“Imagine the scribe” has long been one of the tenets of the art and science called textual criticism. When looking at variations and other difficulties in the biblical text, it helps to think about the work of copying ancient manuscripts in order to figure out why they look the way they do and what the text might have looked like in earlier manuscripts. To what extent is this principle transferable to our thinking about the author/editor who was composing a biblical book, using a combination of existing sources and his own writing? A process of imagination like this might allow us to think about the final form of the text and the earlier sources at the same time. Practitioners of source-criticism have not often engaged in such a practice and when they have it has often been coupled with the placement of the editor/writer in a fairly precise historical context, with a specific historical agenda. Such features may have sharpened the focus of the interpreter, but they had a large chance of being wrong and, because the choice of historical and ideological context was often based upon information in the text to begin with, the arguments were often circular.
The flood story has been a text of great interest to source-critics, because the present text in Genesis 6-8 is incoherent, and it is a fairly easy process to untangle the J and P flood accounts and reconstruct two separate, coherent flood accounts. The divine instructions to Noah from the P source in 6:9-22 are famous. Noah is to take two of each animal onto an ark with three levels that is 300x50x30 cubits. The instructions from the J source in 7:1-5 are less well known, and they are often ignored because the conflicting animal counts, two of each unclean animal and seven pairs of each clean animal, create difficulties for those who want to view the Bible as a seamless whole from its very beginning.
One unresolved difficulty with the source division in the flood narrative is the lack of instructions about the ark itself in the J source. The only two choices concerning this problem are that there were no instructions in J or the author/editor omitted them. The latter is the choice of those who assume the two different flood accounts were close parallels in their components, and that the author/editor would have been troubled by the dissonance of two different sets of ark instructions. This choice has two major difficulties. First, in many places in the Torah the author/editor did not share our desire for consistency and coherence. The separation of the flood narrative into sources is primarily based upon the many examples of inconsistency that we can easily observe. Second, if we start imagining a process in which the author/editor could omit significant amounts of material from the earlier sources, then a reliable process of source division becomes nearly impossible. If the pieces of the J source that we have might be only a fraction of the source, then how can we reconstruct it in any reliable way? Further, if the author/editor could leave out a piece of J to avoid narrative incoherence then why would he not have left out or changed a lot more? For example, the conflicting number of days the flood lasted, 40 days in J and just over a year in P would have been easy to harmonize. The sources must have exerted some kind of control over the process.