1 Introduction

YouFootnote 1 left the kitchen door open and my cat ran out. Needless to say, while walking through the rain trying to find the poor thing, I’m not entertaining exceptionally happy thoughts about you. Do I blame you for what happened? Of course I do! According to the view I will develop here, when I blame you for having left the kitchen door open, I am actively doing something to you, in response to your negligence.

This paper develops, and investigates the credentials of, a novel account of the nature of blame: on this account, blame is a species of performance (henceforth Blame as Performance, or BaP for short). The rationale for the claim that blame is an actionFootnote 2 is speech-act theoretic: it relies on Austinian claims about the nature of performatives and the parallelism between mental and spoken blame (#2). After developing the view, I move on to outlining its theoretical virtues: first, I appeal to general performance normativity to advance a novel view of the normativity of blaming, i.e. what it is to have standing to blame (#3). Second, I give an overview of the difficulties traditional analyses of blame encounter in trying to dismantle blame and show that BaP neatly explains both the intuitions behind, and the difficulties encountered by traditionalism (#4). Third, I show how BaP is better positioned to account for blame’s function than functionalist analyses of blame (#5). Last, I conclude by showing that BaP also nicely accounts for the psychology of blame (#6).

2 Blame as performance

2.1 Methodology

Traditionalism about blame has difficulties coming up with an analysis of blame. The traditionalist model is a dismantling model of analysis: it tries to understand this complex phenomenon by breaking it down into simpler elements that are independently understood. Cognitivist dismantling proposalsFootnote 3 take blaming to be tantamount to judging blameworthy; they are thought by many to fail to capture the affectivity of blaming.Footnote 4 Emotivist accountsFootnote 5 seem to go too far in the opposite direction, and lose blame’s normative diversity in the process: emotional responses do not seem appropriate for some types of blameworthy wrongdoings, such as epistemic, prudential or minor moral wrongdoings.Footnote 6 Conative, desire-based accounts are thought to inherit the ups and the downs of both cognitivism and emotivism.Footnote 7

More recently, blame functionalism (e.g. Fricker 2016; Duff 1986; Sliwa 2021; Vargas 2013) rejects traditionalist methodology: the claim is that one plausible explanation of these historical difficulties lies with the fact that blame is just too complex a phenomenon to afford an (interesting) dismantling analysis. As such, several recent proposals in the literature try to shed new light on the nature of blame by looking at the function of what is taken to be the paradigmatic case of blame: communicative blame. This way of proceeding leaves open the particular attitude or combination of attitudes that constitute blame. In this way, functional accounts can be more flexible. These philosophers advance views about the communicative function of the practice of blame and how this serves to inform us about its nature. One notable worry for this move relates to the fact that attempts to offer a functional analysis of blame commonly start with the communicative function of the practice of blaming. However, it is argued, many (perhaps most) instances of blame are not expressed or communicated. In what sense are those instances of blame communicative? And if they are not, how can blame be essentially communicative in its nature?Footnote 8

This paper investigates a novel lieu in the logical space: it combines functionalist and traditionalist methodology to arrive at a novel account of the nature of blame: Blame as Performance. Like the functionalist, I abandon the ambition to offer a dismantling analysis of blame, and I begin by looking at the paradigmatic case of blame: communicative, spoken blame. Unlike the functionalist, and like the traditionalist, however, my interest lies with the nature rather than the function of communicative blame. That is because, as I will argue, the nature of communicative blame is informative with regards to the nature of blame simpliciter.

2.2 The speech-act theoretic rationale for the performance claim

This section argues that, conditional upon Austinian speech act theory being broadly right about the nature of the speech act of blaming, the phenomenon under discussion—blame—pertains to the category of performances.

To this effect, it will be useful to first have a look at the nature of the speech act of blaming and its relationship to mental blame. To put my cards right on the table, in line with a significant number of philosophers working on the nature of blame (e.g. Fricker 2016; Duff 1986; Sliwa 2021) I would like to suggest that spoken blame is the outer counterpart of inner blame, just like, for instance, assertion is (widely taken to be) the outer counterpart of judgment (e.g. Douven 2006; Williamson 2000). The thought is roughly that, since assertion is our characteristic way of expressing judgments, the one is the inner counterpart of the other. What I want to suggest is that the same motivation will do the needed work for blaming: plausibly, since spoken blame is our characteristic way of expressing mental blame, the one is the outer counterpart of the other. To be clear, the claim here is not that, in either case, the mental and the spoken need to coincide: I can assert that p without judging that p, and the other way around. Similarly, I can blame you in speech but not in thought, and I can mentally blame you without putting it in speech. The claim is rather one of metaphysical parallelism. Here it isFootnote 9:

The Mental Blame-Spoken Blame Parallel (MBSB): Mental blame stands to spoken blame like the inner to the outer.

What we have on this picture, then, it’s a type—blame simpliciter—of which mental and spoken blame are the subvocalized and, respectively vocalized species. Note, further, that if MBSB holds, the nature of outer blame will shed light of the nature of mental blame, since one is the subvocalized incarnation of the other (and the other way around).Footnote 10 Furthermore, the nature of blame simpliciter will be easy to derive by merely removing the species-specific properties—the ‘mental’ and the ‘spoken’. In a nutshell, if we know what kind of entity outer blame is, via the assumption of metaphysical parallelism, we can come to know what kind of entity mental blame is as well, and indeed, what kind of entity blame simpliciter is.

