‘Ransom of His Soul:’ Shaped Text as Medium and Mediator in Byzantium

Shaped text in the Byzantine context has recently received considerable attention from scholars. Yet decorative, non-figural shaped texts remain relatively unexplored. Drawing on the works of Jeffrey Hamburger and Ivan Drpić, this article analyzes an instance of a decorative shaped text in the catena of the Middle-Byzantine manuscript known as Laur. Cod. Plut. 5.9. This paper argues that the shaped text bore a significant purpose and a theological meaning for its producer, Niketas. Far from being merely decorative, the shaped text featured as its own distinct medium and functioned as a soteriological mediator between man and God.


Introduction
The section of patristic commentary, or catena, found on folio 279r of the tenth-century Florentine codex Laur. Cod. Plut. 5.9 differs from the conventional framing configurations of Byzantine catenae because of its irregular shape ( fig. 1). This shaped catena contains the leftover text that did not fit within the main body of the larger catena and is therefore marginalized, quite literally, to the edge of the page. 1 The shape has a curious configuration: despite its purely decorative appearance, it verges on representing, without actually representing, a kind of stylized column or candlestick. Not only does this shape deviate from standard conventions of transcribing catenae in Byzantine manuscripts, it also differs from most of the other catenae in this codex. It therefore stands out as an anomalous configuration, and in this way, it is doubly marginalized. At the same time, the shaped text exemplifies an extraordinary effort on the part of the copyist in arranging the shape and then filling it with the overflowing text of the catena in a way that makes it fit neatly on the vertical axis of the page. Its anomalous position asks questions of today's scholars just as much as it would have of its Byzantine viewers in the tenth century: what may have been the motivation and the meaning behind the copyist's work? To answer this question, it is not enough to compare it with other examples. Even the growing literature on shaped texts in Byzantine manuscripts offers an incomplete explanation of this particular variant because scholars contributing to the discourse tend to focus on figural arrangements-texts that represent a recognizable shape of some material object-and not on decorative shapes. Similarly, studies of ornament tend to occlude text-based configurations. 2 That this trailing remainder of text received such unusual treatment from the copyist is a matter deserving some attention. At the very least, it posits important questions about the role of text as a manipulable medium. Prompted by the works of Ivan Drpić and Jeffrey Hamburger who have dealt with text-based configurations in depth, this article argues that the combination of textual and visual properties in the shaping of this catena constitutes a third medium that amplifies the tension between image and text. The Byzantine concept of adornment (kosmos) played a critical role in the generation of this third medium. In particular, the soteriological or salvific connotations of the act of adornment expressed in the Laurentian codex's dedicatory inscriptions suggest that the copyist and the patron were engaged in something more than mere decoration. It so happens that folio 126r in this same Laurentian codex contains another instance of shaped text, but it differs from the column-like shape on folio 279r in two important ways ( fig. 2). First, the shaped text on folio 126r is not part of a catena but is rather the primary text of the preface to Jeremiah. Second, it is shaped in the form of a decorative cross with triangular vertical arms, whereas the column-like shape on folio 279r bears no recognizable allusion to a symbolic or figural form. It could be argued that the column-like shape does include two horizontal arms that evoke cruciform shapes, but these are only evocative if the rest of the shape is ignored. In contrast with the much clearer cruciform shape on folio 126r, the bulbous features, the flared extremities, and the overall composition on folio 279r are ambiguous enough that the viewer is aware that something is being represented without being able to identify it precisely. Likewise, the perpendicular appendages are enough to dissuade the viewer from identifying the shape as a column in the strict, architectural sense. The catena on 279r is not a title or part of the primary text, as is the case with the shaped text on folio 126r, and its shape is so irregular that it seems to function as a geometric ornament that evokes, but does not quite instantiate, the shape of a cross or an architectural column. Therefore, the cross-shaped example on folio 126r cannot explain the shape of the catena; they are not iterations of each other in either form or substance. The shaped catena is rare and unusual, but it is not entirely unique. Aside from several other iterations of decorative catenae in the Laurentian codex, a closely related example is located in another tenth-century manuscript partially preserved at the Biblioteca Nazionale in Turin: Taur. B.I.2 folio 26v ( fig. 3). 3 Here too, the copyist seems to have rendered the remainder of the catena into a column of shaped text with a decorative, non-figural (i.e., non-representational) outline. Although the page in Taur. B.I.2 is damaged and thus provides only a partial view of this shaped catena, the upper half of the vertical shape is almost identical to the one in the Laurentian manuscript. This coincidence seems to point to a case of idiosyncrasy rather than representing a more general trend among tenth-century Byzantine copyists. Art historian John Lowden supposes that the two manuscripts are either twin codices or, at the very least, that they passed through the hands of the same copyist at some point in their making. 4 In other words, the shaped catenae are unusual enough to likely constitute a personal touch or a kind of maker's mark on the part of the copyist. And this idiosyncrasy adds yet another layer of complexity to an already puzzling visual component.
