What are we doing when we engage in aesthetic discourse—when we discuss something’s beauty, sleekness, elegance, or intensity; when we talk about something’s aesthetic character, discuss how to interpret it, how it works, fits, or fails; and so on?

Aesthetic talk is evaluative, and evaluative talk is commonly distinguished from ordinary descriptive talk in terms of how each affects conversational dynamics. Unlike ordinary descriptive talk, evaluative talk is thought not merely, or at all, to add ordinary descriptive information to the conversational record. In addition, or instead, it expresses feelings, plans, or norms and aims at convergence in such among the interlocutors.

But that alone does not distinguish aesthetic discourse from normative but non-aesthetic discourse. Work on metanormativity typically takes moral thought and talk as paradigmatic of the normative. Pérez Carballo and Santorio’s thesis concerning “conversation about normative matters” is that “In any conversation where certain minimal assumptions are satisfied, it is presupposed that there is a unique normative standard on which the participants’ attitudes ought to converge” (2016: 608). They do not discuss the normative character of aesthetic discourse in particular, presumably because it seems plausible to them that any model of moral normativity will apply straightforwardly to the aesthetic, or if not straightforwardly then with some easy tweaks.¹ Such an approach is encouraged, perhaps, by the prima facie plausibility of expressivism in aesthetics.²

The thought that aesthetic discourse aims at convergence—in judgment, attitude, affective dispositions, etc.—is prominent in philosophy. Some philosophers build convergence into the aims of aesthetic life, arguing that (a) aesthetic experts will agree, will largely agree, or will never disagree. And some argue that (b) in our own aesthetic lives we should aspire to be broad aesthetic experts, seeking out and learning to appreciate a wide variety of items of aesthetic value (Levinson 2002; 2010). It follows fairly quickly from (a) and (b) that aesthetic discourse aims at convergence.

This, of course, is not the only way of arriving at that conclusion. Some follow Kant and say that in making aesthetic claims we demand, expect, or confidently call for agreement from everyone. Andy Egan argues that convergence is the ‘central business’ of aesthetic assertion:

Others see in aesthetic conversation the aim of establishing a ‘community of feeling which expresses itself in identical value judgments’ (Isenberg 1949).

Some philosophers qualify or question convergence but embrace it nonetheless. Lopes (2018) argues that the meaning of aesthetic predicates is fixed by specialized aesthetic practices and argues that there are no general convergence norms across such practices. I may be interested in graceful dance, but unless I am also interested in beautiful tennis, there is no pressure for me to agree with you that Federer’s swing is graceful. I might just be indifferent. However, consensus is required within aesthetic practices (Lopes 2018: 164–80). For Lopes, convergence operates wherever it can, within practices: “… aesthetic disputes are to be overcome, not embraced” (2018: 166).

Nguyen (2019) argues that aesthetic conversations are open-ended and aimed more at the goods of engaging in them than at any desired outcome, for example, a shared or correct aesthetic judgment. This opens the door to a rejection of convergence, but it could just as well leave that door shut, if a hope for convergence is what keeps good aesthetic conversations going. In other work, Nguyen seems to express that hope:

Others take convergence beyond normative and aesthetic discourse and see it as the basic point of conversation. Perez Carballo and Santorio claim that for all conversations “… as long as [the interlocutors] think that there is a point to engaging in conversation, they must think that they ought to converge on some live possibility. Converging on some live possibility is just what the point of conversation is” (2016: 631; see also 634). Others see conversation as fundamentally assertive and define assertion “as committing the speaker to the belief in its propositional content, while simultaneously attempting to get the addressee to believe that content as well” (Farkas & Bruce 2010). The idea goes back at least to Stalnaker, who models conversation on the exchange of assertions, which “reduce the context set …, provided that there are no objections from the other participants in the conversation” (1978: 86). To ‘reduce the context set’ is to converge on information.

Convergence thus features in theories of aesthetic discourse, metanormative theories of normative discourse, and theories of assertion and conversation in general. Returning to aesthetic discourse, the idea we can extract from these views is:

Convergence might follow from general features of conversation, features of normative discourse in general, or features of aesthetic discourse in particular. The claim is not that there is a single unique standard that we must agree upon in every aesthetic conversation. It is that there is a single standard per conversation.

Here I argue that Convergence is wrong, and, therefore, so are the aesthetic, metanormative, and linguistic theories that emphasize it. It misconstrues the fundamental aims of aesthetic discourse and, thereby, overlooks the distinctive character of aesthetic community and power of aesthetic discourse to create it. To establish this, I develop a novel account of those fundamental aims. Although the end of aesthetic discourse is indeed a certain kind of community, it is not an end that requires convergence in aesthetic attitudes. The end of aesthetic discourse is rather a distinctive state of mutual interpersonal valuing between individual appreciators. I will argue that we should replace Convergence with Community:

The result of successful aesthetic conversation is a state of mutual interpersonal valuing among individuals, grounded in their joint aesthetic engagement. Individuals reach this state when they harmonize, or, as I prefer, when they vibe. While sharing sensibilities is often sufficient for vibing, it is not necessary. Philosophers thus tend to mistake the good of harmonizing individuality with the good of shared sensibility.

For a taste of the difference between Community and Convergence consider the following dialogue, supposing that S and A are acquaintances in a casual conversation:

We can easily imagine that the conversation carries on like this. By the lights of Convergence it is not clear that this is a worthwhile or successful aesthetic conversation. S and A do not converge in their views about the aesthetic value of the bridge or the new building, and they are not attempting to converge. By the lights of Community this is an excellent aesthetic conversation. S and A are aesthetically engaging with the bridge and the building, and in doing so they are engaging with and appreciating each other’s sensibilities.

My argument for Community focuses on the idea that aesthetic discourse has a distinctive force, which I describe at three levels: illocutionary, dynamic, and discourse-governing. I begin by surveying some peculiar features of aesthetic discourse (§2). This motivates an investigation of a common but understudied view, namely, that typical aesthetic utterances have directive illocutionary force: What exactly is the ‘directive force’? At the level of illocutionary force, several options stand out: demand, request, recommendation, and invitation (§3). I critically survey each, develop an original account of the felicity conditions of invitation, and argue that the directive illocutionary force of typical aesthetic claims is invitational (§§4–7). The reason, I argue, is that aesthetic discourse is conditioned by the norm of Community, not Convergence (§8). While other illocutionary forces can be deployed to achieve the end of Community, invitation holds pride of place. To make this last point more precise, I develop a formal dynamic account of aesthetic utterances and show how the illocutionary force of invitation serves those dynamics (§9). If this is right, then a range of views in aesthetics, metanormativity, and the theory of communication need to be revised (§10).³

Aesthetic discourse is an eclectic affair. It includes informal discussions about what to wear, read, watch, cook, and play, or how to decorate, design, or arrange. It includes everyday discussions about what is aesthetically good or bad, better or worse, mediocre or mixed. We make recommendations and suggestions; we persuade, encourage, and warn. We say you must see this or that, or you have to check this out. Aesthetic things are said to be essential, vital, necessary. Artists talk to each other about their works in ways that differ from how appreciators, curators, and critics talk about them. Aesthetic discourse happens across a wide variety of practices with an array of aims, from critiquing, reviewing, evaluating, and ranking, to curating, educating, and awarding.

Given this variety, insight into the dynamics of aesthetic discourse requires a narrower target. Specialized aims complicate the conversational contexts of critics, curators, artists, awards institutions, and educators. And it would be a mistake to understand all aesthetic discourse on the model of critic-audience, educator-student, or curator-public. My focus here is on everyday aesthetic discourse on the grounds that, while such asymmetries are often present, they are not baked into the conversational context. It is the relatively simplest and arguably most fundamental case.

Suppose that in the course of a conversation a question comes up about the year a bridge was built. In Deny an addressee, A, responds to a speaker, S:⁴

Here A denies what S says. In a typical conversation where S and A are epistemic peers who disagree, there is significant pressure on S and A to consider who is right and review the matter with the hope of settling it. Compare this with an aesthetic conversation:

Again, A denies what S says. But here there is less, if any, pressure to settle the issue. The conversation can easily carry on without S and A trying to settle their disagreement. S might lower her credence a bit and, perhaps, look more closely at the bridge, if only to affirm her claim.⁵ If a convergence norm explains the character of the non-evaluative disagreement, then something special is going on in the aesthetic case.

