1. The Question

There is a nest of interesting issues concerning our engagement with artworks produced by artists of whose behaviour or character we morally disapprove. One is whether we should engage with their work at all. Another is whether a properly artistic appreciation of works produced by such artists requires us to put our moral evaluation of their behaviour or character to one side. A third is whether doing this is so much as possible. What I have to say might have implications for how we want to answer one or more of these questions, but my concern lies elsewhere.

Expressed somewhat abstractly and loosely, at least to begin with, my question is this. Can the moral failings of artists make their artworks artistically worse? That's to say, can moral defects in an artist's character, or morally reprehensible behaviour on their part, make their works worse, qua artworks, than they would have been otherwise? As this question is currently formulated, the answer would seem to be: yes, obviously. If a comedian's racism brings him to overload a comedy script with racist jokes, then his racism makes the script less funny than it would have been, had his racism not made its way into the text, or had he not been racist at all;Footnote ¹ and since unfunniness is an artistic flaw in a comedy, the said flaw is certainly an artistic one. But this question – in essence, whether an artist's moral shortcomings can make for artistic failings in his works by virtue of being causally responsible for at least some of these works’ manifest, perceptible properties – is not the issue I'm interested in either.

So what is the question I want to address? This. Can an artwork W produced by a morally flawed artist A be artistically worse than it would have been otherwise, merely by dint of his having these moral flaws? To put it another way, can A's moral defects diminish W's artistic success without making a difference to its discernible, manifest properties? Can the mere fact that A has these moral flaws make W a worse artwork, while everything else that could be relevant remains the same? Some philosophers, although they do not quite formulate my question in the way in which I have, seem to occupy a position that entails a positive answer to it. I am very dubious about this, and in this article I want to think through an example in such a way as to evince and explicate my scepticism.

2. Manhattan and Methodology

My example is Woody Allen's 1979 film, Manhattan (1979), whose central relationship is a romantic, sexual one between a 42-year-old, neurotic, aspiring male novelist (Isaac, played by Allen) and, at the start of their relationship, a 17-year-old girl (Tracy, played by Mariel Hemingway). Allen has been accused of having a morally questionable sexual interest in young women: an interest presenting ‘a morass of moral concerns about exploitation and abuse of power’ (Matthes, Reference Matthes2021, p. 21). And critics have pointed to his commencement of a sexual relationship with Soon-Yi Previn – his then partner, Mia Farrow's, adoptive daughter – in December, 1991, when she was 21 and he was 56, as evidence of this. I avoid taking a moral stand on Allen's character or his behaviour in his relationship with Previn. I am just interested in whether, if Allen is morally flawed in the ways alleged, the artistic success of Manhattan is compromised merely by dint of this being so.

Manhattan is a comedy, and so its artistic success depends largely on its being found funny. So my question becomes this: is the film ‘less funny’, and so less artistically successful, merely by dint of Allen's putative moral flaws (Matthes, Reference Matthes2021, p. 12)? As I have noted already, Allen's (alleged) immoral behaviour took place some years after he finished Manhattan, and it is this feature of the case that in fact makes it a useful case study, allowing us to compare our respective interpretations of the film before and after the story of his relationship with Previn broke, with the film's manifest properties remaining constant throughout. If we are now justified in thinking that the film is less funny than it once was, then this would suggest that Allen's (presumed) moral failings alone have made a difference to the film's artistic merit.

An alternative method of considering whether the film is rendered less funny by mere dint of Allen's moral status would be that of ‘spinning the possible worlds’ (Blackburn, Reference Blackburn1984, p. 312): in this case, by trying to imagine a possible world exactly similar to the actual world, including Manhattan's having exactly the same manifest properties, except that Allen lacks the moral flaws he (allegedly) has actually. If such a possible world exists, and if the film is funnier in this possible world than it is actually, then this would show that Manhattan is, indeed, made less funny by mere dint of Allen's misdeeds.

