Film analysis, and multimodal analysis in general, is no longer seen as simply decoding the various semiotic resources involved in the meaning construction, but as asking for inferential processes of reasoning about the best and most plausible interpretation. This interpretation operates according to so-called textual cues, or, in other words, “instructing elements” (cf. Van den Hoven 2012) in the artefact that clearly guide the recipient’s imagination and hypothesis-making (cf. Wildfeuer 2014; Van den Hoven and Yang 2013; Bordwell 1989). This view of film as raising questions and challenges about its content and demanding the active and complex involvement and participation of the recipient for the resolution of these questions is more and more frequently seen from an argumentative perspective in which the cues and elements are described as part of the film’s rhetorical discourse structure. This structure gives clear constraints on the interpretation and at the same time carries out a logical evaluation of the film’s meaning (cf. Alcolea-Banegas 2009). Film thus not only informs the spectator about events, actions, circumstances, and further details of its diegetic world, but also conveys beliefs and justifications of reconstructing these details.
This paper focuses on questions concerning exactly this process of making meaning out of the filmic text by asking for the argumentative patterns that enable the recipient’s inference processes during his/her interpretation. Film as a multimodal discourse, i.e., an interplay of various auditory and visual semiotic resources that create meaning due to their intersemiosis (cf. Wildfeuer 2012a), does not make its content as explicit as verbal language does, but nevertheless conveys reasonable arguments, as van den Hoven and Yang (2013) point out in the context of approaches dealing with visual argumentation:
We can summarize this argumentative approach as the assumption that the rhetor is oriented towards argumentative reasonableness … The meaning of this interpretative principle is that we assume the rhetor to be aware of the fact that his audience expects him to act in a reasonable manner. Therefore all elements in all modes of the discourse can be understood by the analyst as an attempt by the rhetor to keep up the impression that he meets this expectation … (Van den Hoven and Yang 2013: 407)
Argumentation is thus not seen as an internal property of the multimodal text itself, but as the process of inferring the most plausible interpretation of audiovisually expressed standpoints according to rhetorical patterns in the text (cf. Van den Hoven and Yang 2013). This directly corresponds to the basic acceptance that meaning generally depends on interrelations between the text itself and the recipient and his/her interpretation (cf., e.g., Iser 1980). Claims or arguments expressed by multimodal texts cannot, however, be described as identical or equivalent to those expressed verbally (cf. Birdsell and Groarke 1996), since they often lack the explicit mediation of their semantic content. Multimodal texts are therefore “far removed from the ‘standard’ propositional format of an argument” (Van den Hoven and Yang 2013) and their interpretations are always defeasible and abductive (cf. CP). As a consequence, the question of how visual and audiovisual documents argue and mediate purposefully has become a recent topic in argumentation theory and rhetoric. In particular, the question of how their interpretation can be further manifested by a detailed description of the interplay of the various semiotic resources and their meaning making potential is a matter of particular interest, especially with regard to the description of arguments within or evaluations of the textual content.
In this paper, we will present an analytical approach based on recent advancements in contemporary discourse semantics and multimodal discourse analysis that will help us outline the discursive and rhetorical structure of filmic text and to retrace the inference process of the recipient in further detail. This inference process is essentially defeasible and goes back to the notion of abduction introduced by Peirce as a basic logical form (cf. CP; Wirth 2000, 2005). Abduction helps in finding hypotheses and at the same time proving and verifying those hypotheses on the basis of certain logical principles. The logical background which comes into play here and which we will explain in further detail below delivers the necessary foundation to describe the recipient’s inference and interpretation process comprehensively. It has already been successfully applied both to the analysis of verbal (cf. Asher and Lascarides 2003; Hobbs 2003) as well as filmic discourse (cf. Wildfeuer 2014) and has made it possible, particularly for the latter, to substantiate the often required “logic of mapping and modeling that underlies interpretive problem-solving” (Bordwell 1989: 202), which has been described extensively within, for example, the Bordwellian film interpretation framework, but which has so far not been analyzed in further detail.
As a consequence, we will refer back to this logical foundation in further detail in the following. We think that the application of logics to the analysis of film will make it possible to gain further information about how filmic text mediates argumentative patterns and thereby guides the recipient’s interpretation. The aim is to show how multimodal film leads its spectators to acknowledge the argumentative reconstruction of its content by relating the diegetic world to its reality and proving its validity (cf. Van den Hoven and Yang 2013; Van den Hoven 2012). We will therefore pursue this analysis mainly with the help of the so-called logic of film discourse interpretation (cf. Wildfeuer 2014). This framework allows working out so-called logical forms of the discourse that make visible how the semantic content of filmic narrative events can be analyzed. In a second step, it helps to construct the film’s discursive and rhetorical structure by describing discourse relations holding between the narrative events and maintaining the film’s coherence. With the description of this structure, it is then possible to understand how the recipient works through the text and how his/her inference processes are constrained by the various relations which work as the respective textual cues (see above). We think that this offers a more explicit and comprehensive analysis of the argumentative patterns in filmic discourse and their way of influencing the recipient during his/her interpretation.
In the following, thus, we will first give a short overview of how visual argumentation and multimodal analysis as well as a logical basis of film analysis can be brought together effectively and with regard to promising results of a further to be expanded analysis of an example filmic discourse, the short film El Vendedor de Humo (Maestro 2012). Based on a short description of the logical framework and its operating mechanisms, we will then illustrate its work with a more detailed analysis of the film, which features interesting patterns of guiding the spectator to its final interpretation. Concluding remarks will summarize the approach and its benefits and will make further suggestions for an enhancement of the multimodal and logical analysis of filmic argumentation.
2 Visual and filmic argumentation
The interpretation of visual and audiovisual artefacts as argumentative patterns goes back to a general theory of the rhetoric of images which sees these artefacts as activating their recipients’ perception and comprehension by offering deliberately produced opportunities of making meaning (cf. Knape 2005). The same principles can be transcribed to filmic text as both a visual as well as auditory artefact equipped with rhetorical techniques to persuade the recipient. This claim and a more general one concerning all forms of multimodal discourse have been pursued recently by van den Hoven (2011, 2012) as well as Alcolea-Banegas (2009) building upon the debate about “visual argumentation” held by Birdsell and Groarke (1996), Blair (1996), and others (see Van den Hoven and Yang 2013 for further references). In the following, we will give a short overview of this theoretical background and its main hypotheses, which we take as a starting point for a joint approach of film analysis based on multimodal as well as rhetorical and logical principles.
