Naming and Renaming Texts. Rubrics in Middle High German Miscellany Manuscripts

This article analyses rubrics in Middle High German miscellany manuscripts of short texts in rhyming couplets (Reimpaargedichte). A corpus consisting of 1433 rubrics from 68 manuscripts was created to be able to perform this study. As rubrics in medieval manuscripts were not authorial, but composed by scribes, they offer insights into the reception of the texts. This paper analyses their features and functions as a proxy to interrogate the standing and status of Reimpaargedichte between the thirteenth and fifteenth centuries. The main methodology is distant reading, i.e. the application and interpretation of statistical methods on a textual corpus. The features analysed include the length of the rubrics, their level of variation, the presence of author names, and vocabulary. Although no general patterns regarding length nor level of variation were detected, some important conclusions can be drawn: 1. there were no clear markers of literary genre in rubrics; 2. authorship was mostly absent, except for some specific cases of famous authors; 3. relatively stable keywords were used to identify particular texts, but they were more common in manuscripts with narrative texts (Erzählungen) and less common in later manuscripts dominated by the genre known as Minnereden. Furthermore, the analysis revealed that rubrics used a series of linguistic procedures to show that they participated in a different speech act than the main text – they embodied an interaction between scribes and readers, in which the former framed the reception of the work.1


Introduction 2
This article deals with Middle High German (MHG) short texts in rhyming couplets, known as Reimpaargedichte in German scholarship. These texts are extremely diverse in subject matter and style. They were composed between the thirteenth and fifteenth centuries and compiled in large collections between the late thirteenth and the early sixteenth centuries. Nowadays they are part of what Lars Boje Mortensen calls the "open archive" (59), texts regularly referred to by specialists, mostly available in good editions and mentioned in detailed literary histories, but generally unknown and marginal within the field of medieval studies. This article, however, is not primarily concerned with the standing and status of these texts today, but in the period in which they were copied into miscellany manuscripts. It focuses on their Reimpaargedichte's variable rubrics, located in the perimeters and margins of the text. In that sense, rubrics can even be assorted to what Mortensen called the "closed archive," since they are poorly identified and often left unspoken by scholars.
Many factors suggest that, even at the time of copying, Reimpaargedichte were marginal to what people might have considered a literary canon. 3 Firstly, these short texts in rhyming couplets were written in a vernacular language, while most canonical texts in Western Europe were in Latin. This includes not only the classics from Antiquity and the Church Fathers, but also medieval Latin texts that had achieved canonical status. Secondly, they do not usually have enough witnesses to suggest widespread success. Thirdly, they do not show enough influence on other texts, or appear enough in their references, to be regarded as vernacular classics like Dante and Petrarch in Italian or the Roman de la Rose in French. In the German-speaking world, that status could arguably be assigned to Wolfram von Eschenbach's Parzival or Gottfried von Strassburg's Tristan, for example, but rarely to any Reimpaargedicht. Finally, the usually high level of textual variation in the testimonies implies that copyists may not have considered these texts as authoritative sources and felt free to rewrite them.
Nevertheless, saying that they were not canonical texts is a relatively vague statement. Most texts in any culture are not canonical, but are still read and interpreted, and there is a wide spectrum of possible attitudes towards them. If we want to understand how these texts were considered during the Middle Ages, we need to perform a more detailed analysis. One possible avenue to study this issue is through the rubrics attached to them in their manuscript transmission.
Rubrication was a very common practice in medieval manuscripts. In a strict sense, rubrics are just script in red ink. As they were commonly used as headings for texts or sections in medieval manuscripts, the word rubric is used in this article in an expanded sense that includes all headings, even when written with black ink -which only happens occasionally in some later manuscripts. Some examples of rubrics in MHG manuscripts in our corpus can be seen in figures 1 and 2. They exemplify the consistency of this practice over time. The first figure shows a manuscript from the first quarter of the 3. I use the concept of 'canon' as it is articulated in the field of cultural studies by Aleida Assmann: "actively circulated memory that keeps the past present" (98). A reformulation of the concept for medieval literature can be found in Mortensen. thirteenth century, the second from 1512. The script, material and layout has changed considerably, but the use of rubrics as headings dividing the different textual units persists.
