Working out the manuscript affinities from a collation

Yesterday I finally finished collating the 4 editions and a selected 12 manuscripts of John the Deacon’s Life of St Nicholas.  This gives me a Word .docx file with every line of the text, the collation beneath it, and my translation under that.  In the left margin, it gives me a list of significant-looking variants:

I’ve had to recollate the early chapters, because I got better at this as I went on, and the earlier stuff needed to be redone, extra manuscripts added etc.

The text still contains a lot of working notes.  I have already found that it is a mistake to remove these too early.  Keep them to the last, and then remove them all as a specific activity, rather than along the way.

But then the question arises: how do I analyse this data in order to get a stemma out of it?  It’s too big, and I can’t get my head around it.

After some thought, I decided to create an Excel spreadsheet and process the supposedly significant variants into it.  This morning I did so.  I found that this required some intervention.  Actually I had to “simplify” some of the variants as I put them in.  Because unique variants are most likely errors, or mistakes, of no special meaning.  It’s the stuff in common that you need.  So where 3 manuscripts have “meritis” and the 4th has “et meritis”, and the 5th was “procul”, I entered the first 4 all up as “meritis”.

I also ignored variants that were merely endings.  The truth is that all the ending variants probably arise from scribes misreading abbreviations.  There’s just so many!

I then put a column for each manuscript, and put them in.  In the end I only had 19 locations where the text gave clear divergence into families.  On each row I coloured one set of readings in red, and another set in black, just so I could see the groupings (because you just try skim-reading “vocitatur” and “vocaretur”!).  Where a manuscript didn’t have that part of the text, I indicated with hatching.

The result looked a bit like this, except that M was originally on the left and C on the right.

As soon as I did this, I could see the PQO group, and the BGD group, which I was aware of anyway. I drew the vertical black lines to separate the groups.

Then I did some rearranging.  M, which I had thought isolated, I moved to be with W.  C, which I sort of thought was related to O, was now obviously part of the PQO group, so I moved that.

All the same some things do not jump out.  I’d already found that G is actually a copy of B in the first 6 chapters, but then switches to a copy of D!    Indeed the layout on the page is identical.  But that does not jump out from that table.  I’m fairly sure that I can eliminate G.

So … have I learned much?  A bit more than I knew before, perhaps.  But clearly I have a long way to go.


Leave a Reply