encoding complex sources in an apparatus

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

encoding complex sources in an apparatus

ron.vandenbranden
Administrator

Hi,

We're starting an edition project that is building on an existing digital edition in which a number of published versions of a literary work had been encoded using the parallel-segmentation method. Apart from updating the encoding from P4 to P5, a new phasse is planned in which intermediary text versions, such as manuscript versions, proofs, etc. will be incorporated in the edition. As it goes with hand-edited texts, these intermediary text versions are complex sources in themselves. Often, different correction layers can be distinguished, which generally don't allow to make any claims about their internal chronology. I am looking for a way to integrate the encoding of such complex sources in an apparatus, and would like to check if my analysis makes sense.

Suppose we have following text versions:

  Ed1: the first edition
  P2: a proof for the second edition (based on Ed1, with manual corrections)
    P2a: corrections in red pencil
    P2b: corrections in black pencil
  Ed2: the second editon
If version P2 would be included in the digital edition, it seems that it would only make sense to take into account the sum of its correction layers (P2a and P2b). Still, the editor wants to distinguish between the different annotation layers. For this fine-grained distinction, I'm considering the use of the @change attribute to link additions and deletions to the identified revision rounds.

Perhaps an example fragment could clarify things:
  Ed1   The weather was fine.
  P2    The <add change="#P2a">fall</add> <restore change="#P2b"><del>weather</del></restore> was <add change="#P2a">cool, but</add> <subst change="#P2b"><del>fine</del><add>nice</add></subst>.
  Ed2   The September weather was cool, but nice.

If these would be combined into a parallel-segmented apparatus, that could look as follows:
  The
    <app>
      <rdg wit="#Ed1"/>
      <rdg wit="#P2"><add change="#P2a">fall</add></rdg>
      <rdg wit="#Ed2">September</rdg>
    </app>
    <app type="pseudo">
      <rdg wit="#Ed1 #Ed2">weather</rdg>
      <rdg wit="#P2"><restore change="#P2b"><del>weather</del></restore></rdg>
    </app>
  was
    <app>
      <rdg wit="#Ed1"/>
      <rdgGrp type="pseudo">
        <rdg wit="#P2"><add change="#P2a">cool, but</add></rdg>
        <rdg wit="#Ed2">cool, but</rdg>
      </rdgGrp>
    </app>
    <app>
      <rdg wit="#Ed1">fine</rdg>
      <rdgGrp type="pseudo">
        <rdg wit="#P2"><subst change="#P2b"><del>fine</del><add>nice</add></subst></rdg>
        <rdg wit="#Ed2">nice</rdg>
      </rdgGrp>
    </app>.
In this example, readings that actually don't differ are grouped as "pseudo" variants (there are better labels, no doubt) so they can be ignored when they're not relevant to the selected comparison set in the edition. Yet, when one text version is studied in detail, an option could be offered to visually distinguish between the different correction layers.

To complete this idea, I would define the revision rounds in <listChange> elements in the <profileDesc> section in the header, linked to the witness definitions they apply to with a @corresp attribute:
  <teiHeader>
    <fileDesc>
      <sourceDesc>
        <listWit>
          <witness xml:id="Ed1">the first edition</witness>
          <witness xml:id="P2">a proof for the second edition (based on Ed1, with manual corrections)</witness>
          <witness xml:id="Ed2">the second editon</witness>
        </listWit>
      </sourceDesc>
    </fileDesc>
    <profileDesc>
      <creation>
        <listChange>
          <listChange corresp="#P2">
            <change xml:id="P2a">corrections in red pencil</change>
            <change xml:id="P2b">corrections in black pencil</change>
          </listChange>
        </listChange>
      </creation>
    </profileDesc>
  </teiHeader>
Does this all make sense, am I overlooking something, just stating the obvious, or has this been done better before? Any advice welcome!