Given all this, what I will do next is try to shed light on the nature of outer blaming, which will, in turn, inform what I will take to be the nature of mental blame and blame simpliciter.

In order to do that, it will be useful to start by distinguishing between the speech act of blaming and the speech act of explicitly blaming. One can blame another by merely saying something along the lines of: ‘You should not have left the kitchen door open, that was really not very considerate of you! It’s your fault that my poor cat ran out!’. Alternatively, one can also blame another explicitly: ‘I blame you for what happened to my cat!’.

Crucially for my purposes, the speech act of explicitly blaming is notably taken by Austin (1955) to be a performative speech act. Performative speech acts have double direction of fit: When I utter ‘I blame you for being late for lunch!,’ I perform two actions: I describe the world (word to world fit), in that I assert that I blame you for being late for lunch; and I change the world (world to word fit), in that, in making this speech act, I thereby blame you for being late for lunch (now you are blamed). Classical examples of performatives include promising, marrying, ordering and baptizing. When I utter ‘I promise to be there,’ I describe the world and change it at the same time: I will thereby have promised to be there (now I have an obligation to be there), and asserted the fact that I promise to be there. Similarly, when (granted that I have the relevant aut1989hority) I utter ‘I hereby declare you man and wife,’ I will have thereby changed the world—now you are married—and described it—by asserting that I am marrying you (Searle 1989). In Anscombe’s famous comparison, the shopper tries to fit his purchases to his shopping list, while the detective shadowing him tries to fit her list to his purchases (1957: p. 56). Performatives do both.

How do performatives achieve this double direction of fit? Well, as the name suggests, they are speech acts that perform: they are acts that act. Performative utterances are performances of the act named by the performative verb (Searle 1989). In uttering those sentences, one asserts that one is at the same time promising, or christening, or declaring war, and thereby additionally does these things (Ginet 1979; Dummet 1993). There are two actions involved in a performative: the assertion and the action that it performs: the assertive speech act of marrying (describing the worldly fact that I am marrying you) and the marrying itself (the world-changing intervention), the assertive speech act of promising (describing the worldly fact that I am promising) and the promising itself.

If Austin is right, and the speech act of explicitly blaming is a performative, then, it needs be the case, like with all performatives, that there are two actions involved in the speech act of explicit blaming: the assertive speech act (describing the worldly fact that I blame you) and the world-changing act it perform—that of blaming. When I utter ‘I blame you!,’ I both describe what I’m doing—the fact that I blame you—, and doing it: I thereby blame you. There are two acts involved in this speech act, the assertion and the blaming itself. Crucially, if that is the case, blame is an act, a performance. When I blame you for being late for lunch, I am not only performing a speech act, I am actively doing something to you. I am not just having an attitude towards your wrongdoing/making a judgment about it, I’m actively doing something to you in response to it. Just like in the case of promising, marrying or baptizing, if explicit blaming is a performative, then blame is a species of action.

What does this tell us about the nature of mental blame? Recall that, drawing on an analogy with judgment and assertion, I took mental blame to stand to spoken blame like inner to outer. What this amounts to is a metaphysical parallelism claim: mental blame is the same kind of entity as outer blame, only located in a different environment. If that is the case, and if outer blame is an action, mental blame will follow suit: it will be the sub-vocalized counterpart of outer blaming, i.e. a mental act. Again, importantly, inner and outer blame can be present independently of each other. I can blame you in speech but not in my mind, and I can also blame you in my mind but stay silent about it.

Finally, we have seen that, in virtue MBSB, the nature of blame simpliciter will be easy to derive by merely removing the species-specific properties of mental and spoken blame. Thus, again, conditional upon Austin’s view about the speech act of blaming being right, in line with marrying, promising, ordering etc., blame is a performance that can take inner (mental) and outer (spoken) shape.

What kind of performance is blame? In line with Miranda Fircker (2016), I want to say it is a sui generis phenomenon—in my view, a sui generis performance—, that does not admit of a full dismantling analysis. As such, when I say that blame is a performance, I cannot further explain what it doesFootnote 11: that’s just what it is for it to be a sui generis performance.Footnote 12 That is not to say, however, that we cannot say further interesting things about its nature.Footnote 13 What can we say further about the nature of blame as performance? Note that blaming is a three-place predicate: when I blame, I blame someone for something. This suggests that blame constitutively aims at trackingFootnote 14 some normative failure on the part of the blamed. To see this, note that blaming is essentially criticisable (i.e. all instances of blaming are thus criticisable) if it fails to achieve this aim, i.e. if it either fails to hit upon a normative failure, or it fails to target its proper source. Essential criticisability is strong evidence of norm and/or aim constitutivity. Of course, actual wrongdoing need not be present: we are not infallible blamers, I can mistakenly, wrongly blame you. But I cannot blame for nothing, nor can I blame you for doing something perceived to be good: I cannot blame you for saving the world, it is conceptually weird.

On a first approximation, then, blame constitutively aims at tracking normative failure, at tracking norm breach. Aim-constituted activities can occur even when they fail to achieve their constitutive aim. What makes the aim constitutive as opposed to merely conventional is the fact that, necessarily, all occurrences of the aim-constituted activity in question fall short when they fail to reach the corresponding aim. When I mistakenly or wrongly blame you, my blaming will fall short in not having achieved its constitutive aim.