Now consider the same conversational context, but instead of denying what S said, A gives a ‘disengaged’ response. A is silent, silently nods, or simply says ‘oh’.

A default acceptance norm typically governs non-evaluative conversation (Walker 1996). A’s response effectively adds S’s claim to the common ground, as something both S and A accept. Again, we can see a convergence norm operating here. If a communicative aim is to converge in our non-evaluative representations, then it would make sense to have a default acceptance norm in play.

Compare, however, a disengaged response in an aesthetic conversation:

There is something strange in A’s response. It seems that A should engage with S’s claim in one way or another.⁶ Several responses are available:

There seems to be no default acceptance norm. It is not so surprising, then, that intensifiers like ‘totally’, ‘absolutely’, etc. are acceptable, indeed common, in aesthetic discourse. Consider Intensify (Beltrama 2018a):

In contrast, it is odd to indicate such agreement with factual claims.

“Totally” and other English intensifiers are typically inappropriate in response to non-evaluative claims. If a convergence norm explains the existence of default acceptance in non-evaluative discourse, then something peculiar is going on in aesthetic discourse.

Another way in which convergence clearly comes into conversational play is when we question the relevance of our interlocutor’s claim. Suppose we are hanging out and, out of nowhere, I say of a bridge in our midst, “That bridge is beautiful.” In addition to meriting some response or other (as per Disengage), your response should not dismiss my aesthetic claim. Consider:

You might agree, disagree, elaborate, or whatever, but it seems wrong to say “So what?” In contrast, dismissing non-evaluative claims is more apt. Suppose we are hanging out and, again out of nowhere, I say, “That bridge was built in 1967”. Now a dismissive response sounds better:

“So what?” might not always be appropriate, but it is more apt here, when the conversationally-relevant features of the bridge are non-evaluative. “So what?” we might say, “Why are you bringing that up?” A dismissive response is strikingly less apt in the aesthetic case. If a convergence norm governs non-evaluative discourse, then dismissive responses to irrelevant claims would make sense. It matters that we converge, so it matters that I know why you are making an apparently irrelevant claim.

The same effects discussed here occur with other aesthetic claims. Replace “That bridge is beautiful” with “Those clouds are wild” (said of big grey clouds in the distance), or “The bridge is stunning,” “The ocean is so (wonderfully) blue”. But not all aesthetic claims exhibit these effects. As I discuss below, some effects diminish as the likelihood increases that the interlocutors vary in their aesthetic valuing practices. Consider for example replacing in Disengage or Dismiss “That bridge is beautiful” with “The new Ernaux is fantastic,” or “Wow, that ’65 Mustang is stunning.”

There are distinctive features of aesthetic discourse that need explaining. Deny, Disengage, and Intensify show that interlocutors do not converge in the ways they do with non-evaluative claims. Dismiss and Disengage show that aesthetic claims call for a response from the addressee in a way that differs from non-evaluative claims.

These peculiarities of aesthetic discourse alone do not count against Convergence; they suggest at least that aesthetic matters present special obstacles to convergence, and aesthetic discourse deploys special means of achieving it. When it comes to aesthetic claims, there appears to be, to put it roughly, a defeasible presumption of engagement by our interlocutors. We can use aesthetic claims to call for engagement in ways that differ from non-evaluative discourse. Disengaged responses and dismissals are failures to take up the defeasible presumption of engagement and so seem inappropriate; intensification and denial are forms of engagement and so the conversation carries on. How should we characterize this presumption of engagement?

Deny, Disengage, Intensify, and Dismiss make sense if the force of aesthetic discourse is directive. If aesthetic claims tend to issue directives to aesthetically engage, then disengaged responses and dismissals amount to refusals to heed a directive. Responding with an intensifier indicates that the addressee has followed the directive. Denial does not throw the conversation into crisis if the addressee can follow the directive even when they disagree.

Thus it seems that when we make typical aesthetic claims of the form that o is A for some aesthetic adjective ‘A’ and object ‘o’, we tend to communicate a directive to aesthetically engage with o, which, supposing the directive is felicitous, generates a defeasible presumption of engagement. Non-evaluative discourse does not typically issue such directives, and so are not typically responded to in the same ways.

Let’s call this the directive character of aesthetic communication. Many philosophers have emphasized this feature of aesthetic discourse, though few have given detailed accounts. Kant himself emphasizes it when he writes that when someone ‘proclaims’ to others that something is beautiful, “he demands that they agree. He reproaches them if they judge differently, and denies that they have taste, which he nevertheless demands of them, as something they ought to have” (§7, original emphasis). Isenberg writes that aesthetic claims ‘direct our attention’ (1949: 339) because “it is a function of criticism to bring about communication at the level of the senses, that is, to induce a sameness of vision, of experienced content” (1949: 336). Harrison writes that, “The puzzling ‘objectivity’ … of critical judgments flows from the fact that part of their use is to direct other people, whereas ‘I like it’ and its ilk are directively neutral” (1960: 222). More recently, Nehamas (2007) writes that

Javier González de Prado Salas and Ivan Milić (2016) develop a hybrid theory of aesthetic discourse and argue that the illocutionary force of aesthetic claims is typically that of a report and a recommendation. Anthony Cross, drawing on Ziff’s neglected (1966), writes that the critic’s task is to ‘invite’ engagement and to ‘recommend’ acts of appreciation through their aesthetic claims (2017: 305–6).

Accounts of the directive character of aesthetic discourse must answer several fundamental questions, among them: What is the force and content of the directive? Why does a directive with such a force and content feature in aesthetic discourse?

The issue is complicated by the fact that there are two levels at which we might locate the directive ‘force’ of aesthetic discourse. Here I am persuaded by Yalcin (2018), who distinguishes between illocutionary and dynamic force. The illocutionary force of a speech act concerns what the speaker is trying to do in using a certain phrase in a certain conversational context. A parent who needs their teenager to go to the store might use the sentence ‘go to the store’ with the force of a demand to get them to buy some groceries. A person in need of care might use the same sentence in conversation with their friend with the force of a request. One might also use the sentence with the illocutionary force of assertion or inquiry.⁷ Dynamic force is located at a more abstract level and concerns the broader goals of linguistic exchange. The dynamic force of an utterance type concerns “the characteristic kind of change to the state of the conversation the utterance is apt to produce” (Yalcin 2018: 402). Assertion, for example, adds information to the conversation. Inquiry focuses the conversation in a certain way. Dynamic force thus prescinds from speaker intentions in conversational context to describe how various utterance-types tend to affect the course of the conversation. An influential account of the dynamic force of imperatives, for example, is that they update the addressee’s ‘to do’ list (Portner 2004; I draw on Portner’s work below in §9).

To understand what is distinctive about aesthetic discourse, we need to answer these fundamental questions about its force. Given that I want to characterize both dynamic and illocutionary force and explain how these forces interact in aesthetic discourse, there is a question about how best to proceed. I propose we start with the more concrete before moving on to the more abstract. I will start with illocutionary force and study aesthetic discourse in action. This will help us zero in on the character and power of invitation, while exploring some general normative features of aesthetic discourse (§§4–7). From there we can start to appreciate the merits of Community over Convergence (§8) and develop a formal theory of the dynamic force of aesthetic discourse (§9).

The first step, then, is to survey directive illocutionary forces to see what our options are. Mind my use of ‘directive’ here. Imperatives are verb forms that we use to convey directives. We typically associate directive force with the force of commands, which impose obligations on the addressee. The second sentence of this paragraph is a typical imperative. But imperatives do not always have such force; sometimes their force is weaker than a command, and they do not impose an obligation on the addressee. Consider this list (see also Charlow 2014: 542): Come over. (invitation) Take highway 5 S to the 8 E and exit Montezuma Rd. (instruction) Grab some wine on your way. (request) Go on in. The dining room is down the hall. (permission) Don’t drink the whisky on the third shelf. (demand/command) Eat the food. (threat/dare) Drink some Pepto-Bismol and call me later. (recommendation/advice) Rest on the couch and watch a couple TV shows. (suggestion) Get well soon. (well-wish) Die, you horrible cook! (expression of contempt) “Come over” and “Go on in” invite and permit action, respectively, and so seem not to impose obligations on the addressee. Their force is thus weaker than that of a command, but they still involve attempts to influence the addressee’s actions. In what follows I use ‘directive force’ to include the force of both ‘strong’ and ‘weak’ imperatives.⁸

Several of the directives listed above are non-starters for our purposes—permission, threat, well-wish, and contempt rarely if ever feature in aesthetic utterances. But that leaves six other options: demand, instruction, recommendation, request, invitation, and suggestion. Two strike me as unhelpful: instruction and suggestion. In everyday aesthetic discourse we do not typically instruct others to consider the aesthetic value around us, even if that is the typical mode of address for an arts educator, curator, or perhaps critic. And a typical suggestion to φ presents φ-ing as optional, which will not capture the defeasible presumption of engagement.

