The following is a pre-proof version of a chapter (DOI: 10.

1515/9783111001364-
011) in a book that is scheduled to appear in May 2024. Bibliographical details:
Forceville, Charles (2024). “Identifying and interpreting visual and multimodal
metaphor in political cartoons.” In: Manuela Romano (ed.), Metaphor in Social-
Political Contexts: Current Crises. Berlin: De Gruyter.
For verbatim quotations, please check the published version (ChF, 27-1-24).

Identifying and interpreting visual metaphors in political cartoons. Charles Forceville

Abstract
Metaphors abound in political cartoons. This is unsurprising, as visual metaphors are excellent
rhetorical tools to provide, in one glance, an evaluation of a newsworthy person or state of
affairs in the world. Analysts of cartoons, however, are faced with the challenge of examining
and categorizing any visual metaphors in a systematic and replicable manner. In this chapter I
revisit my own earlier, co-authored work on metaphor in political cartoons (Bounegru and
Forceville 2011; Forceville and Van de Laar 2019; Zhang and Forceville 2020), rooted in Black
(1962, 1979), to zoom in on possible criteria for identifying and interpreting metaphors in them.
The chapter, which includes a critical discussion of the visual metaphor identification
procedure (VISMIP) proposed by Šorm and Steen (2013, 2018), ends by giving some practical
advice to aspiring analysts of corpora of discourses featuring visual metaphors.

Keywords Visual metaphor; identifying metaphor; political cartoons; VISMIP; Black’s

interaction theory.

1. Introduction

In open societies the genre of political cartoons has traditionally been granted generous leeway
to present daring, controversial, even offensive angles on the news, since cartoonists’ freedom
of speech-and-depiction is clearly a benchmark for assessing the degree of democracy in a
country (see Pedrazzini and Royaards 2022). A political cartoon, consisting of a picture and,
optionally, short written narratorial explanatory comments and/or utterances by depicted
characters, is expected to provide, in one glance, an evaluation of a specific person or state of
affairs in the world that is in the news at the moment of publication. Creative metaphors are
excellent instruments to produce such evaluations, so it is unsurprising that they abound in
political cartoons (e.g., El Refaie 2003; Bergen 2003; Negro Alousque 2013; Lin and Chiang
2015; Al-Masri 2016; Durović and Silaški 2016; Lan and Zuo 2016; Kwon 2019; Wawra 2018;
Silaški and Durović 2019; Marin-Arrese 2019; Prendergast 2019; Abdel-Raheem 2019).
In creative metaphors (Black 1979) there is a topic (in the terminology of Lakoff and
Johnson 1980: the “target domain”) upon which a specific perspective is provided by
presenting it in terms of something else (the “source domain”). The interpretation of a metaphor
entails deciding which feature(s) is/are to be mapped from source to target. In political
cartoons, a politically newsworthy phenomenon typically functions as the metaphor’s target
domain, which is then coupled with a source domain that strongly evokes the feature(s),
attitude(s) and emotion(s) that capture(s) the perspective the cartoonist wants to present on the
target. Since political cartoons are supposed to comment critically and preferably wittily on the
world, the features mapped from the source usually have negative connotations (Schilperoord
and Maes 2009: 216). The judgment what is to be considered “negative,” to be sure, can be
more or less universal, but may also be specific for the culture in which a cartoon is published.
Whereas the basic template for analysing metaphors in political cartoons is thus fairly
straightforward, in practice analysts are confronted by a number of challenges. On the basis of
my own earlier, co-authored work on metaphor in political cartoon (Bounegru and Forceville
2011; Forceville and Van de Laar 2019; Zhang and Forceville 2020), I will in this chapter
discuss some of these challenges, and reflect on how to deal with them, thereby aiming to help
future analysts of metaphors in political cartoons. A staunch advocate of relevance theory
(Sperber and Wilson 1995), I hold that ultimately all communication is governed by the
relevance principle: “every act of communication comes with the presumption of optimal
relevance to its envisaged addressee” (Forceville 2020: 99). Informally reformulated for the
purposes of this chapter this means that each cartoonist tries to be meaningful to the viewers-
plus-readers of the medium that publishes the cartoon.
The chapter is organized as follows. Paragraph 2 addresses the issue of identifying and
interpreting creative metaphor in general. On the basis of the relevance theory tenet that
discussing the meaning of any message requires taking into account its target audience,
paragraph 3 will discuss what can be predicted about the specific reader-plus-viewership of a
political cartoon. The next paragraph (no. 4) will argue that a metaphor in a political cartoon is
identified and interpreted by a combination of decoding and inferring. By way of an
intermezzo, paragraph 5 critically evaluates strengths and weaknesses of VISMIP, a model
developed to account for visual metaphors. Paragraph 6 argues that metaphor scholars should
pay more attention to the difference between creative metaphors as theorized by Black (1962,
1979) and the orientational and structural metaphors typically researched within CMT. The
penultimate paragraph (no. 7) intends to shed light on the thorny question of how to identify
and categorize metaphorical source domains in corpora of political cartoons. Paragraph 8 gives
some practical advice to aspiring analysts of corpora of visual and multimodal metaphors.

2. Identifying and interpreting a creative metaphor

In Forceville (1996: Chapter 2) I built on Black’s (1962, 1979) language-oriented model to

formulate criteria for construing something as a creative metaphor in such a way that these
criteria work also for visual metaphors – paving the way for modelling metaphor in all
(combinations of) modes. In this paragraph I will expand on my (1996) chapter, slightly
reformulating the criteria to fit my current insights. For anything to be called a creative
metaphor it needs to have the following characteristics:

(1) an identity or similarity relationship is created between two elements or events that, in
the context in which they appear, belong to different categories. While in orientational and
structural metaphors this similarity has often become so entrenched in human experience
that its metaphorical nature tends to escape notice, in creative ones the similarity is
typically not pre-existent but created ad hoc;
(2) one of these elements or events can be construed as the “literal” target, and the other
as the “figurative” source, and the two are, in the context in which they appear, not
reversible;
(3) one or more features, connotations, attitudes, emotions, conventionally associated with
the source must or may be mapped onto the target, which by this mapping is transformed:
we understand this target (temporarily) via the specific perspective evoked by the source.

Let me reflect in more detail on how these characteristics function when analysing something
as a metaphor in a political cartoon. My focus will be on visual metaphors (that is: metaphors
in which the target and the source are exclusively or primarily depicted), but most of the points
made also hold for multimodal metaphors (metaphors in which the target and the source are
exclusively or primarily rendered in two or more different modes: in political cartoons the
visual and the written-language mode).
Ad (1). That the target and source belong, in the given context, to different categories can
also be expressed in terms of there being some sort of “incongruity” between them (as in Šorm
and Steen 2013, 2018). The emphasis on context is crucial, because categorizations are
typically not immutable groupings. Categorizations can be fairly stable and context-
independent, but they can also be created ad hoc. Humans, Lakoff claims, typically categorize
concepts at the basic level: “Basic-level concepts are in the middle of taxonomic hierarchies
and have a great deal of internal structure” that human beings find easy to process – that is,
easy to learn, remember, and use” (1987: 199). He points out that the basic level, which is in
between the superordinate and the subordinate level, is the level of distinctive actions, learned
and named earliest, these names being shortest and most frequently used. It is also felt to be
the most “natural” level of categorizing (Lakoff 1987: 32). Here is an example Lakoff (1987:
46) gives of how the levels are positioned vis-à-vis each other:

