The voice astray: Caroline Bergvall’s Dante

Abstract

This article analyses ‘Via: 48 Dante Variations’ by Caroline Bergvall (1962-) and considers the kind of Dante it performs. ‘Via,’ composed of the first tercet of Dante’s Inferno, accompanied by 47 (or 48) English translations of the tercet ordered alphabetically, was first divulged as a sonic piece and was then published in diverse textual iterations, with variants among the several written forms and between the spoken and written. I explore the links between Bergvall’s project and medieval textual practices and textualities, particularly the creative instability of works that exhibit mouvance and variance within a hybrid aural/textual culture. Engaging with recent critical work on queer philology and with Ovid’s myth of Echo, I consider how Bergvall’s ‘Via’ can be said to perform error, extending the predicament of the Dantean ‘I’ at the opening of Inferno, and how that performance of error may be a source of queer pleasure.

I had started this piece by accident. Stumbling upon Dante’s shadeless souls on my way to other books. Perhaps following a lead, in the dark of dark, the wood of woods, in the sense of panic of the opening canto […]. (Bergvall 2005, 64)

What I put into words is no longer my possession. Possibility has opened. (Howe 2007, 13)

At its heart, this article is a study of errancy and of the possibilities it opens—textually, poetically, and critically. The focus of my analysis is ‘Via: 48 Dante Variations’ by the acclaimed poet, sound artist, and performer Caroline Bergvall (1962-) and its relationship to Dante’s Commedia and the world of medieval textuality more broadly. My enquiry is motivated by a series of questions: What kind of Dante does Bergvall’s ‘Via’ perform? How might we respond to it? And what can Bergvall’s practice in ‘Via,’ in its overlap with medieval modes of creating and transmitting texts, tell us of these imbricated poetic cultures? To this end, I explore Bergvall’s activity as a copyist and compiler of pre-existing textual material, who operates in a multimedia environment. The article considers two main points of interaction between the medieval and contemporary: firstly, how ‘Via,’ in its plural instantiations across the oral/aural and written domains, exhibits the sort of creative instability characteristic of medieval texts that were subject to mouvance and variance; and secondly, the propensity of the work to stray further through Bergvall’s praxis as an artist, which allows us to think Dante, and to think with Dante, differently.

According to Bergvall, as cited above, ‘Via’ began in error when a chance textual encounter, construed as a detour on her ‘way to other books,’ brought her into contact with ‘Dante’s shadeless souls.’ What ensued was a momentary interruption in her journey (‘stumbling’ could imply tripping over the text as much as encountering it) and an alternative direction of movement, the errant pathway having become strangely alluring even though, or perhaps because, it was the way of panic and digression amid Dante’s dark wood (and the reference to ‘panic’ suggests that the infernal landscape is also a soundscape for her.¹)

‘Via’ itself offers an even more immersive encounter with this space through an intricate textual collation, which is also voiced. The text begins with the first terzina of Dante’s Inferno:

Nel mezzo del cammin di nostra vita

mi ritrovai per una selva oscura

che la diritta via era smarrita.

(Inferno I, 1–3, cited from Bergvall 2005, 67)

This is accompanied by forty-seven or forty-eight translations of the tercet in English, ordered alphabetically into a series. We must stipulate forty-seven or forty-eight since the number varies across versions. ‘Via’ is crucially not a single, authorial, text but a composite of performed pieces bridging sound poetry, audio recording, and textual transcription.

In its status as a multiform ‘textual object’ and ‘audiographic’ work (Bergvall 2011c, 244), ‘Via’ is exemplary of Bergvall’s hybrid and experimental modes of production and transposition of ideas across different media. It sits within a remarkable constellation of pieces in which Bergvall has engaged with premodern texts and authors: for example, with Chaucer’s Canterbury Tales in Meddle English (Bergvall 2011a) and Alisoun Sings (Bergvall 2019); and with Old English poetry—the anonymous, tenth-century poem, Seafarer—in Drift (Bergvall 2014), interwoven in that work with a report on the 2011 incident in which a migrant vessel from Libya, bound for Sicily, was left to drift unaided for several days in a busy shipping channel, with those on board left to die. Combining text and image, Drift reconstructs Bergvall’s experience of engaging with the Old English text and its unfamiliar language through a process of quasi-translation. This process is experienced viscerally and creates a sense similar to ‘Via’’s incipit in the pull that language, with attention to its innate plurilingualism, has on her: ‘Calling up the messengers, root-words, stem-sounds, flowering lines that have nurtured and kept the voices that now have me and egg me on into the story’ (Bergvall 2014, 128–29).

As we shall see, like these other creative pieces by Bergvall, ‘Via’ is a striking example of the ‘contemporary medieval in practice’ insofar as it ‘suggest[s] a different kind of research about or conversation with the past: one where we can hold medieval and modern in flux and reciprocity, where the past and present source each other, where the medieval and the modern interact, flow, cross and pass’ (Lees and Overing 2019, 49). Bergvall’s works repeatedly unsettle normative and restrictive categories, including those to do with genre and monolingualism, encouraging her audience to confront and explore indeterminacy and unknowability (see Teich 2020). In the light of this permeability, her work has received critical attention from scholars in a broad range of disciplines: translation studies, conceptual writing and experimental art, gender and queer studies; and interpreting her output demands moving quite fluidly between these disciplinary areas, thinking about where they may intersect.² ‘Via’ is no exception and in what follows I seek to navigate a course from my position as a scholar of Dante and medieval Italian lyric whose perspective on Bergvall—and Dante—has been greatly enriched by the scholarship on conceptual poetry, error, and queer aesthetics that I reference below.

I begin by outlining the different versions of ‘Via’ (sonic and written) on which I concentrate to offer readers less familiar with Bergvall’s piece a point of entry into its fascinating byways. I then situate ‘Via’ in relation to the hybrid aural/textual culture of the Middle Ages and interpret it through mouvance and variance, concepts that medievalists have employed to analyse works that wander between diverse contexts of performance and exist in multiple (textual) iterations. The ‘voice astray’ of my title alludes to the relevance of this medieval context for understanding Bergvall’s contemporary endeavour and, in particular, her sophisticated, meta-textual, performance of the errancy that animates the opening tercet of Dante’s Commedia. This is enacted, I argue, through a repeated and insistent straying (from the Latin extrāvagāre, meaning to wander out of bounds) that resists rectification and is conveyed through the resonating space of Bergvall’s voice and the role of her hand in production. If traditional philological practice could be said to fetishize the source text in its search for an original, ‘Via’ resists analysis through that paradigm and harnesses those pleasures ‘elicited by errors, deviances, and gaps,’ which are associated with ‘queer philology’ (Magnani and Watt 2018, 259). A final section brings Bergvall’s practice of error into dialogue with two figures of linguistic deviation that map the possibilities of a voice which moves outside authorial possession: Ovid’s Echo, and her poetic correlative in Bergvall’s work—the mockingbird.

