Skip to main content
Volume 18

Debates in Aesthetics, Vol. 18, No. 1

Table of Contents

Sarah Kiernan and Claire Anscomb
Read online | Download Pdf

Alessandro Cavazzana
Read online | Download Pdf

Kiyohiro Sen
Read online | Download Pdf

Ian Verstegen
Read online | Download Pdf

Rebecca Wallbank
Read online | Download Pdf

Fotini Charalabidou
Read online | Download Pdf

Sarah Kiernan and Claire Anscomb

↑ Back to top ↑

It is our great pleasure to introduce the 2022 general issue of Debates in Aesthetics (DiA). In this issue there are three original articles, two of which examine issues related to metaphor as understood by Donald Davidson (Cavazzana) and Conceptual Metaphor Theorists and Gestalt Psychologists (Verstegen), while another examines how appreciative behaviour comes to be associated with different categories of art (Sen). There is also an interview with poet Ralf Webb, where he discusses his debut collection Rotten Days in Late Summer and the relationship between authors and readers (Wallbank), and a review of the recently published ‘Looking Through Images: A Phenomenology of Visual Media’ examining how Emanuel Alloa answers the question “what is an image?” (Charalabidou).

In ‘The Dreamwork of Language’, Alessandro Cavazzana offers a fresh approach to Donald Davidson’s 1978 paper What Metaphors Mean. Davidson’s rejection of metaphorical meaning is illuminated, Cavazzana suggests, by placing it within the broader theoretical framework of Davidson’s philosophy of language. He argues that Davidson’s proposal that metaphors do not have any special cognitive content distinct from their literal meaning can be better understood by sufficiently considering Davidson’s distinction between the semantic and pragmatic aspects of language (2022, 15). The central claim here is that the imaginative visions and associations that are stimulated by metaphors are solely the result or domain of the way in which language is used rather than any inherent meaning of language. This conclusion clearly rests on a particular definition of ‘meaning’ and, as Cavazzana explains, this definition originates from a linguistic theory in which meaning and truth are intimately connected. This, of course, is truth-conditional semantics – the position that the “conditions under which a sentence would be true is a way of indicating the meaning of that sentence” (2022, 17). This leads Cavazzana to note the influence of Alfred Tarski’s theory of truth on Davidson’s semantics and explain how Davidson believes that Tarski’s equivocation of truth biconditionals with linguistic meaning can be applied to natural languages as well as formal languages if they are modified to include the context of utterance (2022, 19).

Metaphors seem to present the gravest challenge for the application of truth-conditional semantics to natural languages but, according to Cavazzana, Davidson addresses this challenge by appealing to the inclusion of the context of utterance and construing metaphors as the ultimate examples in which this context takes effect. In short, it is the truth conditions that give a sentence its meaning (not whether or not those conditions are met), and in the case of metaphors where truth conditions are certainly not met, the obvious falseness of the statement leads the listener to look for some other purpose for its utterance. Thus, the allegorical ‘meaning’ of metaphors is solely derived from the context in which they are used and metaphors cannot be said to possess any special cognitive content beyond the literal, according to this theory of language. Cavazzana writes that his assessment of Davidson’s paper differs from analyses which fail to sufficiently consider Davidson’s views on semantics or align his treatment of metaphors with his wider work on the theory of meaning (2022, 14). Such a treatment shows that Davidson’s views are in fact consistent and that his work on metaphors is in no way incongruent with his wider work on semantics.

Sen starts his article by recounting Kendall Walton’s famous argument that there are correct categories for artworks and that the correct appreciation of artworks is based on such categories (2022, 32). As Sen outlines, while Walton offered some factors that count towards it being correct to perceive a work in a particular category, such as which category the artist intended or expected the work to be perceived in, none of these seem to provide necessary or sufficient conditions and nor is it clear how categories originate. Acknowledging the difficulties in addressing the membership question (what determines the correct category of a work?) Sen proposes a more fruitful line of enquiry is to focus on the behavioural question: how are the correct appreciative behaviours related to a category established? In doing so, “the correct category as status would be an incidental outcome of the behaviours in question” (2022, 36). To answer the behavioural question, Sen draws on Francesco Guala’s characterization of institutions as rules-in-equilibrium. Guala’s work explains the aspects of institutions that normatively guide behaviours and incentivize their being followed. To demonstrate that institutions are rules and equilibria, Guala employs the concept of ‘correlated equilibrium’, which is equilibrium, where no one would be better off by deviating alone, based on external signals observed by those involved.

To explain artwork categorization in these terms, Sen assumes that (1) each critic “has an incentive to categorize an artwork in a way that deepens one’s understanding of the work” (2022, 38) and (2) each critic has an incentive to choose the same categorization as others “because working together to criticize a work based on a particular category helps discover meanings and values that one would not notice on their own” (2022, 38). As Sen highlights, while coordination can be easily achieved with a new work that can be assigned to a salient, socially well-established category, this is not so easy with pioneering works, which call for pioneering ways of appreciation. To describe the origin of a category, Sen proposes two phases: the ‘institutionalizing phase’, where critics “attempt various appreciative behaviours” (2022, 39) and the ‘institutionalized phase’ where a certain “strategy is established as the correct appreciative behaviour for the artwork” (2022, 39). By conceiving of correct categories as institutional outputs, Sen proposes that we can account for the stability of categories in addition to the possibilities and difficulties of critical innovation: when institutions are revised this may have been prompted by unusual appreciative behaviours; appreciators need not rely on linguistic abilities to evaluate a work if they recognize what to do; and while the correct meaning and value of an artwork are relative to its history and community of reception, the correct categories are contingent.

Importantly, with this shift in focus, Sen’s work provides a useful framework for addressing questions about whether the appreciative behaviours related to a particular category are appropriate and even ethical. As history shows us, sometimes the ‘correct’ appreciative behaviours developed towards certain kinds of works are formed on the basis of widespread misunderstandings and biases about particular practices which ultimately fail to deepen understanding of the works. Take the case of a categorization like ‘primitive art’ that became popular in the early 20th century. As time has made clear, not only did the appreciative behaviours associated with this category fail to produce a deeper understanding of the works, but they also manifested and perpetuated unethical attitudes towards those who made and were associated with the cultures where the works originated from. So, by focusing on addressing the behavioural question, rather than the membership question Sen’s account offers a valuable new perspective on how categories originate and develop so that certain appreciative standards take hold. In particular, it helps to make sense of how works, which retain all their features, may end up being categorized differently over time, as a result of the emergence of a new category.

In his article, Verstegen sets out to re-examine the relationship between theories of metaphor by Conceptual Metaphor Theorists (CMT), like George Lakoff, and Gestalt psychologists, such as Rudolf Arnheim. This is prompted by several CMT having appropriated Gestalt arguments from Arnheim. As Verstegen goes to on to explain, this ‘translating’ is not easily sustained. While a Gestalt theory features bidirectionality of topic and vehicle of metaphors, CMT features unidirectionality from target to source. That is because for proponents of CMT, expressive language is based on bodily knowledge, which is formed by lived experience (2022, 46). The conceptual significance of expressions like ‘things are looking up’ is borrowed from some aspect of experience relating to space, time, or movement (2022, 47). Metaphors thus reference conceptual schemas, such as GOOD IS UP, based on experience, which, despite stylistic and semantic differences, they can share (2022, 47-48). Such schemas have been used to translate the work of Gestalt theorists. Lakoff, as Verstegen recounts, re-examined Arnheim’s analysis of Rembrandt’s painting Christ at Emmaus, which “in line with the interactive and creative nature of metaphor” kept “the uncertainty of the units of analysis to painting alive” (2022, 53).

Accordingly, in Rembrandt’s painting, for Arnheim the humility of the apostle and the aggrandizement of the servant is directly seen via the phenomenal tension of the two figures within the compositional framework, while for Lakoff, first we need to observe how the scene is structured by and schemas and infer via interpretation which cultural metaphors are being used (2022, 56). Verstegen highlights the lack of fit between these approaches, noting how Lakoff’s results in stock responses that could be prompted by any number of similarly configured compositions. Indeed, as he outlines, for Gestalt theorists, “meaning is embodied by forms in their chosen medium” while for proponents of CMT, “meaning is embodied in sensorimotor experience” (2022, 54). In relation to visual media, for the former, vision is primary and dominates while the body is secondary, but for the latter, “the sensorimotor experiences passively (through inference) dominate over the visual” (2022, 55). As Verstegen concludes, for Arnheim concepts are perceptual, while for Lakoff, perception is more conceptual.

Verstegen’s article helpfully highlights the tensions between these different approaches and clarifies how, despite the attempts of theorists like Lakoff, the analysis of paintings by Gestalt psychologists like Arnheim cannot easily be accommodated by CMT without loss of appreciation for the metaphorical content as particular to the work, or as interactive and creative. As Verstegen indicates, such appropriations serve to provide further support for Arnheim’s more complex approach and his psychology of art. Nonetheless, from an appreciative perspective we might wonder whether there is some benefit to interpreting certain works through stable cultural metaphors. If, as Arnheim suggested, abstracted qualities from the topic and vehicle “continue to draw life blood from the reality contexts in which they are presented” (quoted in Verstegen 2022, 49) then might it not be the case that, if some reality contexts change over time (particularly if they pertain to cultural phenomena for example), works can take on metaphorical meanings not intended by their creators? If perception is penetrable in the way Arnheim’s account suggests, then viewers may not help but be able to see a work as having metaphorical content that its creator could not have. This might even simply depend on one’s cultural background. Depending upon how much one is sympathetic to actual intentionalism, this may somewhat detract from the attractiveness of Arnheim’s approach. It would thus be fascinating to see how these psychological approaches, or combined aspects of these approaches, work in relation to theories in philosophical aesthetics.

We then have an interview with Ralph Webb, whose collection of poetry Rotten Days in Late Summer (2021) was a Book of the Year for both the Telegraph and the Irish Times, was shortlisted for the Forward Prize, and has received many glowing reviews. In an intimate and honest discussion with Rebecca Wallbank, Ralph Webb talks about what it means to be a writer and we are offered a privileged insight into the poet’s personal process. Wallbank’s well-considered questions lead Webb to reflect deeply on the relationships between a poem, its writer, and its readers; this stimulates the question of why and how poetry moves people, as well as touching on philosophical questions about authorial intention, ownership, and appropriation. In particular, the concept of poetic appropriation, in which a reader takes on the poet’s words as they apply to their own experiences or emotions, is given insightful consideration in this interview and Wallbank remarks upon the surprising degree of interpretive freedom that Webb is happy to allow his readers (2022, 67). She concludes that while a poem or artwork “is not ours to take”, the reader may still “bring something of themselves” to their appreciation of it (2022, 68).

Finally, this issue offers a review of Emmanuel Alloa’s book Looking Through Images: Α Phenomenology of Visual Media by Fotini Charalabidou. Charalabidou concisely and methodically summarises Alloa’s insights on the human experience of images before offering some considered critique regarding Alloa’s underlying assumptions – though she emphasises that, ultimately, this book makes an impressive and noteworthy contribution to the study of mediality (2022, 77).

Alessandro Cavazzana

↑ Back to top ↑

Davidson’s insights into metaphor are often treated as an isolated episode, with little regard for his work on semantics. In this paper, I want to reassess What Metaphors Mean (1978) in the light of Davidson’s theory of meaning to explain why he is convinced that a metaphor lacks cognitive content and is devoid of any meaning other than that conveyed by its words in their literal interpretation.

1 Introduction

Donald Davidson’s position on metaphors’ cognitive content can be summarised as follows: metaphors are devoid of any cognitive content in addition to the literal (Davidson 1978, 32).[1] The main goal of this paper is to try to reconstruct, step by step, the argumentation that leads to this conclusion, mainly by contextualizing Davidson’s insights into metaphor in the light of his semantics. This contextualization will help to make Davidson’s point about metaphor more plausible, because it will allow for a better understanding of his argument.

Davidson’s What Metaphors Mean (1978) – often treated as an isolated episode and consequently commented on, and criticized with, little regard for Davidson’s work on semantics – hides a very precise theoretical premiss.[2] In this essay, in fact, Davidson discreetly alludes to the theory of meaning to which he refers, and of which he himself was one of the main exponents (Fogelin 1988, 52). This can be seen in some passages scattered throughout the paper, such as:

Literal meaning and literal truth conditions can be assigned to words and sentences apart from particular contexts of use. This is why adverting to them has genuine explanatory power (for metaphors) (Davidson 1978, 33).

According to Davidson (1978), it is important to distinguish meaning from use, that is, the semantic aspect from the pragmatic aspect.

My disagreement is with the explanation of how metaphor works its wonders. To anticipate: I depend on the distinction between what words mean and what they are used to do (Davidson 1978, 33).

Davidson specifies below the type of usage that he considers suitable for metaphorical sentences: “I think metaphor belongs exclusively to the domain of use [as] it is something brought off by the imaginative employment of words and sentences” (Davidson 1978, 33).

Davidson’s assumption, then, is that one should not postulate a metaphorical meaning, since metaphors lack one. Instead, one must evaluate them exclusively in the context of their use, i.e. according to the conditions under which they were uttered or written. Davidson is also convinced that “metaphors mean what the words, in their most literal interpretation, mean, and nothing more” (Davidson 1978, 32). But what does this remark exactly mean?

