On the history of comparison in folklore studies
It is quite obvious that comparative research has been important in folklore studies in general. Even defining the group that forms the "folk" is the result of a comparative process. The very concept of variation is the product of a comparison.
The historic-geographical method was, for a long time, the predominant pattern for folklore research, and it is still valid as a means of approaching and displaying folklore material. In this paper, I want to discover the dominating background against which this method must be regarded today, what was compared to what and how was the comparison conducted, what were the purposes of comparing, how the method has changed during the last 150 years and what are its future chances.
The study of folklore has, to a large extent, consisted of comparison between one text and another and between one custom and another. The texts and the customs were sought and compared for different purposes. In one case that was influenced by evolution and positivism, the aim was to search for a primordial Indo-European culture from which Western languages and ways of thought were assumed to have developed. In this instance, according to Max Müller (1823-1900), folklore was regarded as a rest of primitive mythology.
Another purpose, evidenced by researchers such as Edward B. Tylor (1832-1917), Andrew Lang (1844-1912) and James Frazer (1854-1941), was to reach a deeper understanding of human culture all over the world. In his essay, "The Method of Folklore", Lang states that comparison is the way of doing folklore research.
Survivals, National Politics and Primoriality
The first recommendation that the folklore of the Swedish minority in Finland be collected was made in 1848 by a teacher at the Vasa Lyceum, Johan Oskar Immanuel Rancken (1824-95). During his time, interest in the folklore of Finnish-speakers was immense among the students at the University of Helsinki. The reason for this interest being the growing nationalistic movement that excited Finnish identity, which was later to create the mental foundation for Finnish independence from Russia. In addition to the Finnish-speaking population, there were also Swedish-speaking people in Finland. During the first decades of the nineteenth century, this was regarded as problematic
J.O.I. Rancken became acquainted with these ideas when he was a student in Helsinki. He became a school teacher of history. Therefore, the historical way of thinking and analysing was well-known to him.
In general, humanistic studies at that time were studies of history. The brothers Jacob and Wilhelm (1785-1863, 1786-1859) and their colleagues traced the etymological roots of the German language, archaeology was a historical discipline rather than a scientific one, there were disciplines such as the history of literature, of arts, of music and of religions. "History" was the overall concept, and so it was to be for more than about a century to come. Implicitly, the historical approach sought to find the origins, the very beginnings. By making comparisons, it was considered possible to reconstruct the evolution of a phenomenon and to determine its form and its time and place of origin. Distribution was eagerly debated. Rancken was trained in this spirit.
By collecting folklore, Rancken considered it possible not only to identify a Finland-Swedish "national" culture, but at the same time to compare cultural traits from a wider geographical region. Thus, Rancken cooperated with Wilhelm Mannhardt (1831-80), when he carried out his work on how the cult of the forest and the field spirits was constructed genetically in the 1870s. It is also important to state that the idea of the Swedish-speakers in Finland as a link between Finnish and Swedish culture became a political factor later on. For political reasons, Finland-Swedish folklorists at the end of the nineteenth century and at the beginning of the twentieth, needed comparison to show that the Swedish-speakers in Finland were implicitly related to the Swedes in Sweden; that they had a right to maintain their own Swedish culture without adopting the Finnish one.
The idea of comparing folklore texts was elaborated by Julius Krohn (1835-88) and later refined by his son, Kaarle Krohn (1963-1933), in the so-called ‘historic-geographical’ method. In the 1880s and 1890s, Julius Krohn studied the Kalevala from a genetic point of view. He compiled songs in Finnish and Estonian and he also included themes and motifs from neighbouring languages. For Julius Krohn the texts published in the Kalevala books were not enough for a starting point for research. Manuscripts on both published and unpublished materials should not be ignored. It was important that only the oldest Finnish variants could be compared with the oldest foreign texts. For him, it was quite clear that the variants were geographically dependent. Geography and topography were important factors when tracing the oldest variants. Krohn organised the texts in geographical and chronological systems in order to discern which was the original and what was a later addition. The results of his studies employing this method were partly politically inconvenient because he could show that the Kalevala was neither purely Finnish, nor very old.
