On the History of the Practice of Fictionality – and the Recurring Problems in its Investigation

Abstract In recent decades, research into the history of fictionality has seen a significant upturn in interest. One promising theoretical foundation for such investigations appears to be the approach commonly known as the »institutional theory of fictionality«. This is based on the premise that fictionality is a rule-based practice determined by conventions which are variable (both synchronically and diachronically), conventions to which authors and readers alike feel committed. The main advantage of this particular theory of fictionality, as far as an analytical approach to the history of fictionality is concerned, is the following: The institutional theory of fictionality is suitable for taking into adequate account the historical variability of terms, concepts and practices by providing a theoretical framework that may be filled with a wide variety of different (kinds of) content. In this way, one may sidestep the danger of examining the history of fictionality in an anachronistic manner, imposing on past times and practices the expectations of a modern perspective. Still, committing to an institutional theory of fictionality avoids only some of the problems all research on the history of fictionality faces. The aim of this article, therefore, is to point out those difficulties which cannot be avoided in such investigations even in the arguably best theoretical conditions of an institutional account of fictionality. To this end, instead of providing an overview of previous research or addressing specific methodological, conceptual or logical problems related, the present essay focuses on recurring and widespread difficulties inherent in both the object of investigation and the various methods of investigating it. The essay is divided into three sections. In the first, a number of problems are addressed that exist regardless of the specific method of investigation chosen. Most epistemological problems result from the fact that written documents must be consulted to make inferences regarding the conventions and practices of the past. In this context, it is not only the sparse tradition that becomes an issue (especially for more remote historical periods) but also the fact that no analysis of written materials can provide direct insight into past practices. Since any social practice, moreover, is in itself a highly complex matter that can hardly be broken down and understood in all of its many aspects – even from an interdisciplinary perspective, which anyway implies its own difficulties such as a frequent lack of uniform terms, et cetera –, such research will only be able, as a matter of principle, to approach past practices more or less closely. Following these general reflections, the article critically examines the two most prominent methods used by those investigating the history of fictionality as an »institution«. These are the analysis of literary texts, on the one hand, and that of poetological texts, on the other. When trying to draw conclusions from literary texts about past practices of fictionality, the focus of much recent research has been on the search for »signposts of fictionality«. The problem with this method is not only that such studies are often at risk of presupposing a positive test result – after all, signposts of fictionality only make sense if a practice of fictionality has already been established – but also that signposts of fictionality are historically variable. For this reason, one cannot simply postulate the validity of present-day signposts of fictionality for historical texts, and conversely, one must also reckon with the fact that other, corresponding signals unknown to us will remain beyond our knowledge. In addition, there is also the more general question of just how different two different practices may reasonably be said to be in order for them to come under the common rubric of a shared »practice of fictionality«. One advantage the analysis of poetological texts appears to have over conclusions drawn from literary texts is that insofar as poetological texts are already meta-textual in nature – as they are texts about literature –, the aforementioned »detour« via an analysis of signposts of fictionality is no longer required. Even such studies, however, are faced with several problems: To begin with, poetological texts are predominantly conceived of as instances of programmatic – and thus as normative, not descriptive – writing. It therefore immediately suggests itself that they should articulate practices desired or demanded rather than depict existing usage. Secondly, poetological texts are written artefacts that, for a very long time, were circulating within a predominantly oral culture. It is therefore arguable whether and to what extent that predominantly oral practice is reflected in poetological texts. Thirdly, poetological texts do not discuss the concept of »fictionality« but, first and foremost, that of »poetry«. The fact that a strongly evaluative component – namely, debates over the value of poetry – is often at the centre of such texts allows the conclusion that what is being negotiated there, rather than an earlier notion of »fictionality«, is an equivalent of the modern concept of literature. By contrast, it seems indisputable that various ways of differentiating between types of texts were, in fact, developed from the earliest times. Fourthly, and considering the fact that in those contexts, debates mainly revolved around such categories as the »truth« and »probability« of a given story or the »inventedness« (i. e., the fictitiousness) of its contents, the question arises, once again, whether these are indeed practices of fictionality we are looking at. This article makes a case for delineating historical terms and practices as accurately and in as much detail as possible, rather than presenting them rashly and reductively, perhaps, as early forms of the institution of fictionality

(especially for more remote historical periods) but also the fact that no analysis of written materials can provide direct insight into past practices. Since any social practice, moreover, is in itself a highly complex matter that can hardly be broken down and understood in all of its many aspects -even from an interdisciplinary perspective, which anyway implies its own difficulties such as a frequent lack of uniform terms, et cetera -, such research will only be able, as a matter of principle, to approach past practices more or less closely.
Following these general reflections, the article critically examines the two most prominent methods used by those investigating the history of fictionality as an »institution«. These are the analysis of literary texts, on the one hand, and that of poetological texts, on the other. When trying to draw conclusions from literary texts about past practices of fictionality, the focus of much recent research has been on the search for »signposts of fictionality«. The problem with this method is not only that such studies are often at risk of presupposing a positive test result -after all, signposts of fictionality only make sense if a practice of fictionality has already been established -but also that signposts of fictionality are historically variable. For this reason, one cannot simply postulate the validity of present-day signposts of fictionality for historical texts, and conversely, one must also reckon with the fact that other, corresponding signals unknown to us will remain beyond our knowledge. In addition, there is also the more general question of just how different two different practices may reasonably be said to be in order for them to come under the common rubric of a shared »practice of fictionality«.
