We talk of verbal, situational, and dramatic irony. How closely related are they? Very closely. I’ll present a theory of these three, plus a few derivatives to be introduced later. The theory won’t cover all or even most of what gets called “irony.” Nor should it. We regularly complain that the idea is invoked to cover “virtually any gap, fracture, resistance, or deferral” (Silk, Reference Silk and Silk1996: 355).Footnote ¹ An attempt to theorize all this inflationary material would be faced with a shapeless mess. So I trade usage for structure. Nonetheless, I aim to cover quite a range of use: from the manifestly ironic “that was a success” said after a disastrous meeting, to complex narratives where reasonable people may disagree about whether the author is ironizing their own activity. I also recognize that there are no sharp boundaries here, and that interesting phenomena will end up as “near misses” from the point of view of my theory; one such near miss gets some discussion.

Why bother with this sort of conceptual geography? Beyond displaying some unobvious connections between the forms that irony can take, there is interest in seeing why certain patterns of ironic narration and ironic drama appeal to fiction makers and their audiences. Many of my examples come from fifth-century Athenian tragedy, and I end with a somewhat extended discussion of irony in film.

Ironic Situations and Their Representation

I start not with irony but with its relative, tragedy. “The play was a tragedy” you say, exploiting an amusing ambiguity. It might be tragic in so far as it is a representation of events which brings out what is tragic about them, or because the play itself (more likely its performance) constitutes something tragic. In saying that events and their representations can be tragedies, we are not ascribing the same property to both kinds of things. There is polysemy here: distinct but intelligibly related meanings.Footnote ² Something like this holds for irony; a play or novel that represents a situation as ironic offers us another form of irony – that, I claim, is dramatic irony. The irony of a situation is one thing, dramatic irony another but closely related thing; it is the representation of an ironic situation.

If this is right, situational irony is the root notion, something I have not yet explained. I start from the thought that situations of irony occur when someone or something (usually some project) fails to meet a goal or expectation.Footnote ³ Often this arises from some limitation of insight, understanding or will, but it can occur simply because of chance, fate, or the misfortunes of a hostile world. It was ironic that the explorers were just a mile from the food store when they gave up; it was ironic that the sign “No shooting” was made unreadable by bullet impacts; it was ironic that the dog bought to protect us from vandalism wrecked the house. More detail will help pinpoint the failure involved; were we just unlucky with our choice of dog or did we thoughtlessly choose a dangerous breed? Either way there is irony there.

Let’s take this crude idea and connect it with dramatic irony. If we have just administered a drug to A that made him fall in love with B, it will strike us as an ironic failure of insight if A then confidently lists (what he takes to be) all the good rational reasons for his affection. Taking a situation not unlike this and dramatizing it, Shakespeare produces dramatic irony when Lysander, whose love for Helena is the product of Puck’s arbitrary magic, says

Lysander’s thinks he knows, and can rationally endorse, the grounds of his attachment, while the audience can see that he doesn’t and can’t. The play instantiates dramatic irony by representing his situation in such a way as to make manifest to us the irony in it. Not every representation of the situation will do that; it might miss the irony altogether and focus on other aspects, or describe it in such laborious detail as to spoil the ironic effect. Those would be failures to make manifest the irony. I don’t have a precise account of success in this regard; it is something like being able to display the irony, just as a painting of a woman in a red dress does not merely represent the redness but displays it. The picture gives us perceptual access to the color, and the play gives us access (not perceptual access admittedly) to the irony it represents. There is a kind of transparency here; when we look at the picture we do not see the dress, but we do, literally, see its color. We perceive it using the same perceptual mechanisms we would use were we to see the dress itself. If brain damage robs me of color vision I cannot recognize the color in either the dress or in the picture. Pictures may also be transparent to other properties. We recognize the outline shape of a familiar figure in the picture using the same recognitional capacity that allows us to recognize the person by their outline when we see them. We also have a capacity (less hard wired presumably) to recognize and respond to the irony in a situation, and when irony works in drama that is because it represents the situation in a way that engages that same capacity for irony detection. A parallel proposal works for tragedy.

This idea certainly needs more development than I can give it here. I will make one point of clarification. It is not part of the claim that our capacity to detect the irony or the tragedy in a situation preceded our capacities to detect irony and tragedy as represented. It is possible that in one or both of these cases the capacity in question developed as a result of exposure to some kind of dramatic or at least narrative representation. I doubt that this “representation first” thesis is true in general, but it is not ruled out by anything said so far, and it is surely plausible that what we recognize as ironic or tragic in real life is at least affected by our experiences of the representation in dramas, novels, and films of these qualities.Footnote ⁴

We have not yet characterized situational irony, for by no means all failures to meet a standard or expectation count as ironic. Lysander’s failure will count as ironic for most of us, partly because he fails to recognize his own failure to meet the standard he sets himself. Failing by self-imposed standards is easily seen as ironic: the police officer who lost control of her vehicle and crashed through the police barrier instantiates a failure we quickly designate as ironic. Was she driving less carefully than the owner of a car who simply crashed into a roadside barrier, provoking little in the way of an ironic response? Probably not, but when the police crash through their own barrier, the standard and its violation become salient in ways that do provoke an ironic response. But this is not the only pathway to ironic failure. Failure is sometimes ironic because of the thin margin – it would not be irony if the explorers had expired a hundred miles from the food store. Sometimes it’s because the one who fails is strongly identified with success in the relevant area, as when the god Apollo fails to arrange human affairs to get the result he wants.Footnote ⁵

