Every once in a long while, a book comes along that challenges the basic assumptions of an entire field. Hartman shows convincingly that much of what historians think happened during the Song dynasty (960–1279) did not occur as recorded, whether in surviving primary sources or in the English- and Chinese-language secondary sources based on them.His key insight? The sources reflect the time in which they were composed—not the times that they purport to describe. Why has it taken so long for Song historians to realize it? We have long known that the written record is heavily edited, as the standard descriptions of how the History Office, first established in the Tang dynasty (618–907), reveal. Every day, two court historians drew up records of the emperor’s actions (“The Diary of Activity and Repose,” qijuzhu), which they then revised into monthly summaries. Combining these summaries with records from other offices and biographical records of high officials, the History Office produced a daily calendar of events, which they then edited into a single Veritable Record, a chronological record of all the events that occurred in a given emperor’s reign (5–6).Hartman devotes a chapter to each of four major sources—The Recovered Draft of the Song State Compendium (Song huiyao jigao), The Long Draft Continuation of the Comprehensive Mirror That Aids Administration (Xu zizhi tongjian changbian), The Chronological Record of Important Events since 1127 (Jianyan yilai xinian yaolu), and The Song History (Songshi). Carefully combining what is known about each book with a rigorous examination of the text, he reconstructs the stratigraphy, or different layers, of each book. This “archaeology” of texts is the book’s primary use of interdisciplinary method. An oft-cited anecdote tells how Song Taizu (reigned 960–976), the founder of the Song dynasty, persuaded his generals to cede their military power to the emperor and his civil administrators. Eight pages of textual archaeology demonstrate that the anecdote had no basis in contemporaneous sources and explain how historians reworded the anecdote to reflect the concerns of each generation (303–311). Throughout the book, Hartman’s reasoning is complex and detailed and his exposition clear and convincing.His findings are most surprising for The Song State Compendium, the source that has the appearance of being the closest to archival materials because it retains the precise formatting and feel of original entries—the very reason why Song historians have relied so heavily on this source. Hartman’s investigations show that the Compendium does not contain an equal number of entries for each year that it covers between 960 and 1224 (43, Figure 1.2). In this extraordinary discovery, Hartman identifies periods of more extensive and less extensive coverage, concluding that all diachronic comparisons are risky. Song historians are well aware that the Compendium was edited at regular intervals, but Hartman identifies some of the historians—including Li Tao (1115–1184), the author of The Long Draft Continuation of the Comprehensive Mirror—who performed the editing. He also explains the nature of their interventions. This chapter alone makes Hartman’s book indispensable reading for anyone working with Song-dynasty sources.Three periods mattered to thirteenth-century historians because they subscribed to a moral narrative of Song dynasty history in which Song Taizu played an important role in creating the dynasty’s institutions (although little information about his reign survives). The second period, the rule of Emperor Renzong between 1041 and 1063, marked the high point of Song governance. In the third period, the machinations of various nefarious ministers (jianchen), starting with Wang Anshi (1021–1086) in the 1060s, led to the dynasty’s ultimate decline (244).By identifying these three tropes, Hartman alerts readers to the topics on which extant sources are least reliable. The question arises, What will the new version of Song history—the one that reads surviving sources against the moral narrative that Hartman has identified—be like? We are very fortunate that Hartman’s next book, entitled Structures of Governance in Song Dynasty China, 960–1279 CE, is already completed and should appear soon.