We've been through all this before, it seems to me, and more than once.
There can be few distinctions more fundamental to textual scholarship than
that between what is actually physically present in the text and what isn't.
A scribe deleting a word and writing another in the margin is definitely not
the same thing as the editor of a text deciding what the scribe ought to
have written but didn't and correcting/emending it. And the TEI provides
elements for both these situations: one uses <add> and <del> (each with a
@hand attribute) for scribal emendations, and <corr> and <supplied> (each
with a @resp attribute) for editorial emendations. The ambiguity of the TEI
"message", to which Michael draws attention, is presumably due to the fact
that in the early heroic days (up to and including P3) the Guidelines did
suggest that one could use <corr> for scribal emendations if one so wished.
This was changed in P4, but obviously not completely enough.
M. J. Driscoll
Fra: TEI (Text Encoding Initiative) public discussion list
[mailto:[log in to unmask]] Pċ vegne af Michael Beddow
Sendt: 26 January 2007 14:00
Til: [log in to unmask]
Emne: Re: scribe / editoral corrections
Elena Pierazzo wrote:
> I keep thinking about the existence of an ontological difference between
> a correction made by a scribe in a manuscript and a correction made by a
> modern editor.
I prefer to keep ontology locked away until I really want to frighten
people, so I'd just call it a plain ordinary difference. But it's a real and
> At present we have just one element for encoding it (<corr> normally
> used together with <orig>) with the possibility of distinguishing who is
> responsible for the correction by the help of an attribute (@resp).
If "we" means the authors of the Guidelines and those who receive their
collective wisdom, then those authors, or at least the authorial persona
that emerges from the merging of disparate strands of work over the years,
speak with forked tongue.
The short description of <corr> reads:
"contains the correct form of a passage apparently erroneous in the copy
where in the case of an original digital edition of a MS, "the copy text" is
everything that the scribe(s) have placed in the witness. This would seem
plainly enough to confine its use to encoding editorial interventions
(though if the copy text is itself a scholarly edition, then this definition
of corr allows for the encoding of such interventions as were made by the
original editor and are hence not error's in the *encoder's* copy text).
That's probably why it can carry a resp attribute. The general idea of resp
is that anything that could embody an interpretation or a decision rather
than an uncontentious transcription needs labelling with a potential
can-carrier. The scribe isn't really a candidate for that role at all.
If we then go to the body of the Guidelines themselves, at 6.4.1. Apparent
Errors (to which the summary points us) we still seem to find the same idea,
with the definition of corr I just cited repeated verbatim.
But read on a bit further, and we find a passage introduced, no doubt with
involuntary humour, by the phrase: "The following discussion may be helpful
when making this decision". The first allegedly helpful notion is that
"where the correction is an addition by a scribe or author in a manuscript
or other primary source (typescript, proof or galley, etc.) then either
<corr> (or <sic>) or <add> might be appropriate, depending on the
But just a minute: this is the first time such a scribal correction or
addition has been mentioned. And in fact, such corrections or additions as
have been so far mentioned have been explicitly *not* scribal. No wonder
then that this "helpful" portion is decked out with a phrase typical of the
sort that makes digital publication of the Guidelines a major contribution
to Health and Safety, since it reduces the risk of collateral damage caused
chucking a weighty tome across the room in frustration: " might be
appropriate, depending on the circumstances".
Thus encouraged, we "might" go to 18.1.3. Correction and Conjecture. And lo,
we find ourselves transported to the land of Something Completely Different.
"The <sic>, <corr>, and <choice> elements, defined in the Common Core module
may be used to register authorial or scribal corrections within a witness."
So: we have two apparently clear statements about the use of corr and
friends, at two different places in the Guidelines, each specifying a
completely distinct use of the elements with no mention on either occasion
that there is an alternative use. And only one of those statements is
reflected in the element summary. This is the TEI equivalent of having a
cake and eating it. A genuine problem is obfuscated by splitting two
essentially distinct and largely incompatible understandings of the same
thing into separate passages, neither of which so much as alludes to the
other, and adverting to the duality only via a single sentence that fobs off
the decision-making it purports to assist by referring to "circumstances".
This sort of thing be a real problem for those who aspire to be True
Believers in the TEI message. Because, all too often, there ain't one.
Though there may sometimes be two, or even more. Those of us who, by
contrast, see the Guidelines mainly as a repository of invaluable but not
always consistent (or even coherent) reflections on how to encode Humanities
texts which inform what must in the end remain our own decisions (reflected
in due course in properly performed and documented customisations or
extensions), find it somewhat less disturbing