Taking Metaphor’s Contextuality Seriously: Sharpening Theory – Reinspecting Methodology

Taking Metaphor’s Contextuality Seriously: Sharpening Theory – Reinspecting Methodology

Petra Gehring



If there was ever a time when metaphors were deemed to be an academic triviality, in which they were even regarded as dispensable because mere decoration, then this time is long past. For decades, we have been experiencing the opposite: There is hardly a phenomenon apart from the classical natural sciences that in a comparable manner attracts so much interdisciplinary research attention—from linguistics and cognitive science via philosophy of science and history of science to semiotics, literary research, cultural studies, political rhetoric as well as the arts together with image and film studies. And as we know, for the computer sciences the algorithmic approach to metaphors has long been a complex research field of its own.


However, there is hardly any other object that from the outset can be modelled as differently as the metaphor—and there is hardly any other that can be addressed so casually and indifferently. And this, too, will have to be conceded: the ubiquitous interest in the metaphorical does not lead to an elucidation of the object field, but rather to a division of the field into “schools” and to an arbitrary expansion of the field. This is a particular difficulty for inter- and transdiscipinary work. Actually, anything can be said about metaphors. Definitions seem so difficult that they are not even requested—and the turn of phrase that the metaphor is itself a metaphor, which has become something of a party joke, has even made its way into academic books.




There are innumerable theories of the metaphor. Many combine linguistic assumptions with cognitive psychology, ultimately with anthropological assumptions. According to this family of theories there are more or less stable, also more or less culture typical semantological patterns in the use of metaphors: “concept systems” [Lakoff] or “stock meanings” [Gentner] (“x is y” or “x is like y”). For their part, these are based on basic non-verbal patterns, “imaginative” clusters [Lakoff] or “mental maps”. As we know, the assumption of such systems, clusters or maps can guide research to—successfully—collect empirical data. Just as there are systems of “literal” meanings (a lexicon in the narrow sense), there are also improper uses of words in general usage or typical phrases aiming at the “metaphorical” transfer of semantic contents; as if there were something like an ‘unofficial’ repertoire of patterns, in part lexical, in part syntactic, or even with ‘stylistic’ function, by virtue of which language users are more or less virtuosos in the use of metaphors.


However, the schemas remain paradoxical. On the one hand, they are abstract, to a great extent they make it possible to take leave of normal interpretations. On the other hand, according to the diverse models it is possible to filter them in concrete experience with respect to a possible ultimate ‘root’—they would then be attributed to bodily experience [Johnson], to space–time forms [Kant] or comparable mechanisms in the brain [Evans and Chilton], or to a determinism of discourse itself, the supra-individual infrastructure of which contains something like “commonplaces” [Schaffer] with underlying effects.


I side with a completely different type of metaphor theory. This family of theories is not mainly interested in calibrating metaphorical findings against the background of a normal language, reducing them so to speak to the smallest common denominator of a quasi-meaning which is conventionalized, though vague and simple, and which they share with other metaphors—for example “rational is up—emotional is down” or “ideas are food” [Lakoff and Johnson]. Nor is the perspective on which I shall now focus interested in ubiquitous, more or less trivial metaphors. Rather, the paradigm is the (astounding) individual case of a distinct metaphor, perhaps even a singular metaphor within the particularity of the context, that is surprising because in it almost everything—normal language, the user—obviously functions by improvisation in favour of a situative meaning of a term (a sentence, a text).


For the moment, I shall call the family of metaphor theories that take metaphorical phenomena in this sense as their point of departure “work-pragmatic” theories. Hermeneutic literature research, text linguistics, which is interested in the elaborate special case, and the analysis of distinct metaphors in the language of theory – i.e., in the ‘thinking texts’ of philosophy (my field of work) – are oriented on such theories of metaphor: Not the cognitive normality of meaning levels that in some way are secondary is studied, but rather the textual exceptions that stand out from it.


