Topos research stands between the disciplines; it is not standardized and transverse to conventional paradigms (cf. [Knoblauch 2000] regarding sociology). In my understanding, this is primarily because its roots lie in antiquity and are embedded in a way of thinking that does not have disciplinary divisions and corresponding knowledge production practices, as we know them today. Within the “philological tradition,” topology has split into different areas, each of which emphasizes different qualities of topoi dependent on the specific (sub)discipline and related research interests. The versatility of the category has led to its fragmentation. In his discussion of ancient rhetoric [Barthes 1985, 138–141] already emphasized the complexity and threefold role of topoi / topology – as method, as storeroom (of possible argumentative elements), and as grid (for finding arguments).[7] However, he did not translate this into a concrete methodology. The most elaborated and important theoretical reference for my approach is Bornscheuer’s dynamic concept of topos. [Bornscheuer 1976, passim] emphasizes that topoi are both “sediment” and “instrument” (i.e. “storeroom” and “grid”). Although Bornscheuer derived and defined the polyvalence of the category of topos very precisely from a theoretical point of view, particularly with reference to the ancient rhetorical tradition, he did not realize nor conceive a transfer into analytical practice. [Knape 2000, 758] notes this lack of practice in the field of topos research as a desideratum and emphasizes the challenges of isolating and interpreting topical structures in a text. However, his approach of a two-axis model influenced by structuralism and rhetoric is linked to a problematic conception of conscious text production, in which he assumes that topoi always exist in a paradigm and are “implemented” during text production [Knape 2000, 756].[8] This entails, among other problems, a tendency to overemphasize the stability of topoi and to overlook their potential to dynamize discourses and systems of knowledge.[9] Following [Bornscheuer 1976, 177], however, topoi can be seen as the source and result of various historical-social forces; they are both influential and changeable, stabilizing and dynamizing. The stable tradition or conventionality of “topoi” is discussed within literary studies, particularly in the reception of Curtius’ influential and polarizing work European Literature and the Latin Middle Ages.[10]

A splitting within topos research was subsequently reinforced in the contrasting of two different qualities of topoi already outlined above – namely that they are a (“material”) “storeroom” on the one hand and a (“formal”) “grid,” a kind of “template” for argumentation, on the other. Studying the travelogues, however, clearly demonstrated how closely these two qualities are intertwined, so that I pursued an integrative approach that would overcome (sub)disciplinary boundaries. The knowledge of experience, orientation and opinion that crystallizes in topoi regulates the “asystematic coherence”  [Bornscheuer 1976, 43] of communication (in everyday language, but also more “aesthetic forms”). This encompasses a broad range of knowledge-related and argumentation practices. It spans from political contexts to constructions of imaginary spaces (such as “India”) to knowledge about genres (like “travelogues”). It is important to recognize that rationality in this context tends to be overemphasized: Argumentative coherence usually is note created purely in a rational or logical way, but based on the (historically and culturally situated) archive of topoi.[11] Thus, topoi generate both argumentative coherence as well as symbolically-imaginarily grounded semantic coherence.

While topos research in its various lines of tradition represents a pool of ideas around topoi, a convincing set of tools to systematize the argumentative-semantic patterns in my study corpus was lacking. In view of this methodological vacuum in the “traditional humanities” and the abundance of material to be handled (40 travelogues, approx. 11,000 printed pages), I relied on ideas from the digital humanities (and corpus linguistics) for conceptual modeling and combined this line of thought with a discussion of topos research. I was concerned with integrating various relevant aspects of the category into a multi-layered conceptualization of “topos” that is connected to, but also distinct from, the rhetorical tradition, and could be translated into a concrete research design. The images contained in the corpus (e.g. photographs) were excluded for pragmatic reasons; all considerations relate to the annotation and analysis of texts.[12]

The results lie on two closely interlinked levels: the (re)construction of topoi and the (re)modeling of topoi as a methodology. An important aspect is the assumption that topoi are defined in relation to other topoi, namely to other topos concepts that have been modeled as instantiations of the category topos. The topos concepts are modeled as elements of a topical constellation and thus lying on the same level. I refer to these constellations using the term topology (and distinguish it from topology as methodology by italicizing it).[13] Two topologies – the “India topology” and the “travelogue topology” – were considered relevant for the corpus of German-language India travelogues around 1900. The India topology is composed of 86 topoi that contribute to the topical stabilization and dynamization of an imaginary “India” around 1900 (see Figure 6). The travelogue topology, on the other hand, forms a constellation of 67 topoi that are considered characteristic of the textual genre “travelogue” around 1900 (see Figure 2).[14] This heuristic differentiation of two topologies allows for the specific characterization of each individual constellation as well as the precise description of the interdependencies between them. Their visualization in clusters opens new perspectives on the underlying principles of their interconnectedness.