Saying exactly what counts as situational irony may turn out to be as hard as theorizing humor, and I won’t try to close in further.Footnote ⁶ We can often say something useful about the relations between things without having a full (or indeed any) account of what those things are. I don’t know what numbers are but I know there are more real numbers than integers – in fact I know precisely how many more. With a rough grasp of situational irony, we can comprehend the thought that dramatic irony is the dramatic (and more generally narrative) representation of ironies of situation. I’ll expand on that a bit by offering a clarification, and a response to an objection.

Representing and Creating Irony

I said we start with ironic situations, which then get represented in a drama (or more generally, some narrative form), and that gives us dramatic irony. But in the case of Lysander’s protestations, there is no actual event that the play is representing to us; the only thing really there is the representation. How can that be?Footnote ⁷ As Georges Rey puts it, “You cannot kiss/touch/kick/amuse something that does not exist; how can you represent ‘it’?” (Rey, Reference Rey2022: 279). This certainly is a problem, but a rather too general one for us to tackle here. While the irony of Lysander’s error is a good candidate for being a nonexistent representable, so are Sherlock Holmes, perfect competition, phlogiston, and heaps of other stuff we seem able to represent; I simply undertake to work the best solution you come up with into the present proposal.

Perhaps we can focus on something more specific to the sorts of situations that come to mind when we discuss the representation of irony – fictions in literature, drama, and film. In such cases, the failure to represent what actually happens is no drawback; we expect that of fiction. What we want fictions to do is to represent what is true-according-to-the-story. And this, it seems, fictions can do automatically. It is sometimes hard to know what is represented as true-in-the-story; garden path plots and unreliable narrators see to that. But if a fiction does represent something as true-in-its-story, it surely cannot do so incorrectly. If I decide that the story does represent Homes as living at 221b Baker Street, it is pointless then to wonder whether it might actually be part of the story that he lives further down the street.Footnote ⁸ While fictions can fail to be true they cannot fail to represent correctly, when we understand what correctness is for the fictional case. The job of the dramatically ironic fiction is to make it true according to the story that the relevant situation is ironic, and it does so simply by representing the situation as ironic.

All very true, up to that last statement. There is a difference between representing Holmes’ address and representing the events of the story as ironic, or tragic, or in other ways that call on some affective or evaluative response from an audience. The story that says killing female infants is good in itself (and not for special reasons, such as, “in the story, females grow up to be mass murderers”) does not thereby make it part of the story that killing female infants is good in itself (see Walton, Reference Walton1994). Nor does a story succeed when it tells us that a featureless plain strewn with rubbish is sublime, or that an uneventful walk round the park is tragic. True, when Doyle tells us that Holmes lives at 221b Baker Street, we would be wasting our time wondering whether he really lives next door. But when a story represents its events as ironic, we may fail to find them so, and consequently wonder whether those events merit being recognized as ironic, concluding that the answer is no. There is reason to think of the sublime, the tragic, the ironic as categories dependent for their applicability within the fiction on what would be a merited response by the reader to the events so represented. If that is right, irony escapes the automatic determination-by-representation that things like Holmes’ address is subject to.

Let’s stay for a moment with the idea that irony depends on its representation. To simplify, consider cases where there really is some event in question, an event being represented as ironic. Where you live does not depend on how your living arrangements are represented. But its being ironic that you live there (next to the enemy you spent your life escaping from) might do. Isn’t there merit in the idea that irony is never “intrinsic to the situation”? Yes there is, but that does not show that it is representation-dependent. Perhaps the irony of a situation often (always?) depends on the wider context in which it occurs, and we may then need a representation (typically a narrative) to pick out for us the subtly relevant aspects of its context to see it as ironic. The conclusion we should draw is that irony is a relational feature of a situation, and relational features of a situation are features of it; the role of representation is to pick them out for us, not to create them.

Another kind of case may also look like representation-dependence. These are cases where the representation does not make the situation ironic, but makes it something I will call ironic-as-represented. Satire often trades on this, taking images out of context in obvious but telling ways, ridiculing the target without expecting us to believe that events went as the satire presents them. A politician with a relaxed attitude to pandemic restrictions might be presented as “in isolation,” followed by a photo of them in a crowd of partying people. To get the point, one does not need to believe that this is a record of some actual event where the politician was supposed to be isolated. In such a case an event is represented as ironic without it being suggested that the event really was ironic, or at least that it was ironic in quite the way represented.

