1.1 Xenographic Annotations: Unequivocal in Theory – Partly Ambiguous in Practice

The discourse tradition of xenographic annotations is made up of a multitude of sub-traditions which are influenced by different cultural environments, schools of thought, assumptions about the duties of a scholarly editor, conjectures about readerly needs, etc. But even though xenographic notes are an “extremely complex, multifaceted genre that resists definition” and are characterised by their “versatility and elusiveness” (Enenkel and Nellen, “Introduction” 59), they are nevertheless governed by a few conventions that – at least in theory – pertain to the discourse tradition as a whole.¹ Firstly, such notes have four main (and partly overlapping) functions:

• (1) emendation and textual criticism, i.e. trying to establish the correct text, justifying one’s textual choices, and recording variants;²

• (2) explanation and information, which comprises annotations that strive to elucidate the text, notes that show readers how they can make practical use of the text in their own life (e.g. by imitating its rhetorical style in their own writing or by following its moral lessons), and annotations that take the text as a starting point to provide readers with different pieces of information and advice that the annotator perceives to be interesting or worth knowing, even if they are not immediately relevant for understanding the annotated text;³

• (3) interpretation;⁴ and

• (4) evaluation, i.e. passing scholarly, moral, or aesthetic judgment on the text, defending the text, anticipating as well as reacting to criticism against the text, and implicitly conferring canonical status on the annotated text by presenting it as worthy of annotation.⁵

Often, a single note performs more than one of these functions. Secondly, xenographic annotations are, in theory, informed by a notion of sound scholarship. In other words, readers can be more or less certain that xenographic annotators provide them with information that they believe to be correct and to be sufficient as well as relevant for the respective target audiences of their editions.⁶ Furthermore, readers can assume that xenographic annotators do their best to make their commentaries as comprehensible as possible. Thus, even though xenographic annotations often present and discuss the different possible interpretations that the annotated text gives rise to,⁷ they themselves usually strive (and sometimes fail) to bring across their point as unequivocally and straightforwardly as possible.⁸ Even if a commentary violates any of the three criteria of adequacy (neither too much nor too little information), correctness, and intelligibility, readers can presume that the respective xenographic annotator acted in good faith and did not willingly mislead them or waste their time. Of course, what counts as (sufficiently) relevant, correct, and intelligible depends greatly on the historical and scholarly context in which an annotation is written.⁹

Thirdly and lastly, the enunciatory, temporal, factual, and (para)textual status of xenographic annotations may be assumed to be rather unambiguous. Readers can be almost certain that the voice speaking in the annotation is that of the real-life editor, that the note should be read as a later comment on an existing text, and that the annotation is part of a factual paratext that exists completely outside of the text that it is commenting on. Taking together the four main functions of xenographic annotations, the notions of sound scholarship and ‘good faith’ that inform such annotations, and their unequivocal status, it can be concluded that, in principle, (1) xenographic annotations are themselves quite unambiguous with regard to their purpose(s), (2) that, for the sake of intelligibility, the rhetorical strategies employed in them are designed to avoid ambiguity, and (3) that they generally strive to elucidate – or even reduce – the ambiguities of the annotated text rather than add completely new ones to it.

In practice, however, things sometimes lie differently. For instance, rather than providing readers with information that is immediately relevant for understanding (or making practical, moral, etc. use of) the text, editors may employ their notes for purposes of social networking with friends, relatives, patrons, and other scholars. Prominent fifteenth-century commentator Filippo Beroaldo, for example, often uses his annotations to extol the influential Bentivoglio family and to introduce personal anecdotes, e.g. about his recent marriage and his wife’s pregnancy (cf. Casella 662; Gaisser, “Teaching Classics in the Renaissance” 8; Krautter 44–52).¹⁰ In fact, Renaissance commentators often take a prominent, overt role in their own annotations and provide a wealth of information about themselves that has little to do with the annotated text or at least their scholarly qualifications (cf. Céard 104–05). In other cases, editors take the work that is to be annotated merely as a pretext to present their own, more or less unrelated, research (cf. Sluiter, “The Violent Scholiast” 193; Dubischar 560). In all of these examples, the function and maybe even genre of the annotations is ambiguated: are they scholarly commentaries, or – depending on the case – attempts at securing patronage, personal memoirs, or independent academic treatises?

On a related but still different note, editors sometimes offer (or are forced to offer) deliberately incomplete commentaries. For instance, in 1939, Ernst Beutler was urged by his publisher to omit all references to the Talmud in his commentary on Goethe’s Faust (cf. Bohnenkamp 129), and Renaissance scholar Cristoforo Landino often abstained from commenting on the homoerotic elements in Horace’s poems (cf. Stadeler 124–29). Most Renaissance commentators of Catullus also chose not to annotate obscene words and bawdy passages (cf. Stadeler 140–41). In these cases, editors are being led more by considerations of what may be annotated than what should be annotated because, objectively speaking, it is relevant for a better understanding the text at hand. Thus, the social and historical context in which an edition appears often results in annotators consciously violating the scholarly soundness of their xenographic annotations.¹¹

Furthermore, the enunciatory status of xenographic annotations is not as unambiguous as it may seem at first sight. Sometimes, editors blur the boundaries between authorial and xenographic annotation by presenting themselves as both the alter egos and the collaborators of their authors, e.g. by pretending that the original writers are speaking through them (cf. Céard 108; Sluiter, “The Dialectics of Genre” 191). Commentators’ voices are further ambiguated by the fact that they usually incorporate material from earlier editions, thus letting both themselves and their predecessors speak at the same time (cf. Kraus 16).

