The Anxiety of Prestige in Stephen King’s Stylistics

1. Introduction

Twentieth-century literary history can often seem enmeshed in an oscillating dialectics of ‘high’ and ‘low’ culture. Horkheimer and Adorno’s “Culture Industry” (Horkheimer and Adorno 1947) and Pierre Bordieu’s La Distinction (Bourdieu 1984) are only two of many notable works in the “Great Divide”, a term popularized by Andreas Huyssen as “discourse which insists on the categorical distinction between high art and mass culture” (Huyssen 1986, vii). Huyssen framed modernism, a paragon of high culture, as displaying an “obsessive hostility to mass culture”, but as modernism ceded to (or merged with) postmodernism, the relationship between “modernism, avantgarde, and mass culture” came to be described in terms of “a new set of mutual relations and discursive configurations” (Huyssen 1986, vii, x). Postmodernism is generally described as embracing “popular,” “mass,” or “kitsch” culture through a variety of ironic strategies, especially pastiche and parody: the “postmodern paradox,” as Linda Hutcheon put it, in which “to parody is both to enshrine the past and to question it” (Hutcheon 1988, 126). While every aspect of postmodernism, including “its very existence,” has “been a matter of fierce controversy,” per Brian McHale, the “term and concept ‘postmodernism’ began to lose traction around the beginning of the new millennium”, and by 2015, “postmodernism, it is generally agreed, [was] now ‘over”’ (McHale 2015, 5) as both an active aesthetic movement and a useful discriminative term. Meanwhile, sociologists have devoted extensive study to a new phenomenon which has emerged since at least the 1980’s: highbrow “snobbery” being replaced by omnivorousness cultural consumption by elites (Peterson and Simkus 1992, Peterson and Kern 1996, Ollivier 2008). As De Vries and Reeves (2022) summarize: “The distinction between ‘elite’ and ‘mass’ consumers once dominated theories of cultural consumption […]. However, over the last quarter century the ‘elite-mass’ hypothesis has fallen out of favour in the sociological literature, largely supplanted by Richard Peterson’s ‘omnivore’ hypothesis”.

Distinctions between ‘high’ and ‘low’ are crumbling not only among readers, but academics, as well. It is now recognized that notions of canonicity and what is considered ‘literary fiction,’ by whom, and when, are highly complex dynamics of social and economic (Bourdieu 1984), gender (Light 2013, 6) and racial (So 2021) concerns. Richard Jean So writes that, “[t]oday, scholars are more interested in studying the porousness and interchangeability of these categories [of high and low], rather than their imagined difference or hierarchy,” and that “[t]he categories of ‘high’ and ‘low’ are still important to cultural scholars; it’s just that the imagined space between them has contracted or at least become altered, shaping the way works of literature are judged and received” (So 2021, 105).

But a major gap exists in many of our narratives about both the Great Divide – discourse based on a categorical distinction of ‘high’ and ‘low’ literature – and the new omnivorousness in cultural consumption which followed: how did popular fiction authors and texts respond to these discourses? While literary modernism and postmodernism basked in prestige throughout most of the twentieth century, how did the so-called mass, popular, or kitsch authors of thrillers, science fiction, romances, horror, comic books, and pulp fiction – unfairly implied as an undistinguished mass by Horkheimer and Adorno’s term, Culture Industry – respond to the dismissal, exclusion, and derision by literary fiction and its attendant gatekeepers of critical acclaim and the canon? Despite the rise of popular culture and popular fiction studies, this story remains largely fragmentary. Ken Gelder writes that “Literary fiction is ambivalent at best about its industrial connections and likes to see itself as something more than ‘just entertainment’, but popular fiction generally speaking has no such reservations” (Gelder 2004, 1). We suspect that this is far from the whole story, however; that many popular fictions have responded to issues of The Great Divide and now culture omnivorousness in a variety of textual ways.

We suggest a new term to explore such commentaries in popular fiction: the anxiety of prestige. We propose the definition: thematic or stylistic textual commentaries by ‘popular’ fiction authors on issues of literary prestige, including critical or parodic portrayals of literary prestige and its gatekeepers, or explicit or implicit attempts by the popular fiction author to attain or achieve higher literary prestige for themself, either by adopting stylistic features of ‘high’ fiction, or asserting the value of ‘popular’ fiction. This definition, while broad, provides us with a starting point to examine a wide variety of textual responses by generally-considered popular authors to issues of literary prestige, often through ambivalent or sometimes even contradictory means: retorts and responses by popular fiction to The Great Divide or the new cultural omnivorousness, which we suggest remains a largely untold story in literary history.