SUPERORDINATE ANIMAL FURNITURE

BASIC LEVEL DOG CHAIR

SUBORDINATE RETRIEVER ROCKER

Whereas conceptual metaphors pertain to the superordinate and basic levels (see the “metaphor
and metonymy index” in Kövecses 2010: 369-375), creative metaphors require the analyst to
focus also, or even primarily, on the subordinate level. Consider the following. Humans, dogs,
and cows all belong to the superordinate category of mammals. The basic level category DOG
comprises, among others, the subordinate categories of retriever, German shepherd, poodle,
and chihuahua. It should not be too difficult to imagine a situation in which we have a metaphor
functioning on the subordinate level, such as “my German shepherd is a poodle,” in which, say,
the conventional feature of slavish obedience associated with the POODLE domain is to be
mapped on the GERMAN SHEPHERD domain; or “my German Shepherd is a chihuahua,” in
which the conventional features of smallness, being prone to trembling, and perhaps silly
barking are mapped from CHIHUAHUA onto GERMAN SHEPHARD. Similarly, in a given
context it is easily conceivable to have a metaphor which should be labelled and analysed as
THIS POLITICIAN IS A POODLE rather than as THIS POLITICIAN IS A DOG. In political
cartoons, a politician can also be metaphorically compared to a member of another subcategory
of human beings, say a film director, a boxer, a captain, or a criminal. I am not sure whether
these latter source domains express basic level or subordinate level categorizations. But
however this may be, metaphors in political cartoons can definitely function at a yet more fine-
grained level, namely when a politician is compared to a specific individual. Again, one could
imagine political cartoons in which the following creative metaphors would occur: THIS
POLITICIAN IS ALFRED HITCHCOCK; THIS POLITICIAN IS MIKE TYSON; THIS
POLITICIAN IS CAPTAIN AHAB; THIS POLITICIAN IS PABLO ESCOBAR. Clearly, in such
cases specific cultural background knowledge needs to be available to the viewer (or analyst)
to be able to interpret the metaphor. That is, metaphors in political cartoons feature source
domains that often represent the subordinate level – or arguably an even lower level. CMT has
not paid much attention to metaphors that are not rooted in embodied cognition. Although
Lakoff and Turner discuss metaphors that occur in poetry, their primary concern is to
demonstrate that we can understand poetic metaphors because they are embedded in conceptual
metaphors – and they do so convincingly. But I propose that analysts of poetic and other
creative metaphors will find Black (1979) a more useful model than CMT. I will come back to
the relation between creative and conceptual metaphors in paragraph 6.
Since pictures and other visuals typically have structure, but no grammar (pace Kress and
Van Leeuwen 1996; Cohn 2013), the identity between a target and a source domain in a
metaphor involving visuals must be established by procedures that are different from those in
language. In Forceville (1996: Chapter 6) I distinguish between three types of pictorial (or
visual) metaphors. In the MP1/contextual type, one of the two metaphorical terms (usually: the
source) is suggested by the visual context in which the other term (usually: the target) is
depicted. The MP2/hybrid type, which Carroll (1996: 214) discusses in terms of “physical
noncompossibility,” visually merges target and source into a single “gestalt” that, in the given
context, cannot exist physically in the real world. The third subtype is a pictorial simile. Here
target and source are visually juxtaposed in their entirety. It is to be noted that in all cases the
resulting incongruity is not in itself enough for postulating a metaphor; rather, it is a clue that
this must/can/may be done.
All of these count as monomodal visual metaphors if no information in other modes would
be necessary to identify the target and the source domain. Multimodal metaphors differ from
monomodal metaphors, I have proposed (Forceville 2006), in that target and source are
exclusively or predominantly cued in different modes. Since print can draw on only two modes
– provided one accepts the catalogue of modes proposed in Forceville (2021) – the only kind
of multimodal metaphor that is possible in political cartoons is the verbo-pictorial variety
(Forceville 1996: 148-161). It makes sense to distinguish between monomodal visual and
multimodal metaphors of the verbo-pictorial variety, since certain groups of people (e.g., pre-
literate children and people that are not familiar with a specific language) might be able to
identify and interpret certain purely visual metaphors, but not verbo-pictorial ones.
 Ad (2) The irreversibility of the two terms is another necessary criterion to call something
a metaphor. Again, in the given context it is clear what is target and what is source (Forceville
1995, 2002). The fact that both the metaphor “that butcher is a surgeon” and “that surgeon is a
butcher” are, in principle, possible does not invalidate the irreversibility principle – simply
because, in a given context only one of the varieties can be sensibly construed. The first variety
may be whispered, say, by one butcher to another butcher about a colleague during drinks at
the Annual Butcher’s Conference, whereas the second variety is gossip more likely to be heard
at the Annual Surgeon’s Conference. If in a situation in which two phenomena are somehow
identified it is impossible to attribute target status to one of them and source domain status to
the other, this simply means they are by definition not to be construed as a metaphor – although
perhaps they can be analysed as a non-metaphorical blend (Fauconnier and Turner 2002; see
Forceville [2013] for examples of non-metaphorical blends in comics).
Ad (3). The interpretation of a metaphor amounts to deciding which feature(s) is/are to be
mapped from source to target. Incidentally, it is worth emphasizing that the conventional A IS
B template to describe metaphor fails to illuminate that it is actually the real or imagined actions
one can perform to, or with, or upon the source domain that are at the heart of the mapping.
The formula A-ING IS B-ING would thus better capture the dynamic nature of metaphorizing
(Forceville 2016: 19). Since usually the mappable features, or actions, are not explicitly
verbalized or signalled in other modes, the interpretation of most metaphors is a matter of
inferring: the addressee needs to combine intra-textual information and extra-textual
knowledge about the (sub)culture within which the metaphorical discourse at stake functions.
The freedom of interpretation of what features, or actions, are to be mapped is robustly steered
and constrained by two factors: (1) the genre to which the discourse in which the metaphor
occurs belongs; (2) the attributes, attitudes, and emotions conventionally adhering to the source
domain in the culture within which the metaphorical discourse appears. As to the former, if we
are faced with a political cartoon, we know that this genre typically urges us to search for
negative features of the source domain that can be mapped onto the target. As to the latter, the
envisaged audience needs to be aware of the qualities that, within the community in which the
metaphor is used, are associated with the source domain (i.e., the source domain’s “implicative
complex,” Black 1979: 28). That said, inasmuch as different addressees in a mass-audience
may not possess, or be able to recruit, the same background information, using a metaphor on
the one hand risks misinterpretation and on the other hand allows for a degree of interpretive
freedom. Therefore, which feature(s) is/are to be mapped from source to target is not something
that can be attested with complete certainty: “Since we must necessarily read ‘behind the
words,’ we cannot set firm bounds to the admissible interpretations: ambiguity is a necessary
by-product of the metaphor’s suggestiveness” (Black 1979: 30).
It is moreover vital to remember that conventional associations can, but need not, be
factually correct. In a discussion of MAN IS A WOLF, Black reminds us that

what is needed is not so much that the reader shall know the standard dictionary meaning of
“wolf” – or be able to use that word in literal senses – as that he shall know what I will call
the system of associated commonplaces […]. From the expert's standpoint, the system of
commonplaces [i.e., the “implicative complex,” ChF] may include half-truths or downright
mistakes (as when a whale is classified as a fish); but the important thing for the metaphor's
effectiveness is not that the commonplaces shall be true, but that they should be readily and
freely evoked” (Black 1962: 40, emphasis in original).