This leaves us with demand, request, recommendation, and invitation. Perhaps there are other options; the list is representative, not exhaustive. However, the four directives surveyed here are, I think, by far the most plausible, and as we saw above there is precedent for characterizing the directive force of aesthetic claims in terms of demand, recommendation, and invitation.

Here are the options, then. For an aesthetic predicate ‘A’, ‘That o is A’ typically communicates the directive to aesthetically engage with o, the illocutionary force of which we might understand as:

How should we understand the difference between these directive forces? Directives have various felicity conditions (Searle 1979: 44). A preparatory condition states that the addressee, A, is able to do what is directed, φ. S’s directive to A to grab the keys is infelicitous if the keys cannot be grabbed. A sincerity condition states that the speaker, S, wants A to φ. And an essential condition states that S’s directing A to φ is an attempt by S to get A to φ.

For my purposes here, it is best to characterize these directives in terms of their presented reason(s), that is, the reason(s) S suggests A has for doing what S directs A to do.⁹ In other words, different directives suggest different relations between the sincerity and essential conditions. S wants A to φ and attempts to get A to φ by saying something in a certain way. In a demand, S presents the authority of their desire for A to φ as the reason for A to φ. In a request, S presents their desire as the reason for A to φ. Recommendations present φ-ing as of value to A. And with an invitation, S presents φ-ing as something either A or S would value A’s doing. I elaborate on these conditions below.

The content of aesthetic directives is the minimal directive aesthetically engage with o. It is not the more specific engage with o as A, for some aesthetic predicate ‘A’. Of course, a speaker might end up communicating the more specific directive when further conditions are met, for example, when we know the addressee will value something’s beauty, or when we are discussing comparative aesthetic claims, for example, when we say that the yard has a cleaner look now or that the dog’s haircut is better this time. The reason is this: either Convergence governs the conversation or it does not. If it does, then specifying how the addressee ought to engage with o is redundant; they presuppose that they ought to converge, so all the directive has to do is direct the addressee to aesthetically engage. If Convergence does not govern the conversation, then there is no presumption that the interlocutors ought to agree on o’s aesthetic character, and so a directive to engage with o as the speaker does is out of place unless the context enjoins it.

What we will study is how Convergence pairs up with each illocutionary force to structure aesthetic discourse. As we will see, broader normative features of aesthetic discourse cause problems for demands, requests, and recommendations. Invitation fares best, but for reasons that cast doubt on Convergence.

We noted above Kant’s claim that (pure) aesthetic judgments “demand” or “require” agreement from others. Kant writes,

Kant claims that a requirement or demand is essential to a judgment by which we proclaim and declare something to be beautiful—so essential that it would not occur to anyone to use the term ‘beautiful’ without having such a demand in mind. Kant puts this in terms of demanding or requiring agreement, having taste, and liking the object, but it makes little sense to issue such demands without also issuing a demand to engage with the object.¹⁰

Kant appears to embrace the view that aesthetic utterances have strong directive force. He also seems to embrace something like Convergence. He thinks that our capacity for aesthetic judgment could be defined in terms of our ability to judge from what he calls a “communal sense”, that is, to judge whether the way one responds with pleasure to certain representations is the way everyone ought to respond:

Here Kant emphasizes the convergent aim of aesthetic communication: the “common” in “common sense” becomes “communal”. By “common sense” Kant has in mind a capacity to demand shared feeling with respect to certain sources of pleasure. Indeed, demanding this is more or less what the judgment of beauty is for Kant.

Let’s use the term “Kantian” for any view that embraces Convergence and holds that the illocutionary force of typical aesthetic claims is that of a demand. (I leave it open whether Kant himself is committed to such a view.)

How plausible is Kantianism? It seems to be unable to account for the conversational patterns surveyed in §2. To see why, though, we need to bring in another feature of aesthetic discourse: aesthetic claims are a paradigmatic way of expressing, and communicating about, our individuality. Through aesthetic discourse we express our sense of humor, our love of art and food, our interest in clothes, décor, music, literature, landscapes, and so on. Our aesthetically valuing these things largely constitutes our individuality, and without aesthetic discourse our ability to cultivate and communicate about our individuality would be severely impaired. Through aesthetic communication, we communicate features of the individual we are, in ways that facilitate others taking up and engaging with our aesthetic claims, and thereby with our individuality, while expressing their individuality in turn.¹¹

Kantianism does not facilitate the expression and mutual appreciation of individuality. For the Kantian, an aesthetic claim communicates, in broad effect, “We presuppose that we ought to converge in our aesthetic attitudes, and I demand that you engage with this thing, which I regard as beautiful.” The speaker need not state the obvious, namely, that the addressee ought to find it beautiful, too. So aesthetic claims effectively demand that the addressee value what the speaker values, in the way that the speaker values it. But this is tantamount to demanding that others adopt one’s own practice of aesthetic valuing, that is one’s own individuality.

However, a demand to do so is too strong; it disrupts a central feature of the practice of aesthetic valuing. The practice of aesthetic valuing is discretionary, in the sense that our aesthetic valuing practices are the product of discretionary choice. Some of our valuing practices are non-discretionary or compulsory: no one gets to exercise discretion in whether love is worth valuing, whether murder or slavery are to be disvalued. In some cases, even questioning whether we should have these attitudes is morally dubious. Aesthetic valuing practices are different; they are discretionary in the sense that they are the product of free choice. This is reflected clearly in our patterns of aesthetic valuing and disvaluing, which vary widely across individuals: Some go in for K-Pop, minimalist sculpture, and fine prints; others go in for Norwegian Black Metal, cityscapes, and gothic design; for others still it is rap, art deco, and sleek motorcycles. If someone who loves minimalist sculpture does not include Norwegian Black Metal in their aesthetic valuing practices, that is ok. Even if, in some sense, they are ‘missing out’ on something, it is up to them whether to include Norwegian Black Metal in their aesthetic valuing practices. It is possible that some items of aesthetic value are such that, in some sense, everyone ought to value them.¹² But that would be a striking exception to the discretionary rule. Using such cases as a model for aesthetic valuing would be wildly distorting of a practice that exhibits the hallmarks of choice.

Given the value of individuality and the discretionary valuing practices that ground it, the Kantian’s demand to effectively share sensibilities would give the addressee reason to ignore or dismiss it as misplaced or presumptuous. If the content and directive force of aesthetic claims give the addressee reason to dismiss or ignore the directive, then Kantianism cannot generate the conversational patterns we observed in §2. A silent response or a dismissal should be just fine.¹³

How can a Kantian respond? To reject Convergence or to weaken the force of the directive is to reject Kantianism. The remaining option seems to be to reject the idea that individuality is worth anything, to think that we should not regard our aesthetic valuing practices as discretionary practices that result in a valuable variety among us. The Kantian must defend the idea that our aesthetic valuing practices are compulsory and ideally uniform practices, comparable to our moral valuing practices. That amounts to a radical revision of our practices, and we should seek an alternative before going down that path.

The most straightforward alternative is to weaken the force of the directive by weakening the presented reason. Demands express what the speaker authoritatively wants. By weakening this to a directive that expresses what the speaker simply wants we get a request.

Aesthetic claims are like requests in important respects. Under common conditions, if I deny a reasonable request to pass the salt, then I am open to a form of criticism. I have not done what was reasonably requested. It is not that I have failed an obligation or skirted duty; I have done something permissible but faulty. I am thereby subject to “suberogatory” critique (Driver 1992). Dismiss and Disengage are normatively similar. Dismissing the aesthetic claim with “So what?”, or by not responding at all, is faulty but not forbidden. I do not have a right to your response and you are under no obligation to give it to me. (Kantianism is vulnerable here, too, in treating the addressee in Dismiss and Disengage as subject to too strong a criticism.)