A drawback with this latter strategy, however, is that such modal thought experiments can be difficult to assess. The method is essentially that of constructing a model of a putatively possible world (in this case, one in which everything remains constant except Allen's moral character or behaviour), and then asking a ‘what if’ question of this world (in this case, whether we would justifiably find Manhattan funnier than we do actually) (Cooper, Reference Cooper2005, p. 336). But we might wonder, first, whether the fact that we can conceive of a world really shows that this world is a genuinely possible one (Wilkes, Reference Wilkes1988, p. 17). And then, perhaps more pertinently for our present concern, we might ask how we could know whether we have answered the relevant ‘what if’ question correctly (Cooper, Reference Cooper2005, p. 342). The construction of the said possible world, and our answer to the question of whether we would find Manhattan funnier in this world, is supposed to constitute data with which to inform our prosecution of the question of whether the film has been made less funny by mere dint of Allen's moral flaws. But it is difficult to counteract the suspicion that our views on whether we would find the film funnier in this world are little more than speculation, and, quite possibly, theoretically driven speculation at that.

So perhaps it is better to stick to the actual world and consider whether we are justified in finding Manhattan less funny than we once did, now that we are aware of the commencement of Allen's relationship with Previn in 1991. At least we have people's actual responses to the film, before and after Allen's relationship with Previn hit the news, and not merely imagined responses, to go on. And this, to be sure, is how the issue has tended to be approached in the literature. Typically, critics point out how, upon finding out about some of Allen's actions – most notably, his relationship with Previn – they respond differently to Manhattan than they once did: they feel much more ‘urpy’ (Dederer, Reference Dederer2017) than they did before they came to this knowledge, and, as a consequence, now find the film (much) less amusing – much less funny – than they did originally (Matthes, Reference Matthes2021, pp. 20–22). I take it that such observations are not intended merely as reports of people's responses to Manhattan, but as evidence that the film is now less artistically successful than it once was: in this case, less funny. Allen's moral character, so these critics assume, bears on how the film should be interpreted. Their thought, reconstructed charitably, is that Allen's moral character plays a part in determining how Manhattan is in itself, such that the film has now taken on a more sinister hue – and, as a result, is much less funny – than it was on first release.

The clearest, most philosophical writer on this issue, Erich Hatala Matthes (Reference Matthes2021), understands the implications of this position very well. Matthes is aware that the case for Allen's ‘moral misdeeds’ making Manhattan less funny depends on its being shown that these misdeeds are responsible for a real change in the film: a change in something ‘found in the work itself’ (Matthes, Reference Matthes2021, p. 29). He is well aware that what needs to be demonstrated here is not merely that our knowledge of Allen's romantic backstory makes us resistant to responding to the film in the way prescribed (i.e., by finding it amusing), but that the source of this resistance is something about how the film really is. Perceptively, Matthes takes this, in turn, to require that the said immorality enters the work itself in a way that has altered or shaped its meaning (Matthes, Reference Matthes2021, p. 25): what is there to be understood in it.Footnote ²

This requires, of course, that it is possible for Manhattan to undergo meaning-change over time, after Allen finished it. Matthes seems happy to grant this, stating that an artist's moral flaws might subsequently ‘alter’ or ‘change’ his work's meaning (Matthes, Reference Matthes2021, pp. 21, 42), and he is happy to use Manhattan as an example of this (Matthes, Reference Matthes2021, pp. 20–22). Now, change (for which ‘alteration’ is a synonym) is a strictly temporal phenomenon: it is nothing more nor less than the temporal variation of the properties of things. So here we have Matthes committing to the idea that Manhattan can change its meaning over time, after Allen finished making it.

Some, no doubt, will recoil from this commitment, but in this paper I am happy to grant it. While the standard view in the philosophy of art is that a work of art cannot change its meaning through time, this view appears to be motivated by a questionable assumption, namely, that an artwork's meaning is akin to the meaning of a dateable utterance: the thing said by the employment of a sentence in a specific context (Levinson, Reference Levinson1990, p. 192). The thing said by a dateable utterance of a sentence in a given context cannot change, and so if the meaning of a work of art is like the meaning of such an utterance, then neither can an artwork's artistic meaning.

