Résumé :
The paper presents the results of a study on the transmission of the
Etymologicum Gudianum. The original manuscript of this lexicon, Vat.
Barb. gr. 70 (d), is still preserved, but a study of its apographs, determining
their stemmatic relationship, is nonetheless necessary for
the constitutio textus, since many folios of d are missing and its text
is in some parts illegible. In spite of this, such an investigation into
the transmission of the lexicon has been hitherto largely neglected.
Moreover, the study of the apographs can also shed light on the role
that the Etymologicum Gudianum played in the Byzantine scholarship
and culture. The paper deals with the manuscripts of the first family
of the lexicon. They all derive from a common ancestor (α) that was
a direct and faithful copy of d. The manuscript Vat. gr. 1708 ( j), the
hyparchetype of the third family, actually belongs to the first family,
too. The third family, then, is not independent, but it only constitutes
a sub-group of the first one. The transmission of the first family is
divided in two branches: Par. gr. 2630 (a) and its direct copy Vind. phil.
gr. 23 (c) represent one branch, while Par. gr. 2631 (b) and j derive
from a common ancestor (β) and represent a second branch. The role
of the manuscript Petrop. gr. 114 + Sinait. gr. 1201 + Sinait. ΝΕ gr. Μ 186
(e) and of its copy Neap. II D 37 (f) is more difficult to assess. Thanks
to significative errors and to other features, such as the order of the
entries and the presence of interpolations, it is possible to determine
that e (and f) belong to the same branch as a (and c). Most probably, e
is a copy of c; however, the text of e (or, rather, of a lost model of it to
be posited between c and e) has been revised and corrected on the basis
of another manuscript of the Etymologicum Gudianum that did not
belong to the first family. All manuscripts discussed were produced in
Terra d’Otranto in the 12th and 13th centuries, with the only exception
of f (mid-14th century).