Furthermore, if I make a request to you, then I have created a reason for you to do what I request only if, at your discretion, you value me as a person. The addressee’s discretionary valuing is central to the normative character of requesting.¹⁴ To illustrate, suppose I am a perfect stranger to you. I knock on your door and request to use your bathroom. My request alone gives you no reason to accede—indeed, I might be disappointed if you turn me away, but doing so is permissible and fine. I requested, you declined, end of story. But now suppose you realize that I am the new neighbor whose bathroom is being remodeled and you come to value me as a neighbor. “Ah, the new neighbor!” you say. Now my request gives you, and I recognize that you now have, a reason to let me use your bathroom, but only because, now at your discretion, you value me or our nascent relationship. It is your right to do neither and to deny my request, though now if you do you are open to suberogatory criticism. “Ah, the new neighbor! Sorry, no.” Our practice of requesting typically involves attempts to present the addressee with non-obligatory reasons to accede in an implicit recognition of the fact that a bare request offers no such thing. When I ask you to pass the salt, it is obvious to both of us that we are dining together. The request is different if I am a stranger at another table.

Aesthetic claims are like requests in this respect, but where the relevant discretionary valuing concerns individuality. Aesthetic claims generate non-obligatory (pro tanto) reasons to engage in ways that depend on the speaker’s attention to the addressee’s discretionary valuing. We know better than to praise a new synth pop album to a friend who we know hates synth pop, or to enthuse about the aesthetics of a pair of sneakers to a flip-flop-loving neighbor who we know could not care less about sneakers. Suppose I am chatting with my neighbor who loves flowers but has no possible interest in the aesthetics of sneakers, and someone walks by with rare Air Jordans. I say, “Wow, those are dope.” My neighbor’s silent or mystified response is permissible and non-faulty because my directive to engage is insensitive to my neighbor’s individuality (or to his openness to individuality). I have thus failed to give my neighbor a reason to respond, and my aesthetic claim is akin to a bare request. But suppose instead that I notice a beautiful flower in the garden. “That’s a lovely flower,” I say. Now a silent response or dismissal carries normative weight.

Like requests, then, we find norms of aesthetic discourse that center around discretionary valuing. We aim to make aesthetic claims that are sensitive to the addressee’s discretionary valuing and so have a chance of generating non-obligatory reasons to engage. In doing so, speakers have to exercise skill in their attention to individuality.

Despite their similarity to requests in these important respects, aesthetic claims are not typically requests. One can add ‘please’ to a request. If “Pass the salt” is a request, then adding ‘please’ is fine: “Pass the salt, please” or “Please pass the salt”. Not so if “Pass the salt” is a demand. But aesthetic directives do not naturally absorb ‘please’ in the way that requests do. Indeed, adding ‘please’ seems to change the meaning of the directive into a request:

Adding ‘please’ seems to somehow weaken the force of the directive too much. Of course, one might explain this away by appeal to conventions of aesthetic discourse, but we can see how aesthetic requests are not directives in another way. A speech act can typically be performed by simply stating its sincerity condition. I can perform the act of questioning—“Where are you from?”—by simply stating the sincerity condition for questions: “I would like to know where you are from.” Likewise, I can state that this sandwich is tasty by saying “I like the taste of this sandwich.” Liking the taste of something is the sincerity condition for predicating ‘tasty’. The sincerity condition for requests is that the speaker genuinely want the addressee to perform the requested action. But I cannot replace “That bridge is beautiful” with any of the ways we might translate it into the sincerity condition for an aesthetic request: “I want you to aesthetically engage with that bridge” or “I want you to engage with that bridge’s beauty” or “I believe the bridge is beautiful and I want you to engage with it.”

Requests, like demands, do not capture the force of aesthetic directives. They share an underlying problem: Aesthetic claims do not merely express what the speaker wants the addressee to do; they also express the thought that it would be good for the addressee to do it. Kantianism tries to capture this with the thought that aesthetic claims communicate a demand, an authoritative desire—it would be good for you to do it because I authoritatively want you to (and we share an interest in converging). The request theory acknowledges that this is too strong, but its fix retreats too far.

The other two options, recommendations and invitations, both strike this balance. Both express a desire for the addressee to φ and, in different ways, the thought that it would be good for the addressee to do so.

Recommendations are commonly associated with aesthetic life. Travel guides, restaurant review sites, book reviews, art criticism—all contain recommendations about what to do, see, hear, experience. Javier González de Prado Salas and Ivan Milić (2016) develop a recommendation account of aesthetic discourse. They argue that the illocutionary force of aesthetic claims is typically that of a report and a recommendation, the content of which is that when appreciating an aesthetic item, “one should exercise a sensibility that would lead to its appraisal as beautiful” (2016: 207, their emphasis). Instead of demanding agreement, as per Kantianism, they recommend agreement by recommending shared sensibilities: “What is recommended … is that the audience is disposed to engage with the object in the same way as the speaker does, so that they get to share her aesthetic appreciation of it” (2016: 209, emphasis added). They thus endorse Convergence.¹⁵

The authors are unclear about how precisely to understand the illocutionary force of recommendations. In places they say, “when the speaker offers a recommendation, she is not describing things as being a certain way, but rather inviting the audience to do something, to adopt a certain attitude,” namely, one that aligns with the speaker’s (2016: 208, my emphasis). This is confusing because recommendations are not invitations—recommendations suggest, advise, or present something as suitable for the addressee; invitations elicit or encourage engagement in an activity. But they use the notion of an “invitation” loosely and construe recommendation in terms that better fit the illocutionary force of advice by including the speaker’s advisory attitude in the illocutionary force. This closely aligns with Hinchman’s (2005a; 2005b) account of advising as “inviting” to trust: “the advisor … represents herself as taking a perspective on the advisee’s self-interest, a perspective which she moreover represents herself as presuming may equal or exceed in authority for the advisee the perspective of the advisee himself” (2005b: 359).¹⁶

In general, to recommend is to propose an activity as suitable for the addressee, but the proposal may or may not account for the addressee’s actual desires. Your physician might recommend that you eat leafy greens despite your hatred of them. Call this an agent-neutral recommendation. They differ from agent-relative recommendations, which take the addressee’s actual interests into account. Suppose you are in line to get some ice cream, and you ask the server what they recommend. The server can offer an agent-neutral recommendation—“Our chocolate ice cream is the best around”—or they can learn something about you so as to be in a position to make an agent-relative recommendation—“Since you hate chocolate and love fruity flavors, try the mango.”¹⁷

Do Gonzáles de Prado and Milić have an agent-neutral or agent-relative notion in mind? On their account, aesthetic claims recommend that you engage with the thing as the speaker does, that aligning your sensibility with the speaker’s is in your interest. But notice that in general two things might be in your interest here: aesthetically engaging with the thing, period, however you please, and engaging with it as the speaker does. And either one could be construed as agent-neutral or -relative.

Gonzáles de Prado and Milić import Convergence into the content of the recommendation. As we have set things up here, Convergence is at work in the background as a discourse-governing norm. So to simplify things, we can leave Convergence where it is, and construe the content of the recommendation as a recommendation simply to aesthetically engage. And here, again, Hinchman is helpful: We can say that the recommendation to appreciate involves the speaker taking a perspective on the addressee’s self-interest that “may equal or exceed in authority for the advisee the perspective of the advisee himself” (2005b: 359). This effectively leaves it open whether the recommendation is agent-neutral or -relative. The recommendation proposes to A that engaging with o is worth A’s doing, in either an agent-relative or -neutral way.

This nicely fits some familiar patterns of aesthetic discourse. Often aesthetic conversations have the character of a joint enterprise where the interlocutors lack a representation of something’s aesthetic character and work together to construct one. Consider:

It is incorrect to characterize Tasting Notes in terms of recommendations to adopt each other’s view. Initially, neither of us has a firm view, and we use aesthetic discourse to jointly determine one. With Convergence in the background, the recommendation simply to engage leaves open the possibility that we will appreciate differently but converge on a shared view and thus satisfy the norm.

