Christine Goldberg, The Tale of the Three Oranges.
Folklore Fellows’’ Communications No. 263. Helsinki: Suomalainen Tiedeakatemia (Academia Scientiarum Fennica), 1997. 268 pp.
Hard (ISBN 951-41-0806-X), FIM 155,
Soft (ISBN 951-41-0807-8), FIM 130,
Available at the Tiedekirja bookstore
Kirkkokatu 14, FIN00170 Helsinki, Finland
fax: +358 9 635017; e-mail: tiedekirja@tsv.fi
Christine Goldberg’s topic is probably the most classical one that one can imagine in the history of folkloristics: a monograph on a wide-spread folktale. It raises ambivalent expectations, because it conjures up phantoms of the comparative, historic-geographic method: archetypes, places of origin, meticulous comparisons of available texts (“variants” placed in a text-critical “stemma”), routes of transmission, and the original idea, moulded time and again through oikotypification and “local redactions”, all concepts which few modern folktale scholars are likely to adopt without modification. Christine Goldberg has the advantage that her object of study, the Tale of the Three Oranges (AT 408), well-known through Sergei Prokofiev’s opera L’Amour des trois oranges, is not only an interesting narrative with a multi-faceted cultural history in Europe and Asia but also a topic of considerable methodological importance because of its research history. Yet it is fair to say that the main interest does not focus on the story itself but on the methodological operations which the analyst must introduce in order to save her credibility vis-á-vis outmoded theoretical thinking.
Christine Goldberg is well aware of her paradigmatically precarious situation. The conclusive chapter, where she sums up her results concerning the age, place of origin and routes of dissemination of the type and subtypes of AT 408, extends to a methodological discussion on the historic-geographic method as utilised and developed by Kaarle Krohn, Walter Anderson, C. W. von Sydow, Jan-Öjvind Swahn, Warren E. Roberts, Anna Birgitta Rooth, Max Lüthi and Goldberg herself, not forgetting certain criticisms expressed both inside and outside the paradigm. She challenges her readers to discuss key methodological issues such as the origin of types: “Whether any given tale type has resulted from a single archetype (Urform), or whether it has been cobbled together, perhaps more than once, from preexisting components, is a question that needs to be investigated and, perhaps, debated” (p. 235). At this point she is ready to make a compromise and abstain from the single-origin orthodoxy of the historic-geographic method in order to allow for a multilinear development of tale types, whereby similar combinations of what she calls “episodes” have been made more than once. Instead of a monogenesis and linear development of the tale, we thus get polygenetic compilations of smaller tale-units in different formats.
The Three Oranges is a case in point: it is a multiepisodic tale and all its episodes except one (the emergence of fairy maidens from oranges) can be found in other tales, in combination with other episodes or as core episodes of independent tales. Four key episodes of AT 408 had existed for a long time before two or three such episodes formed clusters “and spread over various regions”, and when “from these clusters” (162) a complex tale was created by some storyteller(s?). Where and when this happened in the tale’s wide distribution area (India, Persia, the Middle East, the Mediterranean countries from Greece to Portugal, South America), the author is unwilling to say. The tale must have existed for “the past four centuries at least” (237), which is no major concession considering that the tale appears, in all its complexity, in Giambatista Basile’s Pentamerone in 1634/36. The tale has two main subtypes, but their order of emergence, mutual relations and routes of dissemination seem so complex that the author shuns a definitive theory and speaks of “considerable traffic in folktales” (217) instead.
The rich Indian tradition constitutes a special problem: the author remains uncertain whether India was the home country or merely a recipient of some of the constituent episodes of AT 408. She confesses that “… the many points of difference between the Indian tradition and the western tradition, plus the considerable variation in the Indian tales, reduce any conclusions here to mere guesswork. Certainly India is not a candidate for the home of the motif of the trebled fruits. Regardless of where the tale was composed, the fact that each region has developed a distinctive form of AT 408 is an eloquent example of the process of oikotypification that results in regional subtypes attuned to local symbolism.” (148) This comes close to capitulation, and even the last statement, a sensible reference to the milieu adaptation of tale traditions, does not seem to hold the house of “variants” in particularly good order. Later the author concedes that “only a small amount of the variation in AT 408 can be explained through geographically-defined subtypes or oikotypes” (217).
There are basically two ways to deal with a research paradigm which begins to fail. One may loosen the rigidity of its basic tenets, i.e. start making compromises. The risk here is that the paradigm will lose its analytic acumen. Or, one may try to patch up the methodological poverty by offering alternative/complementary methodology, new research angles, which may or may not pertain to the questions posed by the original paradigm. Christine Goldberg chooses to take both approaches. Let us first consider the compromises.