One advantage the analysis of poetological texts appears to have over conclusions drawn from literary texts is that insofar as poetological texts are already meta-textual in nature -as they are texts about literature -, the aforementioned »detour« via an analysis of signposts of fictionality is no longer required. Even such studies, however, are faced with several problems: To begin with, poetological texts are predominantly conceived of as instances of programmatic -and thus as normative, not descriptive -writing. It therefore immediately suggests itself that they should articulate practices desired or demanded rather than depict existing usage. Secondly, poetological texts are written artefacts that, for a very long time, were circulating within a predominantly oral culture. It is therefore arguable whether and to what extent that predominantly oral practice is reflected in poetological texts. Thirdly, poetological texts do not discuss the concept of »fictionality« but, first and foremost, that of »poetry«. The fact that a strongly evaluative component -namely, debates over the value of poetry -is often at the centre of such texts allows the conclusion that what is being negotiated there, rather than an earlier notion of »fictionality«, is an equivalent of the modern concept of literature. By contrast, it seems indisputable that various ways of differentiating between types of texts were, in fact, developed from the earliest times. Fourthly, and considering the fact that in those contexts, debates mainly revolved around such categories as the »truth« and »probability« of a given story or the »inventedness« (i. e., the fictitiousness) of its contents, the question arises, once again, whether these are indeed practices of fictionality we are looking at. This article makes a case for delineating historical terms and practices as accurately and in as much detail as possible, rather than presenting them rashly and reductively, perhaps, as early forms of the institution of fictionality Keywords: literary theory, fictionality, history of fiction In the past few decades, a lot of time and effort have been spent on attempting to investigate the history of fictionality.1 Countless articles and monographies have been written on the existence as well as the absence of fictionality in different historical eras. Considering the reliance of these investigations on diverging conceptual and methodical premises, it is hardly surprising that very heterogeneous conclusions have been reached. As Jan-Dirk Müller points out with respect to the debate on fictionality in medieval times, for example, »different terminological concepts and levels of argument were [often] not sufficiently differentiated […]. Therefore, the discussion is reasonably confusing, and one cannot speak of an assured basis of what one could call ›fictional‹ in the Middle Ages«2 (Müller 2004, 283). This situation is deplorable not only in regard to the rather disparate analyses in research on fictionality in medieval times, but even more so because Müller's findings apply to debates on fictionality in other historical periods as well.
There have, nevertheless, also been developments in the field of theories of fictionality that give rise to hope for a better understanding of the history of fictionality. In the past few years, more and more scholars of philosophy, aesthetics and literary theory have established a consensus that fictionality is an institutional phenomenon -and this seems to be an extraordinary asset in the endeavor to write a history of fictionality. Peter Lamarque and Stein Haugom Olsen spell out the core idea of this approach as follows: »the fictive dimension of stories (or narratives) is explicable only in terms of a rule-governed practice, central to 1 Whenever I use the term »fictionality« (or the corresponding adjective »fictional«), I refer to a property of texts, whereas whenever I speak of »fictivity« (or correspondingly: »fictive«), I refer to a property of objects. 2 »[V]erschiedene Begriffskonzepte und Argumentationsebenen wurden nicht hinreichend unter schieden […]. So ist die Diskussion einigermaßen verwirrend, und von einer gesicherten Grund lage dessen, was man im Mittelalter ›fiktional‹ nennen könnte, wird man nicht sprechen kön nen.« All translations -here and in the following -are mine.
which are a certain mode of utterance (fictive utterance) and a certain complex of attitudes (the fictive stance)« (Lamarque/Olsen 2002, 32, emphasis original).3 In other words, fictionality is neither a feature of a text itself, nor can the author or the reader decide on the fictionality of a text single-handedly. Instead, fictionality is considered to be a social practice governed by certain rules shared by both authors and readers.4 The benefit of an institutional theory of fictionality for an analysis of the history of fictionality is obvious: the institutional theory provides a framework whose content can be filled flexibly. When we try to reconstruct former practices of fictionality, we therefore do not have to hold on to the modern characteristics of fictionality and impose them on historical texts and times since the institutional theory of fictionality explicitly expects that the rules constituting the practice authors and readers have agreed on can vary synchronically as well as diachronically. Thus, we run into considerably less risk of working with anachronistic concepts and simply forcing our modern perspective upon past times.5 In general, an institutional approach hence seems to be a very auspicious way to come to terms with the notoriously difficult analysis of the history of fictionality. An institutional theory of fictionality not only promises the creation of a wider consensus on the phenomenon in question, but also seems to be the best method for doing justice to the object of investigation.
Yet, even advocators of the plausible assumptions of an institutional theory of fictionality -and throughout recent years the debates on the history of fictionality have actually increasingly turned towards this approach -have persistent difficulties with an analysis of the history of fictionality. The main goal of this paper is to demonstrate the problems confronted by this endeavor, even in the arguably best possible conditions of an institutional account of fictionality. With this in mind, I do not want to criticize or make a statement concerning particular debates on specific historical times. I also neglect all methodical, conceptual or logical problems related to individual papers as well as particular established lines of argument in one or the other contributions. Instead, my aim is to point 3 Note that Lamarque and Olsen use the terms »fictional« and »fictive« inverted in comparison with my own usage. 4 An institutional theory of fictionality is thus pragmatic insofar as it conceives fictionality as a rule-governed practice agreed and acted upon by authors and readers. That does not, however, result in the according rules only paying attention to pragmatic aspects of discourse (e. g. speech-acts). In principle, one can easily say that fictionality is an institutional matter, but that the current rules and conventions of fictionality are mainly referential, i. e. that texts are considered to be fictional if and only if they are about fictive objects. 5 For criticism of anachronism, cf. Müller 2004, 283;Gallagher 2006, 338;Meincke 2007, 332 sq. For objections to criticism of anachronism, cf. especially Orlemanski 2019. out difficulties shared by all research into the history of the institution of fictionality.
To this end, I first address several general problems that exist independent of the particular method of investigation. After that, I turn to two of the main procedures used to clarify the question of the history of fictionality: the method of drawing conclusions from literary texts, on the one hand, and from poetological texts, on the other.6 I illustrate the problems arising from an analysis of literary texts with examples taken from the debate on fictionality in the Middle Ages and exemplify the difficulties resulting from an analysis of poetological texts with instances taken predominately from the debate on fictionality in classical antiquity. Once again, it is important to stress that I do not want to present a survey of these particular discussions, but rather to identify recurring problems that emerge from the endeavor of writing a history of the institution of fictionality. Individual contributions that I discuss serve only to demonstrate general issues.