We suggest that digital humanities can help illuminate the anxiety of prestige, especially through its ability to distant read large corpora; as the term ‘mass’ fiction suggests, the corpus of popular fiction is certainly massive. Digital humanities can locate textual evidence more easily, through query of, for instance, thematic portrayal of literary prestige’s gatekeepers, such as literature professors, literary critics, literary awards, and so on. But corpus query can also unearth less obvious textual evidence of the anxiety of prestige through query and modelling of style and change of style, for instance corpus stylistics (Wynne 2006), which can turn up patterns in latent, formal, quantifiable stylistic features. This inquiry can be aided by, and aspire to add to, a growing body of digital humanities studies on the relations between formal textual features and perceptions of literary quality (see Verboord (2003), Hakemulder (2004), Van Peer (2008), Archer and Jockers (2016), Knoop et al. (2016), Piper and Portelance (2016), Underwood and Sellers (2016), Van Cranenburgh et al. (2019), Van Cranenburgh and Koolen (2019), Underwood (2019), Van Cranenburgh and Ketzan (2021), Van Dalen-Oskam (2023)), as well as canon (see Algee-Hewitt and McGurl (2015), Porter (2018)), genre classification (see Rybicki and Eder (2011), Schöch (2017), Underwood (2019)), and linguistic criticism of the writing advice genre (see e.g. Pullum (2004) and Pullum (2015)). We note that while recent work on literary quality is employing sophisticated computational methods that quantify dozens or hundreds of textual features at once (often features which are undefined to the scholar within a ‘black box’ of machine learning), we apply a less sophisticated corpus query method that has the benefit of allowing close reading of definable textual features.

Our term, anxiety of prestige, is coined with a nod to Harold Bloom’s anxiety of influence (Bloom 1997), and our choice of term is somewhat tongue-in-cheek, as Bloom himself was a vociferous critic of popular fiction, as well as of popular American author Stephen King (1947-), the subject of this paper. We suggest King as a major figure in inquiries into the anxiety of prestige, as King began his best-selling career (over 350 million copies sold, per Heller 2016) derided and dismissed by high literary critics, but is now firmly established as a critically-acclaimed American author. King exemplifies, and perhaps contributed to, the current cultural omnivorousness. The writer once so dismissed by high literary critics such as Bloom has been contributing to The New Yorker, a leading arbiter of literary prestige, since 1994, and King won the National Book Award Medal for Distinguished Contribution to American Letters in 2003.

2. Stephen King’s Anxiety of Prestige

King’s fiction contains prodigious quantities of commentary on literary prestige, some of which is too salient to miss, but much of which has so far not been the subject of sustained attention from scholars. Perhaps the most obvious example is Misery, in which the writer Paul Sheldon, who “wrote novels of two kinds, good ones and best-sellers”, has finished his best-selling “series of romances about sexy, bubbleheaded, unsinkable Misery Chastain” and jubilantly resumed his ambitions to write serious literary fiction, despite his audience’s protests: “He could write another […] The Sound and the Fury; it wouldn’t matter. They would still want Misery, Misery, Misery” (King 1987a, 36). Sheldon revels in the completion of his new, ambitiously literary novel, but Sheldon’s aspirations of literary prestige are thwarted when he is kidnapped by superfan Annie Wilkes, who literally chains Sheldon to a typewriter and, under threat of death, forces him to write a new genre novel about her beloved character, Misery. Many more examples from King’s long oeuvre could be named, especially as King made a rather conscious turn to attempt more “literary fiction” in the early 1990s, most notably with Dolores Claiborne (King 1992a). And questions of literary prestige are abundant in King’s fiction to this day. In Rat (in If It Bleeds, King 2020), college English professor Drew Larson, a failed high literary novelist known to “steer clear of popular fiction,” is suddenly seized by the inspiration to write a commercial pulp Western novel. In Fairy Tale, King lightly parodies academia when his teenage narrator adopts an academic career in adulthood: “I am considered quite the bright spark, mostly because of […] an essay I wrote as a grad student. It was published in The International Journal of Jungian Studies. The pay was bupkes, but the critical cred? Priceless” (King 2022, 591).