Here, too, a broad field of various theories is opened. I do not want to explore this field although it would be interesting to take a closer look at the authors relevant to our topic because some of them are hardly perceived in the English-language debate; let me mention only Paul Ricoeur, Hans Blumenberg and Lutz Danneberg. A characteristic aspect is the abstention from any mentalist or anthropological matrix, in general from any systemic matrix: the point is not the concealed rule of a langue. Instead, individual instances of metaphors are examined, text passages, specific events on the level of a pragmatics which so to speak plays with the rule and apart from the rule generates evidences that involve a complex – but: context-dependant – semantology. How this pragmatics operates must yet be elucidated, suffice it to say here that in reading practice it is successful.


If from this perspective assuming a specific ad hoc operation of the metaphor we ask about the “meaning”, three points are central: For the one part, the dimension of the phenomenon is changed. Now the term (for example a focus expression) cannot be viewed in isolation, rather the passage obtains its meaning by virtue of the resonance of its context (which is what it is and nothing else than what it is in precisely ‘this’ text). It is no surprise that instead of representational theories of meaning recourse is taken to a theory of the differential development of meaning and that the text or work pragmatic perspective is akin to what are called interaction theories of the metaphor [Black] according to which the “frame” of an expression belongs to the metaphor; the metaphor actually consists of the divergence between an expression that would be expected in normal language (according to the frame) and this frame itself, one that does not match the expression but rather constrains the activation of alternatives.


Secondly, for this very reason the metaphor cannot be determined in a positive semantic sense as a transfer (of something or other, with “source” und “target” and the like); rather it must be determined in terms of the pragmatics of reading or discourse, which more precisely means: negatively. Specifically, it must be determined in terms of the disruption of context that it constitutes, as a disturbance of a normal understanding, as a passage with ‘open meaning’ that is dealt with in a certain way in terms of a free but not arbitrary decision that is sensitive to the context. And perhaps the point is not a simple decision, but rather a bit of interpretative work, a process of understanding.


And thirdly, in the elucidation of how precisely the metaphor obtains its meaning, this perspective is not interested in the residual, generalizable schema, some discourse behind the discourse (waiting as a sort of secondary system that ‘always’ steps into the breach); rather the interesting aspect is the specific abundance of semantic relations, the mix of particular contexts in the text mobilized by focus, by frame and by their interaction. The bipolar model of “source and target” must at least be complemented by or overlaid with a model of multiple interferences: Even on the “source” side there is a compound situation (focus and frame), and a “target”, that is, the goal area, is not simply an isolated term somewhere in the frame (as suggested by atypical example sentences such as “Henry is a lion”). Rather, what the metaphor ‘says’ or indeed ‘means’ by virtue of the semantic transfers with which it operates must be discovered, in case of doubt by means of semantological–hermeneutic analyses in which interfering contexts again count.




There are of course passages that are metaphorically paler, passages that mobilize relatively little context and interpretative effort: When Friedrich Schiller writes

“that we collect the entire energy of our mind on one focal point and concentrate our entire being in one single force” (1)

(1) Friedrich Schiller: Über die ästhetische Erziehung des Men­schen [On the Aesthetic Education of Man] (6th letter). Stuttgart 2000, p. 27.

there will probably not be much reason to ponder too deeply about the energy and the mind and the focal point (although the constellation, especially when combined with the “concentration” in one “force”, is certainly striking). Nonetheless, we are readily able to come to terms with such passages in the style of practised secondary meaning. (Similarly, from the perspective under discussion, what are called ‘dead’ metaphors are not metaphors at all, rather they are lexicalized so that no transfer or contextualization work at all is performed.) The more genuine the metaphors are, the less linguistic generalizations will help (or “ontologies” that are previously derived in semanticist computer-language terms). When Hans Jonas writes that for the sake of the dignity of the human person we have to


“possess ourselves and not let our machine possess us, we have to get the technological gallop under non-technological control” (2)