Introducing ‘Via,’ in multiple versions

‘Via’ was first performed as a sonic piece at the tEXt02 festival in 2002. A recording of the work had been made in collaboration with the sound artist Ciarán Mahler in summer 2000 and is available online.³ ‘Via’ appeared in print in 2003 in the journal Chain, in a special issue co-edited by Jena Osman subtitled Translucinación, which explores the creativity of translation and its dialogic character (Bergvall 2003a). ‘Via’ was later re-published, in revised form, in Bergvall’s Fig: Goan Atom 2 (Bergvall 2005). There it is accompanied by a prefatory, compositional note and appears alongside several other of Bergvall’s works that similarly originated as installation or performance pieces and were after adapted to the written medium (Bergvall 2011c, 239). While the prefatory note is in prose, the text of ‘Via’ is set out as verse and comprises the first terzina of the Inferno, copied in Italian, and forty-seven translations of Dante’s terzina in English, ordered alphabetically by first word and numbered according to their position in the sequence. Bergvall explains that these translations were transcribed from the holdings of the British Library in the year 2000, thereby setting her enterprise exactly seven hundred years after the fictional date of Dante’s journey into the otherworld (Bergvall 2005, 64). Each translation is followed by the surname of the translator and the year of publication of the edition from which the translation was taken. The excerpt from Dante’s poem is—by contrast—reproduced without an edition, with no punctuation, and with quite a strange attribution. Inferno would normally be referred to as the first canticle, not as ‘Pt. I’ of the poem. This anomaly is just one example of how Dante’s poem is being reframed in ‘Via.’

The ‘Via’ we find in Fig is also markedly different from the Chain version, in which the work is composed of ‘Two Series.’⁴ The ‘First Series’ comprises Dante’s terzina followed by forty-eight English translations of it, the additional translation being by Kathy Acker, which does not appear in the later Fig edition.⁵ A ‘Second Series’ follows which lists the names of the translators chronologically according to the year their translation was published, introduced by a number corresponding to the alphabetized translation, a sequence that is evidently non-chronological and non-linear. As a work, it reveals the Oulipian dimension of Bergvall’s endeavour, the fact that the text was produced following very specific rules and constraints, compiled from a prescribed corpus, and organized according to pre-determined criteria: the alphabetized list (see Perloff 2004 and Brown 2009). The Chain edition also presents the work as the product of a kind of co-authorship since the piece carries the attribution, ‘Dante & Caroline Bergvall,’ as a sort of trans-temporal coupling preceding the title, ‘VIA (48 Dante Variations),’ ‘VIA’ given in this edition in capital letters (Bergvall 2003a, 55–59).

Listening to the sonic version of the piece that is preserved as an audio recording, we hear Bergvall’s voicing of what, in Fig, is characterized as her reading of a written version of the text.⁶ The question of which text she reads is an important one and not easy to answer. The sound recording corresponds much more closely to the ‘Via’ we find in Fig than to the Chain version, except that the title of the work is not pronounced, Dante’s text is also not voiced—only the translations—and the numbers that mark out the list in the print copy are not spoken by Bergvall, suggesting that they are unique to the page. Crucially, the Fig edition which comes third in this series is also not an end point: further iterations of ‘Via’ have followed. Its texts have been written on walls and in gallery spaces⁷; an abridged piece, ‘Via (36 Dante Translations),’ appeared in an anthology of conceptual writing edited by Craig Dworkin and Kenneth Goldsmith (Bergvall 2011b)⁸; and a fourth print edition, ‘VIA: 48 Dante Variations (2000–2020) – A new Inferno’ was published in the ‘irregular’ journal, Bricks from the Kiln #4, in 2020 (Bergvall 2020). This fourth edition reproduces Bergvall’s compositional note from Fig, flanked by an additional ‘Note on VIA 2020,’ in which she explains the dialogic process that led to ‘Via’’s ‘careful revamping’ and ‘new circulation’ for this edition (Bergvall 2020, 287).⁹ In this ‘Via’ the English translations are interleaved with other contributions in the journal, and precede rather than follow the Dantean terzina, which is reproduced on the last page preceded by a single guillemet ‘(»)’: a different ending and another beginning.¹⁰ ‘Via’ has also been continued by other writers in other languages, up to now in Polish and Portuguese versions which largely operate independently of Bergvall’s template (as she herself has referred to it, recognizing ‘Via’’s reproducibility as process more than text in its own right).¹¹ Not only does this possibility of continuation dovetail with multiply-authored medieval works and continuations, like the Roman de la Rose or the endlessly recreated grail cycles of Arthurian romance (cf. Cerquiglini 1989, 60; 1999, 35–36) it also indicates the open and unfinished nature of a work that is both perennially suspended and restarting again nel mezzo (in the middle).

The medieval in practice: ‘Via’’s composition

‘Via’’s continuation as a multi-compositional macro-work or hyper-text is just one way that its production resonates with medieval textual practices, including those practices familiar to scholars of Dante, namely transcription, compilation, and commentary. ‘The dynamics of the parchment page,’ understood in Stephen G. Nichols’s words as ‘an interactive space inviting continual representational and interpretative activity’ can illuminate aspects of ‘Via’’s own composition and performance (Nichols 2015, 39, cited in Magnani and Watt 2018, 258). Dante’s prosimetric work, the Vita Nova, is an illuminating intertext alongside the Commedia with which Bergvall openly dialogues. Dante’s writings repeatedly demonstrate how the process of anthologizing and commenting upon previous texts (in his case, predominantly his own) can create new meanings. Employing citation and autocitation in the Commedia, he stages and rewrites his poetic past, in a text intricately woven with material from scriptural sources, classical texts, romance literature, and medieval lyric poetry.¹² Dante’s ‘book of memory’ (‘libro de la […] memoria,’ Vita Nova I, I, Dante 1932), the Vita Nova, if on a smaller scale, is similarly a hybrid compilation whose novelty (novitas) lies in its anthological and exegetical dimension.¹³

The compositional note which introduces ‘Via’ in Fig acts as a revealing paratext in this regard since Bergvall presents her praxis in terms that echo Dante’s process of composition in the Vita Nova specifically. In his libello (little book), Dante compiles thirty-one lyrics which he had written previously, orders them, and inserts them into a prose narrative, creating his novel vernacular prosimetrum.¹⁴ Given that in Fig ‘Via’ encompasses Bergvall’s note as well as the text, it has a partly prosimetric appearance. In turn, and in parallel with how Dante introduces his textual endeavour (see Vita Nova I, I), Bergvall presents herself in the triple role of scriptor (the one who transcribes the English translations from the British Library’s archive), compilator (the one who collates and orders the transcriptions, in her case, alphabetically), and finally commentator (the one who comments on that compositional process) (Bergvall 2005).

