The authorial fallacy: what literary theory, Roald Dahl, Donald Trump, and artificial intelligence have in common
Johanna Gibson
Search for other papers by Johanna Gibson in
Current site
Google Scholar
Full access

‘I’m not sure I quite know what that means,’ Sophie said.

‘Meanings is not important,’ said the BFG. ‘I cannot be right all the time. Quite often I is left instead of right.’1

‘Well, sir’ said the constable, ‘he’s the man we were in search of, that’s true; and yet he’s not the man we were in search of. For the man we were in search of was not the man we wanted.’2

In 1943, the literary theorist, William Wimsatt, and the philosopher of art and aesthetics, Monroe Beardsley, produced an entry for ‘Intention’ in Joseph T Shipley’s Dictionary of World Literature.3 It is notable that this somewhat revolutionary entry was written at a time of World War and intense scrutiny and criticism of populist rhetoric and cults of personality. At the same time, this critical re-examination of authorial intent was somewhat startling to the dominant traditions of textual interpretation: ‘There is hardly a problem of literary criticism in which the critic’s approach will not be qualified by his view of “intention”.’4 It was also a significant departure from the immutability of the literary canon and critics like FR Leavis who determined an elite tradition and delivered a type of literary criticism that self-consciously rendered a classist account of authors and readers alike.5

After the War, in 1946, Wimsatt and Beardsley published their longer treatment of the question of intention in the now famous essay, The Intentional Fallacy,6 where they maintain that the author’s intention is ‘not a part of the work as a linguistic fact’, it is external and therefore ‘private and idiosyncratic’.7 The work is also contingent, historical, and situated: ‘In the course of years a work may undergo a shift in meaning in some of its words, so that one may have to distinguish between the work “then” and the work “now”.’8 In the original and brief dictionary entry, they make a fundamental statement of the fallacy: ‘The work after being produced must continue to exist independently of the author’s intentions or thoughts about it.’9 Since then the rejection of intention as a critical approach to interpretation has become ‘a stable law of literary interpretation’.10 But does that independent existence extend to influencing the publisher’s thoughts as well?

Eighty years later, with the best of intentions, so to speak, Puffin, an imprint of Penguin Random House, announced it would be editing the works of Roald Dahl for sensitivity. This led to enormous press attention, not only on the plans for Dahl’s books, but also on the role of sensitivity readers more widely. The use of sensitivity readers in publishing has received growing media attention over recent years,11 but with Puffin’s announcement it reached a real crescendo.12 The reports led to reactions from both the Prime Minister13 and the Queen Consort,14 before the ultimate announcement that the original versions would continue to be published.15


In the discourse surrounding the Dahl row, much of the focus has been on the author and his abhorrent views,16 the integrity of the work,17 the market,18 and censorship,19 generating an increasingly agonistic debate settling out along disparate poles of populism. Stopping short of describing a kind of copyright evergreening, David Mitchell argues the decision was largely about refreshing the market for the books and squeezing as much out of the intellectual property as possible. In his scorching criticism of the recycling of intellectual property, he moves through film, television, and streaming, leaving prequels, adaptations, remakes, and other ‘old stories’ in his wake: ‘There is nothing soft about making these changes at all – it is commercially ruthless.’20 He argues that dead authors come with brand recognition and are ‘so much more accommodating to market forces than those pesky living authors with their obstructive artistic concerns’.21

The fact that this action by a publisher is possible, by virtue of copyright,22 is not in itself very interesting or surprising. Adaptations are somewhat reliable intellectual property, and this is indeed the same business model that is playing out in Amazon’s acquisition of MGM23 and adaptation models in streaming more widely.24 And this includes the Roald Dahl literary estate itself. The Roald Dahl Story Company became a division of Netflix after its acquisition by the media company in 2021.25 In view of this acquisition and their emphasis on literary adaptation as a significant streaming property, Netflix was accused of being the momentum behind the plans.26 Use the brand, revise the story.27 However, this was later clarified and denied, with the plans having been underway in consultation with Inclusive Minds,28 which specializes in the children’s book industry, reportedly at least a year before the acquisition.29


But what has received much less attention, if any at all, is the reader. Apart from the comments of outrage, and whether readers can read what they want to read,30 there is no rigorous consideration of the reader’s ‘property’. Where is the consumer perspective in this debate? But it has always seemed that consumer is the wrong word for the enjoyer of copyright. The consumer does not drain and deplete the work. The consumer does not drain and devour. The consumer reads.