The issue of King’s literary prestige, or lack of it, also abounds in King reception. Earlier critics opined on whether King is or is not “literature,” whether he is a “mere” horror or “genre” writer or somehow more “literary” than this label might suggest. The most hyperbolic of such statements came from Harold Bloom, who introduced his edited volume of scholarly essays on King with the sentiment that “King has replaced reading” and that “King’s books […] are not literary at all, in my critical judgment” (Bloom 2007, 2). Further, a 2012 scholarly monograph on King’s magnum opus is titled Respecting The Stand (Paquette 2014), as though 190 pages of literary criticism were required to show why the novel should be respected. Scholars often cannot approach any topic in King studies without some discussion of King’s literary quality, which likewise read as disclaimers or justifications for the scholarly study itself. James Arthur Anderson, for instance, writes that “[i]t is my hope that my application of these theories will […] show that [King] is more than just a horror writer, more than just the creator of ‘popular fiction”’ (Anderson 2017, 8). This attention to King’s literariness or prestige – or otherwise – can also stand in the way of other close readings. For instance, King’s early novel, The Long Walk (King 1979), holds up well as an allegory of the Vietnam War, a fact that can be obscured when appraisals of literary value displace textual attention (see Texter (2007, 47)). King’s retorts to these decades of criticism may be read in his paratextual interviews and prefaces, for instance telling a Guardian journalist that “I have outlived most of my most virulent critics. It gives me great pleasure to say that” (Xan 2019).

More clues to King’s anxiety of prestige may be read in On Writing: A Memoir of the Craft (King 2000), which combines reminiscences of King’s career as a writer with prescriptive writing advice for would-be authors. According to King, adverbs, passive verbs, and adverbially modified dialogue attribution should be avoided, for instance. King is hardly alone in offering such writing advice to aspiring authors, which is arguably a tradition as old as writing itself; Plato himself discouraged the reader from writing at all (Plato 2005, 63)! Writing advice books today could even be considered its own genre (Evans and Kroll 2005). The writing advice in William Strunk Jr. and Edward B. White’s The Elements of Style (Strunk and White 1999), a prescriptive style and grammar guide, has sold over 10 million copies and achieved, per Geoffrey Pullum, “a vice-like grip on educated Americans’ views about grammar and usage” (Pullum 2010, 34). The path that King treads in issuing such advice has been well travelled by other authors and his advice is typical of the genre.

3. Research Aims and Methods

A traditional scholar could easily fill a monograph by close-reading the anxiety of prestige in King’s voluminous fiction (over 60 novels and over 200 short stories, as of 2024), paratexts such as author interviews and King’s commentaries on style in On Writing. But in this paper, we suggest less obvious avenues for unearthing evidence of King’s anxiety of prestige, which, while King-specific in method, could inspire future work in larger popular fiction corpora.

We explore how the anxiety of prestige may be interpreted by comparing King’s writing advice with his own published fiction. These provide small contributions to, specifically, King studies; how did King’s stylistics change over a 50+ year career, and did King actually follow his own advice? But we also hope that our corpus stylistic experiments, applying a mixed-methods approach of close and quantitative or distant reading (Herrmann 2017), may provide models for the study of the anxiety of prestige in popular fiction more broadly.

We first examine the frequencies of word patterns based on King’s advice for writers to avoid: first adverbs, then “Swifties” (adverbially modified dialogue attribution), then the passive voice, all queried in King’s own fiction and comparison corpora. The method is simple corpus query via regular expressions using two widely-used corpus query platforms that pre-process texts by adding part of speech and lemma tags: LancsBox 6.0 (Brezina et al. 2020) and TXM 0.8.1 (Heiden 2010). Both have implemented part of speech tagging using TreeTagger (Schmid 1999), while LancsBox was used in the third experiment because it contains a built-in regular expression for passive constructions. Manual inspection and cleanup of all query results was performed, and visualizations of frequencies were created in Google Sheets.

Martin Eve (2022) presented some initial results and discussion from this paper in The Digital Humanities and Literary Studies, on adverbs in King’s texts. Meanwhile, Hye-Knudsen et al. (2023) have also pursued a very similar research idea, presenting experiments on a corpus of King texts inspired by King’s statements in On Writing, and two of our experiments are identical in aim: experiments on <-ly> adverbs and the passive voice in King’s texts. Hye-Knudsen et al. and we pursue different methods, however, and our papers have considerable divergences; we compare our results with Hye-Knudsen et al.’s in certain sections below.