Rubrics are one of the paratextual elements surrounding works in medieval manuscripts, which also include other features such as colophons, illuminations, initials and decorations. All these elements are important sources to understand the reception of texts throughout the period. As Genette points out, the paratext is "a privileged place of a pragmatics and a strategy, of an influence on the public, an influence that -whether well or poorly understood and achieved -is at the service of a better reception for the text and a more pertinent reading of it" (2). Of course, these medieval paratexts are different from their modern counterparts. Rubrics, unlike most modern titles, were not irrevocably attached to the text and could be adapted and rewritten often. For this reason, they reveal something about how scribes read and understood the works they copied. Like other paratextual elements in medieval manuscripts, they are the result of an active act of reproduction and reception of the text. As I will show, rubrics were explicitly separated from the main text and constitute a different speech act that corresponds to an interaction between the scribe and the readers. They are not really part of the text they refer to, but an external instance in which scribes framed the reception of the text.
The main method used in this article to study these rubrics is distant reading. 4 Distant reading involves the application of statistical methods to study quantities of texts that are difficult to investigate with traditional philological methods or to explore general trends and patterns in a textual corpus. In medieval studies, the available materials are not as abundant as for modern literature, but still large enough to be examined using distant reading.

Sources
The rhyming couplet was arguably the dominant poetic form in German literature from the twelfth to the fifteenth century. It was used from Arthurian epics to theological and didactic treatises. Written in this form, there is a huge group of what Fischer calls Reimpaargedichte, short texts of up to two thousand lines, although rarely more than eight hundred and most commonly between three hundred and six hundred lines. Fischer distinguishes two main categories: Erzählung (narrative) and Rede (discourse), each with their own subcategories.
An important feature of Reimpaargedichte is that they are usually found in miscellany manuscripts. 5 Before being copied into these large collections, many of these texts were probably transmitted in short standalone booklets, although most are no longer extant . During the fourteenth and fifteenth centuries, however, large manuscripts of Reimpaargedichte were quite common. Each codex was produced following certain criteria. In some cases, these criteria are clear, in others they are harder to identify. However, it is fair to assume that the process of copying and adapting these big collections always involved a reason, a goal, and a methodology, even if incidental factors played an important part too.
The corpus for this study comprises 1433 rubrics from sixty-eight manuscripts, which correspond to over eight hundred different works. 6 This sample has been sourced from a number of monographs which compile rubrics of medieval manuscripts and texts (Dahm-Kruse; Klingner and Lieb; Mihm; Moelleken). Additionally, some library catalogues and digital manuscript facsimiles were consulted. 7 The result is not an exhaustive corpus of Middle High German rubrics for Reimpaargedichte, but it includes all the most important manuscripts and can be considered representative. These rubrics have been 4. A recent description and exemplary use of the method can be found in Underwood. 5. I use the term 'miscellany manuscript' to mean a manuscript that collects different texts within one codex. Of course, some of them are collections of very heterogeneous materials (proper miscellanea) while others may be careful selections of texts (collectanea). However, there is no clear dichotomy of practices, but rather a continuum, and we can identify some level of selection in all cases. For this reason, I prefer to use only one of the terms 'miscellany' as a general concept that encompasses all cases.
6. The corpus has been published in the open repository Zenodo: Fernandez Riva. 7. Although it does not include transcriptions of the rubrics, the Handschriftencensus (consulted 04-11-2021) was used as a source for relevant information about the sources and links to digital facsimiles. assembled into spreadsheets with information concerning their palaeographic transcription, an automatically generated lemmatized version, 8 and a standardized modern title of the work. Additionally, the database contains basic information for each manuscript (identification, date of composition, language). For some rubrics, it was not possible to identify a modern title from the sources, so only the rubric and the manuscript are recorded without any other information.
The most common genres in the dataset are Maeren and Minnereden, with over two hundred titles each. Maeren are defined by Fischer as fictional secular narratives with mostly human participants, and can be of many different kinds (for example, humorous or courtly). 9 Minnerede is a late medieval genre with limited narrative content that focuses on love, although there is great variety within it (dialogues, letters, and allegorical dream accounts in the tradition of the Roman de la Rose, among others). The corpus also includes fables, bîspel (akin to the Latin exempla), hagiography and miracle stories. There are also a few texts that belong to completely different genres (epic, lyric poetry) but which have been incorporated in the database because they are attested in the same miscellany manuscripts as the Reimpaargedichte and have rubrics in the same style. It is worth noting that all these genre definitions are modern rather than medieval. The question of whether categories for literary genres appear in the medieval rubrics will be addressed later.