2.1 From images to moving images: Do multimodal discourses argue?
The idea of visual argumentation is, according to Blair (1996), a traceable consequence resulting from a general implication of the concept of arguments, which says that “arguments1 are not necessarily linguistic or verbal arguments. All that is required … for something to qualify as an arguments1 is that reasons be overtly expressed, and that reasons and claim be linguistically explicable” (Blair 1996: 25). 1 Blair therefore defines visual arguments “to be understood as propositional arguments in which the propositions and their argumentative function and roles are expressed visually, for example by paintings and drawings, photographs, sculpture, film or video images, cartoons, animations, or computer-designed visuals” (Blair 1996: 26).
Similarly, Birdsell and Groarke (2006: 103) understand visual arguments as “arguments (in the traditional premise and conclusion sense) which are conveyed in images” and which can mainly be understood with regard to the context in which they are embedded.
These and other approaches to the concept of visual argumentation indeed still question the possibility and manner of expressing arguments visually, but also try to strengthen the complex relation between the two notions of “argument” and “visual” by giving a number of convincing examples (cf. Roque 2012). However, particularly their descriptions of how the visual expression in these examples de facto takes place and can be interpreted are not very detailed and their argumentation often stays uncertain and hypothetic. Neither do they offer a comprehensive linguistic or formal analysis, as required for the general qualification of arguments (cf. Blair 1996; see above). Blair, for example, highlights:
Just how visual images and visual forms in general can and do communicate propositions, just how the important ancillary concept of context is to be understood and how in practice context is to be interpreted and combined with the visual, and just how text and visuals (and sound) interact to produce meaning are all questions which strike me as important, difficult and unanswered … (Blair 1996: 39)
For film, these questions are even more complex with regard to the general interplay of various semiotic resources that together bear a meaning potential to be actualized by the recipient (cf. O’Halloran 2004). Van den Hoven and Yang (2013) therefore underline that multimodal discourse rises a challenge for argumentation theory in that particular these intermodal interactions must be taken into account. We give a first illustration of this challenge and the accompanying necessary principles of interpretation with the following example of the short film El Vendedor de Humo whose main shots are shown in Figure 1.
The film’s story is about a character named the smoke seller (Spanish: el vendedor de humo) who comes to a small village and fascinates the residents with his ability to transform poor and wearisome belongings into good-looking, brilliant things, such as the young boy’s little branch, which is transformed into a colorful toy plane (see images 3 and 4 in line 2 of Figure 1) or the old lady’s clothes, which suddenly become luscious and let her look younger and energetically (images 5 and 6). The magical turning is always accompanied by a purple smoke surrounding the objects during their transformation as well as swinging and lively music from a record player. This music accompanies the whole film, although later, its source is no longer visible and it could thus also be interpreted as non-diegetic, in particular when its rhythm and expressivity change. During the course of the film, the smoke seller is hailed as a hero and brought to the mayor of the village who owns a precious jewel drawing the smoke seller’s attention. Suddenly, he recognizes the precursors of a thunderstorm and starts gathering his equipment and leaving the village. When the rain starts, he is already back on his cart, holding the mayor’s jewel in his hands. In the background, all habitants and objects begin re-transforming and, finally, the smoke smeller is shown as an old, seedy man with a malicious grin in his face.
There is indeed an argumentative structure to be found in the film, showing the story of a mean and sneaky man who is fooling other people by slightly working out his hypocritical behavior as nothing but smoke and mirrors. This structure can be examined not only by reconstructing the film’s diegesis, but also by taking a closer look into its multimodal meaning-making strategies. The film’s narration, for instance, does not include any verbal information apart from the title at the beginning and the credits at the end. Neither the characters speak nor are there any inserts giving further information about setting or time, for example. Every communication between the inhabitants and the smoke seller is realized by gestures and facial expressions; all details of the story are conveyed visually by images, various camera perspectives, motion, etc. It is thus the interplay of the different semiotic resources, their intersemiosis, which carries out the film’s story and provides, at the same time, certain cues to understand what is going on.
It becomes, for example, clear during the reception that the smoke seller’s transformations do not last forever and that some external circumstances determine their durability. A first cue for this is the young boy’s reaction to the re-transformation of his toy plane after falling into a vat. Although this process of plunging into water is only a small detail during a montage of images showing the smoke seller to be celebrated by the inhabitants, a close-up shot of the angry boy and his re-transformed branch shown a few seconds later can be understood as an indication of these circumstances (see Figure 2).
Nevertheless, the recipient has to bring together the two shots and their narrative events and infer a coherent relationship between them that maintains a certain causality between the event of falling into the water and that of being re-transformed into its original state. This interpretation and the recognition of further cues which show similar circumstances or causalities guide the recipient into a concrete direction, namely, the final interpretation of the smoke seller’s behavior and appearance as insidious and hypocritical. Consequently, these cues can be seen as modifying the recipient’s belief about the truthfulness and verifiability of the transformations and thus as parts of the film’s argumentative structure. They help to evaluate the events and details of the filmic story and to verify or withdraw hypotheses about the content in order to come to its most plausible interpretation.
According to Van den Hoven and Yang (2013: 409), these cues indicate that “the rhetor attempts to reinforce or alter the way an audience perceives its reality.” As mimetic and diegetic relations, they explicitly “account for the rhetor’s pragmatic intention, which is the change that the rhetor tries to establish in the audience’s perception of its reality.” By leading the recipient to a critical evaluation of the transformations shown as magical incidences before, the short example already illustrated this change in the recipient’s perception of the whole story convincingly. It is therefore intelligible to describe the film as an argumentative discourse.