Scholars of the Vita Nova have highlighted how by carefully curating and commenting upon the lyric poems he transcribes from his book of memory, Dante asserts textual control and authorship of his work, excluding poems that do not fit his design and reframing some such that they effectively become new texts, distinct from the original rime (free-standing lyrics) (see Barolini 2009, Gragnolati 2010 and 2013, Picone 1977). On the surface, the material Bergvall included in ‘Via,’ in being dependent on a library archive, appears based on objective more than subjective, Dantean criteria. She draws from a collective repository of knowledge rather than an interior memory book. Bergvall also downplays a sense of authorship, representing herself not as an auctor (as Dante ultimately does in the Vita Nova; see Gragnolati 2010), but as one voice in a ‘chorus’ on a par with her quoted sources. Nonetheless, selection is at work: Bergvall has revealed that while an initial search of the British Library catalogue brought up fifty-two published translations of Inferno, this number became forty-seven when she discovered that certain works were either lost or damaged, undergoing restoration, or replicated material from previous editions (see Filreis 2013, and Bergvall 2005, 64). As such, her textual repository is less stable than it first appears and, for anyone consulting the British Library catalogue today, the translation by the Rev. Henry Francis Cary, from 1805, could seem to be a somewhat arbitrary chronological starting point since the catalogue includes translations of the Inferno that were published earlier, like that of Henry Boyd, from 1785, and that of Charles Rogers, from 1782, for example.

This first instance of material variation in a work created in otherwise very controlled conditions highlights her project’s contingency. Bergvall, as we have seen, presents contingency as intrinsic to how ‘Via’ began, and, by portraying her entry into the Commedia’s ‘wood of woods’ as a product of happenstance and as a form of conscious dis-orientation, she recreates something of Dante’s narrative meta-textually from the start. Bergvall’s titular keyword, taken from Dante’s terzina, is polysemous. ‘Via’ can mean ‘way,’ ‘passage (through),’ ‘beginning,’ and/or ‘means.’ Both ‘Via’ and the Inferno begin nel mezzo, in that middle which in Meddle English Bergvall explores as a space of ‘middling’ and ‘meddling’: where ‘Something crosses over comes’ (Bergvall 2011a, 19). Bergvall’s ‘Via’ enacts this crossing, highlighting the motion through and between languages and texts, which makes of poetic form ‘a process rather than an attribute’ (Bergvall 2003b, 327) and invites comparison also with the task of the translator, which as Kate Briggs has remarked creates ‘contacts that fray pathways’ (Briggs 2017, 60). Additionally, the word ‘mezzo’ (Inferno I, 1) carries the sense of linguistic and poetic ‘medium’ and brings us to consider the ways in which mediation works in ‘Via’’s performance on and off the page.¹⁵ Crucially, its varied modes of performance are not hierarchizable and allow for pluralized encounters with the work over time, or perhaps simultaneously. Not all versions of ‘Via’ may be known or accessible to a reader or listener, just as in the Middle Ages it was possible to encounter texts in florilegia or compendia, and thus as fragments; or as performed or read, where reading (lectio) was rarely silent and usually incorporated more than just the sense of sight (see Carruthers 2008, 212–17).

‘Voicing the medieval text’: Orality, aurality, and writing

Bergvall has described her audio works as being in ‘open traffic’ with her writing (Bergvall 2011c, 245), and ‘Via’ certainly fits that view. The work’s dissemination first in oral performance and then through diverse written forms implies that there are fluid margins between voice and text, a fluidity that also characterized the medieval setting with which Bergvall dialogues.

A.N. Doane, in his introduction to the volume Vox Intexta, states that ‘in all the forms of writing in the Middle Ages until at least the later thirteenth century, orality and writing interpenetrate and influence one another in active and vital ways, sometimes cooperatively, sometimes conflictively’ (Doane 1991, xiii). Dante’s works, which bridge the late thirteenth and early fourteenth centuries, speak to this interaction, incorporating a varied and flexible vocabulary to connote his artistic praxis. In the Vita Nova, for example, the author includes the verbs fare (to make), dire (to say), scrivere (to write), and comporre (to compose) to denote the composition of his lyric poems, and he employs the technical verb, ‘assemplare’ (to transcribe), common in Italian medieval notarial practice, to refer to his activity as scribe.¹⁶ In the Commedia, we find the lyrical act of singing (the poem itself ‘sings’ its message in Paradiso II, 2–3: ‘[i]l mio legno che cantando varca’), the act of writing and its instruments (for example, Paradiso XXIV, 25: ‘salta la penna e non lo scrivo’), and a combination of the two (‘S’io avessi, lettor, più lungo spazio | da scrivere, i’ pur cantere’ in parte | lo dolce ber che mai non m’avria sazio,’ Purgatorio XXXIII, 136–38): a fluid mix of writing and singing, addressed to a reader (‘lettore’) who is invited to share in an experience of sweetness which sung (that is, lyrical) language expressly conveys.¹⁷ Within the hybrid textual culture of which Dante’s Commedia is part, an audience might first encounter a text in performance before reading it (Ahern 1997, 217–18). The poet Francesco da Barberino (1264–1358), for example, wrote of having first heard Dante’s Commedia that way (see Ahern 1997, 226), and during the first decades of the poem’s reception, Giovanni Boccaccio inaugurated the tradition of the ‘lectura Dantis’—an explication of the text and public reading—which is still going strong today, more than 700 years after the poet’s death.

Attention to these composite facets of medieval textuality brings out ‘the many ways in which the written word was embedded in orality’ and how ‘the spoken word was implicated in the “literate” domains of writing and visual apperception’ (Chinca and Young 2005, 3). The intermediate mode of ‘aurality’ is especially helpful for thinking about works that simultaneously appealed to both readers and listeners (Chinca and Young 2005, 7), as ‘Via’ also does, reminding us of the multiform character of medieval encounters with texts and artefacts and their varied sensory aesthetics. In Meddle English, Bergvall writes of confronting ancient works and of experiencing ‘the mysterious pleasure grain of the vocalizing, materializing text’ (Bergvall 2011a, 8), a process also alluded to in the preface to ‘Via’ in Fig when she comments that ‘The minutia of writing, of copying out, of shadowing the translators’ voicing of the medieval text, favoured an eery intimacy as much as a welcome distance’ (Bergvall 2005, 65). Moreover, she locates her forty-eighth Dantean variation in an auditory dimension when she states that:

Using calculations set up via his software, he [Ciarán Mahler] unearthed an added line, an imperceptible grain, my voice’s fractals, and we let it run, hardly audible, underneath the structure of the reading voice, inextricably tied to it, yet escaping it […]. (Bergvall 2005, 64).

We can try to listen for this variation in the vocal recording of ‘Via,’ but in Fig we must regard it either as supplementary to the written text or as inherent in it: unheard, imaginable as sound, yet difficult to reproduce without the same support of technology that Bergvall had enjoyed. The ‘added line,’ which is deeply connected to and yet always ‘escaping’ the remainder of the piece, anticipates the destabilizing experience we have if we set the several iterations of ‘Via’ side by side. Doing so, we perceive the many variants, and not just variations, existing between one element of the work and its form elsewhere. No version of ‘Via’ is self-sufficient or closed; instead, the work is in synchrony with the porous, open textuality of medieval texts, especially medieval lyric texts.