We note here in the methods section that our query of words and linguistic patterns which King attributes to “good” and “bad” writing cannot necessarily be naively equated with ‘high’ and ‘low’ literary style, but we attempt to interpret these connections. King has been consistently vocal in his advocacy of popular fiction, even if many of his fictions clearly aim for, or achieve, high literary merit; King made a conscious attempt at more literary fiction in the early 90s, especially with Dolores Claiborne (1992), but such efforts to write more “literary” novels has never been consistent in King’s career, and more straightforwardly entertaining fictions by King have sometimes followed more literary ones, and vice versa. One could certainly interpret King’s specific elements of writing advice as genre- or prestige-neutral; advice for writers to simply write better, regardless of literary aim. But we argue below that King’s writing advice can sometimes be read as exhortations to write in an implicitly more ‘high’ literary way, or that King’s own implementation of his own writing advice can be interpreted as evidence of King’s own high literary aspirations. Tracing King’s writing advice against his own works, then, can provide evidence for interpretations of the anxiety of prestige in King’s texts. If the reader is critical of our comparison of King’s notions of “good” and “bad” writing with ‘high’ and ‘low’ literary writing, we agree that the connection is interpretive and far from unambiguous, and return to this question a number of times below.

4. Corpora

We assembled all 73 novels and novellas solely authored by Stephen King up to 2020. We also separated out “Misery’s Return,” a 9,000 word story-within-a-story pastiche of intentionally “bad” genre writing from King’s Misery, which we treat as a distinct comparator text. Exploring questions about King’s distinctiveness meant that we also needed comparison corpora. For these we selected The Brown Corpus of Standard American English as a snapshot of US English from 1961 (Francis and Kučera 1979) and The Freiburg-Brown corpus of American English (FROWN) as a snapshot of 1992 (Mair 1992). We also assembled a Stephen King Fanfiction corpus containing the first 5,000 tokens from all King-inspired stories on Fanfiction.net exceeding 5,000 words (91 stories in total; 455,000 word tokens); the 5,000 word cut off is arbitrary, and is intended to separate fanfictions which evidence a serious attempt at fiction from the short, sometimes free-form fanfictions on the website. While comparing an author to their amateur literary imitators is a useful foil, a second fanfiction comparison corpus was also desirable for reference (Sigelman and Jacoby 1996). We thus also compiled a corpus of Harry Potter Fanfiction (91 texts, first 5,000 word tokens each), chosen simply as a well-known popular fiction which has inspired many fanfictions. As a final baseline comparison, we assembled a corpus of National Book Award-winning novels from 1974–2020 as our high literary fiction corpus (see section 7). We attempted to control for diachronic change in English by selecting only American authors of roughly the same age (within 10 years) as King, nineteen novels total.

5. Experiments

5.1 Experiment 1: “The Road to Hell is Paved with Adverbs”

King emphatically warns his readers to avoid adverbs, which he sees as a sign of timid writing: “[t]he adverb is not your friend” and “the road to hell is paved with adverbs” (King 2000, 138–139). Such prescriptions against adverbs are common in the writing advice genre, which has drawn the ire of Pullum (2015). Assertions to “avoid adverbs” are also problematic, as So (2021, 129) has shown that one of the core stylistic characteristics shared by best-selling and prize-winning fiction is a “syntactical preference” for adverbs, when compared to a corpus of black writing that was excluded from these canons. Given that King’s work is best-selling, then, we would expect his adverbial prevalence to be similar to other best-selling and prizewinning works.

It turns out that, despite King’s pronouncements, this is indeed the case. Ben Blatt has already made a first contribution to this question; noting King’s advice about adverbs, Blatt (2017) queried adverbs in a large corpus of contemporary fiction, including a King corpus of 51 novels, reporting that King scores average in a selection of authors from Hemingway to E. L. James. We expand this inquiry with a larger King corpus and present data per King novel, to trace diachronic adverb frequency, and trace more of the stylistic devices discussed in On Writing. As shown in Figure 1 and Figure 2, there is statistically significant, but not major variation between the reference corpora, King’s texts, high literary, and, surprisingly, fanfiction,¹ and little variation in adverb usage throughout King’s career. Perhaps ironically, King’s lowest frequency of adverbs is in his first published novel, Carrie (King 1974), while the highest use of adverbs is King’s The Girl Who Loved Tom Gordon (King 1999), published just one year before On Writing. This seems inconsistent with King’s opinion that “the road to hell is paved with adverbs”.