Still, the shaped catenae cannot be dismissed as only the idiosyncratic expression of a single copyist. As in other visual cultures at the time, shaped texts occurring at major points of the script were a common feature in Byzantine manuscripts. This was especially the case with shaped titles at the beginnings of texts and funnel-shaped endings to passages or sections of text. Typically occurring in Byzantine and occasionally in Latin manuscripts, such treatment can be found in countless examples from the Middle-to the Late-Byzantine periods. For example, the fourteenth-century illuminated lectionary of John VI Cantacuzenus includes a large number of such endings that appear with considerable regularity throughout the codex ( fig. 4). In a curious instance of meta-pictorial representation, from an eleventh-or twelfth-century Latin manuscript, Hugh the Painter emphasized his funnel-shaped ending with an elaborate border; he then depicted himself with the manuscript just to the right of the border as if to emphasize his handiwork ( fig. 5). 5 Decorative shaped texts feature prominently and intentionally in both Byzantine and Latin manuscripts, yet it is predominantly the figural or symbolical forms that have attracted the attention of art historians, not the funnels or the non-representational geometric shapes. In his overview of both the Laurentian and Turin codices, Lowden passes over these features with a brief description, giving no indication as to their possible meanings or functions. 6 It is only in recent works by scholars like Ivan Drpić and Jeffrey Hamburger that shaped texts in the Byzantine cultural context have been given greater attention. 7 Through their work, scholars are much better equipped to understand Byzantine attitudes toward shaped texts in general, and geometric non-figural arrangements in particular. Drpić and Hamburger insist that the shaping of texts in the Byzantine context is never insignificant, partly because Byzantine semiotics attributed value to the location of a word in space and partly because ornamentation contributed to the beautification, and therefore the sanctification, of the world. Their contributions show that, at the very least, non-figural instances of shaped texts likely played a much greater role than their decorative appearance suggests at first glance. The evidence presented here will validate many of their assertions, but because they generally work with representational figures (Drpić mostly works with epigrams on artworks and architecture and Hamburger with representational shaped texts), this study will also complicate and move beyond some of their assumptions about shaped texts. By considering a broader view of Byzantine aesthetics and image theory, and by pointing to an important clue in the dedicatory poems of the Laurentian codex, this article provides an explanation for these unusual shaped catenae. In finding that the shaped catenae constituted a distinct medium and that they functioned as mediators of salvation for the patron and copyist, it contributes an overlooked dimension to current discourse on Byzantine visual culture and shaped texts.

(An)Iconicity of Shaped Text
In the introduction to his book Script as Image (2014), Hamburger creates a framework for his study of the text-image relationship in medieval manuscripts. Noticing the interdependence of this relationship, he calls it "iconicity of script." 8 He treats text as a visual medium that is instrumental, expressive, and capable of extending beyond the concepts of signification or symbolism. 9 Such a definition reevaluates the text-image relationship, and it posits shaped text as an autonomous medium different from either the image or the text alone. This distinction is important and useful in clarifying discourse on this topic; otherwise, the text-image dichotomy becomes dependent on one or the other component and restricts discourse to their respective limitations. Here, Hamburger understands medium more directly as an independent material or technique of artistic work. In other words, medium is not just the hybrid or composite outcome of the merging together of text and image, but it is also a substance-an artistic medium-in its own right. 10 What follows from Hamburger's definition is a proposed theory of the "iconicity of script" that acknowledges modalities and analytical systems unique to shaped texts. If the text-image relationship were not treated as its own medium, capable of employing a range of its own modal systems, then it would be restricted to just literary or visual analysis. When it is considered as its own medium in accordance with Hamburger's definition, shaped text allows for discourse that fully acknowledges its complexities as a modality unique to itself even as it is composed of two reciprocating elements-text and shape.