General Problems
Major difficulties in writing a history of the institution of fictionality result from us obviously not being a part of historical institutions of fictionality and therefore having to rely on documents to reconstruct the conventions of former practices. The empirical and methodical challenges emerging from this situation essentially amount to epistemic problems. Firstly, there is the empirical difficulty that, the further we go back in history, the less source material is available to us -and, the less source material at our disposal, the worse our chances of getting an accurate picture of the former conventions and practices of fictionality become. Apart from that, there is, secondly, the methodical challenge of being bound to deduction of former practices of fictionality from texts. Even with our access to literary and poetological texts, perhaps also to reviews that provide information about the contemporary reception, these textual sources obviously do not provide direct insight into former practices of fictionality. As Sonja Glauch puts it with regard to the Middle Ages: Medieval prose literature is still highly situational and rooted in the orality of the recitation. Therefore, the situation of the reception is more significant as a potential signpost of fictionality for medieval literature than for the book-centered modern literature. Of course, the situativity of recitation or reading as well as the other pragmatic framework conditions remain largely inaccessible to today's analysis.7 (Glauch 2014, 406) As underpinned by Manuel Braun, »[t]he practice of medieval literature is […] almost inaccessible to us since it is lost with the situation of the recitation«8 (Braun 2015, 93 sq.). Thus, the open question remains of how close our conclusions from the conveyed texts available to us can ever get to the actual former practices of fictionality. Thirdly, even if the existing sources were maximally informative and numerous, there would still exist the epistemic impossibility of knowing and taking into consideration all the factors relevant for an assessment of a social convention as fictionality in the manner conceived of by institutional theories. Tilmann Köppe is no doubt right in stating that a »comprehensive characterization of a social practice/institution must […] ultimately describe a whole ›way of living‹«9 (Köppe 2014, 420). So, in a nutshell, we are certainly not capable of achieving a complete survey of former institutions of fictionality -even working with an abundant amount of sources and in interdisciplinary groups of research -, and that is why we should concede right from the beginning that our insights will at best be approximations of the actual former practices.
Furthermore, conceptual-methodic problems add to the difficulties already named. One main challenge of an investigation into the history of the practice of fictionality results directly from the fact just outlined, namely that an analysis of a social practice is inevitably an interdisciplinary matter. Interdisciplinary research, however, is often afflicted with the problem that different sciences use different concepts -or that they use the same concepts, but with different meanings. Jan-Dirk Müller, for example, has compiled a long list of recurring conceptual errors made by interdisciplinary analyses of fictionality in medieval times: The opposition of fictional/non-fictional is still mixed up with the opposition of true-right/ untrue-false, especially in studies related to historical science; these consider presentations of historical events that are biased by certain interests to be ›fiction‹. The difference of ›fic-tivity‹ and ›fictionality‹ is still neglected -especially in the debates between historians and 7 »Die Erzählliteratur des Mittelalters ist noch stark situativ und in der Vortragsmündlichkeit ver ankert. Deshalb ist die Rezeptionssituation auch als mögliches Fiktionalitätssignal für die mitt el alterliche Literatur bedeutsamer als für die buchzentrierte Literatur der Moderne. Freilich blei ben die Situativität des Vortrags oder der Lektüre und die sonstigen pragmatischen Rahmenbedingungen der heutigen Analyse weitgehend unzugänglich.« 8 »Die Pragmatik der mittelalterlichen Literatur ist für uns […] nahezu unzugänglich, da sie mit der Situation des Vortrags verloren ist.« Cf. also ibid., 106. 9 »Eine umfassende Charakterisierung einer sozialen Praxis/Institution muss […] letztlich eine ganze ›Lebensform‹ beschreiben«. literary scholars. The concept of ›fictionality‹ measured by the sense of modernity is translated into medieval times. As an antipole to ›fictivity‹, a concept of ›reality‹ is assumed that is essentially geared to so-called ›facts‹ and that blocks out the relation between fictional speech and imaginary arrangements, and, finally, the problem of fictionality is almost exclusively treated as a literary or, even more narrowly, as a narratological problem.10 (Müller 2004, 283) These are, of course, no categorical difficulties that could not be solved as a matter of principle. It is, however, important that these difficulties arise not only in particular papers that may not be well-informed or well-conceived. Consequently, I consider finding a common conceptual basis to be one of the main challenges in interdisciplinary research. As Müller's examples illustrate, agreement on the aspect that fictionality is an institutional matter solves only one of the many conceptual problems related to an analysis of the history of (the practice of) fictionality.
In fact, the aforementioned problem actually becomes greater when we take into consideration the additional fact that, even among literary theorists or philosophers of literature, no consistent answer is found to the question of what the conventions of our contemporary practice of fictionality actually are.11 How can we ever expect to find a convincing answer to the question of the history of the institution of fictionality, which can only be reconstructed interdisciplinarily and 10 »Die Opposition fiktional/nicht-fiktional wird, zumal in geschichtswissenschaftlichen Untersuchungen, immer noch mit der wahr-richtig/unwahr-falsch verwechselt; man betrachtet partei liche oder interessengebundene Darstellungen historischen Geschehens als ›Fiktion‹. Immer noch wird -zumal in der Debatte zwischen Historikern und Literaturwissenschaftlern -die Unter scheidung zwischen ›Fiktivität‹ und ›Fiktionalität‹ vernachlässigt. Ein an der Moderne abgelesener Begriff von ›Fiktionalität‹ wird auf das Mittelalter übertragen. Als Gegenpol zum ›Fiktiven‹ wird ein Begriff von ›Wirklichkeit‹ vorausgesetzt, der im wesentlichen an sog. ›Fak-ten‹ orientiert ist und das Verhältnis fiktionaler Rede zu imaginären Ordnungen ausblendet, und schließlich wird das Fiktionalitätsproblem nahezu ausschließlich als ein literarisches, sogar enger noch: als ein narratologisches Problem behandelt.« 11 Let me mention two examples. There is no agreement on the question of whether the fictivity of at least one object of a text is, according to the current conventions of fictionality, necessary for a text to be fictional (in other words, if the author has to create at least one fictive object to signalize to the reader that she has to receive of the text as fictional). There is also no agreement on the question of which activity the reader is actually supposed to engage in when she receives of a text as fictional. Is she, for instance, only expected to engage imaginatively with the text or also emotionally? The divergence of the opinions on the nature of fictionality increases even more, of course, if we take into consideration the opinions not only of the scholars who believe that fictionality is a matter of institution, but also of those who think that fictionality is solely determined by the intentions of the author or by the reception of the reader or even of those who consider fictionality as a feature of the text itself. from texts, if we cannot even fully agree on the basic rules of our current practice -a practice that we are actually part of? This difficulty is even more pressing if we want to investigate not only when the institution of fictionality was established for the first time, but also when the modern practice of fictionality started. We will probably have a hard time answering the first question, but, as long as we cannot agree on the conventions of our current institution of fictionality, there seems to be no way to come to a mutual agreement concerning the second question. Under the given circumstances, we simply do not agree on the question of what exactly we are searching for in the past when we try to establish the beginning of our modern practice of fictionality.12