There is more to this, however: wrongdoing is not enough for blame. Blame is about the wrongdoer not the wrongdoing. When I blame you for being late for lunch, I don’t merely criticize your action, I also hold you responsible for it, or take this wrongdoing to be attributable to you. Again, blaming is a three-place predicate: I blame you for something. Constitutively, then, blame is a second-personal type of performance: when I blame, I blame someone (or something). In virtue of being a second-personal type of performance, blame is directed at the wrongdoer, not merely at the wrongdoing.

In sum, then, here is the account I would like to put forth:

Blame as Performance (BAP): Blame is a performance constitutively aimed at a wrong-doer for a wrong-doing.

On this view, in virtue of being a performance essentially targeting a wrong-doer, it will be a person-affecting negative-response performance, or retributive practice, analogous with actions like punishment or downgrading.

Importantly, second-personal need not imply other-personal: I can blame myself. Also, second-personal need not imply any actual encounter with the blamed party. I can blame the dead and the distant. When I blame Hitler for the Holocaust, I do something to him even though he’s dead. The same goes when I blame people I have never met, who live on another continent, whose bad acts I've heard about in the news. To see this, note that ‘my doing something to someone S’ need not imply that S is present in my near proximity, that they’re alive, or that I’ve ever met them. I’m doing something (good) to Shakespeare when I put on a festival in his honour; after all, I’m doing something good to his reputation, memory etc. Similarly, I’m doing something (bad) to Celine Dion when I discourage my daughter from listening to her albums, even though I’ve never met Celine Dion.

Before closing the part of the paper that develops the account, I would like to consider the following question: This section has been framed as a conditional throughout: if Austin is right, then blame is an action. What happens if Austin was wrong about the speech act of blaming, though? A few things about this: First, to date, to my knowledge, Austin’s view on the issue has remained unchallenged. I take this to be at least some evidence that not many philosophers thought it outrageously false. Second, note that I don’t need a full blown Austinian view on the speech act of blaming in order to have support for my account: all I need is the milder claim that explicit blaming can, at times, be a performative.Footnote 15 Since that can only be the case if blaming itself is an action, the view is vindicated. Compatibly, it may well be that blaming waivers between descriptive and performative usages. The latter claim—the claim that blaming can be a performative is fairly uncontroversial to my knowledge. After all, it survives the famous ‘hereby test’ for performativesFootnote 16: one can sensically utter ‘I hereby blame you’ in the same way in which one can sensically utter ‘I hereby marry you.’ The latter does not work on pure descriptives: ‘I hereby have blue eyes’ makes no sense at all.

Last and most importantly, I take the hypothesis that blame is an action—rather than a judgment, emotion and so on—to be worth exploring in its own right, independently of the speech-act-theoretic motivation. To see that this is so, it would be useful to read the following sections that outline the many theoretical virtues of the view.

3 What BaP explains: the normativity of blaming

The account of blame proposed in the previous section is independently motivated, based on two very minimal, fairly uncontroversial theses: the inner blame/outer blame parallel and the (at least at times) performative nature of the speech act of blaming. If these two theses hold, the account I have offered above follows.

The next sections will offer yet several other reasons to prefer my account over the competition, all pertaining to BaP’s theoretical fruitfulness. This section focuses on how the account explains the normativity of blaming.

I have proposed that blame is a sui generis species of performance. When I blame you for being late for lunch, I am actively doing something to you: I am engaging in a you-affecting negative-response performance constitutively aimed at a you as a wrong-doer for a wrong-doing of yours. If all this is the case, we are now in a position to develop an account of the normativity of blaming: it will be an instance of performance normativity. One important theoretical upside of this account, thus, is that it can rely on the general normativity of performances in order to explain the normativity of blame. The normativity of performances has been thoroughly theorized by competence-theorists in the last decades.

The most widely discussed articulation of performance normativity in the literature is Ernest Sosa’s AAA–model of performance assessment (Sosa 2007: pp. 22–23).Footnote 17 On this approach, we can assess performances for accuracy, adroitness, and aptness. Accurate performances achieve their aim, adroit performances manifest competence, and apt performances are accurate in virtue of being adroit: their success manifests competence. This AAA-model applies to all conduct and performances with an aim, whether intentional (as in ballet) or unintentional (as with a heartbeat). To see the motivation behind the model, take the case of archery: the success condition in archery is hitting the target. Success can happen by accident: I can hit the target even though I have not shot an arrow before, just by dumb luck. If that happens, my shot will be an accurate shot.

In the case of archery shots by professional archers, though, competence is involved: professional archers have a reliable disposition to hit the target in normal conditions (good lighting, friendly weather etc.). When a competent archer successfully hits the target, the shot is both accurate and adroit. This is not all there is to performance normativity, though: to see this, consider a case in which a professional archer competently releases the arrow, which then heads straight towards to bulls eye; alas, a gust of wind takes it off course. Luckily, a second gust of wind sets the arrow back on its track, and the arrow hits the bull’s eye successfully. In this case, the performance is both accurate and adroit, but its accurateness fails to manifest its adroitness: it is accurate by luck, in spite of the fact that it is competent. Contrast this scenario with one where all goes well: the competent archer releases the arrow, and, in virtue of the manifestation of her competence, the arrow hits the bull’s eye. According to Sosa, when all goes well, performances are not merely accurate and adroit, but also apt: they are accurate in virtue of being adroit.