While the recommendation view gets Tasting Notes right, it struggles with other aesthetic conversations. A French friend tells the story of her American friend visiting Paris with her. They each wanted ice cream on a hot day. She went through the line and got a few flavors, but when her American friend requested a certain combination, the vendor’s response was “Non, c’est pas possible.” It’s impossible; you can’t have that. Let’s adjust the example and consider how it might go in more detail (and maybe move it out of France).¹⁸

Suppose you are having ice cream with some friends, and one of them asks you what you want.

Imagine your friend says this in a non-judgmental and friendly way, indicated perhaps with a light smile or kind gesture. How should you respond? It would be inappropriate to say, “It’s perfectly possible. Just put one scoop on top of the other!” Your friend expresses the thought that peanut butter-lemon is an aesthetically bad combination, so the response that it is physically possible to combine them misses the point. Other responses are intuitively good:

Alternatively:

Does the recommendation view model these dialogues correctly? As a recommendation, the friend’s refusal amounts to this:

You have already made it clear that you value the combination of lemon and peanut butter, so, pragmatically, the friend’s recommendation must be to engage with its disvalue. But you have just told your friend what you would like, that is, what your interests are. So, the recommendation also cannot be agent-relative in the sense that it takes your actual interests into account; it must be agent-neutral in the sense that the friend’s perspective is meant to exceed in authority your own.

Offering such an agent-neutral recommendation is inappropriate, given that you have just said what your interests are. Again, aesthetic valuing is a discretionary practice. Even though your friend is confident that the combination is bad, it is presumptuous for them to suggest that it is in your interest to reject as bad the choice you just made in your interest, while suggesting that their authority on this matter exceeds your own. Now here is the problem for the recommendation view: the friend’s response that it is a bad combination is not inappropriate—it is perfectly acceptable to point out a bad combination in this context. Thus, the recommendation view incorrectly models an appropriate discursive move as inappropriate.

With recommendation as our main directive force, we can appeal to what is in the addressee’s agent-relative interest or what is of agent-neutral interest. As a result, recommendation crowds out the speaker’s individuality, that is, their own individual sense of what is worth valuing. Agent-relative and agent-neutral recommendations crowd that out of the exchange. If the speaker could communicate what they find worth (dis)valuing and invite the addressee into joint aesthetic engagement, then the addressee could take them up on it in the spirit of shared activity and without necessarily changing what they find worth valuing.

Several philosophers have seen a connection between aesthetic discourse and invitation. We already noted Nehamas’s claim that “aesthetic descriptions … issue an invitation to look (or read or listen) for ourselves” (2007: 52). Nehamas further specifies this invitation as an invitation to join a particular community of appreciators who also see something in the aesthetic object and incorporate it into their lives (2007: 80–81). Aesthetic claims are “essentially social” (2007: 77, 85) because they express a hope of establishing communities around the aesthetic object. Nehamas thus seems to accept an idea adjacent to Convergence, namely, that a governing norm of aesthetic discourse is a shared hope for community around an aesthetic object. However, Nehamas firmly distances himself from the Kantian hope for a universal community of shared feeling and emphasizes instead the view that aesthetic discourse establishes smaller like-minded communities: “Aesthetic judgment never commands universal agreement, and neither a beautiful object nor a work of art ever engages a catholic community” (2007: 81). “[N]o community I hope to create around something I find beautiful is ever a universal community” (2007: 82). Nehamas likens his preferred sense of community to “the pagan cults of ancient Greece, which recognized their common concern with the divine despite the different forms in which they worshipped it” (2007: 82). The ‘divine’ is beauty and different ‘cults’ or aesthetic communities find beauty in different things. For Nehamas, aesthetic communities are communities of shared interest in particular aesthetic objects and practices.

There is also a connection between aesthetic discourse and invitation in Ted Cohen’s (1988) work, who holds that aesthetic claims invite others to take aesthetic objects seriously and incorporate them into their lives. In earlier work Cohen (1978) develops the view that metaphors and jokes are invitations. More recently, Ernie Lepore and Martin Stone (2015) develop invitational accounts of metaphor, sarcasm, hinting, humor, and irony—which they construe as invitations to adopt a perspective, explore, engage, or appreciate.

Unfortunately, given my purposes here, none of these philosophers offer accounts of the speech act of invitation or work out the general connection between invitation and aesthetic discourse. However, in several works, Kukla (2018) and Kukla and Lance (2008; 2013) describe some of the distinctive features of the speech act. Kukla (2018) is especially helpful. Kukla notes that invitations “are a fascinating and complex speech act” with “a distinctive normative structure and illocutionary force” which “leave the invitee neither obligated nor with a neutral free choice” (2018: 81). Kukla highlights some of invitation’s felicity conditions and norms of appropriateness governing inviter and invitee. One of the felicity conditions is standing—the speaker must be in a position to issue the invitation. Recommendations do not have this feature: I can recommend that you visit the doctor but I cannot invite you to unless I have the standing to invite, unless, for example, I am the doctor. Kukla does not offer a theory of standing, but it appears to be a matter of oversight regarding exclusive goods. I own my house and so control how people access the goods it provides; the doctor owns the business that provides the goods of her care; a parks management team oversees the use of certain natural resources. Rights to, control over, and management of, the goods generate standing to invite people to engage with them.

Other norms govern the appropriateness of offering and accepting invitations. Even if you have standing to invite, it is typically inappropriate to invite a new friend to your uncle’s funeral. And when an appropriate invitation is issued and accepted, gratitude is called for from both parties: I am glad you accepted the invitation and you are glad that I offered it (as Kukla 2018: 81–85 discusses).

These features of invitation capture some of its distinctiveness, but they do not fully specify the speech act of invitation. To complete the picture, consider informal invitations to do something, to come over for dinner or head to the beach. These are characteristically offered on the grounds that either you would value doing the activity or I would value your doing it. Ideally, it is both: I invite you to do something because I think we would value our doing the activity. If neither you nor I would value doing something, then, other things being equal, an invitation to do it is inappropriate (instead, we demand, dare, challenge, and so on).

Furthermore, when S invites A, S regards A as free to do what A pleases. For example, when “Sit down” is an invitation, it might be glossed as a sincere “Feel free to sit down; it would be nice (for you, for us).” With invitation, S regards A as free to do as A pleases, but this does not mean that S does not care what A does. Through the speech act of invitation, S communicates that they see value in A’s doing the thing and want A to do it but leave it up to A. We can thus characterize the ‘presented reason’ of invitation as communicating S’s desire for A to φ for the reason that A or S would value A’s φ-ing.

Invitation is also infelicitous when the invitee is already sufficiently performing the invited action. If A is sitting down and S says “Have a seat”, then A fails to communicate an invitation. The natural reading is that A is being sarcastic and indicating that A took liberties they did not have. Some actions come in degrees, as does aesthetic engagement. Someone eating a small amount of cake can be felicitously invited to eat the cake. Likewise, someone appreciating a sunset can be felicitously invited to appreciate it as long as the speaker believes they are not fully appreciating it. “So beautiful,” they might say to you, as you are looking at its beauty, “The way it goes from pink to orange.”

Putting these thoughts together supplies felicity conditions for invitations. S’s invitation to A to φ is felicitous only if:

With a clearer picture of the speech act of invitation, we can look at how invitation pairs with Convergence to model aesthetic discourse.

A virtue of this view—indeed, its main attraction—is the way it centers the individual-focused discretionary valuing that features in aesthetic discourse. An invitation is infelicitous unless, per clause (c), the speaker believes that either they, the speaker, or the addressee would value the addressee’s φ-ing. Typically, this clause is satisfied because the speaker believes that either the addressee would value φ-ing or they both would value their φ-ing together. But the speaker cannot arrive at this belief without attending to the addressee’s discretionary valuing. Would this individual value sitting down, coming to the party, having coffee, going on the road trip? This gives us a new handle on Disengage. You know your neighbor does not aesthetically engage with sneakers (and would not be open to doing so), so when you try to invite them to engage with the Air Jordans your invitation is infelicitous. But when you invite them to engage with the flower, you know that it is something they would value doing. The invitation carries through and merits a response.

This allows the invitation view to handle the ice cream conversation better than the recommendation view. On the invitation view, the friend’s claim is an invitation to engage:

Your first response, “Hmm, perhaps … ” takes up this invitation by acknowledging that there might be a way of disvaluing it. But notice that you can take up this invitation on the grounds that the speaker would value your doing so. You do not need to change your aesthetic view or modify what you regard as in your aesthetic interest. In other words, unlike recommendation, your practice of discretionary valuing need not be modified in the slightest to follow the directive. Your second response, “But it’s actually really good!”, declines your friend’s invitation and offers an alternative invitation, which your friend immediately takes up.