For good measure Black points out that in, say, a Native American culture in which wolves are
believed to be reincarnations of dead humans, the mappings from source to target in MAN IS
A WOLF would be radically different than in Western culture (Black 1962: 40) – or perhaps
even be considered a literally true claim rather than a metaphor. Here is another example of
how context affects interpretation: whereas in the decontextualized utterance “that butcher is a
surgeon” the source domain may evoke positive associated commonplaces such as “precision,”
“carefulness,” and “professionalism,” one can easily think of a situation in which a butcher
takes so much time to cut slices of exquisite Spanish jamón ibérico de bellota that a colleague
exclaims “that butcher is a surgeon” in derision rather than as a compliment. Conversely,
whereas “that surgeon is a butcher” may in de-contextualized form be a critical remark, in a
situation in which an operating surgeon needs to firmly amputate a patient’s leg, the metaphor
may well be uttered in admiration. In the last resort context, whether local, text-internal context
or supra-textual, cultural context, is crucial to retrieve a metaphor’s intended interpretation.

3. How do we know what is the envisaged audience of a cartoon?

A cartoonist aims to be relevant (which we can here for simplicity’s sake informally equate
with “meaningful”) to the audience for which he/she/+ has created the cartoon. If a cartoonist
is a regular contributor to a specific newspaper, website, blog, etc., it is to the readers of this
medium that he/she/+ aims to be relevant, which means that these readers need to understand
the critical point the cartoon makes and to find it at least somewhat witty and original. In order
to be successful, that is, the cartoonist first makes, consciously or subconsciously, an estimate
of the background knowledge about persons, objects, actions, and/or events that the envisaged
audience must be able to recruit to assess what is going on in the cartoon. That is not enough,
however. No less important is that the envisaged audience evokes the appropriate emotion,
belief, and/or attitude toward these various phenomena (think of the different beliefs
Westerners and Native Indians hold about wolves; or realize how dragons are scary monsters
in Europe while they bode good fortune in China). The audience thus needs to be able to
recognize the metaphor’s often visually represented target (person, event, scenario) as well as
source (person, event, scenario), but must also be aware of the source’s conventional
connotations to be able to select the ones that must, or can, be mapped. This latter is vital, as
the salient emotions, beliefs, and attitudes pertaining to the source domain are “co-mapped”
onto the target. Metaphor construal can thus go wrong because the viewer (1) fails to correctly
identify the target domain; or (2) fails to correctly identify the source domain; or (3) fails to
correctly identify the salient feature(s) that the cartoonist intends the audience to map from
source to target; or (4) misjudges the emotional/attitudinal value of the feature(s) selected for
mapping from source to target. Presumably the readers of a medium that regularly carries a
cartoon by the same cartoonist (say, a specific newspaper) will not, or rarely, have problems in
this respect. This readership has presumably also stored in the sum total of its background
knowledge – called “cognitive environment” by Sperber and Wilson (1995: 38) – familiarity
with the cartoonist’s style, and will thus for instance know how that cartoonist depicts a specific
politician. But digitization and social media have thoroughly changed who might come to see
a cartoon (Pedrazzini and Royaards 2022: 3). People that do not belong to the envisaged
audience may, subconsciously or intentionally, misconstrue the metaphor. Similarly, if a
cartoon migrates from its original medial context, and thus also from its envisaged audience,
to serve in a new medium, in a new socio-political cultural context, misunderstandings may
arise – at least in the sense that interpretations may differ, in minor or major ways, from how
the cartoonist intended the audience to interpret the cartoon (see also Scott 2022).

4. Identifying and interpreting metaphors in political cartoons: decoding and inferring

Informally phrased, relevance theory suggests that audiences interpret most acts of
communication by a combination of “decoding” and “inferring” pertinent information in the
message at stake. To understand a verbal message, the envisaged audience first of all needs to
understand the words and grammatical structures used. This requires that the audience is in
possession of the linguistic code – that is, knows the language’s vocabulary and syntax. I have
suggested that some visuals, although lacking a syntax and a vocabulary, require “decoding”
just as language does. Logos, pictograms and traffic signs are good examples of such visuals
(Forceville 2020: Chapter 6), which therefore yield explicit information (technically:
“explicatures” Sperber and Wilson 1995: 182). I have also proposed that it makes sense to
extend the concept of “decoding” to the recognition of represented people, objects, events, and
“scenarios” (Musolff 2016). Just as one needs to have learned the meaning of words and
syntactic structures, one may need to learn to recognize for instance politicians in cartoons
(more specifically: as drawn in a cartoon in the style of a particular cartoonist). Someone who
is unable to identify the representation of a given politician as, say, Volodymyr Zelensky or
Vladimir Putin, as Ursula von der Leyen or Marine Le Pen, thus fails at the level of decoding.
I propose that scenarios (elections, eating in a restaurant, courting, shopping, attending a
wedding etc. – see Goffman 1974), too, are to be “decoded”. But usually only part of the
content of a message depends exclusively on decoding, and can thus be “objectively”
understood. Much of what makes most messages meaningful requires the audience to process
the explicit content in interaction with other information so as to infer (as opposed to: decode)
its meaning. Non-explicit meanings are in relevance theory called “implicatures” (Sperber and
Wilson 1995: 35-37). Implicatures are inferred by a combination of textual signals, text-
internal context, pertinent text-external context, and the sum total of what constitutes a
particular addressee’s cognitive environment. Clearly, the first two of these factors are
presumably relatively stable and objectively attestable, while what is “pertinent text-external
context” is far less incontestable, if only because each individual’s cognitive environment is,
ultimately, completely unique. Consequently, not everybody will necessarily infer exactly the
same mapping(s) from source to target in a metaphorical political cartoon. The continuum
between strong implicatures (= strongly implicated meaning) and weak implicatures (= weakly
implicated meaning) in relevance theory is a very useful concept in this respect. The more
strongly an implicature is communicated, the more clear it is that the communicator (here: the
cartoonist) intends the audience to derive it, and thus takes responsibility for this derivation;
the more weakly an implicature is communicated by the cartoonist, the more the responsibility
for deriving it shifts to the audience. If a Dutch friend tells me: “I would never vote for a
populist party!” she would strongly implicate, and thus take a high degree of responsibility for
my understanding so, that she would never vote for Geert Wilders’ PVV or Thierry Baudet’s
FvD. But if I infer, from this utterance, that she would vote for Labour, I will have to take most
of the responsibility for this conclusion, as it is only weakly implicated. For present purposes,
this amounts to the following: features that are implicitly cued in the source domain of a
metaphorical political cartoon are thus on a continuum from being strongly to being weakly
implicated. Whereas strongly implicated features must be mapped from source to target to
make the cartoon relevant, weakly implicated features can be mapped. That is, the degrees of
freedom with which a given metaphor can be interpreted by different people depends on the
strength or weakness of the implicated source-to-target mappings (see also Wilson 2022).

