One of my students attempted a replication of the Gould lake experiemnt
last summer. We reasoned that there were sufficient problems with flying
over water (disturbing polarization cues from lake reflectance;
turbulence air currents & bees drowning etc) which might explain Gould's
result withou entertaining advanced cognitive map type hypotheses.We
therefore repeated the experiment but chose to locate our food source in
a barren area (ploughed field) which remained so for the entire summer
and was regularly checked for presence of any nectar yieding plants. We
found no evidence that bees could not be easily recruited to such a site
through the dance language, despite it being a natural 'desert' and on
any cognitive map hypothesis be discounted by experienced bees as no
point in looking there.
 
As to your proposed experiment, I certainly doubt whether there would be
any semantic accommodation, the parameters of the dance are surely
genetically fixed and the bees cannot learn to adjust these. I do seem
to vaguely recall an old paper. perhaps by von Frisch, talking about
dialects in bees, and which describes precisely the situation you
envisage with two strains united and therefore miss communicating with
each other. Also as bees age, they also overestimate distance flown, and
therefore will indicate a point further away than the source is. This of
course may not matter, as younger recruits following the correct compass
direction, and searching for a particular scent, may happen on the
source anyway. The problem would be more likely for the old bee dancing
and followed by the younger recruit.
 
I would be most interested to have the refernce for your philosopher??
Hope this is of some help.
 
Peter Wright