Following Alcolea-Banegas, we define filmic arguments as “the rhetorical and pragmatic effect[s] of the audiovisual discourse. In its action, this discourse reveals the consistency of rational argument and the efficiency of persuasive force” (2009: 270). However, we do not see these arguments as explicitly propositional arguments with the proposition expressed visually, as defined by Blair (1996; see above). Although we think that paintings, drawings, films, etc., have a general ability of conveying meaning, which we will describe as the film’s semantic content below, their argumentative patterns as shown in the example above become clear only by their discursive structure and the textual cues to be analyzed as prompting the recipient’s inferences. For our concrete example, this means that the interpretation of the narrative event of falling into a vat does not yet include a persuasive structure or argumentative pattern that directly causes a change of the recipient’s perception. Instead, it is the combination of this event with a second (or third) event of the diegesis, that of being re-transformed into a branch, for example, which generates a retraceable comprehension and argumentation structure.
Consequently, analyzing the rhetorical and argumentative structure of multimodal filmic discourse is always a matter of relational meaning-making in terms of explicit relations between events of the filmic story for which a particular rhetorical function can be identified. In the example above, this function is to causally explain the circumstances of the story and to instruct the recipient to draw respective inferences on this. Further rhetorical functions can for example be the conveyance of spatio-temporal circumstances or the comparison of different states or events. In all cases, these rhetorical functions serve as prompting mechanisms to execute specific inferential activities and to construct the film’s general coherence.
These operating patterns and cues have to be examined individually and within the specific filmic context in which they are embedded. We thus conclude that filmic discourse generally has the ability to produce an argumentative structure; its rhetorical functions, however, can only be analyzed with regard to their respective capacity of guiding and persuading the recipient in his/her interpretation.
In the following, we will exemplify how this examination can be pursued effectively with a framework of film interpretation which provides a general approach to the analysis of film discourse relations both from a multimodal as well as rhetorical perspective.
2.2 With logics from A to B
Van den Hoven and Yang 2013 describe the above defined arguments in filmic texts and their effects on the interpretation as a rhetorical situation which relates the multimodal meaning-making strategies of filmic discourse to its argumentative reconstruction (cf. Van den Hoven and Yang 2013). In order to analyze this situation and the recipient’s interpretive acts, they suggest a rhetorical perspective which mainly helps conveying the standpoint and argumentations of the rhetor.
A rhetorical analysis of meaning-making properties generally examines how the interpretation of images and films is determined, for example, by the producer during the production process. Questions of specific decisions during the production as well as the selection of technologies or creative means and their combination are a particular matter of interest here. Furthermore and according to Aristotle, rhetorical competence allows finding persuasive and faith-awakening patterns (the so-called pithanon) in the artefact that cause effects and expectations of the recipient. Examining the rhetorical factor of an image or film can consequently also be seen as analyzing the structure and provision of their interpretation opportunities in terms of their persuasive and argumentative power (cf. Knape 2005), not only focusing on the producer/rhetor, but rather on his/her argumentative reasonableness: “Rhetorical analysis is useful when bringing to light the subtle movements inside argumentative texts, and it transforms itself into a necessary instrument of logical reconstruction. In other words, rhetoric allows us to see which arguments are being constructed and which symbolic elements used” (Alcolea-Banegas 2009: 265).
Van den Hoven and Yang here explicitly highlight that “[t]his should not be misunderstood as being equivalent to the assumption that the rhetor intends to act in a reasonable manner” (2013: 407). Instead, the notions of coherence and relevance are taken into account as those principles which help the recipient to justify his/her reconstruction of the film’s story. Filmic hypothetical meaning making is only possible on the basis of general world and context knowledge as well as with the help of inferences drawn from logical conclusions out of the content. It is, for example, not explicitly shown by the filmic images in El Vendedor de Humo that the process of falling into the water directly leads to the re-transformation of the objects. Instead, the re-transformed branch is interpreted as a result of an event that is not shown within the film; it is the recipient’s interpretation based on a general knowledge that water may sometimes influence the state and condition of some things, which leads to the inference of a causal relationship here. By inferring this relation between the events, these discourse segments become relevant and thus a distinct part of the film’s argumentative structure.
The connection between coherence and relevance as a basic discursive principle has for example been described by Sperber and Wilson (1986) as well as by Asher and Lascarides (2003). In particular, the latter point out that the notion of relevance can be related to coherence in that discourse segments are relevant only if they are connected somehow to other discourse segments (cf. Asher and Lascarides 2003). Both therefore claim “general-purpose inference rules, which apply to any conceptual represented information” (Sperber and Wilson 1986: 176). Consequently, understanding and interpreting a film and its argumentation is a process of both relational meaning-making and abductive reasoning referring to the recipient’s cognitive capacity on the one hand and the contextual circumstances of the artefact on the other (cf. Bordwell 1989; Bateman and Schmidt 2011; Wildfeuer 2014). Van den Hoven and Yang resume this principle for the interpretation of multimodal discourse by pointing out: “[I]f an element conveyed by whatever mode has a strong relevance (conveys indispensible new meaning) as a coherent element in an argumentative reconstruction, then this (element of) the argumentative reconstruction gains a strong validity (and by adding up the argumentative reconstruction as a whole)” (2013: 408).
Hence, a first task during the analysis is to work out which elements are relevant for the interpretation of the discourse, i.e., which elements can be related to each other meaningfully. The rhetorical and argumentative analysis of film (and other multimodal discourses) must therefore directly connect the perspective of the producer as the rhetor conveying argumentative commitments with the perspective of the recipient and his/her interpretation. The interface between these two perspectives is, as we think, the artefact itself which provides the respective relationships within its discourse structure. By retracing these relations as the combining elements of the film’s story, we can then gain more information about the recipient’s activity of interpreting the text and constructing its narrative. 2
The close interrelationships between inferential meaning-making and the construction of filmic narrative have been examined in detail by David Bordwell and colleagues (cf. Bordwell 1985, 1989, 2006, 2008, Bordwell et al. 1985; Thompson 1988), mainly focusing on the basic Russian formalist’s distinction of fabula and syuzhet on the one hand and the patterns of a general narrative logic on the other hand. This narrative logic also explores relations among events in a narrative text, which are, following Bordwell and as illustrated in our example above, primarily causal or feature more abstract, comparative principles. Similarly, contemporary approaches to the notion of narrative see this principle as an explanation of the “different blends of states, actions, and events, different proportions of stereotypic and nonstereotypic knowledge, different strategies for distributing participant roles among individuals and entities in the storyworld, and so on” (Herman 2002: 22–23).
