Very General Hypothesis:
There are more stages of transmission in Anglo-Saxon manuscripts (i.e., between original source and the manuscript we actually possess) than are generally recognized.
Nearly every time I start to work with a text that is not among those most familiar to me, I discover that there is evidence for multiple stages of transmission.
Paul Remley shows that there could be as many as eight stages between the biblical book of Daniel, two Latin canticles (the Oratio Azariae and the Canticum Treum Puerum), and the manuscript texts of Daniel and Azarias.
I'm pretty certain that the OE translation of Chrodegang's Rule was done in the earlier 10th century, between 940 and 950 and then copied again between 984-1006 (when the names from the Old Minster at Winchester were added), before being copied in the 11th century to create the manuscript version we now have.
The Lexomics Research Group has come across some evidence that strongly suggests more than one stage of textual transmission behind a few Exeter Book poems, including Christ III, Guthlac A, and the Descent into Hell.
Beowulf is consistent with being at least a copy of a copy: Lapidge's argument for the exemplar being in a different hand, the problems the scribes have with proper names, the apparent haplography in a few lines, and now some work by the Lexomics group strongly suggests a rather complex textual history.
I think that many scholars (I certainly was one of them) have unconsciously adopted the idea that the earlier textual history of Anglo-Saxon texts was like their late textual history: texts sat un-read and un-revised in a monastic library for centuries, only to be copied over in the 10th century, to then be neglected after the Conquest and sit un-read until, after surviving the cataclysms of the Dissolution, the English civil war and the Cotton Fire, they are finally re-discovered in the 18th century.
Also, because our ability to locate texts in time and space is very limited (our resolution is extremely coarse grained), we unconsciously adopt the position that texts might be handled, copied or read once in a century or so.
Evidence scattered throughout the corpus suggests instead that there was a rich and complex history of textual transmission, modification, editing and augmentation throughout the Anglo-Saxon period. If this is so, then texts may have been very much more "alive" and in use than we have previously thought. They may have been less of a quiet archive than a regularly consulted set of intellectual resources (cf. the heavily glossed texts that are the product of Athelwold's and Dunstan's "Aldhelm Seminar" at Glastonbury).
The texts we have may not represent only a single temporal slice of a culture but instead be multiple re-workings of inherited material, and our authors may have been widely read in their vernacular literature as well as in the Latin tradition.