The dataset also includes 382 occurrences of texts which have no rubrics and yet are witnessed in the same miscellany manuscripts. For the purposes of this analysis, these cases are mostly omitted. For lexical comparisons, the lemmatized version of the rubrics are used.

Systemic Reading and the Reimpaargedichte
The attempt to understand Reimpaargedichte by analysing a large corpus and trying to deduce some general principles is not an innovation of this study. This group of texts has been subject to an analytical method I would call 'systemic reading' -a method located at the crossroads of the creation of a catalogue and the history of literature. The main objective of systemic reading is not the interpretation of individual works, but the creation of an organized typology in order to understand the relationships within a complex literary system composed by a considerable number of texts. Between 1967-68, three important books used this approach: Überlieferung und Verbreitung der Märendichtung 8. Generated using the POS-Tagger for Middle High German by Echelmeyer, Reiter and Schulz.

A new edition of the Middle High German Mären is Ridder and Ziegeler.
im Spätmittelalter (Mihm), Studien zur deutschen Märendichtung (Fischer), and Mittelhochdeutsche, mittelniederdeutsche und mittelniederländische Minnereden (Brandis). The curious emergence of these works at virtually the same time might be explained by the need to explore under-researched literary genres that were considered marginal up to that point, in the cultural context of a general search for change.
In the seventies and eighties, many other books addressed the issues presented by the corpus of Reimpaargedichte with a strong theoretical focus. One of their key concerns was differentiating sub-genres and their features (cf. Grubmüller; Haug; Heinzle; Holznagel; Röcke; Strasser; Ziegeler). However, in the last decade, there has been a renewed interest in cataloguing and ordering the Reimpaargedichte as a more efficient way of understanding them. Two main titles with this perspective appeared in the 2010s: Handbuch Minnereden (Klingner and Lieb) and Geistliches Erzählen: Zur deutschsprachigen religiösen Kleinepik des Mittelalters (Eichenberger).
Against the background of this scholarly tradition, distant reading does not offer a revolutionary perspective. Nevertheless, it is a way of pursuing a trend that is already present in the field, but with different strategies and tools that hopefully can build upon the results of previous approaches.

Research on Medieval Rubrics
During the High Middle Ages, rubrication became a central feature of manuscript production. The proliferation of rubrics in Gothic manuscripts is generally considered part of a change in reading culture and habits, particularly due to the modes of reading typical of cathedral schools and universities (cf. Gumbert; Hamesse). However, this process and its relationship to rubrication is still not completely understood.
Rubrics in medieval manuscripts have been studied from multiple perspectives, although it has never been a mainstream field of research. Many studies have shown that rubrics were not a marginal feature of medieval book production, but a very important component which authors, scribes, editors, owners, and readers considered carefully. 10 If we focus only on Middle High German rubrics, there are two important articles worth mentioning here. Backes compares how rubrics were used in French and German romance. Meyer and Zotz, for their part, identify three functions for rubrics in the sources: to separate, to indicate a text's content, and to index a story. These three functions are similar to modern titles in anthologies, but they tended 10. Some of the most relevant articles dealing with medieval rubrics are Busby; Chavannes-Mazel and Brownrigg; Croenen; Dines; Rouse and Rouse; Rudy. to work in quite a different way in medieval texts. The most visible difference is that medieval rubrics were a lot more inconsistent, in the sense that the rubric for the same work could change considerably from one manuscript to the other. In other words, they were not considered as a standardized name for the text, but as something that could be customized by the copyist. In that sense, rubrics are worthy of study, as they reveal the scribe's understanding of the literary text.

Length
The length of rubrics is a basic way to evaluate how much information they carry. The average length of the title given to these texts acts as a useful baseline for comparison. The modern titles for these texts have been constructed by editors and scholars during the nineteenth and twentieth centuries, sometimes based on the original rubrics, but adapted to the features of modern titles -mostly a short nominal phrase.
The average number of words in modern titles for the texts in the corpus is 3.17, while the average number of words in the medieval rubrics for the same corpus is 7.88. Rubrics have a more complex structure and offer more information than the modern titles. The blue crosses on figure 3 indicate for each manuscript in the corpus, dating from the late thirteenth to the early sixteenth century, the average length of all the rubrics it contains. There does not seem to be any standard length, as they vary widely depending on the manuscript. However, the rubrics in all manuscripts but one are consistently longer than the average length of the modern titles for the same works.