Herman furthermore points out that stories constitute their own logic which must be seen as “an unreplaceable resource for structuring and comprehending experience, a distinctive way of coming to terms with time, process, change” (Herman 2002: 23). Hence, narrative – or, as it is also described, story logic – is the way recipients understand the world around them and make sense of the claims and beliefs conveyed.
This general interconnection of logic and narrative dates back to the classical era of Greek culture, when representations of narrative needed a traceable reconstruction of their probability by means of a rhetorical analysis (cf. Doxiadis 2010). Even then and still today, logic serves as the combining element and necessary amplification of approaching narrative discourse. In contemporary approaches to the analysis of verbal discourse, for instance, this has led to the development of formal frameworks for the description of so-called logical forms that conceptualize in more detail the problems of interpretation of these discourses (cf. Hobbs et al. 1993; Kamp and Reyle 1993; Asher and Lascarides 2003; Hobbs 2003). Asher and Lascarides, for example, provide a set of discourse relations that are supposed to hold between discourse segments in a verbal text because of the truth conditional effects they have on the discourse. Describing these effects and the corresponding relations follows the basic notion of coherence in discourse, formulated as the principle of Maximize Discourse Coherence (MDC): “The logical form for a discourse is always a logical form that’s maximal in the partial order of the possible interpretations …” (Asher and Lascarides 2003: 21).
A compliance of this principle consequently leads to the inference of the maximally preferred discourse relation between two entities in a discourse. For this inference process, Asher and Lascarides provide clear meaning postulates and default axioms that are formulated not only in formal expressions, but also according to basic logical principles of inferences and abductive reasoning, such as Defeasible Modus Ponens or the Penguin Principle. 3 In order to find the respective and maximally preferred discourse relations, these conditions have to be fulfilled within and by the context of the discourse and the recipient’s knowledge. A successful description of the relations between events in a text based on the formal conditions and going back to these fundamental logical principles, delivers, as a consequence, an explicit description of the frequently labeled “constraints” guiding the recipient through his/her interpretation.
In previous work (Wildfeuer 2012a, 2012b, 2014), we have therefore extended the available frameworks for verbal language in order to apply them to filmic discourses and to approach these narratives texts from both a rhetorical as well as multimodal meaning-making perspective. Taking into account the principles of abduction and inferential reasoning, while also including important information from the context and the recipient’s world knowledge and by asking for the film’s plausibility in terms of coherence and structure, the logic of film discourse interpretation allows reconstructing the recipient’s interpretation process both with regard to the distinct analysis of the narrative events out of the intersemiotic interplay of the semiotic resources as well as the discourse relations combining these events into a coherent discourse.
In the following, we will show how this framework helps to work out the argumentative patterns in El Vendedor de Humo by first giving a short overview of the analytical tools of the framework. A detailed analysis of the short film will then describe (1) the general semantic content which can be interpreted out of the multimodal discourse, and (2) further examples of concrete textual cues within the film.
3 A logical and rhetorical analysis of filmic discourse
Argumentation in film as a multimodal discourse, as we illustrated above, can be made visible on the basis of the meaning-making strategies of coherence and structure of the filmic text which, at the same time, elucidate the so-called textual cues that guide the recipient through his/her interpretation. With the help of the shortly to be introduced framework of the logic of film discourse interpretation, we will now illustrate this multimodal and rhetorical examination of the textual cues in more detail and with further examples from the film we already used in Section 2.
3.1 Filmic meaning construction and the logic of film discourse interpretation
As a tool for the analysis of dynamically unfolding discourses, the logic of film discourse interpretation (cf. Wildfeuer 2014) combines a cognitive view of the recipient’s comprehension activity with a linguistic examination of the film’s textual contribution. It thereby provides two individual logics for the interpretation: (1) the logic of information content, which reasons about the discursive’s meaning potential to be inferred by the recipient as narrative events of the film’s story, and (2) the logic of constructing the logical form of this discourse and thereby inferring discourse relations between the events. We will show examples for both parts of the analytical framework, but the main focus will lie on the description of the relations as cues for the argumentative structure of the film. For a comprehensive overview of the whole framework, its theoretical background and further example analyses, see Wildfeuer (2012a, 2012b, 2014).
The logic of information content provides a formal language for the description of the logical forms of filmic discourse as semantic representations of the story. In this part of the analysis, each event which is interpretable both from the interplay of the various semiotic resources as well as from the recipient’s general knowledge about the world and the filmic context is illustrated in a box, as given in the example logical forms in Figure 3.
These boxes display the events as so-called eventualities which can normally be inferred by the recipient, but which can also be verbalised with other predicates. The different lines in the box list the discourse referents which can be inferred from the different contributions of the semiotic resources and which are generally distinguished as visual, [v], and auditory, [a]. It is thus, for instance, possible to describe the film’s main protagonist as represented participant: smoke seller, since the title of the film already indicates this name. The descriptions of the various discourse referents mainly goes back to general description methods in Multimodal Discourse Analysis (cf., e.g., Kress and van Leeuwen 1996, 2001). Every discourse referent in the boxes is tagged with a variable such as (x) or (y), which makes it possible to describe dependencies between the referents. In the example, the only auditory source besides some synchronous noises is music accompanying the images. In general, auditory discourse referents can also be realized by spoken language, sounds or noises.
The description of the eventuality itself as a propositional verbalization is given in bold in the last line of each box. This line symbolizes how the eventuality can normally be inferred from the discourse referents. In the first box, for example, the two discourse referents (x), the smoke seller, and (z), the branch in his hand (which is already surrounded by the purple smoke), play an important role for the inference of the eventuality of the first event. This inference as a defeasible consequence relation is marked by the logical operator |~ in the last line. This formulation indicates that it is these two discourse referents that allow the inference of transform in this context.