Problems Concerning Inferences from Literary Texts
Apart from the general problems just outlined, difficulties also arise from one of the main ways to investigate former practices of fictionality: an analysis of literary texts. A method chosen by many academics in this context is the search for signposts of fictionality. In her paper on ›Fictionality in the Middle Ages‹, Sonja Glauch lists and contextualizes the six textinternal signals most prominently discussed in the respective research: »fantasticism«/unreality, framings (dream, vision), the difference between narrator and author, narrative peculiarities, metafiction/metanarrative utterances and affirmations of truth/references to sources (cf. Glauch 2014, 399-406).13 All in all, Glauch comes to a very sceptic conclusion in regard to the informative value of all of these textinternal signs (cf. ibid.). In my opinion, this is hardly surprising since the whole method of searching for signposts of fictionality is confronted with serious problems -and it is important to note that the reasons for this skepticism not only apply to the discussion of signposts of fictionality in medieval literary texts, but are also decisive for the method of drawing conclusions about former practices of fictionality from textinternal signals in general. In my opinion, the three main problems related to this method are the following: Firstly, whenever we look for signposts of fictionality in premodern literary texts, we are always exposed to the danger of arguing in a question-begging way: a search for signposts of fictionality presupposes (at least in some ways) that an institution of fictionality has already existed. It is, obviously, only in the context of an already existing practice of fictionality that the author has reason for indicating to the readers by signposts of fictionality that the respective text should be considered a fictional text. Without a corresponding institution, signposts of fictionality would be superfluous. Searching for signposts of fictionality harbors the danger of presupposing from the beginning what is supposed to be demonstrated in the end: that a practice of fictionality existed in the respective time period.
Secondly, even if we succeed in avoiding the danger of question-begging, the answer to the question of which textinternal features actually serve as signposts of fictionality is, of course, just as historically variable as the practice of fictionality itself. This has the following consequences with regard to the search for signposts of fictionality in historical literary text: on the one hand, we are, of course, not allowed simply to presume that the signposts of fictionality of today were the signposts of fictionality in former times. So, even if we find signals in historical literary texts that we would consider as signposts of fictionality today, we may not simply conclude that the same signals were considered as signposts of fictionality back then. Harald Haferland, for example, argues that the »incredibility of parts of the arsenal of characters« -an aspect that our current practice of fictionality clearly assesses as a sign of fictivity and therefore as a signpost of fictionality -»is […] a characteristic of the genre [of the text] and not necessarily a sign of fictivity«14 (Haferland 2014, 77, fn. 121). Therefore, at least from textual evidence alone, we hardly stand a chance of deciphering the concrete meaning of the textinternal features -and, the further we go back in history, the less often illuminating authorial or receptive testimony is handed down to us. On the other hand, we also have to expect that some of the textinternal signals that served as signposts of fictionality in former times will remain hidden to us because, from our perspective, they simply do not come to mind as possible signposts of fictionality.
Thirdly, let us suppose, for the sake of argument, that the integration of fictive or phantastic elements (to use a promising candidate) actually was considered to be a signpost of fictionality in former times. The problem is then that the concepts of fictivity and phantasticism are, of course, also affected by historical change. How can we be sure what was considered to be fictive in the Middle Ages, for example? The answer to this question is often not as straightforward as it might appear at first sight. Sonja Glauch emphasizes that in medieval times »the spectrum of things that were actually considered to be impossible and thus could be a signpost of fictionality with regard to the aspect of phantasticism is much narrower than in the modern concept of reality«15 (Glauch 2014, 401). Right in line with this, in her paper on dragons as a challenge for theories of fictionality, Henrike Manuwald is very cautious with her answer to the question of whether or not people believed in dragons in the Middle Ages. She concludes that the only thing one can say with reasonable certainty is that »one cannot exclude that belief from the outset«16 (Manuwald 2018, 78). Therefore, even if we could be sure that fictive elements were a signpost of fictionality in medieval times, not even the appearance of a dragon in a text would allow us to draw the conclusion that the text was (considered to be) fictional.17 The second and the third concerns clearly become more pressing as more time passes between the present and the date of origin of the texts investigated. It is important to note that at least two of the three main historical eras in the discussion about the history of fictionality -classical antiquity, the Middle Ages and the 18th century -are affected by these problems to a large extent.
However, textinternal signals may be the wrong place to look for reliable signposts of fictionality; maybe paratextual signals are more promising. As Nicolas Paige has shown in several publications, a quantitative analysis of paratexts (among other things) can effectively reveal findings concerning the changing practices of authors (cf., e. g., Paige 2017). Nevertheless, it is important to note that Paige's studies focus mainly on the 18th century -a period of time that is quite close to our own (at least in comparison to classical antiquity and the Middle Ages) and thus relatively easily examinable. In reference to further back in history, 15 »Gegenüber der Wirklichkeitsauffassung der Neuzeit ist der Kreis dessen, was tatsächlich unmög lich scheint und über den Phantastikaspekt Fiktionalität anzeigen könnte, viel enger gezogen.« 16 »Die Frage, ob man ›im Mittelalter‹ an Drachen ›geglaubt‹ habe, lässt sich zusammenfassend nur so beantworten, dass man einen entsprechenden Glauben nicht von vornherein aus schließen kann.« 17 It is important to note that this historical change of (the range of) concepts applies not only to the notions of fictivity and phantasticism, but also to the concepts of reality and truth, for example -concepts that have played a vital role in the differentiation of text sorts throughout the centuries. Anne Sophie Meincke, for example, stresses that »the notion of ›historical‹, ›fac-tual‹ truth [may] not be expected to be readily available in the Middle Ages« (»einem für das Mittelalter nicht ohne weiteres als verfügbar zu erwartenden Begriff ›historischer‹, ›faktischer‹ Wahrheit«) (Meincke 2007, 337). Cf. also ibid., 351, and Braun 2015, 98 sq. in contrast, greater problems confront an analysis of paratexts, as explained in the following.18 First of all, one main difficulty that is spelled out by Sonja Glauch, for example, is that »the vernacular writtenness knows […], contrary to the Latin, of no steady paratexts in the form of a titulus until the time of Humanism; therefore, there was no practice (and probably no necessity) to pass on the name of an author outside of the narration reliably«19 (Glauch 2014, 402). Accordingly, at least up till the 15th century, a steady convention of paratexts had not yet been established, leading to the lack of a reliable basis for an analysis of signposts of fictionality in paratexts for this long period of time.