It is clear that Davidson’s position revolves around the verb ‘to mean’. I, therefore, think that it is necessary to contextualize – and not isolate – What Metaphors Mean (1978) in the light of Davidson’s programme about what form a theory of meaning should take.

2 Meaning and truth

According to the ‘narrow’ account,[3] a theory of meaning must be able to explain what it means for an utterance of a language L to be endowed with meaning (Picardi 1999, 13). Let us see, then, the particular declination that this theory assumes within the Davidsonian proposal, where the notions of interpretation, meaning and truth are intimately connected. In what way? During a linguistic communication, the interpreter assigns truth conditions to the utterances produced by the speaker: according to the truth-conditional semantics, giving the conditions under which a sentence would be true is a way of indicating the meaning of that sentence (Davidson 1967, 310).

Davidson develops a theory that he calls radical interpretation, since, according to the American philosopher; “[t]he problem of interpretation is domestic as well as foreign […]. All understanding of the speech of another involves radical interpretation” (Davidson 1973, 313). The question Davidson seeks to answer is: what kind of knowledge must an interpreter possess in order to assign a meaning to each sentence that is uttered by a speaker? If establishing the truth conditions of an utterance is a way of determining its meaning, then what is needed is a theory of truth; i.e. a theory that can answer the question ‘what is truth?’, or ‘what does an utterance need to possess the quality of being true?’. Davidson’s (1967, 309) first move is, therefore, to replace the schema:

(T) the sentence e means that p

With the equivalence scheme:

(T1) the sentence e is true if and only if p.

The problem with the connective ‘means that’ concerns the “anxiety that we are enmeshed in the intensional” (Davidson 1967, 309). Such a scheme, in fact, does not respond to the Leibnizian principle of substitution salva veritate, since it allows the use of terms that have the same extension but differ in intension. For example, the term ‘red’ and ‘the colour of the Chinese flag’ have the same extension, i.e. red. However, let us consider the following sentence:

(T) ‘red is a primary colour’ means that the colour of the Chinese flag is a primary colour.

It is false because it is not true that the remark ‘the colour of the Chinese flag is a primary colour’ provides the meaning of ‘red is a primary colour’ (De Caro 1998, 22).

Here, Davidson calls into question the theses of the Polish logician Alfred Tarski. Tarski’s aim is to construct a theory of truth that: (a) is based on the idea that the truth of a sentence depends on its correspondence with reality, (b) does not apply only to certain sentences, (c) is rigorous and therefore scientifically respectable (Caputo 2015, 81). With regard to point (a), the truth-world correspondence can be summarised by the expression ‘a sentence is true if it designates a subsistent state of affairs’ i.e. if things in the world are actually the way an utterance says they are. However, such expression, is not considered by Tarski to be a satisfactory definition of truth (Tarski 1944, 343). It is much better, then, the following biconditional (Tarski 1944, 344):

(T1) “e” is true if and only if p.

Where does (T1) come from? Tarski asks us to consider a sentence, for example p, and to give a name to this sentence, for example ‘e’. From the point of view of truth conditions, it is clear that ‘e’ and p are equivalent, since ‘e’ is true only if the conditions described by p are fulfilled; e.g. ‘the grass is green’ is true if and only if the grass has a greenish colour. Moreover, according to Tarski, if the appropriate biconditional is provided for each sentence of a language (L), then each biconditional will constitute only a partial definition of truth. Thus, a general definition of truth for a language (L) will be composed of the logical conjunction of all conceivable biconditionals for the sentences of that language (Tarski 1944, 344). The sentences to be inserted in the place of ‘e’ are not true or false per se, but according to the meaning they have in a given language (L). Thus, the scheme (T1) should be relativised by means of the predicate “true-in-L”, where L indicates the language for which the sentence has a truth value (Caputo 2015, 85). The biconditional thus gains the following form, which is the scheme of the so-called Tarskian Convention T (where, as it is well-known, ‘T’ stands for ‘truth’):

(T1) “e” is true-in-L if and only if p.

Moreover, according to Tarski (1944), to avoid the problems caused by semantic paradoxes, such as the antinomy of the liar, it is necessary to construct the biconditionals using sentences that belong to a semantically open language.[4] A language that is semantically open is one in which there is not the predicate ‘true’ (or ‘true-in-L’). The truth predicate must, therefore, be found outside the so-called object language, which is the one we are talking about and to which ‘e’ belongs. It will be the task of the metalanguage, i.e. the language in whose terms we want to construct the definition of truth for the object language, to host the term ‘true’. Which languages lend themselves to the formulation of a definition of truth? That is, which languages are semantically open? Tarski is convinced that it is not possible to construct a definition of truth for natural languages such as Italian or English because of the risk of running into semantic paradoxes, e.g. the liar paradox. The most suitable languages are, therefore, formalized languages like the ones of mathematics, logic, set theory, and so on (Tarski 1944, 348-351).

Let us come back to Davidson, who directly built his work upon Tarski’s. According to Davidson (1967), if providing the truth conditions of a sentence is a way of indicating the meaning of that sentence, then a theory such as the one elaborated by Tarski can represent an excellent model for a theory of meaning applicable to a given language. In this way, an interlocutor, or reader, will be able to use Tarski’s biconditionals to interpret the speaker’s utterances. Evidently, Davidson does not think that the theory applies only to formal languages but is convinced that the biconditionals can also be constructed with sentences from the natural languages (Davidson 1967, 313). He admits that natural languages are endowed with indexicals, such as verb tenses or demonstrative pronouns; i.e. those parts of the sentence whose truth value varies according to the context in which the sentence is uttered. He, therefore, suggests modifying the biconditionals by adding references to time and the speaker (Davidson 1973, 322). The resulting scheme is:

(T1) “e” is true-in-L when spoken by x at time t if and only if p near x at t.

Davidson assumes a situation where the speaker and the interpreter express themselves in two different languages and where the interpreter does not know the speaker’s language. For example, Kurt – a German speaker – utters the words ‘Es regnet’. The radical interpreter does not know German, but his first move must involve the so-called Davidsonian Principle of Charity (Davidson 1967, 1974), i.e. the interpreter must attribute true and consistent beliefs to Kurt whenever it is permissible, since someone is much more likely to believe things he considers true than those he sees as false (Perissinotto 2002). Moreover, according to the Principle of Charity (Davidson 1967), these beliefs cannot be too dissimilar from those of the interpreter since communication can only take place on the basis of a massive agreement between the two parties. The interpreter must, therefore, believe that Kurt believes that ‘Es regnet’ is true. In that case, the evidence available to the interpreter is also of an extra-linguistic nature and includes, for example, the directly observable behaviour of the speaker, or the environmental conditions, within which such utterances are expressed. The collection and analysis of such data will help the interpreter formulate conjectures about what the speaker is saying but will not be explicitly exhibited by the so-called T-sentences (i.e. the sentences of the form ‘e’ is true-in-L if and only if p). In fact, the right side of the biconditional will merely show the appropriate circumstances under which a speaker expresses the utterance, i.e. the supposed truth conditions for ‘e’ (Picardi 1992, 248-249). Davidson, for example, assumes that the available evidence for interpreting Kurt’s utterance ‘Es regnet’, is Kurt’s membership to the German speech community, the fact that Kurt believes ‘Es regnet’ to be true at noon on Saturday, and that it is raining near Kurt at that time and on that day. The appropriate biconditional will have the following form:

(T1) ‘Es regnet’ is true in German, when spoken by Kurt at noon on Saturday if and only if it is raining near Kurt at noon on Saturday.

Davidson acknowledges that Kurt may be mistaken about the fact that it is raining in his proximity, but the intention is to create a theory that maximizes the agreement between speaker and interpreter, with the speaker being as much in the right as possible (Davidson 1973, 323).

3 Limits of the literal

Is it possible, then, to obtain a truth-conditional theory of meaning for all the utterances of a language? Although Davidson states that he wants to do everything possible to dispel Tarski’s pessimism towards the establishment of a theory for non-formalized languages (Davidson 1967, 313), he admits that there are limits. The metaphor represents one of these limitations. From a truth-conditional point of view, a metaphorical sentence is always patently false or true in such an unquestionable way that the interpreter may find identifying its truth conditions superfluous, thus focusing not on the literal meaning but on the speaker’s intended meaning. When Davidson claims that “metaphors mean what the words, in their most literal interpretation, mean, and nothing more” (Davidson 1978, 32), he is alluding to what I have tried to summarise, i.e. metaphors, as sentences of a language, can be treated like any other sentence and put within a T-sentence, but such an operation is almost worthless since, from a literal point of view, no condition can satisfy the truth of what lies to the left side of the biconditional.

In What Metaphors Mean (1978), Davidson rejects some of the main theses advanced by Max Black (1955) in the context of his interactionist account of metaphor. Contrarily to Black, Davidson does not admit: (i) that the metaphorizing term – the focus, in Black’s terminology – is endowed with a special metaphorical meaning (which is in addition to the literal meaning); (ii) that metaphors have a cognitive content that can be true despite the obvious falsehood of the literal meaning; (iii) that the reason why metaphors cannot be paraphrased lies in their being carriers of another meaning, beyond the normal, literal meaning (Leddy 1983, 64). For what concerns (i), Davidson specifies that what the metaphor conveys to a possible interpreter “depends entirely on the ordinary meanings of those words and hence on the ordinary meanings of the sentences they comprise.” (Davidson 1978, 33). This account fits perfectly into his semantic theory and, indeed, in Truth and Meaning (1967), where he recalls the principles of compositionality and contextuality and states that:

we decided a while back not to assume that parts of sentences have meanings except in the ontologically neutral sense of making a systematic contribution to the meaning of the sentences in which they occur. […]Sentences depend for their meaning on their structure, […] Frege said that only in the context of a sentence does a word have meaning (Davidson 1967, 308).

Basically, the meanings of sentences depend on the meanings of the words that compose them, but it is only in their context that a word acquires meaning. This circularity is also expressed in Davidson’s treatment of metaphors. If at the level of the single word, the only possible meaning is the ordinary, literal one, then this meaning will also be conveyed in the whole sentence, and vice versa. In Max Black’s famous metaphor ‘man is a wolf’, ‘wolf’ is the focus of the metaphor (Black 1955). The system of clichés associated with ‘wolf’ and also valid for ‘man’ suggests that a possible secondary meaning for ‘wolf’ is ‘predatory animal’. Thus the double meaning theory, contested by Davidson, considers “the key word (or words) in a metaphor as having two different kinds of meaning at once, a literal and a figurative meaning” (Davidson 1978, 35). However, in Davidson’s semantic account, the double meaning is not permissible. Davidson’s extensionalism stipulates that the reference of ‘wolf’ is the animal, for which the name stands, and not a connotative description (i.e. an intension) of it since, as we have seen, the instruments adopted by Davidson to indicate the truth conditions of an utterance are constructed with the declared aim of curbing intensionality. Thus, if intensionality is to be contained in a sentence, it is clear that, by the principle of compositionality, this must be done starting from the individual words. That is to say that, in order to avoid intensionality, the meaning of words has to be limited to the ordinary one.

While keeping the metaphor at the centre of our discourse, we now turn our attention to the purpose of Davidson’s theory of meaning, namely, the possibility of interpreting the speaker’s utterances. Earlier, it was said that the interpreter’s first move is governed by the Principle of Charity (1974), according to which the interpreter must take the speaker’s utterances to be true. Metaphor overturns this practice since it works when it is considered to be false. Davidson himself suggests that “is only when a sentence is taken to be false that we accept it as a metaphor and start to hunt out the hidden implication” (Davidson 1978, 42). The metaphor is false in such a bizarre way (or true in such a trivial way) that the interpreter cannot conceive it as a source of information from a strictly literal point of view, nor can she attribute to the speaker the paradoxical belief for which the metaphor stands. This ‘irrelevance’ (or non-relevance) of the metaphor (Sperber and Wilson 1986) will then lead the listener to question the real purpose of such an absurd utterance.

4 Conclusion

In conclusion, what the T-sentence succeeds in grasping and evaluating concerns the literal meaning of the metaphor – which is the only kind of meaning Davidson is interested in – and this is exactly what Davidsonian semantics aims at analyzing. It cannot deal with what some critics have called the ‘metaphorical meaning’, since, by Davidson’s own admission, this meaning, if it exists at all, lacks an enunciative form. Through a use of words that involves an imaginative capacity on the part of the interpreter (but also of the speaker), metaphor provokes in the listener or reader a vision, allows for creative elaboration of thoughts, evokes particular connections (Davidson 1978, 47). What a metaphor points out to an interpreter has an extra-linguistic nature that is not verbally delimitable:

there is no limit to what a metaphor calls to our attention, and much of what we are caused to notice is not propositional in character. When we try to say what a metaphor ‘means’, we soon realize there is no end to what we want to mention (Davidson 1978, 46).

Essentially, all that a metaphor makes us imagine is external and does not match any further semantic dimension of the metaphorical sentence.