The problem with this train of thought, however, is that the key assumption here is insufficiently motivated. Why, for example, should we not think of the authorship of a work of art as more akin to the coining of a new word, and so why should we not think of artistic meaning as more akin to word meaning? Words, once coined, are, so to speak, handed over to the linguistic community to use as it sees fit, and this can include altering what counts as their correct use, i.e., changing their meaning.Footnote ³ If, as seems not implausible, works of art, once completed, are similarly handed over to the artworld to appreciate, make sense of, and find meaning in, then they too will be capable of changing their meaning. And what is more, such a conception will enable us to take at face value, and without having to explain away, what seem to be compelling examples in which the meaning of an artwork – what is there to be understood in it – is subject to change in light of historical developments (McFee, Reference McFee1980, §2).

3. An Account of Meaning-Change

Having said this, there is still a major impediment to Matthes's making his case that Manhattan has undergone the meaning-change required to make it less funny than it once was. And in short, it is that of establishing that the relevant datum is best explained in the way he suggests. The now familiar datum in question is this:

1. (D) Knowing what they do now about Allen's subsequent sexual behaviour, people tend to report that they now respond differently to Manhattan than they once did: specifically, they find the film much more sinister and, as a consequence, much less amusing than they did before.

Let us grant (D). Clearly, (D) does not entail that the film really has become more sinister, and hence less funny, than it was at its time of release; that case has to be made by demonstrating that this thesis of meaning-change is the best explanation of our changed responses to the film.

An analogy might help this point to stick. I might be ‘blocked’ from appreciating a performance by the comedian, Michael McIntyre, because he reminds me of an irritating work colleague, but that is not in itself a fault with his performance. The origin of my failure to find the performance funny lies with me, not him; specifically, it lies in my having been distracted by facts irrelevant to the performance's artistic merit. So what Matthes must do is explain why people's recent inability to find Manhattan as funny they once did is not another example of the same phenomenon. Why, then, should we treat people's attested recent ‘urpiness’ at the film, in light of their knowledge of Allen's romantic backstory, as tracking a genuine change in the film itself, rather than something resulting from their having been distracted by extraneous biographical facts.

This is the crux of the issue, and the only way that Matthes can convince us that the film's meaning really has been altered in the way he thinks by what he calls Allen's ‘immorality’ (Matthes, Reference Matthes2021, p. 20) is to give us a convincing explanation of how such meaning-change has occurred. So how does this putative explanation proceed?

The starting point in Matthes's explanation is that an artist's moral shortcomings can be relevant to the content – the meaning – of his artworks only if ‘we can tell a compelling story about how the artist's immorality relates to th[is] content’ (Matthes, Reference Matthes2021, p. 20). So, for example, the reason why Hitler's immoral character is irrelevant to the interpretation of his banal, one might say ‘vapid’, cityscapes is that there is no trace of such immorality in the content of the paintings. What is there to be understood in these paintings is so divorced from the artist's immorality, that there is, so to speak, nothing in them for our awareness of the artist's immorality to get a grip on when we interpret them.

The thought here would seem to be that an artist's real-life moral turpitude can alter his work's meaning in a way that makes it less artistically successful only by being of a similar nature to, and thereby accentuating, a moral flaw that is already present in this work. So, for example, while Aaliyah's insistence that her ‘Age Ain't Nothing but a Number’, in her recording of the song, might set off some alarm bells, perhaps our knowledge that the song was written by R. Kelly, to whom Aaliyah had become secretly married aged 15, ‘give[s] us additional reason not to have the response that the music is aiming to elicit’ (Matthes, Reference Matthes2021, p. 20). Finding the track now sinister and disturbing, we might come to hear the (apparently) infectious groove as part of a project justifying the sexual exploitation of children. In this way, the moral flaws of the song's writer might appear to heighten those in the song to such an extent as to cancel out its musical virtues.

So how does an analogous account proceed in the case of Manhattan? Here, too, Allen's putative moral failings must correspond to a thread in the film itself, supposedly deepening our already held concerns about the film's message to such an extent that our amusement disintegrates. The precise story appears to go like this (Matthes, Reference Matthes2021, pp. 20–21). Manhattan, so Matthes argues, asks us to empathize with – by normalizing and romanticizing – the sexual relationship between the middle-aged Isaac and the seventeen-year-old Tracy. And Matthes's contention is that this aim, already far from straightforwardly achievable due to both the nature of that fictional relationship itself and Allen's habit of casting himself opposite much younger women, is further compromised by his real-life relationship with Previn:

It is the salience of this fact about Allen's relationship with Previn that, Matthes thinks, alters the meaning of the film, making it creepier, more sordid, and hence much less amusing, than we found it back in 1979.