‘Via,’ mouvance, variance

In highlighting the unstable relationship between the oral and written forms of ‘Via’ I consider here, it is productive to engage with the linked phenomena of mouvance and variance. These are terms with which scholars have conceptualized the fluid and heterogeneous character of medieval lyrics. A term first coined by Paul Zumthor in an appendix to his work Essai de poétique médiévale (Zumthor 1972, 507), mouvance describes the variation and mutability to which early lyrics and other texts were subject in their various manuscript iterations, thereby foreclosing any sense of a stable or consistent form of the work (cf. Gaunt 1998, 105). Specifically, mouvance is defined as:

the characteristic of the work which, as such, before the age of the book, results from a quasi-abstraction, given that the material texts which compose the work display, through a play of variants and revisions (variantes et remaniements), something like an incessant vibration (une incessante vibration) and a fundamental instability.¹⁸

In the Middle Ages, the ‘diverse modes of mobility affecting medieval texts, bodies and objects’ (Nelson 2017, 14; 30) produced shifting constellations of texts and a form of textuality which was open, subject to external conditions, and quite contingent. The circulation of manuscripts, the interventions of copyists, and—in the case of the Italian tradition especially—the migration of terms between regional languages meant that texts were labile entities.¹⁹ As Ardis Butterfield notes, ‘Stability is almost accidental, rather than the norm’ for medieval lyric texts and manuscripts (Butterfield 2015b, 335). Dante’s Rime, the author’s unanthologized lyric poems, are a case in point insofar as they circulated quite freely and in a variety of contexts, not always with certain attribution. As Manuele Gragnolati has argued, this freedom and variety ‘allows them to touch upon very different, often contradictory modes, genres, and styles both synchronically and diachronically,’ and the poems can be said to reveal ‘vestiges’ of a ‘less fixed “I,”’ one ‘which could be considered as a trace of a medieval mode of composing, performing, and transmitting lyrics’ (Gragnolati 2022, 13). Bergvall’s ‘Via,’ in the proliferation of ‘I’s that it offers us, all differently lost, allows us to see a trace of this more itinerant and fluid ‘io’ in the Commedia too.²⁰ It reminds scholars of Dante’s magnum opus that its author began life as a lyric poet, that this experience influenced his later composition, and that the Commedia bears its mark textually as well. Even if the Commedia has a somewhat more stable textual transmission than Dante’s lyric poems, partly because of the poet’s use of the terza rima, which ensures a degree of stability, still the Commedia has come down to us in around six hundred manuscripts (a number that rises to eight hundred, if partial copies are included), none of them in Dante’s own hand, and codicologists have identified in these textual witnesses a number of variations and divergent passages (see Zinelli 2021, 63). The language of the Commedia was prone to adaptation according to regional specificities, testifying to the malleability of the vernacular on the fourteenth-century Italian peninsula, with the result that manuscripts compiled in Northern Italy differ from those of Tuscan or specifically Florentine origin, reflecting their scribes’ ‘own linguistic habits and preferences’ (Shaw 2019, 232–33).

Given the important role that copyists had in the transmission of texts, adapting (or contaminating) them through their interventions, it can be appropriate to ask of many medieval texts a similar question to that posed by Simon Gaunt in his study of mouvance in an Occitan canso, ‘Can vei la lauzeta mover’ by Bernart de Ventadorn, whose stanzaic order shifts in the different manuscript witnesses: ‘Whose desire is articulated in Can vei? And in which poem?’ (Gaunt 1998, 100). I approach ‘Via’ taking up Gaunt’s invitation to read the various instantiations of a poem ‘differentially,’ to acknowledge how a rhetoric diversely organized can create an alternative subject and form of desire (see Gaunt 1998, 105; Gragnolati 2022). I do so in order to understand how mouvance operates in ‘Via,’ and how ‘Via’ comes to organize the rhetoric and desire of Dante’s poem differently.

In this sense, most revealing for my analysis is Zumthor’s idea of mouvance as ‘incessant vibration,’ which, together with the likely origins of the noun mouvance in the old French movant and its cognates signifying changeability and inconstancy (Latin mutabilitas), expands its meaning beyond mere movement to ‘the ongoing accretion of a “text in the act of making itself” […] through performance and transcription across the centuries’ (Rosenstein 2010, 1540, citing Zumthor 1972, 73).²¹ Extending the principles of mouvance to his study of medieval (especially Romance) texts that survive in multiple manuscript variants, in Éloge de la variante (In Praise of the Variant). Bernard Cerquiglini argued that ‘variance’ was so widespread and constitutive (‘si générale et constitutive’) in medieval literature that texts were being constantly rewritten. Each manuscript should be considered, reprising Zumthor’s own terminology, as ‘un remaniement, une version’ (a revision, a version), which does not presume a Lachmannian archetype from which all such variations would stem (Cerquiglini 1989, 62; 1999, 38).²² It is precisely this kind of deviation from patriarchal and teleological schema that brings Roberta Magnani and Diana Watt, in their case for a ‘queer philology,’ to situate mouvance and variance in the context of ‘queer’ philological practices, where ‘queer’ is ‘intended capaciously, as the non-normative’:

As the material space of the medieval codex is always already a site of hermeneutic slippage, applying queer theory to the study of medieval textuality is not anachronistic; quite the opposite: it is the fulfilment of the queer temporality of medieval manuscripts, always realized and reified as palimpsests, as mouvance, as variance, as queer. (Magnani and Watt 2018, 260)²³

With Bergvall, we are in the age of the book and print, as well as of digital media and other technologies, not manuscript culture. Nonetheless, insofar as ‘Via’ resists circumscription within any single medium, a form of mouvance operates in the shift between the oral performance of the work and its written counterparts. Meanwhile we can detect variance between the two (or more) physical texts that reproduce ‘Via’ on the page and, more abstractly, in the passage from Dante’s terzina to the multiple permutations of it in English, which both preserve and distort it through time, in a different linguistic context.²⁴ If, ‘every act of transmission is also an act of appropriation of the subject of the poem’ (Gaunt 1998, 101, engaging with Huchet 1987, 44), then ‘Via’ (true to its name) conveys a particular form of Dante to us. It is to the nature of this ‘Dante’ that the remainder of this article is dedicated, outlining the importance of Bergvall’s own errant practice in her transmission of his work, which extends from ‘errors’ in transcription to her embrace of the lability of the voice as an extension of the body.