The analysis of the logical forms makes it possible to give an explicit illustration of the interplay of the various semiotic resources and their functions as well as the emerging interpretations of their intersemiosis. In the example boxes, various referents directly influence the formulation of the eventualities. They are all marked in the last lines of the boxes leading to the inference of the respective eventuality. The auditory referent, the accompanying music, supports the continuity of the whole film and clearly influences meaning construction here, but it is never a distinct part of the inference process describing the eventualities and separating several events from each other. 4
With these descriptions, we thus have a concrete examination of how audiovisual artefacts can communicate a specific semantic content which we described as propositions or predicates and which is still questioned within the context of visual argumentation (see the quotation from Blair 1996 above). Although these interpretations are always defeasible and abductive, they can be taken as normally be inferred by the recipient according to his/her world knowledge and the knowledge about the context of the filmic events. Consequently, both the concept of context as well as the notion of inference are combined into the analysis. Since the eventualities are formulated as abstract but concise formulations of the semantic content, which can conceivably be made by another predicate, we cannot define strong propositions that will be described by each recipient. However, we assume that, similar to the comprehension of verbal texts, these constructions of the semantic representations are guided by both bottom-up and top-down processes that activate a mental representation of what is going on in the film. Van Dijk and Kintsch (1983), for example, highlight that the context of verbal discourse normally affects the identification of specific words, which are then taken for the formulation of the eventualities due to activated pathways. We take the inference and formulation of filmic eventualities similarly as an abstract verbalization of the mental construct.
It would therefore be possible with the provided framework to describe each narrative event in El Vendedor de Humo as a logical form represented by its respective verbalization. This sequence of eventualities can then be seen as a conceptual depiction of the film’s plot or topic structure, which not only illustrates how the film is generally comprehended, but also gives a first impression of how the high number of interpretive possibilities of the resources’ interplay can be successfully narrowed down to a semantic basis from which the argumentative patterns can be analyzed. This meets the requirements that Birdsell and Groarke (1996) set up for any approach to visual argumentation and which we take as equally necessary for the question of filmic argumentation. They highlight that “[a]ny account of visual argumentation must identify how we can a) identify the internal elements of a visual image, b) understand the context in which images are interpreted, c) establish the consistency of an interpretation of the visual, and d) chart changes in visual perspectives over time” (Birdsell and Groarke 1996: 9).
In particular, the identification of the internal elements and their interpretation within the respective context are explicitly satisfied by the analysis showed above. A further examination of all logical forms of the film and other filmic material will deliver the required consistency in the interpretation as well as the possibility to include changes and novelties in and of the material. The combination of both a multimodal discourse analytical as well as logical approach to the analysis of the semantic content of audiovisual discourse thus proves to be very useful and valuable for a further examination of the argumentative patterns to be interpreted from this basis. How this can successfully be operated for our example film will be shown in the section following.
3.2 Multimodal argumentative patterns in El Vendedor de Humo
The second logic of the logic of film discourse interpretation provides a method of relating the various narrative events to each other by inferring film discourse relations between them due to the particular context. As illustrated above, these relations serve as textual cues which give more information about the rhetorical and argumentative structure of the filmic discourse. We will therefore shortly refer to a general description of the film discourse relations available in Wildfeuer (2014) and then show their application to the example film in further detail.
Based on the description of an experimental set of discourse relations for verbal discourse (cf. Asher and Lascarides 2003), the following seven relations have been described for filmic discourse: Narration, Elaboration, Explanation, Result, Background, Parallel, Contrast (cf. Wildfeuer 2014). This choice of relations is motivated by a semantic interpretation with a very restricted notion of meaning which goes beyond compositional and lexical semantics, but is mostly based on general logical principles and relations between entities which we can find as basic conditions for all communicative purposes. The definitions of the relations are available both as meaning postulates and default axioms in formal descriptions which make it possible to constrain the inference process for each relation. For the interpretation of a Result-relation, for example, it is necessary that the discourse gives evidence for a segment β to be caused by a preceding segment α. This is formulated in the equation following. (1)
It can be read as follows: An underspecified discourse relation ? (α, β, λ) holding between the segments labelled α and β in the context of the discourse structure labelled λ is normally inferred as the specific discourse relation Result when a cause in the discourse D is available enabling a reasonable conjunction between the two discourse segments α and β. The abductive inference is marked by the operator>as a defeasible implication to be read as “if … then normally.”
A further necessary condition for this relation to be inferred is a temporal relation organizing the order of the eventualities to be related to each other: (2)
The formal description expresses the necessary temporal order of the events eα and eβ in a linear consequence. The discourse relation Result can only be interpreted if this order is maintained by the sequence of events. 5
Both conditions for Result are fulfilled by the context of the first two example logical forms we analyzed in the previous section. The first event, eπ1=transform, is directly followed by the second event, eπ2=show. The sequence thus corresponds to the temporal relation described in the second equation above. Furthermore, the first event gives a description of a transformation process which can be logically inferred as leading to the second event, namely, that of showing a resulting object of this transformation. The film uses an elliptic depiction of the action process of transforming something into something else which does not show the exact transformation, but rather outlines its result in the shot following. Therefore, the action process visible in the first shot gives evidence for a reason that the purple smoke causes the object shown in the second event. This leads to the inference of a Result-relation, even though the two events are in a spatiotemporal consequence which would normally lead to a Narration-relation (cf. Wildfeuer 2014). Nevertheless, the effects for the Result-relation are stronger than those for Narration, since the eventualities do not exactly overlap (which is the temporal condition for a Narration-relation). This then corresponds to the principle of Maximise Discourse Coherence described above: the maximally preferred discourse relation is interpreted here in order to guide the recipient during his/her interpretation. Consequently, it is more plausible and useful in this context to infer a causal relationship between the events than simply putting them into a spatio-temporal consequence. 6
In a second step, we can now examine this rhetorical function of the causal relationship in further detail in order to gain more information about the argumentative patterns of the filmic discourse. According to the already mentioned principle of maximal coherence, we can assume that the Result-relation is the most preferred relation in this specific context. We can therefore describe this relation as a textual cue guiding the recipient to the interpretation that the transformation of the object into something else by the smoke seller is a verifiable event which has a truth conditional validity due to the inferred semantic content of the available context. This interpretation may change during the further unfolding of the film, but at that moment, the recipient’s hypothesis about the film’s meaning can be strongly uphold. Moreover, further sequences of similar events, all showing transformations accompanied by the purple smoke, prompt the recipient to understand the narrative processes and their causal relationships in a similar way. It is for example likewise possible to interpret the two events displayed in the third line of Figure 3 as causally related by the same circumstances. We can therefore conclude that these textual cues or constraints on the recipient’s interpretation do not only inform the spectator about the story and its circumstances, but also convey the respective justifications for the reconstruction of these details. By inferring various causal relationships between the events, the recipient recognizes the reasonability and suitability of the story and accepts its validity.
