how to mark up a stage direction that also introduces speaker (stage/speaker ambiguity)

[DanilSko]

Sometimes speaker is introduced in a full third-person sentence (first of the two lines):

Гиря вийшов із другої хати:
— Кінчайте, сестриці, бо вже ніч.

Technically this first line is a stage direction (it means 'Girya is entering from the other house'). But it also introduces the speaker (first word). It would be tempting to mark this up as

            <sp>
               <speaker>Гиря</speaker> <stage>вийшов із другої хати:</stage>
            <p>— Кінчайте, сестриці, бо вже ніч.</p>
            </sp>

And this is even allowed by TEI. However, if the speaker name is not on the first place in such a speaker-introducing stage direction, we run into a problem:

Засміявся Панько:
— Ну, це я шуткома... Вип'ємо, дядю Мусію! Вип'ємо, та розкажіть такого, щоб за пупа взяло!..

the first line here just says 'Laughed Pan'ko:', and the second line gives direct speech. But a similar markup in this case is impossible:

            <sp>
               <stage>Засміявся</stage> <speaker>Панько</speaker>:
            <p>— Ну, це я шуткома... Вип'ємо, дядю Мусію! Вип'ємо, та розкажіть такого, щоб за пупа взяло!..</p>
            </sp>

tag speaker is not allowed here after the stage. What's the right way? Marking up everything as stage, although there is a clear speaker name?

[cmil]

The TEI guidelines for the speaker tag describe it like this:

contains a specialized form of heading or label, giving the name of one or more speakers in a dramatic text or fragment

I would say in your examples we don't have that specialised label and therefore would just mark the stage direction without any speaker tag. The important thing would be to give the sp a who attribute.

The Hamlet example in https://tei-c.org/release/doc/tei-p5-doc/en/html/CO.html#CODR suggests that we could wrap character names in stage directions with a name tag.

[DanilSko]

Thanks, Carsten! Yes, you're right, so stage be it! And yes, putting a name tag in it inside the stage seems a nice option :)

[lehkost]

Hello from the side line! 👋 I support Carsten's solution: "The important thing would be to give the sp a who attribute." That's how I usually handled this case in all the corpora I helped building, incl. GerDraCor and RusDraCor.

[lucagiovannini7]

Joining the conversation, since I run almost every day into this kind of problems with my encoding of Baroque plays (whose stage directions are quite... "creative"). I like the solution of Carsten, which basically creates two categories:

1. Only the speaker name = <speaker></speaker>

<sp who="faust">
   <speaker>Faust</speaker>:
   <l>I'll burn my books!</l>
</sp>

2. Speaker name within a sentence = <stage></stage> with a <name> tag inside

<sp who="faust>
   <stage>Enter <name>Faust</name> with a distraught face, and says:</speaker>
   <l>I'll burn my books!</l>
</sp>

What to do, however, when the speaker name is followed just by a word, or a few words specifying the attitude of the speaker, or with a distinct performance-related meaning (e.g. "aside")?

Instinctively, I would mark them as <stage> after <speaker>:

3.
<sp who="faust>
   <speaker>Faust</speaker><stage> aside</stage>
   <l>I'll burn my books!</l>
</sp>

4.
<sp who="faust>
   <speaker>Faust</speaker><stage> looking at his room:</stage>
   <l>I'll burn my books!</l>
</sp>

On the other hand, full-blown sentences (with finite-tense verbs?) should be all marked as <stage> into an <sp>, like in example 2.

What do you think? Is there any formal criteria we can establish to distinguish between these situations?

[DanilSko]

@cmil @lehkost I think Luca has a point. There are more nuanced cases than mine (see Luca's examples above), so the question is still open. There is a grey zone between examples like

Faust (aside)
Faust, aside
Enter Faust, aside
Faust speaks aside
Enter Faust, speaking aside 
...

and so on — a continuous scale 📈 of epification, from mere technical adverbs to wider adverbial clauses to full predicative sentences.

reopening the issue now

[lehkost]

Since the start of the project in 2014, we always tried not to fall into the annotation trap, since we didn't want to provide digital editions in the narrow sense, but research-prone corpora. 😊 So historically, the important thing was to assign proper who IDs to speech acts. This also enables us to add annotation layers later and revise our schema, which we've done in many cases. If a speech is buried in a longer stage direction, we even might have missed it (on purpose), but after handling hundreds of plays myself I can say that these cases are rare. None of the above cases are problematic in this sense, I think.