You should really look at Rebecca Saunders's work on southeastern
pottery.  She was dealing with stamped designs (frequently over-stamped
designs) and developed a methodology for breaking down both the designs and
various physical attributes of the pottery into basic components.  Then she
used cluster analysis to determine whether attributes separated out into
spatial clusters at her site.
 
        She originally published her work in Ceramic Notes through the
Florida Museum of Natural History; you would need to find a library copy,
because that number is no longer in print.  Or refer to her master's
thesis.  It seems to me that you could adapt her methodology for use on
incised designs fairly easily.  Her thesis, although now some years old,
was also considered to be one of the best summaries in archaeology
concerning design analysis and use of cluster analysis in ceramic research.
It got Prudence Rice's approval, which is not an easy thing to do.
 
James Cusick
 
 
james cusick
[log in to unmask]