But the picture is incomplete. Why, exactly, does Allen's relationship with Previn really give us an ‘additional reason’ not to empathize with Isaac's relationship with Tracy? What sort of meaning-change is thereby engendered? In Matthes's view, the context provided by Allen's relationship with Previn transforms the film into a defence of, or apology for, such exploitative relationships, thereby drawing the audience into a project of redeeming Allen himself for his own relationship of the same kind (Matthes, Reference Matthes2021, p. 21; see also Willard, Reference Willard2021, p. 102). It is because the film has come to function, subsequent to its completion, as an apologia that we now have additional, and powerful, reasons not to respond to the work as Allen wants us to, specifically, by regarding Isaac and Tracy's relationship as morally acceptable (Matthes, Reference Matthes2021, p. 22). The audience's sense that it is being manipulated in this way prevents it from finding Manhattan as funny as it once did. The film has become creepier, more sordid, and less amusing.

I have three responses to all of this. First, in §4 I deny that Manhattan attempts to normalize relationships between middle-aged men and teenage girls. It turns out that Matthes's account of how Manhattan changes in light of Allen's subsequent actions is predicated on a misreading – indeed, an oversimplification – of the film. Second, in §5 I show that, even if we accept Matthes’ interpretation of the film, a key element of his account of how it undergoes meaning-change – his claim that Manhattan has come to serve as a defence of Allen's (alleged) misdeeds – cannot be sustained. And finally, in §6 – and this is the criticism that generalizes – I outline two troubling conceptual obstacles that must be overcome before we can make good on a claim that an artist's work can be made artistically worse by mere dint of his misbehaviour or morally suspect character.

4. Understanding Manhattan

On the approach to Manhattan that I favour, there is sufficient unresolved complexity, both within individual scenes and concerning the perspective from which the narrative unfolds before us, to falsify the suggestion that the film normalizes Isaac and Tracy's sexual relationship. The troubling nature of this relationship is not something that the film wants us to sanitise or brush under the carpet; it is something that it grapples with out in the open. Manhattan explicitly treats this relationship as an abnormal one.

Of course, the film sometimes depicts that relationship from the inside, with Tracy and Isaac at times appearing content and in love. In these scenes, the film's use of black and white, together with the Gershwins’ lush, tender music, conveys how the two of them are feeling about their relationship in the moment: carefree and romantic, like lovers in a Hollywood movie. And yet this self-conception is regularly jolted and undermined. For example, Mary, the mature woman for whom Isaac leaves Tracy, pointedly compares him, albeit sardonically, with Lolita's Humbert Humbert in the following exchange:

And later, once Isaac has explained the collapse of his marriage, Mary mockingly replies that she thinks that accounts for ‘the little girl’. Mary, the film makes clear, is someone whose testimony we trust. She is a victim of the narcissistic to-ings and fro-ings of Isaac and his best friend, Yale, and someone who, though sometimes dismissive and blunt, is vulnerable, with a pronounced moral centre. The audience should take these references seriously.

Isaac's conduct is also regularly undermined in the film (Smith, Reference Smith2016, pp. 334–38). Most notably, it is made clear that he is a statutory rapist, since he is sleeping with the seventeen-year-old Tracy, while, within the fiction, the age of consent in New York is eighteen. When Isaac makes a light-hearted, off-the-cuff remark about being concerned that the police might burst in on them, this reflects his own awareness of their sexual relationship's illegality, and we are left wondering how he can be so tone-deaf to the significance of the moral boundary he has crossed. We can only conclude that Isaac is so wrapped up in himself, so wholly self-absorbed, that he is incapable of fully registering the implications of what he is doing.

