Le 29 oct. 2018 à 12:17, Matthew James Driscoll <[log in to unmask]> a écrit :
If the contents are mostly handwritten I would certainly catalogue the entire item as a manuscript – and have done so, such hybrid items not being unknown among the Icelandic stuff I work with. It’s more questionable when the contents are mostly printed, for example when a handwritten leaf has been added to replace a printed leaf which has gone missing (also common enough). Then I would be tempted to treat it as a printed item, but again it would depend on the proportion of printed to handwritten material.
The library of the Zograph Monastery (Mt. Athos) contains convolutes in which manuscripts and imprints are bound together. If these books contained only manuscript material, the parent would be <msDesc>, the cataloguing metadata would fit comfortably into the <msIdentifier> child of <msDesc>, and each of the components of the convolute would be an <msPart> child of <msDesc>. The complication here is that the book is neither manuscript nor imprint, and it incorporates content items of both types.
Calling an imprint an <msPart> seems wrong because it isn’t a manuscript, and, for that matter, this detail makes me anxious about whether the parent should be <msDesc>, since what we are describing, as the catalogued item in the monastery library is not strictly (that is, entirely) a manuscript. Yet <msDesc> and <msPart> seem like the obvious choice if we close our eyes to the incorporation of imprint material into the convolute, and <msIdentifier> seems to be a property of the book as a whole—except, of course, for the fact that it isn’t properly a manuscript.
So: How might we proceed with eyes wide open?