Performing error

Mouvance and variance, as phenomena of mutability and divergence, are closely linked to error. Error should not be taken as something simply negative, in the sense of a ‘mistake’ or ‘fault’; however, it does imply a straying and a wandering, whether as a deviation from a rectilinear path (Dante’s ‘diritta via’ being a prime example), or in a more aimless way, which implies a free-form, directionless sort of movement (see Sng 2010, 3–10, and Davidson 2022, 20). The opening terzina of Dante’s Commedia combines the strangely impersonal line ‘ché la diritta via era smarrita’ (for the straight way was lost) with the active ‘mi ritrovai’ (I found myself), revealing the crux of the poem’s backstory: how has the ‘diritta via’ come to be lost? And who, or what, has strayed—the protagonist, the path?²⁵ Attentive to these kinds of motion and countermotion, Magnani and Watt write the following on a passage from Chaucer’s Troilus and Criseyde in which the author calls on ‘moral Gower’ to ‘correcte’ his ‘book,’ thereby raising a spectre of difference that seals his poem’s deviant character:

Error, which requires re-orientation, or direction and correction, is, nonetheless, also encountered as a potential source of queer pleasure with its ludic scrambling of patriarchal structures rather than their reproduction. (Magnani and Watt 2018, 253–54).

It is almost a truism to say that Dante’s first terzina is about losing direction and going off course, yet a closer analysis reveals the subtleties of this straying and Bergvall’s response to it. Error offers a suggestive counterpoint to the rules and constraints Bergvall worked with in producing ‘Via’ and, I suggest, opens a space for pleasure unbound from heteronormative and rectilinear frameworks of knowledge and textual filiation.

To this extent, I agree with critics who have argued that Bergvall’s work intensifies Dante’s initial state of being lost by stopping his poem’s narrative development in its tracks and making the way forward a seemingly endless return into error (see e.g. Kaplan 2009). Yet I will suggest that error is not only conveyed through the internal repetition and variation at play in ‘Via,’ but performed in the act of the work’s making, since an experiment to recreate Bergvall’s processes of transcription and collation revealed that of the forty-seven translations that compose the main body of ‘Via’ in Fig, only twenty-one are accurately transcribed.²⁶ The ‘variations’ here include the substitution of words by others, inconsistencies in punctuation, indentation, or typesetting, and capitalization (the latter evidently unique to the written forms).²⁷ When the work is voiced in performance, accessible through the sound recording, some of these ‘errors’ are preserved or amplified, and others understandably enter. Such errors are not unique to Bergvall’s project or her praxis—and we can recognize ourselves in all of them—but in ‘Via’ they are marked in ways that I think allow us to regard error as intrinsic to ‘Via’’s poetics as such.²⁸

The first phenomenon we can consider is ‘scribal error,’ a matter of great interest to scholars of medieval textuality, who also attend to its creative possibilities (see Magnani and Watt 2018 and Chaganti 2008). This aspect has received fairly brief critical attention in relation to ‘Via,’ perhaps because in Fig Bergvall insists strongly on accuracy:

My task was mostly and rather simply, or so it seemed at first, to copy each first tercet as it appeared in each published version of the Inferno. To copy it accurately. Surprisingly, more than once, I had to go back to the books to double-check and amend an entry, a publication date, a spelling. Checking each line, each variation, once, twice. Increasingly, the project was about keeping count and making sure. That what I was copying was what was there. Not to inadvertently change what had been printed. To reproduce each translative gesture. To add my voice to this chorus, to this recitation, only by way of this task. Making copy explicitly as an act of copy. Understanding translation in its erratic seriality. (Bergvall 2005, 65)²⁹

Despite Bergvall’s assertions, ‘inadvertent changes’ appear to abound in ‘Via,’ and it is the sense of ‘erratic seriality’ that ends up dominating. That is not to say that Bergvall did not seek to be accurate at every turn, but inevitably—and not so simply—error crept in (cf. Bergvall 2020, 286). For example, in the edition of ‘Via’ in Fig, the translation by Thomas Parsons is misattributed to Seamus Heaney, whose name appears twice: at 6, in error; and at 16, in the correct position. The sixth translation is correctly attributed in the sonic piece so the error seems to have been introduced in the (re-)transcription for the print edition, perhaps replicating the misnumbering that we find in Chain’s ‘Second series,’ where both Heaney and Parsons are numbered 6 in the series. Moreover, there are several instances where a word is mistranscribed. Most striking is the case of Pike’s translation (no. 43), where ‘oath’ is given for ‘path’ and is not altered in the written texts in Chain or Fig, even though some attributions of year are amended which were wrong (or misread?) in the sonic work.³⁰ The errant word is potentially easy to spot given its marked strangeness in English. Oaths are said to be made or broken but are not usually ‘missed,’ suggesting that the straying is of more than one kind. In other cases—for example, in the translations by Robert Durling (no. 14) and Charles Sinclair (no. 15)—‘road’ is given for ‘way.’ ‘Road’ is Heaney’s translation, which comes straight after (no. 16). It is as though the repetition in the sequence is what precipitated the error here. Variation is minimal, but its effects maximal.

Seamus Heaney himself went astray while translating Dante. When Heaney began translating the Commedia in the early 1980s, he intended to do all thirty-four cantos of the Inferno but abandoned the project after completing just three (McCarthy 2008, 3). This forms a suggestive parallel with Bergvall’s text, which gives us forty-eight variations of just the Commedia’s first three lines. Quoting only this short excerpt from Dante’s poem is a gesture that already alters his text quite radically. With Marjorie Garber, engaging with Walter Benjamin, we can recognize the interruptive force of this act: ‘To quote a text is to break into it, to “tear” something out of it.’ Quoting acknowledges the authority of the quoted text but it importantly ‘instates’ that authority ‘elsewhere’ (Garber 2003, 2). It allows us to perceive in how many subtle ways Bergvall’s work performs lostness and eccentricity by tearing the fabric of Dante’s poem and letting the first tercet of the Inferno drift in its emblematic disconnectedness, extending error from intradiegetic narrative experience to form and process.

Even readers who have no knowledge of Dante’s work can recognize that these three lines from Inferno are a fragment, given that they give no sign yet of the encounter with the three beasts to which Bergvall alludes to in her prefatory note (‘The panther the lion the wolf,’ Bergvall 2005, 64) nor of the figure of Virgil, who will rescue the errant Dante from his terror and become his guide in the first two realms of the otherworld. We also do not hear of the ‘good’ that the poet tells us that he found on his journey (‘il ben ch’io vi trovai,’ Inferno I, 8), and in this sense, there is no exit from the ‘selva oscura’ in Bergvall’s text, which lingers instead in that space and encourages its audience to do the same.³¹ While early commentators of Dante’s poem were at pains to point out that the meaning of the adjective ‘smarrita’ (Inferno 1, 3) was closer to ‘misplaced’ than to ‘lost,’³² in ‘Via’ the ‘diritta via’ appears ‘very far away’ (‘ben lontana’: see Chiavacci Leonardi (1991–97, ad Inferno 1.1.3)). We find a form of lostness and errancy that repeats. This effect is intensified metrically given that the nature of Dante’s rhyme scheme, terza rima, means you can only complete the second rhyme (here on with the word ‘oscura’) in the following tercet, which never comes. The text is left suspended, and as David Wallace has commented, the effect is to create a sort of ‘circle in three lines’ (cited in Filreis 2013) or even a spiral, given that the reader or listener moves from one translation to the next without progressing.³³ Since the unrhymed middle line in Dantean terza rima is seen as propelling the poem forward, leaving it hanging intensifies the feeling of stasis.³⁴ It also heightens our attention to the reconfiguration of elements within the English translations of the Dantean tercet, several of which are only two lines long and one of which extends to six.³⁵

