The unearthing of manuscript texts, especially in the last fifty years, has revolutionized the study of early Chinese civilization. Manuscripts, which bear previously unknown texts and unknown forms of known texts, have greatly destabilized our view of the textual canons that have defined the landscape of genre for two millennia. One preliminary way of understanding the precursors of canonical texts is as freely circulating zhang 章 (pericopes; chapters) that became disposed in compendia at random. Although this may partially account for forces at work in text formation, pericopes can also become appended with paratextual markers that guide their interpretation, categorization, and compilation. This dissertation seeks to shed light on these proto-commentarial features, and reconsider how texts and genres accreted and decayed in Chinese manuscript culture. To this end, I examine an array that includes known transmitted texts as well as manuscript texts of previously unknown structure and form. Controversies surround the genre identity of the manuscript texts, while the transmitted texts have been compiled separately in compendia associated with distinct genres. The dissertation’s series of experiments maintains controls of narrative and form: first, all the texts (as contextualized by commentary and paratext) function as sources for the legend of Rui Liangfu 芮良夫, a ninth century BCE noble who spoke out against the government of his contemporary King Li of Zhou 周厲王 (r. 853-842 BCE); second, all the texts, while differing in prosody and form, are written in tetrasyllabic verse. Following a general introduction to the dissertation’s problems and methods in chapter one, chapter two emerges from a philological study and translation of the *Rui Liangfu bi, a Warring States bamboo manuscript in the Tsinghua University corpus. Proceeding from disagreements about whether the text is a shu document or shi poem, I argue that the text’s content and form violates conventions of these genres, at least as these genres are known in transmission. Instead, I identify the texts as “verse albums” on the basis of text-paratext structures in several other manuscripts and one transmitted text. Verse albums function as micro-collections of thematically related literature, and come marked with interpretive paratext whose function is similar to prefaces and synopses found in transmitted canons. Chapter three examines the roots of a historicizing hermeneutic in the Shijing (Confucian Canon of Poetry), in part through the poem Sang rou (Mulberry Shoots) attributed to Rui Liangfu. Here, as with verse albums, I chart the traces of intentionality and proto-commentarial markers that seek to make poems into texts with fixed historical meanings, thus functioning like the narrative paratext found in verse albums of chapter two and the commentarial layers of later canons. Examining the conventions of suites of poems, and the tendency of historical poems to form thematically organized suites, I propose a model by which a historicizing hermeneutic might have spread chronologically and laterally, becoming pervasive in the Mao Shi. The fourth chapter is based on a philological reconstruction and translation of the Yi Zhou shu’s “Rui Liangfu” chapter. The critical translation employs received editions and a version of the chapter preserved only in Japanese manuscripts of the Qunshu zhiyao, a Tang-era encyclopedic compendium lost in China for centuries. In this study, I demonstrate that both versions of the text derive from similar editions; significant lacunae and errors in the two texts show that much of the damage arose in imperial times. Much of this is due to the delegitimization of the Yi Zhou shu, which has long been viewed as the leftovers that remained when Confucius redacted the Shang shu (Revered Documents) canon. While chapters two and three show how paratextual markers might aid a text’s organization and preservation in a collection, chapter four shows how the removal of an editor-figure can induce a text’s decay. The fifth and final chapter examines intertextual relations among the sources studied in chapters two through four. The texts share a great familiarity with a common legend of Rui Liangfu, yet they share almost no word-for-word text with one another. I propose that at least one of the texts results from historical confabulation—a broad and continuous effort to fill historical lacunae that were revealed by a new, systematizing discourse on the cultural forms (and genres) appropriate to remonstrance. I consider also the possibility that the reconstruction or alleged “forgery” of shu documents literature in early and medieval China stems from a similar historiographic impulse.