To conclude, even in the case of xenographic annotations, practices are much more ambiguous than discourse conventions and functions suggest in theory. Nevertheless, these conventions and functions provide a set of criteria on the basis of which one can judge editorial notes from a purely scholarly point of view. For instance, it is possible to call Beroaldo’s digressive notes ‘irrelevant’ with regard to the annotated text and Beutler’s censored annotations ‘incomplete’, and it would be rather difficult to refute this line of argument. In other words, xenographic notes may certainly violate the conventions of their discourse tradition, but this violation always entails that the annotations are ‘flawed’ to some extent since they fail to fulfil their scholarly functions.¹²

In literary self-annotations, this is not necessarily the case – unless they signal that they are meant to be judged by the rules that govern xenographic ones, i.e. that the author strives to provide exactly the information that a competent scholarly editor would supply.¹³ But even in such cases, there is a difference between self-annotations and xenographic annotations: when xenographic notes are found lacking from a scholarly point of view, this only has consequences for how we assess the editor’s expertise. But when the same occurs in self-annotations that strive to be scholarly, it has a negative impact on how we evaluate the entire work because it raises questions about the author’s knowledge and competence in writing about a certain topic. For instance, Byron’s poem “Lachin Y Gair” stresses his Scottish heritage and the fact that he spent a great portion of his childhood in the Highlands. However, in an annotation he incorrectly translates the term “pibroch” as “bagpipe” – a mistake for which he was ridiculed by reviewers¹⁴ and which casts doubt on Byron’s Scottishness, the very aspect around which the poem revolves (cf. “Lachin Y Gair” 31n; CPW 1: 373).

Byron’s note on the pibroch was a self-annotation that set out to imitate a xenographic one and, thus, to adhere to the rule of correctness.¹⁵ It inadvertently violated this rule and hence can be seen as a failed note. However, many of Pope’s and Byron’s annotations explicitly or implicitly signal that they deliberately strive to transform, flout, or undermine the conventions of xenographic annotations for literary, social, political, or other purposes. In these notes, the ‘failure’ to follow the rules of xenographic annotation is not a flaw but a strategy. A strategy that often draws on the inherent ambiguity of the discourse convention of self-annotation and that, in turn, frequently results in new ambiguities.

1.2 Why the ‘Self’ in Self-Annotation Matters, Or: Hotbeds of Ambiguity

This line may puzzle the [future] commentators more than the present generation. – Byron, Don Juan

Alluding to a verse of Mr. Dryden’s not in Mac Flecno (as it is said ignorantly in the Key to the Dunciad, pag. l.) but in his verses to Mr. Congreve: ‘And Tom the Second reigns like Tom the First’. – Pope, Dunciad Variorum

If we found the first note among the editor’s commentary in a scholarly edition of Byron’s Don Juan, we would most likely feel confused or annoyed, and conclude that the editor must have fundamentally misunderstood his or her task. Why annotate a line that allegedly does not (yet) require annotation in the first place? Why speculate on what future scholars and readers will not understand? Why not use your superior knowledge and just spell out the inside joke for the benefit of posterity? Worse still: why alienate those contemporary readers to whom the meaning of the line may not be clear after all? Penned by an editor, this note would violate several of the conventions that govern xenographic annotations, such as relevance with respect to the annotated passage, helpfulness, and clarity. At best, the annotation could serve as an interesting case study of editors who neglect their task to crack jokes and who try to establish a relationship with those privileged readers who already share the editor’s own horizon of understanding. Readers could not even be sure that the allusion that this editor hints at is actually present in the annotated lines – the rather unprofessional xenographic annotator might simply be misinterpreting the text.

By contrast, in the second quote – taken from an annotation on Alexander Pope’s Dunciad Variorum – the author of the annotation appears to have done a pretty good job. At first sight, the note seems to offer a helpful, trustworthy, and easily comprehensible explanation of an intertextual reference, despite being rather rude to a previous editor of the work. If it was a xenographic note, we would have no reason to assume that its author was being ironical or even deliberately misleading; our experience tells us that this is not something that most scholarly commentators are prone to do. However, these two notes were, of course, not written by professional editors but by Byron and Pope themselves (cf. Byron, Don Juan 12.37n; CPW 5: 754; Pope, 1729 Dunciad 1.6n, original emphasis). The fact that the two notes quoted above are self-annotations rather than xenographic ones makes them appear in a completely different light.¹⁶

In the case of Byron, the ‘unhelpful’ note is not the result of an editor’s eccentric and self-absorbed approach to annotating – in which case the note could very well simply be ignored by readers who are trying to make sense of the lines. Rather, its apparent irrelevance is part of the author’s strategy and, thus, an integral component of the meaning of the passage. Byron’s note is appended to a segment of the poem that muses on the fact that some ladies marry “him who scarce pursued at all. / A hazy widower turn’d of forty’s sure” (Don Juan 12.37). The very fact that his annotation cheekily engages readers in a satirical guessing game and that it behaves so differently from a xenographic note adds to its rhetorical force and humour. Byron’s refusal to spell out the allusion – infuriating as it would be in an editorial comment – here serves various social and literary purposes. By its mere existence, the annotation acknowledges that the passage is not meant as a general comment on the marriage market but that it refers to one or more specific incidents. Its ostensible irrelevance points to its actual relevance. Instead of naming names (which would have seemed rather tasteless and might even have caused a duel),¹⁷ the annotation trusts that most contemporary readers are able to decipher the meaning by themselves. Given the scandals that the two possible referents¹⁸ of Byron’s allusion caused shortly before this canto of Don Juan was published, it is likely that many readers were indeed able to understand the joke. The (at first sight) uninformative annotation thus creates a sense of intimacy between Byron and many of his readers, a feeling of ‘us vs. them’ – ‘them’ being the poem’s future readers and those contemporaries who are not in on the joke. This explicit and teasing exclusion of certain readerships, which would be deemed rather condescending and unprofessional in an editorial note, makes for an entertaining and clever literary strategy in an authorial one. Byron’s note invites readers to participate in the meaning-making of the poem, allows a part of them to congratulate themselves on sharing his horizon of understanding, and, perhaps more seriously, makes a point about the ephemeral nature of even the greatest scandals and most widely circulated gossip. Furthermore, by not spelling out the allusion, Byron also allows his readers to decide for themselves how biting and risqué the satire in this passage eventually is. If they disambiguate the lines as referring to the Keppel-Coke marriage (see note 18), the stanza is merely concerned with a rather extreme though ultimately harmless example of a May-December relationship. If, however, the passage is read as a hint at the Hanson-Portsmouth marriage (see note 18), the satire becomes much more serious – a glimpse at some of the darker aspects of the high-society marriage market: fraud, violence, adultery, legacy hunting, and ruthless social climbing. Due to the apparent unhelpfulness of the annotation, the blame for this latter disambiguation is shifted from Byron to the readers. He intimated no such thing; if they want to interpret the passage in this way, it is their own fault. Lastly, the fact that the annotation prefers to remain silent on the background of the passage may also be a self-ironic reference to Byron’s wish to forget or gloss over his own role (marginal though it was) in the disastrous Hanson-Portsmouth marriage (see note 18). In short, the seeming failure of the annotation contributes to its actual satirical effectiveness; it informs by refusing to explain and invites dangerous interpretation without explicitly endorsing it.¹⁹ Furthermore, the – at first sight – irrelevant note is still helpful for future readers: by drawing their attention to the fact that there is indeed an allusion that could be explicated, it encourages them to do their own research and solve the puzzle.