Furthermore, and however strenuously Isaac tries to make light of it, it is true in the fiction that, on having been deserted by his wife, Jill, for her now long-term partner, Connie, he tried to run them both over. Jill's own low opinion of Isaac is presented powerfully, and at length. He has rages, he is a chauvinist, he is narcissistic, and so on. Since we can confirm that at least some of her testimony is true (notably, the attack with an automobile), this strongly suggests that she is likely to be a reliable character witness.

True enough, Isaac has an amusing side to his character, and we can certainly identify with him in many of the situations he finds himself, most notably when Yale recommences his own affair with Mary after she has moved in with Isaac, clearly betraying Isaac's trust. Manhattan's complexity and maturity lies in its recognition that a middle-aged man conducting a sexually exploitative relationship with a seventeen-year-old girl can nonetheless have attractive elements to his character and encounter situations in which we sympathise with him. None of this effaces the film's presentation of the sexual relationship at its centre as an aberrant, morally troubling one. Its achievement is to present this relationship in nuanced ways, including from the perspectives of its participants, hold back from flat-footed moralising, and thereby enable viewers to think through the issues for themselves: a process involving them in reflecting upon their own conflicting and unnerving feelings towards Isaac, as they are both amused and repelled by him. By the end of the film, if things go well, we might emerge with a slightly more profound understanding of what people are like and how they can be.

In short, Isaac is not someone with whom we are supposed to empathize as he conducts his relationship with Tracy, so much as better understand, where this involves getting a sense both of his attractiveness to Tracy and of how they feel, or think they feel, about each other. And this interpretation is rendered more plausible still once we consider the question of whose narrative Manhattan is. Matthes presupposes that the film is Allen's wholly empathetic narrative of a sexual relationship between a middle-aged man and a young girl. We have just seen that the film is much more nuanced than that. But now consider the film's opening, in which Isaac stumblingly attempts to start his novel, a novel about ‘people in Manhattan who … are constantly creating these … unnecessary neurotic problems for themselves ‘cause it keeps them from dealing with … more unsolvable, terrifying problems about … the universe’ (Allen, Reference Allen1982, p. 267). This suggests that what follows this opening scene might actually be the film of the novel that Isaac is writing, rather than Allen's authorial vision of the events which inspired it (Smith, Reference Smith2016, p. 337).

If this is the correct way to read the film – or even just a sanctioned way of reading it that we can hold alongside the interpretation of it as an objective account of events – then it seems to me that Matthes simply cannot be right in thinking of it as being squarely in the business of normalizing or romanticizing sexual relationships between middle-aged men and teenage girls. For once we open up the possibility that the film is Isaac's own version of events – the narrative of narcissistic, statutory rapist – then even those moments in which Isaac and Tracy's relationship are portrayed romantically take on a different character. We begin to wonder whether they are the soft-focus, self-justificatory ‘reminiscences’ of a sexual predator with violent, narcissistic tendencies. Such reminiscences are highly likely to be unreliable. Indeed, we might strongly suspect the narrative that unfolds to be Isaac's confabulated defence of his conduct.Footnote ⁴

On this way of taking Manhattan, as a work acquainting us with Isaac's narrative of his relationship with Tracy, the film's final scene becomes a masterpiece of superficially persuasive, and yet meretricious, unreliable narration. Mary having left him to take up once more with Yale, Isaac suddenly comes to the view that Tracy has always been the woman who has meant most to him, and immediately sets off to her apartment with the aim of reconciling with her. Arriving just as Tracy is leaving for London to train as an actor, Isaac seemingly moves her with his expression of love, and yet Tracy insists that she cannot change her plans, pointing out that it was Isaac's idea that she should study abroad. Isaac, with a growing smile, realises that Tracy is right, and wise beyond her years. Perhaps she has taught him something.