Let us go back to Blame as Performance. I want to argue that Sosa’s AAA structure will give us an account of apt blaming that will be instrumental in accounting for what the literature dubs ‘standing to blame.’ Many philosophers think that blame is a normatively appropriate response only when the blamer has standing to blame. In turn, one widely endorsed condition for having standing to blame is the non-hypocritical conditionFootnote 18: X has standing to blame only if X has not herself engaged in similar wrongdoing in the past.Footnote 19 While the vast majority of blame theorists agree that hypocrisy negatively affects one’s capacity for permissible blaming—indeed, non-hypocrisy is taken by most to be the paradigmatic example of a condition for one’s standing to blame—it has been notoriously hard to satisfactorily account for why this might be so. Most crucially, it has been notoriously hard to offer an account of standing to blame that both naturally drops out of the particular account of blame proposed, and, at the same time, is general enough to accommodate plausible conditions on standing to blame in a non-ad hoc manner (see Tognazzini and Coates 2018 for an excellent overview). In a nutshell, the main problem most accounts on the market face is the following: accounting for the badness of hypocritical blaming is easy: after all, plausibly, it lies in the nature of hypocrisy that, for most phi, hypocritically phi-ing is bad. It is easy then, whatever one’s account of the nature of blame, to predict that hypocritically blaming is a bad thing to do. The issue, however, is explaining why, in the case of blame in particular, hypocrisy is not just bad, but it also affects one’s standing in the way in which it does. After all, this does not seem to generalize in an equally neat way: hypocritically laughing does not affect one’s standing to laugh, hypocritically promising does not affect one’s standing to promise, and so on.

In what follows, I will show how building the normativity of blame on general performance normativity 1. explains the necessity of standing to blame for fully appropriate blaming, and, furthermore, 2. Explains the non-hypocritical condition on standing to blame. If successful, Blame as Performance scores important points on normative fruitfulness.

Just like archery, blaming is a performance with a success condition: blaming is successful when it hits its proper target. I have argued above that blame is a second-person-affecting negative-response performance constitutively aimed at a wrong-doer for a wrong-doing. If this is so, successful blaming will target the blameworthy: when I blame you for being late for lunch, my blaming performance will be successful if and only if you are blameworthy for being late for lunch.

Now, importantly, even if my blaming successfully picks out the blameworthy, it can fail to manifest the relevant competence—i.e., the type of competence corresponding to the type of blame at stake (moral competence for moral blame, epistemic competence for epistemic blame etc.). There are two ways in which this could happen: either I lack the relevant competence (no adroitness), or I have it but my successful blaming fails to manifest it (no aptness: although I am competent, on this particular occasion I only get it right by luck). Apt blame, then, will consist in blaming the blameworthy through the manifestation of the relevant type of blaming competence.

What might the relevant competence be? At a first stab, I want to suggest what is at stake is the normative competence of the relevant type, i.e. a reliable disposition to do well, normatively, for the type o normativity at stake, in normal conditions. To see the rationale behind this, note again that blaming is for wrong-doing, and wrong-doing comes in broad normative types: moral, epistemic, prudential etc. I am not competent to blame for moral wrong-doing if I am not morally competent, as I am not competent to blame for epistemic wrong-doings if I am not epistemically competent. If I am one who is disposed to regularly perform morally bad actions, I am not morally competent, and thereby I am not competent to blame for moral wrong-doings. Similarly, if I am one who is disposed to regularly form beliefs based on astrological readings, I am not epistemically competent, and thereby not competent to blame for epistemic wrong-doings.

Note, though, that one can be generally morally/epistemically/prudentially etc. competent, but not be competent with a particular, more narrowly individuated moral/epistemic/prudential etc. matter. Conversely, one might be generally incompetent with moral/epistemic/prudential etc. normative matters, but do extremely well on a particular, narrowly individuated moral/epistemic/prudential matter. I might be always sloppily late for lunch, although I’m otherwise an extremely polite and dependable person. Or, to the contrary, I might be very punctual, but fail miserably in all other social-normative affairs. When this happens, we will want to say that what matters is the narrowly-individuated competence: I am competent when it comes to punctuality, and therefore I can competently identify instances of wrong-doing in this respect, and aptly blame the wrong-doer.

It turns out, then, that what maters for blame is competence with regard to compliance with a particular norm: only if I am competent vis-à-vis punctuality, can I aptly blame you for being late for lunch. Only if I am competent when it comes to only forming beliefs via reliable processes, can I aptly blame you for trusting astrological readings. Only if I don’t myself binge on fashion shopping, am I competent to blame you for buying your sixth pair of green shoes.

This account nicely explains why standing to blame is necessary for apt blame: apt blame requires that a competence condition on the part of the blamer be met and manifest. Blame can be successful—if I blame the blameworthy—but inapt, in that my blaming the blameworthy does not occur in virtue of a competence of mine with respect to complying with the norm at stake.

