In Shifting Stories: History, Gossip, and Lore in Narratives from Tang Dynasty China, Sarah Allen aims to change the way that scholars conceptualize the Tang tale corpus. By broadening the scope of what is commonly considered a Tang tale to include biji and zhigua i, and by reminding us of the fluidity of these texts, Allen shifts the discourse away from generic and fiction/nonfiction distinctions to an examination of how the building blocks of these tales (i.e., the narrative formulae) were manipulated to produce different narratives for different purposes.

Allen identifies two main types of tales: lore, which relates “incidents that transgress the boundaries of ordinary experience and touch on the occult” (p. 2), and gossip, which includes “tales that relay particular information about particular historical people” who were “presumably known to the tale's audience through reputation if not personal acquaintance” (p. 3). Allen believes that both types of stories served not only to entertain but also to impart lessons about how the world works and to present historical information about the past. Regardless of type, these tales relate private experiences and information that have been excluded from the public record.

In chapter 1, Allen argues that Tang tales originated in gossip among elite men, who were interested in an “exchange of information” rather than in the literary quality of the tales (p. 25). As participants in a manuscript culture, these elite men felt free to co-opt a story and make it into their own. Next, Allen turns to a comparison of three clusters of three analogue tales that narrate the same received story, but vary in details and emphases, in order to show how shared core narrative elements (i.e., story) among the tales in these clusters are reshaped into different tales.

In chapter 2, Allen describes another function of Tang tales as a site for gossip. She again compares several tale clusters against accounts in standard histories, and concludes that these discrepancies represent matters of contention that were debated at the time. She also suggests that some versions might have been written in hopes of influencing history. The examples that Allen provides add to the growing body of scholarship regarding Song historians’ reliance on tales as a source of historical information.

In the next chapter, Allen posits a “chronological progression” from urban legend to subverting audience expectations to “literary play” (p. 124). She is not suggesting that the tale itself “experienced a progressive evolution,” but rather that there was a “shift in writers’ and readers’ attitudes toward the tale” (p. 195). In the first half of the chapter, Allen uses “the adventure formula,” or meeting a stranger, to describe the first phase in this progression, urban legend. The second half of this chapter shows variations of such a formula. Allen argues that these changes did not represent conscious manipulation of the adventure formula by the authors, but rather that they were aware of “a matrix of commonly followed possibilities that defined a range of expectations for what was likely to happen at a given juncture when a traveler meets a stranger” in a tale (p. 146). She then argues that by the late eighth and into the ninth century, tales entered a second phase, in which the “aesthetic and philosophical possibilities” (p. 163) of these inherited narrative patterns were explored and manipulated, often resulting in the subversion of audience expectations. Allen shows how the relatively better-known tales both fit into and deviate from the adventure formula in order to achieve a particular narrative outcome. Allen specifically focuses on the extended interaction between the traveler and the non-human stranger, which allows writers to explore human standards for moral behavior.

In chapter 4, Allen contends that by the ninth century, there was an increased focus on the writers’ skill in crafting an entertaining tale, rather than on his role as a recorder. This represents the third phase of the progression, “literary play.” In these tales, a stranger meets a group of strangers who reveal their true identities (animals, objects, and historical or mythological beings) through puns and riddles expressed through lyric poetry. These riddling tales showcase writers’ talents and challenge the reader to figure out the true identities of the actors before they are revealed within the story. Readers are alerted to this challenge through employment of the recognizable adventure formula.

Chapter 5 complicates the neat narrative of “specific authors bound to stable texts” (p. 201) created by anthologies and modern collections of tales, by arguing that the idea of authorial authority of texts and definitive texts was not firmly in place until, at least, the end of the Tang. Authors felt free to alter received texts depending on their purposes for the narrative, just as gossipers might feel free to change subsequent retellings of an event. It was the information in the story, not the story itself, that was important. Allen sees in this evidence that, throughout much of the Tang, people viewed tales in much the same way as they viewed gossip.

Using a case study of the Yiwen Ji, attributed to Chen Han, Allen argues in chapter 6 that by the mid-ninth century there was a shift to thinking about tales as “(comparatively) closed literary texts, works for which at least part of their appeal lies in their presentation of their content itself and for which the discourse is important as well as the story” (p. 242), as well as a growing interest in attaching texts to particular writers. Chen Han, working within the anthologizing tradition, treated the tales in his collection as texts, rather than as stories to be reworked depending on his whims. Chen Han's anthologizing has broader significance, as well, for the literary history of tales: first, as a declaration of the value of the tale as a literary form worthy of preservation, and second, as a statement about the form itself—that individually circulating tales were gathered in a single collection shows that Chen Han considered them as belonging to a single category of literature.

Throughout her book, Allen is haunted by the specter of fictionality. Early in the introduction, she does an excellent job problematizing Lu Xun's influential opinions that the Tang dynasty was when writers “first deliberately wrote fiction (xiaoshuo),” and that Tang tales represented “deliberate creation,” that is, fiction (pp. 5, 7). Lu Xun's comments have shaped scholarly discourse ever since, resulting in the division of Tang narratives into “fiction and documentation,” which are “characterized respectively also by imagination and reportage, complexity and simplicity, literariness and plain writing” (p. 7). Allen rightly points out that the actual corpus of Tang tales is not as neat as this, and seems to suggest that we should move away from this obsession with the fictionality of Tang tales, in part because it is too restricting. Yet, Allen returns to this question of fictionality again and again, even as she admits that it “matters little” (p. 197). Moreover, in discussions about fictionality, Allen mainly relies on secondary sources on European literature. Turning to definitions of Renaissance-period fictionality as indices by which to judge the inventiveness of Tang Chinese tales remains ultimately problematic.