Volume 13, Issue 2

It is sometimes suggested that deductive reasoning has no place in literary studies, particularly when it comes to the justification of literary interpretations. The thesis is that deductive arguments (almost) never occur and deductive reasoning plays no or at most a marginal role in the actual practice of interpretation. In this essay I will argue that this thesis is false. Using counterexamples, it can be shown that deductive arguments are de facto used in the practice of literary interpretation. Although only an exemplary and not comprehensive investigation of argumentative practices in literary studies can be made here, it can also be made plausible that deductive arguments are not exceptions, but a normal, regularly encountered and legitimate phenomenon of this practice. Furthermore, it can be shown that the thesis criticized here is based on a too narrow understanding of deductive reasoning. Finally, I will argue that (from a methodological point of view) it would be an unnecessary restriction to exclude deductions from our methodological toolbox: to abandon deductive reasoning from the practice of interpretation would be tantamount to renouncing a useful method of justifying (and refuting) interpretive hypotheses. Therefore, a fundamental methodological scepticism about deductive reasoning is unfounded. The essay is structured as follows: First, I will present a current example of how the sceptical attitude towards deductions in literary studies is motivated (Section 1). I will show that the concept of deduction underlying this scepticism is usually based on the assumption that deductive arguments contain ›general rules‹ or ›law-like assertions‹ in their premises. Second, I will confront the concept of deduction assumed there with the understanding of deductive arguments as it is usually assumed in logic (Section 2). It will become apparent that a conception of deductive arguments based on general rules or law-like assertions is (though not false) too narrow. After a short caveat concerning the reconstruction of ›real‹ arguments in general (Section 3), I will present some examples of deductive arguments in interpretations of literary texts (Section 4). These arguments have the logical form of arguments from modus ponens , modus tollens and disjunctive syllogism. The examples not only show that deductive reasoning actually occurs in the practice of literary interpretation, but also make it plausible that deductive arguments are an unproblematic and legitimate method of justifying and/or refuting interpretations. After considering what causes may have led to the dissemination of the thesis that deductions do not occur in practice (Section 5), I conclude with a plea to accept deductive arguments as a legitimate tool of justification among many others (Section 6).

This essay examines the question of whether and under what conditions a fictitious character can be an epistemic authority for (real) readers; more precisely: it asks whether and under what conditions readers can acquire (propositional) knowledge from the character, thus learning something from it. In answering this question, the essay brings together two debates that have so far hardly been related to each other: an epistemological debate on the concept of epistemic authority and a literary-theoretical debate on aesthetic cognitivism, i. e., the discourse about what can be learned from the reception of fictional texts. In order for a person to be an epistemic authority for another person, two conditions must be met: 1) the first person must have an advantage in knowledge over the second person that the second person recognizes and acknowledges as such; and 2) the second person must have appropriate access to this knowledge. In order to clarify to what extent a fictitious character and a real reader can be related in this way, I first examine what it means to attribute knowledge to a fictitious character. To do so, I suggest the following analysis: In story S, character C knows that p if and only if C believes in S that p; p is true in S; and C is justified in S to believe that p (this suggestion, based on the classical definition of knowledge, can easily be adapted for other suggested analyses: all that is required is that all conditions in the analysis – whatever they might be – lie inside the scope of the fiction operator). Furthermore, a knowledge attribution of the form »In S, C knows whether p« is true if and only if in S, C knows that p or knows that not-p. On the question of the correctness-conditions for knowledge attributions of the form »In S, C knows that p« and »In S, C knows whether p«, I will then enter the debate about fictional truths. This is necessary for two reasons. On the one hand, the attribution of knowledge is nothing but the assertion of a particular fictional truth. And on the other hand, an attribution of knowledge involves another fictional fact, namely the fact p (which I call the »underlying fact«). The view that is largely held in the discussion about fictional truth following Lewis is that what is true in a story does not result solely from the explicit assertions in the text, but also from plausible consequences [ Plausibilitätsschlüssen ] that we can be further justified in drawing. More precisely, the following possibilities arise for both facts – the underlying fact as well as the attribution of knowledge: Either the text explicitly contains a reference to the fact. Or it does not contain such an explicit reference, but the question of whether the fact obtains can still be answered on the basis of plausibility conclusions . Or there are no explicit references and plausibility conclusions cannot be drawn. In this case, there is a point of indeterminacy. These distinctions result in a number of possible combinations corresponding to different types of situations, some interesting instance of which I examine in more detail. One case that is especially remarkable is when there is a point of indeterminacy in the text with regard to the underlying fact, which – as I illustrate with an example – does not exclude the possibility that knowledge can be attributed to a character with regard to the proposition in question. The claim is often made about indeterminate passages that not even God can know whether the facts in question obtain – and this is correct. Hence if we are entitled to attribute the knowledge in question to a character, this shows that fictitious characters can not only know more than the reader or the author, but even more than God. Such situations also illustrate that more knowledge does not have to go hand in hand with more epistemic authority. For readers, the indeterminate passage remains unresolvable, and readers cannot learn anything from the character in this regard. This leads me to the question of under which conditions the reader can learn something from a character to whom knowledge is attributed that the reader does not possess. A fundamental problem for the idea that there could be something like a transfer of knowledge between a fictitious character and a real reader is that both belong to different ontological spheres, so to speak: the reader is real, the character merely fictitious. If a character were to be an epistemic authority for a reader, this