Figure 1: Relative frequency of adverbs (per 10,000 word tokens).

Figure 2: Relative frequency of adverbs in King’s texts chronologically (per 10,000 word tokens).

However, these initial results are misleading. As noted by Blatt (2017) and Hye-Knudsen et al. (2023), when King proscribes adverbs, King actually means adverbs ending in <-ly>, e.g. totally, completely, and modestly. This then excludes temporal adverbs and various locative forms. The number of adverbs that are excluded in such filtering vary by author, but Blatt proposes that approximately 10% to 30% of all adverbs are of the <-ly> type (Blatt 2017, 12). In Figure 3 and Figure 4, we show the result of the same query confined to <-ly> adverbs.

Figure 3: Relative frequency of <-ly> adverbs (per 10,000 word tokens).

Figure 4: Relative frequency of <-ly> adverbs in King’s texts chronologically (per 10,000 word tokens).

The data for Figure 3 confirm one of Blatt’s findings: that <-ly> adverbs are significantly more frequent in fanfiction (Blatt 2017, 27), suggesting that King’s and others’ distaste for <-ly> adverbs can be distinctions of ‘good’ vs. ‘amateur’ (or ‘bad’) writing. Consistent with this, <-ly> adverbs are lowest in our ‘high literary’ corpus. Although van Cranenburgh and others cast doubt on the correlation of single stylistic features with literariness measures, this is some evidence that <-ly> adverbs may be a textual marker of low literariness. Our result of 1.159% of <-ly> adverbs in the King corpus is very close to Hye-Knudsen et al. (2023)’s reported 1.11% (with a slightly different King corpus).

Figure 4 expands the corpus and modifies the query presented in Eve (2022), and confirms a new insight into diachronic changes in King’s style also reported in Hye-Knudsen et al. (2023): <-ly> adverbs significantly decline over the course of King’s career, consistent with his advice. It is possible that the changes exhibited over King’s style reflect a broader shift in American fiction or the generic movements with which King is associated. Jack Elliott (2017), for instance, has documented declining adverb usage within a corpus of romance novels over time. However, rather than moving outwards to entire genre study, these results instead also allow us to delve more closely into King’s own anxiety of prestige, specifically in his intentional parody of bad writing: “Misery’s Return.”

In King’s Misery, the violent kidnapper character Annie Wilkes forces author Paul Sheldon to write a new genre story starring her beloved character, Misery, and Sheldon produces “Misery’s Return,” selections of which are spread throughout Misery. Even a cursory first reading of these sections shows a marked increase of egregiously florid or unnecessary <-ly> adverbs: a “stuporously warm West Country kitchen”, “[s]he stood lightly poised,” and “[h]e honked mightily into [the handkerchief]” (King 1987a, 132, 161, emphasis added). Thus, when King parodies bad writing, he augments numerous verbs with an adverbial modifier. King parodying genre writing in this way expresses an anxiety of prestige, with King implicitly placing Sheldon’s true potential as a writer, and King’s own, as above badly written mass fiction.

Hypothesizing why some texts are outliers in adverbial usage should be approached with caution. But it is notable that Dolores Claiborne, King’s nineteenth novel, is the text with the lowest frequency of <-ly> adverbs. This novel was a serious stylistic departure for King and a significant attempt at more literary writing, as discussed below. Dolores Claiborne, the best-selling US novel of 1992, deploys a great deal of phonetic dialect and is written from a single narrative perspective, an unusual feature for King (Smythe 2015). We suggest that here, again, is a marker of King’s anxiety of prestige. Having associated the <-ly> adverb with low, King’s eschews it most in one of his most intentionally literary works.

5.2 Experiment 2: “Swifties,” He Dismissed Quickly

Related to <-ly> adverbs, King urges would-be writers to avoid the “Tom Swiftie”: dialogue attribution with an excessive, absurd, or “purple” (meaning excessive or extravagant) adverb, which eventually took the form of a pun or parody of bad writing. An example of a true, punning Tom Swiftie might be: “‘Pass me the fish,’ Tom whispered, crabbily”. King broadens the purview, though, to include all adverbially modified dialogue attribution: “I can be a good sport about adverbs, though. Yes I can. With one exception: dialogue attribution. I insist that you use the adverb in dialogue attribution only in the rarest and most special of occasions” (King 2000, 140). King illustrates this with:

“Put it down!” she shouted menacingly.