Level of Variation
Another interesting measure is what I call the level of variation (LOV), which measures how stable or variable rubrics are for the same text in different manuscripts. This calculation is important in order to understand in which cases scribes were more creative and in which they just copied their sources. The LOV is based upon the 'Levenshtein distance', a common measure to quantify the dissimilarity of two strings. 11 This metric computes the minimum number of single-character edits (insertions, deletions or substitutions) required to change one word into another. The LOV used in this article calculates an average of the Levenshtein distance for all possible combinations of the rubrics for the same text, in order to assess the variation experienced by the rubrics for one work regardless of how many textual witnesses are extant. In other words, if a text appears in three different manuscripts with one rubric each (A, B, C), then we would need to calculate the individual distance of the sets (A, B), (A, C) and (B, C), then divide by three. However, it is also useful to consider the length of the title to get a better measurement of the actual variation. The change of one word in a two-word title is more significant than one word in an eight-word title. For that reason, the complete function to calculate the LOV proposed here is: the average of the Edit Distance between all possible combinations of the rubrics, divided by the Average Length of those same rubrics. This can be represented in the following notation, where S is the set of all rubrics for the same title, d is the Edit Distance function and L is the set of lengths of all rubrics for the same title.
The bigger the resulting LOV, the more variable and unstable the rubric is in its transmission. A LOV with value zero means that all rubrics for that text are identical (all values of d(x), and therefore D̅ , are equal to zero). For the calculation, I use the lemmatized version of the titles to avoid considering spelling variations as actual differences. Some examples of how the LOV reflects the changes can be seen here: 11. This measure is used, for example, by collation tools such as CollateX (consulted on 4 November 2021). I implemented the Levenshtein Distance in Python following the code in Wikibook (consulted on 4 November 2021).
For this analysis, only texts attested with at least two rubrics in the corpus can be considered, which means using a sub-corpus of only 284 texts. In general, there seems to be no single principle guiding the amount of change to the rubrics. Some scribes just copied the rubrics as they were in their source, some changed them partially, and some completely. However, some behaviours were more common, as shown in figure 4. To create this plot, the results of the LOV were rounded up to the second decimal. A large number of works (eightyeight) have a LOV between 0.1 and 0.6, which means that they change only partially -the rubrics were not completely rewritten, but merely adapted to different degrees. However, there are two individual peaks in the distribution, at 0 (twenty-three works) and at around 0.7 (twenty works), which means that the most common approaches were either to leave the rubric unaltered or to change about half of it. There is another small peak close to 0.9 (seventeen works), which represents changing the rubric almost entirely. A bigger corpus, with more balanced rubrics per manuscript, could improve the results of this exploratory analysis.  Partial changes (between 0.1 and 0.5) are particularly interesting when we consider manuscripts that are closely related, like Cpg. 341 and Bodm. 72. 12 These two manuscripts are very similar and the latter is probably a direct copy of the former. Many of the rubrics for the same works are extremely similar in both manuscripts, but small and noteworthy details were changed. Both have rubrics in the form of rhymed couplets and altering them involved some degree of poetic ingenuity: 13 The example of Marien Rosenkranz is particularly telling of the scribe's agency in the form of minor interpretative edits. As the preceding text, Thomas von Kandelberg, deals with students and its rubric mentions twelve students, the scribe in Bodm. 72 adds "about another student" and not just "about a student" to Marien Rosenkranz. In this way, perceiving a thematic thread in this section, the scribe established an explicit link between the two texts. This is also a clear example of the rubrics' paratextual function, as a speech act between the scribe and the reader. It is also possible to identify other general scribal attitudes in rubrics. For example, in the case of Frauenlist, Bodm. 72 has a title that satirically expands on the women's trickery and cunning, which are the subject of the story ("This book is called women's cunnings, which have many forms"). However, in Cpg. 341, which is characterized by a more clerical attitude (cf. Stutz), the second verse only contains a general religious formula ("God, the Lord, is with us"). Unlike the previous examples, in many cases the rubrics in different manuscripts changed radically. Thirty-two texts have rubrics with a LOV higher than 1, while twenty-eight have a level of variation between 0.9 and 1. Many cases with very high LOV are titles that changed fundamentally between Cpg. 341 and Bodm. 155. These two manuscripts contain many of the same texts in similar order and must therefore be related. However, the scribe of Cpg. 341 composed all rubrics in rhyming couplets while that of Bodm. 155 used summarized and concise phrases. In other words, there can be distinct principles that different scribes followed when creating rubrics for the same collection of texts and based on the same sources.