With regard to the film’s further unfolding and the evolving story, it is interesting to examine some further circumstances and rhetorical relations that give insight into the film’s later argumentation. As we already showed above, it turns into a different evaluation of the main protagonist and his abilities and objectives. This change in the evaluation is, as we assume, equally provoked by the interpretive mechanisms provided by the film’s rhetorical structure. It is for example possible to infer further Result-relations between the events shown in the second part of the film, although their interpretation is more complex. For instance, the two events shown in Figure 2 can also be causally related to each other, since it is possible to infer a re-transformation of the toy plane into the branch caused by the event showing the plane plunging into the vat (see the logical form of eπ3 in Figure 3). Nevertheless, these two events do not follow each other in a direct spatio-temporal relation, since eπ3 is part of a montage of various events showing the inhabitants celebrating their hero, whereas eπ3 (as showing the re-transformed branch) occurs seconds later when the smoke seller has already registered the storm and starts phasing out his equipment (see the two sequences of stills in Figure 4).
Although both conditions for Result are indeed fulfilled, it is definitely a more complex comprehension process of bringing these two events from two different sequences together. Nevertheless, in particular event eπ3 (show) with a medium close-up of the young boy and his re-transformed branch displays a convincing pattern that draws the recipient’s attention back to the small object and its actual state. This is an interesting phenomenon mainly constructed by the specific filmic technique of the camera perspective which also plays an important role with regard to the argumentative power of filmic discourse. Furthermore, this event, as analyzed in Figure 3, is not accompanied by music and thus may gain more attention in contrast to the other events which follow the coherent unfolding of the narrative. In a first step, the event can therefore be interpreted as showing a contrast to the situation before, namely, that of the young boy playing with a toy plane. As a consequence, this contrast can be seen as an indication for something to be happened with this plane. The recipient is thus guided back to the events shown in the preceding context in order to re-evaluate the hypotheses drawn in a first interpretation of this sequence. The specific perspective on the branch and the young boy’s reaction may then lead to a re-interpretation of the spatio-temporal consequence of events shown before by finding a cause for the re-transformation. This causal interpretation is offered by the event of plunging into the water. Whether this new interpretation is then actually drawn by the recipient can only be examined by an empirical analysis. The textual information, i.e., the analysis of the semantic content and the discursive structure, however, make it possible to hypothesize this interpretation.
From a discourse analytical perspective of the dynamically unfolding discourse structure, this re-interpretation is described as a so-called discourse pop because of the leap that the recipient has to make from the last event to an event somewhere in the discourse before (cf. Asher and Lascarides 2003; Wildfeuer 2014). Attaching eπ3 to the discourse, i.e., relating the event of showing the re-transformed branch to the preceding context and thereby interpreting the causal relationship, results from choosing the most preferred relation within this context, which is Result in favor of, for example, a simple Narration-relation. This dynamic interpretation process is again based on logical reasoning about the interdependencies between the events of the text and the conditions that must work for the interpretation of some relations. According to these conditions, it is more plausible to relate eπ3 to an earlier event by the causal relationship than to attach it to the sequence of events focusing on the smoke seller. It is thus again the principle of maximal coherence which influences the discourse interpretation here.
Moreover, there are further textual cues which work similarly, as for example the one given in the two events shown in Figure 5. At the beginning of the film, a small and squinting dog is transformed into a frightening, evil-looking animal. It is appearing later in the film as catching the young boy’s stitch thrown away by the smoke seller and thereby jumping into a well. Again, a few shots later, the dog is shown emerging from the water, now being re-transformed to its original appearance.
The same complex comprehension process is necessary here in order to follow the film’s argumentation structure. It is again a Result-relation that can be inferred as a causal relationship between the event of jumping into the well and thereby plunging into water and the resulting event of being re-transformed. However, the two events do not follow each other directly either; instead, the second event is only shown in a later narrative sequence when the smoke seller is already leaving the village. Nevertheless, at that moment when the re-transformed dog appears, this is again first a contrasting event to the situation before which then guides the recipient back to the preceding context and the possible causal interpretation which has already been introduced by the first example analyzed above. The repetition of the same inference to be drawn from the context thus makes is easier to follow this textual cue as a distinct part of the argumentation.
Interestingly, the importance of the transformation effect of water as a part of the rhetorical strategy is also introduced by the film’s discursive structure in that some events focus the recipient’s attention on the smoke seller’s reaction to the approaching storm. There is for example a specific cue in the discourse structure showing his attitude in a further event to be inferred from the context as directly reacting to the thunder and, as a consequence, stopping his performance. This specific cue can formally be analyzed as a further Result-relation indicating a causal interrelation between the thunder and the smoke seller’s behavior. With particular regard to the following context and the transformations as causally explained by the influence of water, this reaction becomes plausible and allows further conclusions about the story’s unfolding. Particularly, it influences the validity and justification of the recipient’s hypotheses drawn for the first part of the film. Whereas the inference of the number of Result-relations at the beginning all indicate the validity of the smoke seller’s ability, the interpretations in the second part of the film questionize this validity or the validity of its results, i.e., the transformations.
However, this effect of questionizing these aspects is again a further pattern of the argumentative structure, looking ahead to the interpretation of the film’s end, which is depicted in Figure 6.
After leaving behind the village, its inhabitants and their re-transformations which are shown in the background, the smoke seller is shown on his cart, finally (re-)transforming himself into an old man. The events which can simply be described as, on the hand, being the young smoke seller driving the cart and, on the other hand, being an older man driving the cart, can again be related to each other by a Result-relation whose cause can be interpreted as the rain shown in both images. On the basis of the re-transformations caused by the water before, it is now most plausible to infer that the rain brings to light the truth behind all smoke and mirrors and re-transforms the smoke smeller into his original appearance. Although this original appearance has not been shown before, it is interpreted as being the result of a transformation process that must have been happened before the smoke seller came to the village.
This interpretation is then the final resolution of all textual cues and indications which have guided the recipient through the text, giving him/her reasonable arguments to interpret details and appearances to be true. The smoke seller himself, for example, is interpreted as being young and demonstrating magical abilities throughout the whole film, up until his original appearance is shown. The validity of his transformations is already questioned after a few sequences when the first Result-relations indicate a change in the interpretation. These changes already point to the smoke seller’s actual treachery and push the recipient and his interpretation into this direction; the convincing arguments for this interpretation, however, become only clear at the film’s end.