Alongside these varied textual phenomena, we find additional and felicitous kinds of ‘error’ in the sonic performance of ‘Via,’ which highlight the instability of the voice and its meaning-creating potential: ‘the unstable relationship between seen and heard language’ as Alan Golding puts it (2013, 713).³⁶ A ‘contamination’ between words occurs through the play of identical or similar-sounding phonemes, and distortions are produced through the echo they create (see Dolar 2006, cited in Griffin 2015, 94). In the audio recording, we hear instances when Bergvall’s voice falters, or repeats, where she interpolates words that are not actually there, leading her to go back over and say things again. For instance, when reading Musgrave (no. 3), ‘way’ is interpolated between ‘Half’ and ‘over,’ and, in the case of John A. Carlyle (no. 17), ‘in’ is first pronounced ‘within,’ and part of the line is then repeated to correct it. Not only do these forms of linguistic interference highlight the potential for straying inherent in all language and its reproduction, but they expose the ‘limits of the unitary voice’ (Davidson 2022, 11), limits that Bergvall’s praxis, as a multilingual speaker and long-time collaborator in plurivocal projects, so often discloses.³⁷

It is also when listening to Bergvall’s vocal rendition of the work that we can recognize the musicality of her Dantean ‘variations.’ In Fig she comments that forty-seven translations was ‘a fortuitous number that promised a musical structure to the list of entries and helped determine the alphabetical logic of the list’s shifting cadences’ (Bergvall 2005, 64), alluding to Bach’s 48 Preludes and Fugues (see Bergvall 2016). Given ‘Via’’s circling and straying around the incipit of Dante’s poem, we may remain more at the level of the prelude than of fugue; after all, the piece does not come to the completeness that Bach achieves in working through all twenty-four major and minor keys. It instead emphatically resists closure. Nevertheless, the connection Bergvall makes between her variations and Bach’s alerts us to the complex sonic texture of the piece and to the rhythm of language, which as predominantly silent readers of Dante in the present we may underappreciate. In ‘Via,’ there are words that keep returning (‘life,’ ‘dark,’ ‘wood,’ ‘lost,’ ‘road,’ ‘way’) and others that occur just once (‘misplaced,’ ‘maze,’ ‘tangled,’ ‘strife’) but seem to possess the sort of ‘stickiness,’ which Rita Felski has discussed in her work on attachment: the aesthetic experience of ‘finding ourselves not just captured but captivated,’ relishing in ‘the experience of being bound’ (Felski 2020, 6; italics in the original). In ‘Via’ that stickiness may be more an aural phenomenon than a visual one, and there is a sort of vertiginous appeal in listening to the sonic piece, which lasts more than ten minutes and produces a ‘trance-like’ effect through its shifting tones and cadences (see Laynie Browne’s comments in Filreis 2013).³⁸

Together the several versions of ‘Via’ amply demonstrate the extent to which language moves and how far language is movement, returning us to the idea of drift and of meddling. Bergvall has stated that language is a ‘permutational structure,’ ‘open to all sorts of mishaps and games and mistakes’ and that it is a flexible medium that allows for error and resignification in its ‘lack of proper placing’ (Bergvall 2011a, 238). Dante too regarded language—particularly the vernacular language with which he wrote the Commedia—as unstable and unfixed. This fluid language risked unsettling the shape and meanings he aimed to give it, but also, because of that lability, it was malleable and free to follow the speaker’s pleasure (‘così o così secondo che v’abbella,’ Paradiso XXVI, 132). Elena Lombardi has shown that in Dante’s Commedia, it is the ‘syntax of poetry’ that sustains the vernacular language, supporting it and making it endure, but within a flexible rather than a rigid structure that retains language’s instability and changefulness (‘volubilità’) and thus its creativity (Lombardi 2007, 137–38, and 2021, 319–20). Much (lyric) poetry of the Middle Ages foregrounds ‘the subtleties of catching sound and meaning on the page precisely in their looseness, and their profligate iterations’ (Butterfield 2015, 334). ‘Via’ is similarly labile. Serializing the excerpted lines from what is ostensibly an epic and narrative poem, although intrinsically lyric in its texture and mode of desiring and understanding pleasure, Bergvall creates something like a ‘lyric sequence’ through linking together a set of textual fragments.³⁹ If, as Jonathan Culler has argued, lyric is that which does not aim to be consistent (Culler 2015, 77), then there is something quintessentially lyrical in Bergvall’s ‘Via,’ in its resistance to coherence despite its appearance of order in the numerical, alphabetized list.

Crucially, these phenomena linked to error and inconsistency seem to affect writing and voicing alike, although in different ways. They point to where ‘Hesitations, erasures, the accidental details left in become a formal decision which calls up the elusiveness of writing’s corporeal (individuated) and material (instrumental) base’ (Bergvall 2003b, 329). Bergvall implies as much when, in the preface to ‘Via’ in Fig, she states:

There are ways of acknowledging influence and models, by ingestion, by assimilation, by total absorption in the material. To come to an understanding of it by standing in it, by becoming it. Very gradually, this transforms a shoe into a foot, extends copyism into writing, and perhaps writing into being. This whole copying business was turning out to be a hands-down affair. This was an illuminating, if disturbing, development. (Bergvall 2005, 65)

Thus, there is creativity in her acts of ‘recitation’ and ‘copy’ and, also, metamorphosis: turning the existing texts into something new, changing their status, and her own as well. Acknowledging how easily one can slip from copyism to writing, and writing to being, Bergvall attends both to the unsettling nature of her discovery and to the pleasures it brings (it was ‘a hands-down affair’), which carry a transgressive charge. It harnesses a queer potentiality, especially if we consider ‘queer theory’s deep connection with notions of performativity’ and how ‘something about energy, intensity, life, and unpredictability produce[s] the pleasures of eroticism, which resists any easy categorization’ (Cobb 2015, 22–23).

Analysing further Bergvall’s words, and their implications, we can say that Dante too knew what it meant to become totally absorbed in one’s material, to become one’s models and, also, to change them and oneself from the inside. We could think of the lyric poets he encounters in the otherworld, with whom he stages dialogues about what poetry is and how one ought to write, inevitably coming out on top; or of Virgil, who disappears as a character from Dante’s poem at the end of Purgatorio but whose texts remain woven into the fabric of the Commedia until the last; or of the intimately lyrical creations of Francesca, Matelda, and Beatrice, who all embody some aspect of the poet’s lyric past. Beatrice’s case should be considered in light of Dante’s ‘radical […] choice of placing this young, Florentine, bourgeois, vernacular—and, most importantly—woman, and the lyric and erotic world she represents, as the initiator of his journey to the otherworld and as his guide to heaven’ (Lombardi 2018, 117). All of these are examples of how a text can wander or lead (its author) out of bounds (extravagāre). When Bergvall articulates her almost seamless transition from undertaking an apparently mechanical, external exercise to something that felt much more organic, she invites us to reflect on her role in the creation of a work that presents itself, on the one hand, as mere recitation and copy and, on the other, as something which took on new life and changed status, taking her with it.