Comparison and ways of doing comparative analysis were the foundation of Krohn's work. He was a descendant of the scholarly work done at that time. For him, comparison was a means of finding the origin of something, in his case, the first Kalevala-metre poetry. But whilst Andrew Lang strived to understand a custom, poetry or another folklore phenomenon by comparing issues without taking geographical circumstances into consideration, Krohn meticulously paid attention not only to nationally varying texts, but also to variants that differed within one and the same geographical area. Variation in space and time, as it was documented in the archived written recordings, was crucial to Krohn because he could, by taking geography into consideration, trace the historically oldest text.
Thus, comparison is central in the methods employed by both Andrew Lang and Julius Krohn at the end of the 19th century, but what they compared to what, and why they did it, differs. Lang wanted to compare "the ideas which are in our time but not of it". Krohn wanted to compare texts along geographical and historically fixed axes in order to determine which parts of the Kalevala were really Finnish and how old they were. His scholarly work had a worldwide impact upon following generations of folklorists.
The Historic-Geographical Method
Julius Krohn’s colleagues adapted his ideas to their own material. The Estonian, Walter Anderson (1855-1962), for instance, chose the prose narrative classified as "The Shepherd Substituting for the Priest Answers the King's Questions" (AT 922). He published his very thorough study, Kaiser und Abt, in 1923.
Anderson does not mention Julius Krohn, but he thanks Kaarle Krohn and so it is reasonable to think that he was acquainted with Julius Krohn's work, as well. The lack of a list of references makes it impossible to tell what the relationship between Julius Krohn and Walter Anderson really was in connection with this monograph.
Walter Anderson’s study is very detailed. Occasionally, he refers to cultural history, contexts, intertextuality and source critics, although he does not use precisely these terms. Evolutionism and geographical concern have their places in his mind. The result of his study is clearly and explicitly presented. By way of contrast, Anderson’s aim and methodology are implicit as they assume familiarity with a particular form of folkloristics. Given the ideas already outlined on different reasons for comparison within folklore studies, it would not have been out of place for Anderson strictly to explain where he positioned himself, but he obviously did not find this necessary. Just a brief mention of the Finnish school, in the preface, was enough to imply the method and the purpose.
Comparison is central to Anderson’s research. On the one hand, he has to compare several folklore texts with each other in order to define the tale-type on which he wants to concentrate. Thus he finds several hundreds of variants which have some kind of a common core, but, he does not overtly define that core. In a way, he has already made that decision in his choice of the tale-type. For him there seems to be two groups of texts worthy of his interest: the prints and the oral ones. This task of comparison seems to be rather easy to define. On the other hand, he compares the motifs and sub-motifs within the texts, such as the persons mentioned and the riddles posed. However, he does not problematise the nature of these smaller units, nor how one should discern them from non-interesting details. Again it seems that the author anticipated a common knowledge of the method and a generally accepted consensus on terminology.
Walter Anderson's study is frequently mentioned by Kaarle Krohn in his main methodological work Die folkloristische Arbeitsmethode published 1926. He even refers to it as a model. Krohn develops a method for handling large numbers of folklore texts, arranging them geographically and chronologically. His aim is to produce a genetic analysis of a special folklore topic in order to find its origin, and trace its distribution.
Krohn purposefully tries to define which components of folklore he wants to compare and on what level this comparison should be conducted. He also takes into consideration the different conditions connected to and called for by different genres. He develops a general folklore method which is valid for international studies, although he also limits his view of folklore to being a cultural product belonging to his own continent. Krohn is a practical person who provides practical advice, and therefore, he is also a realistic scholar. His aim is clearly formulated: "The purpose of this analytic procedure [...] is the determination of the ur-form of any characteristic of the theme under study as well as the original combination and arrangement of these traits in order to determine the basic form as well as the pattern form in various areas; from the relationship between the basic form and the pattern form, the route of diffusion and development of the tradition in question can be discovered" (Krohn 1971: 63)] and his dream is "the publication of all variants of each tradition in a general language in the most concise possible form, preserving all significant traits" (Krohn 1971: 39)]. This would certainly create a kind of thick corpus!