Beyond the banality of cancellations and rehabilitations, there is something interesting going on here. Is there a reader dimension to authorship? And to what extent can copyright and moral rights protection be construed as protecting the consumer? The episode raises several issues for readers and the quality of consumer information, issues which extend across other creative industries, including film adaptations and similar. What interest does the consumer have, if any, in preserving the integrity of the book? There have been some very clear examples of fundamental and literal changes to the products purchased by consumers, such as the reports of automatic updates to eBooks.31 But what is fascinating is the way in which moral rights might anticipate a kind of ‘consumer protection’ in copyright law, beyond the brand recognition of authorial attribution. It’s not the man we are in search of, it is the work.


What is a sensitivity reader? A sensitivity reader is a type of specialist editor who may bring a particular and personal expertise or experience to reading for offensive and biased material in manuscripts. The sensitivity reader then prepares comments and feedback for the author, just like any editor. Novelist, Juno Dawson, has argued that the criticism is misplaced, and that a sensitivity reader works within a more collaborative process, giving the kind of feedback offered when anyone reads a manuscript before publication.32 And at least one sensitivity reader criticized Puffin, identifying this collaborative relationship as both central to the process and completely overlooked by the publisher in relation to the Dahl books. Virginia Mendez argues that in relation to a Dahl work, a work that is ‘so obviously of its time’, Puffin’s approach is ‘extremely shortsighted’.33 She explains: ‘What Puffin failed to grasp is the golden opportunity to embrace and encourage critical thinking in children.’34

Others, however, have raised questions about an averaging of language and have levelled this criticism directly at the concept of sensitivity reading itself. Or as Lionel Shriver, not unfamiliar with controversy, put it six years earlier, ‘[I]f all the characters speak with the same courtesy, and voice the same standard left-of-centre views, contemporary fiction can’t hope to contribute to the understanding of a world that elects Donald Trump.’35 There are many topics on which Lionel Shriver and I disagree, and I think she has mischaracterized sensitivity reading here as well, but in relation to the specific facts of the Dahl row, the point is well made.


The proposed retrospective rewriting of Roald Dahl’s books is something altogether different from the usual collaborative and iterative editing process with a sensitivity reader. And as such, this episode throws into stark relief the interaction between the integrity of the work, the proficiency of the reader, and the moral rights or legacy of the author. In the focus on language, the primary subject appears to be the integrity of the work and the preservation of a cultural object on the one hand,36 or its renewal on the other.37 But what about the individual reader in this relationship? To what extent can the reader pursue the integrity of the work. As Wimsatt and Beardsley counselled eighty years earlier: ‘Included in the meaning of the words will be their whole history as far as determinable, and all the uses and associations of the words that went to make up their value when the word was produced.’38

Or is the controversy really, once again, all about celebrity, about authorship and authorial legacy? It would seem that the unconscionable views of the author are considered part of the mechanism for making meaning in the work. Notwithstanding the language itself, the surrounding discussions seemed to conflate Dahl the man with the works of Dahl, and imply that it is now somehow the responsibility of the work to account for his behaviour. Once again, Wimsatt and Beardsley frame this consideration: ‘[T]he design or intention of the author is neither available nor desirable as a standard for judging the success of a work of literary art, and it seems to us that this is a principle which goes deep into some differences in the history of critical attitudes.’39 So if ‘[i]t is therefore circular and misleading to speak of judging the work with respect to its success in carrying out the author’s intention’,40 why are we once again focusing on the intention of the author? On that question, copyright provides some clues.