In the case of Pope, the ‘intertextual’ annotation that would have seemed helpful and trustworthy if written by an editor becomes suspicious when one knows that it was composed by Pope himself. At the time when readers encounter this note, they have already made their way through dozens of pages of prefatorial matter (likewise composed by Pope) that teem with ironies, quotes that deliberate misrepresent Pope’s enemies, and just plain fabrications (like the fiction that the Dunciad Variorum is edited by a man called Martinus Scriblerus). Put briefly, even this early on in the footnote apparatus, readers know that they have to take every piece of information provided in the Dunciads²⁰ with a grain of salt. And even if one were to consider the note in isolation, its context and readers’ knowledge about Pope’s political affiliation strongly hint at the fact that this innocent explanation should not be taken at face value.

Appended to the lines “Say from what cause, in vain decry’d and curst, / Still Dunce the second reigns like Dunce the first?”, Pope’s annotation claims that the passage is merely a reference to Dryden’s poem “To My Dear Friend Mr Congreve” (1729 Dunciad 1.5–6). This poem indeed contains the lines “But now not I, but poetry, is cursed; / For Tom the second reigns like Tom the first” (Dryden, Works 4: ll. 47–48). Dryden is here alluding to Poet Laureate and Historiographer Royal Thomas Shadwell being succeeded as Historiographer Royal by Thomas Rymer. (Before he was dismissed in the aftermath of the 1688 Revolution, Dryden himself had held both of these offices.) According to the note, then, the annotated passage in Pope’s satire either ridicules dunce and Poet Laureate Laurence Eusden, who succeeded Laureate Nicholas Rowe in 1718, or Lewis Theobald, whom the three-book Dunciads depict as succeeding Elkanah Settle as king of the dunces. Both of these interpretations are plausible, given that Eusden, Theobald, and Settle are recurring victims of Pope’s jokes throughout the Dunciads. While both of these interpretations are insulting, they are ultimately rather harmless. However, contemporary readers would immediately have discovered a much more dangerous meaning in the lines, one that is not at all addressed in the note. The first version of the Dunciads was published in 1728, just one year after George II had succeeded George I. Is Pope – the highly suspicious Catholic Tory-sympathiser who had extolled the last Stuart queen and who counted several Jacobites among his friends – cursing the ruling House of Hanover? The annotation pretends to forestall such an interpretation, while ironically drawing attention to how forced its own ‘innocent’ disambiguation is, given Pope’s background and the rather obvious topical allusion.²¹ The annotator doth protest too much; the disambiguation becomes a mock-disambiguation, and all three meanings – the two innocuous ones and the dangerous one – are maintained.²² In a scholarly editor, such an absurdly restrictive ‘disambiguation’ would be seen as the result of mere ignorance and a failure to understand the passage and its historical background. In a self-annotator, however, it is deliberately ironic and serves a satiric strategy. Readers are aware that scholars can never fully reconstruct authors’ intentions, opinions, background knowledge, etc. but that authors themselves obviously know what they are trying to say in a passage, what allusions they are making, and what kinds of background information they are drawing from. If their annotations overlook rather evident allusions, propose nonsensical interpretations, and plead general ignorance of a text’s meaning, readers know that they are confronted with an interpretative cat-and-mouse game rather than a scholarly disaster.

The two annotations by Pope and Byron use very different rhetorical means – Byron’s pretends to be irreverently uninformative, whereas Pope’s feigns disambiguation. The aims of both annotations, however, are rather similar. Both invite readers to decide for themselves whether they want to acknowledge the dangerous subtext of the respective poetic passages. Both also rely on assumptions about (contemporary) readers’ background knowledge without which this subtext cannot be uncovered. But most importantly, both annotations can only achieve these aims because readers are aware that they were written by the author rather than a scholarly editor.²³ It is because of their authorship that they must not be seen as irrelevant, incorrect, incomplete, or plainly failed but as creative ways of inviting readers to discover the satiric import of the annotated passages themselves.

