing analysis shows that only in example 6 the rhythmic input is inherently dependent and closely imitates the motion event it describes. The rest of the cases show more or less arbitrary ascriptions between their music and the extra-musical narrative they represent that have become dependent through the programmatic musical work. [(Fig._3)TD$IG] P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 [(Fig._4)TD$IG] 135 Fig. 3. Source-in-target metonymy and metonymic expansion in ‘‘Richard’s at war with himself over who to marry’’. Fig. 4. Target-in-source metonymy and metonymic reduction in the ‘‘surgical bombing’’ example. 2.2. Metonymy The partial representation of a multimodal input requires a cognitive operation that enriches and expands the rendered highlighted part and grants access to the whole conceptual domain in order to make full sense of it. Hougaard (2005) set off an interesting foray in the identification of delimited processes of conceptual disintegration. Splitting (2005:1664) refers to ‘‘the process by which something perceived as an integrated structure gets to be mapped onto two or more discrete structural elements in another mental space and is projected as two or more separate structural items’’. By way of illustration, he explains that in ‘‘Richard is at war with himself over who to marry’’, ‘‘Richard’’ is projected onto two participant roles, ‘‘Richard1 who wants to marry X’’ and ‘‘Richard2 who wants to marry Y’’, who are portrayed as adversaries. Splitting thus elaborates on the conceptualization of the hinted involved structure. Commensurate with this rationale, a metonymy also accounts for this example. Metonymy is a cognitive operation by which we conceptualize one thing by its relation to something else (Lakoff and Johnson, 1980:139; see also Barcelona, 2000, 2011; Ruiz de Mendoza, 2000, 2013). Each self (self1 or ‘‘Richard who wants to marry’’ and self2 ‘‘Richard who doesn’t want to marry’’) constitutes a subdomain of a more encompassing matrix domain (SELF or ‘‘Richard’’). Interestingly enough, Ruiz de Mendoza (2000) has already identified source-in-target metonymies, where the source (self1 / self2) is a subdomain that affords access to the most encompassing domain (SELF) via a metonymic expansion process (see Fig. 3). Source-in-target metonymies provide a cognitively economical point of access to a complex scenario to the extent required for the interpretation of the utterance to be possible. Additionally, Hougaard brings forward a second process of conceptual disintegration that is also amenable to explanation from CMyT. Partitioning selection (Hougaard, 2005:1673) designates ‘‘the selected and projected elements [that] stem from the same tightly integrated source’’. In order to illustrate his argument, this scholar analyzes a chunk of radio conversation where the speaker compares massive bombings in Iraq to those in Dresden during IIWW. This comparison is possible given that both ‘‘Bagdad’’ and ‘‘Dresden’’ belong to the ‘‘bombing’’ frame, and it is meant to highlight the allied intention to cause as much destruction as possible. However, the interlocutor sets up barriers against unwelcomed automatic projections by referring to allied intervention in Bagdad as ‘‘surgical bombing’’ (meaning that it was carried out with precision, contrary to the attack of Dresden). This new feature holds primacy over inferences that follow from the bombing frame that could also apply for the German input. The highlight of a certain feature (‘‘surgical bombing’’) over others (‘‘destructive bombing’’) is also explained by Ruiz de Mendoza (2000) in terms of target-in-source metonymies. See Fig. 4 for a schematic overview of this metonymic mapping, in which a reduction process highlights the target (subdomain ‘‘surgical bombing’’) over the whole source (matrix domain ‘‘BOMBING’’), whereas other subdomains remain obscured: this is the case of bombings in Dresden (subdomain ‘‘massive bombing’’). This specific metonymic mapping steers our attention toward the most relevant aspects of the target, which are seen from the perspective of their corresponding source elements. 2.3. A two-stepwise approach to meaning (re)construction An approach to meaning reconstruction that combines the insights from both the MCIM and CMyT avoids the kind of theoretical limitations I have addressed in the previous section in relation to CBT. First, metonymic expansion and reduction mappings (Ruiz de Mendoza, 2000) are commensurate with Hougaard’s (2005) proposal of two disintegration processes, namely, splitting and partitioning selection. Additionally, a metonymy-oriented account counts on a greater [(Fig._5)TD$IG] 136 P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 Fig. 5. Substitution and metonymy (black for the present multimodal manifestation; white and discontinuous line for absent material in the multimodal manifestation; white and continuous line for accessible information via metonymic expansion). [(Fig._6)TD$IG] Fig. 6. Fragmentation and metonymy (black for the present multimodal manifestation; white and discontinuous line for absent material in the multimodal manifestation; white and continuous line for accessible information via metonymic expansion). interactional dimension with other tropes (as will be shown in the ensuing analysis) with a higher degree of constraint than conceptual blends. In fact, a closer look at the data shows that only metonymic expansion processes are productive meaning reconstruction strategies for disintegrated multimodal products. It only seems natural, given that the hearers are expected to construe a whole story out of a partial verbomusical rendering. Second, it successfully addresses the difference between product and process. That would lead us to conclude that this two-stepwise strategy also goes beyond the scope of CMyT, since it also deals with differences at the dimension of the multimodal manifestation (or product), and not only at the conceptual level (or process). Figs. 5 and 6 show a schematic overview of the two possibilities of meaning reconstruction discussed in this paper. In substitution in connection to metonymy (Fig. 5), partial conceptual material (a’ in black) fully substitutes related partial conceptual material (a in white and discontinuous lines). This partial representation (a’) affords access to the most encompassing domain (A) via a metonymic expansion process. Recall once again the example presented in Fig. 1, ‘‘Don’t bark at me’’. Once that ‘‘bark’’ substitutes human speech verbs such as ‘‘yell’’ or ‘‘shout’’, a metonymic expansion process of the type SOUND FOR PRODUCER affords access to the whole domain ‘‘person’’. The effect of the combined workings of substitution and metonymic expansion is the conceptualization of a bad-mannered person that treats others with the same anger as barking dogs. This meaning (re)construction strategy is followed in examples 1--4 (cf. Table 1), as will be shown in the next section. In turn, fragmentation connected to metonymy (Fig. 6) relates to the way in which partial conceptual material (a’ in black) is projected over the rest of the multimodal representation (A in white). This partial representation (a’) is part of the subdomain (a) or coincides with the subdomain (a) that affords access to the most encompassing domain (A) via a metonymic expansion process. If we consider again the example already offered in Fig. 2, the relevant role of the subdomain ‘‘hand’’ within the broader domain ‘‘worker’’ makes it salient in expressions such as ‘‘I need a hand here’’. A P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 137 metonymic expansion process of the type HAND FOR WORKER makes of ‘‘hand’’ an economic point to access the broader scenario in which a person is able to help in physical labor involving the hands. This meaning reconstruction strategy is present in examples 5--10 (cf. Table 1). Third, CMyT allows us to expand CBT accounts of disintegration processes by offering an expanded inventory of patterns of interaction between metonymy and other tropes (Ruiz de Mendoza, 2007) in multimodal use (Hidalgo and Kraljevic, 2011; Pérez-Sobrino, 2014; Uriós-Aparisi, 2009), as will be shown in section 3. 