Finally, this brings us to point (ii). Davidson does not deny tout court that a metaphor has a cognitive content, but he denies that it has additional cognitive content to its literal meaning. Here, Davidson’s implicit assumption is that only sentences are vehicles of such content.[5] If what a metaphor brings to the interpreter’s attention is not reducible to a propositional form, then Davidson also rules out the possibility of further cognitive content in addition to the literal.

At this point, one could ask why Davidson did not formulate his theory of metaphor more explicitly and closely related to his theory of meaning. The answer is contained in this sentence: “I think metaphor belongs exclusively to the domain of use. It is something brought off by the imaginative employment of words and sentences” (Davidson 1978, 33).

Davidson is not formulating any theory of metaphor. He is trying to say that, if the solution of scholars is to postulate a metaphorical meaning, then this particular kind of meaning does not fall into the realm of semantics and, if it does not fall into that realm, it is not something that can be examined with his theory of meaning. In this respect, Davidson is absolutely right, and what I wanted to show here is that his position on metaphors is perfectly consistent with his semantics.

Nevertheless, Davidson offers some alternatives to truth-conditional semantics. He puts on the table some arguments against metaphorical meanings that do not rely on his semantic theory. For example, postulating a new or extended meaning makes it difficult to explain in what way the metaphorical interpretation starts from the original (or literal) meaning (Davidson 1978, 34). Another point against metaphorical meanings lies in the existence of the so-called dead metaphors: if a new meaning has totally replaced the literal one, it is not so simple to say why we no longer consider dead metaphors as metaphors (Davidson 1978, 37-8). Davidson here wants to show that the arguments supporting the existence of metaphorical meanings are weak even when analysed with approaches other than that of truth-conditional semantics.

One of the suggestions offered by Davidson is to compare metaphors to pictures:

How many facts or propositions are conveyed by a photograph? None, an infinity, or one great, unstatable fact? Bad question. A picture is not worth a thousand words, or any other number. […] What we notice or see is not, in general, propositional in character (Davidson 1978, 47).

For Davidson, the working mechanism of a metaphor is comparable to the Wittgensteinian seeing-as; we do not see that Juliet is the Sun, but rather we see Juliet as the Sun. Metaphors make us see one thing as another through a certain literal statement that stimulates an imaginative insight (Davidson 1978, 47). This is also the reason why, for Davidson (1978, 32), metaphors cannot be paraphrased (point iii): the mental images they provoke are not verbally translatable.

This last suggestion, about imaginative insight, has been fruitfully picked up and developed in the field of both visual (Carroll 1994)[6] and verbal metaphors (Carston 2010; 2018).

In the end, Davidson’s great merit was to heavily influence and direct the debate on metaphor, wrenching it away from semantics and fruitfully delivering it to the fields of pragmatics and cognitive linguistics.[7]


Black, Max, ‘Metaphor’, Proceedings of the Aristotelian Society (1955) 55, 273-94.

Black, Max, ‘How Metaphors Work: A Reply to Donald Davidson’, Critical Inquiry (1979) 6:1, 131-43.

Caputo, Stefano, Verità (Roma: Laterza, 2015).

Carroll, Noël, ‘Visual metaphor’, in Jaakko Hintikka (ed.), Aspects of Metaphor (Dordrecht: Kluwer Academic Publishers, 1994), 189-218.

Carston, Robyn, ‘Metaphor: Ad Hoc Concepts, Literal Meaning and Mental Images’, Proceedings of the Aristotelian Society (2010) 110, 295-321.

Carston, Robyn, ‘Figurative Language, Mental Imagery, and Pragmatics’, Metaphor and Symbol (2018) 33:3, 198-217.

Cavazzana, Alessandro, ‘Immagini (per l)e Parole. La Metafora Visiva tra Occhio Innocente e Immaginazione’, Rivista Italiana di Filosofia del Linguaggio (2017) 11:2, 109-22.

Cavazzana, Alessandro, ‘Imagining: The Role of Mental Imagery in the Interpretation of Visual Metaphors’, in András Benedek and Kristóf Nyíri (eds.), Perspectives on Visual Learning, vol. 3: Image and Metaphor in the New Century (Budapest: Hungarian Academy of Sciences/Budapest University of Technology and Economics, 2019), 71-82.

Cavazzana, Alessandro, and Bolognesi, Marianna, ‘Uncanny Resemblance. Words, Pictures, and Conceptual Representations in the Field of Metaphor’, Cognitive Linguistic Studies (2020) 7:1, 31-57.

Davidson, Donald, ‘Truth and Meaning’, Synthese (1967) 17:3, 304-23.

Davidson, Donald, ‘Radical Interpretation’, Dialectica (1973) 27:3-4, 313-28.

Davidson, Donald, ‘On the Very Idea of a Conceptual Scheme’, Proceedings and Addresses of the American Philosophical Association (1974), 47, 5-20.

Davidson, Donald, ‘What Metaphors Mean’, Critical Inquiry (1978) 5, 31-47.

Davies, Stephen, ‘Truth-Values and Metaphors’, The Journal of Aesthetics and Art Criticism (1984) 42:3, 291-302.

De Caro, Mario, Dal punto di vista dell’interprete. La filosofia di Donald Davidson (Roma: Carocci, 1998).

Dummett, Michael, The Logical Basis of Metaphysics (Cambridge: Harvard University Press, 1991).

Fogelin, Robert J., Figuratively Speaking (Oxford: Oxford University Press, 1988).

Gentile, Francesco P., Talking Metaphors. Metaphors and the Philosophy of Language (PhD dissertation, University of Nottingham, 2013).

Goodman, Nelson, ‘Metaphor as Moonlighting’, Critical Inquiry (1979) 6:1, 125-30.

Leddy, Thomas, ‘Davidson’s Rejection of Metaphorical Meaning’, Philosophy and Rhetoric, (1983) 16:2, 63-78.

Lepore, Ernest, and Stone, Matthew, ‘Against Metaphorical Meaning’, Topoi (2010) 29:2, 165-80.

Levinson, Jerrold, ‘Who’s Afraid of A Paraphrase?’, Theoria (2001) 67:1, 7-23.

McGonigal, Andrew, ‘Davidson, Metaphor and Error Theory’, in Kathleen Stock and Katherine Thomson-Jones (eds.), New Waves in Aesthetics (London: Palgrave Macmillan, 2008), 58-83.

Moran, Richard, ‘Seeing and Believing: Metaphor, Image and Force’, Critical Inquiry (1989) 16:1, 87-112.

Perissinotto, Luigi, Le vie dell’interpretazione nella filosofia contemporanea (Roma-Bari: Laterza, 2002).

Picardi, Eva, Linguaggio e analisi filosofica. Elementi di filosofia del linguaggio (Bologna: Pàtron Editore, 1992).

Picardi, Eva, Le teorie del significato (Roma-Bari: Laterza, 1999).

Reimer, Marga, ‘Davidson on Metaphor’, Midwest Studies in Philosophy (2001) 25, 142-55.

Reimer, Marga, and Camp, Elisabeth, ‘Metaphor’, in Ernest Lepore and Barry C. Smith (eds.), The Oxford Handbook of Philosophy of Language (Oxford: Oxford University Press, 2006), 845-63.

Sperber, Dan, and Wilson, Deirdre, Relevance. Communication and Cognition (Oxford: Blackwell, 1986).

Tarski, Alfred, ‘The Semantic Conception of Truth and the Foundation of Semantics’, Philosophy and Phenomenological Research (1944) 4:3, 341-76.

  1. Let us consider the metaphor ‘my brother is the black sheep of the family’. According to Davidson, it makes no sense to distinguish – as many other scholars do – a literal meaning (i.e. the patent falsity that a metaphor like this one expresses) and a metaphorical meaning (the intended meaning, i.e. the fact that my brother is considered as different, bad, worthless, etc. by the rest of the family). For Davidson, the metaphorical meaning simply does not exist. Metaphors show us something not by conveying a well-defined propositional content but by triggering our imagination through the ordinary meanings of words and sentences. In Davidson, the parameter of the cognitive content of the metaphor is the core of what some critics have called error theory (McGonigal 2008). According to the error theory of metaphor, metaphorical statements are not meaningless but literally false. McGonigal tries to defend the following position: “if a sentence used metaphorically is true or false in the ordinary sense, then it is clear that it is usually false. […] most metaphors are false” (Davidson 1978, 41). McGonigal’s polemical target is the radical alethic pluralism defended by Crispin Wright (McGonigal 2008, 76-80). Davidson’s non-cognitivist position is also called causal theory, since a metaphor causes a vision “[making] us attend to some likeness, often a novel or surprising likeness, between two or more things” (Davidson 1978, 33). For a general framing of Davidson’s theory in the analytical debate on metaphor, see Reimer and Camp (2006, 854-8).

  2. It is worth summarizing some of Davidson’s critics to show that his insights about metaphorical meaning have almost never been framed within his semantic theory. Max Black (1979), for example, defends his own interactionist account of metaphor against Davidson’s criticism, but his apologia never really contextualizes the sense of ‘meaning’ as understood by Davidson. According to Nelson Goodman (1979), instead, metaphor operates through a mechanism of label application: contrarily to Davidson, he believes that a term is taken from its literal use and applied in a novel way to a new object. Among Davidson’s critics, Richard Moran (1989) disagrees that the message of a metaphor is difficult to delimit from a verbal point of view. For Moran, grasping the meaning of a metaphor means instead selectively limiting interpretation to the right similarities. Jerrold Levinson (2001) replies to Davidson by comparing metaphors to exclamations: like exclamations, metaphors have meanings in context that go beyond the meanings of their constituent words. This further meaning is partly propositional and non-propositional and can, therefore, be paraphrased. The non-propositional part, characterized by an illocutionary force, can still be described (Levinson 2001). Stephen Davies (1984), on the other hand, largely adheres to Davidson’s position, arguing that the truth value of metaphors is linked to their literal meaning, which is the only meaning metaphorical utterances have. Among those inspired by Davidson, at least in terms of their denial of metaphorical meaning and metaphorical truth, see also Lepore and Stone (2010). With other methods and other purposes besides those of this paper, the only work which attempted to contextualise Davidson’s insights on metaphor in light of his theory of meaning was Gentile’s PhD dissertation (2013).

  3. For Michael Dummett there is a meaning-theory and the theory of meaning. The former is specific to a specific language: “[it] is a complete specification of the meanings of all words and expressions of one particular language” (Dummett 1991, 22). While the latter has to define what general principles are needed to build the former. The task of the theory of meaning, according to Dummett, is to provide an explanation of how language works, i.e. to explain what happens when a speaker utters a sentence (in a certain language) in the presence of a competent listener (Dummett 1991, 21).

  4. Tarski (1944) recalls the liar paradox in this way. Consider the sentence “(a) is not true”. Let us give a name to this sentence, calling it (a). The resulting biconditional is as follows: “(a)” is true if and only if (a) is not true. The contradiction is obvious, which is why, according to Tarski, an object language must be semantically open, i.e. it must not contain the predicate “true” (Tarski 1944, 347-348).

  5. For a critique of this position see Reimer (2001). Reimer (2001) summarises Davidson’s argument as a modus tollens: [ ( p ” q ) ^ ¬q ] ” ¬p. Where: p = a metaphor has a special cognitive content; q = it is possible to provide this (presumed) content through a literal expression. So, Reimer’s reconstruction of Davidson’s argument sounds like: if a metaphor has a special cognitive content, then it is possible to provide this (presumed) content by means of a literal expression, but it is not possible to provide this (presumed) content by means of a literal expression, so a metaphor does not have a special cognitive content. Reimer disagrees with q; she believes that cognitive content is not the only content that can be expressed verbally (Reimer 2001, 145).

  6. For a critical reading of Carroll’s account (1994) of visual metaphors, see Cavazzana (2017). For the importance of imagination in the perception of visual metaphors, see Cavazzana (2019), as well as Cavazzana and Bolognesi (2020).

  7. I would like to thank the anonymous referees for their precious comments.

Kiyohiro Sen

↑ Back to top ↑

It is widely acknowledged that categories play significant roles in the appreciation of artworks. This paper argues that the correct categories of artworks are institutionally established through social processes. Section 1 examines the candidates for determining correct categories and proposes that this question should shift the focus from category membership to appreciative behaviour associated with categories. Section 2 draws on Francesco Guala’s theory of institutions to show that categories of artworks are established as rules-in-equilibrium. Section 3 reviews the explanatory benefits of this institutional theory of the correct category.


It is widely acknowledged that categories play significant roles in the appreciation of artworks. Categories affect interpretation: we can assume that the setting of a Western film is the United States during the Wild West, even if that is not stated explicitly in the story. Categories affect evaluation: if it is a musical, we cannot view it as silly when characters suddenly start singing and dancing. Not only genres but other types of categories too — form, style, movement, tradition — can have similar effects[1].

Here, I would like to take categories in a broader, more general, way, that is as points of view, ways of seeing, or frameworks for appreciating artworks. A work of art can, in principle, be perceived from several different categories. For example, we could judge Stanley Kubrick’s 2001: A Space Odyssey (1968) as a mystical and sublime masterpiece as a science fiction film, or we could judge it as a dull and uninspired waste of time as a romance film. However, the latter judgment would miss the point or be less appropriate than the former. Walton (1970) famously argued that there are correct categories for artworks, and the correct appreciation of a work is based on such categories. This raises the question, what determines the correct categories?