(2) Hans Jonas: Warum die Technik ein Gegenstand für die Ethik ist: fünf Gründe [Why technology is an object of ethics. Five reasons]. In: Hans Lenk, Günther Ropohl (eds.): Technik und Ethik. Stuttgart 1987, pp. 81-91, here p. 91.

technology and the gallop and control (together with the question of possession and the machine) are comprehensible, but the metaphor unfolds in a complex manner. In this case, knowledge of discourse and knowledge of the text are needed in order to understand details; indeed, a comprehensive consideration of pragmatic aspects of complete text understanding is also required.

This brings me to my point: the contextuality of metaphors, at least of those metaphors that cannot be attributed to a quasi-lexicon of secondary or tertiary meanings of the focal expression (in the style of a “focal-point metaphor”, “gallop metaphor”) or to a paraphrase that is once and for all derived by abstraction, then to operate as a “concept system” [as Lakoff and Johnson might put it]—in our examples this could perhaps be: “mind is collected energy” (for the quotation from Schiller) or “technology is a bolting riding horse” (Jonas). It seems to me that passages that are metaphorically outstanding have earned more than that. If we wanted to quarrel, we could claim that Lakoff and Johnson are not interested in metaphors at all, at best they are interested in the ‘principle of metaphor’ as a mental function completing the faculty of language and providing comprehensive order and foundation for the semantic economy of language.





However, I am not really intent on a controversy on the theory of metaphor. Rather, I would like to make it clear that the context-oriented perspective, even though it drives us into complications, is methodologically more fruitful for the concrete approach even in text linguistics, more so in interpretative disciplines ranging up to image, film and other multimodal artefacts. The shift of perspective advocated here is certainly not trivial. It requires us


  • to focus on the individual aspects of the particular occurrence,
  • to free ourselves of ideas that, though they have with a long-standing tradition, are (overly) abstract,
  • to make advance decisions to specify exactly what is to be studied under the title “metaphor” depending on the corpus and the research question,
  • and to communicate this to the academic reader.


The object metaphor is modified when it is assumed that a given context includes both normality and deviation. It is then possible to become more precise.


Among the points that have to be renounced is the talk of what are called dead metaphors, which can only be distinguished from non-metaphors by means of etymological recollection or other dissociation practices. It seems to me to be evident that something that in the situation of use does not require transfer is not a metaphor. The other way round, no fixed inventory or lexicon of metaphors can be derived from language reality: at least in principle, every term can become a metaphor by virtue of the manner of use. If an inventory list of metaphors is needed, it must be established item by item for a specific corpus. We must free ourselves of the idea of a main meaning—the notion that the “source” of metaphors is always more concrete than the target—or that what is transferred is always perceptual attributes of what the focal expression designates (consider simply the “focal point” in the Schiller example, which conceptually brings the entire construct character of theoretical physics to bear with a view to what is collected in the mind).


The radically pragmatic criterion of a context violation or disruption of context requires us to make further specifications; it is the characteristic of genuine metaphoricity (to be used in the identification of metaphors), and goes beyond a mere context-driven shift from the literal main meaning to secondary meanings. The self-evidence of a “disruption” implies a yes or no decision on normality (either exceptional contextualizing efforts are required or not); but on the other hand, it is obvious that the disruptions in question can be more or less dramatic so that we would have to speak of normalities in the plural. A phrase that in one text, relative to the normality of genre and discourse, stands out as solitary and strange, may fail to generate a disruption in another text because the phrase in question has a terminological sense (having been previously defined) or it is ubiquitous and thus has no genuine or hardly any genuine metaphorical value.