3. Multimodal meaning reconstruction strategies: towards the construction of an inventory Although there is yet no published research work on verboaudial metonymies, there is some preliminary literature on verbopictorial metonymies (cf. Forceville, 2009b; Hidalgo and Kraljevic, 2011; Uriós-Aparisi, 2009) and verboaudial metaphor (Author, 2014; Forceville, 2009a; Zbikowski, 2002). This kind of research supports the productivity of the application of tools for analyzing verbal communication to multimodal environments. This working assumption has been labeled equipollence hypothesis in Ruiz de Mendoza and Mairal (2011). This methodological stance states that, in order to advance our knowledge in the field, the analyst should test whether linguistic processes that have been attested in one domain of linguistic enquiry may also be (at least partially) active in other domains.7 In this connection, I argue that the application of the equipollence hypothesis in a reverse way, i.e., from non-verbal to verbal communication, makes it possible to enhance the scope of current verbal-based accounts of these two tropes in CL. Therefore, this work contributes to further the body of knowledge on CMyT by looking at multimodal manifestations of this trope. For the sake of explanatory clarity, I have singled out specific details of each of the analyzed examples in order to build an inventory of different strategies of meaning reconstruction. By no means should the ensuing list imply that these categories are mutually exclusive. In fact, given that the selected verbomusical examples are rich and complex in nature, it is usual that one strategy leads to another over the course of the musical work (as will be made apparent in section 3.3.4). 3.1. Multimodal metonymic echoing vs. multimodal metaphtonymy In this section I discuss two types of conventional associations between the musical work and the extra-musical reality to which it refers. Conventionality, as understood by Mailman (2009:406), arises from the hearer’s perceived similarity between the musical work and the underlying program. The achievement of such analogy greatly depends on the amount of shared heard surface features (in terms of melody, dynamic, rhythm, etc.) between music and extra-musical narration, making of ‘‘conventionality’’ a gradable concept. At one extreme of the continuum, the most direct form of perceived similarity is musical echoing, an operation whereby music directly imitates a sound from the physical world. Take for example the musical analogy between the chirp of birds and violin trills from Vivaldi’s Four Seasons, The Spring (example 1). In this example, the musical birds replace the birds from the real world (substitution) via direct imitation. Given their manifest similarity, it becomes clear that the musical rendering is inherently dependent of the real chirp of birds. Echoing (Ruiz de Mendoza, 2011) is a cognitive operation whereby the close imitation of one conceptual domain (violin trills) straightforwardly replicates another domain (the way birds sing). This ostensive similarity helps hearers to construct a broader extra-musical reality: as shown below in Fig. 7, the violin trills (which echo the warble of birds) metonymically stands for birds as animals; what is more, these animals can also metonymically stand for spring, for the bird chirping is a central defining feature of this season. 7 For instance, the application of the equipollence hypothesis within the Lexical Constructional Model (LCM; Mairal and Ruiz de Mendoza, 2009; Ruiz de Mendoza and Mairal, 2011) has allowed the identification of metaphor and metonymy in various degrees of abstraction and in interaction (the so-called metonymic and metaphoric complexes) as underlying inferences and other kinds of construal at both argument structure and higher levels of linguistic description (e.g., pragmatic implication, illocutionary meaning and discourse). [(Fig._7)TD$IG] 138 P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 Fig. 7. Echoing and metonymic expansion in Vivaldi’s Four Seasons, Spring (black thick arrow for substitution and echoing; simple black arrow for metonymic mapping and mode-shift). In a less direct manner of imitation, a musical feature (melodic line, rhythm, harmony, etc.) can resemble a paradigmatic feature (such as the shape or trajectory) of a soundless object in motion. Consider the case of Schubert’s Gretchen by the Spinning Wheel (example 2). The image-schematic structure that is relevant for a spinning wheel (circular shape, never-ending circular pattern) is put in correspondence with the image-schematic structure of the melodic line, so that it closely resembles the constant and cyclical motion of wheels. In terms of conceptual disintegration networks, this example is similar to the previous one: the musical rendering substitutes the extra-musical narrative of the wheel on the grounds of a relationship of inherent dependence between the paradigmatic defining feature of the physical wheel and its musical version (which has consequently adopted a constant and cyclical melodic pattern). It can be argued that this ascription is somehow more arbitrary than the example discussed above: how can music univocally refer to a soundless event? Although this case could be analyzed as an instance of metonymic cueing (to be discussed in the following subsection), it is worth noticing that the association between the melodic theme and a wheel is also less optional than in the examples 3 and 4 discussed below, given that the repeated undulant melody and the wheel trajectory share a common ground property (namely, the constant and regular motion). This example shows that there are not clear-cut boundaries between the categories of the MCIM (as schematized in Table 1) and that this model offers a continuum of different possibilities of conceptual (dis) integration networks, rather than a closed typology. All in all, this example clearly differs from example 1 in terms of the meaning reconstruction operations at play. The type of mapping based on the perceptual similarity between the trajectory of a wheel and the melodic line combines metaphor