Sorting Out Donkey Skin (ATU 510B): Toward an Integrative Literal-Symbolic Analysis of Fairy Tales1

This article debates the merits of fairy tale interpretive frameworks that privilege the psychological and symbolic, versus those that utilize a literal and feminist orientation. Using ATU 510B as a test case, for its intriguing blend of real-world elements and the fantastic, the author suggests that a synthesis of literal and symbolic theories allows for the fullest understanding of the polyvalent meanings of tale, which is particularly problematic due to its depictions of incest. Drawing examples from canonical as well as contemporary versions of ATU 510B, various psychoanalytic and feminist interpretations of the tale type are put to the test, and ultimately combined to reach a more productive framework. ATU 510B, “Peau d’Asne” (also “Donkey Skin” and previously “The Dress of Gold, of Silver, and of Stars [Cap o’ Rushes]”), exists within a rich, international oral tradition and has attracted literary rewriters, from Charles Perrault and the Grimm brothers, to modern American novelists and short story writers. The tale has been problematic for publishers as well as scholars, however, due to its overt references to incest. As Kay Stone observes, ATU 510B rarely appears in collections “since the heroine is forced to leave home to avoid her father’s threats of an incestuous marriage” (1975, 46). Scholars of this tale type must decide how to interpret the tale’s elements, ranging from those that appear in real life—family relationships, rings, and dresses—to those that are clearly fantastic, like the garments that shine as brightly as celestial bodies. Interpretations of this tale tend to focus either on its manifest or latent content; however, exclusive attention to surface details instead of deeper symbols, or vice versa, restricts the potential meanings of the tale and its possibilities to address a wide range of experiences. The purpose of this inquiry is to illuminate this full range of meaning in ATU 510B and, in doing so, to outline a flexible interpretive methodology that can better account for these multiple interpretive levels. Classificatory scholarship on ATU 510B tends to focus on its divergences from the related tale type ATU 510A, “Cinderella.” Hans-Jörg Uther’s updated The Types of International Folktales describes ATU 510B’s plot generally: a king promises his dying wife that he will marry someone as beautiful as her (or who fulfills another condition), who turns out to be their daughter. The daughter delays the wedding by asking her father for magical garments, often three beautiful dresses and a coat or covering of rough fur or wood. The daughter escapes to another kingdom, works in the castle, and enchants the resident prince in her dresses. She taunts him with her secret identity, retaliation for his rude treatment of her servant persona. Finally, she slips him a token or he uncovers her, and they marry. The versions that Uther lists span the British Isles, Baltic states, Scandinavian countries, Germanicspeaking countries, Romance-language Jeana Jorgensen Sorting Out Donkey Skin 92 countries, Mediterranean countries, East European states, Slavic states, Middle East, some parts of Asia, and Europeancolonized locations in North and South America.2 Christine Goldberg’s monograph on ATU 510B gives an outline of the tale’s plot that positions the “Unnatural Father” and the “Love Like Salt” episodes as equivalent, for either one can provide the motivation for the daughter to flee and pose as a servant. Goldberg asserts that, because “the Unnatural Father (motif T411) is a character in AT 706, The Maiden without Hands,” the incestuous father’s presence cannot be used to distinguish ATU 510B from other types (1997, 31). Instead, she argues, “The essence of the Donkey Skin tale—its identifying qualities—are the heroine’s disguise and her position as a servant” (1997, 31). Yet the same could be said of ATU 510A, or ATU 923 (“Love Like Salt”).3 Thus the reason for the heroine’s flight must inform any consideration of this tale type as distinct from related tales. As the threat of incest links many variants found under these tale types that chronicle the rise of an innocent persecuted heroine in disguise, an analysis of one should be informed by the others.4 D. L. Ashliman’s analysis and collection of Indo-European incest tales relies on intertextuality to discuss the tales in light of one another. His summary of ATU 510B on his webpage “The Father Who Wanted to Marry His Daughter” is useful, as it does not rely solely on motifs to define the tale, but instead breaks the tale down into plot elements rather like Holbek’s moves.5 Ashliman’s comprehensive analysis