There is a basic ambiguity in the entire study, because it takes the concept of natural tale types very seriously but lays the main stress “on the tale’s particular composition and the extent of its variation”, not on “its original form and its history” (237). This shift of focus, probably aimed at creating a distance to the traditional historic-geographic method, raises the question, however, whether such sporadic, thin and contextless material is fit for the study of (real) composition and (real) variation (cf. below)? Despite the effort, the distance to the historic-geographic method remains close indeed because of the role given to Walter Anderson’ s unpublished study on The Three Oranges from the 1950s, made available by the Archives of the Enzyklopädie des Märchens in Göttingen. The author compares her results with Anderson’s from a position of superiority, the grand old master of the method being unable to defend his unfinished manuscript. This ethically somewhat problematic constellation has two consequences: we may follow not one but two analyses of largely the same material and the dominant line of thought becomes historic-geographic after all, also in Goldberg’s case, because she systematically reproduces Anderson’s Urtext and comments on his reasoning. Control studies are rare in the humanities, but here we have a pair.
The lucidity of Anderson’s frequency-analytic method comes through fairly well although his analytic text remained fragmentary. Where Anderson operated with numbers, Goldberg has three vague categories: a feature may appear “in most examples”, “in a substantial number of texts” or “in occasional variants” (61). Where Anderson had close to 500 variants meticulously analysed, Goldberg has 350 (150 full texts, 150 “summaries” plus fragments), “three-quarters of which have been coded individually” (61). The bookish filter through which this material comes creates a distance between the original, presumably oral, performances and the available texts. “Summaries” do not tolerate minute analysis, as the author admits. The orality of the material is obviously a problem which could have been discussed from the point of view of empirical field work on storytelling. Somewhat surprisingly, the author does not choose (or find?) any cases of particularly well-documented tales, i.e. examples based on good contextual information about storytellers, their composition techniques and performance situations, for close scrutiny. The interest in form, composition and variation might have profited from case analyses on a qualitatively higher level. Now the storyteller remains a purely hypothetical integer.
The study of form presupposes a respect for “all variants” even more than the customary historic-geographic reconstruction of archetypes. Yet the author seems to need fewer variants than e.g. Anderson: “I found that amassing more and more variants was an exercise in diminishing returns: the second hundred examples more or less duplicated the first hundred. This feeling was confirmed when I read Anderson’s analysis of nearly five hundred variants. Beyond a certain point, more examples do not lead to new insights, or even to firmer foundation for conclusions” (54). If this impression is due to contextless, “bad” variants, it could be said. Otherwise the compromise may be difficult to accept in a study where both composition and variation are in focus and where each tale must be read and accepted as a result of a particular performance. The essentialist choice of material is also in conflict with the stress laid on the tale’ s endless ability to vary and its episodes’ tendency to derive from and recombine with other tales.
Goldberg’s comments on the previous research on The Three Oranges and affiliated tales are mostly critical, often quite refreshingly so. From the control study point of view it must be said that her results do not generally corroborate Anderson’s analysis. Stith Thompson also comes in for his share of criticism because of his handling of the tale type(s) in question. For some reason, dead scholars deserve less discretion than living: critical references to the latter are occasionally (88, 218) made without name and bibliographic entry.
Yet the old paradigm proves a formidable adversary as soon as one does not abandon it totally but is satisfied to “renew” or “complement” it. Inherited arguments slip in inadvertently and the vocabulary maintains the bondage to established conventions. An example is “archetype”, which Goldberg continues to use despite critical statements concerning its value (234). The idea that archetypes may be feasible in small areas, with tales showing less variation and/or single-episodic structure, represents actually a capitulation of the method which was originally developed for complex narratives found in geographically wide areas. Goldberg deserves credit for a clear stand against the text-critical model of comparison: “A narrator remembers, or prefers, the dramatis personae from one source, and the riddle questions from another. This idea explains why the stemma model, which works so well to explain the relationship among the exemplars of written texts, so often fails to apply to the variants of oral traditions” (228). Other arguments may be adduced to this conclusion, one being the different forms of oral variation (see my article in Journal of Folklore Research 23/1988: 105124).
Actually, the historic-geographic method does not tolerate variation despite its insistence on maximal availability of variants. Both Anderson and Goldberg avoid the construction of Urtext or “normal forms” when variation becomes really big. Even the tale type falls apart at episodes which vary too much. The amount of unexplained variation swept under the carpet in favor of “essential” traits used for actual hypotheses is so large that the analytic job must be regarded as half-done at best. The material consists of “variants” wide apart in the point of time and geographic location of their emergence. Such a thin net does not represent real variation in the sense of cultural relevance; behind each text there must be hundreds or thousands of comparable performances which may be radically different but which the analysis does not reach. The organic variation to be found through multiple documentation of narratives by one storyteller or in a cohesive group or region where people have the possibility of contact and exchange is not reflected in the bodies of mostly contextless variants created for the historic-geographic comparison. Even the question of “what is it that actually varies” tends to remain open because of the thinness and artificiality of the body of material.