Thus the invitation view has no trouble explaining why each response is good. The view allows us to maintain our practices of discretionary valuing while exploring other options. Spontaneously, adventurously, or experimentally agreeing to an activity is exactly the kind of thing we do in response to invitations. Doing so might change your sensibility as a result—“Wow, lemon and pear is way better than peanut butter and lemon!”—but you did not change your choice in the hope that it would, or because you modified your interests and rejected your choice.

Another way to highlight the difference between the recommendation and invitation views is to look at how they construe the disapproval that would be merited if one were to straight decline the invitation or ignore the recommendation.

How should the friend or an observer feel in response to your abrupt refusal? If the friend has just issued an agent-neutral recommendation, then your refusal amounts to a refusal to take advice about what is in your agent-neutral interest.¹⁹ Little has happened to shake whatever confidence your friend has in issuing the agent-neutral recommendation. They might feel slighted due to your failure to recognize their trustworthiness, but from their point of view it is really your loss. “All the worse for you on such a hot day,” they might think.

But if the friend appropriately issues an invitation, then your stark refusal amounts to your ignoring the invitation without offering either a reason to refuse or an alternative invitation. In that case your response merits suberogatory criticism, and that is exactly what we should feel in such a situation, either as the friend or as someone observing the exchange.

The difference here is due to the fact that recommendations typically reveal a reason to do something provided you have no reason to distrust the recommender. Your trust in the recommender is not trust first and foremost in a valuing individual as such but in a person as a source of evidence. If you are offered a recommendation and have no reason to distrust the recommender, then there is some defeasible normative pressure to at least consider the recommendation. Flat out declining the recommendation amounts to a slight, due to a failure of recognition, a failure to recognize a person’s status as a trustworthy source.

But that is very different from a failure to take up an invitation to joint activity that engages each individual as such. Construing aesthetic claims as recommendations fails to highlight the individual-focused discretionary valuing that is the heart of aesthetic life. Aesthetic discourse engages us first and foremost not as trustworthy sources of evidence but as individuals with sensibilities and styles. Whether and how we engage with someone’s aesthetic claims depends on our sensitivity to each other’s modes of discretionary valuing. The invitation view gets that exactly right.

The invitation view thus has unique resources to capture the sense of connection that aesthetic discourse can generate among the interlocutors. When aesthetic invitations are rightly pitched and received, they generate reasons not just to engage with the aesthetic object but to value each other as individuals. To see this, let’s suppose that S says ‘o is beautiful’ to A and thereby invites A to aesthetically engage with o. There are two grounds on which A might take up the invitation and each generates additional reasons for interpersonal valuing. Suppose that A takes up S’s invitation to aesthetically engage with o because:

Ideally, both (i) and (ii) hold: S offers and A takes up the invitation because they would value their joint aesthetic engagement. To get these results A and S have to exercise sensitivity to aesthetic and interpersonal reasons to value. S has to exercise sensitivity to A’s individuality—to her modes of discretionary valuing—and A has to be responsive to S’s invitation and to the reasons S’s claim generates. In taking up these reasons, S and A have reasons to value each other. The result, when they do, is a community of individuals whose aesthetic valuing practices are mutually supportive.

When individuals interact in such a mutually supportive way—when they harmonize or, vibe—they form a special kind of community. Some individuals are fitted to one another; their practices of aesthetic valuing are mutually supportive and enriching. Two or more individuals vibe when their practices of aesthetic valuing are mutually supportive. Maybe they value the same things, appreciate each other’s style, or gain something from each other’s aesthetic insights. We see this in the way bands interact, when band members with different skills and styles work together to create something good. We see it at good dinner parties, when for example a strong and welcome personality is repeatedly met with a dry humor. We see it in discussions about literature or fashion, when people differ in what they value but benefit nonetheless from aesthetic conversation. We see it in the way a person is inspired by another’s style, in fashion, sports, freestyle battles, improv acting, and many other aesthetic interactions and practices. Vibing, then, is a dynamic interpersonal state that promotes aesthetic discourse by promoting the expression of and attention to individuality. Aesthetic invitations aim at this harmony of individuality, and in this way, the invitational force of aesthetic directives centers the discretionary valuing at the heart of aesthetic life.

Let’s summarize §§4–7: I argued that invitation is the illocutionary force that best captures the directive illocutionary force of typical aesthetic claims. Demands and requests fail to express what it would be good for the addressee to do. Recommendations achieve this, but their agent-neutrality is limiting. Aesthetic discourse is a communicative practice in which we direct each other to exercise our special discretionary valuing capacities, in a way that reveals something valuable about the individuals who have and exercise those capacities. Given the variety of individuals, and the nature of the valuing states they aim to enter, we tend to aesthetically address one another invitationally, enticing others to attend to and engage with what we notice and to express themselves in turn. We use aesthetic utterances to invite each other to exercise our discretionary valuing capacities in the hopes that we will vibe, or achieve a sort of community of individuals.

We should thus predict that aesthetic conversations are non-defective when they are rightly aimed at this state of mutual valuing of individuality. This suggests that Community best captures the end of aesthetic discourse.

Now recall:

Community and Convergence entail different constraints on which aesthetic conversations are worthwhile. Convergence restricts the sensible conversations to ones whose participants have reason to think that their discretionary valuing practices align. Recall Egan (2010) who claims that convergence is “the central business” of aesthetic discourse. Egan writes,

Community says that individuals in aesthetic conversation should vibe. It allows for sensible aesthetic discourse among non-convergent discussants so long as they reasonably believe they might reach a state of mutual valuing of individuality. If converging is necessary and sufficient for vibing, then the two norms are equivalent. So the question is whether we can vibe without converging, and whether we can converge without vibing. I will argue that we can do both.²¹

Let’s start with the claim that convergence is necessary for vibing. Are there worthwhile aesthetic conversations in which the parties either do not converge, or reasonably take themselves not to? Consider a range of common scenarios:

People can vibe in Full Agreement, Verdict Agreement, and Verdict Disagreement. We frequently disagree about what to aesthetically value and how to do so, but our disagreements do not always preclude our vibing. I might value you as an individual even when we disagree about whether something is aesthetically good, whether an album is worth listening to, how to interpret a poem or film, how to dress or decorate, or whether a restaurant is good, and so on. We need not approach such discussions as attempts to make the other more like ourselves. And we need not always end such discussions thinking each other is defective, incompetent, or imperceptive. It is enough when we each have an appreciative and insightful sense of the valuing individual each other is. We enter states of mutual valuing in a range of non-convergent scenarios: critics whose reasons or conclusions differ, artists whose sensibilities diverge, friends who enjoy disputatious aesthetic discussions, people on social media. Community thus allows for worthwhile aesthetic discussions in Verdict Agreement and Verdict Disagreement. In other words, it allows for sensible conversations between interlocutors who express different aesthetic views, as long as they can reasonably expect to vibe. It thereby recognizes the fact that so many of our aesthetic conversations are not about reaching consensus but about exploring nuance in our sensibilities, our interpretations, our attention to detail, our aesthetic actions—in how we exercise our special discretionary valuing capacities.