Kind regards,

Ron

-- 
Ron Van den Branden
CTB   - Centrum voor Teksteditie en Bronnenstudie / 
        Centre for Scholarly Editing and Document Studies
KANTL - Koninklijke Academie voor Nederlandse Taal- en Letterkunde / 
        Royal Academy of Dutch Language and Literature
Koningstraat 18 b-9000 Gent Belgium
E-mail : [hidden email]
Reply | Threaded
Open this post in threaded view
|

Re: encoding complex sources in an apparatus

ANTONIO ROJAS CASTRO
Hi there,

Yes, it makes sense to me. However, I would use @hand and @rend instead of @change. I would describe each hand in a <handNote> in the header. I might be wrong but when I see @change I expect a numeric value such as a phase (1, 2, etc.).

Best,



2016-05-26 15:33 GMT+02:00 ron.vandenbranden <[hidden email]>:

Hi,

We're starting an edition project that is building on an existing digital edition in which a number of published versions of a literary work had been encoded using the parallel-segmentation method. Apart from updating the encoding from P4 to P5, a new phasse is planned in which intermediary text versions, such as manuscript versions, proofs, etc. will be incorporated in the edition. As it goes with hand-edited texts, these intermediary text versions are complex sources in themselves. Often, different correction layers can be distinguished, which generally don't allow to make any claims about their internal chronology. I am looking for a way to integrate the encoding of such complex sources in an apparatus, and would like to check if my analysis makes sense.

Suppose we have following text versions:

  Ed1: the first edition
  P2: a proof for the second edition (based on Ed1, with manual corrections)
    P2a: corrections in red pencil
    P2b: corrections in black pencil
  Ed2: the second editon
If version P2 would be included in the digital edition, it seems that it would only make sense to take into account the sum of its correction layers (P2a and P2b). Still, the editor wants to distinguish between the different annotation layers. For this fine-grained distinction, I'm considering the use of the @change attribute to link additions and deletions to the identified revision rounds.

Perhaps an example fragment could clarify things:
  Ed1   The weather was fine.
  P2    The <add change="#P2a">fall</add> <restore change="#P2b"><del>weather</del></restore> was <add change="#P2a">cool, but</add> <subst change="#P2b"><del>fine</del><add>nice</add></subst>.
  Ed2   The September weather was cool, but nice.

If these would be combined into a parallel-segmented apparatus, that could look as follows:
  The
    <app>
      <rdg wit="#Ed1"/>
      <rdg wit="#P2"><add change="#P2a">fall</add></rdg>
      <rdg wit="#Ed2">September</rdg>
    </app>
    <app type="pseudo">
      <rdg wit="#Ed1 #Ed2">weather</rdg>
      <rdg wit="#P2"><restore change="#P2b"><del>weather</del></restore></rdg>
    </app>
  was
    <app>
      <rdg wit="#Ed1"/>
      <rdgGrp type="pseudo">
        <rdg wit="#P2"><add change="#P2a">cool, but</add></rdg>
        <rdg wit="#Ed2">cool, but</rdg>
      </rdgGrp>
    </app>
    <app>
      <rdg wit="#Ed1">fine</rdg>
      <rdgGrp type="pseudo">
        <rdg wit="#P2"><subst change="#P2b"><del>fine</del><add>nice</add></subst></rdg>
        <rdg wit="#Ed2">nice</rdg>
      </rdgGrp>
    </app>.
In this example, readings that actually don't differ are grouped as "pseudo" variants (there are better labels, no doubt) so they can be ignored when they're not relevant to the selected comparison set in the edition. Yet, when one text version is studied in detail, an option could be offered to visually distinguish between the different correction layers.

To complete this idea, I would define the revision rounds in <listChange> elements in the <profileDesc> section in the header, linked to the witness definitions they apply to with a @corresp attribute:
  <teiHeader>
    <fileDesc>
      <sourceDesc>
        <listWit>
          <witness xml:id="Ed1">the first edition</witness>
          <witness xml:id="P2">a proof for the second edition (based on Ed1, with manual corrections)</witness>
          <witness xml:id="Ed2">the second editon</witness>
        </listWit>
      </sourceDesc>
    </fileDesc>
    <profileDesc>
      <creation>
        <listChange>
          <listChange corresp="#P2">
            <change xml:id="P2a">corrections in red pencil</change>
            <change xml:id="P2b">corrections in black pencil</change>
          </listChange>
        </listChange>
      </creation>
    </profileDesc>
  </teiHeader>
Does this all make sense, am I overlooking something, just stating the obvious, or has this been done better before? Any advice welcome!