5. Intermezzo: an evaluation of VISMIP

The previous paragraphs summarize and delve deeper into the model for analysing visual and
multimodal metaphor I have developed and refined over the past decades. It has been suggested
by Šorm and Steen (2013, 2018) that this model does not provide sufficient guidelines for the
identification of visual metaphors, motivating them to present an alternative model, called
VISMIP. This model is rooted in MIPVU (Steen et al. 2010), a method for identifying and
interpreting verbal metaphors. I agree with Šorm and Steen that it is important to be as explicit
and precise as possible in identifying visual (and multimodal) metaphors as possible, but I have
serious problems with their approach. In this paragraph I will reflect on what I see as its
shortcomings.
Šorm and Steen find fault with both Kaplan’s (2005) and my model. I will here restrict
myself to their criticisms of my work. The authors object that I “do not explicitly propose a
step in which the analyst tries to establish a general understanding of the picture. Instead, in
the first steps […] the analyst is supposed […] to indicate the two terms of metaphor (Forceville
1996, 2002, 2008)” assuming “a general understanding of the visual,” without including “a
separate, explicit step to establish that [general understanding]” (Šorm and Steen 2018: 54). I
do not think the authors do me justice here. Forceville (1996) is a monograph exclusively
devoted to visual and verbo-visual metaphors in the genre of advertising. This awareness of
genre steers and constrains the interpretation of the visuals or visuals-plus-language (Forceville
2020: Chapter 5). That is, the “general understanding” of each of the visuals discussed in
Forceville (1996) is that it is an advertisement. If we are familiar with the genre, we know that
a (commercial) advertisement always “makes a positive claim about brand X” (Forceville
1996: 104).
Figure 1. French print advertisement for Clerget Shoes presenting the visual metaphor SHOE IS TIE, ±
1985 (Forceville 1996: 110)

A key idea in MIPVU and VISMIP is that to warrant metaphor construal there must be
some sort of incongruity between two elements. Šorm and Steen claim that I do not “propose
an explicit step pertaining to the discovery of incongruous elements” (2018: 57). I find this an
unfair criticism. Here is the beginning of my analysis of the first example discussed in
Forceville (1996: 109), here reproduced as figure 1:

We see immediately that there is something odd about this shoe: it is located in a place
where we would not ordinarily have expected it. Since we trust that the communicator of
this message, the advertiser, is trying to be optimally relevant, we expect that the odd
position of the shoe is not a mere whim. We realize that the shoe is depicted in a place
where we ordinarily would have expected something else namely, a tie. Therefore this
shoe is not merely a shoe, but also suggests a tie. Put differently, the viewer is invited to
understand and perceive the phenomenon SHOE not in its usual, “literal” sense, but in
terms of the very different phenomenon TIE.

Surely, this passage amounts to “conceiv[ing] visual incongruity as visual elements that seem
distorted or out of place” (Šorm and Steen 2018: 57). The authors go on to claim that “we
should also consider the possibility that visual incongruity takes another form than a violation
of physical reality, and that more than one type of visual incongruity thus exists” (57). I agree
– which is precisely why I distinguish between the contextual, the hybrid, and the simile types
of incongruity in visual metaphors. Šorm and Steen further insist that “we need a clear set of
guidelines for deciding about when exactly visual incongruity occurs” and “about how we can
decide on the units that could be the candidates for some form of comparison” (2018: 57). In
my view, however, there simply is no failsafe procedure to decide a certain picture constitutes,
or contains, a metaphor. One reason is that the analyst of visual metaphor is not helped by a
grammar or vocabulary – a crucial resource in Steen et al. (2010) for identifying verbal
metaphor. There must thus be other reasons to construe the combination of two elements in a
picture as a metaphor. As indicated in paragraph 2, identifying a constellation of two visual
elements as a visual metaphor begins by postulating an identity relationship between two
elements that, in one way or another, is incongruous; that is, in the given context they typically
belong to different categories. So far I agree with Šorm and Steen. Whether these two elements
then must be construed as a metaphor and, if so, which of the two then is the target and which
the source, can only be resolved by the relevance principle: only by understanding one of the
elements metaphorically in terms of the other, the picture as a whole makes sense. The decision
which of the two is the target and which is the source is – at least in the case of advertisements
and political cartoons – quite reliably governed by knowledge of the genre to which the visuals
belong: advertisers want to make a positive claim about their product or brand; cartoonists want
to make a critical, preferably witty, observation about a politician or a political state of affairs
in the world. This means that in almost all cases the target domain is the product advertised (in
advertising) and a phenomenon in the domain of politics (in cartoons), respectively. The
construal of a metaphor is thus as much a top-down as a bottom-up process. After identifying
a picture (or picture-plus-text) as a political cartoon, we go on to spot some sort of incongruous
identification of two entities (“A IS B”). We then can (hopefully) resolve that incongruity by
understanding one of the two entities (the target) metaphorically in terms of the other entity
(the source) by mapping pertinent features, attitudes and emotions from the source onto the
target. All this happens, typically, in a split second. The analyst can then account for the
metaphorical construal post hoc by pointing at the details of the incongruous relation between
the two entities, and at necessary (that is: strongly implicated) and optional (that is: weakly
implicated) mappings from source to target.
To conclude this intermezzo let me take a closer look at the three case studies analysed in
Šorm and Steen (2018). In the first, a Dutch political cartoon by Tom Janssen (3 February 2003,
reproduced here as figure 2), we see the extreme-right wing member of parliament Geert
Wilders, recognizable thanks to his distinctive hairdo, sitting in a chair. The heading can be
translated as “The Islam debate,” while the text on Wilders’ chair means “direction”. I agree
with the authors that we need to construe the metaphor GEERT WILDERS IS A DIRECTOR
here. The incongruity resides, indeed, in Geert Wilders being seated on a typical “director’s
chair,” and having a megaphone, an attribute stereotypically associated with directors, lying
next to him. In the view of the cartoonist, the politician Wilders thus “directs,” that is,
dominates, the Islam debate instead of being one of the participants in this debate. In my model,
the first step would be to register the incongruity between “being a politician” and “being a
director”. Given that this is a political cartoon it is highly likely that GEERT WILDERS is the
target domain and DIRECTOR the source domain – indeed reversing these would not make
sense. Viewers need to recruit from their background knowledge that a director is (supposedly)
the boss on the film set – Black would discuss this in terms of being part of the “implicative
complex”/”associated commonplaces” of the DIRECTION domain – and that this is the feature
that the cartoonist intends us to map onto Wilders, as without this mapping the metaphor, and
indeed the cartoon, is not relevant to the envisaged audience. While a bossy director is more or
less acceptable on a film set, members of parliament, by contrast, have to debate on a basis of
parity, and therefore cannot one-sidedly impose their views. Consequently, the bossiness of the
director acquires negative connotations in the mapping onto Wilders, and it is this that gives
the cartoon its genre-related critical perspective. Incidentally, while Šorm and Steen (2018)
mention the appearance of the text “direction” on the chair in passing, they fail to take it into
account in their analysis. For viewers who are not familiar with the attributes of a prototypical
film director (chair, megaphone), the word “direction” is crucial for identifying the metaphor’s
source domain, and for these viewers the metaphor would thus be multimodal rather than
monomodally visual – a distinction that is crucial in my work but that Šorm and Steen (2013,
2018) do not make.
Figure 3, a cartoon by Arend van Dam, portrays an anxious-looking man in front of a
computer screen, caught in a web. A big menacing spider crawls toward him, presumably about
to attack and eat him. By and large I agree with Šorm and Steen’s analysis that “the man and
computer in combination with the verbal text ‘privacy’ [underneath the man, ChF] are likely
to represent ‘on line privacy’. […] the spider accompanied by the verbal text ‘WWW’ [on the
spider, ChF] is likely to represent the World Wide Web” (2018: 76). Rephrased in terms of my
model, the incongruity of the man-with-computer in a spider’s web alerts us to construe a
metaphor that could be verbalized as MAN-WITH-COMPUTER IS FLY CAUGHT IN A WEB,
while the WORLD WIDE NETWORK IS A SPIDER’S-WEB. But note that the textual
components (“privacy” and “WWW”) are likely to be necessary for many viewers to interpret
the cartoon. Whereas the visual “web” is, in the context of a political cartoon, presumably
interpreted as the source domain coupled with the target domain “World Wide Web” rather
than vice versa even without the letters “WWW,” the word “privacy” surely cannot be
dispensed with for correct interpretation of the metaphor. Without “privacy” we might still
understand that the WWW threatens computer users, but the nature of the threat could be of
another kind, for instance, the WWW might be seen as being a pervasive usurper of computer
users’ time and attention. A more complete verbalization of the metaphor would thus be
COMPUTER-USER’S PRIVACY IS A FLY CAUGHT IN (WORLD WIDE) WEB. In this view, the
verbal component thus importantly contributes to the construal of the metaphor – which, again,
is thus a multimodal rather than a monomodally visual metaphor.