The comparison between xenographic and authorial annotations can be summed up as follows: while xenographic notes usually determine meaning(s) and either simply register existing ambiguities or disambiguate them altogether, authorial ones proliferate meanings and introduce completely new ambiguities. In other words, in xenographic annotations the use and creation of ambiguity can generally be seen as a flaw or at least a strange aberration, but in authorial annotations ambiguity becomes a strategic device. This device often relies on the fact that self-annotations are inherently ambiguous themselves. For one, there are often two different possibilities of describing the temporal status of authorial annotations: are they synchronic asides embedded in the temporal frame of the main text – similar to parentheses? Or are they diachronic, later comments on an existing text? Furthermore, self-annotations are ambiguous with respect to the question whether they are (1) factual explanations of a fictional text or (2) themselves (partly or entirely) fictional or (3) a factual sign that the annotated text is to be read as (partly) factual as well.²⁴ Closely related to this is the question of their (para)textual status: are self-annotations really paratextual features, i.e. external to the main text, or are they an integral part of it?²⁵ Lastly, there is an ambiguity of voice or, put differently: an enunciatory ambiguity. Who do we ‘hear’ in an annotation? The real-life author, a fictionalised version of that author, the narrator (who may likewise be aligned with the author or not), or an entirely new fictional character?²⁶ And, in the two latter cases, in what relationship do these fictional(ised) voices stand to the author – are they mouthpieces, antagonists that are being ironised, or maybe something completely different?

To make matters even more complicated, many works deliberately obscure the authorship of their notes and either pretend that the author’s notes were written by someone else or that someone else’s notes were written by the author (cf. Venturi, “Introduction” 20–21). Erasmus’s Praise of Folly is a famous example of the first case; it claims that the notes were composed by Gerardus Listrius, but they are generally believed to be Erasmus’s own (cf. Griffiths 108; Slights, “Edifying Margins” 710–11). In Byron’s and Pope’s works, we are sometimes confronted with the second case, i.e. annotations written by someone else but presented as if they were the author’s own. This goes for the annotation on the setting of Lara that Byron had asked his friend John Cam Hobhouse to write (cf. BLJ 4: 143–44; 146), and, more importantly, for the notes that William Warburton contributed to Pope’s 1743 Dunciad in Four Books.²⁷ After Pope’s death, Warburton signed his annotations in the 1751 edition of the work, but, as Valerie Rumbold points out, his authorship attributions are not entirely trustworthy, given that Warburton had a “vested interest in emphasising his importance to the project” and might have overstated his involvement (editor’s introduction to Dunciad 2).²⁸ Hence, discussing self-annotations often entails discussing which annotations can actually be seen as such.²⁹

Apart from their ambiguous temporal, enunciatory, (para)textual, and factual status as well as their sometimes-equivocal authorship, self-annotations also evoke two different discourse traditions at the same time. On the one hand, authorial annotations still conjure up the scholarly, xenographic model and thus raise certain expectations of what a note may (and may not) do. Hence, even when reading self-annotations, the knowledge about the discourse conventions of xenographic annotation is (potentially) always at the back of readers’ minds. On the other hand, readers are aware that, due to their completely different relationship to the annotated text and its author, self-annotations do not necessarily fulfil the same functions as editorial notes and, hence, are not governed by the same conventions as they are. Thus, readers of authorial notes have the double awareness that these notes imitate the discourse tradition of xenographic notes while at the same time constituting a separate discourse tradition to which the conventions of factuality, unambiguousness, and direct relevance with respect to the annotated lemmata do not necessarily apply.

As shown in the examples from Pope and Byron above, self-annotations usually follow a specific pattern: (1) they explicitly or implicitly evoke one or more functions that a xenographic note would fulfil and then (2) use various textual strategies in order to (3) either indeed accomplish these functions or to actually perform very different ones. Pope’s note quoted above, for example, evokes the xenographic function of explanation (here: identifying an allusion), uses the strategy of ironic and misleading disambiguation, and thereby achieves various satirical aims. As we will see in the course of this study, the degree to which a specific self-annotation still adheres to the discourse conventions of xenographic annotation can vary greatly: some authorial notes follow them so closely that they can be mistaken for those of a scholarly editor, while others overtly transform, flout, and subvert them.

As the examples drawn from Don Juan and the Dunciads above have shown, it would be overly simplistic to just call these latter kinds of authorial notes ‘failed’ annotations or parodies of editorial annotations. Though making fun of the conventions of xenographic annotations is a common strategy in self-annotations, the main function of such apparently bungled notes is only rarely to simply parody scholarly notes. In other words, in self-annotations, the subversion of the discourse conventions of xenographic notes is generally a means to an end, not an end by itself.³⁰

Any given self-annotation may have intratextual as well as socio-pragmatic functions. The former refers to how a note influences the meaning of the annotated passage and sometimes even the text as a whole. It can straightforwardly explain and support the meaning, but more often alters, expands, or even contradicts it. The socio-pragmatic functions are concerned with how the annotation interacts with the world outside the text – serving to portray the author in a certain light, enacting his/her relationship with the public, and fulfilling practical purposes like flattering patrons and insulting enemies. In the Byron example above, the intratextual and the socio-pragmatic functions are closely intertwined. With regard to the former, the note points readers to the presence of a topical allusion within the text, thereby emphasising and expanding the meaning of the annotated passage. The socio-pragmatic functions of the annotation are to comment on the high-society marriage market as well as on the ephemerality of gossip, to establish a shared horizon of understanding with contemporary readers, to tease future audiences about their alleged inability to make sense of the passage, and, thereby, to still give them a hint that there is indeed an allusion that can be deciphered provided that they do the necessary research.

While in this specific note all of the functions complement one another, there are also many annotations in which the functions are unrelated or in which they even contradict each other. The latter is the case in the example drawn from Pope’s Dunciads which partly stresses the innocence of the annotated passage but which, by its very silence on the political allusion, also ironically confirms that the lines have a much more dangerous meaning. A note may also have different functions for different readerships (see esp. chapters 2.2.2.3 and 3.2.2). The ‘double-directedness’ of self-annotations at the text and the world outside the text, and the various (sometimes contradictory) functions that a note can fulfil simultaneously add even more ambiguities to this far from unequivocal discourse convention.