This paper argues that the correct categories of artworks are institutionally established through social processes. Section 1 examines the candidates for determining the correct categories and proposes that this question should shift the focus from category membership to appreciative behaviour associated with categories. Section 2 draws on Francesco Guala’s (2016) theory of institutions to show that categories of artworks are established as rules-in-equilibrium. Section 3 reviews the explanatory benefits of this institutional theory of the correct category.

1 The correct category and membership

Walton sketches four factors that “count toward its being correct to perceive a work, W, in a given category, C.” (1970, 357):

(i) “The presence in W of a relatively large number of features standard with respect to C. […] it has a minimum of contra-standard features […] .”

(ii) “W is better, or more interesting or pleasing aesthetically, or more worth experiencing when perceived in C than it is when perceived in alternative ways.”

(iii) “the artist who produced W intended or expected it to be perceived in C, or thought of it as a C.”

(iv) “C is well established in and recognized by the society in which W was produced.” (Walton 1970, 357)

Above all, Walton emphasizes the importance of the criterion (iii), that is, the author’s intention (cf. Walton 1973). The idea that the author’s intention determines the correct categories of a work of art is deep-rooted and held even by those who otherwise prefer non-intentionalist approaches (cf. Levinson 1996, 188-189; Davies 2006, 233).

There are reasons and cases to doubt the intentionalism on categorizations. Categorical intentions are insufficient insofar as categories are frameworks for interpretation and evaluation. Allowing artists to solely determine the category of a work would grant the artist’s intention implausibly strong powers in determining the work’s meaning and value. That is, the artist could set a self-serving bar for her work. It is odd that a work that is only incoherent and sloppy should be interpreted as a symbol of something significant or evaluated positively simply because the artist intended it to be a piece of Absurdist fiction. Moreover, we should consider cases where the author’s intention fails. When the artist attempts to create an artwork that belongs to a given category C, the categorical intention alone is not sufficient for the attempt’s success (cf. Mag Uidhir 2010).

To regard the categorical intention as a necessary condition is too restrictive in the following respects. As it is evident in practice, an artwork often belongs to a category that the author did not intend when it was created. Reading Raymond Carver’s work as minimalist literature or listening to Portishead as trip-hop involves categorizations that the authors did not intend or openly rejected. Moreover, an artwork may belong to a category that the author cannot intend. Reading Kafka’s or Dostoevsky’s works as existentialist literature or watching Tarkovsky’s films as slow cinema involves categorizations established long after the creation of these works. Since such categorizations occur in practice, theoretically dismissing these categories as incorrect would be a cost to pay. I would instead look for a theory of categories that approves various critical practices (cf. Gaut 1993, 605).

The four factors listed by Walton would be helpful as a heuristic for identifying the correct categories. However, none of them seems to provide logically necessary or sufficient conditions for category membership. Moreover, Walton (1970) treated categories as existing options and said little about the origins of categories. For example, how a category is associated with a particular set of features and what it means for a category to become established remains open.

As Walton (1970, 362) acknowledges, tightening up the conditions that determine the correct categories would be difficult. It is also quite possible that different categories have different conditions of membership or different priorities of factors. Therefore, the theory of correct categories needs a different approach. I propose that the question of what determines the correct category of a work (the membership question) should be replaced by how the correct appreciative behaviours related to that category are established (the behavioural question). By centering the behaviour, the question of category membership can be skipped.

As mentioned in the introduction, correct categories have normative roles in appreciation. The fact that a work belongs to a particular category is not just a matter of membership. It is also about undertaking certain appreciative norms. This normative aspect of categories can be described as a combination of the Membership rule and the Status rule.

(M) If X has the property F, then the correct category of X is Y.[2]

(S) If the correct category of X is Y, then the one who appreciates X should Z.

Z will be filled with a particular set of appreciative behaviours associated with category Y. If the work belongs to a specific category, one would be granted to assume that the setting is the United States or prohibited from evaluating it as silly just because characters start singing and dancing. This is analogous to the case of money: if X has specific historical and physical property F, X is money (Y), and if it is money, X is accompanied with rights and obligations concerning a set of economic behaviours (Z).

Here, importantly, the two rules above can be converted into a more straightforward norm of behaviour (Guala 2016, chap.5). Namely, a Regulative rule.

(R) If X has the property F, then the one who appreciates X should Z.

Here, the problem of membership to the correct category is skipped, transformed into a problem about appreciative behaviours. Then the correct category as status would be an incidental outcome of the behaviours in question.

Thus, the explanatory buck is passed to how the regulative rules under each correct category arise and become established. I am going to argue that the theory of institutions explains it.

2 Categories and institutions

Francesco Guala, in Understanding Institutions (2016), characterizes institutions as rules-in-equilibrium. By integrating the rule theories and equilibrium theories of institutions, Guala’s approach compensates for the shortcomings of each. If institutions are rules, we can explain the aspects that normatively guide behaviours. However, if they are merely rules, and not equilibria, the difference between followed institutions and those that are not is unclear. If institutions are equilibria, we can understand whether they are followed by seeing whether there are enough incentives. However, if they are merely equilibria, and not rules, we cannot explain the aspects that guide behaviours. Guala employs the concept of correlated equilibrium in coordination games to show that institutions are rules and equilibria.

In a practical sense, it does not matter whether a traffic law requires driving on the right or the left. What matters is to avoid the situation in which each driver chooses their lane freely and creates chaos. An agreement like driving on the right side in this country solves this coordination game. In this case, the strategy of following the agreement together (the “correlated” strategy) becomes the equilibrium of the situation in that no one would be better off by deviating alone. That is why people follow the agreement. A correlated equilibrium is an equilibrium based on external signals that are observed by each player. In this case, the signal is the public agreement. Additionally, driving as it has been agreed becomes a normative rule for each driver, or “player”, to follow. The rule guides behaviours, and whoever violates it receives penalties imposed for interfering with the traffic flow. Penalties reduce the incentive to deviate and reinforce normativity. In this way, certain behaviours are maintained within a society. According to Guala, institutions are rules-in-equilibrium.

Correlated equilibrium can also be applied to cases with asymmetric payoffs. Suppose that when a couple decides where to go on a date. Player 1 wants to go to A, while Player 2 wants to go to B. At the same time, however, neither of them wants to choose selfishly to give up their date. Figure 1 is the payoff matrix for this case, and each cell shows Player 1’s payoff on the left and Player 2’s payoff on the right. In this situation, for example, the agreement of deciding where to go by tossing a coin can be a signal of coordination. Suppose the probability of getting heads or tails is one in two for each, and they will go to A for a head and B for a tail. Participating in the coin toss can be a correlated equilibrium because deviating and going where they want to go on their own would not bring a bigger payoff. It is also desirable in terms of equality.

Figure 1. Correlated equilibrium for games with asymmetric payoffs

How is the artwork categorization game played? Let us assume two things. Firstly, each critic has an incentive to categorize an artwork in a way that deepens one’s understanding of the work, that is, in a way one prefers. This is not an assumption that all critics seek to maximize the value of a work. The category that helps me notice a work’s severe flaws and points out how it fails to achieve its goals is my preferred one because it deepens my understanding of that artwork. Secondly, each critic has an incentive to choose the same categorization as other critics because working together to criticize a work based on a particular category helps discover meanings and values that one would not notice on their own.[3] These assumptions lead to a situation that is analogous to the couple deciding where to go on a date, i.e. an asymmetric coordination problem.[4]

If the work to be appreciated is a typical one, coordination will be easy to achieve thanks to precedent works (Xhignesse 2020, 479).[5] It is not difficult to assign a new work to salient, socially well-established categories, in which the connection between standard/variable/contra-standard features and the categories is solid. There is no difficulty in categorizing Christopher Nolan’s Tenet (2020) as a science fiction and action film. The real problem occurs when pioneering works require pioneering ways of appreciation. According to Xhignesse, “their existence calls for a theory of the art-kind which they pioneer, […] they call for the development of conventions” (2020, 474). Works such as Marcel Duchamp’s Fountain (1917), Alejo Carpentier’s The Kingdom of This World (1949), Herschell Gordon Lewis’s Blood Feast (1963), and The Sugarhill Gang’s Rapper’s Delight (1979) require more than just putting them into existing categories at the time of their creation. It is not difficult to imagine that categories that are well established today also had pioneering works that called for the new categories at the beginning of their history.

As we saw in the previous Section, the choice of categories can be converted into the choice of appreciative behaviours. With this in mind, the origin of a category can be described in two phases. In the institutionalizing phase, critics attempt various appreciative behaviours. As mentioned above, critics, who confront each other, have incentives for both their preferences and agreements. Over a pioneering work, some prefer to Z1; while others prefer to Z2. Here, the critics should be taken in a broad sense. The theoretician-inclined artists would often be the first critics of their works, promoting specific behaviours. Categories like Nouvelle Vague and the Readymade were thus theoretically driven by the artists. In the institutionalized phase, a particular strategy is established as the correct appreciative behaviour for the artwork: collectively choosing that behaviour becomes a rule-in-equilibrium.

Meanwhile, the categories behind these appreciative norms were named conceptual art, magic realism, splatter film, or rap music. It helps us solve coordination problems smoothly based on precedents when we encounter new works with similar properties. This background is also consistent with what Walton points out as the causes of perceiving works in specific categories (1970, 341-342): (a) our familiarity with other works, (b) categorization by critics and others, and (c) the context in which we encounter the work.

However, suppose the incentive to choose the individually preferred category is much greater than the incentive to choose the same category as other critics. In that case, the correct category might not take hold. If both players prioritize their preferences over an agreement when deciding where to go, the only equilibrium is to give up on their date. The uniqueness of equilibrium remains unchanged even when the game is extended by adding the strategy of participating in a coin toss. Here, the couple is in a so-called prisoner’s dilemma:[6] even though the strategy of participating in the coin toss is superior in terms of payoff, players cannot but act selfishly (Figure 2). If the artwork categorization game has such a payoff setting, then no one needs correct categories.

Figure 2. Prisoner’s dilemma in a selfish couple

There are three possible responses to this supposed objection. Firstly, such a payoff setting is questionable in that it is at odds with the existing practice of critical consensus on the interpretations and evaluations (cf. Walton 1970, footnote 21). It precisely shows the situation where there is no disputing about taste, and it is difficult to recognize the significance of critical debate. Secondly, even if the incentive to choose the individually desirable categorization is great, it is not an immediate threat to establishing correct categories. The correct appreciative behaviour is established in a reflective equilibrium through more detailed negotiation instead of a random coin toss. Even if an inappropriate interpretation or evaluation is made, the features and facts that do not fit it do not disappear. This prompts re-categorization, which eventually leads to a better category by equilibrium. The reached consensus will be superior at least to the expected value of the coin toss as far as the payoff is concerned. Thirdly, Guala (2016, chap.6) shows a path to convert the prisoner’s dilemma into a coordination problem by considering the costs (penalties) associated with breaking the rule. In this regard, there could be institutions that would be useful for prisoner’s dilemma situations too.

3 Explanatory benefits and consequences

One of the advantages of identifying the correct categories as institutional outputs is that we can observe the categories’ stability as well as the critical innovation’s possibilities and difficulties. When a specific categorization of a work is firmly established as a rule-in-equilibrium, the attempts at unusual appreciative behaviours are often weeded out without having much impact. However, in rare cases, they may prompt a revision of the institution. Evolutionary games may explain this. When an evolutionarily stable strategy is given, even if a few new players choose a strange strategy, the stable state is maintained without a proliferation of players choosing such mutable strategies.[7] However, a highly applicable strategy (in our case, a highly attractive categorization) may arise as a small number of mutations and eventually lead to another equilibrium that is superior to the current rule-in-equilibrium.

The advantage of adopting Guala’s account of the institution over other accounts is that it does not require the player to display a high degree of linguistic ability. There is no need to collectively recognize that the category to which a work belongs is called ‘horror’ if one can recognize what to do. We recognize that the purpose of the work is to frighten the viewer, so it should be evaluated according to the effectiveness of the means. We recognize the precedents and conventions, which shape how we interpret that artwork. Like the etiquette of standing on one side when using an escalator, the rule-in-equilibrium does not always have a name. Jan Švankmajer’s works are authentic when seen as a category that evokes disgust through vulgar eating scenes. We do not need the category’s proper name to make such an evaluation. Of course, since having a name is advantageous to the operation and modification of the institution, naming and describing informally accepted categories will be one of the critics’ essential tasks.

If the correct categories are institutional outputs, then the correct meaning and value of a work of art are relative to its history and the community of reception. In this regard, the stable state of a game often has path dependence on its initial conditions. Suppose a situation where a coordination game is being played repeatedly in a group. The strategy chosen by a randomly selected player is influenced by the ratio of strategies in the group: if many people use Windows, I choose Windows; if many people use Macintosh, I choose Macintosh. If the percentage of Windows users is above a certain level, the randomly selected player has more incentive to choose Windows. Eventually, an equilibrium will be achieved where most people choose Windows. The same thing can happen with Macintosh since the achieved equilibrium depends on the initial state at a given point in time. Although there are many other factors to consider in the real world, this is partly how VHS defeated Betamax and Blu-ray defeated HD DVD. Likewise, the categorization of an artwork that was salient in the early days of its publication may affect the correct categories, later established. In this respect, the current correct categories are contingent and, like social institutions, not always the best. Tracing these origins and transitions, and then documenting the facts, will be one of the essential tasks for art history.[8]


Abell, Catharine, Fiction: A Philosophical Analysis (Oxford: Oxford University Press, 2020).