Vocabulary
To understand the use of language in the corpus of rubrics, lexical choices are key. The five most frequent lemmas in the entire corpus are: 1-sîn (408): verb, 'to be' in the infinitive 2-diser (376): pronoun, 'this' (many variant declinations and spellings possible, ditz, diz, etc.) 3-daz (186): article, 'the' 4-mere (157): noun, 'story/narrative' 5-hie (112): noun, 'here' Using only the four most frequent words (MFW), it is possible to compose the beginning of a typical rubric: ditz ist daz maere… ("this is the story…"). It is possible to arrange the most frequent words by manuscript, which offers a more precise overview. A sample of the spreadsheet with that information is shown in Figure 5. Each lemma is paired with the number of times it appears in the rubrics of the corresponding manuscript. The lemmas are ordered from most to least frequent. As expected, the MFW are usually function words (articles and prepositions) that are very common in any given sample of language. However, some prominent non-function words occur significantly above average. Furthermore, even the common function words can offer some insight into the structure and semantic of rubrics. There are four types of words that feature prominently in the spreadsheet: the word von; demonstrative pronouns ditz/diser; copulative and other verbs that allow characterization or identification; and words that refer to the 'genre' of the text.
The best way to visualize the importance of these terms is to convert the spreadsheet into a plot as shown in figure 6. The x-axis represents the manuscripts (the columns of the spreadsheet), ordered by approximate date of composition. The y-axis consists of the numbers 1 to 14 in reverse order, denoting the word's rank among the MFW of each manuscript, as indicated by the points on the graph. A colour and shape is assigned to each of the four categories of words introduced above. In this way, it is possible to compare and easily visualize the relevance of the MFW according to their type.
The word von is at the top of the list for most manuscripts. This preposition was used to indicate the subject of the text and could be translated as "of " or "about," like the de so common in Latin rubrics. Some rubrics have an introductory statement like ditz ist von ("This is about..."), but many others just offer a prepositional phrase like von dem ritter mit der halben bir ("About the knight with the half pear") or von dem armen Heinrich ("About poor Henry"). Unlike modern titles, scribes considered it meaningful to use this preposition in rubrics, making it explicit that the rubric mentions the subject of the text. This practice means that rubrics did not only have the function of describing the content of the text, they also clearly stated the function itself.
Very prominent in most manuscripts, ditz/diser was a demonstrative that worked almost as a deictic marker pointing at the text: ditz ist das mere von… ("This is the story of..."). The word linked the rubric to the text and the reader to the page -ditz denoted the text as it was inscribed in front of the reader. Modern titles conventionally do not explicitly point to the text with a demonstrative in this way. Medieval rubrics, on the contrary, were constructed as if they were directly addressed to the recipient, separated from the text and pointing at it. In other words, modern titles and medieval rubrics have some similar functions, but rubrics made that function explicit by the use of prepositions and demonstratives.
Just like von or ditz, there are some verbs, more common in the earlier manuscripts but still relevant in the later ones, that explicitly proclaim that the rubric is the place where the text is being named or described. It is common to encounter sîn (to be), usually combined with von or some similar structure to introduce the topic of the text ("This is about…"). Not as common, but also present, is the verb heißen ("to be called"), which assigned a title or name to the text. Finally, the expression Hier hebt sich an, which could be loosely translated as "here begins" (literally "here X raises itself "), is also frequent. In this case, the expression has a very strong demonstrative value. The hier refers to the semantic universe of space and the page more than a performance situation, which could be implied with jetz ("now").
With verbs like heißen or sich anheben, rubrics perform a complex speech act that establishes a name for the text. A rubric like Diß büchlyn heißet der dogende krancz / Vnd leydet zu des hymmels dancz ("This small book is called the garland of virtue / and leads to the dance of heaven") is constructed as a statement that includes the name of the text: der dogende krancz, the garland of virtue. These kind of rubrics are common, especially for Maeren.