To conclude, in xenographic annotations, ambiguity is a bug; in authorial ones, it is a defining feature. Xenographic annotations are usually unambiguous in themselves and strive to determine or at least to record possible meaning(s); authorial ones are frequently ambiguous and often serve to further proliferate meanings.³¹ Self-annotations set out by invoking one or more of the functions of editorial annotation. They then use a multitude of textual strategies to fulfil, transform, or subvert these functions for ulterior aims, thereby interacting both with the annotated text and the world outside it. These strategies and the aesthetic, satirical, social, etc. purposes for which they are used – the how and the why of self-annotation – are at the heart of this study which sets out to provide a systematic overview of Pope’s and Byron’s practices of self-annotation (for more details on the focus of this book, see chapter 1.5.3).

1.3 Existing Attempts at Categorising Self-Annotations

Since the publication of Genette’s pioneering Seuils, a considerable number of studies have stressed the manifold ways in which authors’ annotations interact with, and contribute to, the meaning(s) of their texts.³² Yet, so far only a handful of them have attempted to comprehensively categorise the notes of individual writers, literary movements, or even of self-annotated literature as a whole. Such categorisations, however, are essential for gaining insights into the immense (and perhaps surprising) diversity of self-annotations and for comparing the uses of authorial notes across different authors, genres, and times. Furthermore, even the existing handful of categorisations cannot (for the most part) fully do justice to the complexity of the self-annotations that they are covering.³³

As one of the few attempts to classify self-annotations, Ottavio Besomi’s article “L’autocommento nella Secchia rapita” is concerned with the notes in Alessandro Tassoni’s famous mock-epic – a work that was well-known to Pope (see p. 62 below). Besomi’s essay, however, is mainly concerned with categorising the pieces of information that the notes provide (i.e. what kind of knowledge do they explicitly contain?) rather than the functions that these notes serve (i.e. what do the notes do to the meaning of the work and the author’s interaction with the world outside the text?).³⁴ Furthermore, Besomi’s categorisation partly blurs the line between the purposes of the notes and the textual strategies employed in them (e.g. ‘direct addresses to the reader’).

William Slights in “The Edifying Margins of Renaissance English Books” proposes fifteen functions of marginal notes, regardless of whether they are authorial or xenographic.³⁵ He also acknowledges that a note often does not serve only one of these functions but that “many of the more significant contributions of marginalia to particular texts result from subtle combinations of these purposes” (686–87). Despite the helpfulness of Slight’s list for gaining an overview of the uses of early modern marginalia, I have indicated above why it is problematic to discuss authorial and xenographic notes under the same heading. The very same note, e.g. ‘This is nonsense’ (which, in Slights’s terminology, might be categorised as ‘correction’ or as ‘parody’), can have different functions depending on whether it was written by the author of the annotated text or by someone else. In the case of an authorial annotation, readers would have to come to terms with the apparent self-contradiction and can read it, among other things, as an isolated case of playful self-irony, later repentance for one’s youthful writings, or a sign that the annotated text as a whole should not be taken seriously. If the same note was written by someone else (e.g. an editor), readers can, theoretically, just dismiss the note as subjective and unfounded, with no bearing on the annotated text whatsoever.

Maxine Hancock, in her study of Bunyan’s marginal notes, identifies four functions and two effects of self-annotation. The functions are “to refer, to index, to interpret, and to generalize” (Hancock 123). Like Slights, Hancock stresses that

some of the most complex and interesting interactions between marginal notes and narrative text occur when marginal notes function in more than one way at a time, as when reference notes offer interpretation. (Hancock 133)

The effects of authorial marginal notes can be divided into “text-reflexive”, i.e. notes that “modify, intensify or ameliorate the effect” of the annotated passage, and “text-extensive”, i.e. notes that “invite the reader’s attention to move beyond the narrative” (Hancock 123). Text-reflexive annotations are hence concerned with intratextual patterns and the interpretation of the text, while text-extensive notes are related to signification, representation, and intertextuality (cf. Hancock 134). Hancock aptly shows that even brief marginal notes often serve a number of different purposes, and she likewise recognises the double-directedness towards the text and the world outside the text that characterises annotations. Nevertheless, she does not mention that notes can also be employed for socio-pragmatic functions. The prevalence and importance of such ‘social’ annotations will be shown in the course of the present study.

Andréas Pfersmann’s article “Éléments pour une approche typologique des notes infrapaginales” offers six main functions of self-annotations in general, without focussing on a specific author or genre (cf. Pfersmann 66–88).³⁶ These categories are again comprised of many subcategories. Even though Pfersmann’s approach takes an important step in stressing the social and ‘practical’ dimension of self-annotation, it is still mainly concerned with categorising the information that is explicitly given in a note, thereby neglecting cases in which one or more functions are implicitly fulfilled by the apparent performance of another. For instance, a note may – on the surface – be used to merely inform readers about the presence of an allusion (information), but this information may – depending on the context – be used to reinforce the meaning of the work or, by contrast, to undermine it (function).

Lastly, in his study of Rousseau’s La Nouvelle Héloïse, Yannick Séité lists nine different functions of self-annotations in this novel.³⁷ Later in his book, he also suggests a second way of categorising the annotations in the Nouvelle Héloïse, this time with more focus on the socio-pragmatic aspect of the notes and the occasional antagonism between the notes and the main text.³⁸ Séité’s proposal of two different models indicates how difficult it is to systematise the functions of self-annotations. The complexity of authorial notes, which often pretend to do one thing while fulfilling many different functions in addition to, or even instead of, their alleged aim, is also acknowledged by Séité. Referring to his first categorisation (pp. 292–93), he points out that it cannot account for the actual intricacy of self-annotations:

[I]l est une ultime caractéristique des notes de La Nouvelle Héloïse dont le tableau ne rend pas compte: la complexité de leur structure; le fait que, pour beaucoup d’entre elles, elles ne remplissent pas une et une seule fonction, et, plus encore, ne se limitent pas à un seul effet mais en visent bien souvent plusieurs. […] [S]i le but apparent d’une note est peut être bel et bien à chercher du côté d’un désir de faire passer un savoir, son effet, voire son but réel, peut être d’un tout autre ordre. (Séité 297; 323, original emphasis)³⁹