Predictably, this technique can slide toward outright misrepresentation; at least, there are cases where it is hard to know whether we have irony-as-represented or the (mis)representation of a situation as genuinely ironic. Suppose that scenes in a straight-faced documentary are ordered to seriously imply that a politician arrived to give a moral boosting speech at a recession-hit supermarket, during which the cash register was stolen – when in fact the theft actually took place days earlier. It is not unreasonable to think of that as misrepresentation of irony. By contrast, in a film where there is obviously and openly a good deal of humorous license taken in editing, the judgment might be that this is merely the creation of irony-as-represented; seen that way, the work evades the charge of misrepresentation.Footnote ⁹

Rethinking Dramatic Irony

An objection to my proposal is that it so widens the scope of dramatic irony that it changes the subject. I said that there is dramatic irony where there is a narrative representation that makes manifest to an audience the irony of a situation. But the descriptions of irony one finds in dictionaries of literary terms, introductory texts, and sometimes in scholarly works speak of it as a situation where the audience knows more than some character.Footnote ¹⁰ As stated, this wildly over-generalizes; in almost any narrative with multiple characters and scenes, at least one character will not know something that is available to the audience. Little of this would strike us as dramatic irony. Traditionalists may amend their account in a way that reduces the distance between their theory and mine. They may say that the representation of a disparity in knowledge will count as dramatic irony only where a character’s false belief threatens their prospects with failure. This is how dramatic irony is often exemplified. In Kyd’s The Spanish Tragedy there is dramatic irony in the Viceroy’s mistaken belief that his son is dead (he is a prisoner), not merely because we know something the Viceroy does not know, but because the treacherous Villuppo has, “with an envious forged tale,” used this false belief to turn the Viceroy mistakenly against a rival.

That is a better theory, but it still limits dramatic irony to failures of knowledge, when other kinds of failures are as capable of being responded to by audiences as ironic, and as deserving of that response. Nor is this a bare possibility; scholarly studies of the dramatic irony in narrative often ignore the official restriction to epistemic failure and emphasize other things that compromise effective agency, good sense, and good morals. Here are some recent examples from the literature, briefly described.

A study of the New Testament finds dramatic irony in the gospels at points where the disciples’ many failures are represented. While some of this is failure to understand Jesus’ message and is covered by standard, epistemically focused accounts of irony, it also includes not observing food prohibitions, neglecting to “assign credit where credit is due,” and exhibiting “consistently inept behaviour” – failures certainly, but not failures to know things (Miller, Reference Miller2012). An essay on Conrad finds a consistent thread of dramatic irony in his work that highlights the failures of well-intentioned interventions in the lives of others (Paltin, Reference Paltin2014: 99). A study of Washington Square similarly identifies Jamesian ironies as occasions when a character’s “stratagems” lead unintentionally to failed relationships or loss of fortune (Maine, Reference Maine2016). Looking back a bit further, in a chapter on dramatic irony in medieval romance, Dennis Green notes the likely awareness in an audience of the time that the boasting of Keu (a character in the work of Chrétien de Troyes) will be followed by ignominious defeat (Green, Reference Green1979: 256).Footnote ¹¹

I cite these as cases of dramatic irony that are not accounted for by saying that we know something the character does not know. I don’t deny, however, that the idea of knowledge is relevant in various ways to their workings. In some cases dramatic irony does highlight ignorance, without that being its primary purpose. Taunting Tiresias with his blindness highlights what Oedipus does not know, but is significant primarily for its undercutting of Oedipus’ presumptions of superiority in judgment. In Heracles, Lykos, soon to be killed by Heracles, insists that “He is not here and will never return,” thus displaying his ignorance, but more importantly his illusion of mastery.Footnote ¹² And boasting followed by defeat is dramatically salient in Green’s example because it is indicative of deficiency of character, not of knowledge. It is also true in some of these cases that if the agents involved had known how things would turn out, they would have acted differently and thus avoided failure. But to locate the irony in this gets it round the wrong way. Had I known that my car would crash, I would not have made the journey; that does not mean the crash represents a failure of knowledge rather than of driving skill. As a last ditch defense one might insist that failures of character are, at bottom, failures of knowledge: being dangerously over-confident is a matter of not knowing the limits of your powers. This confuses a trait with its effects. Lacking diligence, I fail to know various important things. But that defect of character cannot consist in what I fail to know, for failures to know can be caused in many ways, including bad luck.Footnote ¹³

Perhaps the most telling evidence for the recognition of dramatic irony falling outside the bounds of the received epistemic account comes from what is said to be the source of that account, Bishop Thirlwall’s essay on Sophoclean drama of 1833 (Thirlwall, Reference Thirlwall1833).Footnote ¹⁴ The essay is entirely consistent with the view I am proposing according to which epistemic failure is simply one aspect of, and sometimes merely a pointer to, ironies located across a broader spectrum of failures. Thirlwall emphasizes the contrast between Oedipus, confident in his powers when pleaded with by the citizens for relief from their suffering, and Oedipus, ruined by his exercise of those powers against the background of “his headstrong, impetuous character” (500). Here we have, he says, an instance of the “the fragile and transitory nature of all mortal greatness” not unlike the collapse of states when apparently at their most powerful (489). While Thirlwall notes the play’s emphasis on “how different all is from what it seems” (496), the irony he points to is mostly evident in the play’s representation of failure of agency and ambition – failure certainly made vivid to us by our superior knowledge.Footnote ¹⁵ In such cases lack of knowledge is a pointer to the irony, it is not what constitutes the irony. Occasionally, later writers have described dramatic irony in just this way. J. A. K. Thompson said that the method of Aeschylus is “by playing … upon the contrast between the knowledge of the spectator and the ignorance of the agonists, to drive home the irony of the situation,” and Dennis Green notes that dramatic irony “depends for its effects on the superior knowledge of the listeners” (Thomson, Reference Thomson1927: 53; Green, Reference Green1979: 250 (my emphases)).Footnote ¹⁶ Both formulations make the disparity in knowledge the means by which the irony is made vivid; they do not locate the irony in the disparity.