What about Strong Disagreement and Estrangement? Community will often fail in Strong Disagreement, but perhaps not always. Perhaps I can disagree with you and think you are making a mistake but also think that it is an interesting mistake that plays into your sensibility and style in curious and appreciable ways. In other words, your mistakes might reveal something valuable about your individuality. Only in Estrangement does the presupposition of Community clearly fail.²³

Convergence and Community agree on Estrangement and Full Agreement, but Convergence construes Verdict Agreement, Verdict Disagreement, and Strong Disagreement as defective conversations or ones where the interlocutors should prioritize changing each other’s view. Community thus captures more of our non-defective aesthetic conversations. But it also outperforms Convergence at what it is designed to do, for even when we do aesthetically agree, agreement is not what aesthetic discourse is fundamentally about. Interlocutors who know they aesthetically agree about a song, painting, or novel will continue on and on, sharing their respective insights, commenting on each other’s interpretations, and revealing disagreements that lead to further insights. Convergence struggles to make sense of this. Why should it matter that we converge on the slightest little details, especially when pressing further will likely reveal ways in which we do not converge? By the lights of Convergence, it looks like we are testing our bond to its limit, or even trying to break it. Any answer from Convergence will take some finessing, but the answer from Community is straightforward: Digging further into the details reveals further nuance in our individualities and promises to deepen our mutual appreciation whether or not we converge.²⁴

This suggests that we seek something other than convergence in aesthetic discourse, which casts doubt in turn on the thought that agreement is sufficient for vibing. Agreement alone will not necessarily satisfy Community. Some philosophers have imagined a world in which everyone agrees on what has or lacks aesthetic value. Nehamas (2007) writes,

Nehamas seems to think that fully converging with others would be a ‘frightful’ and ‘desolate’ state of affairs, that is, give one reason not to aesthetically engage. Nehamas suggests that what is required is something that agreement may or may not provide, namely, insight into the way that aesthetic claims articulate a broader sensibility (2007: 85–86) or a ‘style’. If that is right, then agreeing about an aesthetic claim is also insufficient for vibing. We also have to be attracted to the individuality expressed by the claim.

To endorse Community and reject Convergence is not to claim that aesthetic discourse is possible in the total absence of aesthetic agreement. If there were no agreement at all among individuals—shared aesthetic predicates, basic shared interests in aesthetic valuing in general, mutually familiar aesthetic responses, generally shared and mutually intelligible practices of listening, seeing, making—then their valuing practices would be mutually unintelligible. Neither of them would be in a position to form beliefs about whether the other would value aesthetically engaging. As a result, all aesthetic invitations would be infelicitous. In this way, some commonality in aesthetic practices is an important enabling condition of aesthetic discourse.

But so many philosophers have treated this enabling condition as the main point of aesthetic discourse, as though aesthetic conversation begins with supposing similarity and ends by affirming it. But that is not how so many of our aesthetic conversations go. Community explains why we let so many ‘disagreements’ stand in aesthetics, why critics, artists, and designers who have sharply differing aesthetic views nonetheless admire one another and engage in disputatious aesthetic conversation with no hope of convergence. The fact that we disagree on aesthetic matters is often neither here nor there as long as we can find ways of valuing each other’s individualities. But we cannot always get there—our styles clash, our aesthetic commitments and interests diverge. We might push our aesthetic disputes further and try to find a way in, or go our separate ways. In this light, we can see a wider range of aesthetic disagreements as being not about achieving convergence but about finding or perhaps creating enough aesthetic common ground to value each other as individuals even when we are very different.

Invitation holds a special place in the economy of aesthetic discourse, but its pride of place is not wholly exclusive of other illocutionary forces. An art teacher’s aesthetic claims might resonate with instruction, given the goals sought by student and teacher; a critic’s claims might invoke recommendation, given the critic’s relationship with their audience; and a friend might seem to voice a demand when she knows you would love something or seeks agreement. And as we have noticed, many philosophers have thought that aesthetic discourse is directive while, at the same time, offering a wide range of perspectives on that directive force. When it comes to the directive character of aesthetic discourse, why is there so much variety in both theory and practice?

To answer this question, we have to ascend to a more abstract description of aesthetic discourse. We can describe the directive force of aesthetic utterances at a higher level of abstraction to appreciate that there is a sense in which they do not have a force as finely articulated as invitation or demand. This dynamic ‘directive core’ is retained across different contexts, and in different contexts, different illocutionary forces can step into the appropriate dynamic role.

We can model the dynamic force of aesthetic discourse by drawing on formal tools in dynamic pragmatics. The analysis follows in the tradition of Stalnaker (1978), Lewis (1979), and many others and targets how successful utterances affect the conversational state by adding information to it, thereby ‘updating’ the ‘common ground’.²⁵

I will ultimately work with a modified Farkas and Bruce (2010) model that has six elements, but it helps to work our way up to that by first considering simpler models. Following Stalnaker (1978) and others, conversations develop against a background of propositions that each interlocutor accepts. The propositions a participant is committed to are their ‘discourse commitments’ (DC). The ‘common ground’ (cg) of the conversation is the information the participants share—that is, the intersection of the sets of discourse commitments. And the context set (cs) is the set of all worlds that are compatible with the common ground. A context, C, is represented by these three elements:²⁶

Conversations unfold as participants add to or subtract information from the common ground. As Stalnaker writes, “To engage in conversation is, essentially, to distinguish among alternative possible ways that things may be. The purpose of expressing propositions is to make such distinctions” (1978: 85) And we typically do so by making assertions:

By making an assertion, the speaker adds its content to their discourse commitments, and provided there is no objection, it enters the common ground. We can capture a salient conversational effect of assertion, then, by modeling how assertions change the context. The conventional effect of assertion in context Ci is Ci+ S=Co, where ‘S’ is a sentence and ⟦S⟧ is the proposition expressed by ‘S’:

Informally, ASSERT says that an assertion of the sentence ‘S’ adds ⟦S⟧ to the speaker’s discourse commitments and to the common ground.

One problem with this model is that it does not capture the ‘proposal’ character of assertion. ASSERT simply adds ⟦S⟧ to the common ground, but in normal conversations when a speaker utters a declarative sentence (with a falling intonation contour), their speech act raises the question whether what they said is true. The addressee can then agree or disagree, silently nod, and so on. To capture the proposal character of assertion, Farkas and Bruce (2010) propose adding a ‘table’ that represents the current topic of discussion and a ‘projected set’ that represents how the common ground will change. Together these elements capture the proposal character of assertion: the table represents what’s at issue and the projected set represents how that issue will likely be resolved. A context, C, now includes these two additional elements:

Farkas and Bruce (2010) model the effect of a typical (falling) declarative sentence, ‘S’, as uttered by a speaker, sp, as follows. The conventional effect of an assertion in context Ci is Ci+ S=Co:

The utterance indicates the speaker’s commitment to the proposition expressed by S, ⟦S⟧, raises the issue whether ⟦S⟧, and by entering the projected set, establishes the expectation that ⟦S⟧ will enter cg and become a mutual commitment. The model thus captures the conversational expectation, voiced initially by Stalnaker, that an unchallenged assertion will enter the common ground. Assertions thus project confirmation and are accepted by default.

Beltrama and Rudin (2019) find a problem with the model. Drawing on Beltrama (2018a; 2018b), they argue that the model makes the wrong predictions for declarative ‘subjective’ sentences like “The movie was awesome”, because such sentences do not update in the same way as declarative ‘objective’ sentences like “The movie was set in 1995”. They note three ways in which the conversational effects of subjective sentences differ from the effects of ‘objective’ sentences. Subjective sentences do not exhibit the same automatic common ground update as objective sentences, their denials do not as frequently provoke conversational crisis, and unlike objective sentences they license distinctive response particles (e.g., ‘totally’). As surveyed in §2, aesthetic utterances also exhibit these and other effects. Formally, a sentence S is ‘subjective’ iff for a speaker sp:

In other words, a sentence is subjective if, in a given context, what is communicated when one speaker uses the sentence is not the same as what is communicated when another speaker uses it.²⁷ Not so with objective sentences. A sentence S is not subjective iff whoever utters it in a given context communicates the same thing:

To capture the update effects of subjective sentences Beltrama and Rudin import this distinction into a model of assertion. They define a unified model of assertion that, they argue, predicts different effects for subjective and objective sentences. For a sentence, S, a speaker, sp, and a context, C:

This says that an assertion adds the speaker’s acceptance of ⟦S⟧ to their discourse commitments, raises the issue as to whether all interlocutors agree, and projects a common ground update in which they do.

With objective sentences ⟦S⟧sp=⟦S⟧sp’, so the same proposition that enters the speaker’s discourse commitments (DCsp,o) also gets tabled and projected—this mirrors the original Farkas and Bruce model for assertion. Subjective sentences, however, are such that ⟦S⟧sp≠⟦S⟧sp’, so the proposition that gets tabled and projected is not ⟦S⟧ but the speaker-relative ⟦S⟧i1+…+in. Subjective sentences thus table the issue as to whether the addressee also accepts ⟦S⟧, and ⟦S⟧ⁱ¹ will not enter the cg until the addressee indicates agreement. The model thus represents the automatic common ground update for objective sentences and explains why subjective sentences require more explicit agreement.

