Christian de Pee has written an extraordinary book, yet one plagued by a deep, persistent flaw. The book does not concern so much middle-period weddings, although we learn a great deal about them, but texts and how we use and learn (or mislearn) from them. In De Pee's words, “This is a book about the practice of text … and in the process reassesses the relationship between the Middle-Period text and the practice of the historian” (p. xii).

The heart of the book is De Pee's masterful textual analyses. Over four chapters, he explores (1) ritual manuals; (2) wedding correspondence and nuptial songs; (3) calendars, almanacs, miracle tales, and medical texts; and (4) legal codes and contracts. In chapter 1, De Pee focuses on a debate between practical accommodation and classical imperatives. The mature Tang, he argues, saw “the ancients” as “forbiddingly remote” and so impractical; thus, they sought accommodation between the ritual heritage and practice (p. 23). By the eleventh century, however, “scholars were gaining confidence in their ability to recover ancient ritual” (pp. 22–23) and denounced the Tang accommodations. He links this change to the fascination with the material remains of the ancient world, specifically to the bronzes that fascinated men such as Ouyang Xiu (1007–72): “The cauldrons and beakers created a connection between Song literati and the ancients that was both historical and timeless” (p. 23). It was Family Rituals, identified with Zhu Xi (1130–1200), that rescued ritual reconstruction from “the ponderous, bewildering archaism of the Northern Song” and made it manageable (p. 82).

As the ancient inscriptions were transcribed to paper and reproduced in anthologies, De Pee insightfully argues, they became something different when removed from their original ritual and physical contexts. The texts of ancient bronzes transcribed to paper lost their original power. Moreover, this change applied to the manuals, for what they transcribe resembles only superficially real practice. The historian, he asserts, must bear in mind that the manual outlines how the wedding couple underwent ritual transformation, both as individuals relating to others and as sexual beings intended to prolong the patriline. The danger lies in ascribing a literal, invalid universalism to the text. Ritual was diverse, as real individuals sought to practice it in culturally and economically diverse contexts, and the texts that survive for the modern historian are only a veneer of what was truly done. Reflecting on the forces that allow texts to survive, De Pee reminds us, “The sole historical value of Song ritual manuals lies in their documentation of an attempt at ritual hegemony by the new elite … and in their preservation of occasional popular customs” (p. 85).

De Pee follows a similar line of analysis in the three following chapters as he explores the relationships between wedding correspondence (chapter 2), calendars and such (chapter 3), and law (chapter 4) and a hypothetical reality that he admits cannot be fully reconstructed. We know what the surviving corpus tells us, but that corpus is both particular in its perspective and incomplete in its totality. In emphasizing how the record both preserves only a portion of what once existed and overwhelmingly reflects the emergent Neo-Confucian orthodoxy, De Pee makes an important historiographical assertion, which scholars must take into account. It is the eternal conundrum of historiography. Moreover, through his extraordinary command of the literature, De Pee is solidly positioned to make his argument. His familiarity with the texts is faultless and enables him to suggest what might be missing and what it might mean to our understanding.

But here he goes astray. De Pee builds on the work of others; he is not the first to address wedding practice or ritual, or the first to consider the meaning of ritual in China's middle period. Nor does he claim to be. What he does claim, however, is to be the first to read the record correctly. In assessing the work of his predecessors, De Pee observes,

He decries, too often in very personal terms, the “structuralist assumptions” and “universalist concepts” (p. 17) that he asserts undermine the work of others. In his uncompromising criticism of the work of others, he misses an equally important historiographical truth: historical truth is a very elusive goal, which is rarely, if ever, attained and can only be approximated through the many surviving avenues. Like those whom he criticizes, De Pee is only bringing “another” perspective, albeit one a very important and brilliantly presented one, to our search for the “truth” about China's middle period.