Figure 2 shows the twelve cluster of the travelogue topology. These clusters are the result of conceptual modeling. Two topoi associated with the same cluster are distinct but functionally similar. Subsequent to the (re)construction of the topoi, the (conceptual) cluster analysis allows for their systematization in order to gain a better overview of the connections and characteristics of the specific topology. The clustering of the two different topologies revealed clear differences in the ways the respective topoi act.[15] In the clustering of the travelogue topology, the phenomenon emerged that some topoi belong to more than one cluster. This coincides with the increasingly clarified view of the polyfunctionality of topoi, i.e. the fact that topoi are argumentatively open. That means that their argumentative character can manifest in “different directions” (described in rhetoric as the “in utramque partem principle”  [Bornscheuer 1976, 42]). However, these argumentative “directions” are not only “pro” or “contra,” but – in my combinatorial approach – depend on the other topoi with which they occur. This approach echoes earlier findings of topos research, but significantly expands and nuances them.

These forms of combination have been central to the investigation of the corpus and I have modeled them as “topical collocations.” The relevance of this category was clear relatively early on, as my aim was to operationalize the observed combinatorial character of topoi. The modeling of bridge topoi, in contrast, is a particularly striking example of how a new phenomenon and the associated category can become visible in the research process. Based on the finding that some of the topoi belong to several clusters, I defined the concept of bridge topoi and applied it as a category in the following. This is an important step in the process: An emergent phenomenon is conceptualized and the resulting concept is transformed into a category in the moment in which it is systematically applied. This application of the category consisted of systematically checking for each topos whether a multiple assignment to the clusters is plausible or not on the basis of the annotated examples (and their frequency). This resulted in a deeper knowledge of the travelogue topology and thus also a new understanding of the genre. The approach of (re)defining travelogues in its basic characteristic and genre principles via clusters and associated topoi could probably be transferred to other genres. The assumption is that a text-type-specific investigation of topoi and their clustering would also be informative for other corpora.

While the (re)construction of the topoi, the clustering and the collocation analysis concern the level of the realm of phenomena, I have (re)modeled topology as methodology step by step in parallel. The starting point was the topos research mentioned above and the aim of developing a systematic methodology. In addition to the more precise modeling of some general functional principles of topoi (reduction and amplification, distribution, combination), the study resulted in ten operationalization steps. This result is both a model of topology as a practiced methodology (“model of”) and hopefully a “model for” future research designs in (digital) topos research. The operationalization steps condense the methodological insights gained in the course of the process – in other words, they are an idealized and abstract representation of the research process. The modeling of these steps and their practical implementation simultaneously led to the application of the methodology that I developed and used in parallel. Their order does not necessarily correspond to this chronology but can also deviate depending e.g. on the hypotheses or the research question. The first five steps are particularly important with regard to category development and application. (1.) definition of the corpus, (2.) definition of the relevant topologies, (3.) (re)construction of topoi, (4.) cluster analysis, (5). collocation analysis. The steps 6 to 9 are less relevant for understanding the main analysis results but concern their further interpretation.[16] The last operationalization step (10.) reflects the dynamics of remodeling topology as a methodology while applying it.

In the following, I discuss the four relations already presented using concrete examples from the travelogue corpus of the underlying study. The principle behind their selection is that they are as representative as possible of the relation type to be shown and relatively easy to understand, also without deeper knowledge of travelogues or India.

The category topos is closely related to the concept topos. Both lie on a mediating level, oriented upwards towards the textual surface and downwards towards the endoxa, an assumed discursive deep structure (see Figure 3). As already mentioned, topoi can be (re)constructed from their manifestations on the surface of the text. In this process, certain recurring patterns, which I have called “manifestation patterns,” can often be recognized among larger numbers of manifestations. Several of these manifestation patterns can be subsumed under one topos or, conversely, the manifestation patterns can be understood as forms of expression of the topos. The level of manifestation patterns is an intermediate level between the category topos and the surface of the text. As outlined above, the relevance of topoi is based on both recurrence and coherence – both together determine their relevance in a concrete discursive situation. In Figure 3, the category topos shows up on the axis linked to the form of a concrete concept anchored in an underlying structure.

The too much Europe topos appears in different manifestation patterns such as “too much civilization,” “too much industry,” “bad European influence,” etc. In the naming of the topos, the too much was included because the topos is particularly frequently appearing in this line of argument (not just as “European influence”). In some cases, it is linked to the influence of British colonial domination.[17] In other cases, it is not explicitly related to Europe, but to “civilization” in general. Criticism of industrialization and modernity are closely linked in this manifestation pattern, which is, e.g., related to living and working conditions.[18] Certain major cities are perceived as particularly Europeanized, while others, such as Jeypur, do not yet seem to be affected by the process.[19] The topos thus appears on the text level in a wide range of forms in which it crystallizes and is at the same time anchored in various discursive contexts that inform the corpus (colonialism, tourism, etc.).