“Give it back,” he pleaded abjectly, “it’s mine.”

“Don’t be such a fool, Jekyll,” Utterson said contemptuously. (King 2000, 140–141, emphasis added)

Query reveals that King has avoided these specific phrases almost entirely in his own writing.² Having decried such adverbial modification under most circumstances, King nonetheless admits that he still occasionally uses the form:

And here’s one I didn’t cut…. not just an adverb but a Swiftie: “Well,” Mike said heartily…. But I stand behind my choice not to cut in this case, would argue that it’s the exception which proves the rule. “Heartily” has been allowed to stand because I want the reader to understand that Mike is making fun of poor Mr. Olin. Just a little, but yes, he’s making fun. (King 2000, 344, emphasis in original)

As a next step, we wished to query Swifties in King’s texts, which could be operationalized in a number of ways. Lessard (1992) designed a Swiftie-generating computer program. Litovkina (2014) writes that more recent examples of Swifties do not strictly require an adverb. While canonical Swifties contain an element of humor, we simply query the basic adverbial construction that King decries. All of King’s examples follow a precise word order: Direct Speech → Noun/Pronoun of the speaker → Attribution Verb → <-ly> adverb. The frequency of this form is shown in Figure 5.

Figure 5: Relative frequency (per 100,000 word tokens) of the Swiftie construction.

These results are consistent with King’s perception of the Swiftie (adverbially modified direct discourse attribution) as a marker of bad writing. King’s fiction and Brown score similarly, the high literary texts use the construction far less frequently, while fan fiction displays a high prevalence. As with adverbs, “Misery’s Return” scores the highest. In King’s case, this is strong evidence that the use or avoidance of the Swiftie construction can be considered a marker of the anxiety of prestige.

A closer inspection of this Swiftie construction in the comparison corpora underscores its association with prestigious, high literature. A number of the National Book Award winners eschew the construction entirely, perhaps an indication that these writers have absorbed the collective (if questionable) stylistic wisdom of the writing guide genre. While examples from fanfiction would raise the ire of a writing teacher – “Vernon boomed happily,” “Carlos yammered ecstatically” – the majority of Swiftie constructions are mostly, by themselves, aesthetically inoffensive and found in many professional comparison texts; it is rather the high frequency of them in fanfiction that correlates with low prestige.

Within King’s oeuvre, this Swiftie construction clearly decreases over the course of his career (Figure 6). King’s earlier, journeyman works employed this Swiftie construction far more frequently, but this decreased over time as he developed the stylistic aesthetics eventually expressed in On Writing. Interestingly, the highest result, The Long Walk, was King’s fifth published novel but first written novel, begun in 1966–67 during his freshman year at the University of Maine (King 2000, 428–432), bolstering the impression that King as a younger man dabbled in the Swiftie, but quickly decreased its usage. The next highest result, The Running Man (King 1982), was also written before King’s first published novel, Carrie (King 1974). The Swifties in these early works are, for the most part, not purple prose – e.g. “said casually”, “said cheerfully”, “thought bitterly” – it is again the frequency which is notable. Some of the Swifties do, however, read as what many would consider bad prose. Twice in The Long Walk (King 1979), direct speech is introduced by “shrewishly”: “Barkovitch screamed shrewishly” and “Garraty said shrewishly”. Similarly, in The Long Walk, King broke his own rule against the use of elevated vocabulary (which Hye-Knudsen et al. (2023) explore further), writing that “McVries said sententiously”; a word that query reveals King never used again. All of this suggests that King formed his disdain for this kind of Swiftie (adverbially modified discourse attribution) very early in his career.

Figure 6: Relative frequency (per 10,000 word tokens) of the Swiftie construction in King’s texts.