On my account dramatic irony is sometimes constituted by a failure of knowledge but sometimes by a failure of will, or integrity, or luck.Footnote ¹⁷ A good model for this is functionalism about the mind, though the two views don’t stand or fall together.Footnote ¹⁸ There may be no one physical state that is pain, but there may be a physical state that is pain in humans, or pain in this human, or pain in this human now. Pain on this occasion is whatever on this occasion occupies the functional role definitive of pain. Rather than specify a functional role for irony we may take a shortcut suggested by some earlier remarks, and say that irony is what merits being recognized as ironic. This is of course no definition of irony and is not meant as one, but it helps us in the following way. Earlier I gave the example of Lysander’s protestations to Hermione. What exactly is the irony in Lysander’s situation at this point in the play? It is whatever it is about that situation that is correctly recognized as ironic. The thing that, on this occasion, merits that recognition is Lysander’s lack of knowledge, and that is the irony on this occasion. On some other occasion, in some other drama or novel, what merits recognition as irony will be some other failure: of will or duty or simply of fortune.Footnote ¹⁹

It needs to be acknowledged that epistemic cases of dramatic irony are not on all fours with these other kinds; they are, for a start, disproportionately represented in drama. There are many such cases in the first Oedipus, and its status as a foundational drama may help explain why epistemic cases are abundant in the subsequent practice of ironic representation and so often the focus of theoretical attention. There are more general reasons why epistemic cases absorb so much of our concern.Footnote ²⁰ First, situations where we know something that another does not know are among those we find most arresting and important, in part because of the possibility that this opens up for deception. Social and often cooperative beings though we are, deception is still a tempting way out of many difficulties and being able to detect deception protects us from many disasters. Disparity in knowledge, which is a requirement for deception, is unsurprisingly something that attracts our attention and our emotional response, both when encountered in reality and when represented in fiction.Footnote ²¹ Second, epistemic failure on the part of a character is essential for the working out of that popular trope of narration, recognition, as when Electra spends many lines grieving over what she takes to be the ashes of Orestes, before realizing that he stands living before her.Footnote ²² Third, as indicated earlier, failures that are not primarily epistemic do often involve failures of knowledge at some level. If we fail to be vigilant in acquiring and maintaining knowledge, we are likely to fail in our quest, and the immediate cause of failure may be some piece of information we missed, though the underlying failure was one of attention. So an approach to dramatic irony that focuses exclusively on knowledge will often have a superficial purchase on cases that really ought to be treated in another way.

Engineering, Clarifying, Simplifying

I’ve argued that the literature of dramatic irony doesn’t limit itself to epistemic cases, and often ignores a definition that I’ve said is unserviceably narrow. Perhaps that definition is just ideology, floating free of the facts about what concept “dramatic irony” locks onto in practice, and those facts are all in accord with my proposal, which has nothing revisionary about it. However, it would be unrealistic to suppose that a definition so widely affirmed as the standard epistemic one has no impact on practice, and that if we all recognized its inadequacy nothing would change. I have given reasons why many, perhaps most, cases of dramatic irony actually dealt with in the literature are epistemic. But one additional reason is surely that the standard definition orients us toward such cases and makes nonepistemic cases less visible – though not entirely invisible, as we have seen. So my proposal to dump the definition would, if accepted, probably change something in the way we identify and analyze cases of dramatic irony. Does the proposal in that case count as conceptual engineering, an activity motivated by the idea that the concepts we use are not always those best able to do the jobs we want of them, and should sometimes be replaced by better ones?Footnote ²³ Sometimes it is hard to tell engineering from clarification, which tells us that the concept in question is in good order but needs to be better understood and applied. David Chalmers and Andy Clark (1998) proposed that memory-impaired Otto counts as believing P as long as P is written in his notebook. Were they re-engineering the concept of belief? Chalmers says no, it was an attempt to show that the un-engineered concept of belief does, surprisingly, count Otto as believing P; we were confused about what belief covers, and they were helping us to a clearer understanding.Footnote ²⁴ I am unsure whether my proposal counts as engineering, clarifying, or something else. Note that dramatic irony is unlike Otto’s beliefs. People who deny that Otto believes what is in his notebook are not generally influenced by a definition, defective or otherwise, of belief; most of us don’t have a definition of belief to appeal to. What seems to be happening with dramatic irony is that there is a default tendency to lock to a concept that accords with the standard definition, but to shift to something more like the one I propose when one finds interesting cases that don’t fit the epistemic model. If that’s right, what I am proposing is that we dispense with the current default concept and begin our investigations with my more liberal one, which will then no longer be merely an optional refuge for those who find the epistemic account too constraining. We might describe this as conceptual simplification.