Reckoning with the “disruption” thus makes our vision more flexible—but it also means that we have to speak about questions of calibration. Comparisons, say between metaphors in poetry and metaphors in academic language, are not impossible, but they become more difficult. Specifically, texts or corpora in which metaphorical phrases are used at a high frequency, perhaps even in every sentence, are from the very beginning effectively ‘noisier’ than are texts or corpora in which the use of metaphors is a rare exception. The principle of the contextual disruption is thus itself already context-dependent: It can be ruined by a complete lack of adherence to meaning (for example baby talk) or by inflation (for example bombast). However, the tools of analysis can be readjusted to look for fine differences instead of the great deviation. Even a bombastic text in which clichés are amassed can on second sight contain artful disruptions that among all the blatant clichés function metaphorically and trigger genuine transfer.


Overall, it seems to me that taking the context as the point of departure for conceiving and interpreting the metaphor does not per se privilege drastic metaphors. Genuine metaphors (metaphors that bring about a clear disruption of the “normal” context to be expected in discourse) do indeed stand out more clearly in reading. But they are simply more spectacular. As such they must not necessarily be semantologically more ramified than paler metaphors that are less conspicuous in context, that tend to appeal more to convention and thus seem to be less in need of recontextualization. Thus, there may be the case of a ‘concealed’ genuine metaphor—detectable by an analysis of corresponding sensitivity—just as well as a term that stands out all too clearly (for example flowery language) can be revealed to be not a metaphor, but simply a set phrase.


Taking the example of academic usage of “source”, Hans Blumenberg showed how to distinguish between jargonizing and astute remetaphorization, using the expression as a set phrase or terminologically—it depends completely on the context.(3) On the one hand, this context has a certain character, its own normality (including genre and discourse); on the other hand, on the basis of the biases corresponding to this normality analysis has to detect how a specific text passage in fact ‘plays’ with this normality. Very normal listeners and readers are experts in this. Metaphor research that is not comparably adaptive proves to be more stupid than the world.



(3) Hans Blumenberg: Quellen, Ströme, Eisberge [Sources, streams, icebergs]. Berlin: Suhrkamp 2012.

The contextuality of the metaphor is thus a fact derived from the chosen analysis perspective and the pragmatic definition associated with it—and also from the interest in an interpretation of the individual case that is as extensive as possible. Furthermore, from this perspective contextuality has to be regarded not as something that might tend to disrupt analysis so that it has to be eliminated by abstraction, but rather as an essential feature of metaphor, one that is central to analysis; I consider this to be a productive demand, a good ‘imposition’ on the methodological discussion. In making this claim, I do not want to say that speaking of “transfer” (which is itself often a metaphorical manner of speaking) and the positivist isolation of focal expressions together with certain semantic “patterns” associated with them from a given corpus are wrong approaches. But measured on real meaning in usage, they are at the same time too narrow and too broad. They detect trivial instances, but overlook sophisticated cases, so that in an ‘open-ended’ search for metaphors (that is, in large corpora) they do not quite make contact with those passages that function metaphorically.




After this long approach, let me come to the actual theme of this conference: what about metaphoricity in images, in art, in cinema? Can the form and manner of functioning of the linguistic phenomenon of metaphor be recognized in such non-verbal or multimodal object fields—such that we do not have to speak of the phenomenon in terms of a vague analogy? It seems to me that an answer to this question depends immediately on the selected concept of metaphor and thus on the selected type and level of methodological approach.


You will not be surprised to hear that I again caution against an overly simple understanding of metaphor (“transfer”) and in general against a semanticist interpretation according to which within a pictorial or cinematic depiction certain meaning-like contents of individual symbols or icons are overlaid or replaced by others. This involves a danger of reduction. In the worst case, the idea of an image language or cinematic language (or simply a work language) with symbol-like elements is imposed on the works in question; by virtue of these elements, the representation is said to evoke “meaning” as if it consisted of terms referring to a kind of non-verbal lexicon or perhaps to corresponding verbal concepts.