In the frequency of Swiftie constructions, Figure 6 shows that there is a distinct point of division in King’s texts. The break occurs in 1992 with the publication of Gerald’s Game (King 1992b, in May) and the aforementioned Dolores Claiborne (King 1992a, in November). These novels, importantly, were attempts by King to move away from the (inaccurate) label of horror genre writer and write more prestigious, literary works. Although King had previously written works that were narrated in omniscient third-person and that followed a number of characters’ thoughts in each novel via free indirect discourse (with occasional first-person narration for stories within stories, diary entries, etc.), Gerald’s Game and Dolores Claiborne were attempts by King to follow a single character’s voice. Gerald’s Game features a woman who is handcuffed to a bed and must escape, alone with her thoughts, narrated in the third person and eventually first person. Dolores Claiborne goes a step further, with the entire novel narrated in the first-person voice of the eponymous Dolores, a 65-year old widow. In this text, King phoneticizes the speech of the narrator throughout (e.g. “he ast me” for “he asked me”), uses frequent contractions (dropped ‘g’s in <-ing> words: “lookin’”, “givin’”), and vernacular exclamations of “Gorry!”. This “single point of view is a huge change for King,” observes James Smythe, who notes “the semi-phonetic nature of the text” (Smythe 2015). These novels from 1992 also mark a turning point in King’s characterization and portrayals of women. Carol Senf (1998), for instance, has praised the realist psychological portraits of female characters in these novels. Heidi Strengell (2005, 16) further writes that “since the publication of Carrie (1974), King has been blamed for depicting women characters as stereotypes,” but notes that, “especially since Gerald’s Game (1992), he has more consciously concentrated on women, the emphasis shifting from child characters to women characters”. Senf, in a feminist analysis of the two novels, writes that she finds herself “applauding King for the risks he has taken in Gerald’s Game and Dolores Claiborne” and praises his “shift in perspective and his ability to create strong, plausible women characters” (Senf 1998, 105).

The low frequency of the Swiftie construction in Gerald’s Game and Dolores Claiborne and the subsequent decline in this form over the remainder of King’s career can be read as an indication of King’s intensified literary ambitions in these particular novels, and the anxiety of prestige. On the other hand, it could be hypothesized that Gerald’s Game and Dolores Claiborne feature a lowered frequency of Swiftie constructions because, being single-character studies, they have only a small quantity of direct speech. If there is little quoted dialogue, it would follow that fewer Swifties would emerge. But this is not necessarily the case. We estimated the quantity of direct speech in King’s fiction via a simple query: word tokens between left and right quotation marks (Figure 7).³ By this estimate, Gerald’s Game does indeed have the lowest volume of direct speech (4.23%) of any of King’s novels, which makes sense, as much of the dialogue in this novel is presented indirectly in the memories, fantasies, and hallucinations of its protagonist, who is trapped alone in a bedroom. Dolores Claiborne, however, while on the low end of dialogue by volume (10.86%), is slightly higher than a number of other earlier King novels – The Eyes of the Dragon (King 1984), The Tommyknockers (King 1987b) – and is only 1% lower than Cujo (King 1981). This suggests that the frequency of Swiftie constructions in a text by King cannot necessarily be directly correlated merely with lower quantities of direct speech.

Figure 7: Estimate of direct discourse word tokens as percentage of novel, using regular expressions and quotation marks.

This new evidence – low Swifties in novels aiming to be high and literary, and the low Swiftie query result not explainable by low amount of direct speech alone – underscores the close reading impression that Swifties in “Misery’s Return” appear stark and deliberate. The overbaked adverbially modified speech attributions in “Misery’s Return” – e.g. “he whispered strengthlessly” – also do not appear anywhere else in King’s writing.

The question remains, though, as to the extent that King associates such “bad” writing with genre fiction, whether the two are separable, and thus, whether our queries truly reveal an anxiety of prestige, or merely an anxiety of King’s notions of “good” and “bad” writing, that are distinguishable from the style of high, prestigious literature. First, in On Writing, King frames his disdain of Swifties by noting their historical origin in juvenile genre fiction and dime novels (King 2000, 125–126). Second, it is at a point where King veers away from his own generic stylings that the Swiftie construction declines, giving evidence of a conjunction of high prose style with new high literary genre modes. This is complicated, though, by the fact that even when King later returns on occasion to generic horror writing after 1992, the Swiftie construction is nonetheless used less and less often. The interpretation we suggest is that while King initially and historically associates Swifties with “bad” writing within generic moods, after 1992, even when returning to various genres, King aims for a higher literary prose style.