and metonymy, in line with Goossens (1990) and Ruiz de Mendoza’s (2000) notion of metaphtonymy. A conceptual cluster puts in perspective a prominent attribute of the source domain (constant and cyclical motion of wheels), which is mapped onto the target (music) thus highlighting a corresponding attribute (namely, the constant and undulating accompaniment ostinato).8 Once the analogy between the central feature of wheels and the melodic line has been 8 This metaphorical mapping based on perceptual resemblance is commensurate with Grady’s (1999) resemblance metaphor. Grady established that resemblance metaphors are very basic mappings between perceptually similar entities. [(Fig._8)TD$IG] P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 139 Fig. 8. Substitution and metonymic expansion of the metaphoric source in Gretchen by the Spinning Wheel (black thick arrow for substitution; simple black arrows for metonymic mappings; white arrow for resemblance metaphoric mapping and mode-shift). established through metaphor, a metonymic expansion process takes place in the metaphorical source domain in order to access the wider context suggested by the lyrics, i.e., a girl sitting by a spinning wheel (see Fig. 8). In particular, this example is discussed among Ruiz de Mendoza’s (2000) patterns of interaction under the label of metonymic expansion of the metaphorical source. The metonymy (PROPERTY FOR OBJECT, whereby ‘‘constant and regular motion’’ stands for ‘‘spinning wheel’’) provides a cognitively economical point of access to a complex scenario. Therefore, the metonymy has the function of developing the point-of-access subdomain to the extent required for the metaphor to be possible. Examples 1 and 2 fit into the same category because echoing and metaphor are two conceptual operations of the A IS B form. In spite of their generic form, and their similarity in terms of conceptual disintegration networks (cf. Table 1), echoing and metaphorical resemblance are not the same kind of conceptual correspondence: in echoing, one conceptual domain (A) imitates a different conceptual domain (B); in metaphor, on the contrary, we use certain aspects of the source domain (B) to talk, reason and understand some features of the target domain (A). 3.2. Metonymic cueing As put by Mailman (2009:406), ‘‘the more unconventional a musical work is, the more it demands that narration be actively supplied, either by its perceiver or a third mediating party’’. This is precisely the effect of cueing (Ruiz de Mendoza, 2011:105), a basic cognitive operation whose function is to allow the addressee to select relevant conceptual structure and organize it for [(Fig._9)TD$IG] Fig. 9. Linguistic cueing between the text and the music and metonymic expansion in Berlioz’s Fantastic Symphony (black bold arrow for cueing; black simple arrows for metonymic mappings and mode-shift). 140 P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 the interpretive task that is taking place. This notion, which has been attested so far in linguistic examples (cf. Ruiz de Mendoza, 2011: ff., Ruiz de Mendoza and Galera, 2014:87 ff.), proves equally relevant to the analysis of multimodal examples (with some adjustments, as shown below). More specifically, this paper deals with the connection between cueing and multimodal metonymy (although cueing applies invariably in all meaning construction tasks). By way of working definition, multimodal metonymic cueing can be characterized as a two-step process where the addressee (1) selects (partial) relevant structure in one mode and (2) makes it to coincide with a subdomain that accesses all the information of the most encompassing domain rendered in a different mode via metonymic expansion. Berlioz’s Symphonie fantastique is an illustrative example of the interaction between an instrumental piece and an independent textual program. As stated by the composer in the notes accompanying each of the five movements, ‘‘the following programme must therefore be considered as the spoken text of an opera, which serves to introduce musical movements and to motivate their character and expression’’ (Berlioz, 1845). Under the influence of opium (in the 1855 version), a young and sensitive artist (Berlioz himself), experiences a series of visions (the different movements of the symphony) in which his beloved figures as a theme or idée fixe (see example 3). The idée fixe recurs in every movement as a haunting obsession (although each time in a different form). One of the characteristics of instrumental pieces is that they count on two independent components (verbal text and musical part) that render disintegrated versions on their own of the musical work. Text and music are thus processed in chronological fashion, where the music substitutes the linguistic program once the instrumental interpretation begins. However, notice that the fact that they are independent does not mean that they are not related. In fact, the narration is essential to select and activate relevant information in the musical part; and therefore, it becomes subsidiary to the music to set the construction of the musical narration. Given that they are not inherently dependent, it is necessary to refer to the program provided by the composer in order to select the relevant structure from the music to decode the extra--musical narrative. Hearers need to take into consideration the composer’s obsession for his beloved in order to interpret the idée fixe conveyed by the instrumental part. In this case, the program (Berlioz, 1845) linguistically cues ‘‘a certain quality of passion, but endowed with the nobility and shyness which he credits to the object of [the artist’s] love’’ in the music. Once the idée fixe has been recognized, the hearer needs to undertake a series of cognitive processes to interpret the theme as the beloved, and then the symphony as an autobiographic discourse of ‘‘an artist gifted with a lively imagination’’ who has ‘‘poisoned himself with opium’’ in the ‘‘depths of despair’’ because of ‘‘hopeless love’’ (Berlioz, 1845). This metonymic mapping is multimodal, since it expands a linguistic subdomain, i.e., the verbal allusion to the beloved, into a musical narrative, i.e., the recurring idée fixe (see Fig. 9). It can also happen that the music hints the construction of an entire textual narrative. For instance, this is a recurrent resource in movies to set the mood of a determined scene or to further characterize the protagonists. However, there are much more creative manifestations of music which are illustrative to this study. One outstanding example is the adaptation of Edith Piaf’s ‘‘Non, je ne regrette rien’’ in Christopher Nolan’s (2010) movie Inception (example 4). [(Fig._10)TD$IG] P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 141 Fig. 10. Musical cueing between Edith Piaf’s Non, je ne regrette rien and the extra-musical narration and metonymic expansion in Inception (black thick arrow for cueing; black simple arrow for metonymic mapping and mode-shift). Inception is a multidimensional dream-within-a-dream action film in which a main character switches between reality and several levels of dreaming in order to