On the reading that I have been exploring, however, this narrative is a piece of make-believe – one which has a wonderful relationship between equals coming to an end, its participants parting as friends – that its author, Isaac, cannot recognize as such (Smith, Reference Smith2016, p. 338). If we understand the subtlety in the film – the way in which it invites us to consider the perspective adopted to be that of an unreliable narrator, not that of the film-maker – then we can get a feel of this perspective from the inside, while distancing ourselves from it by stepping back and seeing it for the dangerous and self-serving confabulation that it is. Once we read the film in this way, we can see it as inviting us to reflect upon Isaac's own version of the story, understood as such, without authorial comment; and this we do by, so to speak, trying this perspective on for ourselves and then evaluating it. It is not Manhattan that acts as a defence of, or apology for, Isaac's relationship with Tracy, but Isaac's narrative within the film: the narrative that unfolds after the film's opening scene. And so, with this in mind, we can interpret Mary's reference to Nabokov as, not merely a humorous episode in Isaac's own narrative, but as a clue to what this film is really doing. Like Lolita (1959), Manhattan can plausibly be read as bringing us right up, and uncomfortably, close to a sexual predator, letting us in to his point of view without demonising him or overtly sermonising. This makes the film unsettling – scary, even – but in no sense a defence of a statutory rapist and an exploiter of girls. On this way of interpreting the film, it no more normalizes or romanticizes sexual relationships between middle-aged men and teenage girls than Apocalypse Now serves as a defence of the United States’ involvement in the Vietnam War.

The moral here is that Matthes's account of how Manhattan has come to be more sinister, and hence less funny, in recent years relies upon a superficial and somewhat rudimentary reading of the film. As we noted in §4, the crucial claim that Manhattan has come to function as an apology for Allen's own sexual abusiveness relies upon the interpretation of the film as normalizing and romanticizing sexual relationships between middle-aged men and teenage girls. But as we have seen, it is precisely this interpretation that is flawed. So Matthes's account of how the film can be rendered less funny by mere dint of Allen's moral flaws tumbles upon its first step.

5. The Account Rejected

Having made this point, I shall now put it to one side. For even if we accept Matthes's reading of Manhattan, there is nonetheless compelling reason to reject his account of the mechanism by which the film has changed its meaning in such a way as to become less funny than it once was. As I explained in §2, I would not want to rule out the possibility that artworks can change their meaning, post-completion, tout court. But Matthes's story, according to which the film undergoes meaning-change by virtue of becoming an apology for, or defence of, Allen's subsequent alleged misdeeds (Matthes, Reference Matthes2021, pp. 21–22), does not stand up to scrutiny.

Matthes's explanation of how Allen's supposed misdeeds have come to alter Manhattan's meaning makes much of that thing called ‘context’. ‘It can be the context’, he says, ‘as opposed to the intention of the artist, that makes the work function as the defense of the artist's moral misdeeds’ (Matthes, Reference Matthes2021, pp. 22). But what is this context that has supposedly turned the film into a defence of something that happened some twelve years after the film was released? Well, presumably, it is constituted by the following two facts: that relatively well-informed cinema-goers are now aware of, among other things, Allen's commencement of a sexual relationship with the twenty-one-year-old adoptive daughter of the woman he was in a long-term relationship with; and that, by and large, they morally disapprove of this relationship. So I suppose that what is supposed to have happened is something like this. Noticing, as Matthes would have it, that Manhattan normalizes sexual relationships between middle-aged men and teenage girls, the participants in the aforementioned context collectively come to adopt the interpretation of the film briefly introduced in §3 above: that is, they come to view it as a kind of morally disreputable ‘redemptive project’ (Matthes, Reference Matthes2021, p. 22): something that, in effect, functions as a defence of Allen's subsequent sexual relationship with the twenty-one-year-old Previn.

Clearly, such an account relies on a response-dependent conception of artistic meaning. Roughly, an artwork's meaning, according to such a picture, is what it would be judged to be by an appropriate subject in appropriate conditions. Such a view, properly worked out,Footnote ⁵ strikes me as highly plausible. Echoing John McDowell, I suspect that we cannot construct a conception of artistic meaning that is detached from the idea of an artwork's seeming to have it.Footnote ⁶ But notwithstanding this element of agreement, it is numbingly unclear why an appropriate subject, properly situated, would judge Manhattan to be the ‘redemptive project’ that Matthes believes it to have become. For it is mysterious what would form the basis of such a judgement. To be sure, someone convinced that the film presents Isaac and Tracy's relationship as non-exploitative might use it as an example to illustrate why, in real life, the age-difference between Allen and Previn is not in itself evidence that Allen was guilty of sexual misconduct. But using the film in this way would not show it to be, in itself, a defence of the latter relationship. There is complexity here that is worth exploring.