The axis of the horizontal relations can operate in different ways and thus be subdivided into two parts (see Figure 4). In the case of the (re)construction of topoi, the relation betweens the individual topos concept (as instantiation of the category) and the other elements in the respective concept system (the “India topology” and the “travelogue topology”) are important. The basic logic of horizontal relations is that the combination of concepts – modeled as topical collocations – can be described on the basis of a clear differentiation of the concepts. The combinations become visible by distinguishing the individual concepts. Furthermore, on the level of the text surface, the relations between manifestations of topoi can be described more precisely. In this context, I have proposed the concept of the “topos profile” of texts.[20] Many passages had to be analyzed to obtain an overview about the occurrence of topoi in order to systematize the collocations. During the (re)construction process, it is not immediately obvious whether a particular recurrence pattern is to be classified as a collocation (i.e. two topoi) or as one topos. The demarcation between two topoi or the subsuming of different manifestation patterns under one topos is a question of the granularity of the category and the understanding of the concept in relation to other concepts within the system.

An example illustrating the challenges in differentiating recurrence patterns and defining them as one distinctive topos concept is the topos of inexpressibility that is present in various genres and transhistorically widely spread. However, according to [Curtius 2013, 159] it is debatable whether we are dealing with one topos or whether several topoi should be differentiated. He states that the “root of the topoi [...] is ‘emphasis upon inability to cope with the subject.’ From the time of Homer onwards, there are examples in all ages. In panegyric, the orator ‘finds no words’ which can fitly praise the person celebrated.” In contrast, other studies, especially those dealing with travelogue corpora, have reconstructed only one topos, with slight nuances in the names (e.g. “indescribability topos” in [Fischer 2004, 288]; “topos de l’indicible” in [Choné 2015, 251–257]). The topos or the topoi (depending on the perspective) revolve around the paradox that they express that something is inexpressible. In the travelogue corpus, different manifestation patterns (something is “indescribable,” “untellable,” etc.) could be identified leading to the differentiation of two topoi: inexpressibility by words (“Unsagbarkeit”) and unrepresentability (“Undarstellbarkeit” = not representable in any medium). This differentiation is justified by their strong recurrence and argumentative differentiation in the corpus. In some cases, inexpressibility increases intermedially, in the sense that something cannot be expressed in any medium – a frequent manifestation pattern of this collocation is “neither brush nor pen.”[21] In other cases, inexpressibility by words can be overcome because visual media offer a way out of the dilemma of unrepresentability.[22] Regarding the horizontal layer of the text surface and the analysis of topos profiles the texts by [Halla 1914] and [Litzmann 1914] are for example very contrasting: While in Halla’s text manifestations of the color topos have a significant frequency and dominate the text, in the case of Litzmann it can be observed that almost the entire repertoire of topoi of the two topologies is distributed over the text.

The relation types of the vertical and horizontal axis are focused on one central category and concepts all belonging to this same type of category – namely “topos.” In the heuristic space, these two axes are closely interlinked. The heuristic space involves furthermore the diagonal axis, thus those relations between categories belonging to different classes. As described above, the conceptual clustering revealed various phenomena, which in turn led to the formation of new categories and the corresponding expansion of the toolkit: (a) Bridge topoi emerged, which belong to two or more clusters. By modeling this category, the argumentative polyfunctionality of topoi becomes visible. (b) Different cluster types appeared, which only become evident in the comparison of the two topologies and through which, in turn, the specifics of these can be captured more precisely.[23] Of course, the dynamics in category formation do not need to be so open. For example, one could think of a follow-up study that continues with the 150 topos concepts (or a part of them, for example individual clusters, etc.) and the methodology outlined in a selection of the 10 operationalization steps.[24]

Looking at the following concrete textual example, I want to illustrate the added value of this approach with respect to the categories of clusters, bridge topoi, and collocations.