At about the same time as Krohn was formulating his folklore methodology book, Walter Anderson was ready to extract and summarise his method into a dozen pages which he published in Handwörterbuch des deutschen Märchens in 1934. The fact that he published this concentrated set of directions for use, reflects a change of mind. He no longer expected everybody to be acquainted with his method. He divides it into twenty-one strictly limited phases, according to which folk-tales can be studied. There are three unconditional principles: every single existing recording of a tale must be considered no matter whether the text is printed or oral and irrespective of language. Comparison should be made trait by trait in as unpreconceived a manner as possible. And, the time and place of the recording must be taken into consideration. The aim of a study conducted by this method is a monograph, in which one tries to determine the original form, i.e., a common archetype from which all variants derive, the place and time of origin, local redactions and ways of distribution.
This article tempts one to see this method as a finished entity. However, even in this set of directions Anderson does not problematise comparison. The main lines along which comparison is done are geographical borders, but comparative views are also needed when the student tries to analyse the content of the texts. Here, however, an inexperienced folklorist receives no help from Anderson’s article, because the basic elements, like an episode, a motif or a trait, are never described although these concepts are crucial to him. Anderson clearly sees the limits of his method. He constantly shows that the study of folklore is empirical work. It is a practical task with many concrete problems to be solved.
Anna Birgitta Rooth
The historic-geographical method had a great impact on Nordic folkloristics. However, it also met with critics. One of the scholars who questioned it was Carl Wilhelm von Sydow (1878-1952) in Lund, who wrote an article titled "Finsk metod och modern sagoforskning" as if the Finnish method and modern research were opposites. He opposed the idea of using the historic-geographical method for producing monographs on tale-types, but preferred the study of the life of the tales. One of his successors, Anna Birgitta Rooth, carried on a controversy around his thoughts when, in 1951, she published her dissertation The Cinderella Cycle. She emphasises that the method chosen must be appropriate for the questions the student wants to answer. Thus, monographic studies employing the historic-geographical method serve one purpose, studies of life of the tales another. Rooth studied a group of related types of a tale instead of only one type, but still she sticks to the classical purpose that "of elucidating the mutual relationships and development". She stresses studies of geographical distribution, the principal motifs, variants and detail motifs. For Rooth, the historic-geographical method seems to be too schematic, various views have to be considered in a more comprehensive manner. For instance, she points out the importance of regarding the tale as a logical composition of interior organic elements when trying to determine the original form and place. She also disagrees with existing ideas that all chimaeras originate from within the Indo-European sphere of culture. By comparison, Rooth ranks geographical distribution high, but supplements it with aspects of general cultural history and ethnological circumstances in a more elaborate way than either Anderson or Krohn. In this way, she sheds light on the value of folklore studies for analysing the relationship between different cultures, which neither Anderson nor Krohn mention so clearly. In her bibliography, there is no mention of either Krohn or Anderson. It seems that she is acquainted with their investigations through other scholars like von Sydow, Thompson and Kiefer.
Matti Kuusi's Redaction Analysis Applied by Inger Lövkrona
Some thirty years later, Inger Lövkrona published her doctoral dissertation Det bortrövade dryckeskärlet (ML 6045). She studies the legend of how a human being steals a cup from the other world. First Lövkrona tries to place the legend in a research tradition by considering different names for it. This procedure aims, by comparison, at defining the legend from similar, but unrelated legends. At the same time, Lövkrona also considers genre terminology, function and previous research. Her book contains a detailed discussion of the aim and method of the study. The method of analysis, Lövkrona states, is dependent on the principles according to which the material has been collected, i.e., almost no choice is left but to analyse the legend according to "the model of the Finnish school, the so-called geographic-historical method" in a monograph. Her aim is to shed light on the distribution and origin of the legend. Lövkrona regrets the limitation of the material, because consideration of function, meaning, storytellers, situations of actualisation, narration and communication is nearly impossible. She regards structuralism as an aid to verifying the results reached by the historic-geographical method and its product, the monograph, as a means of understanding the historical and cultural contacts and complexes better. Furthermore, she wants to combine the writing of a legend monograph with thoughts on cultural areas, boundaries and distribution. Finally, she wants to test the historic-geographical method on material which was collected according to its own particular rules to see where it needs to be modified.