Secondly, the first concern regarding the search for textinternal signposts of fictionality mentioned above also applies to the search for paratextual signposts: the danger of presupposing the existence of an institution of fictionality, whereas the investigation of the signposts is supposed to prove the existence of the according institution in the first place.
Nevertheless, at least one paratextual element seems to be quite informative and revealing in regard to the question of the history of the institution of fictionality: the indication of the literary genre. Today, this paratextual element is usually a very reliable signpost for the sort of text we hold in our hands. When we buy a book with the label »science fiction«, we already know that this is a fictional text without having read a single word. In fact, it seems legitimate to assume that the indication of the literary genre has always served as a very effective means of differentiating various sorts of texts. At the same time, the usage of a generic indication in the paratext also seems to be a reliable signpost of the existence (or at least the development) of particular literary practices associated with certain literary genres. A paratextual signal would simply make no sense if the author could not 18 In this chapter, I concentrate solely on the problems arising from the search for signposts of fictionality in paratexts (or, to be more specific, in peritexts). The following chapter addresses »quasi-poetological« paratexts such as prefaces.

»[D]ie volkssprachige Schriftlichkeit kennt […]
, anders als die lateinische, bis zum Humanismus keine stabilen Paratexte in der Art eines titulus und damit gar kein Verfahren (und wohl auch kein Bedürfnis), einen Verfassernamen außerhalb der Erzählung sicher zu tradieren.« As Paige observes, »many novels [of the eighteenth century still] contain no paratextual indications as to their fictional status« (Paige 2017, 507). Glauch also emphasizes the negative consequences of the lack of this practice for a possible distinction between the author and the narrator of a text -a distinction that is very often mentioned as a signpost of fictionality in modern times as well (cf. Glauch 2014, 402). In addition, »[s]ince the narrators transmitted the plots and were hardly ever their creators […], they were generally just as disoriented with regard of their status of reality as their audience« (Haferland 2014, 62; »Da die Erzähler Tradenten und kaum je einmal auch Schöpfer der Plots sind […], stehen sie dem Wirklichkeitsstatus grundsätzlich ebenso desorientiert gegenüber wie die Hörer«). count on the fact that at least some of his readers actually understood the difference in the practice of writing and reading the text indicated by the generic indication. Even in times without an existing paratextual convention, the generic identity of a text could, in all likelihood, be inferred from textinternal signals such as the theme or the structure of the plot. Accordingly, Sonja Glauch comes to the following conclusion: »In analogy to the function of the paratexts and to the conventions of the literary market today, affiliation to a certain genre could have been a very effective signpost for fictionality or factuality in the Middle Ages«20 (Glauch 2014, 406). Yet, Glauch has to restrict her assumption immediately: »(O) f course [this is only the case] if this genre was defined on the basis of fictionality or factuality«21 (ibid.). At this point, the matter becomes more complex because we are, once again, in danger of arguing in a question-begging way: we cannot use a certain generic indication (may it be explicit or implicit) as a signpost of fictionality unless we have already proven that this specific genre was written and meant to be received in terms of fictionality. How could this be done? Either by an investigation of textinternal signposts of fictionality, such as the fictiveness of the objects or the truth content of the story -in which case we would gain nothing from the generic indication because it would basically throw us back to the problems of textinternal signposts discussed previously -or by an analysis of poetological comments, which I will discuss in the following chapter.
Before that, however, I have to address one final problem concerning the affiliation of a text to a certain genre as well as textual phenomena like signposts. As is evident from the above-mentioned difficulties, there is no simple correspondence between historical genres and their textual presentation, on the one hand, and our current genres and their textual presentation, on the other. Nevertheless, research has repeatedly ascribed fictionality to textual and genre-specific phenomena that, upon closer look, sometimes differ in crucial aspects from what we consider to be a fictional text and the corresponding signposts of fictionality today. I will pick out a prominent example of this course of action and concentrate on the concept of »functional fictionality« introduced in research on medieval texts. However, my observation also applies to similar approaches (in studies on the Middle Ages as well as in studies on other periods of time).22

»Analog zur heutigen Funktion von Paratexten und von Konventionen des literarischen
Markts dürfte im Mittelalter die Zugehörigkeit zu einer Gattung ein sehr wirksames Fik tio na litäts-bzw. Faktualitätssignal abgegeben haben«. 21 »[F]reilich nur, sofern diese Gattung durch Fiktionalität oder Faktualität bestimmt war.« 22 Consider, for example, the attempts to locate fictionality somewhere in the trias of fabula, argumentum and historia, or to associate it with the concept of integumentum. For an illuminating discussion cf. Müller 2004, 288sqq., andGlauch 2014, 397 sq, 410 sq. As Sonja Glauch explains, the main difference between »autonomous« and »functional« fictionality is the following: »Autonomous fictionality corresponds to the modern concept of fictionality. Functional fictio denotes the practice of a poetic-rhetoric arrangement of narrative material that was historically believed in«23 (Glauch 2014, 393 sq.). Jan-Dirk Müller emphasizes in a similar way that medieval research »has identified the asymmetric relation between factivity and fictionality and contrasted different types of feigning, such as the rhetorical orchestration of given (or supposedly given) material (›functional‹ fictionality) and its free invention (›autonomous‹ fictionality)«24 (Müller 2004, 282 sq.). The extent to which the concept of a »functional fictionality« actually differs from our current understanding of »autonomous fictionality« is best seen in Glauch's summary that »functional fictio is neutral in regard to fictionality«25 (Glauch 2014, 396), and that »functional fictio is obviously not fictionality in the narrower sense«26 (ibid.).27 The difficulty arising from the ascription of fictionality to phenomena like these is then evident: it is at least unclear (to state it carefully) whether we are really talking about two different occurrences of the same phenomenon -namely fictionality and the according institution. What is very obvious, in contrast, is that »functional fictionality« differs significantly from the modern concept as well as the modern practice of fictionality.28 So the crucial question is: Is »functional fictionality« a medieval form of the phenomenon we call »fictionality« nowadays -or are the differences so decisive that we should refrain from calling the medieval phenomenon in question »functional fictional-23 »Autonome Fiktionalität entspricht […] dem Fiktionalitätskonzept der Moderne. Als funktionale fictio wäre die Praxis zu bezeichnen, einen vorgegebenen, als historisch geglaubten Stoff dichterisch-rhetorisch zu gestalten.