In turn, note that the view also nicely explains the non-hypocrisy condition on standing to blame: competence with the relevant norm implies reliable disposition to comply with the norm in normal conditions. Hypocritical blamers are, in contrast, by stipulation, people who have themselves engaged in similar wrongdoing in the past or have a disposition to do so. Note, too, that in order for standing to blame to be missing, it needs be that the blamer has engaged in similar wrongdoings to quite a significant extent/has a significant such disposition: if I’ve only been late for lunch once in my entire life, I still have standing to blame you for being late; if I am unreliable when it comes to coming on time, I don’t have standing to blame you. The reliability condition on competence delivers this very result: when I am often late for lunch myself, I am not competent with respect to following the norm of punctuality, and, thereby, I cannot aptly blame you for being late for lunch.

4 What BaP explains: the history of the analysis of blame

The next two sub-sections look at what this account of the nature of blame explains when it comes to cognitive, emotive and conative phenomena associated with blaming. Importantly, the ambition of this section is not critical: I am, at least for the most part, not going to aim to significantly advance the state of the art concerned with difficulties for extant accounts of blame. Rather, I will survey the relevant literature to the aim of arguing that Blame as Performance is in a good position to explain both the advantages and the weaknesses of the competition.

4.1 Blame traditionalism and directions of fit

I would like to go back to the history of the analysis of blame project. Several people think blaming is roughly tantamount to judging blameworthy (e.g. Haji 1998; Hieronymi 2004; Scanlon 1986; Smart 1961; Watson 1996; Zimmerman 1988). According to Zimmerman, for instance, when we blame someone, we judge that there is a “discredit” or “debit” in his ledger, that his “moral standing” has been “diminished” (Zimmerman 1988: p. 38). In a similar vein, on Gary Watson’s account, to blame someone is to judge that they have failed with respect to some standard of excellence.

There are, though, two notable worries for cognitive accounts, concerning the necessity and, respectively, the sufficiency direction of analyzing blame in terms of judging blameworthy. If successful, these objections seem to suggest that blaming and judging blameworthy, while they may often go hand in hand, survive perfectly fine without one another. Consider, against the sufficiency direction: I can judge you blameworthy for being late for lunch and still not blame you—after all, I’ve done it myself in the past, and it’s not that big a deal anyway; nobody’s perfect (Kenner 1967; Coates and Tognazzini 2012). Conversely, against the necessity direction, consider irrational blame; classical cases involve benighted moral agents: I see how you could not help but do what you did, you’re blameless in virtue of your social determinations, but I cannot help but blame you for it (Pickard 2013).

The account of Blame as Performance has the theoretical resources to explain this phenomenon: judgments of blameworthiness, while often associated with blame, are metaphysically distinct entities. It is easy to see what went wrong if we go back to the judgment/assertion parallel: the speech-act correspondent for a judgment that you are blameworthy is not the speech act of blaming, but the speech act of asserting: ‘You are blameworthy.’ The latter is distinct from the speech act of blaming: it is not a performative, but an assertive. It merely describes the world, it does not affect it. It only has a word-to-world direction of fit.

Since blame is a second-person retributive performance, though, it will be often associated with judgments of blameworthiness: retribution constitutively aims to match desert. Normally, when I do something negative to you in response to your being a (perceived) wrong-doer, I aim to get it right: I aim to do it in virtue of your actually being a wrong-do-er. This model, however, suffers exceptions, and the classical counterexamples to cognitivism track these exceptions: in particular, cases where I blame irrationally, i.e. in spite of the fact that I don’t judge you blameworthy, will be explained on BaP as garden variety cases of irrational action against one’s best judgement. Also, the classical cases put forth against sufficiency have straightforward explanations on BaP: first, if blame is an action, in line with other actions, it is hardly surprising that I can judge that you are worthy of phi-ing and yet not do phi to you. In particular, this will easily be the case when I lack the authority to do phi: I can judge that you are worth punishment, for instance, but since I lack the authority to punish you, not act on my judgment. In line with this general model, counterexamples to the cognitivist’s sufficiency claim, unsurprisingly, ride on lack of standing to blame.

Let us move on to emotions and desires: The disconnect between blaming and judging blameworthy led some people to believe blame was not a cognitive affair, but rather an emotion or set of emotions.Footnote 20 When we blame, they argue, we feel resentment, indignation, anger and the like for some wrongdoing (Bell 2013; Strawson 1962; Wallace 1994; Wolf 2011).

Worries have been put forth for the necessity direction of emotivist analyses. The problem is that all these emotions, while characteristic to instances of blame for fairly serious breaches of moral norms (i.e. murder), are not plausibly necessarily present in other, less highbrow types of blame. Take, for instance, moral blame for minor wrongdoings: I’m 5 minutes late for lunch. Yes, I did break my promise, so if I did it out of sloppiness, and you care about punctuality, you may end up blaming me. Still, anger, resentment and indignation (no matter how ‘light’) seem like a disproportionate reaction. Consider, also, the case of epistemic blameFootnote 21: if you’re looking straight at the vase but believe there is no vase in the room, all else equal, you are epistemically blameworthy and subject to proper criticism: ‘Can’t you see the vase? You’re looking straight at it!’. However, it seems rather peculiar for me to resent you or feel indignation in relation to your epistemic failure, absent other normative considerations. To the contrary, your epistemic unresponsiveness is rather hilarious. Similarly, take prudential blame: I can surely blame you for wasting your money on your sixth pair of green shoes, while not feeling particularly emotional about it.

