Consider the following (from FDT's Henry V):
La main, de hand. Les doigts, le fingres.
Je pense que je suis le bon écolier. J’ai gagné deux
mots d’anglais vitement. Comment appelez-vous “les
ongles”?
(http://www.folgerdigitaltexts.org/?chapter=5&play=H5&loc=ftln-1336)
Ideally we would enclose all foreign words in the <foreign> tag (in this example we consider "de hand" and "le fingres" to be English/non-foreign). We would like to enclose "des ongles" in <q> tags. If <foreign> can contain <soCalled> or <mentioned>, what is the case against being able to do so for <q>?
We find there to be a slippery slope between <soCalled> and <mentioned>, so we prefer to use a generic <q> for all quotes throughout the project. For example, in Hermia's line below, is "little" in quotes because it was mentioned by Helena, or because Hermia is so-called little?
“Little” again? Nothing but “low” and “little”?
(http://www.folgerdigitaltexts.org/?chapter=5&play=MND&loc=line-3.2.343)
There are two reasons why <q> is not permitted within <foreign>. Firstly <q> is an "inter" level element, which can appear between paragraphs as well as within them. Allowing it within <foreign> would therefor be quite a major change to its content -- it would have to allow paragraphs and lists as well. Secondly, <foreign>-ness is regarded as subsidiary to <q>-ness. This is not so arbitrary as it sounds: it's quite likely that a
<q>
will contain both foreign and non foreign material, whereas the reverse seems less probable. Also, foreignness can most economically be indicated by an attribute.I would tag the speech you quote as follows
~~~~~~~~
<p xml:lang="fr">La main, <mentioned xml:lang="en">de hand</mentioned>. Les doigts, <mentioned xml:lang="en">le fingres</mentioned>. Je pense que je suis le bon écolier. J’ai gagné deux mots d’anglais vitement. Comment appelez-vous <mentioned rend="quoted">les ongles</mentioned>?
~~~~~~~~~
[Tho this doesn't of course indicate that what is presented as "French" here wouldnt go down so well in the hexagon ("vitement", forsooth)]
Which brings us to the slippery slope of <mentioned>. You're quite right in thus characterising it, but you're not required to make this subjective and perhaps over subtle decision. You can simply use <q> for anything wrapped in quote marks -- effectively as a synonym for <hi rend="quoted"> if you like : many people do.
Last edit: Lou Burnard 2013-03-13
I am not sure why this ticket has been opened again. Assigning to Sebastian to explain himself.
Michael asks about this today, and I see no response from Folger, so re-opening this to allow comment.
The release notes for 2.4.0 say:
A number of changes to loosen content models, mostly driven by experience of TCP EEBO project:
allow <foreign> to contain <q> (per FR 442)
But I don't see the change in the guidelines.
that was a mistake in the release notes, I am afraid. a misreading
of the output of this ticket.
Apologies, that seems to have been in error. What now is the status of this ticket. Changing from GREEN to AMBER.
Discussed by Council 11/13. Concluded that
<foreign>
is not intended to be used as structural wrapper element like this, and that LB's reformulation should be acceptableLast edit: Sebastian Rahtz 2013-11-11
This case is exceptional, and there are workarounds, so it's not worth fighting over. Consider this response merely for the record.
Often, the foreign phrase is a subset of the speech. For example, Lafew in All's Well That Ends Well:
And Parolles:
Here we need a container, probably <foreign>.
Applying tags consistently is one of our aims. In my mind, someone who wants to search for foreign words/phrases should not have to search in two places: foreign tags and xml:lang attributes. Now, if we can apply the xml:lang attribute to all <foreign> elements, we don't have this problem.
But, again from All's Well:
and a few lines later:
In the first example, you have a foreign phrase that comprises the whole speech. According to LB, we should use xml:lang in the <sp> element. In the second example, we need to use the <foreign> tag. Why not use the <foreign> tag for both? Besides, this is a made-up foreign language. When the language cannot be identified, or when it is a corruption of a known language, how do we apply the xml:lang attribute?
I think we felt that
<foreign>
is a short-hand for when you just want to quickly marksomething in an unstructured way, a bit like
<hi rend="foreign">
. Using xml:lang is more generic and extensible. So we suggest using @xml:lang everywhere, on<seg>
if its not distinguished in any other way, and don't use<foreign>
at all in your context.there are provisions in xml:lang for unidentified languages, made up languages, and private variants
Last edit: Sebastian Rahtz 2013-11-15
You most certainly can use xml:lang on the <foreign> element. In fact, that's what the Guidelines recommend, if you have no other tag available to carry the attribute. You can also (though the Guidelines don't say this) use <foreign> with the xml:lang attribute redundantly within an element the entirety of whose content is in a foreign language. So I am afraid I am still failing to understand what the problem is here.
The problem came when we had a foreign language we couldn't identify and didn't know how to apply the xml:lang attribute. If there are provisions for unidentified and made up languages, we'll look into that.
The original problem was that we can't use the <foreign> tag redundantly when the content contains a <q> tag. But, like I said, it's a minor issue that rarely occurs for us.