« For a fundamental criticism with regard to this differentiation, cf. Reuvekamp-Felber 2013, 435, fn. 56. 24 Man »hat das asymmetrische Verhältnis von Faktizität und Fiktionalität herausgearbeitet und unterschiedliche Typen des Fingierens zwischen rhetorischer Inszenierung von vorgegebenem (oder als vorgegeben angesehenem) Material (›funktionaler‹ Fiktionalität) und dessen freier Erfindung (›autonomer‹ Fiktionalität) voneinander abgehoben.« 25 »Funktionale fictio ist fiktionalitätsneutral.« 26 Während »funktionale fictio offensichtlich keine Fiktionalität im engeren Sinne darstellt«. 27 Glauch as well as Müller argue that, accordingly, this »medieval sense of fictionality« permits degrees of fictionality resulting in the existence of hybrids of fictionality and factuality (cf. Glauch 2014, 400 sq.;Müller 2004, 286, 295). In modern research on fictionality, in contrast, the idea of a gradual concept of fictionality is usually rejected since fictionality is generally considered to be bound to categorical [!] intentions (cf. Kablitz 2008, 17). 28 Cf. Haferland, who emphasizes that medieval literature »achieves a remarkable autonomy of narrating without immediately including a status of fiction« (Haferland 2014, 76 sq.; »dass sie eine bemerkenswerte Autonomie des Erzählens erreicht, ohne dass diese gleich einen Fiktionsstatus einschlösse«).
ity«?29 I personally tend to be rather restrictive in this regard. By using the same concept of fictionality for two phenomena, we obviously want to stress the similarity of the events in question. Therefore, when they differ in too many aspects, I consider adherence to the concept of fictionality regardless of fundamental differences to be more confusing than helpful.30 To sum it up, the above-mentioned problems raise doubts whether former practices of fictionality can be reconstructed in a thorough analysis of literary texts alone. The method of inferring premodern practices of fictionality from literary texts confronts us with serious, intractable problems concerning the investigation of textinternal signposts as well as of paratextual signposts. As already indicated, by consulting poetological texts as well, we may stand a better chance of coming to a conclusive result, so I will discuss the method of inferring former practices of fictionality from poetological texts in the following chapter.

Problems Concerning Inferences from Poetological Texts
Poetological texts seem to be a more appropriate object of investigation for an understanding of former practices of fictionality since their analysis certainly does not pose the problems related to signposts found in literary texts. Poeto-29 This is, of course, related to the more general question of how different two practices may be until we should stop calling both of them »practices of fictionality« -a question this paper can not solve. 30 Even in regard to Jan-Dirk Müller, who uses the term »fictionality« with all due caution, one could question whether it makes sense to stick to this concept under the given circumstances. Müller argues that a contract of fictionality slowly became established in the 12th century. Nevertheless, he does not fail to highlight that »compared to the leeway of a modern author, what fictionality allows for in the case of Chrétien [de Troyes] is obviously different and more restricted« (Müller 2004, 292; »Dabei ist der Spielraum dessen, was Fiktionalität bei Chrétien zuläßt, doch offenbar ein anderer und enger begrenzt als der eines modernen Autors«). »Chrétien remains bound to a certain ›historical‹ basis, but, in comparison to the parameters for historians, those for a narrator's fantasy are greatly extended« (ibid., 293; »Chrétien bleibt an bestimmte ›histor-ische‹ Vorgaben gebunden, doch ist der Spielraum des Fingierens gegenüber dem Geschichtsschrei ber um ein Vielfaches ausgeweitet«). Müller therefore considers Chretien's work only to be »fictional in a historically specific, namely medieval, sense of the term« (ibid., 294; »schafft er einen in einem historisch spezifischen, nämlich mittelalterlichen Sinne fiktionalen Text«). Once again, it is clear that what Müller describes is rather different from the modern concept as well as from the modern practice of fictionality. Therefore, one could ask if the phenomenon in question is in fact a form or practice of »fictionality« at all. logical texts are already meta-texts about literature, so we do not have to rely on inferences from time-variant signposts to learn about a former practice of fictionality. Therefore, in this regard, poetological texts are not only more informative, but also more dependable. Nevertheless, there are also a number of problems related to the method of drawing conclusions from poetological texts regarding the history of the practice of fictionality.31 One major difficulty with inferences from poetological texts is the potential (or even probable) gap between the poetological discourse, on the one hand, and the actual practice with literary texts, on the other. As Tilmann Köppe rightfully observed, what is explicitly spelled out is usually only what is not taken for granted, whereas institutionalized practices are informed by a »matter of course«, that is, by an often unconscious (»intuitive«) knowledge of rules. One has to expect, therefore, that what is explicitly said does not correspond precisely to the conventional rule or does not represent it precisely.32 (Köppe 2014, 421) But if poetological texts were more often than not programmatic and normative instead of descriptive, that is, if they stipulated a desired practice rather than described an actual practice, it is questionable how reliable our inferences about former practices of fictionality from analysis of poetological texts can be as well. At least, we cannot simply presuppose that the practices of fictionality discussed in poetological texts automatically represent former, stable institutions of fictionality after all.