Several people in the traditionalist literature have proposed conative analyses of blame. The most influential account of this sort is George Sher’s (2006) conative account. Sher (2006) endorses a “two-tiered” account of moral blame “which takes it to consist of a characteristic set of affective and behavioural dispositions that are organised around a characteristic type of desire-judgment pair” (2006: pp. 14–15). What Sher ads to judgments of wrongness is a backwards-looking desire “that the person in question not have performed his past bad act” (2006: p. 112). On top of this, it should be the case that this judgment-desire pair triggers dispositions to behave in characteristically emotional ways, such as feeling badly about not getting what one wants, publicly expressing the unsatisfied desire etc. Thus, Sher suggests that the negative emotions typical of blame are an example of how one feels badly when one’s desires are frustrated.

Unsurprisingly, this type of view is also bound to inherit the objections to the necessity of judging blameworthy for blaming (nested in the possibility of irrational blame). Furthermore, worries have been expressed that there are cases when the relevant desire that you have not done what you did is absent while blame is present. After all, it may well be that you lie to me, I judge you are blameworthy and blame you for it, but I don’t desire you hadn’t done it because, by mere accident, I end up better off because of your lie. Furthermore, it looks as though in cases of non-moral blame, the relevant desire is usually not present, if not even outright inappropriate: I don’t have any desire concerning your epistemic or your shoe shopping blunders (why would I care?), but it still seems perfectly appropriate to blame you for your irresponsible behavior.

Against the sufficiency direction involved in Sher’s view, Angela Smith argues that in some cases, say in “the reactions of a mother whose son is blameworthy for [a] crime” (Smith 2013: p. 35), the relevant judgment-desire pair might be present without blame. Furthermore, note that, compatibly, Smith’s mother may also display some of the dispositions to behave in characteristically emotional ways, such as feeling miserable about what happened, publicly expressing the fact that she wished the son did not do it etc.

I want to argue that BaP explains what went wrong with both the emotivist and the conativist analyses: the problem here is that there is a mis-match of direction of fit. Emotions, like cognitions, tell us something about how the world is (pleasant, unpleasant—mind-to-world fit) and, like desires, they motivate us to act in various ways to change the world (world-to-mind). First, if blame is an action, it misses the characteristically cognitive mind to world direction of fit, in that it does not describe the world, it only affects it. Further on, when it comes to changing the world, the important difference between emotions and desires on one hand, and blaming on the other concerns the success condition: emotions and desires motivate us to change the world, while blaming successfully does so. In this, the account of blame as performance explains why emotions (anger, resentment) and desires (that the wrong-doer had not done the deed) can come apart from blaming, but often go hand in hand: first, blame does not require valuations of state of the world: this is what the case of irrational blame teaches us. That being said, in line with other retributive actions, it is often associated with—indeed, often triggered by—bad feelings about what happened. Second, while desires and emotions merely motivate one to respond to wrongdoing, blame successfully does so: in many cases, the motivation and the success will come together: I will be motivated to respond to your wrong-doing, and I will also successfully do so. This, of course, though, need not be the case: motivation and successful action can come apart.

4.2 Blame functionalism and the mental/spoken blame parallel

In view of the history of unsuccessful dismantling analyses of blame, more recently, several people (e.g. Bennett 2013; Duff 1986; Fricker 2016; McKenna 2012; Sliwa 2021; Smith 2013; Vargas 2013), propose to shed light on the nature of this phenomenon by looking into the function the practice of blame serves for us. Instead of identifying blame with any particular attitude or combination of attitudes, functional accounts of blame identify blame by its functional role.

What unites all functional accounts of blame in the literature is their focus on the practice of communicating blame. The thought is that identifying the function of our practice of blaming is going to shed light on the nature of blame. How is this supposed to happen? Blame is internally very diverse. So maybe it is a practice essentially defined by its function, which, in turn, is multiply realizable. Take a paradigmatic case of functional item: for instance, keys. They come in many shapes and colors, and it’s hard to find a common denominator that is also in any way informative. What unifies the tokens of the type is rather their function: opening/closing locks. As opposed to keys, bachelors, for instance, have an informative common metaphysical denominator: they are unmarried men. The thought, then must be that blame is more like keys than like bachelors. What is defining of the practice is its function, and the latter can be fulfilled in many ways—blame is internally diverse.

According to Fricker (2016), the function of communicative blame is epistemic: it resides in its power to increase the alignment of the blamer and the wrongdoer’s moral understandings. It aims to bring the wrongdoer to see or fully acknowledge the moral significance of what they have done or failed to do. By communicatively blaming you for leaving the kitchen door open, I aim to make you see that what you did was wrong—you were not very considerate, and now my cat ran out!—and feel remorse for having done it. This aim need not, of course, always be present as an intention in the psychology of the communicative blamer; rather, according to Fricker, the aim is a function of the type of speech act blame is, the nature of its illocutionary point.

This alignment of moral understanding, in turn, has a social function: it is aimed at correcting the future behavior of the wrongdoer. I blame you for leaving the kitchen door open in the hope that you will recognize how this was not very considerate, and be more careful in the future. Finally, Fricker thinks that all other types of blame can be understood as derivative of communicative blame.