So where is Roald Dahl in all this? Well, we can read all about it. Just like copyright’s somewhat tautological rendering of originality,41 intentionality is similarly wrought in traditions of literary theory. In other words, both originality and intentionality are created by the work, or more specifically, in the ‘work’ of the work. In this respect, originality is taken for authorship, intentionality (and authorial voice) is taken for meaning – but, reflecting upon the intentional fallacy, neither set of conflation of terms reflects the internal evidence of the work. However, it does reveal the mechanics of copyright and its authorial legacy. The editing of Dahl demonstrates this perspective quite clearly. Is it really about the editing of language? Or is it about recuperating and refreshing the reputational legacy of Dahl the author, and the crucial value in Dahl the brand? This conflation of the work and the man is powered by the resilient affection for intentionality and its proficiency in building a story of intangible value. It is notable that a renewed attention to intentionality is perceptible in the anxiety around present incursions of artificial intelligence in visual art and other works of authorship. How can we trust the authenticity of art generated by artificial intelligence, unless we can infer intentionality on the part of the ‘author’. This is the intentional stance that is recruited in building trust as well as authenticity in human interactions with artificial intelligence. But what is this intentionality and does it have anything to do with the actual machinations of copyright?

Through the intentional fallacy, and a similar operation in copyright as well, the author is made by the work, is internal to the work. More significantly and precisely, the construct of the author and the concept of intentionality are both revealed by what the work does.42 In a sense it is a kind of intentional stance facilitated by the perceived authenticity of the work. And this has significance across the range of intellectual property rights and their objects; means, if you will. As this issue goes to press, arguments have just been heard in the Supreme Court in Thaler v Comptroller-General as to why an artificial intelligence machine should be recognized in law as the actual deviser of an invention.43 Intentionality is deployed in defence of the machine, including in the submissions, a reference to the machine’s own identification of the commercial value of the invention.44 In other words, if a machine intends, then surely a machine devises? The trouble is that it is possible to infer this intentionality, whether copyright or patent, only through what the intellectual property artefact does.

Indeed, this is the tautological impossibility of intentionality. Is it both imputed and ‘enforced’ through the reader, not the author. What Wimsatt and Beardsley have to say about poetry is strikingly resonant in this context of the interpretation of the inventor. They describe judging poetry as ‘like judging a pudding or a machine. One demands that it work. It is only because an artifact works that we infer the intention of an artificer.’45 In other words, what the artificer or maker intended (whether that maker is an author, a deviser, an inventor, and so on) is irrelevant. Intentionality has nothing to do with the work that is done by that artefact, whether that artefact is a poem, a Roald Dahl story, or indeed, a food container.


Wimsatt and Beardsley are usually associated with New Criticism, and the intentional fallacy was certainly adopted by this tradition of criticism, which has been largely overtaken by reader-response theories. However, the intentional fallacy is perhaps also an unintentional theory of readership. The authors suggest that, ‘The meaning of the work is all that the ideal reader can find in it; and the ideal reader is the one fully aware of the accepted value of the words and least influenced in his interpretation by idiosyncratic associations.’46 However, they accept that readers are never ideal, ‘and no two readers will ever read the work in exactly the same way’, leading them to conclude that ‘the meaning of the work lies within an area of readings’.47 The author becomes ‘a witness to the meaning of his work and one may even grant special validity to idiosyncratic associations of the author’, but they emphasize that, ‘This kind of evidence, however, must be distinguished from the author’s “intention”.’48 Stanley Cavell’s critique of the intentional fallacy assists here. Cavell argues that investigations into intention actually direct the reader ‘further into the work’.49 Thus, the reader may ‘accept the words, or refuse them; wish for them, or betray them’, but will always be ‘implicated’.50 This somewhat many and varied contribution to meaning is actually very relevant in an intellectual property context. Jacques Derrida accepts intention as persistent, but no longer authoritative: ‘[T]he category of intention will not disappear; it will have its place, but from this place it will no longer be able to govern the entire scene and the entire system of utterances.’51 Making meaning in the work is an act of readership. And the author is yet another reader.