Davies, Stephen, ‘Authors’ Intentions, Literary Interpretation, and Literary Value’, British Journal of Aesthetics (2006) 46:3, 223-247.

Gaut, Berys, ‘Interpreting the Arts: The Patchwork Theory’, The Journal of Aesthetics and Art Criticism (1993) 51:4, 597-609.

Guala, Francesco, Understanding Institutions: The Science and Philosophy of Living Together (Princeton: Princeton University Press, 2016).

Laetz, Brian, ‘Kendall Walton’s ‘Categories of Art’: A Critical Commentary’, British Journal of Aesthetics (2010) 50:3, 287-306.

Levinson, Jerrold, The Pleasures of Aesthetics: Philosophical Essays (Ithaca: Cornell University Press, 1996).

Mag Uidhir, Christy, ‘Failed-Art and Failed Art-Theory’, Australasian Journal of Philosophy (2010) 88:3, 381-400.

Walton, Kendall, ‘Categories of Art’, Philosophical Review (1970) 79:3, 334-367.

Walton, Kendall, ‘Categories and Intentions: A Reply’, The Journal of Aesthetics and Art Criticism (1973) 32:2, 267-268.

Xhignesse, Michel-Antoine, ‘What Makes a Kind an Art-kind?’, British Journal of Aesthetics (2020) 60:4, 471-88.

  1. Inherent categories such as the author, performer, period, region are beyond the scope of this paper. Media understood as a purely physical vehicle should also be set aside in this sense. Here I am concerned with acquired categories that cannot be determined by historical facts alone. A film made by Hitchcock can avoid being a suspense film but cannot help being a Hitchcock film. In this respect, whether it is a Hitchcock-style film is not inherently determined.

  2.   Here, the property F should not be understood only as perceptual properties of a work. It also includes relational properties, such as being the product of a successful intention or having a causal connection to precedents. Repeatedly, it is hard to find a condition that is determinative for category membership, but the approach of this paper has the advantage of leaving property F open. My approach is compatible with a position like Laetz’s (2010), which is sceptical of equating the correct category with the category to which the work belongs.

  3.   Abell (2020) identifies ‘the communication of imaginings’ (32) as a coordination game to be solved and argues that the contents of fictional utterances are institutionally determined. Although her interests and assumptions differ from mine, some of my arguments overlap with Abell’s insofar as categories have interpretive roles.

  4.   Xhignesse (2020, 477-8) is hesitant about my assumption here, and argues that many artistic practices are not like solving a coordination problem. Such remarks do not threaten the argument of this paper because my argument is more limited than his, and I only argue that one of these practices, the critical categorization of works, is a coordination problem.

  5.   To the question of what makes a given kind an art-kind, Xhignesse (2020) takes an approach like mine, appealing to the existence of social practices. However, while I rely on Guala’s institution, Xhignesse prefers Millikan’s convention as the explanans.

  6.   There is a dominant but inefficient equilibrium in the prisoner’s dilemma, so it is not a coordination game because the problem of choosing among multiple equilibria does not arise. In the classic example, the choice to remain silent together is efficient but not an equilibrium.

  7. Evolutionarily stable strategy (ESS) was introduced by John Maynard Smith and George R. Price to explain natural selection. A strategy S is evolutionarily stable in a particular group if and only if every player selects S; selecting alternative strategies would not improve one’s reproductive fitness; that prevents mutant strategies from invading.

  8. I would like to thank John O’Dea for helpful discussions. Thanks also to anonymous referees for insightful comments.

Ian Verstegen

↑ Back to top ↑

Cognitive psychology has expressed its debts to Gestalt Psychology and Conceptual Metaphor Theorists (CMT) such as George Lakoff have expressed debts to Gestalt psychologists, like Rudolf Arnheim. However, there are prima facie obstacles to this easy genealogy, especially the Gestalt preference for an interaction theory of metaphor. This paper addresses these issues by, firstly, revisiting the discussions of metaphor by Gestalt-oriented psychologists and comparing them to CMT. Secondly, the paper discusses the ways in which CMT has appropriated Gestalt ideas, usually as a ‘translation’, but not a true assimilation. Lastly, the paper focuses on Lakoff’s discussion of a static image using CMT and uses the insights of Gestalt-theoretic critiques of CMT to explain its shortcomings in the visual domain.

Fig. 1, Rembrandt, Supper at Emmaus, 1648, Louvre (Wikimedia: image in public domain)

1 Introduction

Conceptual Metaphor Theory (CMT) has proven to be a highly durable theory of expressive language. It explains expressive language as something based on bodily knowledge, that is formed by lived experience. Cognitive psychology has expressed its debts to Gestalt Psychology and Conceptual Metaphor theorists, such as George Lakoff, have expressed debts to Gestalt psychologists, like Rudolf Arnheim. But how is Gestalt linked to CMT and to what degree is Lakoff ‘Arnheimian?’ In particular, what are we to make of Lakoff’s renderings of spatial analyses of paintings that have been undertaken in the manner of Arnheim in his own language of CMT?

The argument of this paper is that Arnheim’s position is more complex than CMT’s, and should not be regarded as a simple stepping stone towards it. Indeed, there is more work in the domain of contemporary metaphor theory, equally inspired by Gestalt theory, that has developed important critiques that lend further support for Arnheim and his psychology of art. By clarifying Arnheim’s position on metaphor, we can see more clearly some of the simplifications and only apparent improvements that CMT has contributed to metaphor theory.

This paper answers this question through a series of discussions. Firstly, it compares CMT with the Gestalt-Interaction theory of metaphor (Glicksohn 1994; Goodblatt & Glicksohn 2003, 2010) to identify its consistent critiques, which include the need to recognize the bidirectionality of metaphor and the productivity of metaphoric imagery. Secondly, it reviews the appropriations of Arnheim and Gestalt psychology by CMT researchers. Lastly, it extends the critique of CMT by the Gestalt-Interaction theory in the domain of spatial imagery.

2 CMT Versus Gestalt on Metaphor

In traditional metaphor analysis, a metaphor like ‘man is a wolf’ is parsed into its tenor (or topic), ‘man’, and vehicle, that is ‘wolf’, enlivening our sense of man and vice versa. However, CMT is not only set to analyse language but also to argue for the metaphoric structure of experience itself. As an example of Lakoff’s approach to metaphor, the expression ‘things are looking up’ reflects the conceptual schema (known as a cross-domain mapping) ‘GOOD IS UP’. Here, the source domain invariantly maps (verticality) onto the target (value) and borrows conceptual significance from some aspect of experience relating to space, time, or movement.

Not only the conceptual power of the metaphor flows in one direction but also becomes somewhat ‘Kantian’, in the sense that a category is invoked and then becomes the actual meaning. That is, a metaphor would reference a conceptual schema based on experience. In a sense, the conceptual schema is more real than the metaphor because it animates the metaphor through lived experience. Once it is invoked, the metaphor merely effects the mapping of the mind. For this reason, cross-domain mappings can be interpreted as mere ‘stock responses’ (Tsur 2000). For example, ‘things are looking up’ is no different from ‘we’re rising to the top’ because they have the same schema, that is ‘good is up’. However, stylistically and semantically, they are vastly different.

This is a problem for the phenomenology of poetic language. It has always been accepted that metaphor is a literary device that is used to surpass language’s literal semantic repertoire. Arnheim (1966) explained that metaphors make concepts tangible by referring to sensory experience thereby giving the words tangible meaning. But this is different from CMT according to which the body merely serves as an ultimate source of meanings resulting in a sedimented semantics. Similarly, if metaphors direct us to the schematic mapping of experience, this thwarts the creative and expressive elements of metaphors, which are often incongruous and search after new meanings rather than stock responses. Generally, metaphors are dynamic and mutually inflecting. A novel metaphor would not work, not due to a sensed aptness, but merely because the overlaying conceptual schema are relatively similar in structure.

Specialists have pointed out further problems to do with poetic language, which stress the inadequacy of bodily schemas to service all poetic metaphors and, furthermore, the bidirectional mapping of topics and vehicles (Brandt 2013; Goodblatt & Glicksohn 2017). Some of the inadequacies of CMT have given rise to development in ‘blending’ theory, associated with Gilles Fauconnier and others (Fauconnier & Turner 2002). But this development is to bypass the venerable Interaction theory of metaphor, which gave rise to the Gestalt-Interaction theory of metaphor (Glicksohn 1994; Goodblatt & Glicksohn 2003, 2010) based on Gestalt precedents (Asch 1961; Arnheim 1966). Features of a Gestalt theory that are lacking in CMT include bidirectionality of topic and vehicle and, a related problem, a novel outcome of metaphoric signification. CMT has a permanent feature, that is the unidirectionality from target to source of metaphors. For Lakoff (2006, Lakoff & Turner 1989), the metaphor always passes from conceptual to perceptual. In this claim, he explicitly built this theory in contrast with the Interaction theory associated with I. A. Richards (1936), Max Black (1979), and the Gestaltists, for whom there is a back-and-forth movement from topic to vehicle (Goodblatt & Glicksohn 2003, 2010). Lakoff and his collaborators always resisted the Interaction theory of metaphor (Johnson 1987, 69-70; Lakoff & Turner 1989, 131-133).

The poetic element in a metaphor arises from the structural conflict caused by forcing heterogeneous elements or components of reality – that is, tenor and vehicle – into a whole. Unity for this whole is only attained through a retreat to the more abstract level of common expressive qualities. Arnheim writes,

negatively, the reality-character of the components is toned down and…positively, the physiognomic qualities common to the components are vigorously underscored in each. Thus, by their combination, the components are driven to become more abstract; but the abstracted qualities continue to draw life blood from the reality contexts in which they are presented (Arnheim 1966, 279).

A good example is given by Chanita Goodblatt and Joseph Glicksohn (2003); that is, John Donne’s poem The Bait (1633). The conceit is that love is described as fishing. Donne writes of the ‘enamour’d fish’, who will ‘amorously to thee swim’. On a standard reading that is consistent with Lakoff’s approach, the target domain of courtship is made more familiar with the language of fishing; the poetic metaphor must be enriched with the conceptual schema of fish pursuing bait. In this way, “men become suffused with the predatory attributes of carnivorous fish feeding on human bait” (Goodblatt & Glicksohn 2003, 215). However, the opposite also occurs. As indicated by the ‘enamour’d fish’, “fish are suffused with sexual desire” (Goodblatt & Glicksohn 2003, 215). The theme of courtship is not the only thing enriched; our sense of a fish’s behaviour is as well. Indeed, in Goodblatt and Glicksohn’s Gestalt-interaction theory of metaphor (Glicksohn 1994; Goodblatt & Glicksohn 2003, 2010), the process of arriving to a satisfying interpretation of the metaphor is akin to problem solving and far from simply tracing the target back to schema.

In these brief examples, we have seen that there is a long-standing contrast between the Interaction theory of metaphor in its Gestalt guise (Asch 1961; Arnheim 1966) and CMT. Turning to the visual arts, we would expect that the problems of stock responses, bidirectionality and the novelty of metaphors would be relevant in some ways.

3 CMT Appropriations of Gestalt Phenomena

A number of CMT researchers have appropriated Gestalt arguments, putting them on a proper, modern footing with the principles of conceptual metaphor theory. This, of course, suggests that Gestalt psychologists’ method of explaining is inadequate. Typically, the relationship is one-sided. Arnheim’s argument is rescued by CMT and the operation is considered one of ‘translation’ (Lakoff 2006). Mark Johnson was the first to do this in the Body in the Mind (1987, 76-79). This one-sided translation is not only true for metaphors, which have already been analysed, but also extensions of CMT to the visual arts. Subsequently, such translations have been affected by Lakoff’s theory (2006) in visual art and Maarten Coegnarts (Coegnarts & Slugan 2022) in film studies.

Because Lakoff and Johnson configured schemas as ‘Gestalts’, it was clear that they were seeking a more general basis for experience that could be applied to other domains. Because Arnheim worked predominantly on visual images, it was natural to test the relationship of the new CMT with such images. In 1987, Johnson considered Arnheim’s (1974) famous circle on a square. The most dramatic juxtaposition came in Lakoff’s (2006) re-examination of Arnheim’s (1969) analysis of Rembrandt’s Christ at Emmaus (1648) (Fig. 2).

Here is how Arnheim analyses the picture:

the religious substance symbolized by the Bible story is presented through the interaction of two compositional groupings. One of them is centered in the figure of Christ, which is placed symmetrically between the two disciples. This triangular arrangement is heightened by the equally symmetrical architecture of the background and by the light radiating from the center. It shows the traditional hierarchy of religious pictures, culminating in the divine figure. However, this pattern is not allowed to occupy the center of the canvas. The group of figures is shifted somewhat to the left, leaving room for a second apex, created by the head of the servant boy. The second triangle is steeper and more dramatic also by its lack of symmetry. The head of Christ is no longer dominant but fitted into the sloping edge. Rembrandt’s thinking strikingly envisages, in the basic form of the painting, the Protestant version of the New Testament. The humility of the Son of God is expressed compositionally not only in the slight deviation of the head from the central axis of the otherwise symmetrical pyramid of the body; Christ appears also as subservient to another hierarchy, which has its high point in the humblest figure of the group, namely, the servant (Arnheim 1969, 269).