There is a perceptible distinction between rubrics that provide a name for the text, commonly introduced by the verbs heißen or sich anheben, and rubrics that provide keywords, commonly introduced by the preposition von or the verbs sîn or sagen. A clear example of the latter is the expression von der minne ("about love"): "Dis mére seit von der minnen" (Frau Minne warnt vor Lügen in Straßburg, Cod. 94), "Dise mére ist von der minnen" (Sekte der Minner in idem), "Dise mére mahte Meister gotfrit von strazburg vnd seit von der minne" (Herzmaere in idem), "Von der minne" (Ochse und Hirsch in the Leone manuscript), "Ein spruch von der mynn" (Der unentwegte Liebhaber in Cpg. 313).
Finally, there are many words that might designate 'genre:' mere, spruch, rede, buoch, bispel. 14 The first, mere, is a very common MHG word which means narrative, news, story, something that someone tells. It is the origin of the modern genre category Maere. Originally, it had the meaning of 'well known/famous' and derived to something that is worth telling. Spruch is related to sprechen, speak; it means something that is said. Rede, related to the verb reden, also means to speak or talk. However, it can be used to define a language or the general ability to utter speech. Buoch derives from the same root as English 'book' and originally designated the surface upon which something is written. In this way, considering etymology, buoch comes from the world of writing, while mere, spruch, and rede from that of orality. Finally, bispel (less common than the others) is the Middle High German translation of Latin exemplum and was used for short narrative didactic texts.
Mere is the most common of these terms with 181 occurrences. The use of spruch and buoch are the next most popular, albeit trailing considerably with sixty-two and sixty incidences respectively. Rede is used in only twenty-two rubrics and bispel in just three cases. It is unclear if there were any overarching principles directing their use.
A possible answer appears when considering the approximate date of production of the different manuscripts. Mere appears mostly in the earliest manuscripts, while the word spruch and rede appear mostly in the later medieval productions. Buoch is evenly distributed. Bispel is not considered as there are only three cases in the whole corpus.
In other words, the period of production seems to have had more influence on the lexical choices to refer to the work than did its genre. Maeren were more common up to 1400, Minnereden afterwards. There are many examples of works which were described as mere in early manuscripts and as spruch in later ones, contradicting a possible use of those terms to identify genres. For example, the rubric for the Maere "Des Mönches Not" was "Ditz ist ein schoner mere gnvc / Wie ein mvnch ein kint trvc" in Cpg. 341 (first quarter of thirteenth century) and "Ain ander spruch" in Cgm. 5919 (composed 1500-10). The words mere, spruch, buoch and rede did not generally designate genre, but were synonyms for 'text', and their use varied according to time and place. Of course, it is possible that some manuscripts did use these and other words in a systematic sense to indicate different kinds of works. For example,  has postulated the hy-14. I normalize the spelling not according to the standard Middle High German, but to the prevalent forms in the palaeographic transcriptions of the rubrics. This also helps avoid confusion between mere (term used in the rubrics) and Maere (modern genre designation). pothesis that Cpg. 341 only used variants of buoch (bvch or bvchel) for the works that were transcribed from standalone copies. The only word that (in the few occasions in which it appears) has a clear meaning related to its genre is bispel. The didactic narratives inspired by Latin exempla were clearly identified as distinct from mere. For example, the Leone manuscript introduced the compilation of short texts under the rubric: daz buoch nennen die werlt daz sagen von bispeln und von mern ("the book is called the world and deals with bispeln and with meren").
Apart from function words and the four terms for 'text' just discussed, other lexical items, such as adjectives and nouns, are worth studying, as they usually refer partly to the subject of the texts. The most common of these in the corpus are vrouwe (woman, 93); got (god, 61); schoene (beautiful, 60); ritter (knight, 59); guot (good, 58). That women were the most predominant subject is not surprising. This trend stems from the centrality of women and love in Minnereden, but is also due to the presence of some religious Marian poetry in the corpus. Knights also figured prominently, representing a widely occurring character, although far behind vrouwen. God is in second place, which was the result of the amount of religious literature and the use of invocations not related to the content in the rubrics.
However, one of the most interesting words in the list is schoene ("beautiful"). This word was used in two different ways. There are titles like von ein schoene vrouwe ("about a beautiful woman"); ditz ist von der gans / daz was ein schoene jvncvrowelin ("this is about a goose, that was a beautiful maiden"). Here schoen was used to describe characters or events in the text. However, more common is the expression ditz ist ein schoene mere ("This is a beautiful mere"), as in Ditz ist ein schones mere / von einem ritter lobere ("This is a beautiful mere / about a praiseworthy knight"). In these cases, the word did not refer to the content of the text; it highlighted the function of the rubric as an evaluative instance. The rubric was a place where literary works were judged and reviewed. Like the demonstrative pronoun ditz, the schoen increased the distance between rubric and text -the rubric was part of a different voice evaluating the literary work.