includes ATU 510B in addition to ATU 706, brothersister incest tales, and other tales that are not as easily classified, reminding us that the boundaries between tale types are fluid. In this paper, however, I primarily limit my analysis to examples of ATU 510B. I utilize some of the well-known texts found in Ashliman’s online tale collection, “The Father Who Wanted to Marry His Daughter”: Charles Perrault’s (French) “Donkeyskin,” the Grimms’ (German) “All-Kinds of Fur” (“Allerleirauh” in German), Thomas Crane’s (Italian) “Fair Maria Wood,” and “Broomthrow, Brushthrow, Combthrow” from Austria. Ethnographically-collected texts in my analysis include Laura Gonzenbach’s Sicilian “Betta Pilusa” (in Zipes 2004, 5258), E. T. Kristensen’s Danish “Pulleru” (in Holbek 1998, 552-53), Alessandro Falassi’s Italian “Donkey Skin” (1980, 4245), James Taggart’s Spanish “Cinderella” (1990, 106-109), and Ibrahim Muhawi and Sharif Kanaana’s Palestinian “Sackcloth” (1989, 125-30). Finally, I also draw on recent literary rewrites of the tale: Robin McKinley’s Deerskin (1993), Jane Yolen’s “Allerleirauh” (1995), and Terri Windling’s “Donkeyskin” (1995). Though the above-mentioned tales are the only texts I will directly reference, I also make use of Marian Rolfe Cox’s 1893 collection of Cinderella-type tales and Anna Birgitta Rooth’s 1951 The Cinderella Cycle. Here, I take the view that it is more useful to view literary and ethnographic texts on a spectrum than to separate them completely in analysis.6 Interpretive Frameworks How we interpret fairy tales depends a great deal upon how we conceive of their relationship to reality, and the nature of this Sorting Out Donkey Skin 93 relationship remains contested. Although fairy tales’ content may not be precisely mimetic, the scholarly treatments of them have often employed the metaphor of a mirror to describe a wide variety of relationships between these texts and reality. For instance, Stephen Swann Jones (2002) hypothesizes that fairy tales are cultural and psychological mirrors, while Kate Bernheimer’s edited book Mirror, Mirror on the Wall: Women Writers Explore Their Favorite Fairy Tales (1998) is based on the premise that fairy tales offer people, especially writers, valuable sites of self-reflection. Bengt Holbek, by contrast, observes that it does not appear possible to keep the interpreters out of the interpretation, and hence that “texts are in fact mirrors in which we see our own faces rather than anything else” (1998, 402). Fairy tales do not always serve as simple mirrors, however. As Cristina Bacchilega demonstrates in Postmodern Fairy Tales, the mimetic strategies of fairy tales are ideologically motivated, making it crucial to ask not only how the fairy tale is framed, but also who is holding and manipulating that frame. Therefore, it is necessary to ask about cultural context, and to determine when fantasy can be interpreted as mirroring reality or distorting it. That fairy tales have some connection to reality is undeniable. Vladimir Propp, for instance, notes: “Obviously, the tale is born out of life; however, the wondertale is a weak transcript of reality” (1984, 84). This remark implies that while tales may not be mirror-images of reality, they are still informed by human experience. Similarly, Lutz Röhrich argues that: “Reality underpins even fantasy; not even fantasy is independent of the social conditions in the narrator’s real life” (1991, 192). Holbek goes further in his Interpretation of Fairy Tales to claim that fantasy may even be the primary instrument through which social conditions can be discussed, mediated, and escaped (1998). Social context is an important part of folkloristic analysis, but it is not always clear whether this context should privilege internal (relating to the psychology or emotions of characters and narrators) or external reality (historical or actual events). This conundrum creates the divide between symbolic and literal readings of fairy tales. Often the issue of performing literal or symbolic reading has broken down along the lines of those who see characters as exhibiting human emotions and reactions, and those who see them as flat details or psychological devices in service of a larger narrative. The question of whether characters in fairy tales represent real human beings and can be analyzed as such is subject to great variation. In the assorted versions of ATU 510B, for instance, we see behavior that ranges from earthy and human to abstract and archetypal.7 This range is necessarily related to the tension between communal tradition and individual innovation, which is at the heart of folklore performances. Naturally, diverse narrators and audiences will relate to their characters