Lakoff (2006) appreciates this description but wishes to translate it into the cognitive schemas of his theory. Here is the outcome:

a grouping is the imposition of CONTAINER schema, a bounding of a region of space with figures contained within. Arnheim describes two such schemas, one without the servant boy and one with him. In the inner CONTAINER schema, Christ is the center and highest. The metaphors interpreting this arrangement are IMPORTANT IS CENTRAL and DIVINE IS UP. Not only is Christ, the divine, the highest, but he is looking up, toward the divine God. In the upper grouping, the servant boy appears. He is painted as being in the middle of an action, serving Christ food. This puts him socially below Christ, but Christ is painted as below him, the metaphor being HUMILITY IS DOWN. The same metaphor interprets the structure of the servant boy’s body: he is bowing, tilting his body down toward Christ, showing his humility. The action of serving Christ food is metaphorical for serving Christ. The light emanating from Christ instantiates one of our culture’s basic metaphors for God: God is the source of what is good, in this case the source of light, which is interpreted via two conventional metaphors: MORALITY IS LIGHT and KNOWLEDGE IS LIGHT. The image schemas structuring the painting are oriental: HIGH-LOW, TWO CONTAINER schemas, two CENTER-PERIPHERY schemas, and LIGHT-DARK. Our conventional cultural metaphors apply to these schemas structuring the painting, to give it a meaning expressing an important aspect of the Protestant religious tradition: The ordinary person serves Christ in all humility, while Christ, the most important figure as the source of goodness and knowledge, sets the example, showing his own humility relative to people, and looking upward to God (Lakoff 2006, 156).

Following the theory of CMT, Lakoff is explaining how the composition imposes various structures or schemas (high-low, container, center-periphery, light-dark), and these are interpreted via various metaphors (‘divine is up’, ‘humility is down’). Understanding the meaning of the composition requires mapping the space via a conceptual metaphor.

Part of an apparent lack of fit between CMT and Arnheim’s Gestalt theory is perhaps due to the fact that CMT was developed for language – discrete words – and not the analogue medium of the image. However, although the result is enlightening, one can also see that it is a different kind of explanation. In line with the interactive and creative nature of metaphor, Arnheim keeps the uncertainty of the units of analysis endemic to painting alive, and we are meant to sense the competing pulls on the various figures. With Lakoff’s analysis, each figure fulfils its role in a tight system. Insofar as any number of similarly configured compositions could lead to the same meaning, the response is indeed stock.

4 CMT and Visual Images

Although Lakoff (2006) thought that his and Arnheim’s (1969) approach were more or less compatible, Arnheim shows instead how there is an active mutual accommodation between figure and ‘container’ between visual elements and their light symbolism or moral humility. In line with the Gestalt-Interaction analysis of Donne’s poem, we might say that we lose a sense of the painting when we list each figure’s relationship to its source schema. Christ is humble relative to the servant and, simultaneously, a distinguished figure to the disciples.

Fig. 2, after Arnheim (1969, 268, fig. 71a)

Interestingly, Arnheim (1969) provides a diagram that shows the tension or dynamics between the dominant line of sight and one subordinate line. Turning to Lakoff, it is as if Arnheim’s diagram of the ‘structural skeleton’ of the composition has become his actual interpretation. However, for Arnheim, the diagram is only meant to help identify the real phenomenal tensions that we are meant to see visually – “the barest scaffold of Rembrandt’s painting” (1969, 278). Here we run up into a different interpretation of the meaning of ‘embodiment’. Let us attempt to clarify this further.

For the Gestaltists, meaning is embodied by forms in their chosen medium (picture, sculpture, etc.) and, for conceptual metaphor theorists, meaning is embodied in sensorimotor experience. One has recourse to the visual medium in which the idea is realized and the other has recourse to the body in which the idea is realized. As it was made clear with the example of metaphors,

if the tenor is viewed through the filter of the vehicle, as suggested by Interaction theory, metaphor comprehension will be dependent on perception and on perceptual qualities, and not on conception and on semantic features (Glicksohn 1994, 231, my emphasis).

In order to make sense of this we have essentially three variables: the visual stimulus, sensorimotor sensations, and visual sensations. The way that theorists describe our relation to the work can be summarized in these three scenarios:

  1. The expression is only seen by activating present sensorimotor sensations
  2. The expression is inferred by imagining past sensorimotor sensations
  3. The expression is seen directly in the visual material, without the body

There are some theorists that believe that we see the tension because of sensory-motor facts. This was the old empathy theory, held by Hermann Rorschach (1942) in his Psychodiagnostics, among others. Subsequently, a version of it was developed by Heinz Werner (Werner & Wapner 1954) and others, as the sensory-tonic field theory of perception. This theory was mostly out of favour but somewhat revived in mirror neuron accounts of perception. In that case, owing to the intimate connection between visual and action neural pathways, seeing a work of art can simulate virtual motor activity in the form of motor-evoked potentials (inter-muscular stimulation). It is the evocation of sensorimotor sensations that makes the visual forms expressive; they are only expressive secondarily.

On the other hand, for CMT writers, the tension is related to concepts derived from the body via schemata. The spatial schemata is recognized in the image and interpreted as a bodily metaphor. Although the body is necessary for cognition, it is only a means to an end. Sedimented ideas of sensorimotor knowledge lead to inferences of metaphoric meaning.

Finally, for Arnheim and the Gestaltists, we see the tension visually as a phenomenal object, contrary to the empathy theory in which our motor simulation of bending communicates weight to a pictorial form. Secondarily, there can be a kinesthetic participation with the object (Arnheim 1966), in which case, sensorimotor sensation reinforces the visually perceived tension. However, it is principally a visual experience.

Focusing mostly on the second two options, we can see that they differ in the meaning of embodiment and the role of the body. According to CMT, the sensorimotor experiences passively (through inference) dominate over the visual, while, from the Gestalt position, vision is primary and dominates, while the body is secondary. In this difference, Lakoff might remind us somewhat of Gombrich (1960). For Arnheim, concepts are perceptual; for Gombrich, perception is intellectualized (Pizzo Russo 2005).

This schematism helps us understand the difference between a CMT and Gestalt explanation of a work of visual art. In Arnheim’s account of Rembrandt’s painting, we see directly the humility of the apostle and the aggrandizement of the servant via the phenomenal tension of the two figures within the compositional framework. In Lakoff’s account, we first need to observe how the scene is structured by schemas and infer via interpretation which cultural metaphors are in play.[1]


Arnheim, Rudolf, Toward a Psychology of Art (Los Angeles: University of California Press, 1996).

Arnheim, Rudolf, Visual Thinking (Los Angeles: University of California Press, 1969).

Arnheim, Rudolf. ‘Perceptual analysis of a Rorschach card’, in Arnheim Rudolf (ed.), Toward a Psychology of Art (Los Angeles: University of California Press, 1966) 90-101.

Asch, Solomon E. ‘The metaphor: a psychological inquiry’, in Mary Henle (ed.), Documents of Gestalt Psychology (Los Angeles, University of California Press, 1961), 324-333.

Black, Max, ‘More about Metaphor’, in Ortony, Andrew (ed.), Metaphor and Thought (Cambridge: Cambridge University Press, 1979), 19-43.

Brandt, Line, Metaphor and the Communicative Mind (Newcastle: Cambridge Scholars Publishing, 2013).

Coegnarts, Maarten, and Slugan, Mario, ‘Embodying Meaning Visually: From Perceptual Dynamics to Motion Kinematics’, Art & Perception (published online ahead of print 2022). doi:

Fauconnier, Gilles, and Turner, Mark, The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities (New York: Basic Books, 2002).

Glicksohn, Joseph, ‘Putting interaction theory to the empirical test: Some promising results’, Pragmatics & Cognition (1994) 2, 223-235.

Gombrich, E. H., Art and Illusion: A Study in the Psychology of Pictorial Representation (London: Phaidon, 1960).

Goodblatt, Chanita. and Glicksohn, Joseph, ‘From Practical Criticism to the Practice of Literary Criticism: Metaphor Comprehension Online’, Poetics Today (2003) 24, 207-236.

Goodblatt, Chanita, and Glicksohn, Joseph, ‘Conversations with I. A. Richards: The Renaissance in Cognitive Literary Studies’, Poetics Today (2010) 31:3, 387–432.

Goodblatt, Chanita., & Glicksohn, Joseph, ‘Bidirectionality and Metaphor: An Introduction’, Poetics Today (2017), 38:1, 1–14.

Johnson, Mark, The Body in the Mind: The Bodily Basis of Meaning, Imagination, and Reasoning (Chicago, IL: University of Chicago Press,1987).

Lakoff, George, ‘The Neuroscience of Form in Art’, in Turner, Mark (ed.), The Artful Mind. Cognitive Science and the Riddle of Human Creativity (New York, NY: Oxford University Press, 2006), 153-169.

Lakoff, George, and Turner, Mark, More than Cool Reason: A Field Guide to Poetic Metaphor (Chicago: University of Chicago Press, 1989).

Pizzo Russo, Lucia, ‘Percezione e immagine nella rappresentazione artistica’, Rivista di Estetica (2005) 45, 245-268.

Richards, I. A., The Philosophy of Rhetoric (New York: Oxford University Press, 1936).

Rorschach, H., Psychodiagnostics: A Diagnostic Test Based on Perception (New York: Grune and Stratton, (1942).

Tsur, Reuven, ‘Lakoff’s Roads not Taken’, Pragmatics and Cognition (2000) 7:2, 339-359.

Werner, Heinz, and Seymour Wapner, ‘Studies in physiognomic perception: I. Effect of configurational dynamics and meaning-induced sets on the position of the apparent median Plane’, Journal of Psychology 38 (1954), 51-65.

  1. I am grateful to Joseph Glicksohn and Chanita Goodblatt for discussing this paper.  

Rebecca Wallbank

↑ Back to top ↑

In the autumn of 2021, I sat down over zoom to interview Ralf Webb, whose debut collection of poems Rotten Days in Late Summer was published by Penguin that year. It was shortlisted for the Forward Prize for Best First Collection and is a Telegraph and Irish Times Book of the Year. Webb’s poems and critical writing have appeared in the London Review of Books, Test Centre, The Poetry Review and the Guardian. During 2020-21, he ran PoetryxClass, an Arts Council-funded reading group and seminar series and, in 2022, he will be a writer in residence at the Jan Michalski Foundation in Switzerland.

Rotten Days in Late Summer (2021) narrates a raw, complex and deeply personal account of youth, love, class, grief and death. The Guardian described the work as ‘impressive… tender, unflinching’ (Wilkinson 2021); Alex Dimitrov described it as ‘direct and heart-breaking’ (Webb 2021). I am neither a poet nor a poetry reviewer, but I found the collection strikingly honest, even brave. There are poems within this collection that resonated with me, educated me, and really moved me in a way that no other written work had for a long time. This is why I reached out to Webb for an interview. I wanted to know what his intentions are when writing, whether there is something specific that he wants the audience to feel or experience, whether he feels that there is something that the audience owes him, and whether there is something that he feels he owes the audience in turn. At the very least, it seems that, as audience members, we have a responsibility to do the work justice. This responsibility reflects and, indeed, hinges on the possibility of trust between poet and audience (Nguyen 2021). The nature and significance of such trust is one of the central themes of this interview. Within this interview, I also tied these themes to broader questions concerning artistic license, artistic creation and the concept of ‘poetic appropriation’ introduced by Anna Christina Ribeiro (2009).

I want to start with a two-part question. The first part: what is your intention when writing? Is the intention primarily to get the audience/reader to feel something? Is it to get them to believe something? Is it perhaps both of those things?

Alternatively [and this is the second part of the question], is your intention when writing less about the audience and more about you – less about what you hope the audience will experience and more about what you hope to gain as a writer. Is it more an opportunity for you to articulate something (in an almost cathartic manner) than it is an opportunity for an audience to hear something?

I’m focused on both the reader and myself, but to what degree depends from poem to poem. [In some ways, the question is hard to answer because I often think that the reader is myself].

I think the origin of my desire to write comes from a desire to communicate, but it’s a very abstract, unformed, amorphous desire. It’s not like I want to communicate to a particular group of people, or that I want to reach a particular set of people.

You asked about wanting to get people to believe something or wanting to instil them with a particular affect. It’s rarely that specific. There are some poems that I have written, which are more politically oriented; for sure, there is a desire to get the hypothetical reader thinking about things in a certain way, but it’s not to get them to believe in something. The intention may be to get them to reconsider something that they might take for granted, but that is really on a poem by poem basis. I don’t believe in the idea that poems are just arguments, statements of belief, or assertions of some particular set of values.

A lot of what you communicate with your poetry is very personal. Actually, a few of your poems made me cry. This actually happened to me whilst I was at the hairdressers. I had to stop reading and apologise to them. [laughs]

[Laughs] Well, that may be the best review I have ever had.