deal with an unresolved past issue. Since there are at least four levels of dreaming at play, it gets difficult to determine the exact dreaming level in which the character is as the story develops. In order to ease things up, there is a musical iconic motive (a slowed-down version of Edith Piaf’s song instrumentation in terms of a brass instrument fanfare) that helps the viewers to detect the moment in which the protagonists leave the real world and enter in one of the dreaming dimensions.9 Hence, the musical cue sets up the environment where the action occurs. In this case, the music substitutes the audiovisual narration as soon as the scene has turned from the reality to the dream dimension. Much in the same line with the previous example, music and audiovisual narration become subsidiary to each other in order to understand the story. Furthermore, a metonymic expansion process can provide viewers with additional extra-musical information, such as the level of dreaming the protagonists are (the plot involves at least four), by modifying musical features such as rhythm, instruments or harmony (Fig. 10). At first sight, metonymic cueing could seem to fit into the analysis of examples 1 and 2 analyzed earlier on, since it also involves a combination of substitution and metonymy. However, if we recall the classification offered in Table 1, a major difference arises: metonymic cueing here refers to ad hoc combinations of a partial input and a matrix domain that have become dependent through the disintegration process, whereas in echoing the partial representation is inherently dependent on the domain to which it belongs. On the other hand, metonymic cueing as a process is compatible with regular multimodal metonymies (examples 5--10) insofar as the metonymic mapping connects a subdomain with the matrix domain that is rendered in a different mode. However, note the difference in kind of multimodal expression that prompts the metonymic mapping. In metonymic cueing, the subdomain belongs to an entity ‘‘a’’ (in example 3, the linguistic reference to the beloved) that is substituted for an independent yet related entity ‘‘a’’’ (the idée fixe) that provides a vantage point to develop the musical program ‘‘A’’ via metonymic expansion (the character of the beloved, who paves the way to the understanding of the symphony as an autobiographic discourse). In multimodal metonymy, however, the motivation of the metonymic mapping is not substitution but fragmentation, since the subdomain ‘‘a’’ is already part of the same domain ‘‘A’’ that is accessed directly via metonymic mapping. 9 This musical strategy can be traced back to Saint-Saëns’ famous piece ‘‘Turtles’’ from The Carnival of the Animals (1886). As hold by Mordden (1980:237), Saint Saëns borrows two racy themes from Offenbach’s musical comedy, Orpheus in the Underworld (one of them the famous cancan), and slows then down to a snail pace in order to express the drowsiness of the turtle. 142 P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 3.3. Musical source-in-target metonymy As mentioned elsewhere, multimodal source-in-target metonymies involve a cognitive operation that enables a subdomain rendered in one mode (in this case, music) to stand for the most encompassing domain conveyed in a different mode (in this case, an extra-musical narrative). The partial multimodal representation that activates a metonymic mapping is rendered through fragmentation, and it can either fully coincide or be part of the metonymic subdomain that affords access to the matrix domain through a metonymic expansion process. As evidenced in this paper, metonymy plays a key role in program music, since it allows composers to portray the different characters intervening in the extra-musical narrative on the basis of an incomplete musical element.10 In the following I provide a detailed taxonomy of different types of multimodal metonymy on the basis of the type of rendered musical element (i.e., musical motive, rhythmic motive, key and scale). 3.3.1. Via musical motive As shown in the previous discussion on Berlioz’s idée fixe in example 3, it is usual in program music to identify a character with a musical theme, i.e., a combination of melody and rhythm. Sometimes the theme is so representative that a fragmented version of the theme (that is, a motive) is enough to give access to the matrix domain (or the character it represents). Consider Papageno’s theme in Mozart’s The Magic Flute (example 5a). The well-known theme is entirely displayed at the beginning of the opera in the aria ‘‘Der Vogelfänger bin ich ja’’ (‘‘I am the bird-catcher’’). In this musical piece, Papageno introduces himself to the audience and plays with his flute the characteristic tune with which he hunts birds. Later on, only the ascendant flute semifuses (highlighted in red in example 5a) are played in ‘‘Wie stark ist nicht dein Zaubertone’’ (‘‘How strong is thy magic tone’’, example 5b). In this scene, Prince Tamino spends the night waiting outside Sarastro’s temple. He has sent his friend Papageno to rescue Pamina, Tamino’s beloved fiancée. The Prince is alone and full of despair until he hears Papageno and Tamina’s voices inside the temple. The musical rendering of the plot reunites in the aria ‘‘Wie stark ist nicht dein Zaubertone’’ Tamino’s flute theme with Papageno’s characteristic motive from the 10 There are some precedents in the literature that hold for the possibility of a musical piece to stand for the personality and emotional life of a person (Blechner, 2001; Meyer, 1956). P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 143 previous aria ‘‘Der Vogelfänger bin ich ja’’ (the ascendant semifuses; also highlighted in red in example 5b). What is more, even Tamino’s lyrics in the opera reinforce this identification: ‘‘Ach, das ist Papagenoston!’’ (‘‘Ah, this is Papageno’s tone!’’). In this case, the audience is capable or matching the ascendant semifuses motive to the character Papageno thanks to the previous musical context, which has made the musical motive to become dependent of the bird-catcher character. This iconic ascendant flute motive is enough for Tamino and the audience to identify Papageno (hidden in the darkness) via the double metonymy MOTIVE FOR THEME FOR CHARACTER. In this mediated metonymic mapping, the MOTIVE rendered in ‘‘Wie stark is nicht dein Zauberton’’ provides the hearer with a vantageous point of access to the reconstruction of the character Papageno (which is not present in the scene because he is still trapped inside Sarastro’s castle). The double metonymy here recovers the whole theme from the earlier aria ‘‘Der Vogelfänger bin ich ja’’, which in turn stands for Papageno, the bird catcher in Mozart’s opera (see Fig. 11). 