Graham McFee, albeit tentatively, describes a more plausible example, again from Allen's oeuvre, of how a film's meaning might evolve or otherwise change after it has been finished:

Maybe, although the details require greater working out. In particular, what needs to be established here is that what McFee describes as ‘a new social poignancy’ really represents a change of meaning in the sense with which we have been operating hitherto: what I described in §2 as a change in what is there to be understood in the film itself. This additional social poignancy, at the moment somewhat under-described, could turn out to be a change, not in what is there to be understood in the film, but just in its significance for us. Significance, thus conceived, is a phenomenon akin to sentimental value: something altogether less objective and more personal. A change in the film's significance for us consists, not in a real change in the film itself, but merely in a change in how we feel about it or regard it. To a use an ordinary-language distinction, it is a change not in the meaning of the film, but in its meaning for us.

Nonetheless, McFee's example does have some appeal philosophically. Crucially, though, this is in no small part precisely because it does not deploy the conceptual machinery dubiously appealed to by Matthes. Notably, McFee holds back from describing the scene in question from Sleeper as a ‘critique’ of, or ‘satire’ upon, the Watergate scandal, and this is all for the good. For these descriptions do not bear directly on Sleeper's meaning, but upon how it should be categorized: that is to say, upon how the film should be ‘classified or taken in some specific or general way’, upon ‘how it is to be fundamentally conceived or approached’, or upon what the work is for, ‘on a rather basic level’ (Levinson, Reference Levinson1996, p. 188).Footnote ⁷ Now, while a film's meaning can, perhaps, change post-completion, its currency evolving over time within the relevant cultural context, what the film is for, its categorial nature, cannot. This is fixed by the maker's ‘categorial intentions’ (Levinson, Reference Levinson1996, p. 188): his intentions as to what kind of film it is. In short, while Sleeper's meaning might evolve over time, the fact that it is a comedy, rather than a critique of, or a satire upon, the Watergate scandal cannot. These facts were fixed by Allen in 1973 and cannot be changed.

Well, the same goes for Manhattan. It cannot function as a defence of, or an apology for, anything unless Allen made it so, since being a defence of, or an apology for, something are categorial properties, if anything is. So it follows that ‘context’, post the film's completion, cannot turn the film into either a defence of, or an apology for, Allen's subsequent sexual behaviour, specifically his putatively immoral sexual conduct with Previn. Hence, since Allen himself could not have given the film this function (because the said behaviour occurred after he made the film), there is simply no available explanation of how the film could have come to function as Matthes claims it to.

6. Meaning-Change: The Very Idea

Let us now move on from the question of the cogency of Matthes's specific proposal as to how Manhattan has become less funny by mere dint of Allen's (supposed) misdeeds. For even if his account of how this has happened were more plausible than it is, there would still be two conceptual obstacles to be overcome before we would be able to grant that the requisite kind of meaning-change has, indeed, taken place.

The first can be approached by revisiting an observation from §3: the mere fact that audiences report finding Manhattan less funny since they became aware of Allen's putatively immoral behaviour – the datum I labelled ‘(D)’ – does not in itself demonstrate that the film really has undergone the kind of meaning-change Matthes and others claim for it. As we noted then, what needs to be established is not merely that people with this knowledge tend to respond differently to the film, but that these altered responses track a genuine change in the film itself. To put it another way, what must be demonstrated is that knowledge of Allen's moral shortcomings has become necessary to understand something about the film itself, i.e., to grasp something of its meaning. And this is an extremely difficult claim to secure.

Clearly, what has to be established here is that this case is significantly different to those we want to describe as cases in which certain worldly knowledge serves only to distract us from responding to an artwork as we should (as in §3's case, in which Michael McIntyre's resembling an annoying colleague puts me off finding him funny). Why doesn't knowledge of Allen's allegedly immoral behaviour count as information that is likewise irrelevant and distracting?