Even though this is an extremely important insight, Séité’s book overstates the degree of Rousseau’s uniqueness and innovation. For instance, he asserts that Rousseau was the inventor of the complex note and that he was the first and maybe only author to use self-annotations to address personal messages to a small group of insiders rather than to his whole readership (cf. Séité 298; 313). By contrast, I wish to argue that self-annotations in general exhibit a considerable degree of complexity (albeit not always to the same extent as those in La Nouvelle Héloïse) and that Pope’s and Byron’s are among the most intricate examples of the discourse tradition of self-annotations. Furthermore, similar to Besomi, Séité mixes categories: some of his criteria refer to the explicit information contained in the notes (e.g. information about Switzerland), some to the textual strategies (e.g. annotations that tonally clash with the text), and some to actual functions (e.g. preventing readers from becoming too invested in the sentimental letters).

As already indicated, the five approaches just outlined suffer from several drawbacks. Besomi and Pfersmann mainly focus on the information that is explicitly given in an annotation rather than its actual functions, thereby implicitly evaluating self-annotations according to the criteria of xenographic notes. Likewise, in combining authorial and editorial annotations in the same scheme, Slights does not consider the fundamental differences between these two discourse traditions. Slights, Hancock, and Séité recognise that a single authorial note can serve many different functions, but none of them ‘translate’ this insight into a categorisation that considers the complexity of self-annotations, nor do they investigate why and how authorial annotations lend themselves so well to this multifunctionality.

The question is, how can one satisfactorily categorise something that is intriguing precisely because it resists categorisation? The most promising approach seems to be to analyse the textual strategies that lead to the multi-functionality of self-annotations and to describe patterns of functions and strategies that frequently co-occur. I have therefore chosen a tripartite categorisation in this study. First, I sort Pope’s and Byron’s self-annotations according to the function(s) of xenographic annotation that they appear to mimic (e.g. explanatory notes, emendatory notes, or evaluative notes). Then, I structure these notes depending on the main textual strategy (or strategies) employed in them (e.g. using manipulated quotes or employing various fictional annotator personas). Lastly, I categorise these annotations again into the main functions that are served by these strategies (e.g. both reinforcing and disowning the attack of a passage). This may lead, for instance, to the investigation of Pope’s self-annotations that mimic both emendatory and interpretative xenographic notes (see chapter 2.3). In this specific case, only one main strategy is employed, namely the attribution of these notes to two annotatorial personas – Bentley and Scriblerus. The functions that these notes serve are, first, to ambiguate the two personas themselves and, second, to use the annotations signed by these ‘annotators’ to ambiguate both Pope’s Dunciads and his public image as a whole. This tripartite categorisation – mimicked function(s), textual strategies, actual function(s) – frequently results in overlaps and the repeated mention of a single function in various contexts. However, such repetitions are useful in that they show, for example, when an author is especially preoccupied with a certain function and when vastly different textual strategies serve the same function.

Pfersmann’s and Slight’s categorisations of the functions of (self-)annotation derive from a large sample of vastly different texts, while Besomi, Hancock, and Séité all focus on one writer and work, respectively. This makes it hard to use either of these approaches for the aim of learning more about the similarities and differences of authorial notes from different authors, genres, and periods. Furthermore, neither method can fully grasp how variegated practices of self-annotation actually were (and still are) – the former only outlining what this discourse tradition can do in general and the latter showing what one author does but not how this author’s self-annotations differ from those of his contemporaries. For this reason, the present analysis focuses on two writers, Pope and Byron, and embeds their approaches to authorial annotation in the context of the (self-)annotatorial practices of their respective ages. This will, of course, still be insufficient for a comprehensive overview of all the strategies and functions of the discourse tradition of self-annotations as a whole (i.e. irrespective of author, time, or genre). However, the study of these authors will show that even in a case in which two writers wrote in the same language and lived merely one hundred years apart, were often preoccupied with the same genre (satire), and in which one saw the other as his great idol and model, their notes vastly differ from one another, thus attesting to the multifariousness of the discourse tradition of self-annotation. Furthermore, the analysis of their two very dissimilar approaches to self-annotation will allow me to develop broader analytic categories (e.g. the degree to which the notes subvert the main text) that can be employed for comparing annotatorial practices of different authors, genres, and periods.

The close readings of Pope’s and Byron’s notes will be guided by two interconnected theoretical frameworks: the conceptualisation of self-annotation as a discourse tradition which is still informed by xenographic annotations but strategically deviates from its conventions in a multitude of ways (see above), and a fine-grained approach to the phenomenon of ambiguity.

1.4 Self-Annotations and Ambiguity

Ambiguity relates to self-annotation in four ways: (1) as a characteristic of this discourse tradition as a whole, (2) as a starting point for self-annotations, (3) as a strategy in them, and (4) as an outcome of them. These four aspects are located on conceptually different levels. The first point has been discussed above and refers to the highly ambiguous enunciatory, temporal, (para)textual, and factual status of self-annotations, their sometimes-equivocal authorship, and their ambiguous relationship with the conventions of xenographic annotations. Here, ambiguity appears as an inherent property of a discourse tradition. An individual author does not produce this ambiguity but rather employs (and thus: perpetuates) an already-existing ambiguity. The second aspect points to the fact that annotations (be they authorial or xenographic) often react to ambiguities within the annotated text; they take them as a starting point for explanation, interpretation, disambiguation but also – in the case of self-annotation – further ambiguation. In this case, ambiguity is a feature of the annotated text; it is created by the author and can be located in a single word or passage but also in a work in its entirety. Thirdly, self-annotations often use textual strategies that involve ambiguity, e.g. irony, addressing different readerships, performing two contradictory functions at once, proposing mutually exclusive interpretations, etc. Here, ambiguity – or, rather, ambiguation – is a rhetorical and literary device employed for a specific aim. Lastly, and perhaps most importantly, based on the three former aspects, self-annotations create yet further ambiguities. These ambiguities can be relevant on a local level (i.e. with respect to the annotated passage) but also on a global level by complicating how we interpret the whole annotated text and even an author’s entire œuvre and public image. Thus, ambiguity as an outcome of self-annotation touches on many different conceptual levels: it can raise questions about the meaning of a single word in a text; complicate our notion of a work’s genre, tone, and factuality; and even have far-reaching social implications (e.g. when an author uses a facetious note to cast doubt on his usually sombre or morally impeccable self-presentation).

