seg type for syllogisms

[jeffreycwitt]

It would be a interesting feature to support a markup for syllogisms. It could be something like:

<seg type="syllogism" ana="#AAA1">
<s type="major">Major</s>
<s type="minor">Minor</s>
<s type="conclusion>Conclusion</s>
</seg>

This might be particularly useful in logical texts books, but I have instances in sentences commentaries (for instance see text of http://scta.info/resource/pgb1q6-thpoeg).

The difficult would be balancing this with competing encodings. Include these level of detail of encoding could cause conflict with recording apps, quotations, and citations. Though I still think it could be handled.

Perhaps the bigger question is whether supporting structures at this level, we are opening committing to supporting all kinds of other structures that exist at a similar level of granularity (for examples sentences, consequences, antecedents. etc.

If the goal is for LombardPress-Schema to become the gold standard for encoding scholastic texts, such that all these texts become interoperable, supporting structures like this seems desirable.

We could start by just supporting this structure and then work our way toward supporting other common structures as use cases emerge.

Adding this support, shouldn't be a breaking change, but just a new feature. At the very least it is something to think about for the 1.1.0 release.

[stenskjaer]

I definitely see the strength in this. But it also worries me that it will, as you mention, be very difficult for us to determine which types of such meta-analysis goes into the text and which don't. Especially because they may overlap in complex or unpredictable ways. I think resorting to references to different parts of a coherent structure in a situation where you may have two or more types of meta-analysis that intersect may make it very unwieldy to handle as an encoder and maybe also during processing.

On the other hand, I already want to (and actually do) something similar in annotating the overall structural elements in a text (rationes and section numbers).

I don't know if we should revive some of the thoughts about referring to text spans (like we have talked about in the context of source references). Imagine we could have an unambiguous system for referring to text spans either within the transcription file itself or in an separate file. In that way each transcription could have an arbitrary number of meta-analyses superimposed on it, without having to have any worries about them conflicting.

This of course has several problems:

1. How do you handle the fact that the texts are still often in a fluid state? If I decide to make a change in my transcription that results in a reordering of words that make all existing references to that passage invalid or incorrect.
2. How do we handle this material in processing? You still want the meta-analyses made available in some of the representations that you get out of the transcription, and how is that done in a good, robust way, when the analysis is separated from the transcription?

By separating these things you would also make it possible for any scholar to impose his own meta-analysis on any text. Applying or representing that analysis on an existing transcription would then just be a matter of combining the "base" transcription with the analysis file.

Would this then also make up for a flexible way for annotating materials too? I, Jeff and Luciana could each our own commentaries to Jeff's transcription, and in an interface the user would then be able to choose which analysis they choose to overlay the text...

I know this is getting a bit rambling now, but I'm pretty convinced that if the challenges of finding a good way of annotating spans, that could have some very interesting perspectives.

[stenskjaer]

Okay, I have poured a bit more thoughts into this and developed a short description of the concept if you keep the analysis in a separate file. See https://github.com/stenskjaer/notes/blob/master/text-spans/meta-analysis.md

I am prepared to see this concept become dismantled, so "ready, aim, shoot!"

[lucianacioca]

I'm not sure if my comment is still relevant, but here it is :) That's a very good idea, I'm in favour of it. Both syllogisms and the shorter versions, the consequentiae, could use this type of support. I am also thinking that there are introductory parts at the beginning of questions in the sentences commentaries which, as far as I've seen, tend to have the same structure, with the question, 3 arguments for, and an oppositum, and then the very quick summary of the articles. So, very nice!