Mostly Lövkrona works according to the Anderson-Krohn recommendations, but she also deviates from them. Her most important deviation from the classical historic-geographical school is inspired by Matti Kuusi (1914-98). She bases her investigation of the origin of the tale on an analysis of redactions. In his article, "The Bridge and the Church" published 1974 in English and in 1980 in an enlarged Finnish version titled "Suomalainen tutkimusmenetelmä" ["The Finnish Research Method"], Kuusi seems to have given up with the historic-geographical method. Unlike his forerunners, he uses very limited material consisting of twenty-eight variants from the coast of Ingria. The article presents different viewpoints from which to examine the legend in order to answer the question as to why Jesus prefers a bridge before a church and even maintains to have been sheltered by the bridge, but prosecuted in the church. Kuusi’s article is a pedagogical example of how to apply a method. He states, therefore, that he has maintained several of the phases of analysis that scholars do usually not retain in a publication, just because he wants to show how his method works. Kuusi does not seek to discern the original variant as we have come to expect from the historic-geographical method. His main concern is on the content or perhaps even on the meaning of the text. He calls his method the ‘text critical method’.
For Kuusi, the comparative moment is crucial; just as it was for the Krohns, for Anderson and Rooth. However, for Krohn, comparison is important not mainly for the sake of geographical distribution, but in order to point out textual change without regarding place at first hand. In this article, we might see signs of the development of recent thoughts distinguishing between variants and variation.
Kuusi’s redaction analysis has helped Lövkrona to find the centre of the distribution, i.e., where the motif is logical, epically correct, and best designed. Centre–periphery thoughts are dominant in this way of adapting the historic-geographical method to redaction analysis. The next step is to date the redactions and to decide on their mutual age. Finally, the analysis of redactions leads to a hypothesis of origin and distribution.
To sum up, at the beginning of the 1980s, it was still possible to think that the historic-geographical method was worth using in folklore studies. Like Rooth Lövkrona is aware that this method has been criticised, but she defends it on the grounds that the investigated material was collected using this method. She also takes advantage of new ways of studying folklore. Whereas earlier efforts concentrated on text-centred methods, she discusses a performer-centred type of folklore research. Whereas Rooth hinted at the connection between anthropology and folklore, Lövkrona does not exceed the limits of contemporary folkloristics, but rather utilises other kinds of folklore research, such as views of structure, function, message and meaning. Comparison is central for her analysis when she selects the chosen legends from other texts, and when she tries to determine which motifs constitute the legend. Comparison is also important in the redaction analysis and thus in the construction of the original redaction and when she tries to distinguish types and motifs.. The world wide comparison found in for instance Anderson does not occur in this study.
Three Variants of the Historic-Geographical Method
Anderson, Rooth and Lövkrona claim to use the Finnish or the historic-geographical method. Anderson and his followers build up their studies in a similar manner. They take their readers more or less directly in medias res. Rooth arranges her dissertation differently discussing her ends and means in more detail. Anderson does not find it necessary to describe his method, whereas such a presupposition is missing in Lövkrona's book. She provides a thorough account of her version of the method, which is a combination of Anderson-Krohn’s and Kuusi’s models. Rooth and Lövkrona defend their methods as if they were not widely accepted. Rooth and Lövkrona also clearly state the aim of their studies; such statements are omitted by Anderson.
The material in all three studies consists of prose narratives. Anderson and Lövkrona have studied legends, Rooth a Märchen. For Anderson and Rooth the need for a world-wide gathering of material is obvious, for Lövkrona, the Nordic recordings are central.