The only available answer, as I have just suggested already, is this: because knowledge of Allen's behaviour is, while seeing the resemblance between McIntyre and an irritating colleague is not, required to understand something in the relevant work (in one case a film, in another a comedic performance). But how can this be argued for?

Presumably, by drawing upon what we have seen Matthes describe, in §3, as ‘a compelling story about how the artist's immorality directly relates to the content of their work’ (Matthes, Reference Matthes2021, p. 20). But once we strip away the implausible claim that Allen's subsequent putative misdeeds transform how the film is classified, turning it into a redemptive or defensive exercise, all this leaves us with, at best, is the observation that there is a correspondence between the film's subject-matter and Allen's subsequent sexual behaviour. This, indeed, is the only sense in which it is incontrovertibly true that ‘the artist's immorality directly relates to the content of their work’ (Matthes, Reference Matthes2021, p. 20).

However, this fails to speak to what we need: namely, an explanation of why what we know about Allen's relationship with Previn enables us to better understand Manhattan, or, to adopt Matthes's own argot, of why this knowledge gives us an ‘additional [normative] reason’ (Matthes, Reference Matthes2021, p. 21) not to respond to the film in the way it prescribes. The opponent here will insist that the mere fact that the film and Allen's subsequent sexual behaviour embody the same or similar themes is not enough to show that our understanding of the film is enhanced by an awareness of this extra-textual knowledge. Or, to put it Matthes's way, the mere fact that subjects are now motivated not to respond to the film in the way prescribed, and so find it less funny than before, does not in itself show that they are thereby acting on a normative reason (i.e., a consideration that genuinely counts in favour of so responding). All in all, the ‘compelling story’ about how Allen's own sexual behaviour and the film are related falls well short of demonstrating that Manhattan's meaning has been altered in light of the said behaviour. In essence, it just leaves us with (D), which we all agree upon, but whose recognition does not even start to justify the claim that the film's meaning has changed in the way suggested.

The second conceptual obstacle to making good this claim concerns the very idea of work-meaning. As I pointed out in §2 and reiterated in §5, in order for the case to be successfully made that Allen's subsequent behaviour brings about genuine meaning-change in Manhattan, the film's meaning must be something there to be discovered in the work through successful interpretation of it. That is, our resistance to responding to it as we are prescribed must be rooted in how the film is in itself, as Matthes himself realises (Matthes, Reference Matthes2021, p. 29). But sharpening our conception of work-meaning in this way only goes to show how difficult it is for Matthes to establish that Allen's putatively immoral actions have genuinely effected meaning-change in the film. The idea that such alteration in the film's meaning has, indeed, taken place certainly appears plausible if ‘meaning’ is what I in §5 termed ‘significance’. It is much less plausible once we think of meaning as we must do so here, as something there to be retrieved from the work itself.

As if to lay bare just this problem, Matthes's analogy between the meaning of a work of art and the ‘meaning’ of an athlete's success falls utterly flat. Hoping to convince the reader that an artwork's meaning can be modified by the artist's immoral behaviour, he points out that the cyclist, Lance Armstrong's, doping has ‘alter[ed] the meaning of his success’ (Matthes, Reference Matthes2021, p. 22). But this just sees him slide from the notion of meaning that matters in the artistic context – namely, as something there in the work itself to be understood – to the more personal notion of something's significance that we have just put to one side. Now that we know that Armstrong consistently cheated throughout his career, we think of his achievements very differently; but these achievements do not have meaning in our more robust sense. These achievements certainly have significance for us; they matter to us. But they do not mean anything.

None of this shows that there could not be a convincing case made for Allen's claimed misdemeanours altering Manhattan's meaning in such a way as to render it less funny than it was on first release. But Matthes, for sure, fails to make such a case, and his failure is illustrative. What would need to be shown in a case like this is that our knowledge of the relevant real-life facts actually enhances our understanding of the work itself, where this notion of understanding tracks something there to be retrieved from the work, rather than a notion of ‘meaning’ that collapses into mere significance. Demonstrating that an artwork's meaning, in the former, more robust sense, has been genuinely changed by an artist's moral shortcomings is hard work, indeed.