3.3.2. Via rhythmic motive A specific rhythmic pattern can also be deployed to structure a character or an event. Let’s consider the use of the rhythmic motive in ‘‘In the Hall of the Mountain King’’, from Grieg’s orchestral suite Peer Gynt (example 6). According to Ibsen’s plot, the music describes the flight of the main character, Peer Gynt, from the mountain wherein the trolls live. The key feature to interpret this fragment is the identification of the main rhythmic theme in terms of human footsteps. This matching is accomplished via a metaphorical mapping that establishes a correspondence between the musical theme and human footsteps. The metaphorical correspondence is possible due to the high resemblance between the sound and dynamics of human footsteps and the highly rhythmic musical theme. In terms of the relationship of dependence between the physical activity and the deployed rhythmic motive, the same observations made for example 2 apply here. One the one hand, human footsteps and the rhythmic motive are inherently subsidiary to each other because the similarity is based on experiential conflation, and the external contextual cues (such as the title and the knowledge of Ibsen’s story) reinforce this identification. One the other hand, it could also be that the identification of one in terms of another is indirect and needs external cues (although it is also true that the matching is less arbitrary than those in examples 3 and 4). Its borderline status is captured in Table 1. Once the hearer establishes an experiential correlation between the musical theme and a series of footsteps, a metonymic expansion process allows RHYTHMIC THEME (or FOOTSTEP via metaphorical mapping) to stand for PERSON IN MOTION (as shown in Fig. 12). This connection allows us to establish correspondences between musical progression and human motor behaviors. 3.3.3. Via musical scale Musical scale also contributes to deploy a program, although in a much more subtle way. Contrary to musical and rhythmic motives, harmonic relationships are not always directly apprehensible to lay hearers. However, there are a few [(Fig._1)TD$IG] 144 P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 Fig. 11. Double metonymy MUSICAL MOTIVE FOR MUSICAL THEME FOR CHARACTER (black simple arrows for metonymic mappings and mode-shift). musical examples that can be analyzed in this fashion. A rather illustrative use of musical scale to create a broader extramusical environment is to be found in ‘‘Laindenotte, Empress of the Pagodas’’, from Ravel’s Mother Goose (example 7): Ravel takes advantage of the pentatonic scale (in F#) to evoke an adequate Asian atmosphere to frame the story of Laindenotte. A pentatonic scale is a five notes per octave musical scale which provides the harmonic base for most of the Asian traditional instruments.11 In the example under scrutiny, Ravel leaves the traditional heptatonic system (at the basis of the rest of the pieces in Mother Goose) and productively exploits the centrality of the pentatonic scale to Asia in order to construct a vision of the non-Western world and its inhabitants. This strategy is acknowledged by scholars such as Locke (2011:217), who claims that the incorporation of ‘‘distinctive scales, harmonies, orchestral colors and other features that had previously been associated with exotic realms’’ are compositional techniques used to evoke an image of ‘‘exotic otherness’’ (Locke, 1991:261). More specifically, Locke (2011:228) identifies the pentatonic scale, especially focusing on the intervals of fourths and minor thirds, and rhythmic figures that are more impressions than melodic as the most recurrent compositional techniques to evoke the Far East12 (see Fig. 13). 3.3.4. Via musical key Let’s turn now our attention to the use of musical key in relation to metonymy. Consider the following extract (example 8) from Chrissochoidis and Huck’s (2011:65) analysis of Wagner’s Lohengrin: 11 To mention but a few, the major pentatonic scale is the basic scale of the music of China and the music of Mongolia; in turn, the Japanese shakuhachi flute plays a minor pentatonic scale. 12 Other well-known examples are Debussy’s Pagodes and Saint-Saëns’ Samson and Dalila. [(Fig._12)TD$IG] P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 145 Fig. 12. RHYTHMIC MOTIVE FOR PERSON (IN MOTION) multimodal metonymy in ‘‘In the Hall of the Mountain King’’ from Grieg’s Peer Gynt (black simple arrows for metonymic mappings and mode-shift). [(Fig._13)TD$IG] Fig. 13. Metonymy PENTATONIC SCALE FOR ASIA in ‘‘Laindenotte, Empress of the Pagodas’’, from Ravel’s Mother Goose (black simple arrow for metonymic mappings and mode-shift). (8) Choice of key also underscores the conjugal state of each couple, which rests on beliefs of trust and fidelity. Elsa’s A-flat major is the closest possible to Lohengrin’s A major yet stands the furthest away in the circle of fifths, the half-tone friction between the two tonal plateaus projecting Wagner’s core idea of the impossibility of a human--divine union. It becomes evident from this paragraph that the authors ascribe a metonymic relationship between each key and their corresponding opera character: A major stands for Lohengrin, whereas A-flat major stands for his beloved Elsa (Fig. 14). MUSICAL KEY FOR PERSON thus contributes to the proper identification of each of the characters involved in the scene. Although the attribution of a key to an opera character is a rather discretionary decision on the part of the composer, it makes the characters and their corresponding musical key become dependent on each other for the identification to take place. Curiously enough, this example could also be analyzed from the point of view of metaphorical correspondences. In this regard, the structural relationship between both musical keys (close in the pentagram but far away in the circle of fifths) is put in correspondence with the relationship between the lovers (emotionally close but physically distant) through the PEOPLE ARE MUSICAL KEYS metaphor. Wagner plays with this idea later on in the opera, when he makes Lohengrin’s theme modulate from his key (A Major) to Elsa’s key (A flat Major). This unison aria could reflect a physical union, once that Lohengrin has abandoned his divine status. The ambivalence of the MUSICAL KEY FOR PERSON and PEOPLE ARE MUSICAL KEYS metaphors to account for this example could obey to genre specificity. Given the length and complexity of Wagner’s musical work, the proper understanding of the extra-musical narrative could be accomplished in two steps: (1) a metonymic mapping facilitates the [(Fig._14)TD$IG] 146 P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 Fig. 14. MUSICAL KEY FOR PERSON metonymy in Wagner’s Lohengrin (black simple arrows for metonymic mappings and mode-shift). identification of the protagonists in terms of a key; and then (2) a metaphor helps to develop the psychological characterization of these protagonists and the events they take part in by means of the mapping of the structural relationship between keys onto the lovers’ relationship. 