The problem for aesthetic discourse is that the model puts ⟨S,⟦S⟧i1+…+in⟩ on the table and projects common ground update of ⟦S⟧i1+…+in across all interlocutors. This effectively imports the influence of Convergence into the model. One way to fix this is to table and project ⟨S1,⟦S1⟧i⟩ for some sentence S¹ related to S, where the relation between S and S¹, is constrained by Community. In other words, S¹ is an appropriate aesthetic sentence to utter in response to the speaker’s utterance of S:

This sets the expectation that the addressee will respond to the speaker’s utterance of S with a sentence, S1, about the object of their joint aesthetic attention. Of course, it is permissible that S=S1 and that S≠S1.

While Almost Aesthetic is an improvement, it does not capture the directive character of aesthetic discourse. It does not represent the fact that aesthetic discourse issues directives to aesthetically engage. For all the model says, aesthetic discourse might communicate the mere expectation of a response that puts ‘⟨S1,⟦S1⟧i1⟩’ on the table, where the expectation is little more than a convention of aesthetic discourse.

To capture the directive character of aesthetic discourse, then, we must modify the model to represent imperatives. Following Portner (2004) and (2013), we can complete the model by adding a ‘to-do list’ (TD) function that assigns properties to individuals. Portner (2013) characterizes the pragmatic function of imperatives as follows:

In other words, imperatives update participants’ ‘to-do’ lists. If I say ‘Come in!’ to you, then TD assigns the property ‘Comes in’ to your ‘to-do’ list, ‘TD(addressee)’. Let’s add the sixth and final element to our model:

With this on board, then, let’s use ‘I^ae’ to symbolize the aesthetic imperative, aesthetically engage with o and say that the canonical discourse function of an aesthetic sentence includes adding I^ae to TD(addressee).

The final wrinkle to iron out concerns the fact that the aesthetic imperative is weak, not strong. Consider these two imperatives (adapting Portner 2018: 307–8):

I⇓ is a strong imperative, typically said with a falling intonation. The speaker adds it to their to-do list as something the addressee should do and projects its addition to the common ground. Portner (2018) formalizes the effect of a falling imperative, I⇓, in context C as C+I⇓=C’, where:

In other words, a falling or strong imperative adds the content of the imperative to the speaker’s list of things the addressee shall do and projects mutual agreement along those lines.

I⇑ is a weak imperative, typically said with a rising intonation. The speaker typically communicates that the addressee add it to their to-do list as something they shall do and projects their doing it into the mutually assumed to-do list. Portner (2018) formalizes the effect of a rising imperative, I⇑, in context C as C+ I⇑ = C’, where:

In other words, a rising imperative updates the addressee’s to-do list with something for the addressee to do and projects that this updated to-do list will become common ground.

With this in hand we can lay out our final account of the dynamic pragmatics of aesthetic discourse. The account, in brief, is this: An aesthetic claim adds a speaker-relative proposition to the set of discourse commitments; updates the addressee’s to-do list; puts an issue on the table concerning an addressee-relative proposition; and projects a common ground update of an addressee-relative proposition and the addressee’s to-do list:

The canonical discourse function of an aesthetic sentence, S, is to update the speaker’s discourse commitments with their acceptance of ⟦S⟧; update the addressee’s to-do list with the aesthetic imperative Iae; raise the issue as to whether there is an aesthetic sentence S¹ whose content is judged true by the addressee; and to project a common ground update to that effect, along with an updated to-do list for the addressee.

If the addressee follows the directive, then they introduce a new aesthetic sentence S¹ and the dynamics repeat: the original speaker becomes the addressee with an updated to-do list. The conversation is governed by Community and carries on as long as the interlocutors vibe—that is, until they are estranged, or so similar that they get bored, or become interested in something else. In this way, aesthetic discourse is like an improvisational game. You do something that you hope gets me into the game. If your invitation is good and I am into it, then I toss out something that I can commit to but that I also hope keeps the game going. If what I say is good enough, then you take it up in turn and the dynamics continue as long as we are vibing. We are, in a sense, playing.

The effectiveness of invitational illocutionary force in aesthetic discourse shines here, and we can appreciate this vividly now. Consider a speaker S saying ‘o is beautiful’ to an addressee A. If Community governs aesthetic discourse, then understanding that an aesthetic sentence has been uttered should trigger the norm (if it is not in play already). If S and A were in a position to believe that they would value their joint aesthetic engagement with o, then they would immediately begin to achieve what they presuppose they ought to achieve in speaking that way—they would be on their way to vibing. The felicity conditions of invitation are designed to put S and A in exactly that position. Recall that speaker S’s invitation to addressee A to φ is felicitous only if:

If S’s saying ‘o is beautiful’ were also a felicitous invitation to A to aesthetically engage with o, then S would thereby both issue a weak imperative to engage with o and communicate their belief that either S would value A’s engaging with o or A would (or both). In other words, S communicates their belief that A’s (or their joint) aesthetic engagement with o would be good. A’s taking up the invitation effectively affirms the worth of A’s aesthetically engaging with o in light of S’s claim about it, affirming S’s belief that either S would value A’s engaging with o or A would (or both). This amounts to an affirmation of Community and the beginning of its realization, viz. the formation of aesthetic community. The illocutionary force of invitation thus plugs directly into the dynamic force of aesthetic discourse. But we can also appreciate that Aesthetic Discourse is not absolutely wedded to invitational illocutionary force—invitation itself does not feature in the dynamic pragmatics. Thus if other forces in various contexts serve Community, then they are fair game. The only change might be a slight shift in the dynamics from a weak to a strong imperative.

The account developed here assumes that there is a connection between speakers’ understanding of aesthetic sentences and their presupposing Community. What could the connection be? The answer requires more care than we have space for here, but I will sketch a thought: Aesthetic sentences deploy aesthetic predicates, which pick out aesthetic values. To answer the question, then, we need a theory of aesthetic value. If the good of aesthetic community were at least partly constitutive of aesthetic value, then to understand that aesthetic value is at issue would be to understand that aesthetic community is in the offing. There could then be a direct connection, through speakers’ understanding of aesthetic value, between understanding aesthetic sentences and presupposing Community. The next step, then, is a communitarian theory of aesthetic value.²⁸

Community supersedes Convergence, and while there is an open-ended ‘directive core’ of aesthetic discourse, invitation holds pride of place. Engaging in some normative discursive practices is worthwhile even when we know we will not converge, and so a general convergence norm does not govern all of our discursive practices. Metanormative theory is thus not as tidy as many seem to suppose, and aesthetic discourse is a counterexample to convergence theories of communication.

One consequence of the view presented here is to sever the tie between aesthetic utterances and a common way of thinking about assertion, according to which “an assertion is defined as committing the speaker to the belief in its propositional content, while simultaneously attempting to get the addressee to believe that content as well” (Farkas & Bruce 2010: 92). While Aesthetic Discourse does represent the speaker as committing to a proposition, it does not represent an attempt to get the addressee to believe the proposition. This theory of aesthetic discourse does not exclude this as a goal of an aesthetic conversation, and Community welcomes it insofar as agreement promotes vibing, but it is not a feature of the dynamic force of aesthetic discourse. Securing agreement can be an optional goal of aesthetic conversation.

So perhaps aesthetic utterances are not assertions. Some might take this to show that aesthetic discourse is impoverished, or lesser than fully assertoric discourse. But perhaps, instead, it reveals the mistake of modeling so much of human communication on assertion.²⁹ Built into the idea of assertion is not merely the thought that I commit myself to something, but also the idea that I am trying to get you to commit to the same thing. This second property of assertion befits discussion of moral or purely factual matters because we need to share plans and representations. But sharing these things is not intrinsically good. We need to coordinate to act well and get what we want or need. We converge on pictures and plans to avoid danger, put food on the table, keep the kids alive, build the house. But what is the instrumental value of agreeing about the aesthetic? Agreement might help if we need to design a living space, build a bridge, or put on a show. But what if we don’t? Why should it matter that we have the same aesthetic beliefs or value the same aesthetic items? Aesthetic agreement has no general instrumental value. But therein lies its beauty: Aesthetic conversation is intrinsically good exactly when it is vibing, and vibing just is the point of it. Aesthetic discourse is an end in itself.