Figure 2. Cartoon by Tom Janssen, 3 Figure 3. Cartoon by Arend van Dam, date
February 2003 (Šorm and Steen 2018: 48) unknown (Šorm and Steen 2018: 75)

Figure 4 is a print advertisement described by Šorm and Steen as “Slices of toasted

bread pop up from Bolletje toasted bread package” or “Slices of toasted bread in Bolletje
toasted bread package” (2018: 79). The authors conclude that because there is no incongruity
“there are no elements to be marked for metaphor” (2018: 81). This conclusion is in my view
too quickly drawn. While it is possible to consider figure 4 as completely “literal” depiction,
the authors’ first description, including the verb “pop up” already suggests that one can also
consider this depiction metaphorically, verbalizing the metaphor as BOLLETJE TOASTED
BREAD PACKAGE IS POP-UP TOASTER, on the basis of an identification of a Bolletje toasted
bread package and a pop-up toaster. This metaphorical reading would be triggered by the
incongruity of two pieces of toast “floating” in the air halfway above a bread package, defying
the laws of gravity. The mapped feature would be “yielding fresh toast,” the claim being
something like: Bolletje toasted bread comes out its package in such a fresh and crisp manner
that it is as if it is toasted on the spot. Since this metaphor can be construed (and my hunch is:
was meant to be construed) this example shows that incongruity may be subtly conveyed, and
not registered as such by all viewers. Indeed, while incongruity characterizes most metaphors,
it is not a necessary criterion for metaphor (Lankjær 2016: 119; Forceville 2016: 25-26). If
VISMIP rules out examples like this, it provides an impoverished view of what is, or can be,
metaphor.

Figure 4. Advertisement for Bolletje toasted bread, date unknown (Šorm and Steen 2018: 79 )

6. Conceptual versus creative metaphor?

Lakoff and Johnson (1980) do acknowledge that not all metaphors are conceptual, embodied
ones. In Chapter 21 they discuss the metaphor LOVE IS A COLLABORATIVE WORK OF ART,
which in Black’s (1979) model would be called a creative one. But in their book this is an aside.
One of the downsides of the enormous popularity of the CMT paradigm since the publication
of Lakoff and Johnson (1980) is that the kind of creative metaphors theorized in Black’s (1979)
“interaction theory” has, outside of my own work, for decades been virtually ignored (but cf.
Müller and Kappelhoff 2019). For analysts of metaphor the distinction is an important one, but
Lakoff and Turner’s (1989) claim that many creative metaphors are manifestations of
underlying conceptual ones has blurred rather than clarified the differences. It is of course true
that the creative metaphor GEERT WILDERS IS A POODLE is a manifestation of the structural
metaphor PEOPLE ARE ANIMALS, but the latter verbalization does not even begin to capture
what, in the context in which it is used, makes this metaphor meaningful (for some more
discussion about the difference between conceptual and creative metaphor, see Forceville
2016).
 Now the analysist of corpora of metaphors in political cartoons may run up against the
following problem. Cartoonists want to be original and not simply copy the metaphors that
fellow cartoonists have used, and thus have an interest in finding unique source domains. By
contrast, corpus-oriented metaphor scholars want to find patterns in metaphor use, and thus are
typically interested in recurring source domains. But the source domains used by cartoonists
to provide a perspective on a given target domain in metaphorical political cartoons are not
always unique, while different cartoons drawing on the same, limited set of conceptual source
domains do not necessarily lack originality. How can this conundrum be resolved?
We need to realize that the success of a good political cartoon depends to a considerable
extent on the specific style of depiction and the specific context of the scenario created by the
artist. A metaphor that could be characterized at the high level of PEOPLE ARE ANIMALS only
becomes witty and appealing at the level of WILDERS IS A POODLE. Some metaphors are
analysable simultaneously at different levels of abstraction, while others are better analysed
either at the general, higher level or at the specific, creative level. To pursue an earlier made-
up example: POLITICIAN X IS CAPTAIN AHAB may be analysed as a creative metaphor (which
will require us to complete the source domain scenario by, minimally, finding a target-domain
equivalent for MOBY DICK in the source domain scenario) or as a manifestation of the
conceptual POLITICAL ACTIVITY IS A JOURNEY metaphor (on the basis that in the scenario
in which there is a captain, there must be a ship, and therefore an implied journey, with a goal).
Metaphors that draw on a conceptual metaphor may, of course, be creative, too. As an example,
Lakoff and Johnson mention “imaginative expressions” of the THEORIES ARE BUILDINGS
metaphor such as “his theory has thousands of little rooms and long, winding corridors” and
“he prefers massive Gothic theories covered with gargoyles” (1980: 53). So, it will depend on
the analyst’s research question at what level of abstraction metaphors in cartoons (or other
genres) are to be analysed (for a similar way of reasoning about the “right” level of abstraction
in verbalizing metaphors, see Musolff [2016: 13]).