3.3 Summary of the analysis
With the analytical framework introduced above, it has now been possible to work out the argumentative patterns within the film’s discursive and rhetorical structure. By outlining the intersemiotic construction of the narrative events on the one hand and their mostly causal interrelationships on the other hand, it has become clear how the film pushes the recipient’s argumentative reconstruction. Interestingly, this reconstruction is mainly based on a number of Result-relations worked out as the concrete textual cues within the example film. This corresponds to the general assumption that additional content to be attached to narrative texts is not exclusively temporal in nature and that it is “much more plausible … that we infer a causal link in virtue of the presence of a rhetorical link between the propositions” (Asher and Lascarides 2003: 7). However, comprehensive analyses of further filmic discourses (cf., e.g., Wildfeuer 2014) have shown that all other film discourse relations as shortly introduced above work similarly as causal, associative and elucidatory links, each describing other levels of details of the respective discourse and with different effects for the interpretation. It is thus absolutely necessary to further examine argumentative patterns in filmic discourse by an empirical analysis based on the framework suggested above.
4 Conclusion: It’s all about logics?!
We have shown in this paper how a more detailed and formally supported analysis of the meaning-making strategies in multimodal texts enables a reliable examination of the argumentative patterns and the rhetorical structure of filmic discourse. With the provided tool of the logic of film discourse interpretation based on fundamental logical principles of abductive reasoning and inferential interpretation, it has been possible to work out how the short film El Vendedor de Humo as a first illustrating example guides the recipient through the interpretation of its intersemiotically constructed content and thereby provides explicit textual cues that acknowledge the validity and reasonability of the film’s argumentation.
Consequently, this analysis represents a further approach to the systematic rhetorical analysis of multimodal argumentative discourse which is still a matter of particular interest in argumentation theory, visual communication and multimodal linguistics. Whereas further accounts mainly question the actual interpretation of how meaning in multimodal discourse is made and can be represented linguistically, the provided tool offers a more formal and detailed description of how filmic semantic content can be summarized in abstract propositional descriptions. These descriptions can then in a further step be taken as a basis for the question of how filmic and other multimodal discourses argue without directly delivering propositionally expressed arguments. As has become clear in the example analysis, these arguments can be worked out by the analysis of the film’s relational meaning-making strategies in terms of concrete discourse relations holding between events in the film’s narrative. These relations can at the same time be interpreted as concrete textual cues that instruct the recipient to draw diegetic inferences about the content.
As a result, the examination resembles only partly approaches from argumentation theory which concentrate on the rhetor as the producer of the text aiming at influencing and persuading its audience (cf., e.g., Van den Hoven 2012; Van den Hoven and Yang 2013). Instead, our approach focuses on the textual artefact as the interface between the rhetorical strategies of the text and its contextual interpretation by the recipient. It therefore follows the suggestions for a rhetorical analysis given, for example, by Van den Hoven and Yang (2013) and combines the analysis of the kairos as the specific occasion in which the text has been produced with an examination of the recipient’s meaning making strategies as the perspective of the audience as well as with an analysis of the form of the discourse (cf. Kuypers 2009; Van den Hoven 2012).
As the theoretical and methodological starting point, logics has successfully proven its accountability and applicability to these questions. It has been possible to make the question of multimodal meaning-making based on abductive and defeasible reasoning more reliable and to a lesser extent attributed simply to the work of either the producer or the recipient. It is now de facto possible to constrain the immense number of potential interpretations in order to elucidate exactly those elements that play a significant role within multimodal and filmic argumentation.
Further interesting aspects which could be worked out during this examination were, for example, the various functions specific filmic techniques have in this meaning-making process. In El Vendedor de Humo, the camera perspective as well as the music play an important role for the interpretation and evaluation of specific narrative events and their relations to each other, which has only partly been mentioned and analyzed above. It would be interesting here to ask in more detail for the specific tasks of these resources and their role for the argumentation. So far, we could only work out that they help focus the recipient’s attention on some small details of the story which, in a further step, are relevant for the interpretation. Whether these resources can be further analyzed, for example going back to traditional descriptions of rhetorical devices within the categories of logos, pathos and ethos (cf. Aristoteles; Sandys et al. 1970), must be examined elsewhere.
However, since our approach delivers explicit and verifiable results for the question of how multimodal discourse conveys argumentative structures, it can be applied to these and further empirical analyses of filmic and other multimodal texts and may thereby serve as a starting point for more comprehensive approaches to argumentation and rhetorics.
Aristoteles. 1999. Rhetorik, Gernot Krapinger (ed.). Stuttgart: Reclam.Google Scholar
Asher, Nicolas & Alex Lascarides. 2003. Logics of conversation. Cambridge: Cambridge University Press. Google Scholar
Bateman, John A. & Karl-Heinrich Schmidt. 2011. Multimodal film analysis: How films mean. London: Routledge. Google Scholar
Birdsell, David S. & Leo Groarke. 1996. Toward a theory of visual argument. Argumentation and Advocacy 33(1). 1–10. Google Scholar
Birdsell, David S. & Leo Groarke. 2006. Outlines of a theory of visual argument. Argumentation and Advocacy 43. 103–113. Google Scholar
Blair, Anthony. 1996. The possibility and actuality of visual arguments. Argumentation and Advocacy 33(1). 23–39. Google Scholar
Bordwell, David. 1985. Narration in the fiction film. London: Routledge. Google Scholar
Bordwell, David. 1989. Making meaning: Inference and rhetoric in the interpretation of cinema. Cambridge: Harvard University Press. Google Scholar
Bordwell, David. 2006. The way Hollywood tells it: Story and style in modern movies. Berkeley: University of California Press. Google Scholar
Bordwell, David. 2008. Poetics of cinema. London: Taylor & Francis. Google Scholar
Bordwell, David, Janet Staiger & Kristin Thompson. 1985. The classical Hollywood cinema: Film style and mode of production to 1960. London: Routledge & Kegan Paul. Google Scholar
Doxiadis, Apostolos. 2010. Narrative, rhetoric, and the origins of logic. Story Worlds 2. 77–99. Google Scholar
Herman, David. 2002. Story logic: Problems and possibilities of narrative. Lincoln: University of Nebraska Press. Google Scholar
Hobbs, Jerry R. 2003. Discourse and inference. Information Sciences Institute. http://www.isi.edu/~hobbs/disinf-tc.html (accessed 18 October 2017).