It is no wonder that the paradigm has become context-resistant. The justification for genetic links between texts must be sought in a declaration of independence and cross-cultural identity for the tale under study. Christine Goldberg believes in natural cross-cultural tale types but she also senses the difficulties; about a handful of possibly affiliated tales she states: “It is not easy to determine which correspondences among these tales represent ’’genetic’ connections (i.e. mean that the tales are part of a single tradition) and which are coincidental” (185). Coincidence is hardly an explanation. When criticising the dating of a subtype of the Cinderella tale complex by Anna Birgitta Rooth, Goldberg comes up with a better suggestion: “I am disinclined to credit modern tales with such great age (which implies great stability throughout that age). Just as the entire tale of The Three Oranges is not necessarily directly related to the Asian tradition that utilises its episodes V and VI, the entire AT 511A does not necessarily predate the Egyptian Two Brothers tale. A more conservative (and to me more credible) explanation is that several of the same motifs are configured together in both the ancient and the modern traditions” (187).
In a nutshell, Christine Goldberg’s main contribution to the historic-geographic method has been formulated by herself: “The difference of opinion between Krohn, who wanted each folktale studied separately, and von Sydow, who thought that groups of closely-related tales should be considered together, reflects the different opinions of these two scholars as to the natural separateness of the tale types. To this disagreement, which can only be settled on a case-by-case basis, I would add a third suggestion: that (at least for some tales including AT 408), in addition to variants of the complete tale type, examples of its episodes in other narratives need to be considered” (237).
This realisation brings us across the border between compromise and alternative research. In much of what Christine Goldberg writes, there is a predilection for basic images which dominate human comprehension and bring forth “motifs” and “episodes” capable of joining such images time and again with narrative plots and with each other. From this angle (and without mentioning Ernst Robert Curtius) she composes small chapters about the landscape in The Three Oranges: water and trees; souls in trees and plants; citrus trees and fruits; and repeated transformations (ch. 5). The method is not “motif history”; rather I would call it motif phenomenology. It certainly helps the reader to get a glimpse of a vast cross-cultural array of meanings and it serves the purpose as long as it stays near the texture of the tale(s) under study. But as soon as “trees” are discussed in a most general way and all “water” is made inhabited by spirits, the risks of the old phenomenology of religion (e.g. by Gerardus van der Leeuw) come to mind. The vastness of relevant materials makes the treatises look like midgets and the generalisations too simplified.
Beyond a mere summing up of results and a rehearsal of memory studies, the last chapter of the book fortifies the wall against methods not adhered to by the author. “Performance” and “context” have been kept at a distance throughout the book, and on the last page we read: “Such a tale [AT 408] is bigger than any particular social context, and its obvious general appeal even challenges the notion of cultural context” (238). The real actor is not the storyteller nor the audience nor the culture which the text may reflect, it is simply “it”, the tale itself, the subject of its own history. By context is meant “the context of the other variants of the same tradition”, a papery concept indeed.
In natural science, shift of paradigm often is a point of no return for the old paradigm. Whole libraries tend to become obsolete overnight. Fourteen years ago the slow development of folkloristic methodology made me write: “The ’’shift of paradigms’ may be fairly dramatic even in the humanities, although if it may not lead to the extirpation of the old paradigm in the long run. To some extent humanistic paradigms seem to be complementary rather than mutually exclusive; this is a deviation from Kuhn’s concept.” (Studia Fennica 27/1983: 19.) To me personally, the text-critical model of the historic-geographic method is stone dead and unable to explain oral variation. Yet, I can understand that scholars lacking empirical experience on the process of oral composition are enticed to apply it. My prediction is, however, that future generations will work more with “thick” materials and organic variation than with “thin” materials and artificial variation. If compromises can save a paradigm, which I generally doubt, parts of the historic-geographic methodology may survive refashioned into a tradition-phenomenological, cross-cultural comparative method. The purpose of that method will be to remind us of the universals, the traditional elements constitutive of human culture in general and capable of continuous configuration and return. In that perspective, even The Three Oranges may have a story to tell and Christine Goldberg’s important study may prove to be ground-breaking.
Lauri Honko
University of Turku
FF Network No. 14
(December 1997): 2326