would be a case of a transfictional epistemic authority , which must be distinguished from »ordinary« epistemic authorities as well as from fictitious epistemic authorities and from epistemic authorities for fictitious truths . I propose to analyze transfictional epistemic authorities using the make-believe theory and the extended-pretense operator: When readers find themselves in extended pretense and pretend to be part of the fictitious world, they become at least imaginatively capable of interacting with the characters, so that the characters can become imaginary epistemic authorities for the readers. I also discuss the cognitivism debate and argue that the (fictitious) knowledge of a character can affect not only intra-fictional but also extra-fictional objects and truths. A main objection to the cognitivist view that readers can acquire propositional knowledge of reality from reading fictional texts is that fictional texts are not reliable sources and that the beliefs the reader may form through reading cannot be justified. I reject this objection and argue that readers can also acquire knowledge about reality through the attribution of knowledge to fictitious characters or even from speech acts that the characters make in the story. Finally, I will deal with a possible objection that the epistemic authority that a character can have is completely parasitic on that of the author: the objection here is that if readers learn something, it’s actually from the author. In contrast, I argue that fictitious characters can acquire an independent epistemic authority that cannot be reduced to that of the author.

The positioning in space and time of performed narration in theater poses a specific challenge to classical narratological categories of structuralist descent (developed, for example, by Gérard Genette or Wolf Schmid, for the analysis of narrative fiction). Time is the phenomenon which connects narratology and theater studies: on the one hand, it provides the basis for nearly every definition of narrativity; on the other, it grounds a number of different methodologies for the analysis of theater stagings, as well as theories of performance – with their emphasis on transience, the ephemeral, and the unrepeatable, singular or transitory nature of the technically unreproducible art of theater (e. g. by Erika Fischer-Lichte). This turn towards temporality is also present in theories of postdramatic theater (by Hans-Thies Lehman) and performance art. Narrating always takes place in time; likewise, every performance is a handling of and an encounter with time. Furthermore, performed narration gains a concrete spatial setting by virtue of its location on a stage or comparable performance area, so that the spatial structures contained in this setting exist in relation to the temporal structures of the act of theatrical telling, as well as the content of what is told. Both temporal and spatial structures of theater stagings can be systematically described and analyzed with a narratological vocabulary. With references to Seymour Chatman, Käte Hamburger and Markus Kuhn among others, the contribution discusses how narratological parameters for the analysis of temporal and spatial relations can be productively expanded in relation to theater and performance analysis. For exemplary purposes, it refers to Dimiter Gotscheff’s staging of Peter Handke’s Immer noch Sturm (which premiered in 2011 at the Thalia Theater Hamburg in cooperation with the Salzburger Festspiele), focusing on its transmedial broadening of temporal categories like order, duration, and frequency, and subsequent, prior, or simultaneous narration. The broadening itself proves feasible since all categories of temporal narration can be applied to performative narration in the theater – at times even more fruitfully than in written language, as is the case, for example, with the concept of ›duration‹. The concept of ›time of narration‹ too can be productively applied to theater. Whilst a subsequent narration is frequently considered the standard case in written-language narratives on the one hand – a conclusion that is, however, only correct if the narrator figure and narrative stand in spatiotemporal relation to one another, i. e. if a homodiegetic narrator figure is present – it is commonly held that in scenic-performed narration, on the other hand, the telling and the told take place simultaneously. The present contribution argues against this interpretation, as it stems from a misguided understanding of the ›liveness‹ of performance. ›Liveness‹ refers only to the relationship between viewers and performers and their respective presence, but not to their temporal and spatial relationship to the told. Rather, the following will argue that the time of narration in theater (as well as in film) stays unmarked in most cases. It is possible, however, to stage subsequent, prior, or simultaneous narration, too. Immer noch Sturm is one example for a performed subsequent narration. For audiovisual narration, then, a special case of iterative narration (telling once what happened n times) can be identified, which is to tell a few times (n minus x) what happened n times. As an additional category for the analysis of narrative temporality in audiovisual narrative media, I propose what I venture to call ›synchronized narration‹, in order to describe the specificity of spatiotemporal relations in performance. In synchronized narration, two or more events (that happen at different places or times in the narrative world) are shown at the same time on stage. This synchronized performance of several events is only realizable within the audiovisual dimension of spatial narration and not in written-language based narration. Furthermore, for narrative space relations the categories ›space covering‹, ›space extending‹, and ›space reducing narration‹ are suggested in order to analyze the relationships between discourse space and story space(s). Discourse space emerges in the concrete physical space of the performance when narrativity is present. Within this discourse space any amount of story spaces (with any expansion) can emerge. However, whilst in time-extending narration the time of the telling is longer than the time of the told, in space-extending narration the told space is bigger than the space of the telling. This principle is analogously valid for time-reducing or space-reducing narration. The transmission and media-specific broadening of temporal and spatial narratological parameters reveals how time and space form a continuum and should thus be linked and discussed alongside one another in analytical approaches to narrative artifacts. The staging of Immer noch Sturm actualizes a metaleptic structure, in which temporal borders are systematically dissolved and the overstepping of spatial borders becomes an indicator for the merging of different temporal levels. Referring back to established narratological parameters and developing analogous conceptual tools for narrative space facilitates a comparative analysis both of specific narratives and of narrative media and thus not only offers a productive challenge of classical narratological parameters, but allows to investigate and construct a holistic – if culture-specific – overall view of narration.