Given the effects of mouvance and variance and the numerous errors or slips in transcription or pronunciation, we can say that Bergvall’s voice is one that ‘both repeats and alters’ the Dantean text she reads, (re)producing the variations we hear, see, and experience (Griffin 2015, 76). This dynamic can be taken as a meaningful performance of the process of translation, particularly in a ‘prismatic’ sense, that is ‘opening up the plural signifying potential of the source text and spreading it into multiple versions, each continuous with the source though different from it, and related to the other versions though different from all of them too’ (Reynolds 2019, 3). Yet it suggests something more as well. If in ‘Via’ Bergvall ‘intervenes with a woman’s voice and a creative rearrangement of a predominantly male textual tradition’ (Bermann 2014, 287), her freedom to play with the distinctive tones and accents of the English she speaks, which carry the traces of her multilingual belongings, also troubles the hegemony of that language (see Bergvall 2018a, b; Leong 2011). Together these phenomena allow us to consider Bergvall’s relationship to the Ovidian figure of Echo, whom Miranda Griffin declares ‘is vital […] for understanding the way in which language is transformed by rewriting’ and whom critics increasingly acknowledge as possessing an unexpected potential to disrupt the presumed stability of language and the conveyance of meaning’ (Griffin 2015, 68). The last part of this article is consequently dedicated to a discussion of Bergvall’s ‘Via’ in relation to the voice, astray, whose divorce from the realm of authorial utterance can paradoxically free the speaking subject to create new and unexpected meanings, enabling a particular pleasure—in the proper sense of diversion—that comes from resisting linearity and teleology.

The voice astray: Playing Echo, or Bergvall as ‘mockingbird’

Insofar as Bergvall’s part in ‘Via’ was, in her words, ‘[t]o reproduce each translative gesture. To add my voice to this chorus, to this recitation, only by way of this task’ (Bergvall 2005, 65), she appears to be a ‘rote reciter’ of others’ words, with no authentic discourse of her own (see Reed 2007). Yet the qualification Bergvall adds, ‘or so it seemed at first,’ hints at how things are more complicated and that there can be no stable, one-to-one correspondence between a word or phrase and its reiteration elsewhere.

Echo offers one way of thinking through this phenomenon of linguistic deviation. In book III of Ovid’s Metamorphoses, Echo is introduced in terms of her garrulousness and skill in speaking (Met. III, 362–65, see Knoespel 1985, 7), a trait that leads to Echo’s punishment when Juno discovers that Echo has distracted her with long speeches while her husband, Jupiter, was fornicating with other nymphs. In retribution, Juno takes from Echo her potential for original utterance, meaning that she can only repeat the last words that others have spoken (‘illa parata est/ exspectare sonos, ad quos sua verba remittat [she is ready to await the sounds to which she may give back her own words], Met. III, 378–79, Ovid 2014). For Dante too, as is evident from a passage of the Paradiso that evokes Echo through circumlocution, Echo is a figure for a particular mode of utterance, one presented as a sort of sonic reflection, conjured through the image of a double rainbow, which importantly incorporates a vibrating motion and not just symmetry:

nascendo di quel d’entro quel di fori,

a guisa del parlar di quella vaga

ch’amor consunse come sol vapori.

(Paradiso XII, 13–15)

[the one without born of the one within, like the speech of that wandering nymph whom love consumed like the sun does vapours.]

The adjective ‘vaga,’ from the Latin vagare, in conjunction with the final line of Dante’s tercet, succinctly conveys that part of the Ovidian myth that recounts Echo’s fated desire for the twinned figure of Narcissus (trapped in a visual rather than an auditory reflection), whom she can never bring to recognize her desire, and which leads to her being consumed by love for him. For Dante, Echo is the one who ‘wanders’ with desire and whose voice wanders with her, her body withered to stone and then lost completely, leaving only voice behind it (‘vox manet,’ Met. III, 339).

At first glance, Echo is a figure who consequently stands for an erasure of subjectivity and loss of agency: unable to invent words of her own, she is totally dependent on what others say and has the ‘briefest use’ of speech (‘brevissimus usus,’ Met. III, 367). At the same time, as Gina Bloom has argued, precisely because Echo ‘has only “use” of speech,’ not ownership of it, ‘she reaps the benefits of a property that is not hers’ (Bloom 2016, 136). An example often cited is the moment in Ovid’s narrative when Narcissus cries out ‘huc coeamus’ (‘let us meet here,’ Met. III, 386), which Echo returns as ‘coeamus’ (Met. III, 387), meaning ‘let us meet/copulate,’ foregrounding the sexual significance of the verb (see Knoespel 1985, 8 and Bloom 2013, 164). In this example, she can be seen to profit from Narcissus’s speech since the word she repeats unexpectedly conveys her desire, and the distortion of language is not without meaning.⁴⁰ Moreover, as Jhumpa Lahiri has reflected on Echo’s significance for understanding what it means to translate; whereas Ovid’s Narcissus is transformed into a beautiful but silent flower:

It is Echo who looks beyond herself, who sings alongside others, who survives him, and whose voice resonates and remains. Her story and her resilience remind us that translation—which simultaneously repeats, converts, reflects, and restores—is central to the production of literature, not an accessory to it. (Lahiri 2022, 58)

The same could be said of those works of ‘unoriginal genius,’ which create precisely in being ‘citational and often constraint-bound poetry’ (Perloff 2010, xi).⁴¹ Appropriating and repeating the words of others does not preclude creativity, but that creativity does not come from a perfect correspondence between an ‘original’ utterance and its copy, but in the more ripple-like effect of the echo that includes transversal—that is, errant—movement and not just replication. To this extent, I find suggestive Bergvall’s own characterization of how the voices of conceptual women writers, who create via ‘accumulation and collation’ of pre-existing material, gain traction by having the discourse they repeat resonate through them, thereby warping the fabric of language:

The authorial voice multiplies its effects by explicitly acting as an empty intermediary, a ventriloquist, a mockingbird. But this bird distorts and misuses. It imagines one-to-one as a friction, not an equation. (Bergvall 2012, 18)