If I take the contextuality of the metaphor as the point of departure together with the pragmatics of understanding resulting from the manner in which the metaphor is relationally embedded, a different perspective emerges. This, too, will have to do with conventions, but not with conventionalized semantics; rather it could address expectations of use that so to speak remain on the near side of the threshold to a fixed symbolic system—but can quite readily react ad hoc to meaningful interrelations between signs that are appropriately constellated (or indeed to a refusal of interrelationships). There is thus a text-like normality, for example in audio-visual reception, in the form of non-verbal sequences in pictorial or cinematic usage, perhaps also in the expected “and so on” of the graphics of a comic. In a pragmatically extended sense of the word text (perhaps we should speak of “texture”), there is also a chance that something irregular can emerge, the possibility of a disruption—a disruption that I do not have to reconstruct as an interruption or change in the “meaning” of a single pictorial element or cinematic element in focus; rather, I can read it as a phenomenon of correspondence and interaction taking place primarily between contexts that in this case are from the very beginning non-verbal.


Let me cite two examples; they may well be quite simple, but they can show where the difference is located. They are two famous graffiti by the street artist Banksy. The first work is called Napalm Print, the second Flower Thrower.


As the quality of the reproduction is not relevant here I cite both works via one of the first matches on “google-Bilder” [German account]:
http://hanguppictures.com/artists/banksy/banksy-signed-prints/napalm [15.3.2017] (Napalm Print); http://www.michaeloart.com/banksy-his-most-powerful-pieces/ [15.3.2017] (Flower Thrower).

It can be seen immediately that they are very austere arrangements made up of disjunct, even picture-like elements. In my view, they function metaphorically, but do so much in a language-like manner, but rather because they have a “text-like” design. It is not because meanings are exchanged, replaced, transferred in the picture that the body of the fleeing Vietnamese girl and the bouquet of flowers are understood to be non-genuine elements, but rather because the artist creates such a clear context that the disruption of context stands out with a maximum of clarity. The result is something like a genuine metaphor; moreover, it fulfils all of Black’s conditions for interaction between focus and frame: significant aspects are transferred such as nudity vs. clothing, face vs. mask in the case of the first picture, harmlessness vs. danger, unfitness for throwing vs. throwability in the case of the second. But also the other way round: the clowns are disfigured, dehumanized creatures, the militant activist could just be revelling. Focus and frame reciprocally compromise each other in their pictorial contents: Is the face full of panic itself a clown’s mask? Are not bouquets of flowers the Molotov-cocktails of friendliness? And so on…


Moreover, both works are able to integrate further contexts into the intensive metaphorical dynamics (in the case of the “Flower Thrower”, for example, this obviously includes the materiality of the wall on which the graffito is painted, the place, the city, and so on).


I do not wish to interpret the two works in greater detail. But I do think that they can be ‘read’ as metaphors. And I also think that this is not primarily due to meanings that are symbolically prompted and then transferred, but rather to a certain measure of interrelationality—operating beyond written language, forming a kind of ad-hoc context—and that this interrelationality for its part makes a contextual disruption possible that then effectively commits the use of the picture to oscillating efforts at recontextualizing. The examples of graffito, comic and film stipulate a minimum condition marking the limits of all possible metaphoricity of non-linguistic metaphors. There has to be a texture with a sufficient drive to meaningful interrelation by virtue of which the expectations in using the image are normalized, guided and so clearly pre-structured that the “focus” is not simply a figure against a background, but is seen as part of a “frame” out of which it can fall. In the case of the two graffiti by Banksy, we observe the picture-like scenery with many spontaneously interpretable picture elements that trigger a visual cliché that all too clearly can for its part be disrupted. What is not familiar enough, cannot be disrupted. Consequently, non-conventionalized pictorial, audio-visual (or multimodal) contexts cannot originate metaphors.





Translated from German by Donald Goodwin



  • There are currently no refbacks.

Copyright (c) 2017 Gehring Petra

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.


ISSN 2567-9309

mediaesthetics – Journal of Poetics of Audiovisual Images

Cinepoetics – Center for Advanced Film Studies