Another problematic aspect to be taken into consideration is the fact that premodern poetological texts are written texts that dealt with written texts in an environment of orality. At the beginning of the 19th century, about 75 % of all adults were still illiterates. Nevertheless, myths, fairy tales, etc. existed as oral heritage and were certainly passed on in according practices of story-telling and reception. The question is, therefore, if and to what extent poetological texts reflected this overwhelmingly oral literary practice. In general, it does not seem implausible to assume -in addition to the supposable difference between theory and practice 31 What I will say about (the analysis of) poetological texts also applies to (the analysis of) »quasi-poetological« paratexts such as prefaces. Cf. e. g. Kuhn 2018 for a detailed investigation of the three prefaces of the Robinson Crusoe series. 32 »Explizit gesagt wird i. d. R. nur das nicht Selbstverständliche, institutionalisierte Praxen sind aber durch ›Selbstverständlichkeit‹, d. h. ein oft unbewusstes (›intuitives‹) Regelwissen, ge prägt. Man muss daher damit rechnen, dass das explizit Gesagte gerade nicht dem kon ven tionel len Regelfall entspricht bzw. diesen gerade nicht repräsentiert.« Cf. ibid., 421 sq., for further problems correlated with the interpretation of historical sources. mentioned above -a probable difference between orality and literacy as well. In any case, one should keep in mind that poetological texts -even when they were meant to be descriptive -probably depicted only a small segment of the former practices of fictionality. This skepticism regarding the range of poetological discussions on former practices of fictionality increases even more if one argues -as has been done repeatedly -»that writtenness is […] a necessary […] condition for conscious fictional narration«33 (Ridder 2002, 30). Under this premise, the practices of fictionality discussed in poetological texts would have been the exception to the rule of a predominantly oral practice of narration in which the question of fictionality probably did not even arise. As Harald Haferland puts it: In times of solely oral narration, fictionality is not ascribed to narrative text since it is not, and could not be, consciously known as text. […] Recognizing or identifying a text as a text -and even more as a fictional text -is part of the abstraction of distinctive literality. Therefore, in the sense of terminology, fictionality is excluded from narrative folklore.34 (Haferland 2014, 57 sq.) In any case, even if one does not follow Ridder's or Haferland's considerations on the immediate connection of fictionality and writtenness, one has to keep in mind that the method of investigating poetological texts probably permits only rather limited inferences about former practices of fictionality -if they existed at all in an overwhelmingly oral context.
Another difficulty that comes along with drawing conclusions about the history of the practice of fictionality from an analysis of premodern poetological texts arises from the fact that these texts -of course -do not refer to the term »fictionality«. They typically do not center on the concept of »literature« either, but rather on »poetry«, which causes the following two problems. Firstly, up to the 18th century, poetry was generally written in verse (and meant to be recited orally) due to reasons of aesthetic appreciation (cf. Althaus 2011, 275 sq.). From a modern perspective which typically ascribes fictionality to prose literature, it is questionable whether poetry in this sense can qualify as fictional -and, accordingly, whether the practices associated with poetry can qualify as related to the modern practice of fictionality. Secondly, many poetological texts place a discussion of the value of poetry at their center. As Wolfgang Rösler has shown, Gorgias and Aristotle aimed to »free poetic ›untruthfulness‹ or even ›lie‹ resolutely from its negative assessment and to turn it into a positive one«35 (Rösler 2014, 375). With reference to fictionality in Early Modern times, Tilmann Köppe assesses in a similar way: »What is characteristic of the Early Modern discourse on fictionality is the domination of statements on the value or purpose of poetry«36 (Köppe 2014, 437). Many other examples could be cited to prove a strong evaluative component in poetological texts, and this indicates that, were we to replace the historical concept of »poetry« with a modern concept, »literature« would probably be more appropriate than »fictionality« (or »fictional texts«). What poetological texts discuss, therefore, seems mostly to be practices of literature, but not practices of fictionality.37 It is revealing in this regard that Wolfgang Rösler, who has emphasized the discovery of fictionality in classical antiquity in many papers since 1980, expresses at one point that he considers classical poetological texts as well as texts of contemporary literary theory to be »reflection[s] on fundamental questions regarding the functioning of literary [!] communication«38 (Rösler 2014, 363).
One last problem involving inferences from premodern poetological texts concerns the question of which parts or subjects of these texts are perceived by their modern interpreters as discourse on fictionality. A closer look at the papers of Wolfgang Rösler reveals that, according to him, the »discovery of fictionality in ancient Greece« came about »in the context of the development of the singer as the medium of divine inspiration in originally oral poetry to the autonomously creative author of literature«39 (Rösler 2018, 54). Whereas »-according to these programmatic testimonials [i. e. the poetological sources] -it was its divinely conveyed and guaranteed element of truth«40 (ibid., 52) that was constitutive of poetry before Aristotle (as, for example, in Homer's epics), Aristotle then devel-35 »[D]ie dichterische ›Unwahrheit‹ oder gar ›Lüge‹ entschieden von negativer Bewertung zu befreien und ins Positive zu wenden.« 36 »Kennzeichnend für den neuzeitlichen Fiktionalitäts-Diskurs ist das Dominieren von Aussagen über den Wert oder Zweck der Dichtung« [italics on »Diskurs«/»discourse« deleted]. 37 Cf. also Glauch, who ascertains in regard to medieval poetological texts that »there are obviously no contemporary self-descriptions and no contemporary concepts of a literary theory that clearly aimed at fictionality« (Glauch 2014, 387; »dass es offensichtlich keine zeitgenössischen Selbstbeschreibungen und keine zeitgenössischen Konzepte einer Literaturtheorie gibt, die eindeutig auf Fiktionalität abzielen«). 38 »Reflexion über fundamentale Fragen des Funktionierens literarischer Kommunikation«. 39 »Entdeckung der Fiktionalität im antiken Griechenland, insofern sie sich im Zuge der Entwick lung vom Sänger als Medium göttlicher Eingebung in ursprünglich mündlicher Dichtung hin zum autonom schaffenden Autor von Literatur vollzog.« 40 »Konstitutiv für epische Dichtung ist also -nach diesen programmatischen Zeugnissen -ihr göttlich vermittelter und garantierter Wahrheitscharakter.« oped -with reference to Gorgias and his thoughts on the genre of tragedy -a new definition of poetry. He relieved poets from telling (with divine inspiration) what really happened, seeing their duty instead in telling (by ways of their own human creative powers and use of the principle of mimesis) what could have happened. Therefore, the poets' area of responsibility was no longer truth, but verisimilitude. This is what Rösler calls the »criterion for fictionality established by Aristotle«41 (ibid., 54), explaining that »Aristotle does not stop at treating fictionality as just a property of texts«42 (ibid., 57), but that he -like Gorgias before him -takes »the process of poetic communication«43 (ibid.) into consideration. Since poetry, now regarded as a product of human origin, had lost its supernatural truth content, »it was possible to think of it as a (legitimate) act of deception […] on the part of the author, a deception a smart recipient was willing to expose himself to. Thus, the ›institution‹ of fictionality was grasped theoretically for the first time«44 (ibid., 61).