However, Hamburger's approach has limitations. "Iconicity of script" corresponds most effectively to texts shaped to resemble a recognizable form resulting in the treatment of text as an icon or image. The phrase implies that script or text has been turned into image, and it therefore prioritizes the visual and, specifically, the iconic element over any other. "Iconicity of script" becomes one of many different forms of iconicity, or, put another way, "iconicity of script" treats the shaped text as a form of visual art that happens to be constructed of letters and words rather than paint, clay, marble, or any other medium. By prioritizing the iconic quality or the recognition of form in his juxtaposition of script and image, Hamburger leaves unexplored the non-figural, geometric shapes such as funnels at the ends of passages or, in the case of the Laurentian codex, the column-like shape on folio 279r.
Hamburger's term "iconicity of script" fails to describe many cases of non-figural shaped text because it assumes that script or text, as the medium, acquires the modality of icon. In such cases, this definition contradicts itself. On one hand, it claims to treat the shaped text as a distinct medium different from either text or image, and on the other hand, it prioritizes the visual component as the output and text as the medium that assumes a visual mode. An image constructed from script is still an image. Thus, the column-like catena on folio 279r approaches something equivalent to the "iconicity of script" but manifests itself as non-representational ornament rather than icon or image. In this case, the phrase "aniconicity of script" comes to the fore as an appropriate, if also a paradoxical, neologism.
The distinction between icon and aniconic script may be subtle when applied to shaped texts, but it bears important connotations for the study of the shaped catena. Juxtaposing folio 126r in the Laurentian codex against the column-shaped catena on folio 279r exemplifies the different values at play. Insofar as the former text assumes or at least suggests the shape of a cross, it acts as a medium used to construct a recognizable image. And, indeed, the cross shape should be thought of as an image according to Byzantine post-iconoclastic image theory. After the iconoclastic controversy of the eighth and ninth centuries, the Byzantines thought of the cross as an image (i.e., icon) of the True Cross rather than an ornament or symbol. 11 Thus, the Byzantine beholder of the shape on folio 126r would consider it an image, and this justifies employing the phrase "iconicity of script" to describe the effect. However, the shape on folio 279r does not represent anything concretely recognizable. Despite its evocation of crosses or a column, it is a shape for the sake of a shape, or rather, it is a self-referencing shape because it does not signify anything in the natural world other than itself. It is as much a text-made-shape as it is a shape-made-text, but it is not an image, since it does not represent any single object in particular, nor is it just text. It truly becomes its own medium. Folio 279r thus furthers Hamburger's claim that shaped text constitutes its own medium, but at the same time, it extends the concept "iconicity of script" beyond the limits of representational form -beyond iconicity itself.
Thanks to these inherent tensions, shaped texts such as the catena in question never actually realize the transformation of text into image. It is indeed shaped text, and therefore inherently visual in essence, but the shape never achieves iconicity, that is, representability. Without a referent, non-iconic shaped text does not "represent" in the strict sense of the term; in other words, it does not make any specific referent "present again." This can be shown more clearly using other examples of decorative shaped texts such as the funnel-shaped endings that proliferate in Byzantine manuscripts. Such funnels are categorically aniconic even if they may resemble or imitate pieces of jewelry and tassels on liturgical vestments. They do not fulfill the role of image-at least not to the degree that the shape represents an object or entity found in the natural world. This resulting aniconic script, or "aniconicity of script" to reformulate Hamburger's phrase, becomes a significant, if overlooked, feature of shaped texts. As a medium in its own right, aniconic script lends itself to various circumstances that call for non-representational form to resolve tensions between text and image. An example of this can be found in manuscript E.H. 2878 held at the Topkapı Palace Museum. The fifteenth-century illumination of a lion constructed entirely of words within a contour line is an example of shaped text that resists fully becoming image ( fig. 6). 12 The lion bears no direct reference to the content of the text from which it is composed; it reads: "prince of the believers, ' Alī son of Abū Țālib; may the generosity of God be upon his victorious countenance, and may God be pleased with him." Instead, the image refers to the rich cultural iconography of lions in Islamic art as representations of power and right to authority. 13 At the same time, the copyist refers to the general tension inherent in Islamic culture between iconic and aniconic representation. It is well recognized that Ottoman visual culture was not entirely aniconic, but tensions nevertheless existed within the larger cultural sphere, and these tensions were greatly amplified in literary circles among copyists of religious and semi-religious texts. In this literary sphere, aniconism was occasionally tested and sometimes broken. Thus, this fifteenth-century manuscript possibly exemplifies an artist pushing cultural resistance to figural representation to the very limit.