Kind regards,

Ron

-- 
Ron Van den Branden
CTB   - Centrum voor Teksteditie en Bronnenstudie / 
        Centre for Scholarly Editing and Document Studies
KANTL - Koninklijke Academie voor Nederlandse Taal- en Letterkunde / 
        Royal Academy of Dutch Language and Literature
Koningstraat 18 b-9000 Gent Belgium
E-mail : [hidden email]



--
Reply | Threaded
Open this post in threaded view
|

Re: encoding complex sources in an apparatus

ron.vandenbranden
Administrator

Hi Antonio,

On 26/05/2016 16:23, Antonio Rojas Castro wrote:

Yes, it makes sense to me.


Ok, that's reassuring, thanks.

However, I would use @hand and @rend instead of @change. I would describe each hand in a <handNote> in the header. I might be wrong but when I see @change I expect a numeric value such as a phase (1, 2, etc.).



I've hesitated about <handNote> as well, but I think the description as <change> allows for a more exact identification of correction layers in specific versions of a text. When seeing an addition in the encoded text, I'd be more interested to know when it had been added, rather than knowing just the details of the hand. In fact, I think both can be combined, by defining the different hands in <handNote> elements, and associating the <change> elements with the correct hands via the @who attribute:
  <teiHeader>
    <fileDesc>
      <!-- ... -->
      <sourceDesc>
        <listWit>
          <witness xml:id="M1">draft manuscript</witness>
          <witness xml:id="Ed1">the first edition</witness>
          <witness xml:id="P2">a proof for the second edition (based on Ed1, with manual corrections)</witness>
          <witness xml:id="Ed2">the second editon</witness>
        </listWit>
      </sourceDesc>
    </fileDesc>
    <profileDesc>
      <handNotes>
        <handNote xml:id="au-r" medium="red-pencil" scribe="author">author's annotations in red pencil</handNote>
        <handNote xml:id="au-b" medium="black-pencil" scribe="author">author's annotations in black pencil</handNote>
        <handNote xml:id="corr" medium="red-ink" scribe="corrector">corrector's annotations in red ink</handNote>
      </handNotes>
      <creation>
        <listChange>
          <listChange corresp="#M1">
            <change xml:id="M1a" who="#au-r">authorial correction layer in red pencil</change>
            <change xml:id="M1b" who="#au-b">authorial correction layer in black pencil</change>
          </listChange>
          <listChange corresp="#P2">
            <change xml:id="P2a" who="#au-b">authorial correction layer in blue ink</change>
            <change xml:id="P2b" who="#corr">correction layer by corrector in red ink</change>
          </listChange>
        </listChange>
      </creation>
    </profileDesc>
  </teiHeader>
This seems more expressive, since:
    -one hand can be associated with several correction layers in the history of a text. In the slightly expanded version above, revisions in distinct stages of a text (M1b and P2a) can be ascribed to one single authorial hand (au-r).
    -the fact that <listChange> elements can nest, allows for grouping related revision rounds, and linking them with the precise text witness in which they occur

Again, this is just my interpretation so I might be wrong. Thanks for pointing this out, anyway!

Best,

Ron
Reply | Threaded
Open this post in threaded view
|

Re: encoding complex sources in an apparatus

Lou Burnard-6
Hi Ron

Hi Ron

Thanks for this excellent example, possibly worth stealing for the
Guidelines. However, I'm a bit curious about the fact that none of your
<change> elements references the first edition (Ed1). Is that because
the assumption here is that we are transcribing everything but that
edition? Or is there some kind of default assumption that any text not
assigned to a specific change is derived from that edition? That
probably needs to be clarified somehow.