These four aspects taken together can show how different ambiguities located on different conceptual levels may interact with one another. For instance, the very fact that a note is an authorial instead of a xenographic one raises the question of who exactly is speaking in it – is it the real-life author or a fictional annotator persona whose opinion might be ridiculed by the author? This note might then be used to address an ambiguity in the text and offer an elucidation. The enunciatory ambiguity of the note might be employed by the author to ambiguate the irony or seriousness of the note – is the elucidation to be taken as a serious explanation by the author or as a nonsensical proposal by a fictional persona that the author ironises? This ambiguity of the note, then, ambiguates the annotated text by giving rise to various interpretations of the annotated passage, which, in turn, may raise questions about the author’s philosophical and political outlook. (For concrete examples of such cases, see chapter 2.3.)

In order to be able to describe these interrelations between the status of a note, its starting point in the main text, the strategies it uses, and the outcomes it produces, a framework is needed that is detailed enough to precisely describe why a certain textual phenomenon has different meanings, and what these different meanings are. At the same time, this framework has to be flexible enough to be applicable to ambiguity phenomena on different conceptual levels and of different sizes (ranging from a morpheme to a whole discourse tradition) and to cases in which the multiple meanings of one part of the text have ramifications for a completely different part of the text, thereby ambiguating it as well. This framework is based on the approach to ambiguity developed by the Tübingen Research Training Group “Ambiguity: Production and Perception”.⁴⁰

1.4.3 Analysing Self-Annotation Through Ambiguity & Ambiguity Through Self-Annotation

For the purposes of the present study, it is important to describe how exactly ambiguity is created and for which purposes it is used. Both of these aspects are covered by the Tübingen model of ambiguity which covers a great number of parameters that are indispensable for an in-depth analysis of different ambiguity phenomena (for an overview, see Winter-Froemel and Zirker passim; and S. Winkler, “Exploring Ambiguity” passim).

On the most basic level, one can distinguish between ambiguities in the language system and ambiguities in concrete utterances (cf. Winter-Froemel and Zirker 286). The first “represent a characteristic of the abstract linguistic items […] which may – or may not – materialize in human communication” (286). In actual communication, these ambiguities are usually not even noticed because they are disambiguated by context (cf. 286). For instance, the word ‘bat’ is ambiguous in the language system and can either refer to an animal or to an implement used in baseball. However, if a baseball coach were to say ‘we have a game tomorrow, so don’t forget your bat again’, the potential ambiguity of ‘bat’ would, most likely, not be noticed because the context disambiguates the word. Nevertheless, one of the players might joke: ‘I didn’t forget it; it flew to Transylvania’. In this case, the ambiguity in the language system would be actualised in discourse. Yet, ambiguity in concrete utterances does not always rely on ambiguities in the language system. For example, the sentence ‘while you were sleeping, I cleaned the whole flat’ is not ambiguous in the language system. Nevertheless, it is ambiguous because it can be interpreted either as mere statement or as a reproach (cf. 287–88). In a similar vein, self-annotations can, among other things, (1) react to an element in the main text that is ambiguous in the language system but not in this concrete utterance and draw attention to/actualise its ambiguity, (2) address an element in the main text that might be read as ambiguous (either in the language system or in this concrete utterance) and in some way react to its ambiguity (e.g. by disambiguating it), and (3) give rise to ambiguities that do not exist in the language system at all but depend on context.

In the ‘bat’ example above, the coach (who unintentionally utters something ambiguous) and the player (who deliberately reacts to an ambiguous utterance) epitomise two further aspects: (1) the production and the perception side of ambiguity and (2) the difference between strategic and non-strategic uses of ambiguity (cf. S. Winkler, “Exploring Ambiguity” 3). Even though in this concrete situation the coach was not aware of the fact that he uttered something ambiguous, he can be seen as the producer of ambiguity in this example. This means that the moment of the production of ambiguity always refers to “its first appearance in the given context […], no matter whether its first appearance reveals the ambiguous nature of the item” (Hartmann et al. 12). The coach produced the ambiguity, but he did so unwittingly; thus, his utterance can be seen as an example of non-strategic production of ambiguity. Strategy here refers to the question whether the ambiguity of an utterance “serves the function of a means to reach a particular goal in communication” (Hartmann et al. 12). Especially in literary studies, it is often a matter of debate whether the ambiguity of a given textual element was strategically produced by the author (and hence, intended to be understood as ambiguous by readers), or whether the text’s audience is reading more into the work than there actually is. This problem is sometimes satirised in self-annotations, which, among other things, can suggest deliberately outlandish interpretations (thus finding additional meanings in a passage that, at first sight, looked quite inambiguous) or, in turn, claim that an annotated section only has one completely unequivocal meaning (thus feigning ignorance of this section’s actual strategic ambiguity).