Iser, Wolfgang. 1980. Interaction between text and reader. In Inge Suleiman & Susan R. Crosman (eds.), The reader in the text: Essays on audience and interpretation, 106–119. Princeton, NJ: Princeton University Press. Google Scholar
Kamp, Hans & Uwe Reyle. 1993. From discourse to logic: Introduction to modeltheoretic semantics of natural language, formal logic, and discourse representation theory. Dortrecht: Kluwer. Google Scholar
Knape, Joachim. 2005. Rhetorik. In Klaus Sachs-Hombach (ed.), Bildwissenschaft: Disziplinen, themen, methoden, 134–148. Frankfurt: Suhrkamp. Google Scholar
Kress, Gunther & van Leeuwen. Theo 2001. Multimodal discourse: The modes and media of contemporary communication. London: Arnold. Google Scholar
Kress, Gunther & Theo Van Leeuwen. 1996. Reading images: The grammar of visual design. London: Routledge. Google Scholar
Kuypers, J. A. 2009. Rhetorical criticism: Perspectives in action. Lanham, MD: Lexington. Google Scholar
Maestro, J. (dir.). 2012. El Vendedor de Humo. Spain: PrimerFrame. https://www.youtube.com/watch?v=dwWqMgddes4 (accessed 18 October 2017).
O’Halloran, K. 2004. Visual semiosis in film. In Kay O’Halloran (ed.), Multimodal discourse analysis: Systemic functional perspectives, 109–130. London: Continuum. Google Scholar
O’Keefe, D. J. 1977. Two concepts of argument. Journal of the American Forensic Association 13. 121–128. Google Scholar
O’Keefe, D. J. 1982. The concepts of argument and arguing. In J. R. Cox & C. A. Willard (eds.), Advances in argumentation theory and research, 3–23. Carbondale, IL: Southern Illinois University Press. Google Scholar
Peirce, Charles S. 1931–1966. The collected papers of Charles S. Peirce, 8 vols., C. Hartshorne, P. Weiss & A. W. Burks (eds.). Cambridge: Harvard University Press. [Reference to Peirce’s papers will be designated CP followed by volume and paragraph number.]Google Scholar
Roque, Georges. 2012. Visual argumentation: A further reappraisal. In F. H. Van Eemeren & B. Garssen (eds.), Topical themes in argumentation theory: Twenty exploratory studies, 273–288. Dortrecht: Springer. Google Scholar
Sandys, John Edwin, Edward Meredith Cope & Aristoteles. 1970. The rhetoric of Aristotle, John Edwin Sandys (ed.). New York: Olms. Google Scholar
Sperber, Dan & Deirdre Wilson. 1986. Relevance: Communication and cognition. Oxford: Basil Blackwell. Google Scholar
Thompson, Kristin. 1988. Breaking the glass armor: Neoformalist film analysis. Princeton, NJ: Princeton University Press. Google Scholar
Van den Hoven, Paul. 2011. Iconicity in visual and verbal argumentation. In F. H. Van Eemeren, B. Garssen, D. Godden & G. Mitchell (eds.), Seventh international conference of the international society for the study of the argumentation proceedings, 831–834. Amsterdam: Sic Sat. Google Scholar
Van den Hoven, Paul & Ying Yang. 2013. The argument reconstruction of multimodal discourse, taking the ABC coverage of president Hu Jintao’s visit to the USA as an example. Argumentation 27. 403–424. CrossrefGoogle Scholar
Van Dijk, Teun A. & Walter Kintsch. 1983. Strategies of discourse comprehension. New York: Academic Press. Google Scholar
Wildfeuer, Janina. 2012a. Intersemiosis in film: Towards a new organization of semiotic resources in multimodal filmic text. Multimodal Communication 1(3). 276–304. Google Scholar
Wildfeuer, Janina. 2012b. More than WORDS: Semantic continuity in moving images. Image and Narrative 13(4). 181–203. Google Scholar
Wildfeuer, Janina. 2014. Film discourse interpretation: Towards a new paradigm for multimodal film analysis. London: Routledge. Google Scholar
Wirth, Uwe. 2000. Die Welt als Zeichen und Hypothese. Perspektiven des semiotischen Pragmatismus von Charles Sanders Peirce. Frankfurt: Suhrkamp. Google Scholar
Wirth, Uwe. 2005. Abductive reasoning in Peirce’s and Davidson’s account of interpretation. Semiotica 153(1/4). 199–208. Google Scholar
Argument1 is a description according to the concept of O’Keefe (1977), who differentiates this form of an argument against the concept of arguments2 as an “overt disagreement … between interactants” (O’Keefe 1982: 11). Blair (1996: 24) uses the first argument here, “because visual arguments are more plausibly akin to reasons for claims (arguments1) than to open disagreements between interacting parties (arguments2).” Due to space constraints, we are not further referring to the various discussions of the concept of visual argumentation and its acceptance here.
We describe the reasoning process as an explicit narrative construction, since the recipient mainly tries to find spatial and temporal as well as causal circumstances within the film and its story. This comprehension of basic narrative principles is, as we assume, available in all feature films that count as narrative or expository texts. We do not take into account documentary and other non-fictional films here.
For a comprehensive explanation of how the various events are separated and how other discourse referents influence the inference of the logical forms, see the detailed analyses in Wildfeuer (2014, 2012a).
The reversed order of the events would lead to the inference of an Explanation-relation which has as a necessary condition to be fulfilled the temporal consequence before(eα, eβ) (cf. Wildfeuer 2014). As a consequence, the specific order of the events described in these definitions is of considerable importance for the interpretation of filmic discourse. One of its basic qualities is linearity which does not logically include a chronological temporal sequence of the events. A precise formulation of the temporal relationships and conditions is therefore absolutely necessary.
The framework of the logic of film discourse relations also provides a formal description tool for the illustration of inferring the discourse relations between the events. We forgo this description here due to space constraints. For a comprehensive overview of how the discursive structure of flmic discourse can be graphically and formally illustrated, see Wildfeuer (2014).