differently, with several levels of projection and empathy. Given this range, it is most productive to view the behavior and humanity of fairy-tale characters on a spectrum, extending from conceptual to concrete. Since the way in which we see fairytale characters—as symbols or humans— carries consequences for how they are Jeana Jorgensen Sorting Out Donkey Skin 94 interpreted, this dichotomy of views on the nature of fairy tale characters has given rise to two distinct traditions of interpretation: psychological approaches, which interpret characters conceptually; and literal approaches, which interpret characters as concrete. I shall give an overview of both approaches in the scholarship on ATU 510B before presenting a more subtle synthesis of interpretive methods.8 Psychological Approaches to the Tale Scholars interested in symbolism claim that literal readings of fairy tales are simply not adequate, since the fantastic elements of tales must be accounted for. Symbolic readings of folk narrative can take many forms, of which the psychological approach is but one. However, given the comparative prevalence of psychological approaches in fairy tale studies, I have chosen to foreground these theories in my discussion of symbolic approaches.9 According to Donald Haase, psychological approaches work on the assumption that “fairy-tale plots and motifs are not representations of socio-historical reality, but symbols of human experience that provide insight into human behavior” (2000, 404). This approach is useful in explicating incredible aspects of tales not found in real life. As Holbek notes: “If the meaning...of the marvelous features in fairy tales cannot be disclosed by a study of their historical origins, some kind of synchronic approach to interpretations becomes a necessity” (1998, 259).10 Psychological approaches to ATU 510B, however, do not always focus on the fantastic elements of the tale. More often, the interpersonal relationships within the tale are the focus of interpretation, though how they are interpreted often breaks down along Freudian or Jungian lines.11 Freudian or psychoanalytic approaches to ATU 510B privilege the Oedipal drama, positing that the fairy tale is told from a child’s point of view. Any projection within the tale is thus on the part of the child protagonist. A few examples from related tale types will illustrate this tendency. Dundes, Holbek, and Bruno Bettelheim all focus on explaining the incestuous father-daughter relationship from the daughter’s perspective. Dundes begins his study of ATU 706, “The Maiden without Hands,” with a discussion of the Electra complex, relating a comment by early psychoanalyst Riklin on tales where fathers want to marry their daughters: “the initial death of the mother (queen) reflected wishful thinking on the part of adolescent girls who, in terms of the Electral complex, wanted to replace their mothers vis-à-vis their fathers” (1989, 138). From this view, the cycle of father-daughter incest tales is not actually about real-life incest, but rather the daughter’s desire for her father, disguised and embedded within the tale’s plot. Other Freudians arrive at similar conclusions. Bettelheim states that Cinderella’s “degradation—often without any stepmother and (step)sisters being part of the story—is the consequence of oedipal entanglement of father and daughter” (1975, 245). Ben Rubenstein connects young Cinderella’s “phallic strivings and penis envy” to “the sexual pursuit of the daughter by the father,” whom the daughter desires (1982, 225). In contrast, Jungian approaches to fairy tales eschew the infantile sexuality hypothesis, focusing instead on Sorting Out Donkey Skin 95 universalized masculine and feminine values. Marie Louise von Franz’s approach is typical in that she does not address incest or conditions of power and abuse in real life in her discussion of ATU 706, “The Maiden without Hands.” All of the characters in this tale represent inner figures within a psyche, with the father as the destructive animus and the future husband as “a collective dominating positive spirit” (1993, 95). Marion Woodman also views the fatherdaughter tension as part of a soul’s journey. In Leaving My Father’s House: A Journey Toward Conscious Femininity, she uses the Grimms’ version of ATU 510B as an organizing metaphor for every woman’s voyage from patriarchal trauma to contact with the eternal feminine.12 This analysis focuses on collective experiences with individual testimonies woven in. The fairy tale is meaningful only insofar as it expresses the Jungian paradigm of feminine development that Woodman follows. Holbek surveys Freudian, Jungian, and other psychological approaches to fairy tales, claiming to