By creating a visual rendering that can situate itself, even if precariously, within the restrictions of the immediate cultural context, the copyist seems to offer a commentary on what should and should not be regarded as figurative representation. At the very least, the copyist is acknowledging the tensions inherent in the cultural setting and offers up a visual format that nestles somewhere within the fissure of text and image. This is not a conversion of text into image or else the image would find itself infringing upon the conventions and practices employed by Ottoman copyists of religious and semi-religious texts. 14 At the same time, it is not simply text because it has been manipulated to such a degree that it compromises legibility to facilitate an altogether independent meaning. The artist visually flexed and contorted the grammatical and orthographic composition to accommodate a visual demand. It would therefore be entirely plausible to claim that this example is neither text nor image rather than to say that it is both; it emphatically compromises its position as either one of the two elements and becomes some third element in its holistic entirety. It employs the "iconicity of script" at the same time that it relies on the "aniconicity of script"-even if it is still formulated, paradoxically, as a text-image composite. In other words, the combination of text and image has become something more than the sum of its parts.
A Byzantine example of this paradoxical composite is John the Grammarian's inscription over the Chalke gate in Constantinople that he used as a replacement for the icon of Christ during the iconoclastic controversy of the eighth and ninth centuries. The pro-iconoclastic poem reads: They who speak of God write/depict Christ in gold and contemplate [Him] not with the material [eyes] but rather through the speech of the prophets; for faith in God is the hope of those who speak in like manner. They trample openly upon the resurgent error of those who make images, as it is an abomination to God. In agreement with them, they who wear the crown gloriously raise the cross high with pious resolve. 15 Drpić notes that this poem was rendered in such a way as to create an acrostic in the middle of the paragraph that takes the shape of a cross ( fig. 7). The resulting cruciform shape in the center of the inscription utilized the arrangement of letters to spell out the self-referencing phrase "to pathos elpis" (the cause of hope). 16 The "cause of hope" refers not only to the cross for which the phrase acts as a structure but also to the True Cross. At this time, when iconodule image theory was only just being developed by St. John of Damascus, St. Theodore the Studite, and the other defenders of images, the iconoclasts conceived of the cross as the only acceptable figure to be displayed in a religious setting because they treated it as a non-representational or non-iconic figure even if technically the cross shape referred to the prototypical True Cross. 17 Iconoclasts replaced some of the most prominent images in churches with crosses, and Drpić is convinced that the cross thus visualized in John the Grammarian's inscription was somehow highlighted, perhaps by gilded letters, to ensure that it was recognized by passersby. 18 This inscription again demonstrates the great advantage of the shaped text as medium. It avoided charges of being regarded as an image so long as it remained aniconic script. Simultaneously, the entire inscription is more than merely text since it directly replaced an image previously deployed over the gate. In this way, it took on the function of an image, and it ensured its recognition as such by the clever rendition of (possibly gilded) letters and the poetic formulation of the acrostic that spells out the cruciform shape both figuratively and literally. At first, this example may evoke the cross shape on folio 126r of the Laurentian codex where the cross was regarded as an image or icon of the True Cross by the post-iconoclastic Byzantine viewer. 19 However, John the Grammarian's poem with its prominent cross was conceived of very differently by iconoclasts in the ninth century; for them, the cross in the inscription was fundamentally non-iconic. It came to occupy a liminal position between non-figural shapes and fully figural representations such as icons of Christ or the saints. In other words, the iconoclasts introduced the cross as a third category that functioned as image but was not conceived of as an idolatrous image by definition. The iconodules later pointed out just how inconsistent and contrived this formulation really was and established their own definitions of images. More to the point, iconicity is thus shown to be a fluid concept subject to the vicissitudes of a developing visual culture. A cross in one context is regarded as an image, and in another context it is not. These two examples, the illumination from the Topkapı manuscript and the iconoclastic inscription of John the Grammarian, demonstrate the critical function of shaped text in historical context. They avoid direct identification as fully iconic representations, and yet, they reach far beyond the capabilities and properties of text by functioning as images. Compositional or visual context similarly affects the fluctuation of iconicity that can be observed in particular examples of shaped text. The iconicity of an ornamental shape might be heightened when it appears next to non-decorative undifferentiated text, but deemphasized when juxtaposed against an illumination. In the twelfth-century Curzon Cruciform Lectionary now located in the British Library, the entire text is rendered in cruciform shape ( fig. 8, fig. 9). This codex is post-iconoclastic, and therefore, according to contemporaneous Byzantine theory, the cruciform shape constitutes an image. Indeed, its iconicity is emphasized by the decorative florets that draw the viewer's attention to the corners and therefore to the cruciform shape itself. 