Antony Duff has proposed a similar understanding of the aim of blame, according to which it is “an attempt to communicate to the wrong-doer a moral understanding of his wrong-doing; to bring him to recognize his guilt and repent what he has done (Duff 1986: p. 70).

More recently, Sliwa (2021) also proposes a function-based account of blame. On her account, the latter consists in creating common knowledge of reparative duties and claim-rights for wrongdoing.

In sum, then, according to these philosophers, the function of our practice of blaming is mainly epistemic: it aims at effecting changes in the epistemic status of the participants—be it their moral understanding or their knowledge of post-wrongdoing duties and rights.

Other alternative popular functionalist views on blame take the relevant function to be one of expressing protest or disapproval in the face of wrongdoing. According to Michael McKenna, for instance, blame is conversational, and thus functions to continue a conversation started by the blamee’s wrongful action. McKenna claims that all reactive attitudes and their expressions serve this function. According Smith (2013), blaming functions as a form of protest. For Smith, when we modify our attitudes and intentions towards the blame, and this serves a particular function, namely that of protest, it counts as an instance of blame.

There are two main worries for blame functionalism: one is methodological, the other one concerns the details of the accounts under discussion.

I will start with the latter: Criticism and blame are two distinct normative categories; furthermore, the corresponding speech acts can also come apart. I can criticize you (‘What you did was wrong’) but fail to blame you (…although I must say I’ve done it myself in the past, I can sympathise!). Interestingly enough, note that criticizing in speech is weaker than blaming: When I utter ‘I blame you for phi-ing!’ I both blame and criticize you. The converse need not be the case: I can criticize you for phi-ing without blaming you (after all, nobody’s perfect).

Now, note that the functions proposed in the literature for the practice of blaming—epistemic functions, protest functions, communicating disapproval functions—are perfectly well served by the weaker speech act—the speech act of criticizing. That is because, just like blaming, criticizing communicates to you my normative take on things. In a nutshell, by merely criticizing you for having phi-ed, I can: increase the alignment of our moral understandings and your moral knowledge, protest your actions, express my disapproval, without necessarily also blaming you for what you did. If that is the case, though, blaming you—on top of mere criticism—seems like a normative overkill. If the function is perfectly well served by mere criticism, why have a stronger practice to the same effect? Consider, for a visual analogy, asserting in normal tone and shouting. The function of asserting that p is to inform you that p. Furthermore, in normal conditions, assertion is perfectly capable to fulfill its function. It seems, then, unnecessary (not to mention outrageously impolite) to shout that p at you, if what I’m up to is merely informing you that p. More needs to be the case for the stronger speech act to be warranted: maybe I also want to signal an emergency (‘Your train is leaving!!!!’), for instance. Whatever it is, shouting for mere informing is overkilling. Similarly, blaming rather than merely criticizing for generating normative knowledge or understanding, or for communicating disapproval or protesting is normative overkilling.

Criticism and blame can come apart. Furthermore, there are instances in which criticism is appropriate while blame isn’t; i.e., instances in which not only they do come apart, but they should come apart (Kelp and Simion 2017). Actions are often performed in the public sphere and, as such, are observable by others, who may pick up the forms of behaviour exhibited. When you fail to call Uno when playing your penultimate card in a game of Uno, and so violate a rule of the game, this may be observed by someone else who will pick up your behaviour and, as a result, may violate the rule in the future, too. By allowing for criticisms of actions that violate specific norms we can work against the spread of norm-violating forms of behaviour. Since this is a good thing, it makes sense for us to allow for such criticisms. At the same time, we may also want to grant that a norm has been broken blamelessly by the agent. We do not want to hold the norm violation against her: she was blamelessly ignorant, things were blamelessly out of control and so on. If so, there is excellent reason for us to allow criticisability relative to a specific norm and blamelessness relative to the very same norm to coexist.

In sum: criticism and blame can come apart and there are instances in which they should come apart; also, the communicative function at stake is fulfilled by mere criticism, which is the weaker speech act. If this is so, looking at the communicative function will shed little light on the nature of blame: after all, similarly, we are not going to understand the nature of shouting by looking at the function of informing.

But couldn’t someone like Fricker reply that mere criticism, while reaching the aim of bringing about alignment in moral understanding, it does so via the wrong means?Footnote 22 After all, according to Fricker’s view, the communicative function of blame is to bring about alignment in moral understandings via inspriring remorse in the blamed party. The thought then would be that what makes blame so effective in particular contexts is precisely this way of bringing about alignment in moral understanding. While criticism (of someone's conduct) may well serve the function of bringing about alignment in moral understandings, maybe it cannot reliably serve the function of bringing about said alignment via inspiring remorse.

I don’t think this reply will work well, for three main reasons: first, building the means to alignment in moral understanding in the content of the function leaves the account poorly motivated: recall that, according to Fricker herself, the final point of all this is to correct the future behavior of the wrongdoer. To this end, an alignment in moral understanding seems sufficient, no matter how it may be brought about—be it via remorse or via, say, brainwashing. Second, considerations pertaining to the effectiveness of one means over the other are inconsequential, insofar as the function is fulfilled often enough (this, indeed, is the case with functions in general); in turn, whether criticism fulfills the function often enough is an empirical question. One thing that seems plausible, though,—which brings me to my third point—is that the difference between the effectiveness of the two—blame and criticism—cannot be very large, since the two speech acts are most often indistinguishable—with only one notable but infrequently encountered exception: explicit blaming.