reject each one owing to his inability to “pronounce any one of them more ‘right’ than the other” (1998, 319). However, in his own analysis, Holbek resorts to what Vaz da Silva criticizes as “fairly standard Freudian symbolism” which attempts “to reduce all symbolic expressions to emotional impressions” (2000, 7). The tendency for psychological approaches to be reductionistic is one problem with applying them to fairy tales, which are by nature polyvalent. Additionally, most of the psychological theories applied to folklore were created to be used in therapy and not the interpretation of cultures and texts. So this must also be taken into account. Thus, the origins and uses of psychoanalysis in particular should be questioned before their relevance to fairy tales is unconditionally accepted. For example, the Electra complex is instrumental to the psychoanalytic interpretation of ATU 510B and kindred tales, and is an especially contested idea. Kilmartin and Dervin point out that, because Freud did not see the developments of males and females as analogous, “the Electra complex and the Oedipus complex should not be represented as parallel” (Kilmartin and Dervin 1997, 269). In fact, they continue, “the Electra complex is actually not much more than a footnote in psychoanalytic history, rejected by the father of the field (Freud) and later ignored by the very person who coined the term (Jung)” (Kilmartin and Dervin 1997, 269). Today, even Freudian scholars such as Alan Dundes have questioned whether the Electra Complex’s underlying notion of “penis envy,” Freud’s idea that a girl would reject maternal identification and desire to be with/like the father, “isn’t simply a form of male projection” (1982, 220). Casting the existence of penis envy into doubt thus threatens the legitimacy of the Electra complex, which would invalidate most psychoanalytic interpretations of ATU 510B. In addition to problems with the terminology associated with psychoanalysis, there are feminist concerns about the validity of psychoanalytic principles. Initially articulating his “seduction hypothesis,” which concluded that his female patients actually were telling the truth when they said they had been sexually molested, Jeana Jorgensen Sorting Out Donkey Skin 96 Freud later shifted his views, dismissing these claims as fantasies (Warner 1994, 350). This shift has had repercussions for the fields of psychology, therapy, and any discipline that employs Freudian ideas. The authors of “The Emergence of Child Sexual Abuse from the Shadow of Sexism” examine “the attitudes and prejudices that blamed the victim, minimized her experience, and held her accountable” in light of “Freud’s legacy... which has pervaded our conception of child sexual abuse for nearly 100 years, informing medical, legal, and psychiatric treatment” (Bayer and Connors 1988, 12). In the evolution of Freud’s ideas from the seduction hypothesis to the Oedipus and Electra complexes: “The real experience of sexual abuse, which was the basis for the earlier theory, was turned into a fantasy of longing and seduction in which the child is transformed from a victim of adult power to a willing participant in the sexual fantasy” (Bayer and Connors 1988, 13). It is an interesting twist on wish-fulfillment that Freudian ideas like the Electra complex “reinforce feelings of self-blame on the assumption that children’s feelings are responsible for the alleged trauma” (Bayer and Connors 1988, 14, italics in original). In the case of ATU 510B, the protagonist’s desire for her father’s love can be seen as the motivation for the tale’s plot, but only if the entire tale is understood as fantasy, not mimesis. That is, any incestuous feelings on the part of the girl toward her father would seek an outlet through a fantasy that exculpates her from responsibility for taboo desires. Projection and identification present further problems with the psychoanalytic approach to fairy tales. For instance, Francisco Vaz da Silva asserts that projection cannot be attributed to fairy tale characters at all. Instead, he declares, “there cannot of course be such a thing as projection on the part of a fairy tale character” which means “projection is to be ascribed to narrators identifying with” characters (2000, 5). In Vaz da Silva’s view, then, projection can be seen as an artificial construct imposed by narrators upon characters, or worse, as a theoretical frame forced by the interpreter. Another concern with projection is the difficulty of determining who is doing the projecting upon whom. Specifically, the premise that fairy tales are told from a child’s perspective raises issues of culpability and authority. Maria Tatar notes that people are drawn to readings