20 Yet the text never really reaches iconicity; it never fully represents its referent. This contradiction is driven home by the repetition of cruciform shapes that dilute the iconicity of the cruciform text. Because the post-iconoclastic image theory of the Byzantines leaves open the question regarding degrees of iconicity, this accommodation allows for a theorization of aniconic shaped text even if such a theory was not explicitly articulated by Byzantine aestheticians. There can be little doubt that the cruciform text on folio 1r of the Curzon Lectionary visually differs from the cruciform text on folio 126r of the Laurentian codex. In the former, the text is framed by a decorative border, and even if the text is technically cruciform in shape, it is the dominant colorful border that visually constitutes the shape of the cross-one can almost imagine how the text would spill out into an amorphous pile of letters had it not been for the solid ornamental border that props up the entire composition. Compared with the cruciform text in the Laurentian codex, the framed text on folio 1r of the Curzon Lectionary seems to be constricted rather than shaped. It can hardly be described as shaped text, let alone given the status of iconic script. By contrast, a Gospel lectionary at Dumbarton Oaks (Dumbarton Oaks MS 1) contains a reverse example on folio 145r ( fig. 10). Here, it is the text that surrounds the image of Christ crucified on the cross. Similar to the Curzon Lectionary, the lectionary at Dumbarton Oaks includes cruciform-shaped text across the majority of its pages, but on this particular folio an image of the Crucifixion is imbedded into the text. The effect of this composition is such that the text functions as a virtual container for the image-a reliquary of text containing a relic/image, and, as so often happens with reliquaries, this container imitates the shape of the relic inside. 21 Clearly, text can function as a containing medium no less than the ornamental framework on folio 1r of the Curzon Lectionary functions as a container of text. However, this example again shows that shaped text loses some of its iconicity when juxtaposed against an image with a stronger degree of representability. On any other page, the cruciform text could be regarded as an image, but here, in contrast with an image of the crucified Christ, the shape of the text loses some its potency as an image and cedes iconographical primacy to the fully figural image. At once iconic and aniconic, the viewer is thus reminded of the fluid property of shaped text, but this does not mean that text and illumination cannot operate on equal terms. In fact, the copyist and illuminator of the Curzon Lectionary forced shaped text and illumination into cooperation on folio 196v where text and illuminated ornament function symbiotically to complete the shape of the cross ( fig. 11). Having copied the last part of the text but finding the cruciform shape unfinished, the copyist collaborated with the illuminator who completed the composition by adding the ornament at the bottom. The Curzon Lectionary beautifully exemplifies the various unique properties of shaped text even as it highlights the limitations of its iconicity.
Because shaped text in these examples oscillates between iconicity and aniconicity, its fluidity made it a versatile and powerful tool in the hands of skilled copyists and artists. More importantly, the hypothesis that aniconic shaped text constitutes its own medium explains why non-figural shaped text, such as the column-like shape on folio 279r of the Laurentian codex, does not just look different than the rest of the text on the page, it also functions differently. While the exact function and meaning of the shaped catena will be explored in the following two sections, the premise inherent in the concept of aniconic script foregrounds the possibility that shaped text as a medium functions independently of either the content of the text or the shape.

The Medium and Function of Shaped Text
Text maintained a special status in Byzantium, especially when it signified imperial or holy status (emperors, saints, Christ, etc.). In notable surviving instances, Byzantine artists demonstrate a heightened awareness of text in their work by the special attention they gave to names and their location. The hierarchic spatial demands made of names sometimes required artists to manipulate written compositions by sacrificing or downplaying grammatical and linguistic structures.
Antony Eastmond brings several of these examples to the fore in his study of monograms, and two particularly striking cases that also involve shaped text will serve to underscore the Byzantine treatment of names. 22 The first instance involves a dedicatory poem on a sixth-century reliquary cross ( fig. 12). 23 Written out in two hexameter verses, the poem features the emperor's name, Iustinus, in the middle of the second verse: "Ligno quo Christus humanum subdidit hostem | dat Romae Iustinus opem et socia decorem." 24 Spelled out in a straight line, the name does not occupy a particularly significant place in the composition. However, the name occupies a prominent place on the arm of the cross where it becomes the first word encountered from the left side. To ensure that the name appears exactly in this position, the artist constricted the preceding eight words into the vertical portion of the cross. After this apportionment, the text was spatially configured to physically fit into the cruciform arms of the reliquary allowing the name, Iustinus, to feature in a prominent position on the object. This second revision came at a significant cost to legibility because the entire first verse and a portion of the second verse were packed into the upper arm of the cross. In mathematical terms, the artist constricted slightly over sixty percent of the text into just twenty-five percent of the space leaving about twelve percent of the text for each of the remaining three arms. Put this way, there seems to be little doubt that the artist here has prioritized certain inherent non-textual qualities of script over the convention and function of text.