Maybe, then, we can keep the functionalist analysis proposed in its broad shape, but abandon the details—i.e. the particular communicative functions identified? One initially promising way to do this, for instance, might be to identify a second-personal function of communicative blame: rather than targeting the normative properties of the target action (wrongness etc., which, if I am right, is the proper function of criticism), on this view blame targets the normative properties of the blamee (fault, responsibility for wrong action). In this, blame’s function will be distinct from that of criticism, and therefore better identify the phenomenon we are interested in.Footnote 23

This brings us to the methodological worry: focusing on the communicative nature of an important part of the practice may easily be misleading. Undeniably, we sometimes (often!) only blame people in our heads. Furthermore, we sometimes blame them in speech while mental blame is absent. If that is the case, by looking at the function of our practice of blaming at large, we risk identifying a function that is only associated with communicating blame, rather than blaming. To see the worry further, consider the parallel case of judgment and assertion. Just like mental and spoken blame, the one is the inner counterpart of the other. Now, plausibly, the (main) function of assertion is communicative: generating a particular epistemic standing in the hearer (knowledge, true belief etc.). This, however, will not yet tell us anything interesting about the nature of judgment, in virtue of the fact that judgment does not share this communicative function. To the contrary, this particular communicative characteristic is what sets judgment and assertion apart.

What can do the trick, however, in shedding light on the nature of judgment, is an account of the nature of assertion: if assertion stands to judgment like outer to inner, judgment is sub-vocalised assertion. Take a X view of the nature of assertion, then, and you have a view on the nature of judgment: subvocalized X. That is precisely the framework that BaP relies on for blame: Bap stats with the nature, not the function, of vocalized blaming, and derives the nature of mental blame accordingly.

Furthermore, BaP can also easily make sense of the intuitions behind the several functionalist views we have been looking at: after all, if I am right and blaming is a second-personal retributive performance, it is hardly surprising that it will, in line with other retributive performances like punishment and downgrading, often serve several communicative functions: to communicate that the wrong-doer is being held responsible, that the normative landscape is changed in the light of the wrongdoer’s wrongdoing, etc. Also, in virtue of being a second-person performance, like punishing and downgrading, it will be an efficient tool in the service of protest and expressing disapproval against the wrong-doer. Importantly, though, all the relevant functions will be targeting the wrong-doers rather than the wrongdoing: blame, as opposed to criticism, is always about the doer not the doing. In this, BaP accounts for the fact that criticism can be appropriate when blame is not: cases of blameless norm-violation are cases in point.

5 Concluding remarks

I have developed a view according to which blame is a species of performance. Crucially, the argument for the view relied on a minimal number of widely accepted claims pertaining to the nature of the speech act of explicit blaming and the metaphysical parallelism between mental and spoken blame. In this, the account proposed comes with a high score for prior plausibility. Furthermore, I have argued that BaP is theoretically fruitful in many further ways: first, it gives us a straightforward way to understand the normativity of blaming. Second, it neatly explains both the advantages and weaknesses of the most prominent competing accounts.

To conclude, I want to briefly suggest that the view is also empirically adequate, in that it explains several psychological phenomena often associated with blaming. A few central cases discussed in the literature include:

  1. (1)

    The force of blame: it seems that, when we are blamed, we tend to not remain indifferent; furthermore, several people think that blame has the power to affect the normative landscape (in the same way as e.g. promising), by creating new obligations/permissions (e.g. Sliwa 2021). The view proposed here nicely explains the force of blame, i.e. why it is not easy to disregard: after all, on this account, when I blame you, I am actively doing something to you, I am not merely judging or feeling something about you. Other’s retributive actions against us are less easy to ignore than their judgments and feelings about us. If this is not obvious yet, consider the difference in force between: my judging/feeling that you deserve punishment and my actually punishing you. Furthermore, BaP also (arguably, uniquely) neatly explains the widely spread intuition that, after blaming, the normative landscape is crucially affected: emotions, judgments and desires do not affect the common normative ground, in virtue of either not having the right direction of fit (judgments) or not entailing a success condition (emotions, desires); actions do: they change the world successfully.

  2. (2)

    Blame’s internal metaphysical diversity: we can blame by explicitly uttering ‘I blame you for being late for lunch!,’, but we can also blame by merely saying something like ‘You should not have come so late to lunch, it was not nice of you to make me wait!;’ Blame as Performance also makes easy sense of this phenomenon: recall the distinction form Searle, between performatives and performances. I can blame you by uttering ‘I blame you!,’ but I can also blame you by merely saying something like ‘That is a terrible thing to do!.’ When I utter ‘I blame you!,’ my speech act is a performative, with a double direction of fit. When I utter ‘That is a terrible thing to do!,’ I blame you without, at the same time, describing what I am doing; my speech act is a performance, with a world-to-word direction of fit.

  3. (3)

    The normative diversity of blame (we can blame people for moral, prudential, epistemic etc. failures). Blame as Performance has no trouble accommodating different kinds of blame: moral, prudential, epistemic etc. Recall the model I have proposed: blame is a second-person-affecting negative-response performance constitutively aimed at a wrong-doer for a wrong-doing. The relevant wrongdoing can be moral, prudential, epistemic or of any other normative kind.