Communicative (Verbal) Irony

So far I have accounted, in outline at least, for the relation between situational and dramatic irony. How do we bring verbal irony into the picture? First, by changing the name. “Verbal” limits us to communication in language, and ironic communication is available in linguistic and nonlinguistic forms. I turn to you with a look of delight as an extra, day-long departmental meeting is announced. I pretend to be delighted by the news, thereby vividly illustrating what a failure of judgment it would be really to be delighted by it.Footnote ²⁵ I could have said “How wonderful,” another way of pretending delight. The two cases, one involving bodily expression and the other language, are both cases of communicative irony. You are being communicatively ironic when you pretend to assert something, to ask a question, to make a request, to be delighted or affronted or amused, while evidently not really doing or being those things, and with the aim of making evident the shortcomings of really doing them or being them. When the banished Hippolytus asks his father whose house he can go to, Theseus replies “Someone’s who likes to entertain seducers of their wives,” making it clear that the only positive answer available is one no one could seriously assert.Footnote ²⁶ The communicative ironist is like the actor who pretends to insult or harangue or fight by doing something which is rather like insulting, haranguing, or fighting – enough like these things that the audience can be counted on to know what action is being pretended.

Thinking of irony as a mode of pretense is not the only option. One ancient view treats ironic utterance as “meaning the opposite of what one says.” The reasons for rejecting this view are widely known and strike me as conclusive.Footnote ²⁷ A much better option is the echoic theory, due to Dan Sperber and Deirdre Wilson.Footnote ²⁸ There is no space here to review the debate between the pretense and the echoic theories. The present proposal for a general theory is in a sense a challenge to the echoic theorists – can they develop a theory comparable in generality and explanatory power to this one? I leave it to them to answer.

Earlier I joined others in noting that irony is a very stretchable concept, rejecting the ambition to find a theory that covers all or even most of what goes by that name. But we need to recognize “near miss” cases, those we may think of as “having something of irony about them.” I’ll describe a near miss as regards communicative irony, and say why it falls just outside the scope of my theory.

Sometimes narrators tell us that they don’t know something which we readers would like to know. Not all such protestations are even close to irony, for a story may be told by a narrator whose assertions of ignorance are seriously meant and easily accepted. “I wanted to know – and to this day I don’t know, I can only guess,” says Conrad’s Marlow, acknowledging the entirely plausible limits of his understanding (Lord Jim, chapter 7). On other occasions, however, an apparently omniscient narrator (at least, a narrator who clearly knows many things) declares their ignorance in a way we can hardly take seriously. Trollop does this sometimes, and Dennis Green finds a supply of cases in medieval romance: a narrator who cannot describe a combat because no third party was present, though this has not stopped the narrator from reporting other things that were similarly unobserved; a narrator who disclaims knowledge to “avoid the charge of obscenity, knowing full well that the listeners’ imagination will more than do his work for him” (Green, Reference Green1979: 234–235). In such cases it is natural for the reader to think of the narrator as pretending to have a limited perspective on the action, when in fact they could tell us more if they chose. To count strictly as irony for me, this would have to be done in the service of making it vivid how absurd it would be really to claim ignorance in this way. I don’t think that is quite what is going on. True, there is something odd about a narrator who otherwise shows all the conventional signs of omniscience, suddenly declaring themselves ignorant, and readers will often have some awareness of this, though few would articulate the point while reading. But the purpose of these ironic disclaimers of knowledge, as Green’s careful analysis of cases illustrates, is not usually that. It is rather to make some rhetorical point about the action or the characters in it, as when an affirmation “is weakened by being turned into mere surmise” (Green, Reference Green1979: 234). Still, I think it is fair to count such cases as near misses from a pretense-theoretic perspective; they are cases where the narrator engages in a nondeceptive pretense of failing to know something, a pretense of having a perspective more limited than the one they actually have. Such a display might well provoke a response in the audience not unlike that merited by cases that more clearly meet the conditions for irony.

Solipsistic Irony, Expressive Irony

I said we should concentrate on communicative irony, a category broader than verbal irony. But there are cases of verbal irony that are not communicative, and they are immensely important in drama and other places where irony is represented. Socrates’ irony nicely illustrates the pretense account; he pretends to admire the intellectual powers of his conversational partners, to plead with them to treat his own very limited intellect with more gentleness, to look for enlightenment from them. John Ferrari describes this irony as “solipsistic”: exactly not intended to be understood by those he is speaking to, or by anyone else in his audience. Some, like Laches, never do get the irony, and those who do get it take themselves to be uncovering a slight of hand rather than grasping something they were honestly if somewhat ambiguously presented with. As Ferrari notes, this is not uncommon; concealed irony is the norm where retribution is likely (Ferrari, Reference Ferrari2008).Footnote ²⁹

It may please Socrates to speak over the heads of his fellow dialogists, but for some agents in drama, not being understood is an absolute necessity. Inviting Agamemnon into the house where, as we know or suspect, he will be murdered, Clytemnestra orders richly dyed cloth to be laid in his path:

Clytemnestra is merely pretending to an attitude of admiring welcome. From her point of view to see things that way would be absurd; he has, after all, killed their daughter. The justice that ushers him in will in reality be of an entirely different kind. Her utterance is richly ironic, though intended not to be so understood by Agamemnon.