Terminological problems are debated to some extent. However, a strict definition of a motif is missing, even though this concept lies at the core of all comparisons. Presumably, there was a commonly accepted definition of the issue. Rooth, on the other hand, pays the concept some attention. Previous research is handled in all three studies: Anderson provides 23 lines, Rooth a little more, Lövkrona some four pages. Anderson does not argue with any other student of folklore, whereas Rooth and Lövkrona take part in an ongoing debate fairly actively.
Anderson presents extremely detailed lists of motifs in long catalogues. Rooth, on the other hand, presents her tale types in a more reasoning manner. Lövkrona does not appreciate the catalogue as a means of displaying her findings. She concentrates on the redaction analysis. Lövkrona regards comparison as a means of shedding light upon structure, function and the meaning of the tale and not as a result in itself. Thus the comparative method has changed during these years from being a way of describing a tradition to being a vehicle for asking subsequent questions about the life of the tradition.
Anderson and Lövkrona concentrate a huge amount of information into tables and lists and partly into maps as well. Rooth describes her findings in a fluent text. But whereas Anderson and Rooth describe how the tale and legend are put together in different times and places, Lövkrona discusses tradition. All three scholars also seek some kind of origin and they are, to some extent, able to state where and when their legends originated.
Each of the three scholars have their special areas of interest, Anderson on the problem of self-correction, Rooth on the connection between folklore and ethnology-anthropology and Lövkrona on how to widen and improve the classical historic-geographical method. For all of them, comparison is crucial; they all roughly follow Krohn's recommendations. It seems, however, that the method has developed in two directions: a more restricted and a more liberal form. For instance, Lövkrona restricts her study of the geographical distribution of tradition to northern Europe, but she is more liberal in finding new factors for comparison, such as structural elements and redactions. Although the historic-geographical method had fallen into disrepute, according to, among others, Rooth and Lövkrona, we should not expect very many studies conducted according to this method any longer. Nevertheless, they are still published, roughly in accordance with Krohn's scheme.
Folklore Comparison Reviewed
I have demonstrated that comparison has been used for different purposes in different times. There have been scholars like Rancken and his fellow students at the end of the last century who wanted to compare for political reasons, to draw a line between "us" and "them" or to show that "we" strictly speaking belong to "them" and not to "those". This was partly repeated during the Fascist period and is familiar for modern identity studies. There have also been scholars who looked upon folklore as a set of survivals impossible to understand without comparison to parallel phenomena. In this case, a universalistic comparison such as that advocated by Andrew Lang was accepted. This kind of study is still conducted today.
Comparison was the most significant method of conducting folklore research within the historic-geographical school. The aim of comparison was to trace the origin and genesis of folklore phenomena according to the general view of humanistic studies as historical studies, with folklore representing the remains of a lost golden age of culture. This school, however, differs from the universal and all-embracing comparison for various reasons. In order to conduct comparative research, it is necessary to define comparative concepts that are so well-determined that they make a comparison between different cultural phenomena possible. The comparative concepts are, in fact, the criteria for comparison; they must be observable criteria suitable for empirical investigations which are built on the assumption of compatibility. They must have a mutual intension. Thus, definitions of folklore items are needed. Without them it is not possible to state whether two folklore phenomena have so much in common that a comparison is even possible from a common perspective. In the studies investigated here, Julius Krohn was aware of these problems, when he realised that the texts for comparison had to fulfill certain conditions. Kaarle Krohn discussed these problems implicitly when he problematised the relation between different genres, asked what a motif was, tried to stick to empirically determinable factors for comparison or reflected on phenomena that seemed superficially alike. But only those folklorists who advocated the historic-geographical method reflected upon the process of comparison.
Recently, Christine Goldberg has made a new effort to escape the limits posed by the classical historic-geographical method. In her study of the tale of the three oranges from 1997 (AT 408), she remains faithful to the method. She states that it is "the only way to give definition to a tale", that is, to demonstrate the uniqueness and the boundaries of each individual tale. Nevertheless, she is critical of the method. She abandons the central aims of the classical method, i.e., to find the primordial place, time and form and the distribution of a tale. She clearly realises that philological studies of the history of language inspired the scholars who originally designed the method. However, folklore studies require a variety of approaches. According to her, the first folklore students concentrated too much upon distribution and origin. For her, it is crucial to understand the mechanism that makes a tale hold together for centuries. Therefore, she says, it would be better to ask questions concerning stability and variation, because the method is particularly suited for the study of form. Thus, according to Goldberg, the historic-geographical method is useful, especially for the investigation of details, similarities and differences.