The use of past tense in narrative discourse at first glance seems to imply that the narrated events are lying in the past as compared to the act of narration. However, this intuitive notion was doubted by several scholars in the mid 20th century, among them Käte Hamburger, Harald Weinrich, Émile Benveniste, and Ann Banfield. This article investigates the temporal constitution of literary narratives from a perspective of the biological evolution of Human cognition. My analysis begins with Hamburger’s most disputed claim that in epic fiction the past tense »loses its grammatical function of designating what is past« and proceeds by testing a derived hypothesis against a cross-cultural sample of (mostly oral) folklore. Hamburger denied any temporal relation between the speaker and that which he speaks of, assuming instead a fictitious neverland in which the narrated events are situated and to which the preterit refers. If Hamburger’s model is correct, so the derived hypothesis goes, then the use of past tense in fictional narratives is merely a cultural convention, and different narrative traditions should expose different conventions. Indeed it can be shown that in other cultures stories are told in the present tense, in infinitive verb forms, or in forms indicating abstractness or remoteness. It can be followed that Hamburger was right at least in presuming that reference to the past is not a necessary constituent of verbal storytelling. Actually, instead of referring to the past, the epic preterit rather seems to indicate a change in the modality of speaking, thus adhering to the category of grammatical mood rather than tense. In some languages of oral cultures, however, this presumed mood shows up instead as a way of indicating the source of information. This kind of source information – fully grammaticalized in a quarter of the world’s languages – is called ›evidentiality‹ by linguists. In a phylogenetic perspective on the evolution of cognition, source information only becomes necessary with extended inferential and communicative capabilities and may thus have emerged as a cognitive tool in early humans when entering the ›cognitive niche‹. Evidentiality markers in language may thus be the linguistic reflex of a very ancient cognitive scope category in the innate architecture of the human mind, one which served to separate first-hand experience from reported knowledge. In oral storytelling, evidentiality is marked not only by specific verb forms but also by specific formulas (›they say‹/›it is said‹), intonations, or rhetorical devices. From this perspective, the phenomenon observed by Hamburger and others can be said to originate in the beginning of Tradition – that is, of verbal transmission of cultural knowledge. My hypothesis is that literary narratives in literate cultures still use this ancient cognitive scope operator of ›tradition‹ when employing the epic preterit. Admittedly, in literate cultures it often suffices to put »A novel by« on the title page in order to signal the categorical otherness of narrative fiction. Yet still, authors employ additional means to evoke the atmosphere of a ›murmuring conjuring‹ – as Thomas Mann once called it – that creates the impression of an objective world of tradition behind the individual story told. I point toward examples in literary first-person narratives, because homodiegetic narration – in contrast to Hamburger’s classical case of heterodiegetic narration – shows a continuous spatio-temporal relation between speaker and that which he speaks of and thus requires additional means or efforts to signal a break between the ordinary world of first-hand experience and the world of the literary. Since Hamburger once treated the epic preterit as a signal of fictionality, I briefly discuss the notion of fiction in the last paragraphs of my paper. I consider ›fictionality‹ to be a late cultural concept in literate societies that is not identical with the cognitive category of ›tradition‹ but is ultimately made possible by the existence of the latter.