Ironic deception in this mode is a high wire act, largely confined to drama. To be taken in by a speaker we have to think them sincere; revealing the irony of your utterance makes that impossible. You have to be confident that your addressee is sufficiently disadvantaged by obtuseness or ignorance that they have no chance of getting the irony. If anything much hangs on the success of the lie, better not take the risk. In drama it works, or can do, because the author can ensure that the audience is primed by knowing something crucial that the person spoken to does not know, as we, unlike Agamemnon, know the intentions of Clytemnestra. And this gives us one more reason for thinking that dramatic irony that consists in disparity of knowledge, while not the only form that dramatic irony can take, is so common. It is the kind of dramatic irony generated by hidden acts of verbal irony.

Clytemnestra’s irony, like that of Socrates, is not communicative. What sort of irony is it? It is expressive irony, irony that is expressive of an ironic state of mind in the way in which a drooping posture is expressive of a sad state of mind. Just as your posture expresses sadness without you using it to communicate anything, Clytemnestra’s words express, or are symptomatic of, her ironic state of mind – her vivid awareness of the absurdity of really meaning what she is pretending to mean. What she must count on is Agamemnon’s insensitivity to this information. When it comes to a sad face or posture one would have to be very insensitive or distracted not to recognize it as a sign of inner sadness. But in other cases recognition can depend on possession of contextual information not widely shared. Knowing the couple well, I see a glance between them as expressive of hostility, something a stranger might easily miss. Agamemnon similarly lacks vital bits of contextual information possessed by the audience.

We might have offered a different account of the case, saying that Clytemnestra performs a dual speech act: a lying assertion for her husband and an ironic performance for the audience of the play. But this is not pantomime, and she is not addressing the audience. Still, the irony of her speech may affect an audience in dramatically crucial ways. The careful phrasing given to her words does, I think, serve to connect the audience more closely, and perhaps uncomfortably, with her murderous project. For they understand what her words express, and understand that the soon-to-be-victim Agamemnon does not understand this.Footnote ³¹

Are the cases of Socrates and Clytemnestra different? Not obviously; both have audiences – Clytemnestra a theatrical one and Socrates a readership. But I agree with Ferrari that there is an important difference between the dramatic and the literary cases:

Quite what makes for the difference is a bit unclear. As already indicated, the issue is not whether a character may address the audience directly; Clytemnestra is not doing that any more than Socrates is. But it is surely right to emphasize the difference between textual representation on the one hand and dramatic presentation on the other, where spatial and temporal proximity to the staging of the action encourages the sense that one is a side-participant. Here I will say only that the remarks of both Socrates and Clytemnestra are expressive of irony without being communicatively ironic, that the remarks of both are understood by their (extra-fictional) readers/viewers as ironic, but that Clytemnestra’s words have the power to affect the target audience in a way that Socrates’ do not.Footnote ³²

Communicative and Dramatic Irony Compared

Abandoning subtle distinctions and returning to the big picture, let us say that communicative irony provides another way for us to represent the irony of a situation. Feigning delight at the announcement of the meeting, I represent myself as failing to meet a goal or expectation; failing, in this case, to see how boring the meeting will be. If I say, “You sure know a lot” to the bore who regales us with his knowledge, I represent myself as fascinated by trivia. If I say “That’s great driving for you” as we watch the police car crash through the police barrier, I play the part of someone failing to recognize failure when I see it. This last case layers the ironic effect, being ironic commentary on an ironic situation. In these instances I represent myself as doing something I am not actually doing, in the cause of expressing my sense that really doing it would be an egregious failure of judgement.Footnote ³³

If communicative and dramatic irony are both representations of situational irony, how do they differ? Communicative irony is self-reflexive, dramatic irony isn’t. To say or do something communicatively ironic is to perform an action that is a pretense of exemplifying some (appropriate) sort of failure or limitation: failing to recognize bad driving when you see it, failing to recognize how boring the speaker is. By contrast, the writer of a play that represents the police crashing through their own barrier represents the police failing, not herself. What if I write a play in which I am the character who crashes the barrier? Yes, in that case I ironize myself, but not in the act of composing the play. That act of presentation is not a pretense of exemplifying any kind of failure on my part, though the result of the act is a recipe for providing someone (an actor) with a way to pretend to exemplify such a failure.

While communicative irony involves representing yourself as failing, its ultimate purpose may be different. I may be hoping to bring to mind some other real person who does or might respond in the way I am pretending to respond. Knowing your musical tastes, I say “I’m looking forward to the Abba tribute concert this week.” You are the target of my irony, and I make you the target by presenting myself as someone with your dubious taste in music. The same can happen with dramatic irony; the failure of a character may be intended to bring to mind the failure of some real, extra-textual agent.