Another striking example exists at the church of Saints Sergius and Bacchus in Constantinople. Here, a dedicatory inscription to the founders Justinian and Theodora runs along the interior nave of the building just above the cornices of the capitals ( fig. 13, fig. 14). In this twelve-line hexameter poem, the name of Justinian appears in the third verse while Theodora's name appears in the tenth verse (i.e., the third-from-last verse). This somewhat awkward separation of the donors' names is explained by the actual positions of the names in the architectural context. The two names appear on the north and south sides of the nave facing each other and correspond to the gendered spaces of the imperial church. 25 In this case, the names in the inscription would also correlate with the persons of Justinian and Theodora as they would perhaps stand in the church's gallery during the liturgy-directly above their names in the inscription. Whether or not the names were gilded or somehow emphasized by color has not been determined, but this remains a possibility. As with the inscriptions of John the Grammarian on the Chalke Gate and the cross reliquary, the Byzantine artists here have again prioritized the spatial configuration of the text over literary content. These examples suggest that the literate Byzantine viewer could reasonably expect certain texts to defer to visual demands associated with hierarchies of space, sometimes at the cost of legibility. If these expectations were also applied to script in other contexts, they could account for the funnel-shaped endings and other variously shaped texts that occur in Byzantine manuscripts. The notion that Byzantine viewers might place visual demands on text in both of these contexts-the monumental and the literary-is certainly a reasonable proposition. Most significant in these examples of dedicatory inscriptions is the way they convey the Byzantine approach to text that was meant to fill a physical space and fulfill a visual role. To this end, the very structure of Byzantine poetic compositions seems ideally suited for rendering text into visual form. Each example-John the Grammarian's inscription, Justin's cross-reliquary, and the inscription in the church of Saints Sergius and Bacchus-employed a strict poetic formulation either in the form of hexameter verses or a dodecasyllabic composition. Byzantine hexameter and dodecasyllabic verses are rather rigid structures for poetry, and some scholars have seen in this rigid structure a certain lack of creativity. However, these poetic restrictions conform exceptionally well to geometric outlines in physical space. 26 This is most clearly evident in their employment on the reverse side of a reliquary-enkolpion now located in the Moscow Kremlin Museum ( fig. 15, fig. 16). Here, the eight lines of poetry that list the various relics inside the reliquary fit neatly within its square form. Because each verse is naturally longer or shorter than the others, the inscriber repositioned, omitted, or conjoined certain vowels, consonants, and some conjunctions in order to ensure that each line begins and ends at the edges of the reliquary's frame. Drpić calls this technique the "verse-filling asyndeton" where asyndeton refers to the intentional elimination of conjunctions. 27 He explains that it lends itself strategically to spatial contexts, and it solves a persistent problem encountered by virtually all copyists who transcribed text onto physical surfaces in medieval productions. Other solutions to configuring text to meet visual demands were also employed, but less effectively. A fourteenth-century Latin leaf containing a litany from the Book of Hours, now held at the Chazen Museum of Art in Madison, provides one such example ( fig. 17). Here, the copyist filled the blank spaces at the ends of shorter verses with an ornament that was gilded and painted in red and blue tempera. In both the reliquary-enkolpion and the litany leaf, visual cohesion of text on the surface was of great concern to each of the artists, but they addressed the demands of visual structuring in different ways. Unlike his Latin counterpart, the Byzantine metalsmith avoided extraneous elements in the reliquary largely because his dodecasyllabic composition, aided by "verse-filling asyndeton," was structured in a way that facilitated the resultant shaped form. More than just a coincidence of poetics and shaped surface, this inscription demonstrates a cultural approach toward orthography and linguistic composition that is exceptionally fluid and thoroughly based on the visual expectations within Byzantine culture.
All of the instances of Byzantine shaped texts described above allude to a cultural approach toward the text-image relationship that treats shaped texts as a kind of hybrid medium. If these shaped texts were considered purely through textual analysis with no consideration for the visual component, a large portion of their meaning and agency would become instantly weakened. The visual properties simply cannot be communicated by non-visual means such as aural reading of the text. More specifically, the poem of John the Grammarian loses virtually all of its potency when the cruciform formation of the acrostic is not also considered. Similarly, both of the dedicatory inscriptions-one to Justinus on the cross reliquary and the other to Justinian and Theodora at the church of Saints Sergius and Bacchusmust be considered in their respective contexts in order to fully acknowledge the locative significance of the patrons' names. And finally, the inscription on the reliquary-enkolpion emphasizes how efficiently Byzantine prose facilitated the rendition of poetic language into visual form.