Lou


On 26/05/16 23:46, ron.vandenbranden wrote:

> Hi Antonio,
>
> On 26/05/2016 16:23, Antonio Rojas Castro wrote:
>>
>> Yes, it makes sense to me.
>
>
> Ok, that's reassuring, thanks.
>
>> However, I would use @hand and @rend instead of @change. I would
>> describe each hand in a <handNote> in the header. I might be wrong
>> but when I see @change I expect a numeric value such as a phase (1,
>> 2, etc.).
>>
>>
>
> I've hesitated about <handNote> as well, but I think the description
> as <change> allows for a more exact identification of correction
> layers in specific versions of a text. When seeing an addition in the
> encoded text, I'd be more interested to know when it had been added,
> rather than knowing just the details of the hand. In fact, I think
> both can be combined, by defining the different hands in <handNote>
> elements, and associating the <change> elements with the correct hands
> via the @who attribute:
>
>   <teiHeader>
>     <fileDesc>
>       <!-- ... -->
>       <sourceDesc>
>         <listWit>
>           <witness xml:id="M1">draft manuscript</witness>
>           <witness xml:id="Ed1">the first edition</witness>
>           <witness xml:id="P2">a proof for the second edition (based
> on Ed1, with manual corrections)</witness>
>           <witness xml:id="Ed2">the second editon</witness>
>         </listWit>
>       </sourceDesc>
>     </fileDesc>
>     <profileDesc>
>       <handNotes>
>         <handNote xml:id="au-r" medium="red-pencil"
> scribe="author">author's annotations in red pencil</handNote>
>         <handNote xml:id="au-b" medium="black-pencil"
> scribe="author">author's annotations in black pencil</handNote>
>         <handNote xml:id="corr" medium="red-ink"
> scribe="corrector">corrector's annotations in red ink</handNote>
>       </handNotes>
>       <creation>
>         <listChange>
>           <listChange corresp="#M1">
>             <change xml:id="M1a" who="#au-r">authorial correction
> layer in red pencil</change>
>             <change xml:id="M1b" who="#au-b">authorial correction
> layer in black pencil</change>
>           </listChange>
>           <listChange corresp="#P2">
>             <change xml:id="P2a" who="#au-b">authorial correction
> layer in blue ink</change>
>             <change xml:id="P2b" who="#corr">correction layer by
> corrector in red ink</change>
>           </listChange>
>         </listChange>
>       </creation>
>     </profileDesc>
>   </teiHeader>
>
> This seems more expressive, since:
>     -one hand can be associated with several correction layers in the
> history of a text. In the slightly expanded version above, revisions
> in distinct stages of a text (M1b and P2a) can be ascribed to one
> single authorial hand (au-r).
>     -the fact that <listChange> elements can nest, allows for grouping
> related revision rounds, and linking them with the precise text
> witness in which they occur
>
> Again, this is just my interpretation so I might be wrong. Thanks for
> pointing this out, anyway!
>
> Best,
>
> Ron
>
Reply | Threaded
Open this post in threaded view
|

Re: encoding complex sources in an apparatus

ron.vandenbranden
Administrator
Hi Lou,

The example was clearly made up; I'm still waiting for a detailed
overview of the scans for all text materials this early in the project.
If something useful comes out of this, and if a Dutch text could be of
use for the Guidelines, I'd gladly share this.

Nevertheless, the fact that in my example Ed1 and Ed2 are not featuring
in the lists of changes, is on purpose. Since those are published
editions of the text (with all traces of revisions edited out), they
don't contain any revisions themselves. They will differ from each
other, and from the other versions, but these differences will become
apparent in the apparatus. On the other hand, while the exact details of
a change in an intermediary version like P2 are not relevant to the
apparatus (which refers just to P2), they can still be recorded in the
@change attribute for the addition/deletion/... within the apparatus:

     <app type="pseudo">
       <rdg wit="#Ed1 #Ed2">weather</rdg>
       <rdg wit="#P2"><restore change="#P2b"><del>weather</del></restore></rdg>
     </app>

If the three text versions in this apparatus entry are compared, this
will produce no variants (since the restored deletion in P2b would just
distort an account of the actual variation). Yet, when version P2 is
rendered in isolation, that restored deletion would be rendered, though,
and probably distinguished from corrections of P2b that occur in P2.