7. How to identify and categorize metaphorical source domains when analysing political
cartoons?

When in political cartoons a politician or other well-known figure is portrayed as if he/she/+

were an identifiable other (kind of) person, or as engaged in an activity which is not typical of
what politicians routinely do – namely, say, doing their work as politicians – this is a strong
cue that the politician is portrayed metaphorically. When analysts examine a corpus of
metaphorical cartoons, one goal is presumably to chart the frequency with which a specific
source domain is recruited to depict a politician, or a political state of affairs, from a certain
perspective – the latter usually supposed to evoke a negative evaluation. This requires the
analyst to take several decisions.
Since the question at stake is typically how a certain politician or political event is
characterized, the target domain tends to be fairly unambiguously attestable as somehow
pertaining to the realm of politics. By contrast, the issue which source domains are to be
distinguished, and how they are to be verbalized is a much more thorny problem. As we have
seen above, a source domain can be categorized at various levels of abstraction. On the whole,
the larger the corpus of data analysed, the more sense it makes to allow for fine-grained
distinctions between source domain categories, as a larger data set makes it more probable that
a given category has a substantial number of items, whose inventory thus is worthwhile and
may even yield statistical significance. Let me revisit how this issue was approached, implicitly
or explicitly, in Bounegru and Forceville (2011), Forceville and Van de Laar (2019) and Zhang
and Forceville (2020).
In Bounegru and Forceville (2011), we selected political cartoons from all over the world
via the Cagle cartoon website using “financial crisis” as the key word and restricting the time
slot to 15-31 October 2008. This yielded 30 cartoons, of which we decided that 21 contained a
metaphor (confusingly and mistakenly the number 25 also appears in the paper – a remnant
from an earlier draft, when we identified four cartoons as metaphorical that we later discarded).
We resolved to label the target domain FINANCIAL CRISIS, which was typically visually
represented by a metonym (e.g., a bank, or a banker, or a $/€/¥ sign). We postulated the
following source domains: (a) CATASTROPHE/(NATURAL) DISASTER (10x); (b) ILLNESS/
DEATH (4x); (c) BEGGING (4x); OTHERS (3x). We developed the following criteria to identify
these domains. To categorize a cartoon as (a), the crisis had to be visually presented “in terms
of a potentially lethal threat that was caused by forces of nature or by other circumstances
beyond the control of human beings” (Bounegru and Forceville 2011: 214); to categorize it as
(b) the crisis had to be understood “in terms of a potentially lethal illness, and of (the risk of)
death brought about by causes other than the forces of nature” (217); and under (c) “we grouped
cartoons in which the need for money (itself a metonym for everything one can buy with it)
caused by the crisis is presented in terms of begging for money or food” (218). For present
purposes, it is important to emphasize that several of the metaphors involved a form of
personification. Moreover, to be able to assess that a given person was, specifically, a banker
or a businessman we relied on the visual presence of mobile phones, cigars and suitcases as
stereotypical accoutrements of these professionals.
Forceville and Van de Laar (2019) hypothesized that analysing metaphors in cartoons
featuring the Dutch populist politician Geert Wilders would yield a good impression of
Wilders’ public image. We expected many of the metaphors to deploy the same source domain.
Our corpus consisted of cartoons published in “the best Dutch political cartoon of the year”
books published between 2006 and 2017. To enable assessing how unique (or not so unique)
Wilders’ metaphorical portrayals were, we checked the metaphors used in cartoon depictions
of all politicians who were party leaders of a Dutch political party at the same time as Wilders.
Admittedly, we did not explicitly discuss the procedure of identifying a given constellation as
metaphoric – but what we did do was score a politician as metaphorically represented if that
politician was depicted (or verbally labelled) as being, or acting, in a role different than that of
a politician, thereby resulting in an “incongruous” characterization. Sometimes this was done
by creating a “hybrid,” for instance in the form of an animal with a specific politician’s head,
but more often the politician appeared in an incongruous “scenario,” performing actions not
associated with the profession of being a politician – which would, in my terminology, be a
contextual metaphor (as in figure 2). In our corpus, Wilders was depicted more often than any
other party leader, and also depicted more often metaphorically than any colleague. Somewhat
to our surprise, in the 72 metaphorical depictions of Wilders, (only) nine source domains
occurred more than once: FOULNESS-EXUDING CREATURE – the “foulness” comprising
feces, vomit, and inappropriate language (5x), BOMB (3x); DOG (3x); BABY/CHILD (3x);
HELMSMAN (2x); PUPPET PLAYER (2x); SPOUSE (2x); HORROR-FILM MONSTER (2x); and
CRUSADER (2x). Thus, Wilders was depicted metaphorically no less than 48 times by means
of source domains that occurred only once in our corpus.
Here are some other thoughts. First, it needs to be acknowledged that the FOULNESS-
EXUDING CREATURE source domain may be considered a bit of a stretch. This is a reminder
that metaphors are, ultimately, construed by viewers and analysts, and that not everybody
would agree on metaphorical construal in a given case. The somewhat strained formulation of
this source domain was chosen to enable us to group feces, vomit, and foul language as
manifestations of a source domain that could, thereby, be labelled at a higher level of
abstraction. Second, of the nine recurring source domains, only FOULNESS-EXUDING
CREATURE, BOMB, PUPPET PLAYER, and HORROR-FILM MONSTER were exclusively used
for Wilders, and not for any other politicians (in total we found 244 metaphorical depictions of
party leaders, including the 72 of Wilders). This shows how important it was to compare the
metaphorizations of Wilders with those of other politicians, as several source domains were
not unique for Wilders. Third, in order to say something about the characterization of a
politician, only making an inventory of the metaphorical source domains deployed in the
cartoons is not enough; it is no less important to consider the feature(s) that is/are mapped from
source to target. Whereas in the case of FOULNESS-EXUDING CREATURE, BOMB, HORROR-
FILM MONSTER, these are typically always negative ones, a DOG can be aggressive, obedient,
as well as loyal and protective, while a CHILD/BABY can be irritating, innocent, naïve, as well
as vulnerable and cute. In various cartoons featuring POLITICIAN IS DOG or CHILD, then, not
necessarily the same feature from the DOG/CHILD domains was mapped in each case. In short:
in the last resort, the specific context of the metaphor in the cartoon needs to be taken into
account by the analyst when making claims about the nature of the source-to-target mappings.
Zhang and Forceville (2020) analysed a corpus of 69 Chinese and 60 American political
cartoons that had as their topic the Sino-US trade conflict. The corpus was compiled by using
the time frame 1 March 2018 – 31 December 2019 on the Cagle cartoon website, selecting the
“USA” option, and then further filtering the results by using the keywords “US,” “Chinese
economy,” and “tariff”/“tariff war.” We followed a similar procedure for the China Daily
website.
We began by examining the cartoons first separately and then together without any
formal identification procedure. We focused on the visuals, taking into account verbal
information only when necessary to identify visually represented phenomena. We then grouped
the metaphors we found in four dominant scenarios (that is: scenarios that occurred at least five
times in the 129 cartoons): the target domain “trade conflict” was presented in terms of the
following source domains: (1) a violent confrontation between China and the USA; (2) a sports
match, game, or play involving the two countries; (3) a journey that was in one way or another
impeded, inconveniencing one or both of the countries; (4) a natural disaster. We decided to
label these metaphor scenarios as follows: (1) TRADE CONFLICT IS WAR/FIGHT; (2) TRADE
CONFLICT IS GAME/SPORTS/PLAY; (3) TRADE CONFLICT IS INTERRUPTED JOURNEY;
and (4) TRADE CONFLICT IS NATURAL DISASTER. We initially agreed only on the
categorization of about two thirds of the metaphor scenarios. The mismatches turned out to be
due to the following reasons: some instances were scored as belonging to scenario 1 by one of
us and to scenario 2 by the other; one of us had not identified scenario 4 at all; and sometimes
more than one metaphor occurred in a cartoon. As for the last issue, when a cartoon that, in our
first round, yielded two or more metaphors, we decided, after discussion, which was the
dominant metaphor. We moreover finetuned our criteria for categorization as follows. We
identified metaphor scenario 1 when at least one of the two parties in the conflict (often
depicted metonymically; see Zhang & Forceville 2020 for details) used a weapon which was
aimed at the other party or when there was wrestling or kicking between them. Scenario 2 was
considered to apply when a typical sports/game/play attribute (e.g., a tennis racket), location
(e.g., a golf court) and/or action (e.g., juggling) was visually depicted. Scenario 3 required the
visual presence of a vehicle (e.g., a car, a boat, an air balloon) and/or a path (e.g., railroad
tracks, a labyrinth) in a situation in which physical progress was impossible, dangerous, or
seriously hindered. Scenario 4 was deemed pertinent when the consequences of some force of
nature were mapped onto the trade conflict domain. These criteria helped us considerably to
account for, and agree among ourselves on, our categorizations. That said, it is important to
observe that we restricted our categorizations to the four scenarios above as these were the only
ones that had at least five manifestations. Conceptual metaphors that surfaced less than five
times were TRADE CONFLICT IS BUILDING/BURDEN/FALLING DOWN and THE ECONOMY
IS PLANT/MACHINE, while the personification metaphor WORLD IS PERSON (only in the
Chinese cartoons) cross-cut all these categories. Just as in Forceville and Van de Laar (2019),
moreover, there was a whole range of metaphors that appeared only once or twice and verged
toward being creative in the sense of Black (1979). Examples include TRADE CONFLICT IS
OPENING FORTUNE COOKIES; ECONOMY IS SANTA CLAUS; TRADING PARTNERS ARE
(POTENTIAL) LOVERS. However, it needs to be realized that (much) larger corpora might have
yielded more instances of such metaphors.
To do more justice to the overall meaning of the cartoons, finally, we also discussed other
meaning-making mechanisms besides metaphor and metonymy, specifically intertextuality,
idioms and expressions, themes and topoi, and comics conventions.