In the same essay, ingeniously entitled ‘The Conceptual Twist: A Foreword,’ Bergvall refers to Luce Irigaray’s notion of ‘mimicry’ (mimétisme), which the French philosopher has proposed as an intermediate strategy in the struggle to overcome a patriarchal order that systematically diminishes the voice of women, insofar as mimicry ‘already means to convert a form of subordination into an affirmation, and thus to begin to thwart it’ (Irigaray 1985, 76.)⁴² Bergvall’s lexicon of distortion and misuse pushes Irigaray’s strategy a stage further, however: what the mockingbird borrows is not left unchanged, and error and misprision inhere in its voice, which resists conformity to the voices it channels as its own. The word ‘friction’ dovetails with Zumthor’s concept of mouvance as ‘incessante vibration’: friction denotes the resistance a ‘body meets with in moving over another body’ (OED), which slows passage forward. It also signifies the energy generated by rubbing language(s), or versions of language, together. Perhaps most importantly, the ‘intermediary’ through which the voice passes, which Bergvall characterizes as ‘empty,’ is neither immaterial nor transparent, just as in the aural version of ‘Via,’ we can perceive the aleatory and somatic dimensions of language that mark out Bergvall’s own voice: rhythm and intonation, the movement of her tongue, breath.⁴³ In that work Bergvall may well be acting in the guise of the mockingbird, harnessing the potential that gesture has for generating creative multiplications and acts of friction, her own voice resonating through the existing voices of the translators, thereby also changing them.

If Echo is a figure who lays bare the radical instability of all signifiers and the ‘confusing entanglement of meanings’ that follow (Nouvet 1991, 107), then ‘Via’ is a text that shows us language’s innate propensity toward misprision and drift. For medievalists, the mockingbird could contribute an additional ornithological voice to Sarah Kay’s taxonomy of parrots and nightingales in her study of troubadour quotation—in which the ‘parrots’ way’ implies verbatim repetition of a text and foregrounds questions of knowledge and writing, while the ‘nightingales’ way’ underscores affect, via a process of recreative adaptation and lyricization (Kay 2013, 11–13). According to Garber, ‘Quotation resides somewhere in the territory between the aphorism and the echo—which is to say, between the essay and the lyric voice’ (Garber 2003, 2). ‘Via’ appears to inhabit that middle space as well, diffracting Dante through its ‘translation multiples’ (Szymanska 2019) and offering its audience a particularly open and dispersive kind of non/commentary on the opening lines of his poem. As such it is a striking example of ‘deformative reading,’ an imaginative response to a work of the imagination that encourages us to ‘rethin[k] our resources of interpretation’ (see Osman 2022, 208 and Samuels and McGann 1999, 28).⁴⁴ Bergvall’s authorial status in relation to ‘Via’ remains elusive, yet she gains from that elusiveness an ability to intervene laterally, obliquely, adding her own ‘conceptual twist’ to the tale of language. Bergvall’s is a ‘strange passage,’ one that Susan Rudy, writing on ‘Via’ in dialogue with Eve Kosofsky Sedgwick’s concept of ‘ardent reading’ (also known as ‘bad reading’), has argued creates ‘queer openings’ in the text for readers like her ‘to move through’ (Rudy 2020, 170), inviting readers into those ‘sites where the meanings d[on’t] line up tidily with one another’ (Sedgwick 1993, 4, cited in Bradway 2017, xxxiii). Interpreting ‘Via’ after Echo, and in dialogue with Bergvall’s figure of the mockingbird, shows us how much Bergvall’s voice matters, and how the ‘friction’ generated by her creative process, by the errors and distortions that language passes through and relays, is an important element of what her work conveys to us.

Conclusion: Straying farther

Through Bergvall’s ‘spiral retelling’ of Dante’s poem, to borrow a phrase from Édouard Glissant’s Poetics of Relation (1997, 16), we are invited to reflect again, and more deeply still, on the straying depicted in the opening lines of the Commedia.⁴⁵ We are also encouraged to consider why error has such a strong hold on us. Earlier I cited Magnani and Watt on the nature of ‘queer pleasure with its ludic scrambling of patriarchal structures rather than their reproduction’ (2018, 253–54). It would be hard to think of a more apposite expression of ‘Via’’s iterative, non-replicatory, and unstable ‘variations’ on Dante, of its insistent meddling with the geometric, linear trajectory implied by the sequentiality of the text and the ‘diritta via,’ which by this account is always already lost. While taking nothing away from the seriousness and erudition of Bergvall’s enterprise, I believe we can recognize in it a more ludic aspect, one in synchrony with the body and the voice, which, rather than shadowing the material they draw on, ‘come to an understanding of it by standing in it’ (Bergvall 2005, 65), meddling ‘copyism,’ ‘writing,’ and ‘being,’ and hinting at the pleasures of allowing oneself to be absorbed, or carried away, by a text and its effects on us. Bergvall’s ‘Via’ is the ‘way’ of error. Its resistance to closure and its openness to continuation keeps ‘the prospect of further errance’ in play (Sng 2010, 10), re-/dis-orienting the space of Dante’s text to make it inhabitable by others too (see Rudy 2020 and Robinson 2009).

The Dante that Bergvall’s ‘Via’ performs, for all the intensification of lostness and errancy the work effects, is also paradoxically a freer Dante insofar as it is a Dante uncoupled from the text of the Commedia that follows the opening terzina and from the seven centuries of commentary surrounding the poem. Bergvall’s Dante is not the monolithic authorial figure so often presented as being in control of himself and his œuvre, the latter often seen as totalizing and coherent, with little room for error (Gragnolati, Lombardi, and Southerden 2021). Resisting the sort of teleology that is often ascribed to Dante’s poem, Bergvall gives back to Dante a waywardness that the remainder of the Commedia does not wholly expunge, but which is not usually considered in relation to its protagonist.⁴⁶ If, as suggested above, Bergvall’s operation in ‘Via’ is one receptive to queer potentialities, these may include those inherent in Dante’s own text, which have been sensitively delineated by scholars such as Bruce Holsinger, Gary Cestaro, and Manuele Gragnolati (see Holsinger 1996, Cestaro 2010 and 2021, and Gragnolati 2013). Most recently, and with particular relevance to Bergvall’s ‘Via,’ Cestaro has argued that the ‘selva oscura’ of Inferno I is itself a ‘queer space’ connected to male erotic vulnerability through the Virgilian figures whom Dante names in the very same canto of the poem: the lovers Euryalus and Nisus in particular, whose exploits in Aeneid IX–XI are set in ‘a dark wood that hosts a series of queer narratives, […] lost paths’ (Cestaro 2021b). From this perspective, Bergvall’s choice of the first three lines of the Commedia would not be coincidental. We have seen how, as an audience, we can encounter ‘Via’ in manifold ways, choosing to enter or leave it in different places, or finding ourselves where we did not expect. We do so always with a sense of being (again) astray. Captivated by ‘Via’’s spiralling narrative, our way through Dante’s text is necessarily mediated and indirect, and yet, like Bergvall, we move ever deeper into the ‘selva oscura,’ rather than beyond it. Untethered from purpose, resolutely straying, the Dante whom Bergvall’s work creates is similarly errant and unruly.