I wondered whether – because the poems are so personal – you ever feel that the wider reader owes you something: a specific kind of attentive stance or an interpretative duty?

And, I guess, there is a related wider question emerging here as well: is there a kind of interpretative freedom that the readers are permitted, or is the meaning more fixed? Are we allowed, as readers, to bring ourselves (and our interpretations) to it, or is the reader’s task more about ‘getting it right’?

That’s super interesting. The first part of your question was whether I feel the reader owes me anything. I don’t think so; I don’t think the reader owes me anything. I think there is a list of things I would want, hope or perhaps expect from the reader, but I don’t think I’m owed anything by virtue of how personal some of the poems are or seem to be.

But, you know, that has been put to the test a little bit for me with this debut collection. I’ve had to encounter readers’ real life opinions about my work, which actually exists in the world, and they’re not universally complimentary. [Laughs] I’ve read a couple of unflattering things about the book online and I’m just like, ‘but did you just… did you think about it properly?’ It’s interesting. In a way, I am expecting that if they spend longer with the book, then they’ll think it’s better.

So, what do I want from the reader? I guess, what I want from the reader is for them to approach the book, or the work, in good faith; for them to approach the work on its own terms. As a reader, I think you should come to the work without prejudice, essentially, and treat it according to its own terms. So, if the reader came to the book and was hoping for poems about birds and got really annoyed that there were no poems about birds, then they wouldn’t be treating it on its own terms. That’s an extreme example but, hopefully, it illustrates my point!

Can you remind me what the second part of your question was?

Sure! So, a lot of your poems are quite personal and they reference things that you’ve experienced. I wondered whether there is still room for the reader to bring their own interpretation here.

Yeah, I think the answer is: absolutely. I feel creative writing as a whole (and probably all other art forms) is an inherently mysterious process. I can’t explain what goes on between the brain and pen or keyboard, and I cannot always explain the origin of a particular poem. I can’t even capture how much of it I’m in control of. It is inherently mysterious. So, regardless of how much you craft something – for me – there should always be room for subjective interpretation.

Interesting, ok.

A good case in point: Rotten Days in Late Summer was recently reviewed in The Poetry Review (2021), and the reviewer picked up on a recurrent image throughout the book – that of mouths, open mouths – that I was completely unaware of and that even my editor was completely unaware of.

Ok. So, in your view, we can [as readers] bring something to the work, and there is some interpretative freedom – that is, the reader can ‘bring themselves to the poem’. Interesting. I wonder whether this is only permissible to a certain degree.

Let’s turn now to another somewhat related, strand of thought within the philosophy of poetry that I would like to hear your thoughts on. It argues that good poetry speaks to the audience; it generates in the reader the sense that they could have written it themselves, if only they had the talent to do so (Ribeiro 2009). The idea is that the words express something that we have felt or thought, or are on the cusp of feeling of thinking, but we just don’t know how to articulate it. So, poetry is meant to reflect our experiences or emotions.

I’m raising this strand of thought because I wonder whether it is too much to say that such an experience is the mark of good poetry. I worry that it seems to encourage the external reader to use poetry for their own personal gain. I worry that it is a problematic attempt by the wider reader to take the poem for themselves and make it theirs when it is not; it’s the poet’s. The work is the poet’s own personal expression, and maybe we [those who are not the poet] should just appreciate it for what it is.

Perhaps, the matter depends on the particular piece. But surely not all poetry is good in virtue of its power to relate to our own experiences or make us ‘feel seen’.

One term that is used in this context is ‘poetic appropriation’. [To clarify, this is the idea that it is acceptable, even desirable, for the reader to appropriate the poet’s words and approach the poem in terms of how it bears on their own lives]. Poetic appropriation is described as a positive thing. I just wonder whether the term appropriation kind of already indicates that it is maybe something that we need to be cautious of.

[Laughs] Sorry, I realise that I’ve asked a question, and then I’ve told you what I think about it! That couldn’t have been more leading!

This reminds me of a quote, which I like using at any given opportunity, partly because I’m lazy, and partly because I think it’s a really good quote. It is by poet Ilya Kaminsky, who had a book out called Deaf Republic (2019). The quote is: “a great poet is someone who is able to speak privately to many people at the same time” (Kaminsky, 2013). I kind of bastardised it recently into the idea that ‘poetry is a private language shared’, which sits adjacent to what you’re saying. The idea being, that the poem is a medium able to articulate that which seems inarticulable. I think there is a lot to that idea.

That said, I think that there’s no way of judging the ‘success’ of a poem, whether a poem is ‘good’ or ‘bad’ by this measure. Everyone thinks different things. Everyone has different opinions and different life experiences. There are some people for whom the poem will just not connect – you know – it won’t articulate anything for them. I guess what I’m saying is that we should be wary of presuming there’s an objective set of experiences, affects, or desires that cannot under normal circumstances, be articulated, but that poetry is somehow uniquely able to express.

Okay. Yes, that makes sense. Maybe I can concede a little.

Well, no. Not at all. The term ‘appropriate’ is – I guess – quite a loaded one.

I think part of the worry here – which the term ‘appropriation’ brings out – is that there is a danger that the reader/audience member would just be taking on board the poet’s experiences as their own, using the poem for their own personal benefit in a manner that fails to respect the poet and do justice to the poem.

I think you are right. There is some merit in a work of poetry being able to speak to people. I guess I wondered if there is a limit to which people can do that – whilst still respecting the very unique and personal thing that the poet has expressed.

Wow! That’s a very interesting idea. I can definitely see how ethical limits could apply when you’re looking at poems, whose subject matter lies at the intersections of race, class, gender and various kinds of marginalised, lived experiences. An example, I was trying to form in my head was like, if, say, a working class poet is talking about pretty extreme levels of poverty in their poem and then a very wealthy reader is like: ‘oh, that reminds me a lot of a bad afternoon I had when the fancy wine place ran out of my favourite bottle of £60 chardonnay’. Does that apply as an ethical limit? Does that make this person a ‘bad’ reader? It makes them a comical and ridiculous reader, but a ‘bad’ one? I’m not sure what I think about that. It opens up knotty questions to do with how or if you can quantify oppression and empathy.

Speaking generally, though, I don’t know that there are any limits.

That’s really interesting. So, I’ve asked you a little bit about whether we, as audience members, have duties to poets. However, do you, as a poet, ever feel you have a duty to the audience in turn?

So – as a bit of background – something that I’m particularly interested in is the concept of trust and the levels of trust between the artist and the audience. It’s been claimed that the artist might often trust the audience that treats their work with a kind of interpretive and attentive justice (Nguyen 2021). Similarly, it has also been claimed that the audience might trust that the artist has created a work that is worth investing time in and is also ‘authentic’ in a certain sense (Ibid). The idea is that the audience members trust that the poet hasn’t simply sold out to the publisher’s demand. And I guess I wondered if you felt any pressures or any duties of these kinds to an audience.

On a related note, I would also be interested in hearing your thoughts on what ‘authenticity’ might mean in this context. Your work is quite personal and reflective of your own experiences. Would you feel comfortable writing in a manner that’s less reflective of your own experiences, perhaps more fictional? Is writing in a manner that is personal and reflective of experiences that have been lived by you part of what it is to be authentic? (Sorry, huge cluster of questions! [Laughs]

That’s a huge question. The first part of your question is around audience trust and whether I feel that I owe the audience something. I don’t know. I mean – in real terms – it can be quite difficult to create the conditions, within that setting, to form a proper connection with the audience to create something authentic or to create something ‘real’. I think that’s because in my experiences of those types of settings (I’m thinking of real-life readings, or on Zoom) are always quite vexed. People are coming to them with different sets of expectations and sometimes with vested interests. And these expectations are all articulated by a set of social codes that some may feel alienated by, that some may be in a position to manipulate and exploit, that some may be actually in the process of creating – social codes which are very much to do with class.

But if we’re talking in the abstract, like, you know, the ideal audience and the ideal reading situation and that kind of thing, I think, yes, my aim would be to deliver the work or orate the work in a way that can sustain some kind of affectual connection with the audience. I think it does require an element of trust in them, and they will have to have a certain amount of trust in me. Poetry readings are chaotic situations. Anything could happen; and anything does happen. Like, really weird stuff happens at poetry readings.

And do you think that poetry is particularly well-suited to talking about the personal?

I feel that poetry is such an exceptionally strange form, which I think has suffered for decades for not being demarcated into genres. One could demarcate different genres and sub-genres. There are competing sets of aesthetics, different philosophies around how language itself should be treated within poetry – philosophies which are by turns metaphysical, scientific and political – and these can be based on inherited manifestos from tutors at universities; or from organic, historic evolutions within certain scenes, over years, and in response to completely varying canons and alternative canons and schools and cliques… So, it’s all quite weird and complicated and there’s a lot of tussling over how poetry ‘should be’.

However, I think poetry is particularly well-suited for talking about the personal. And there’s obviously a long lineage of it, going way back to Sappho or Catullus, and then tracing it forward to the Romantics, and then on through the twentieth century. There’s a long lineage of relatively compressed expressions of personal experience.

But I don’t think writing about one’s personal experience is the same as writing auto-biographically or factually. I think there’s a difference. I forget which poet said it – maybe it was Sharon Olds – who said that her poems are ‘apparently personal’, which I think is pretty accurate. I think you can be authentic and personal without being factual.

Oh, interesting. I hadn’t thought about it like that before, but that makes sense. I guess the poet (famously) has a licence to play with the facts and, perhaps – if I understand you correctly – by doing so, they can often create or express the thing that they are trying to say most clearly. That’s really interesting.

Ok, so I have one final set of questions. I wanted to ask you about your creative process; how you play around with different literary techniques and the decision-making process that you use to implement these techniques. Do you play around with degrees of opacity? I guess every decision you make is loaded, including all the decisions about the formal characteristics of the poem, such as spacing and where to place the lines on the page. You have large amounts of freedom, so how do you make those decisions, and what motivates them?

Poetry has – to borrow a phrase I used before – a kind of micro-architecture.

So, even the placement of commas or dashes – all of that – is contributing to the greater whole of the poem. Which is actually odd in a way, considering the aural/oral origins of the form. I guess, my decision-making process around the poem’s micro-architecture is both oral and visual.

For me it is really a question of playing around endlessly, or almost obsessively, with line breaks, rearranging words, and setting new words next to each other. I find it very calming, and it is quite a mechanic process.

I was perhaps unduly influenced by a documentary I once saw about William Carlos Williams. There’s a scene about how he worked on a typewriter and how he was obsessed with the way it emulated machine work. That always stuck with me. There is this kind of machine-like quality to the constant rearranging and shifting and changing parts within the poem. This is where it often feels to me that ‘craft’ is involved, that poetry does involve the execution of a craft-bases skillset that you could spend a whole lifetime honing.

In terms of being opaque and whether that is a choice, it definitely can be at times. The feeling that you want to make a particular passage sound a bit more mysterious, ha! I don’t know if that necessarily equates to good writing, but it does happen. John Ashbery once said something about the meaning of a piece of writing being a frequency that you can tune into. I agree with that.

Would you say that the decisions you make are quite instinctive (even if you go back on them and change your mind)? I assume you are not following rules or a set standard; you just go on what looks or feels good.

Yeah. It is quite instinctive. But, there are – in the English language anyway – ghosts.

There are forms, which haunt any writing of poetry, particularly poetry that has a first person pronoun, such as the Sonnet. These ghosts can be useful. They can be ignored but not entirely exorcised.

When you are writing you are aware that there exist certain forms that can, if you want, be used very loosely. You can mess them up or work with them or against them. So, these forms become something like a leash; one that lets you roam around without getting lost.

Fascinating. So, how do you know when you are ready to finish writing? How do you know when you are done?

Yeah, that is a good question. I mean, everyone’s process will be different, but there’s a feeling I have where. I think a poem is finished; where it kind of becomes inert to me. I’ve used the word ‘ossified’ before, where it kind of becomes static. So, it holds no charge for me any longer, and then I think it’s done, which is not the same as saying that I think it’s good. It’s a feeling that I’ve either gone too far, and now it is dead, or just far enough, and there’s nothing more I can give.


One of the things that I found most striking about my interview with Webb is that he permits the audience to a higher level of interpretative freedom than I anticipated. In many ways, poetically-naïve audience members, such as myself, should be relieved as there is little pressure to get it right. Here, the meaning of the work is a ‘frequency’ which we are simply welcome to ‘tap into’. The trust that Webb places in the audience primarily relates to an expectation that the audience will approach the work with the requisite level of attention and respect, and the correct interpretation of the work is somewhat secondary. There is something wonderfully liberating about this. However, how liberated are we exactly? And what are the philosophical implications of this freedom?

In turning to the first of the above questions. I cannot help but wonder if there is a helpful comparison to be drawn between poetry and works of scored music.[1] Performers of a work of scored music must meet a certain standard of correctness when interpreting the score if the performance is to count as a genuine performance of that work. But, some works of scored music have a notation which permits – and even encourages – a high level of interpretative freedom within the framework set out.[2] Such works contain fewer instructions and may even contain the instruction to improvise for a few bars. Other works are captured by a notation, which leaves the performer very little scope to play. I wonder if the audience members of a work of poetry are somewhat comparable to performers of a work of music, and the words on a page are somewhat comparable to notational scores. If so, it remains the case that there is a framework of interpretation that the audience of a work of poetry ought to respect in order to engage with that work genuinely, and the degree to which they have interpretative freedom may vary. I am not claiming that written poetry is a notational instruction nor that readers of poetry are performers.[3] The claim is simply that a helpful comparison might be found here.