8. To conclude: some practical advice to aspiring analysts of metaphors in political

cartoons

There is no simple procedure for identifying, labelling, and interpreting visual and multimodal
metaphors in political cartoons (or in other genres and media, for that matter). We need to be
reminded here of Black’s warning that “there is […] a sense of ‘metaphor’ that belongs to
‘pragmatics’ rather than to ‘semantics’ – and this sense may be the one most deserving of
attention” (Black 1962: 30). That said, let me here summarize the insights presented in this
chapter in the form of some recommendations.
 Cultural provenance of the cartoons. Only analyse cartoons from a language and culture of
which you have native or near-native knowledge. Recognizing intertextual references (ranging
from Moses to Zeus, from Superman to Kim Kardashian, and from the Holocaust to Game of
Thrones) may be indispensable for identifying and interpreting a metaphor. A complicating factor
is that the connotations of source domains may change over time – think of the Twin Towers pre-
and post-9/11 and the reputation of famous people before and after revelations of their
misconduct. Furthermore, the implications of the date of publication must be pondered: if a
cartoon is published in response to a newsworthy event, cartoonists will automatically assume
that this event is foremost in their audiences’ minds, so that they need not verbally or visually
spell out this event in their cartoon. But with the passing of time, such knowledge and awareness
inevitably fade, which may require the analyst to delve into archives (and necessitate glosses in
scholarly papers discussing these cartoons).
Access to the corpus. Ideally, the cartoons in the corpus are examined in their original
context. Often, this is practically impossible, and cartoons need to be collected digitally. It is
important to remain aware of the medium in which the cartoons were first published, as this tells
a lot about their envisaged audience. Make sure you retain access to your entire corpus. In the
case of digital data, it is advisable to download, and properly tag, all cartoons straightaway.
Target and source domains. Opt for a very limited number of target domains, preferably just
a single one, and verbalize each target domain in such a way that it can capture a fairly broad
variety of manifestations and simultaneously ensures your research project has a clear focus.
Think carefully about which source domains you are going to use to categorize the metaphors,
and at what level of abstraction these source domains are going to be examined and labelled.
Deciding on source domain categories as well as their verbalizations needs as much as possible
to be done on the basis of intersubjective criteria: “if X (or X and Y) is/are visually (or verbally,
in the case of multimodal varieties) present in a metaphorical source domain, it is to be classified
as belonging to category I; while if S (or S and Z) are visually (or verbally, in the case of
multimodal varieties) present in a metaphorical source domain, it is to be classified as belonging
to category II”. Decide how you are going to deal with situations in which two or more metaphors
occur in a single cartoon.
Verbalization of the metaphor. Verbalize metaphors in the conventional TARGET A IS
SOURCE B form, as only this enables counting the frequency with which a specific source
domain occurs. That said, never forget that a metaphor can be verbalized in different ways, but
that each verbalization steers interpretation: there are subtle but important differences between
ARGUMENT IS WAR, ARGUMENT IS A BATLLE, and ARGUMENT IS A FIGHT; and between
TRADE CONFLICT IS A SPORTS MATCH, TRADE CONFLICT IS A COMPUTER GAME, and
TRADE CONFLICT IS PLAY. The larger the corpus, the more it makes sense to make more fine-
grained distinctions. And of course be aware that metaphors invite mappings of activities rather
than of attributes adhering to SOURCE B onto TARGET A.
What, in cartoons drawing on the same domain category, is mapped? While deciding what
is target and what is source suffices for identifying a metaphor (and grouping metaphors with the
same source domain), interpreting a metaphor means deciding which feature(s), emotion(s), and
attitude(s) can/are to be mapped from source to target – and a given source domain may cue
different mappings in different contexts (and to some extent: by different audiences; see Plug
[2017] for examples of contested metaphors in political cartoons).
Non-metaphorical meaning. Even though metaphors usually capture the key meaning of
political cartoons, there often are other things going on that contribute to meaning-making. There
are two reasons for keeping this in mind. In the first place, being strict about what part of the
overall meaning is, and what part is not, captured by the metaphor ensures that “metaphor”
remains a well-defined and useful analytical tool. In the second place, this awareness will alert
metaphor aficionados that there is more to meaning than just metaphor.
Clearly, much more thinking needs to go into fine-tuning how visual and multimodal
metaphors can be analysed in political cartoons, but hopefully this chapter provides helpful
starting points. I look forward to further developments in this field.

Acknowledgment. I am indebted to two anonymous reviewers for providing pertinent comments

and criticisms on an earlier draft of this chapter.