In my opinion, there are two ways to interpret Rösler's thoughts. One very restrictive way with regard to the application of the concept of fictionality would say that fictionality is not the object of Rösler's considerations at all and that Aristotle is rather talking about poetic creativity, permissible objects in poetry (probable courses of history instead of true historical events), the truth value of poetry and maybe even its truth claim. But none of this automatically implies a »discovery of fictionality« -neither as a concept nor as a practice. Of course, Aristotle (among others) clearly differentiates between sorts of texts -those, for example, that talk about true historical events and others that do not. This does 41 »[Das] von Aristoteles aufgestellte Kriterium der Fiktionalität«. It is remarkable that Catherine Gallagher argues in a very similar way for »the simultaneous appearance of fictionality and the novel« (Gallagher 2006, 340) in the eighteenth century: »The widespread acceptance of verisimilitude as a form of truth, rather than a form of lying, founded the novel as a genre […]. It also created the category of fiction.« (ibid., 341) Yet, Gallagher explicitly distances herself from arguments like Rösler's: »As the example of Don Quixote demonstrates, there were novels before the eighteenth century, and as the citations of Aristotle and Sir Philip Sidney indicate, the components for an understanding of fictionality were also available. And yet, these did not gel [sic] into either a common knowledge of the concept or a sustained and durable novelistic practice until they coincided in the eighteenth-century English novel« (ibid., 345). not mean, however, that one must jump to the conclusion that Aristotle differentiates between fictional and non-fictional texts.45 Accordingly, one could say that neither Aristotle nor Gorgias nor Rösler himself is, in fact, talking about an »institution of fictionality«. What they are discussing is an institutionalized understanding and the respective practices associated with certain, mostly poetic, texts, but not an institution of fictionality (and, of course, least of all the modern institution of fictionality).

»Aristoteles
The other, less strict, way to interpret Aristotle's and Rösler's thoughts assesses that they do talk about fictionality and an institution of fictionality, but that this concept and the institutionalized practice of fictionality are relevantly different from our complex modern concept and practice. In this case, we arrive at a similar general problem already faced with regard to the concept of »functional fictionality«: Are the two phenomena in question -the modern concept and institution of fictionality, on the one hand, and the historical concept and institution, on the other -similar enough to evaluate them as two different occurrences of one and the same entity? Or is the historical phenomenon actually different in kind? Once again, I would personally prefer to refrain from conceptual unification and, instead, to stick to what we can say without reserve: 45 This kind of conclusion, though, is not only problematic in Rösler's argumentation, but seems to be a quite wide-spread phenomenon. To give another example: in Tilmann Köppe's investigation of fictionality in Early Modern times, the section on text production starts as follows: »In the source material, numerous signs of a differentiation between text types are found, a differentiation that draws on text production and attributes the characteristic of being invented to fictional [!] texts« (Köppe 2014, 424; »Im Quellenmaterial finden sich zahlreiche Hinweise auf eine Textsortenunterscheidung, die an der Textproduktion ansetzt und fiktionalen Texten das Merkmal zuspricht, erfunden zu sein«). Köppe goes on to present these production-related signs of a differentiation of text types and comes to the convincing conclusion that »[t]he early modern poetological discourse knew about the differentiation between ›invented‹ and ›not invented‹ stories, and [that] there is at least some evidence that texts were counted as poetry because or insofar as the depicted events were invented« (ibid., 426, emphasis original; »Der neuzeitliche poetologische Diskurs kennt die Unterscheidung zwischen ›erfundenen‹ und ›nicht erfundenen‹ Geschichten, und es gibt zumindest Hinweise darauf, dass Texte zur Dichtung hinzugezählt werden, weil oder insofern das Dargestellte erfunden ist«). What Köppe shows, therefore, is that, in Early Modern times, there was a differentiation of text types based on the criterion of inventedness. What he does not show, however, is that this differentiation equals the differentiation between fictional and non-fictional texts (as already presupposed in the first quote). The problems in this equivocation become apparent when Köppe mentions »the complicated case of a ›partially‹ invented story« (ibid.; »den komplizierten Fall der ›teilweise‹ erfundenen Geschichte«) -a hybrid form that seems to undermine the alleged differentiation between fictional and non-fictional texts based on the question of inventedness.
We know (thanks to Rösler and others) that a conventionalized practice of text production and reception (i. e. institutionalized rules and conventions about how to write and read texts) goes way back to classical antiquity, maybe even farther. We also know that awareness of the existence of different kinds of textswhich implies that they had to be produced and received differently -dates back to ancient times as well. Moreover, we have also learned that discussions on the features distinguishing these different kinds of texts (truth vs. falsehood or lie, fictive vs. real objects, truth vs. verisimilitude, etc.) started a very long time ago. As to the question of whether or not we want to call these former practices »practices of fictionality«, I consider one thing to be significant: in the face of all the recurring problems mentioned above that are related to an investigation of a history of the practice of fictionality, describing historical concepts and practices as accurately and in as much detail as possible and highlighting and carving out their similarities and differences with our modern phenomena seem to be more important than giving in to the urge of harmonizing rather different concepts and practices through time.