3.4. Multiple source-in-target metonymies The previous discussion of examples 5--8 has revealed that different elements of a certain musical theme (namely, musical motive, rhythmic theme, scale and key) can provide access via metonymic expansion to the most-encompassing domain. However, it may well happen that several of these musical elements are simultaneously activated through multimodal manifestations and metonymically projected onto the extra-musical narrative. Multiple source-in-target metonymies (Pérez Hernández, 2013, in application to illocutionary speech acts) challenge the traditional version of metonymy as a singular mapping from a single source subdomain onto the corresponding target domain for which it stands. One revealing case of study is that of sound logos. A sound logo is a short distinctive melody or sequence of sounds played at the beginning or ending of a commercial. Sound logos are increasingly becoming a vehicle to convey a memorable message to targeted consumers, taking advantage of the powerful memory sense of music. They often combine with textual and/or visual logos in order to reinforce the recognition of a brand. For the sake of illustration, consider the Nokia13 basic tune (example 9a). The most recent cellular phones display more sophisticated versions of the tune, including harmonic background or real instrument performance (e.g., by a piano, as in example 9b). The conglomerate nature of brands makes it possible for individual brand elements to stand for the brand either in isolation or in combination. In the case of Nokia, BASIC TUNE, PIANO TUNE, and LOGO (a combination of text and visuals, in terms of the use of a specific typographic font) are equally viable source domains to stand for the more encompassing domain BRAND (Fig. 15). The higher the number of brand elements (in this case, tunes, name and typographic traits) that get activated simultaneously, the easiest the process of brand recognition will be. In order to survey an equivalent conceptual mechanism in classical music, the notion of leitmotif arises as crucial. The technique is notably associated with the operas of Richard Wagner (although the notion is derived from Berlioz’s idée fixe, as mentioned in section 3.2). By way of reminder, a leitmotif is a musical term referring to a short, constantly recurring musical phrase associated with a particular person, place, or idea. The motive should be carefully designed in order to 13 Both the Nokia tune and logo (see Fig. 15) are copyright of Nokia Corporation and here used and reproduced with permission of Nokia. [(Fig._15)TD$IG] P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 147 Fig. 15. Multiple source-in target metonymy BASIC TUNE/PIANO TUNE/LOGO FOR BRAND (NOKIA) (black simple arrows for metonymic mappings and mode-shift). retain its identity even if it is modified on subsequent appearances in terms of rhythm, harmony, orchestration or accompaniment. In line with our previous analysis, a leitmotif (either based on melody, rhythm, harmony or a combination of the three) could be well studied from the perspective of the examples in section 3.3 as a metonymic source domain that grants access to the target domain for which it stands for, in the form of the metonymy LEITMOTIF FOR PERSON/EVENT/etc. Interestingly enough, there are also instances where one person or extra-musical event is expressed through a variable number of different leitmotifs. Cone (2009:80--81) identifies up to three different leitmotifs referring to Siegmund’s sword in The Ring of the Nibelungs. In example 10a, Wotan invokes for the first time the sword at the end of The Rhinegold: ‘‘So grüss ich die Burg, sicher vor Bang’ und Grau’n’’ (‘‘Thus I greet the castle, safe from fear and dread’’). Later on at the beginning of The Valkyries, this leitmotif is recovered in Sigmund’s vocal line as he reflects on the promised sword (example 10b): ‘‘Ein Schwert verhiess mir der Vater, ich fänd’ es in höchster Noth’’ (‘‘My father promised me a sword, which I would find in my greatest need’’). The Sword is brought up a third time at the end of the first act of The Valkyries when Sigmund pulls it out from Hunding’s house tree and names it ‘‘Nothung’’ (or ‘‘needful’’, given that this is the weapon that he needs for his forthcoming fight with Hunding) (example 10c). Notwithstanding obvious differences in terms of rhythm, melody and harmony, the fact that these three leitmotifs have in common (1) the falling octave and (2) the triad-based structure following the falling octave makes them eligible subdomains to stand for the more encompassing domain SWORD (Fig. 16). Interestingly enough, the Sword leitmotif harmonic structure (major triad arppegiation) is also present in the Rhine and Mother Earth leitmotifs, thereby serving as a cognitively economical point of access to these three wider domains. Similarly, the felicitous coincidence in the brass instrumentation of both Nothung and Rhine leitmotifs respectively points [(Fig._16)TD$IG] 148 P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 Fig. 16. Multiple source-in-target metonymy LEITMOTIF 1/LEITMOTIF 2/LEITMOTIF 3 FOR SWORD in Wagner’s The Ring of the Nibelungs (black simple arrows for metonymic mappings and mode-shift). toward the source of the metal with which the sword is forged and the nature of the Rhinegold, respectively. The greater context of the opera sheds light on the complex and intertwined connections between leitmotifs that, in the end, musically link the sword to its source and physical material. It is worth mentioning that these links are also of metonymic nature: MAJOR TRIAD ARPPEGIATION FOR NOTHUNG/RHINE/EARTH and BRASS FOR NOTHUNG/RHINEGOLD. 4. Conclusions The study of examples of program music reveals that the process of meaning (re)construction can be reduced to a finite set of patterns of conceptual integration and conceptual interaction. The major claim of this paper is that meaning (re) construction is a two-step process where the specific set of conceptual integration patterns structuring the multimodal manifestation triggers an array of cognitive operations at the conceptual level necessary for the inferential task to take place. In particular, I have focused on the patterns of conceptual disintegration that pave the way to the activation of metonymic reasoning (in isolation or in combination with other conceptual mechanisms). A combined approach to the study of meaning (re)construction proves efficient to cover in finer detail a series of deficiencies found in CBT and CMyT. For instance, it goes beyond CBT inasmuch as it provides a clear-cut distinction between the product (multimodal expression) and the process (cognitive operation) of disintegration. Furthermore, the taxonomy provided by the MCIM proves relevant to detect different instantiations of conceptual disintegration. As shown in Table 1, two variables have been offered to classify different types of ‘‘disintegrated’’ conceptual packages. According to the degree of integration, a difference has been established between cases of substitution, examples 1--4, and fragmentation, examples 5--10, which otherwise would have been considered equivalent examples of selective projection in CBT. The second variable is the kind of dependency between the represented input and the whole domain to which it belongs (absent in CBT). It has been claimed that it becomes crucial to further classify cases of conceptual disintegration: for instance, examples 1--4 would be considered equivalent from a CBT-oriented account since they all render instances of conceptual disintegration. However, it has been demonstrated that they differ in the relationship of dependency between the present