Consequently, Goldberg uses the idea of comparison in conducting her research, but her goal is not to find the primordial factors mentioned above. Instead, she addresses the issues concerning the boundaries of a tale. She asks "how far actual texts can differ from the tale type and still be part of it, and how close they can seem to the tale type without belonging to it" (ibid.: 62). Her main problem concerns the factors that hold a tale together. Her method consists of the comparison of repeated images and themes found in the variants of the tale of the three oranges.
Christine Goldberg is a faithful, but not an uncritical, admirer of the historic-geographical method. She maintains the comparative grasp, but subverts the whole idea of the Anderson-Krohn method. She refuses to ask questions concerning primordial place, time and form, and concerning distribution. Instead, she uses the comparative method for discovering the meaning of the tale and to reveal the factors that have held the tale together. Consequently, she does not end up with a place or date of origin, but with a statement that explains the tale’s form as a result of the harmony of the themes, motifs and images found in it. "Tales, including AT 408, are elastic", Goldberg maintains. "They can be altered in many ways to suit many tastes. To some extent these tastes are shared by members of one or another region or culture; to some extent they are idiosyncratic" (ibid.: 238). The distance from Anderson and Krohn is so great that it is difficult to recognise the classical method of folklore in Goldberg’s research. However, she opens up different points of view when she mentions that problems of variation and stability should be solved. These items are, as a matter of fact, central in folklore studies which are grounded on folklore performance and which aim to explain the meaning and function of variation, variation being a concept filled with connotations of comparison.
To conclude, the process of comparing, despite its crucial role in folklore studies, was not and has not become the subject of general theoretical interest amongst folklore students. An exception to this general rule is Lauri Honko’s article "Types of Comparison and Forms of Variation". In this article, the author tries to summarise the kinds of criticism the historic-geographical method has received and delineates one category of "opponents", including scholars such as Carl Wilhelm von Sydow, and another group of "proponents", who try to develop the old method into a model more suitable for recent practice within folklore studies. Among these "proponents", Honko lists Jan-Öjvind Swahn, whom I have not addressed here, but in a forthcoming paper, and Matti Kuusi.
However, these efforts are not enough, according to Honko. He looks for a general theory of comparison within folklore studies, and outlines a model consisting of three types of comparison which would, consequently, lead to three types of meaning.
The tradition-historical dimension "reigned supreme and was the only form of comparison recognised by the traditional ‘Finnish’ method" (ibid.: 120). Among other things, Honko blames it for being neither satisfactory nor sufficient, because the scholars did not problematise the process of cultural loans well enough to make this kind of comparison meaningful.
Honko also refers to comparison from the perspective of traditional phenomenology. However, to develop his idea on a general theory of comparison, Honko finds it necessary to problematise the concept of phenomenon. Honko maintains that this concept and the individual traits and limits of every phenomenon depend on "some kind of scholarly agreement". Consequently, the definitions are merely temporary. Moreover, when searching for the meaning of phenomena, the folklore student should be aware that they "have been lifted from their natural context and placed side by side with similar manifestations derived from entirely different milieus and contexts" (ibid.: 114).
Honko’s third type of comparison is the tradition-ecological perspective. Here the material is limited to a specific geographical area with either a macro or a micro ecological aspect. Here it is important to realise how a tradition has adapted to a specific environment. The variation thus achieved is no longer structural, but systemic. This means that it is realised not only by the scholar comparing on a structural level, but also by the individuals who tell the stories or sing the songs. >From this comparative perspective, it is possible to study not only the variants, but also the variation and different meanings that different variants have according to the circumstances of performance.
Dette er en kortversion av en lang artikkel som inngår i boken Thick Corpus, resultatet efter Folklore Fellows' Summer School 1999. I den finnes noter og referanser.