For once Medlands is right, sort of... Accidents happen. Maybe the
malignant jackass in the White House will stumble into one soon and do
right by the US people for a change.
The erroneous identification of Ermesenda as the daughter of Pierre of
Melgueil married in the 1080s to Guillaume V appears to be a misplaced
Occam's-razor assumption, based on a primary source (the dispute
resolution I mentioned before). There are no doubt many other similar
mistakes embedded in the woodwork of medieval genealogy.
The cartularies of Montpellier and Maguelone are not sources for these
curious misidentifications - both ladies occur in charters only by given
name as the mothers of their respective sons.
A solution to the cousin-marriage scenario for Guilelma the daughter of
Ermesenda with Bernard of Melgueil was canvassed in *Histoire générale
de Languedoc*, but unfortunately with the result of introducing new errors.
The only rational solution I can see is that Ermengarde and Ermesenda
were not from the Melgueil family at all, while Guillaume V of
Montpellier was married first to a daughter Pierre of Melgueil (given
name unknown) and then to Ermesenda who became his widow, the mother of
all his recorded children.
Peter Stewart