On several occasions the scribe of a thirteenth-century manuscript left gaps (marked here using square brackets) in the text they copied:

Mest alle men he yeueþ drynke. of one deofles [senthe]He schal him cunne schilde wel. yef he him [wole hi þenthe]

(Almost all men he gives drink of a devil's [draft],He must know to shield himself well if [he wishes to think] for himself.)¹

Such gaps stand out as a category of "error": the white spaces that they leave visually emphasize the moments when copying is disrupted. But, as Daniel Wakelin notes in his work on scribal correction, the intrusion of errors into our reading—particularly visually disruptive errors such as these—also serves to remind us that most of the time scribes copied faithfully. Errors reveal the care and attention that usually went into the writing process. It is only when scribes miswrote or chose not to write that the thoughtfulness underlying their copying practices is revealed. In particular, the decision to leave blank space rather than produce "bad" copy may be construed as an act of care. Through the study of errors—and the corrections that may follow them—comes a history of (largely successful) scribal copying.² Further, gaps like the ones above are not primarily the result of a failure to write. That is, the decision not to write is not born out of a lack of ability; rather, it results from a [End Page 381] disruption in the act of reading. This scribe, or a scribe further back in the text's transmission history, chose not to write because they found the words of their exemplar to be inscrutable, illegible, nonsensical, or simply missing.³ It was the inability to make sense of the exemplar that precipitated the decision to leave a gap.

Soon after this couplet from the early Middle English Poema Morale was copied, someone filled in the missing words (within the brackets).⁴ The added words were not composed ad hoc to make up for the discrepancy in rhyme and meter. Rather, the corrections align reasonably well with the textual tradition as it is found across other extant copies of the poem.⁵ What these corrections provide, then, is an example of the semi-traceable use of this copy in parallel with another, perhaps the original exemplar now deciphered, perhaps a further copy encountered in passing. Either way, the corrector must have read their exemplar successfully in order to make up for the defective text. Such gaps, together with the words that fill them, demonstrate the reflexivity of scribal work: reading is a necessity of textual transmission. One book generates another, as a copy cannot be made without an exemplar, written or spoken. Indeed, an established scribal exemplar may be one of the few manuscripts that we can confidently assert was read in its entirety. While this point may seem pedestrian, its implications are fundamental to understanding the process of copying as untapped evidence for the history of reading. However, what scribal reading practices consisted of is difficult to ascertain in most moments when reading occurred successfully. How does one detect the use of no-longer extant works, the only trace of which is a new copy of the text itself? Just as with scribal copying, perhaps it is only in moments of error or variation that the processes by which scribes read are revealed.

Since Bernard Cerquiglini identified the positive implications of variance—its potential to create "joyful excess" in the transmission of [End Page 382] texts—scholarship has been attentive to the ways variation reveals scribal practice and belies lines of textual transmission.⁶ The ability of scribes to take a text and make it their own, along with the "intellectual and even creative engagement" that this represents, has been brought to the fore.⁷ However, just as each variant reading of a text can be used as evidence for the scribal act of writing, putting pen to paper with the intention of correcting, altering—perhaps minutely—or transferring truly, then it is also evidence of the scribal act of reading that...