We must distinguish an act of communicative irony from two other things. The first is the act of representing communicative irony. We have the latter when the maker of narrative has a character speak ironically. To do that is not automatically a way for the author to speak ironically, for the author’s perspective may be clearly dissociated from that of the character, or at least not obviously aligned with it. There is more on the irony of characters and the perspectives of authors in the next section, but for now I ask: how can a maker of narrative be communicatively ironic? The prospects may seem poor, for the maker would have to be appearing to engage in some apparently serious act, but in fact be pretending to engage in it, with the intention of pointing to the ridiculousness of really doing that thing. And ironic filmmakers do not pretend to make films; they really make them, as ironic novelists really write novels. To answer this we need to recognize the variety of forms that pretense can take. Someone who speaks ironically, and hence in pretend mode, may really be asserting what they say, as with an example I gave just now: “You sure know a lot,” said to someone too keen to share their knowledge. While the ironists must be engaging in a pretense, they need not be pretending to assert something. They may instead be engaging in a pretense of manner, pretending to assert enthusiastically, admiringly, or with the hope of eliciting further information.Footnote ³⁴ The ironic playwright is writing a play and not pretending to write one; what they may additionally be doing is pretending to write a play that has deep philosophical significance. In the cause of that pretense, they invest the play with lots of superficially profound but in reality – as the intended audience will recognize – shallow and pretentious dialog. It is the same with Clytemnestra’s address to Agamemnon. She really does welcome Agamemnon, for welcoming is like promising: saying the right words in the right context constitutes welcoming, whatever the beliefs and desires of the speaker.Footnote ³⁵ What she pretends to is the possession of the beliefs and desires this act of welcoming suggests.

We must also distinguish between an utterance which is communicatively ironic and one which is not meant ironically but which constitutes a situational irony. Good examples of this occur in an experimental study by Gibbs and colleagues (Reference Gibbs, O’Brien and Doolittle1995): participants judged an utterance of “I would never be involved in any cheating” to be ironic if the speaker unknowingly contributed to someone’s cheating in a class exam. The utterance is ironic, but not meant ironically. It is ironic because a situation where someone’s claim never to have cheated is based on a failure to recognize their own unwitting participation in cheating is an ironic one.Footnote ³⁶

Communicative Irony and Romantic Irony

The characters in a drama may act out situations which exhibit dramatic irony; they may also, through their words or behavior, give us instances of communicative irony. The examples from narrative art I have given – Socrates, Clytemnestra – have been problematic because of their expressive but noncommunicative status. But there are many straightforward examples where characters speak ironically in ways that are as obvious to their addressees as they are to us. “To be sure – our discordancies must always arise from my being in the wrong” says Emma to Mr. Knightly, meaning to make him aware of errors in his own thinking (Emma, vol. 1, chapter 12). This is, of course, not a real act of ironic communication, since speaker and hearer are mere fictions. It’s no different on stage, as an actor who speaks the ironic line is not speaking ironically any more than the actor playing Hamlet kills Polonius. Actors do occasionally speak ironically on stage; “Wann geht der nächste Schwan?” says the Wagnerian tenor to the audience as the swan is hauled off stage too early. But the actor is simply speaking on his own account and outside the confines of the drama. Sometimes dialog in pantomime presents itself as the utterance of the actor, commenting on how poor the jokes are. But this is just more fiction; it is fictional that the actor’s speech expresses contempt for the jokes (though it may also be true that she does think the jokes are terrible). On the other hand, we are often able to identify places in fictions where what the characters fictionally assert is indicative of the maker’s own attitude, one they would like us to share, or where an ironic remark of a character seems to express an ironic thought of the maker. I will consider a case that is more complicated: it is one where the maker’s ironic posture is revealed, and is intended to be revealed, by a straightforward, nonironic remark of the character. In such cases the maker is communicatively ironic, but in a very indirect way.

I’ll present three examples of communicative irony in narrative art, one from literature and two from film. Given our emphasis on the ironic potential of the narrative arts, that hardly needs a justification. But there is something else in favor of this approach. Evidence-based inquiry into irony comprehension has so far mostly concerned on-the-fly processing of conversational remarks, where responses are measured in hundreds of milliseconds.Footnote ³⁷ Artistic contexts do make comprehension harder; they also offer more opportunities for lengthy reflection and finely graded judgment based on repeated exposure to the work, the sorts of processes not easily captured in experiments. In these contexts ambiguity as to an ironic intent can take its place among the difficult and intriguing problems posed by a work of art, and here irony sits squarely in the domain of the aesthetic. Irony in serious narrative is irony seen in slow motion.