The question of interpretation shows that the percipient (i.e. hearer, reader, beholder⁴⁸ etc.) of an ambiguous utterance also has to be taken into account when analysing ambiguity. There are many different ways in which percipients may interact with ambiguity:

• Download Figure
• Download Figure
• Download figure as PowerPoint slide

View Full Size

• Download Figure
• Download Figure
• Download figure as PowerPoint slide

• Download Figure
• Download Figure
• Download figure as PowerPoint slide

View Full Size

• Download Figure
• Download Figure
• Download figure as PowerPoint slide

• Download Figure
• Download Figure
• Download figure as PowerPoint slide

Like the producer of ambiguity, the percipient may thus strategically or non-strategically react to the producer’s (likewise strategically or non-strategically) ambiguous utterance. Non-strategic perception refers to all cases in which percipients are not aware of strategically produced ambiguities (case 1.1) and to cases in which they remain unaware of a non-strategically produced ambiguities (cases 2.1 and 2.3). In all other cases, i.e. when percipients either in some way recognise, and react to, a strategically produced ambiguity or deliberately misconstrue a non-strategically produced ambiguity, they are engaging in strategic perception. Example 2.5 is a special case because here the strategic perception of ambiguity still accidentally leads to a (non-strategic) misunderstanding.

The perception-side is especially intriguing with regard to self-annotations because these contain different kinds of perceptions of ambiguity. Examples include (1) authors’ strategic self-perceptions of the ambiguities they strategically produced in the annotated text, (2) perceptions (both strategic and non-strategic) by fictional annotator personas (e.g. finding absurd additional meanings in the annotated text that are quite obviously precluded by the context), (3) notes quoting actual critics as percipients (who may have found ambiguities that were not strategically produced or missed some that were), and (4) notes that anticipate and try to guide perceptions by real-life critics (e.g. by clarifying the meaning of a certain passage to prevent misunderstandings). In all cases except the third, the author takes on the double role as the producer and the percipient of ambiguity.

An especially fascinating point about self-annotations is that, in their enactment of the perception of ambiguity, they often employ a great deal of irony. For instance, they may claim that a certain passage was by no means intended to be ambiguous and that it should only be read in some way, while the great majority of readers recognise that this passage is actually highly (and strategically) ambiguous. Pope’s ‘innocent’ annotation on “Still Dunce the second reigns like Dunce the first” quoted above is an example of such a note – unconvincingly disambiguating a passage that may both plausibly be read as an intertextual reference and a dangerous political remark.

Such cases highlight the need for an analytic parameter that is not yet included in the Tübingen model of ambiguity: the difference between overt and covert strategic productions and perceptions of ambiguity. ‘Overt’ here means that the producers and percipients of an ambiguous utterance are explicit about their strategic (i.e. conscious and intentional) use of, or reaction to, ambiguity. ‘Covert’ means that they are not. An example of a covert use of ambiguity is (1.2) named above. The producer says, ‘Wow, you’re being really diligent today’, thus strategically producing an ironic utterance⁴⁹ that is ambiguous with respect to the question whether it is a genuine compliment or a hidden reproach. The percipient is aware of this ambiguity but only answers ‘Yes’, disguising this awareness. The producer may then respond, ‘Don’t get me wrong, this was not meant as a criticism. I really appreciate your help’, again disguising the fact that the initial utterance was indeed meant as a reproach. Both producer and percipient here try to hide their strategic use of ambiguity, and – just like Pope in the example above – maintain the ambiguity by pretending to disambiguate. In many of such cases (both in every-day communication and literature) it is difficult and sometimes even impossible to conclusively decide whether an ambiguous utterance was (1) strategically but covertly produced/perceived, or (2) indeed non-strategically produced/perceived. This problem can be termed the ‘ambiguity of ambiguity’– the ambiguity of whether or not something was intended to be ambiguous by a speaker.

Especially in literary texts, questions of perceived/non-perceived, strategic/non-strategic, and overt/covert productions and perceptions of ambiguity have to be answered both with respect to the external level of communication (between author and readers) and the internal level(s) of communication (between narrators and narratees or characters among each other) (cf. Winter-Froemel and Zirker 303–04; 322–23). For instance, authors may strategically produce an ambiguous utterance, while the characters in the work remain unaware of the fact that they utter (or, rather, are made to utter) something ambiguous.

The examples that I have used in this section have been concerned with ambiguities on the word- or sentence-level only. However, the textual elements that can be ambiguous range from a mere morpheme to a whole complex of “thematically, structurally and/or functionally linked texts” (e.g. a whole discourse tradition) (Hartmann et al. 14–15). Likewise, as briefly alluded to in the discussion of Rimmon’s approach to ambiguity, an ambiguity can be local or global; this concerns the question up to which level the ambiguity is relevant (cf. Hartmann et al. 13; Ebert 16). For instance, a single self-annotation, appended at the very end of a literary text and saying ‘but this was all hackneyed nonsense’ has bearings on the interpretation of the text as a whole, ambiguating it with respect to the question whether readers should still interpret it as a serious literary endeavour or rather as mere nonsense or a parody. This question about the range of an ambiguity is often, but not always, intertwined with the question whether the ambiguity is resolved at some point (e.g. through context or through metalinguistic strategies by which the ambiguity is explicitly addressed) or whether it is maintained throughout the text (cf. Winter-Froemel and Zirker 315).

The summary of the Tübingen model of ambiguity has shown why this fine-grained approach is helpful for analysing self-annotations and, in turn, why the field of self-annotations is particularly intriguing when one sets out to study how ambiguity is used in literary texts. Self-annotated works contain both the strategic production and the strategic perception of ambiguity; they show how authors pretend to explain the meanings of their works, how authors react to critics’ interpretations of their works, and how authors try to (mis)guide future readings of their works. Self-annotations are the confined spaces where many different ambiguities interact and even the briefest annotation is able to ambiguate a whole work, discourse, or genre. Pope’s and Byron’s self-annotations use and create ambiguity in numerous highly inventive ways. However, both the strategies in, and the function of, their ambiguous as well as ambiguating notes are still understudied.

1.5 Ambiguous Self-Annotation: The Cases of Pope and Byron