Another aspect of my interview with Webb, which I found particularly interesting, was his openness to poetic appropriation. We are welcome to bring something of ourselves to his work. We are permitted – even encouraged – to let the work speak to our experiences and shape our understanding of what we have been through. In light of Webb’s response, I have a greater understanding of Riberio’s (2009) account of poetic value. Nevertheless, there is a part of me that wants to express caution. I still wonder if there is a limit to which audience members can permissibly appropriate the poem, particularly if what the poet has expressed is personal. I want to avoid a philosophical approach, which seeks to use the artist or their work for our own personal gain. In philosophy of art, we tend to ask: ‘what can I learn from this artwork?’, or ‘what can I gain from it?’. However, an artwork is not always for us. It seems possible that some artworks are created primarily for the artist as a simple act of expression. When this happens, as audience members, we are lucky to gain access to the artist’s expression. We are lucky that the artist trusts us, but the work is not ours to take.


Goodman, Nelson, Languages of Art: An Approach to a Theory of Symbols (Indianapolis: Bobbs-Merrill, 1968).

Hamilton, Andy, ‘The Art of Improvisation and the Aesthetics of Imperfection’, British Journal of Aesthetics (2000) 40: 1, 68–85.

Kaminsky, Ilya, ‘Of Strangeness That Wakes Us: On mother tongues, fatherlands, and Paul Celan’. Poetry Foundation. (published online 2013) accessed 24 May 2022.

Kaminsky, Ilya, Deaf Republic (Minneapolis, MN: Graywolf Press, 2019).

Kivy, Peter, The Performance of Reading: An Essay in the Philosophy of Literature (Malden, MA: Blackwell Publishing, 2006).

Nguyen, Thi C. (2021). ‘Trust and Sincerity in Art’. Ergo. 8:2. doi:

Ribeiro, Anna Christina, ‘Toward a philosophy of poetry’, Midwest Studies in Philosophy (2009) 33:1, 61-77.

The Poetry Review ‘Rotten Days in Late Summer’, The Poetry Review (2021) 132 accessed 24 May 2022.

Webb, Ralf, Rotten Days in Late Summer (Dublin: Penguin Random House, 2021).

Wallbank, Rebecca, ‘A Literary Aesthetic of Perfection and Imperfection’, in Andy Hamilton and Lara Pearson (eds.), The Aesthetics of Imperfection (London: Bloomsbury, 2020).

Wilkinson, Ben, ‘The Best Recent Poetry – Review Roundup’ The Guardian (published online 2021) accessed 24 May 2022.

  1.  See Wallbank (2020).

  2.  See Goodman (1986), Hamilton (2000).

  3.  For this kind of claim, see Kivy (2006).

Fotini Charalabidou

↑ Back to top ↑

Looking Through Images: A Phenomenology of Visual Media” is the title covering the front of one of Emmanuel Alloa΄s books. As a phenomenological reading of it indicates, the title reveals much and raises enough curiosity to be characterized as enigmatic. On the one hand, it offers insight into the book’s two main themes; it announces that at stake is a phenomenology of images, according to which images are looked through, as well as a phenomenology of visual media in general. Furthermore, it provokes the suspicion that the former is to be put in methodological priority over the later. On the other hand, it creates curiosity as to how this “looking through” is conceived, along with how to understand the manner in which the description of this phenomenon can shed light upon the experience of visual media as well.

As Alloa sets up the main philosophical question to be addressed in his book, he draws from the observation that in our contemporary lives we are abundantly met with images (2021, 1). As he furthermore notices, although we seem to hold an intuitive understanding of what it means for something to be an image, we would be hard pressed if asked to articulate it. He therefore asks: What does the pictoriality of images consist of? What is the function of an image? What is it that makes images, images? These function in the text as different ways to ask one and the same question: “What is an image?” (2021, 2).

The book consists of five chapters. In the first chapter, Alloa offers an interpretation of Plato’s historical move from Sophistic and Eleatic philosophical doctrines, towards the demarcation of what he characterizes as the “true philosophy of the image” (2021, 16). According to this interpretation, Plato aimed to pinpoint the manner in which the Sophist is lead to hold simultaneously that images shall be considered as being and that they should be thought of as not being. This polarization is a consequence of presuming both that images essentially are the same as their models and that, in parallel, they are not, in the sense that they seem to differ from them. Plato understood that the Sophist argues under the premise that “the measure of the being of the image is the entity it depicts” and that it is this assumption which lays ground for this opposition between the two theories of images as being and not being (2021, 18).

As Alloa writes, Plato’s obtained awareness of this ontological commitment made by the Sophist, alongside his conviction that all speech refers to the being, allowed him to alternatively employ the categories of methexis and difference (2021, 19, 20) He thus conceived of an image as being, inasmuch as it is essentially similar and different from its model, instead of opposite to it (2021, 19-34). Diving into the inquiry about the ways in which various images are similar and different with respect to their models bears the seed of awareness of the truth that such research is possible only in the event of us as preconceiving images as concerned with their own appearance. This reading of ancient Greek philosophy prompts Alloa to set the description of the phenomenal qualities of images as the historical starting point of reflection which concerns the concept of the image (2021, 34-39).

The second chapter offers a roundly informed reconstruction of Aristotle’s definition of the medium, which is historically marked as the first attempt to construct a general media theory and which took on perception as its starting paradigm. As Alloa writes, in the Aristotelian description perception can only take place through a medium, which connects the perceiving eye with the perceived object, while maintaining their distance (2021, 81, 82). This medium is thought of by Aristotle as a space, which is characterized by the capacity to receive and transmit forms, without the matter of things from the object of perception to the sense organ, thus giving rise to appearance (2021, 83-85). Importantly, Alloa also highlights, that in the Aristotelian doctrine any medium is far from being theorized as transparent. It rather “shines through” things and endows their appearance with its own character (2021, 76-80).

In the third chapter, he provides an interpretation of the historical reception of Aristotle’s notion of the diaphanous medium, as a “history of forgetting the medium of appearances” (2021, 11). As he suggests, theories of perception throughout the Middle Ages and till the Renaissance received and interpreted the concept of the diaphanous medium by the influence of metaphysical, epistemological, theological, as well as structural reasons (2021, 106-147). More particularly, they were polarized between endorsing the transparency thesis, namely a conception of images as defined by what lies “behind” them, or the opacity thesis, which posits that images are fully determined by their material objecthood. They thus proposed various accounts about the optical, physiological, and physical laws of perception and at the same time forgot about appearance and mediality (2021, 147-155).

In the fourth chapter, Alloa provides a brief introduction to Edmund Husserl’s phenomenology, with a special focus upon his transcendentalism about appearance. As he points out, the theory at issue specifies that the ego brings into appearance things that either exist or do not exist (2021, 178). Additionally, it draws from a presupposition of three moments of phenomenalization: namely the addressee (ego); the mode of appearance (cogitatio); and the appearing object (cogitatum). It also acknowledges a need for a medium to connect the cogitatio with the cogitatum. On the face of it, the doctrine proposes, that it is the ego that mediates between the other two moments of appearance (2021, 171, 172). It becomes evident, then, that the Husserlian theory about appearance is starkly divergent from the Aristotelian, as presented in the second chapter, yet remarkably it remains highly influential in contemporary thought. It is for these reasons, that Alloa worries that Husserlian transcendentalism about appearance stands in the way of reconciling our current phenomenological theory with the Aristotelian theory of the medium (2021, 173).

In the rest of this chapter, Alloa aims to secure the capacity for such a reconciliation. To this purpose, he presents Sartre’s theory, which proposes that consciousness of images posits its object as non-existent (2021, 178). He thereafter highlights and interprets the difference between Sartre’s and Husserl’s views. It is in this manner that he ends up endorsing an understanding of the object of appearance as neither actively represented by consciousness, nor purely given to it after all (2021, 180). He thus appeals to the need for a corresponding medial phenomenology and revises Husserlian transcendentalism. In three additional units of this chapter, Alloa furthermore introduces further contemporary doctrines– those of Eugen Fink, Jacques Derrida, and Maurice Merleau-Ponty – which have already sought after a construction of such a medial phenomenology (2021, 192-208).

Alloa’s interpretation, alongside his joining of traditional phenomenology with the Aristotelian theory of the medium, contains playful and insightful moves which feel philosophically fertile and exciting. His thorough historical knowledge concerning disputes around images and the Aristotelian theory of the medium, and his convincing interpretation of these allows for a philosophy of images to finally attend to and understand its own origins and internal struggles, so as to become mature enough to exceed itself and grow further. Growth, here, is understood as the rise of willingness to admit that everything in the history of our conception of images indicates that it just makes sense to think of them as media in the Aristotelian sense of the term. It also amounts to the enhancement of the courage which is necessary for describing the character of the medium, which essentially eludes the focus of attention.

The fifth chapter bears the heading “Media Phenomenology” and is intended to offer a phenomenology of images (2021, 11). In this part, Alloa introduces a distinction between so called “replete” and “discrete” media. He specifies that, in the case of the former, the medium is inseparable from the content, which implies that, in the event of it bringing about the slightest phenomenal difference, this difference shall hold significance in its overall transmission. In contrast, in the case of the latter, the medium of transmission rather disassembles, transports and reassembles the content at hand and, in this respect, it is separate from the transmitted content (2021, 228-233). Along with this conceptual distinction between the two kinds of media, Alloa specifies that images are to be submitted under the category of replete media and deserve to be phenomenally described as such (2021, 233-238). Alloa distinguishes between these kinds of media in a clear way and convinces us that this move is necessary for better understanding the plurality of media, images included.

In order to phenomenally describe images as replete media, he sets to reject the thesis that images can be defined by reference to necessary and sufficient conditions. Instead of opting for a theory about such conditions, he draws inspiration from Nelson Goodman’s methodological shift from the question “What is art?” to the question “What are the symptoms of art?” and implements the same idea (2021, 238-242). He thus shifts the attention towards the development of a “symptomatology” of iconicity, namely to the provision of a list of symptoms of images, the presence of which rather indicates that one is possibly met with an image, without necessarily – let alone sufficiently – entailing so. As for three characteristic examples, Alloa lists “framing”, “synopticity”, and “presentativity” (2021, 242-275). This symptomatology is developed by reference to various works in the literature concerning the special properties of images and is remarkably careful and perspicacious. It does great justice to first person phenomenal experience of images and transmits the wild joy accompanied by the accomplishment of articulating phenomena which resist being articulated.

Upon critical reflection, it is worth pointing out that Alloa seems tempted to identify either replete or visual media with iconicity. This inclination is rendered most evident by choosing the heading “Media Phenomenology” for a chapter which mainly concerns the phenomenology of iconicity. Furthermore, he lists several kinds of replete media, such as images, voice, dance, and gestures and – even though he does not expand on this – one imagines that he would admit that there are various kinds of visual media as well, such as images, videos, holograms etc. However, identification of the symptomatology of images with either the symptomatology of replete or visual media in general implies that the genus is considered the same as one of the kinds that falls under it and this is a consequence to be avoided. Additionally, it runs against the premise that – ordinarily – the concepts of iconicity of visual media are employed differently.

In addition to an identification of these concepts being somehow arbitrary, the effort to account for the symptomatology of iconicity leads to the unwanted inference that the symptoms of iconicity are identical with the symptoms of either all replete or all visual media. As a consequence, we find ourselves lacking criteria for a “differential diagnosis” of replete and visual media in general on the one hand, and for each visual medium separately – including images – on the other hand. This implies an incapacity to account for what makes images images and to fulfill what was set as a main goal in the introduction of the book. Αs Alloa himself mentions: “If everything is an image, nothing is an image anymore […].” (2021, 239).

Furthermore, if we endorse Alloa’s premise that images constitute one amongst a plurality of replete or visual media, then we must also accept both that all images are replete or visual media, and that not all such media are images. With that being assumed, it would be preferrable to obtain knowledge about the symptoms of iconicity, starting from the symptoms of replete or visual media, rather than the other way around.

As for one final remark, in the introduction of the book Alloa promises to dismiss the view that it is possible to account for the conditions of necessity and sufficiency for images. However, the rejection of this thesis functions as a working assumption in the text rather than as an inference (2021, 12). As a consequence, for the reader who is relatively unfamiliar with philosophy of images, it may not be all that clear why, for example, the symptoms of framing and synopticity together are not sufficient for iconicity.

These points of criticism deserve to be seen as indicative of the potential to immerse ourselves in Alloa’s thought, to cherish his insight upon images’ replete mediality, and to work our way towards the formulation of new theoretical attempts which do justice to images’ functioning as media, while simultaneously rendering clearer the distinctions amongst the various kinds of visual media, thus defining images more precisely. Under these considerations, I wish to maintain that the book at hand offers ample opportunity for discussion and to recommend it for thinkers interested in the question of images and mediality.