“In the realm of the Maharajas and Rajahs, you get to know the truly Indian India, quite untouched by English civilization, the dreamland where you experience fairy tales at every turn, you only need to wander through the cities.”  [Litzmann 1914, 61]

In this highly condensed passage, a relatively large number of topoi appear in reduced form. According to the examples above, the collocation of the three topoi too much Europe, untouchedness, and authenticity can be considered as already familiar. According to Figure 2 showing the clusters of the travelogue topology, these topoi all belong to the tourism cluster. Untouchedness furthermore bridges to the singularity cluster. The authenticity topos (in the original Echtheit) is particularly polyfunctional – it belongs to five clusters (besides tourism and singularity also language reflection, recognition, writing programmatic / self-presentation). These topoi are all elements of the travelogue topology. At the same time, we can also observe manifestations of topoi of India topology in the passage cited above: The dream topos and the fairy tale topos, both belonging to the irreality cluster which also includes 1001 nights, magic, fantasy and miracles (see Figure 6). Topoi of this cluster are often combined with topoi of the princes / wealth cluster. To illustrate the quantitative dimension of topoi, one case can be mentioned in which the recurrence of a pattern was not frequent enough to be considered as a topos: The pattern that the (Maha) Rajas were only “pseudo kings” or “puppets” of the British recurs in the corpus.[25] Thus, a potential collocation pattern between the princes / wealth cluster and the “puppet pattern” of the colonialism cluster seemed to emerge, but this could not be confirmed (due to relatively low frequency).

While the space of heuristic relations refers to topoi, topologies and associated categories within the topology as methodology, the space of epistemological relations is concerned with the relations between different category systems. Figure 7 depicts various categories of other frameworks to which the category “topos” is related. This kind of epistemological demarcation of a category can be thought of both synchronically and diachronically. In addition, the epistemological type of relations can be split up further: (a) There can be relations between two different conceptualizations of the category “topos.” Here the naming of the category coincides, but the conceptualizations are different. An example of such a different modeling approach working with the supposedly “same” category (identical label, i.e., “topos”), is the SAToRBase project. As part of the project and network, a database for topoi in French literature was set up, but with a significantly different understanding of topos than that of the underlying study. Anchored in “structural narratology,” the SAToRBase project aims to create a thesaurus of narrative recurrence patterns conceived as topoi [Sinclair 2003] [Dubost 2021]. In contrast, I argue that topoi are argumentative patterns rooted in a historically and culturally situated doxa and are not just “plot elements.”[26] (b) In the second case, the names of two supposedly different categories can differ, but the conceptualization of the content of the category can be (almost) congruent. This can be observed, for example, for the category of the “collective symbol,” which is strikingly similar to the understanding of “topos” represented here.[27] (c) In the third case, there are categories with distinct labels and different conceptual content. This is the most common case and the one outlined in Figure 7. Relating the conceptualizations in the sense of a demarcation of a category vis-à-vis the “categorial neighbors” has two advantages: on the one hand, it sharpens the categories of the different frameworks; on the other hand, it enables well-founded combinations of methodologies.

In studying the history of the concept of topos, it became clear that some basic assumptions of ancient rhetoric are not compatible with the attempt of operationalization. This, above all, concerns the fact that topoi are situated at the level of inventio (the first production stage of a speech before dispositio, elocutio, memoria, and actio) and are conceptualized as an element of a repertoire or paradigm [Knape 2000]. Within the (re)modeling process, topoi were taken out of this narrow assignment. Instead, it was emphasized that topoi are not exclusively situated in a paradigm, but (in the logic of rhetoric / reformulated in structuralist terms) lie at an intermediate level between a deep structure (inventio / paradigm) and a surface structure (elocutio / syntagma).[28] As already indicated, the conceptualization of its categorial neighbors and counterparts has an effect on the notion of “topos” itself. This concerns among others the intersections to “narrative” and “trope”  [Nate 2013], to “conceptual metaphor”  [Lakoff and Johnson 1980], or to “frame” as category within frame semantics [Ziem 2014]. The closer a related category is conceptually, the more precision potential an epistemological relationalization has. This concerns, in particular, the tradition of discourse-linguistic topos research [Wengeler 2015], which has close connections and similarities (analysis of argumentative recurrence patterns as topoi) but presupposes a clearly different concept of argument and chooses a different solution in dealing with the opposition of “formal” and “material” topoi.[29]

The added value of the approach of topology as a methodology and of considering its epistemological relations is that a differentiated description of methodological contexts is possible. For example, [Buzard 2001, 4] defines the touristic pattern that travelers want to move “just off the beaten track” as a “master trope.”[30] In the corpus investigated, four different topoi of the “travelogue topology” are closely associated with this figure of argumentation (too much Europe, labyrinthine alleys, authenticity, untouchedness).[31] This (re)construction makes it possible to describe the recurrence pattern of the touristic discourse “just off the beaten track” much more precisely as a specific form of collocation that can manifest in different ways. Particularly frequent or popular topoi have been mentioned and discussed in studies about travelogues (idyll, paradise, inexpressibility, etc.), but so far, no attempt has been made to consistently apply topology as a methodology for investigation and to systematically (re)construct a conceptual system of topoi. In the perspective proposed here, a non-systematic (re)construction risks being subject to a “confirmation bias” and mainly recognizing those patterns of recurrence that have already been discussed in previous research.