5.3 Experiment 3: The Passive Voice Should Be Avoided

In On Writing, King exhorts the would-be writer to avoid passive verbs, which he contends are “weak”, “circuitous”, and “frequently tortuous, as well” (King 2000, 122). As with his warning against adverbs, King hedges this advice, specifying that he

“won’t say there’s no place for the passive tense. Suppose, for instance, a fellow dies in the kitchen but ends up somewhere else. The body was carried from the kitchen and placed on the parlor sofa is a fair way to put this, although ‘was carried’ and ‘was placed’ still irk the shit out of me” (King 2000, 122).

Nonetheless, King’s opinion is clear: overuse of the passive voice is characteristic of bad writing.

Such warnings against passive verbs are a staple of twentieth-century writing advice, from Edwin Woolley in 1907 via George Orwell through William Strunk (Zwicky 2006). However, as Pullum (2014, 61) notes, “there is rampant confusion about what ‘passive’ means linguistically”, as “contrary to popular belief, passives do not always contain be and do not always contain a past participle”. Pullum sternly admonishes writing advice authors for their “extraordinary level of ignorance of simple facts” and laments that “the state of the general public’s education regarding the notion ‘passive voice’ is nothing short of disastrous” (Pullum 2014, 64, 67). King at least provides correct examples of passive verbal phrases, unlike many of the writing advice offenders castigated by Pullum. But King, like most of his writing advice forebears, means be verbal phrases when stating “avoid the passive”, and his examples of bad passive phrases in On Writing fall into two categories: future tense (e.g. “the meeting will be held at seven o’clock”) and past simple (e.g. “the body was carried from the kitchen”). Querying and classifying the tense of passive verb forms in the Brown Fiction corpus suggests that past simple passive verbs make up the large majority of passive verbs found in fiction, and that future tense passive verbal phrases are rare (Table 1).⁴

Table 1: Passive verb forms in Brown Fiction corpus.

Passive verb forms	Brown Fiction
Present Simple	63
Present Continuous	0
Present Perfect	34
Past Simple	700
Past Continuous	1
Past Perfect	154
Future	0
Future Perfect	0
Total	952

As a next step in investigating whether the types of passive verbal phrases that King warns against display variance in King’s fiction and are observably more frequent elsewhere, we queried passive be-verb constructions in the corpora (Figure 8) and the trend over the course of King’s writing career (Figure 9).

Figure 8: Passive verbal phrases (with word forms of be), per 10,000 word tokens.

Figure 9: Passive verb forms in King corpus, per 10,000 word tokens.

These results show a low variance in use of be passive phrases in texts as disparate as National Book Award winners and Harry Potter fanfiction, suggesting that despite the common advice to “avoid passives,” they remain a widespread feature of English writing, as Pullum (2014) suggests, and a poor indicator of differential literariness. Furthermore, although there is a steady and marked decline in be passive use over the course of King’s career, it is hardly substantial, and some of the later texts feature significantly more passives than a number of the earlier books. Our results align with the results of Hye-Knudsen et al. (2023), who formulated this experiment by comparing passive to active clauses, and similarly show a slight decline in King’s career over time. Together, this is evidence that passives, in general, do not seem to serve as good indicators of high and low literary language.

6. Conclusion and Future Work

This paper has introduced a term, the anxiety of prestige, along with a proposed definition, above, to serve as a starting point in the analysis of a still largely unexamined phenomenon in literary history: textual responses by widely-considered ‘popular’ fiction authors to issues of literary prestige.

Our experiments provide contributions to King studies, in particular, the nascent digital King studies (see Blatt (2017), Van Cranenburgh and Ketzan (2021), Hoover (2021), Hye-Knudsen et al. (2023), Dorothy Henriette Modrall Sperling (2024), Ketzan (2024)), but also hope to contribute to future investigations of the anxiety of prestige in popular fiction broadly. Digital humanities may be well suited to this task, most simply in the location of textual thematic evidence in larger corpora, but also, as we have attempted to show, through corpus stylistics. Future work could also attempt to locate veiled or explicit antagonism to the act of criticism itself (Eve 2016) within popular fiction, perhaps through suggestions by narrators or characters that books should not be “dissected” through critical theory, but merely enjoyed.

7. Data Availability

Due to copyright restrictions, the full corpus cannot be made available publicly. Frequencies and results of queries can be accessed at https://github.com/erikannotations/King_data.

8. Acknowledgements

The authors would like to thank the editors, peer reviewers, and audience of the CCLS 2024 conference in Vienna for their many insightful comments.

9. Author Contributions

Erik Ketzan: Conceptualization, Writing - original draft

Martin Paul Eve: Writing - original draft