The first ‘property’ of reading is the originality of the work itself. The concept of originality, the ultimate authorial concept, is similarly qualified by the reader. A curiosity in the jurisprudence on originality is the notion that the author ‘can stamp the work created with his “personal touch”.’52 It functions as a stamp only because it is read; that is, it is repeated, and thus recognized. Already the possibility of originality demands a certain recognition and identity that can be verified in reading, indeed, that comes about only in reading.

Then there is the work itself. Wimsatt and Beardsley maintain: ‘The evaluation of the work of art remains public; the work is measured against something outside the author.’53 Indeed, something very similar is integral to the operation of copyright as well. In order to particularize the expression it is obligatory to look not at the author’s intention for that expression, but at the reader’s meaning. Taste, for example, is not that of the author’s plan and design, but ‘particular to the person tasting the product concerned’.54 Expression is thus an issue for the reader – both in terms of its scope and in terms of its preservation. And as such, the reader has an explicit stake in the integrity of the work. What the work does in the world, is actually intended, as it were, by the reader.

But given the particular circumstances of the Dahl affair, and the emphasis on the integrity of the works, the most complex readership ‘property’ may be deciphered in the author’s moral rights and the relationship with the reader. The foregoing analysis of the relationship between meaning and intentionality, and the conflation of the authorial witness with authorial intention, is of immediate relevance to considering the way in which the reader might be ‘protected’ by the author’s moral rights. Indeed, it is through the intentional stance of the law that the reader is necessarily recruited in attempts to characterize artificial intelligence as author. But it is also in the intentional stance of the law that the author is in an inextricable time share, so to speak, with the reader.

First, any infringement of these rights necessarily, perhaps somewhat obviously, arises only in the publication of the work.55 It is not enough simply to offend the author with a mutilation of a private copy. The treatment must be in a relationship with the public. Thus, it is not simply an affront to the author, at least not alone. Rather, it is necessarily a misrepresentation to the reader that the author seeks to redress, albeit preserving their own reputational interests at the same time. Second, that the law explicitly provides for rights in the integrity of the work, in addition to the rights of attribution (both to be named and to correct the record), suggests that the law itself makes a distinction between the work and the author. In this way, there is an important separation between the author’s public identity and the work’s integrity, in that the scope of the latter is defined by the reader (in publication) and the author (as witness).

The fallacy of the intentional fallacy is that it perhaps answers the wrong questions. Intentionality in the work itself ‘lies within an area of readings’,56 including the reading by the author, by the law, and by the reader. The intentional stance in copyright law is instrumental in promoting the legitimacy of the law. At the same time, the intentional stance has been co-opted into artificial intelligence and robotics: first, in the development of public trust and in the mainstreaming of the technology and trust; and second, in the service of campaigns for authorship and inventorship. The trouble is, as Cavell might say, the latter is looking for intention in all the wrong places.57


‘Read on! Please read on! Don’t stop!’58

Returning to Dahl then. What might this mean for the readers of Charlie and the Chocolate Factory? Obviously they cannot enforce Roald Dahl’s moral rights, nor can they enforce rights in the works which they do not have. But in the very existence of moral rights there is an implied, normative obligation to the integrity of the work, and in the making of a work the reader is paramount.59

But it does mean that in its practical application and delimitation, arguably the law shapes a normative expectation of the integrity of the work once published, in conjunction with an interest in the preservation of the provenance of the work. This is in part sustained by the visual hegemony and auteurial emphasis of copyright law. But it is also sustained by the intentional fallacy, and the impossibility of imputing that desired authorial intention to the work as already made and without the author as the witness of that making.

And to some extent, there is an uncomfortable value in Dahl’s errors. There is a value in human error, in getting it wrong, in letting people make mistakes, and then generating the reading around that work. This is the work doing its work. Indeed, it is the authorial fallacy that means there is so much importance attached to getting it right, and a complete disregard for reading it right … and left, and all around. It is once again the socio-cultural context of copyright’s auteurism that makes fans of authors, users of makers, and paupers of us all.

So let us read. Or let us not. It is somewhere within that range of reading that the work lies.

March 2023