The discussion of genre theory since the middle of the twentieth century has been marked, among other things, by two conflicting endeavors: on the one hand, there are attempts to semantically regulate generic terms for the purpose of a clear scientific language by defining them as types of texts (cf. Fricke 2010, Fishelov 1991, Zymner 2003); and on the other hand, there are attempts to accustom literary studies to the factual indeterminacy of generic terms and to comprehend this indeterminacy by adapting Wittgenstein’s concept of family resemblance or the semantics of prototypes (cf. Strube 1986, Hempfer 2010a). The essay is intended to show that both positions – that of strictly regulating and of merely descriptively comprehending practice – as they have been advocated so far hinder the foundation of genre-historical research. This essay aims to more precisely formulate the productive ideas from both positions and thus to provide a robust foundation for the historical study of genres. This foundation is to be structured by two core questions that until now have not been clearly separated: first, how one ought to define, before and during historiographic and empirical work, the concept of the genre that is to be examined; and second, how one can synthesize the historically changing semantics of a genre at the end of this empirical work. Answering these two questions becomes a problem when one is dealing with historically discontinuous and heterogeneous genres. This essay pursues this problem using the example of novella history, because research here is particularly divided as to how the concept of the novella can best be understood. The guiding idea of the essay is to reverse the usual sequence of steps: in genre theory, the dogma is often advocated that the concept of genre must be defined a priori , i. e., before the investigation begins and independently of empirical work. The first section of this essay rejects this dogma and recommends that we make the use of the concept of a specific genre dependent on the concrete epistemological interest of the investigation. With the help of an argument by Kendall L. Walton, the essay shows that aesthetic innovations and the historical development of literary procedures can only be understood if a hermeneutic interest in the references of individual texts to historically established generic expectations is at the forefront of the investigation (cf. Walton 1970). The essay recommends that generic historiography serve this hermeneutic interest. It aims to answer the first core question of how the concept of genre ought be used before and during historiographic work for two central procedures: for reconstruction from, on the one hand, historical poetics (second section); and, on the other hand, from text groups (third section). In both procedures, the thesis is that the concept of the genre to be examined should not be regulated before or independently of empirical work by defining a concept of a text type. In the case of the reconstruction of generic expectations from text groups, one is therefore dependent on a variant of the so-called inductive procedure. In order to rid the inductive method of its methodological deficiencies, the article reformulates this method in terms of quantity theory in such a way that the procedure fulfills the essential requirements. The problem with the inductive procedure is that, in the case of inconsistent historical use of names for genres, it is not suitable for inferring generic characteristics from text groups. This problem is particularly pronounced with novellas, because the texts historically referred to as novellas seem to have only one thing in common, namely, that they have been referred to as novellas. The set-theoretical reformulation of the inductive procedure should allow the reconstruction of the generic expectations historically associated with the name of the genre by making use of classificatory predicates. However, it is not the concept of genre that this procedure aims to determine prior to historiographic work. Rather, the aim is to relate the set of texts that were communicated as novellas in the relevant historical situation to classificatory sets by means of intersections. Each of these intersections – for example, the intersection of the texts referred to as novellas and of fictional journal prose – can be described with regard to the relevant text characteristics of the defined set (for example, fictional journal prose). It is important not to summarily define such intersections as novellas, as is often done. Such a definition would result in a classificatory concept of the respective genre, which in turn would no longer be suitable for comprehending the historical use of the generic label and hence the semantics of a genre in historical literary communication. This essay thus provides an operationalizable reconstruction of the intuition, only vaguely formulated in genre theory so far, that in conducting research one must mediate between inductive and deductive procedures, between being oriented toward the material and reflecting on systematic concepts (cf. Hempfer 1973, 128 sqq., Mueller 2010). The fourth and final section of the essay seeks to answer the second key question of how to move from heterogeneous empirical findings to a concept of genre that brings these findings together. This question is divided into two subquestions. On the one hand, the question arises as to how the connection between factual and conceptual history can be conceived: whether such a connection exists cannot, this essay argues, be decided on the basis of genre theory. Rather, this is a problem of conceptual and factual history that can only be solved hermeneutically. Two different genre understandings are connected in terms of conceptual history only if the later conception is meaningful for the earlier conception. They are not connected by a concept of text type that has been previously defined quasi as a substance. On the other hand, the question arises as to the concrete cases of literary application for which the results of a history of genre should be integrated into a single concept of genre. These are primarily acts of interpretation and codifying mediations of knowledge, in which generic knowledge becomes manageable in being comprehended by generic concepts that are as simple as possible. However, formulating generic knowledge in terms of concepts is principally associated with a loss of historical accuracy. In order to avoid returning to historically inadequate concepts of genres, the essay develops a model in which a concept of family resemblance oriented toward Wittgenstein and already roughly transferred, by Strube, to genre theory is specified and differentiated with historical indexes (cf. Strube 1986). The essay argues that such a specific concept of genre is most likely to serve the initially developed hermeneutic interest and, at the present stage of theory elaboration, represents the best compromise between historical accuracy and practical applicability.