of fairy tales that blame children’s sexuality for what befalls them “in part out of a desire to avoid facing the ‘unpleasant truths’ that emerge once we conceded that some of the events staged in fairy-tale fictions can be as real as the fantasies they seem to represent” (1992, xx-xxi). In tales like ATU 510B, where the protagonist is threatened with incest that she supposedly desires from a psychoanalytic view, the incest that is present in the text is thereby dismissed as a subconscious fantasy, not something that might be present in real life. If tales of abuse are not, as psychoanalytic interpreters assume, told from the children’s perspective, then they mean something very different about power relations. Jack Zipes claims that fairy tales, along with other folk narrative genres, express an adult perspective on power and the family (1995, 220). That is, they rationalize the self-serving actions of adults, while portraying Sorting Out Donkey Skin 97 the consequences for—not the desires of—children. Interestingly, Otto Rank, a member of Freud’s circle in Vienna, viewed father-daughter incest stories as fantasies from the father’s perspective (1992), hinting at a more culture-reflective view of the tales’ content. Jungian interpreters have, at times, completely ignored any reality of violence, instead viewing the fairy tale as a journey of the soul, disconnected from any reality of sexual abuse. Woodman, for example, acknowledges that physical incest occurs, but uses ATU 510B to address psychic incest, which she believes is more prevalent and requires women to re-evaluate how they relate to femininity and masculinity—rather than addressing the underlying social and sexual problems of living in patriarchal societies. At their worst, psychological analyses of fairy tales not only ignore realities of abuse, but also rationalize them. For example, Bettelheim’s Freudian explication of the protagonist’s punishment in ATU 510A, “Cinderella,” uses the problematic idea of masochism: “deep down a child who knows that she does want her father to prefer her to her mother feels she deserves to be punished for it—thus her flight or banishment, and degradation to a Cinderella existence” (1989, 246). This mentality can be used to justify violence against children, as in Ben Rubenstein’s analysis of his young daughter’s behavior in the context of the Cinderella story and psychoanalytic theories of masochism. He writes that his daughter’s “unconscious constructed a beating phantasy [sic] arising sharply out of her castrated feelings” (1982, 226). Further: “It was obviously her wish that I would come and beat (attack) her. I had, many times, been quite impressed by the obviousness of this wish” (1982, 226-27). This interpretation exemplifies many of the problems with psychoanalytic approaches to fairy tales—its dogmatic lens that departs from the text, and its reliance on the ideologically skewed roots of psychoanalysis—in addition to offering a disturbing glimpse of how a fairy tale can be read to justify one’s own power by blaming the victim. Literal Approaches to the Tale By contrast, scholars such as Jack Zipes reject psychological approaches to fairy tales in favor of highlighting the realities of “child abuse, neglect, and abandonment” in fairy tales (1995, 220). Literal-minded scholars are also aware that psychological interpretations can detract from the sobriety of problems like sexual abuse by transforming them from social crimes into harmless and universal developmental issues. A literal interpretation of ATU 510B highlights the incest motif as reality, not symbol. Though in many versions of ATU 510B the incestuous act is threatened rather than carried out, I consider even the threat of sexual violence to be damaging and thus worthy of discussion.13 In this light, the magical elements become metaphors for the heroine’s experience of abuse, as well as reflections of a patriarchal society that condones such abuse by veiling or ignoring it. Often, literally-oriented scholars draw attention to the problems inherent in viewing the father in the tale as faultless, as psychoanalytic interpretations seem to portray him. Indeed, all motivations—of characters in tales, and of the interpreters themselves—are questioned. As Alice Jeana Jorgensen Sorting Out Donkey Skin 98 Miller points out, psychoanalysis has generally had a one-sided view of the parent-child relationship: “The way parents actually feel about their children is brought out very clearly in fairy tales...In psychoanalytic literature, on the other hand, parents’ feelings towards their children are hardly ever the subject of research” (1984, 237). Projection and power remain key issues in literal interpretations of