Is this clarification sufficient (and valid), or should it be stated
explicitly in the header somewhere?

Best,

Ron


On 27/05/2016 10:46, Lou Burnard wrote:

> Hi Ron
>
> Hi Ron
>
> Thanks for this excellent example, possibly worth stealing for the
> Guidelines. However, I'm a bit curious about the fact that none of
> your <change> elements references the first edition (Ed1). Is that
> because the assumption here is that we are transcribing everything but
> that edition? Or is there some kind of default assumption that any
> text not assigned to a specific change is derived from that edition?
> That probably needs to be clarified somehow.
>
> Lou
>
>
> On 26/05/16 23:46, ron.vandenbranden wrote:
>> Hi Antonio,
>>
>> On 26/05/2016 16:23, Antonio Rojas Castro wrote:
>>>
>>> Yes, it makes sense to me.
>>
>>
>> Ok, that's reassuring, thanks.
>>
>>> However, I would use @hand and @rend instead of @change. I would
>>> describe each hand in a <handNote> in the header. I might be wrong
>>> but when I see @change I expect a numeric value such as a phase (1,
>>> 2, etc.).
>>>
>>>
>>
>> I've hesitated about <handNote> as well, but I think the description
>> as <change> allows for a more exact identification of correction
>> layers in specific versions of a text. When seeing an addition in the
>> encoded text, I'd be more interested to know when it had been added,
>> rather than knowing just the details of the hand. In fact, I think
>> both can be combined, by defining the different hands in <handNote>
>> elements, and associating the <change> elements with the correct
>> hands via the @who attribute:
>>
>>   <teiHeader>
>>     <fileDesc>
>>       <!-- ... -->
>>       <sourceDesc>
>>         <listWit>
>>           <witness xml:id="M1">draft manuscript</witness>
>>           <witness xml:id="Ed1">the first edition</witness>
>>           <witness xml:id="P2">a proof for the second edition (based
>> on Ed1, with manual corrections)</witness>
>>           <witness xml:id="Ed2">the second editon</witness>
>>         </listWit>
>>       </sourceDesc>
>>     </fileDesc>
>>     <profileDesc>
>>       <handNotes>
>>         <handNote xml:id="au-r" medium="red-pencil"
>> scribe="author">author's annotations in red pencil</handNote>
>>         <handNote xml:id="au-b" medium="black-pencil"
>> scribe="author">author's annotations in black pencil</handNote>
>>         <handNote xml:id="corr" medium="red-ink"
>> scribe="corrector">corrector's annotations in red ink</handNote>
>>       </handNotes>
>>       <creation>
>>         <listChange>
>>           <listChange corresp="#M1">
>>             <change xml:id="M1a" who="#au-r">authorial correction
>> layer in red pencil</change>
>>             <change xml:id="M1b" who="#au-b">authorial correction
>> layer in black pencil</change>
>>           </listChange>
>>           <listChange corresp="#P2">
>>             <change xml:id="P2a" who="#au-b">authorial correction
>> layer in blue ink</change>
>>             <change xml:id="P2b" who="#corr">correction layer by
>> corrector in red ink</change>
>>           </listChange>
>>         </listChange>
>>       </creation>
>>     </profileDesc>
>>   </teiHeader>
>>
>> This seems more expressive, since:
>>     -one hand can be associated with several correction layers in the
>> history of a text. In the slightly expanded version above, revisions
>> in distinct stages of a text (M1b and P2a) can be ascribed to one
>> single authorial hand (au-r).
>>     -the fact that <listChange> elements can nest, allows for
>> grouping related revision rounds, and linking them with the precise
>> text witness in which they occur
>>
>> Again, this is just my interpretation so I might be wrong. Thanks for
>> pointing this out, anyway!
>>
>> Best,
>>
>> Ron
>>