References

Abdel-Raheem, Ahmed. 2019. Pictorial Framing in Moral Politics: A Corpus-Based

Experimental Study. New York: Routledge.
Al-Masri, Hanada. 2016. Jordanian editorial cartoons: A multimodal approach to the cartoons
of Emad Hajjaj. Language and Communication 50. 45–58.
Bergen, Benjamin. 2003. To awaken a sleeping giant: Cognition and culture in September 11
political cartoons. In Michel Achard & Suzanne Kemmer (eds.), Language, Culture, and
Mind, 23–35. Stanford: CSLI Publications.
Black, Max. 1962. Models and Metaphors: Studies in Language and Philosophy. Ithaca NY:
Cornell University Press.
Black, Max. 1979. More about metaphor. In Andrew Ortony (ed.), Metaphor and Thought, 19-
43. Cambridge MA: Cambridge University Press.
Bounegru, Liliana & Charles Forceville. 2011. Metaphors in editorial cartoons representing the
global financial crisis. Visual Communication 10. 209-229.
Cohn, Neil. 2013. The Visual Language of Comics: Introduction to the Structure and Cognition
of Sequential Images. London: Bloomsbury.
Đurović, Tatjana & Nadežda Silaški. 2016. Multimodality and the construal of reality in
political cartoons: The case of Serbia-EU relationship. Facta Universitatis, Series:
Linguistics and Literature 14(2). 117–128.
El Refaie, Elisabeth. 2003. Understanding visual metaphors: The example of newspaper
cartoons. Visual Communication 2. 75–95.
Fauconnier, Gilles & Mark Turner. 2002. The Way We Think: Conceptual Blending and the
Mind’s Hidden Complexities. New York: Basic Books.
Forceville, Charles. 1995. (A)symmetry in metaphor: the importance of extended
context. Poetics Today 16. 677-708.
Forceville, Charles. 1996. Pictorial Metaphor in Advertising. London: Routledge.
Forceville, Charles. 2002. The identification of target and source in pictorial metaphors.
Journal of Pragmatics 34. 1-14.
Forceville, Charles. 2006. Non-verbal and multimodal metaphor in a cognitivist framework:
Agendas for research. In Gitte Kristiansen, Michel Achard, René Dirven & Francisco Ruiz
de Mendoza Ibàñez (eds.), Cognitive Linguistics: Current Applications and Future
Perspectives, 379-402. Berlin: Mouton de Gruyter.
Forceville, Charles. 2008. Metaphor in pictures and multimodal representations. In Raymond
W. Gibbs, Jr. (ed.), The Cambridge Handbook of Metaphor and Thought 462-482.
Cambridge: Cambridge University Press.
Forceville, Charles. 2013. Creative visual duality in comics balloons. In Tony Veale, Kurt
Feyaerts & Charles Forceville (eds.), The Agile Mind, 253-273. Berlin: Mouton de Gruyter.
Forceville, Charles. 2016. Visual and multimodal metaphor in film: charting the field. In
Kathrin Fahlenbrach (ed.), Embodied Metaphors in Film, Television and Video Games:
Cognitive Approaches, 17-32. London: Routledge.
Forceville, Charles. 2020. Visual and Multimodal Communication: Applying the Relevance
Principle. Oxford: Oxford University Press.
Forceville, Charles & Nataša van de Laar. 2019. Right-wing politician Geert Wilders
metaphorized in Dutch political cartoons. In Encarnación Hidalgo Tenorio, Miguel Ángel
 Benitez Castro & Francesca De Cesare (eds.), Populist Discourse: Critical Approaches to
Contemporary Politics, 292-307. London: Routledge.
Goffman, Erving. 1974. Frame Analysis: An Essay on the Organization of Experience.
Cambridge: Harvard University Press.
Kaplan, Stuart J. 2005. Visual metaphors in print advertising for fashion products. In: Kenneth
L. Smith, Sandra Moriarty, Keith Kenney & Gretchen Barbatsis (eds.), Handbook of
Visual Communication: Theory, Methods, and Media, 167–177. Mahwah, NJ: Lawrence
Erlbaum.
Kress, Gunther & Theo van Leeuwen. 2006. Reading Images: The Grammar of Visual Design
(2nd edn.). London: Routledge.
Kwon, Iksoo. 2019. Conceptual mappings in political cartoons: A comparative study of the
case of nuclear crises in US-North Korean relations. Journal of Pragmatics 143. 10–27.
Lakoff, George. 1987. Women, Fire and Dangerous Things: What Categories Reveal about the
Mind. Chicago: University of Chicago Press.
Lakoff, George & Mark Johnson. 1980. Metaphors We Live By. Chicago: University of
Chicago Press.
Lakoff, George & Mark Turner. 1989. More than Cool Reason: A Field Guide to Poetic
Metaphor. Chicago: University of Chicago Press.
Lan, Chun & Danyu Zuo. 2016. Pictorial-verbal metaphors in Chinese editorial cartoons on
food safety. Metaphor and the Social World 6(1). 20–51.
Lankjær, Birger. 2016. Problems of metaphor, film, and visual perception. In Kathrin Fahlenbrach
(ed.), Embodied Metaphors in Film, Television, and Video Games, 115-128. London:
Routledge.
Lin, Tiffany Ying-Yu & Wen-Yu Chiang. 2015. Multimodal fusion in analyzing political
cartoons: Debates on U.S. beef imports into Taiwan. Metaphor and Symbol 30. 137–161.
Marin-Arrese, Juana I. 2019. Political cartoon discourse: Creativity, critique and persuasion.
Cultura, Lenguaje y Representación 22. 117–134.
Müller, Cornelia & Hermann Kappelhoff. 2018. Cinematic Metaphor: Experience – Affectivity
– Temporality. Berlin: De Gruyter.
Musolff, Andreas. 2016. Political Metaphor Analysis: Discourse and Scenarios. London:
Bloomsbury.
Negro Alousque, Isabel. 2013. Visual metaphor and metonymy in French political cartoons.
Revista Española de Lingüística Aplicada 26. 365–384.
Pedrazzini, Ana & Tjeerd Royaards. 2022. Cartoons on trial: a case study integrating
discursive, legal and empirical perspectives. Humor, ahead of print (25 pp).
https://doi.org/10.1515/humor-2022-0016
Plug, José H. 2017. Attacks on the cartoonist’s strategic manoeuvring: An argumentative
analysis of criticism on political cartoons. In: Cornelia Ilie and Guliana Garzone (eds),
Argumentation across Communities of Practice. Multi-disciplinary Perspectives, 317-
337. Amsterdam: Benjamins.Prendergast, Muireann. 2019. Political cartoons as
carnivalesque: A multimodal discourse analysis of Argentina’s Humor Registrado
magazine. Social Semiotics 29(1). 45–67.
Schilperoord, Joost & Alfons Maes. 2009. Visual metaphoric conceptualization in editorial
cartoons. In Charles Forceville and Eduardo Urios-Aparisi (eds.), Multimodal Metaphor,
213-240. Berlin: Mouton de Gruyter.
Scott, Kate. 2022. Pragmatics Online. London: Routledge.
Silaški, Nadežda & Tatjana Đurović. 2019. The journey metaphor in Brexit-related political
cartoons. Discourse, Context & Media 31. 1–10.
Šorm, Ester & Gerard Steen. 2013. Processing visual metaphor: A study in thinking out loud.
Metaphor and the Social World 3. 1–34.
Šorm, Esther & Gerard Steen. 2018. VISMIP: Towards a method for visual metaphor
identification. In Gerard Steen (ed.), Visual Metaphor: Structure and Process, 47-88.
Amsterdam: Benjamins.
Sperber, Dan & Deirdre Wilson. 1995). Relevance: Communication and Cognition (2nd edn.).
Oxford: Blackwell.
Steen, Gerard J., Aletta G. Dorst, J. Berenike Herrmann, Anna A. Kaal, Tina Krenmayr &
Trijntje Pasma. 2010. A Method for Linguistic Metaphor Identification: From MIP to
MIPVU. Amsterdam: Benjamins.
Wawra, Daniela. 2018. Multimodal literacy: Meaning negotiations in political cartoons on the
refugee crisis. System 77. 10–18.
Wilson, Deirdre. 2022. Communication, comprehension, and interpretation. In Herbert L.
Colston, Teenie Matlock, and Gerard J. Steen (eds.), Dynamism in Metaphor and Beyond,
143-155. Amsterdam: Benjamins
Zhang, Cun & Charles Forceville. 2020. Metaphor and metonymy in Chinese and American
political cartoons (2018-2019) about the Sino-US trade conflict. Pragmatics & Cognition
27. 476-501.