ATU 510B, though seen from a different angle than in psychological interpretations. Two approaches to literal interpretations of fairy tales are the historical and feminist perspectives, which are complementary but with differing assumptions and goals. I characterize as “feminist” those approaches that not only focus on women’s perspectives, but also actively theorize the connections between gender, sexuality, society, and power.14 Some of the historical approaches to incest tales border on myth-ritualism in their attempts to elucidate the inclusion of incest in folklore; Dundes characterizes these approaches as attempting to explain away the “monstrous” incest motif with rationalizations that reach far back into ancient times (1989, 136-37).15 In Holbek’s methodological remarks on socio-historical approaches to fairy tales, he concludes that all the sources he surveyed “pointed to a connection between the contents of the tales and the living conditions of the narrators” (1998, 400).16 Helen Pilinovsky deals with the literal possibility of incest in her essay “Donkeyskin, Deerskin, Allerleirauh: The Reality of the Fairy Tale,” in which she traces recent literary transformations of ATU 510B as well as the symbolism found within oral tradition. Pilinovsky focuses on the tale’s potential to convey the heroine’s positive responses to danger: “Although the princess in each version of tale type 510B does face the possibility of sexual abuse, she is portrayed as being able to avoid it through a combination of unlikely luck...and unrealistically achieved accomplishments” (2001). Pilinovsky does not take into account psychoanalytic scholarship, dismissing the possibility that the heroine desires her father’s attention by stating that abuse is never the victim’s fault. She rejects psychoanalytic theory on the basis of its “number of problematic qualities,” only acknowledging “examples in the genre of the fairy tale have been said to serve a therapeutic psychological function” (2001). Literal approaches tend to follow the culture-reflector theory of folklore, which, in the case of ATU 510B, would state that the manifest details of the tale reveal something about the culture in which it is told. D. L. Ashliman, in his essay “Incest in Indo-European Folktales,” refers to the cultural context which sustains the plot of ATU 510B: “Reflecting the patriarchal values of the society that used them, these folktales seldom challenge a father’s authority to do with the members of his household whatever he pleases” (1997). This is congruent with the context-sensitive approach that Ibrahim Muhawi and Sharif Kanaana use in their study of Palestinian-Arab folktales, which include a variant of ATU 510B, “Sackcloth.” Muhawi and Kanaana interpret the heroine’s plight thus: “the sexual awareness begins even before the girl leaves home, producing feelings of confusion, shame, and guilt, especially Sorting Out Donkey Skin 99 since she seems to arouse a most unnatural passion in her father. Hence her desire to cover her body completely, so as to appear to be not only of the opposite sex but also a horrible freak whom no one would want to touch” (1989, 145). Some literally-oriented scholars incorporate psychological approaches, even while acknowledging skepticism toward the power structures contained within. Ashliman states, “folktales dealing with father-daughter incest often reflect a psychological projection of unresolved Oedipal issues” (1997). Further, “the description of the fleeing daughter’s ‘rescue’ gives further credence to a psychological projection interpretation of type 510B tales” because “the man who discovers and ultimately marries the runaway princess closely resembles the girl’s own father. He too is a king, exerting despotic, patriarchal authority over his household” (1997). Ashliman also examines the rhetorical devices that transfer culpability from the father or obscure his role in the incest. Examples include the promise extracted by the king’s dying wife to remarry under certain conditions, or the daughter’s initiative in trying on her mother’s ring (or shoe in some versions). Literal interpretations convincingly highlight power imbalances within patriarchal family structures, but they do not account for all of the tale’s features. Forcing certain details that appear in any one version of the tale to conform to a case study of incest is difficult at best. Isolated aspects make sense, but others are too marvelous or too unique to fairy-tale structure to be interpreted in a realistic light. For instance, why would a tale depicting the trauma resulting from incest seek to exculpate the perpetrator, and not punish him as so many female villains are cruelly punished in other tales?