The cases of communicative irony that I will discuss also fall into a subcategory we have not yet considered: romantic irony. On his topic I will be more than usually prescriptive, dialing down the idea of romantic irony associated with Friedrich Schlegel’s valorization of works that are conscious of their own “inability to represent the absolute” (Rush, Reference Rush2016: 66). Untroubled by my inability to represent the absolute, I focus on the modest suggestion that “[a] specifically Romantic irony can therefore be said to be present when texts become self-reflective about their construction as texts and authors show genuine scepticism about their own aesthetic control of their products” (Handwerk, Reference Handwerk and Brown2000: 206, my emphasis). Taking up this idea, we can isolate a form of romantic irony that turns out to be a special case of communicative irony. It is the ironic communication of one’s (partial) loss of authorial control, a loss typically due to pressures from outside the work, notably in externally generated norms or expectations. In extreme cases these pressures provide strict limits to what the artist can do, in others they may have no more force than that of convention, though conventions can be difficult to ignore. In one way or another, there are times when local and contingent constraints on narrative making will be adhered to but not be welcomed. If we think of romantic irony as the practice of conforming to expectation while indicating in subtle or not so subtle ways that conformity represents to some degree a failure of your artistic ambitions, the communicative irony claimed for my examples will count as romantic. As already indicated, I don’t claim that this covers anything more than a small corner of all that has been called romantic irony. But instances of it are important for scholars and ordinary readers when they occur, and they fit well into my overall theoretical structure.

This abstract discussion badly needs the support of concrete cases. My first example is taken from a scene in Douglas Sirk’s film There’s Always Tomorrow (1956); I’m indebted here to a fine study of cinematic irony by James MacDowell (Reference MacDowell2016). Lying between two of Sirk’s better-known movies, All that Heaven Allows and Written on the Wind, There’s Always Tomorrow is more restrained; a small-scale film, notably for Russell Metty’s cinematography. In brief outline, West Coast businessman Cliff Groves (Fred MacMurray), somewhat taken for granted by his family, is tempted to leave for elegant executive and old flame Norma Vale (Barbara Stanwyck). On the pleading of the Groves’ older children, Norma, who has long been in love with Cliff, refuses him and leaves for New York. The final scene shows a subdued Cliff return home to a more than usually attentive family. MacDowell finds irony in the final moment when the youngest child, Frankie, regarding her parents, remarks “They make a handsome couple don’t they,” to which the older children register a smiling assent. On MacDowell’s view, the film’s appearance of providing this statement as emblematic of a reunion to be celebrated is misleading, and what we have is “a pretence of trying to salvage a conventional ‘happy’ concluding image from the meagre materials available” (MacDowell, Reference MacDowell2016: 80). Frankie’s observation is perfectly serious, but its discordance with Cliff’s too-evident discomfort suggests that the words have been put into her mouth as ironic acknowledgment that, while this may be what is expected, it will hardly do.

My second example comes from a better-known and much theorized source: Hitchcock’s The Birds (1963). At various points in the film Hitchcock shows signs of ironizing the project on which he appears to be embarked, that of creating a realistic and frightening story of nature turning again humankind, most notably in the gas station scene, where a bird attack results in a series of explosions, death, and injury. As the scene develops, we get a series of reactions shots from Melanie (Tippi Hedren) watching the spread of a petrol fire; in each her face appears in a static and stylized pose representing an exaggerated expression of horror. Here and in a few other, less obvious places Hitchcock seems to be standing back from his project, suggesting to his audience that he does not take this very seriously, and neither should they.Footnote ³⁸

The (putative) irony of these two scenes is similar in an important way: in both, the nonironic reaction of a character is to be treated as a surrogate for the ironic communication of the maker. Frankie is serious in what she says, but (on the ironic reading) we are invited to imagine the film’s maker saying the same thing ironically, merely pretending to celebrate the ending he has crafted. Melanie, we may assume, is (fictionally) seriously horrified by the mayhem, but her odd expression invites us to imagine the maker expressing an ironic mock-horror at the scene. This pathway to irony is important in film, and there is a contrast here with literature, where direct authorial intervention that communicates ironic distance is easier. At the end of Northanger Abbey, we are told that “Henry and Catherine were married, the bells rang, and everybody smiled”; we are, says our narrator, “hastening together to perfect felicity” (chapter 31). In good romantically ironic style, this cheerful ending is provided out of conformity to expectations Austen does not endorse; she pretends to be someone silly enough to suppose that endings of perfect felicity are what is needed in literature, even if done in a rush. Because the narrator speaks directly to us, there is no need to look for a character whose words are to be taken as the clue to what is communicated.

Summary

My title may be thought misleading; the general theory I have offered does not cover everything that we have some right to place under the heading of irony. But general theories are not always universal; the general theory of relativity does not, so I’m told, apply at very small scales. Starting from the idea that ironic situations are among those that exhibit failure to meet some expectation, goal, or standard, I proposed that dramatic and communicative irony are distinctive ways of representing situational irony. I argued that while awareness of irony often depends on a rich understanding of the context in which it occurs, it is not the product of how we represent that situation. I defended a very inclusive account of dramatic irony and argued that, despite its deviation from the way dramatic irony is usually defined, it is consistent with our practices of irony identification and with historically influential sources. I introduced the idea of expressive irony – irony expressed in words but not meant to be understood – and emphasized its significance in live drama. I showed how romantic irony, in one of its many guises, constitutes an indirect kind of communicative irony, conveyed through the nonironic speech or behavior of a character. I concluded with a brief exposition of three cases of this romantic/communicative irony in narrative fiction. I hope this provides a structure for thinking about core areas within irony’s domain; extending the theory will need careful judgments about what should count as irony, what merely as irony-related, and what should be abandoned as the product of an inflated vocabulary.