Authorship
For the most part, rubrics do not mention author names. Reimpaargedichte are either anonymous or contain the name of the author in the text itself, especially in the prologue or epilogue. The authori-al attribution of the same text might even differ in different witnesses. In some special cases, however, rubrics identify the author of the works, which implies that these poets were particularly well known. The most common case is Der Stricker. He was sometimes mentioned in the rubric in addition to the name for the corresponding text, as in Hie sait der strickere von dem / konige Salomone ("Here Der Stricker talks about Salomon," Bodm. 72). In other cases, the author's name is mentioned, but the text itself is described only in very general terms or not at all: Hie ist des Strickers mere / Got bvz vns vnser swere ("Here is a story from Der Stricker; let God heal our pain," Cpg. 341). This suggests that he was a known author and his name was even more important than the identification of the text itself.
Another interesting case regarding authorship is the German translation of the Disticha Catonis, one of the most canonical texts in Western Europe during the Middle Ages, as it was widely used in educational contexts to teach Latin and as a source of moral instruction. This text was usually mentioned in the rubrics as Cato, to whom the collection was unjustly ascribed in the Middle Ages. The name of the alleged author was used as the name of the text itself, which was not an uncommon metonymy for famous works in the Middle Ages. In our corpus, it also happens to Freidank, whose work, even if not as famous as Cato's, was sometimes just referred to by his name -for example, der fridang in the Leone manuscript. One of the rubrics for Cato in our corpus specifically referred to the use of the text in educational contexts, which can be read as a mechanism of referring to the canonical status of the Latin source: Ditz buoch heizet Katho / vnde liset man ez in der schvele do ("This book is called Cato, and is read in school," Cpg. 341)

Conclusions
In this article, I have tried to implement a distant reading approach to MHG rubrics in miscellany manuscripts. This methodology indicates that length and syntactic structures were highly variable features of these rubrics. The attitude of the scribes towards the rubrics in their sources was also not uniform. There were many common practices: copying the rubric as it was, rephrasing it, or replacing it completely. In general, it is possible to conclude that the diversity was considerable and specific conventions for particular times, plac-es and even for each manuscript were normal. In this sense, each witness deserves to be evaluated as a specific case.
Nevertheless, there is one thing that remains consistent in most manuscripts: the general principle of constructing rubrics in such a way that they are explicit about the communicative act between scribe and reader they are performing. In other words, rubrics were the place where scribes deliberately presented the text to the readers. This makes sense when the context of creation and the performative function of rubrics are considered. In contemporary literature, title and text are the creation of the same author, so the communicative act between author and reader starts with the title. In medieval manuscripts, on the contrary, the rubrics were not part of that communicative act between author and receiver. Instead, scribes assumed the role of intermediaries between author and reader, and they introduced the work. Rubrics were external to the text and one of their functions was to make this communicative act between scribes and readers explicit. They are the place where scribe and reader meet.
Going back to the original question of what information regarding the status of these texts can be extracted from rubrics, it is possible to draw some conclusion from our analysis. 1. Rubrics do not identify literary genre. The words used to refer to the texts changed with time and did not correlate with internal textual features. There might have been some implicit conception of different genres that elicited different kinds of attitudes, but rubrics do not provide any information in this regard. 15 2. The level of variation suggests that the texts were not easily identifiable by a particular name. However, in many cases, even when the rubrics changed, some important keywords that serve the purpose of identifying the text remained consistent. Furthermore, the use of verbs like heizen or sich anheben points to some concept of name or title for the text. In this respect, there is also a fundamental transition between the early Maeren and the later Minnereden. The manuscripts of the thirteenth and fourteenth centuries that contain predominantly Maeren use more precise words to identify specific texts. The later manuscripts, containing mostly Minnereden, do not possess this feature. They often display generic constructs which they applied for different texts, sometimes with the sole purpose of enumeration. This suggests that the narrative Maeren were, in a way, more 15. An analysis of these terms in Middle High German texts was performed by Düwel.