Open stenskjaer opened 6 years ago
I'm not sure I understand when exactly should we use the cert attribute. So if there are some uncertainties with the reading, we add cert=high/medium/low, if there is no doubt as to what the reading is for, we don't add the attribute?
So far it's only a suggestion. We need to agree on whether we want to add it to the next version of the LombardPress specification.
But what I thought was to add it when you are unsure about the what actually is going on in a reading.
If you are unsure about what a given word is, we would normally use <unclear>
.
But what I am in doubt about here is whether the "quando" with "quia" above is actually to be interpreted as an example of a "correction-substitution". And hence the indication of certainty on the rdg
element in stead of just parts of it.
Make sense?
It does, yes. But if the certainty of a reading is high, why the need to express the attribute at all? Or does 'high' suggest a type of certainty that is not entirely certain? How do we compare a reading that has no cert attribute and one that has a cert=high value?
My first thought was that "high" is not the same as 100% sure. So if no cert
, then there is absolutely no doubt (to the editor) about the type of reading.
If you select cert='high'
, you are pretty sure, but maybe not 100% sure.
But we might also agree that "high" is the default, and that to not indicate any cert
is the same as indicating cert='high'
. It's up to us what we think would be most intuitive and clear.
Ah, I see. So the very presence of the cert attribute is indicative of a doubtful reading, but with its respective degrees. I think this could work.
Take this:
I have encoded this as follows:
But I am not entirely sure that the addition of the "quia" above the line actually indicates that the scribe wants to substitute it with the "quando". So what I am in doubt of is the whole reading, not mere the reading of one of the words (if I were in doubt about the reading of fx. "quando", I could wrap that in
<unclear>
.So my first idea would be to wrap the whole reading in
<unclear>
like so:But that gives a lot of possibly problematic conditions in the processing (every
<rdg>
element must be checked for a wrapped<unclear>
. So why not just use the@cert
attribute on the reading.I would suggest that
rdg
MAY have@cert
with one of the following three values:high
,medium
, orlow
. The default is none.This would make the following encoding viable:
I find this to be both clear, semantically clean and in accordance with the intention of the TEI Guidelines (http://www.tei-c.org/release/doc/tei-p5-doc/en/html/ref-rdg.html).
It would be possible to also use it on the
lem
with a slightly different semantics, as that would modify the certainty of the choice of established text rather than a reading.Any thoughts???
EDIT: Maybe I should add that I am using it currently in some of my editions, to try testing the idea.