These examples express a symbiotic relationship between Byzantine literary and visual cultures. They point to a broader and higher purpose of sacred communication in which the application of text to surface was done with great consideration for spatial and visual context. These examples also provide evidence of an attitude in which matters of grammar and syntax give way to visual cohesion-where priority is given to the text-image dyad rather than to composition or readability. Thus, the visual and functional demands of each of these objects ensured that text was manipulated, shaped, and treated as if it were a medium in its own right. Perhaps more importantly, shaped text was used to imbue words with meaning that extended well beyond their semantic quality.

Shaped Text as Mediator
Shaped texts that take the form of non-figural, decorative elements are somewhat more cryptic than ones which take representational or even symbolic configurations such as the cruciform acrostic in John the Grammarian's inscription or the cross reliquary with the name Iustinus. Still, the principles at play in the text-image relationship apply also to these instances. Decorative shaped texts carry meaning rooted in their function as a distinct medium (shaped text) that is not constrained by orthographical or syntactical regulations. However, non-figural shaped texts rely much more heavily on the aniconic properties of the text-shape dyad than figural shaped texts that contain either a clear visual referent (e.g. John the Grammarian's acrostic) or a semantic referent (e.g. names in inscriptions). In other words, the manipulation or shaping of text is more easily justified when the desired outcome is the shape of a cross or the even distribution of text across a square-shaped reliquary-enkolpion, or when a particular name is to be positioned in a desired location. However, a non-figural, decorative shape employs different parameters and therefore requires a different set of justifications.
The column-shaped catena on folio 279r that initiated this foray into Byzantine shaped texts contains a referent and a meaning that is much more difficult to uncover. However, two dedicatory poems in the same codex provide important clues. These poems allude to the codex and its maker and refer to the decorative program on the codex's pages; in fact, they directly mention "adornment." The introductory poem to Isaiah refers to a person by the name of Niketas who was clearly involved in the making of the codex, likely as a patron and possibly as the author of the poems (fig. 18). The preface to Isaiah reads: "Isaiah…|… left precepts for profit in living | Which Niketas… | Has brought together in this place in adornment | As token of his faith, and ransom of his soul." 28 The poem to Ezekiel bears a similar description: "… [Ezekiel] showed so clearly the knowledge of the complete end | Of things that he could write down even the measures | Of arrangements for the future in the words | Which thus a dutiful soul adorns in hope." 29 These passages make it clear that Niketas, who was involved with the production of the manuscript, placed significant emphasis on its adornment. By doing so, he hoped that this adorning would be perceived by God as a token of his faith and a ransom of his soul, and thus, become a salvific act. The manuscript contained illuminations of each of the four prophets, though only one survives today. These illuminations are also decorated with a luxurious colored border very similar to the decorative border surrounding the introductory poem to Isaiah. The passages in the poems that mention adornment likely refer to all of these instances as a whole: the illuminations, the luxurious borders, and most certainly the decorative shaped texts that occur on folios 279r and 126r. These latter texts should not by any means be discounted as acts of adornment simply because they are constructed of words rather than paint and gold flakes. On the contrary, the attitude of Byzantine copyists toward shaped texts strongly suggests that Niketas includes the instances of shaped text in his codex as part of a conscious effort that contributed to the overall adornment of the codex.
Even if it may have not been planned initially, the ornamentally shaped catena nevertheless required a significant degree of effort and careful alignment. In like manner, the shaping of the preface to Jeremiah in the form of a cross on folio 126r exemplifies an effort that extends beyond the utilitarian layout of the rest of the text. These two examples point to deliberate efforts on the part of the copyist to give the text no small degree of aesthetic prominence. When Niketas asks God to pay attention to his act of adorning the codex, he presents his work in overt terms as a visual or aesthetic production rather than as a literary one. In other words, it is the visual elements in the codex that constitute for Niketas the intended emphasis and the salvific purpose of the manuscript.