input and the domain to which it refers: in examples 1 and 2, they share a relationship of inherent dependence that makes the identification easy and direct; by contrast, text and music become ad hoc dependent on each other in examples 3 and 4, thereby steering the hearer to rely on extra-contextual information to establish the connection with originally independent inputs. The two-stepwise approach to meaning (re)construction also advances our knowledge on musical metonymy. For instance, it helps to build an index of multimodal metonymies that share a common underlying reconstruction process (metonymic mappings) triggered by different multimodal products (patterns of disintegration). As a result, the following inventory of multimodal metonymies is offered (see Table 2). The application of the theoretical postulates of this paper to the non-verbal part of communication proves efficient in at least two additional respects. On one hand, it builds a stronger theory of conceptual integration and completes CMyT, since both have traditionally been restricted to the study of linguistic examples. Complementarily, it also advances the theory of multimodal metonymy, which so far has only attracted scarce attention, and exclusively within the visual realm (cf. Forceville, 2009b). P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 Table 2 Summary of the types of metonymy discussed in this paper. Conceptual operation Multimodal metonymic echoing Domains and mappings Subdomain: a’ (a musical item that imitates or echoes a sound from the real world, a). Matrix domain: A (the real producer of the sound). The metonymic subdomain and the matrix domain are partially or totally rendered in different modes. Multimodal metaphtony my Subdomain: a (a central feature of a real world element). Matrix Domain: A (the element). This matrix domain becomes the source of a metaphor used to think and reason about the musical piece, which is the target. The metonymic subdomain, the matrix domain/metaphoric source and the metaphoric target are partially or totally rendered in different modes. Multimodal metonymic cueing The metonymic subdomain and the matrix domain are partially or totally rendered in different modes. Multimodal multiple source-in target metonymy a echoes A B 1 a a: constant and regular motion B: understanding of *(black arrow for metonymic mapping and mode-shift; white arrow for metaphoric mapping and mode-shift) as a spinning wheel a: the beloved A a’: a cues a’ A: Life of the artist (Hector Berlioz) *(black arrow for metonymic mapping and mode-shift; interrupted-line white arrow for cueing operation) A a: a A: Papageno *(black arrow for metonymic mapping and mode-shift) 1 Matrix Domain: A (the character the musical element represents). A: Birds A: spinning wheel Matrix Domain: A (the character the musical element represents). Subdomain : a (a musical element) 2 2 Subdomain : a (a musical element) 3 3 Subdomain : a (a musical element) a’: a’ *(black arrow for metonymic mapping and mode-shift; interrupted-line white arrow for echoic operation) Subdomain: a (a musical element). The metonymic subdomain and the matrix domain are partially or totally rendered in different modes. Example a: chirp of birds A Subdomain: a’ (a musical item that cues a central feature of a real world element). Matrix Domain: A (the element). Multimodal source-in target metonymy Schematic representation* 1 a: a1 2 a: a 2 A 3 a: a3 A: Siegmund’s Sword The metonymic subdomains and the matrix domain are partially or totally rendered in different modes. *(black arrows for metonymic mappings and mode-shift) 149 150 P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 Further research should test the compatibility of this approach to multimodal meaning (re)construction practices in linguistic environments (in a reverse application of the equipollence hypothesis). Additionally, other strategies of meaning making should be investigated (such as that wherein patterns of conceptual integration trigger metaphoric mappings, as hinted in Author, 2014), in order to verify whether the two-stepwise model is a feasible and efficient approach to every kind of meaning construction operations. Acknowledgements I am deeply indebted to Lorena Pérez Hernández and an anonymous reviewer for their insightful critical remarks, which have notably helped me to improve this manuscript at different stages. The responsibility for any remaining mistakes is solely mine. This research has been carried out with the support of a FPU grant AP2010-5172 (Spanish Ministry of Education). I am also grateful to Nokia Corporation for granting permission to reproduce their tune and logo. References Bache, Carl, 2005. Constraining conceptual integration theory: levels of blending and disintegration. J. Pragmat. 37 (10), 1615--1635. Barcelona, Antonio (Ed.), 2000. Metaphor and Metonymy at the Crossroads: A Cognitive Perspective. Mouton de Gruyter, Berlin/New York. Barcelona, Antonio, 2011. Reviewing the properties and prototype structure of metonymy. In: Barcelona, A., Benczes, R., Ruiz de Mendoza, F.J. (Eds.), Defining Metonymy in Cognitive Linguistics: Towards a Consensus View. John Benjamins, Amsterdam & Philadelphia, pp. 7--58. Barthes, Roland, 1965. Elements de sémiologie. Gonthier, Paris. Berlioz, Hector, 1845. Program for the Symphonie Fantastique. http://www.hberlioz.com/Scores/fantas.htm (Retrieved 09.05.14). Blechner, Mark J., 2001. The Dream Frontier. The Analytic Press, Hillsdale, NJ. Bloom, Harold, 1973. The Anxiety of Influence: A Theory of Poetry. Oxford University Press, New York. Brower, Candace, 2000. A Cognitive Theory of Musical Meaning. J. Music Theory 44 (2), 323--379. Chrissochoidis, Ilias, Huck, Steffen, 2011. Elsa’s reason: on beliefs and motives in Wagner’s Lohengrin. Cambr. Opera J. 22 (1), 65--91. Cone, Edward, 2009. Sigfried at the Dragon’s cave: the motivic language of The Ring. In: Morgan, R.P. (Ed.), Hearing and Knowing Music: The Unpublished Essays of Edward T. Cone. Princeton UP, Princeton. Cook, Nicholas, 1992. Music, Imagination and Culture. Oxford University Press, Oxford. Coulson, Seana, 1996. The Menendez Brothers Virus: Analogical Mapping in Blended Spaces. In: Goldberg, A. (Ed.), Conceptual Structure, Discourse, and Language. CSLI, Palo Alto, CA, pp. 67--81. Cox, Arnie, 2011. Embodying Music: Principles of the Mimetic Hypothesis. J. Soc. Musical Theory 17 (2). http://www.mtosmt.org/issues/mto. 11.17.2/mto.11.17.2.cox.pdf (Retrieved 09.05.14). Fauconnier, Gilles, Turner, Mark, 1998. Conceptual integration networks. Cogn. Sci. 22 (2), 133--187. Fauconnier, Gilles, Turner, Mark, 2002. The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities. Basic Books, New York. Forceville, Charles, 2009a. The role of non-verbal metaphor sound and music in multimodal metaphor. In: Forceville, C., Urios-Aparisi, E. (Eds.), Multimodal Metaphor. Mouton de Gruyter, Berlin/New York, pp. 383--400. Forceville, Charles, 2009b. Metonymy in visual and audiovisual discourse. In: Ventola, E., Moya, A.J. (Eds.), The World Told and the World Shown: Multisemiotic Issues. Palgrave-McMillan, London, pp. 56--74. Gibbs, Raymond W., 2000. Making good psychology out of blending theory. Cogn. Linguist. 11, 347--358. Goossens, Louis, 1990. Metaphtonymy: the interaction of metaphor and metonymy in expressions for linguistic action. Cogn. Linguist. 1, 323--340. Grady, Joseph, 1999. A typology of motivation for conceptual metaphor. Correlation vs. resemblance. In: Gibbs, R.W., Steen, G.J. (Eds.), Metaphor in Cognitive Linguistics. John Benjamins, Amsterdam/Philadelphia, pp. 79--100. Hidalgo, Laura, Kraljevic, Blanca, 2011. Multimodal metonymy and metaphor as complex discourse resources for creativity in ICT advertising discourse. In: Gonzálvez García, F., Peña Cervel, S., Pérez Hernández, L. (Eds.), Metaphor and Metonymy Revisited Beyond the Contemporary Theory of Metaphor, vol. 4. John Benjamins, Amsterdam/Philadelphia, pp. 153--178. Hougaard, Anders, 2002. Conceptual Disintegration as a Principle of Meaning Construction. PEO (Pre-Publications of the English Department of Odense University). Hougaard, Anders, 2005. Conceptual disintegration and blending in interactional sequences. J. Pragmat. 37 (10), 1653--1685. Lakoff, George, Johnson, Mark, 1980. Metaphors We Live By. University of Chicago Press, Chicago. Lakoff, George, Turner, Mark, 1989. More than Cool Reason: A Field Guide to Poetic Metaphor. Chicago University Press, Chicago. Larson, Steve, 2012. Musical Forces: Motion, Metaphor, and Meaning in Music. Indiana University Press, Bloomington. Larson, Steve, Johnson, Mark, 2004. Architectural metaphors in music discourse and music experience. Yearbook of Comparative and General Literature Vol. 50. Mutability. Architecture, Music and the Chicago School, University of Indiana Press, Bloomington, pp. 141--154. Locke, Ralph, 1991. Constructing the oriental ‘Other’: Saint-Saens’s ‘‘Samson et Dalila’’. Cambr. Opera J. 3 (3), 261--302. Locke, Ralph P., 2011. Musical Exoticism: Images and Reflections. Cambridge University Press, New York. Mailman, Joshua, 2009. An imagined drama of competitive opposition in Carter’s Scrivo in Vento. Music Anal. 28 (2--3), 373--421. Mailman, Joshua, 2012. Agency, determinism, focal time frames, and narrative in processive minimalist music. In: Klein, M., Reyland, M. (Eds.), Music and Narrative Since 1900. Indiana University Press, Indiana (Chapter 6). Mairal, Ricardo, Ruiz de Mendoza, Francisco J., 2009. Levels of description and explanation in meaning construction. In: Butler, C.S., Martín, J. (Eds.), Deconstructing Constructions. John Benjamins, Amsterdam/Philadelphia, pp. 153--198. Meyer, Leonard, 1956. Emotion and Meaning in Music. University of Chicago Press, Chicago. Mordden, Ethan, 1980. A Guide to Orchestral Music: The Handbook for Non-Musicians. Oxford University Press, Oxford. P. Pérez-Sobrino / Journal of Pragmatics 70 (2014) 130--151 151 Oakley, Todd, 1998. Conceptual blending, narrative discourse, and rhetoric. Cogn. Linguist. 9 (4), 320--360. Pérez Hernández, Lorena, 2013. Illocutionary constructions: (multiple source)-in-target metonymies, illocutionary ICMs, and specification links. Lang. Commun. 33, 128--149. Pérez-Sobrino, Paula, 2014. Multimodal cognitive operations in classical music. VIAL 11, 137--168. Ritchie, David, 2004. Lost in space: metaphors in conceptual integration theory. Metaphor Symbol 19, 31--50. Ruiz de Mendoza, Francisco J., 1998. On the nature of blending as a cognitive phenomenon. J. Pragmat. 30 (3), 259--274. Ruiz de Mendoza, Francisco J., 2000. The role of mappings and domain in understanding metonymy. In: Barcelona, A. (Ed.), Metaphor and Metonymy at the Crossroads. Mouton de Gruyter, Berlin/New York, pp. 109--132. Ruiz de Mendoza, Francisco J., 2007. High level cognitive models: in search of a unified framework for inferential and grammatical behavior. In: Kosecki, K. (Ed.), Perspectives on Metonymy. Peter Lang, Frankfurt/Main, pp. 11--30. Ruiz de Mendoza, Francisco J., 2011. Metonymy and cognitive operations. In: Benczes, R., Barcelona, A., Ruiz de Mendoza Ibáñez, F.J. (Eds.), What is Metonymy? An Attempt at Building a Consensus View of the Delimitation of the Metonymy in Cognitive Linguistics. John Benjamins, Amsterdam/Philadelphia, pp. 103--124. Ruiz de Mendoza, Francisco J., 2013. Linking Constructions into Functional Linguistics: The Role of Constructions in RRG Grammars. In: Nolan, B., Diedrichsen, E. (Eds.), Studies in Language Series. John Benjamins, Amsterdam/Philadelphia (vol. in preparation). Ruiz de Mendoza, Francisco J., Galera, Alicia, 2014. Cognitive Modeling. John Benjamins, Amsterdam/Philadelphia. Ruiz de Mendoza, Francisco J., Mairal, Ricardo, 2011. Constraints on syntactic alternation: lexical-constructional subsumption in the LexicalConstructional Model. In: Guerrero, P. (Ed.), Morphosyntactic Alternations in English. Functional and Cognitive Perspectives. Equinox, London, UK/Oakville, CT, pp. 62--82. Spitzer, Michael, 2008. A metaphoric model of Sonata form: two expositions by Mozart. In: Mirka, D., Agawu, K. (Eds.), Communication in Eighteenth-Century Music. Cambridge University Press, Cambridge. Turner, Mark, 2001. Cognitive Dimensions of Social Science: The Way We Think About Politics, Economics, Law, and Society. Oxford University Press, Oxford. Uriós-Aparisi, E., 2009. Interaction of multimodal metaphor and metonymy in TV commercials: four case studies. In: Forceville, C., Uriós-Aparisi, E. (Eds.), Multimodal Metaphor. Mouton de Gruyter, Berlin/New York, pp. 95--118. Wimsatt, William Kurtz, 1954. The Verbal Icon: Studies in the Meaning of Poetry. University of Kentucky Press, Kentucky. Zbikowski, Lawrence, 2002. Conceptualizing Music: Cognitive Structure, Theory, and Analysis. AMS Studies in Music. Oxford University Press, New York. Secondary references Example 1. Vivaldi, Antonio, 1725. Le quattre stagioni (The four seasons): Spring, 1st mov. Example 2. Schubert, Franz, 1814. Gretchen am Spinnrade (Gretchen by the Spinning-Wheel), D.118 (Op.2). Example 3. Berlioz, Hector, 1830. Symphonie fantastique (Fantastic Symphony). Example 4. Piaf, Edith. 1956. Non, je ne regrette rien (‘‘No, I regret nothing’’). Example 5. Mozart, Wolfgang Amadeus. 1791. Die Zauberflötte K. 620 (The Magic Flute): ‘‘Der Vogelfänger in ich ja’’ (‘‘I am the bird-catcher’’) & ‘‘Wie stark ist nicht dein Zauberton’’ (‘‘How strong is thy magic tone’’). Example 6. Grieg, Edvard, 1888. Peer Gynt: ‘‘I Dovregubbens hall’’ (‘‘In the Hall of the Mountain King’’). Example 7. Ravel, Maurice, 1910. Ma Mère l’Oye (Mother Goose) M.60: ‘‘Laindenotte, Imperatrice des Pagodes’’ (Laindenotte, Empress of the Pagodas). Example 8. Wagner, Richard, 1848. Lohengrin WWV75. Example 9. Nokia tune. Based on Francisco Tárrega’s 1902 ‘‘Grand Vals’’. Example 10. Wagner, Richard, 1874. Der Ring des Nibelungen (The Ring of the Nibelungs) WWV86. Paula Pérez-Sobrino is currently a PhD student at the University of La Rioja (Spain), where she is about to finish her doctoral thesis with the support of and FPU scholarship (Ministry of Education, Spain). Her main field of interest is multimodal metaphor across different genres, such as printed advertising, commercials, music and cinema. She is also concerned with the study of Conceptual Theory of Metonymy and other cognitive operations involving text, music and sound.