When Niketas is mentioned adorning the manuscript as a token of faith and a ransom of his soul in the dedicatory poems, the author of these passages, presumably Niketas himself, integrates the act of adornment or beautification with soteriology (the doctrine of salvation) and beneficence. Such integration was common in Byzantium in the tenth and eleventh centuries when the manuscript was written. It involved the Byzantine understanding of kosmos-the most common word for "adornment"-as a conceptualization of artistic or creative activity that moved beyond utility toward an aesthetic and spiritual ideal. 30 Drpić explains that "to encase or frame a sacred object with a lavish material kosmos (adornment) was understood by the Byzantines as a way to gain access to and partake, as it were, of the object's sanctity." 31 At its most profound stage, this concept applied directly to epigrams, which became regarded as the highest category of adornment "above any material kosmos." 32 According to Drpić, "what made [words] such an exceptional cosmetic medium was the fact that they were wrought not from earthly materials-gold, silver, gems, pearls, and the like-but from the supremely precious stuff of logos (word)." 33 This attitude among the Byzantines explains why the shaped text occurring in the Laurentian codex mattered in the overall decorative program and how the shaped words may have constituted a sacred medium on par with paint or even precious metals. Given that, as a whole, greater attention was given to the illuminations in the manuscript, it is not clear whether Niketas would have considered text to be a medium that superseded paint or gold, but the very idea that text played a role in adornment and therefore as part of the ransom of Niketas's soul indeed raises the medium to a precious status.
The convention of treating words with particular reverence developed from a theological and philosophical discourse that described words as images and as clothes or garments of thoughts. This convention was notably employed by John of Damascus in his refutation of iconoclasm, but a general reverence for words can also be found in other cultures of the Mediterranean. 34 Byzantine theologians who dealt with aesthetics-Dionysius the Aeropagite, John of Damascus, and Maximus the Confessor, among others-formulated their theories of beauty based on the idea that the created world bears witness to or speaks of the transcendent beauty of God. Maximus the Confessor follows this conceptualization when he writes that "creation participates in the beauty and being of God in an iconic or refractive manner, whereby God's own beauty and image are passed through the cosmic order, causing the mind to 'ascend' in contemplation of the transcendent." 35 The principle of aesthetics in Byzantium should therefore be understood as an ontologically based conception that bears soteriological implications. Beholding things that are beautiful causes the mind, and therefore the soul or nous, to ascend toward God.
When Niketas has his manuscript adorned with illuminations and decoratively shaped texts, he participates in a soteriological activity that beautifies the work created. In transforming a blank space on parchment into an aesthetic medium, Niketas moves his soul closer to divinity via the senses that behold the manuscript. To put it somewhat more directly, Niketas participates in the sanctification of the created world by his efforts in adorning the manuscript. By his patronage and efforts, the decorative shaped text becomes a soteriological mediator that alludes to and leads the manuscript's reader, as well as its maker, toward the divine beauty, toward God. In this way, theological and philosophical theories of beauty in Byzantium gave decorative elements a purpose and function equivalent to salvific activity. To pen a decorative element onto parchment was to mediate the sanctification of one's soul.

Conclusion
Starting with the Florentine manuscript with its column-like catena, the examples of shaped text discussed in this article provide evidence that written language was employed in Byzantine literary circles as a visual medium that carried potency and meaning in the larger culture. Shaped text relied on both the iconic and aniconic qualities that lent it the status of being its own unique medium. Never entirely independent of the properties of text and image, it nevertheless expressed itself in more fluid ways without certain limitations and constraints. Furthermore, shaped text and especially decorative shaped text carried soteriological significance as mediator of divine beauty and of salvation.
To this end, the patron of the Florentine manuscript, Niketas, expressed his salvific aspirations as being directly linked to his adornment of the manuscript. Most importantly, his dedication provides evidence that the Middle-Byzantine approaches to written language were full of nuance and theological implications. The functional fluidity that shaped text provided for copyists and artists made this a potent medium in Byzantine visual culture with various modalities. Furthermore, situating shaped text in the context of Byzantine aesthetic and philosophical theories explains the function and significance of the catena in the Laurentian codex. The shaped text of the catena is a visual device capable of producing its own set of meanings wholly distinct from literary content and figural representations. As a part of the Laurentian codex's decorative program, the column-like catena embodied the soteriological aspirations of its producer and therefore functioned as a mediator of salvation.

About the Author
Mateusz Ferens is a Ph.D. candidate at the University of Wisconsin, Madison. He studies the art and visual cultures of the Byzantine and post-Byzantine worlds, medieval theories of optics and perception, and cross-cultural interaction. He is currently writing his dissertation on the transition of Byzantine culture to early modernity. The text of this work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0). All images are reproduced with the permission of the rights holders acknowledged in captions, or are reproduced under license, and are expressly excluded from the CC BY license covering the rest of this publication. These images may not be reproduced, copied, transmitted, or manipulated without consent from the owners, who reserve all rights, and outside the terms of any specified license.