Dear Neel,

thank you for replying! Sorry for my own late reply. I've been thinking 
about this.

> [...] at some point someone
> must cite a version!
> As you note, CTS URNs also allow that, so I think your important
> question really is more about
> best scholarly practices than technological protocols:  when is it
> appropriate to
> cite at a notional level, and when not?  I think that some of the
> perceived difficulty
> in this practice is rooted in the ways we have
> used critical editions in the past,
> since critical editions already represent reconstructions that may be
> pretty distant
> from any concrete version known in MSS or papyri.

I completely agree that the key issue does not lay in the CITE protocol 
but in our traditional scholarly practices.

> urn:cts:latinLit:phi0959.phi005:27
> is an assertion that "line 27" is a valid reference to a "notional" work
> -- as you
> succinctly express it, this is a formalization of ubiquitous practice
like
> "Ov. Rem. 27".

True. In fact, this practice is so ubiquitous that I think that the
urn:cts:latinLit:phi0959.phi005:27
type of URN (without a reference to a specific digital version/edition) 
will be very widely used -- more widely than the type
urn:cts:latinLit:phi0959.phi005.specificversion:27

CITE would would not be to blame for this, since it provides two ways to 
cite passages and leaves the choice to scholars:
(1) urn:cts:latinLit:phi0959.phi005:27
(2) urn:cts:latinLit:phi0959.phi005.specificversion:27

I think that (1) is a digital replica of the traditional way -- a lost 
opportunity --, while (2) is a more "philological" way to cite texts.

I like (2) better than (1) because I think that digital philology should 
help philology to achieve one of its goals, that is to give us a more 
open/plural vision of what a text is.

Again, this does not have to do with the CITE protocol per se, but with 
how we'll use it. It's why I think it might be worth mentioning on the list.


> There is no "binding" of notional works to any specific version; there
> is no instantiation of an
> archetype.

This is yet another point. Regardless of the scholar's choice is between 
(1) or (2), in the current implementations of the CITE protocol the 
conventional codes for authors/works/numbering are based on the the PHI 
and TLG Canons (in "tlg0001", the string "tlg" refers to the TLG Canon).

A URN like
urn:cts:latinLit:phi0959.phi005:27
does not imply that the text of that verse is
"Vitricus et gladiis et acuta dimicet hasta"
as opposed to, say,
"Vitricus et gladiis et acutis dimicet hastis".

In this sense, I agree that there is no instantiation of an archetype.

However, I'm wondering whether that URN somehow implies that verses 
25-26 are part of the text (they are not spurious), and that the Remedia 
Amoris are a work of Ovid.

I've been thinking a lot in this couple of days about what that "somehow 
implies" might mean -- if it makes any sense at all.

So, this is what I mean when I think of a macro-archetype of the whole 
Greek and Latin literary corpus:
- the "base/standard" verse/chapter order of a text is that of the 
PHI/TLG version of that text;
- the standard poem/chapter boundaries are those existing in those 
PHI/TLG versions;
- the standard (conventional) authorship of works is that proposed in 
the PHI/TLG Canons, etc.

Not because CITE establishes a text, but because its standard 
implementations establish and formalize an author/works/chapter 
numbering canon.

Again, I think that if all future digital philologists will be fully 
aware of the _completely conventional_ nature of those URNs (as you, 
Neel, pointed out), everything will be perfectly fine, and we'll have a 
powerful tool to use.

As you write, tlg0012.tlg001 (tlg0012=Homer; tlg001=Ilias) should not be 
understood to imply that an individual named "Homer" actually existed, 
and that he is the author of a work entitled "Ilias". It is just a 
conventional code to point to a cultural object.

I guess, then, that my questions here are: will this awareness always be 
full in all future digital philologists? Will the (hopefully) ubiquitous 
implementation of a PHI/TLG-based CITE exert some sort of long-term 
cultural influence in the direction of assuming that a Homer existed and 
that he composed the Iliad?

Which basically means: how can we help in minimizing the risk and 
maximizing the opportunity?

Paolo



Il 14/09/2015 00:51, Neel Smith ha scritto:
> Thanks for the questions with helpful specific examples.
>
> CTS URNs let you identify passages and works with the degree of specificity
> of your choosing.  In your example, the notional-level URN
>
>
> There is no "binding" of notional works to any specific version; there
> is no instantiation of an
> archetype.  So what do we do if, as in your example, two editors
> disagree about a "notional" text,
> with one editor rejecting line 27 and another accepting it?  Since
> "notional" texts mean
> nothing without some concrete supporting evidence,


sgn


>
> The other helpful examples you mention are similar in the sense that I
> think the
> explicit semantics of the CTS URN shines a light on traditional citation
> practices
> that are sometime more ambiguous than we have realized.  In the Homer
> Multitext project,
> I deal routinely with differing versions of the *Iliad* that might have
> plus/minus
> verses, or have verses in different orders, for example, and it has been
> wonderfully
> helpful to be able to express that, and to automate comparison across
> specific versions
> of texts.
>
> It's tempting to overload *citation* with various kinds of *analysis*,
> such as attribution,
> but we should keep those distinct.  Whether you think Aeschylus is the
> author of the
> Prometheus Bound or not makes no difference in how we *identify* the
> work:  it's the
> use of common identifiers that lets us identify what we interpret
> differently!  So
> in the greekLit namespace, urn:cite:chs:textgroup.tlg0085 refers to a
> group of works
> traditionally called "Works of Aeschylus", but implies nothing about
> your views on attribution:
> urn:cts:greekLit:tlg0085.tlg003 can be our common identifier for
> "Prometheus Bound" whoever
> you think the author is.
>
> I hope this quick reply makes some sense.
>
> Neel Smith
>
>
>
> On 9/13/15 2:51 PM, Paolo Monella wrote:
>> Dear all,
>>
>> I'm reading scholarship and technical reports on the CTS (Canonical
>> Text Services) protocol, and I have a question that I thought would be
>> best answered/discussed on the list.
>>
>> How does the CTS protocol deal with the possible variations of texts
>> in the philological community?
>>
>> I'll make one main practical example (preferring - hopefully - clarity
>> over brevity):
>>
>> Verses 25-26 of the Remedia Amoris by Ovid are suspected to be
>> spurious, so they are sometimes published in brackets in editions such
>> as the online PHI 5.3 in http://latin.packhum.org/loc/959/5/0#0
>>
>> puer es, nec te quicquam nisi ludere oportet:  23
>>      Lude; decent annos mollia regna tuos.     24
>> [Nam poteras uti nudis ad bella sagittis:      25
>>      Sed tua mortifero sanguine tela carent.]  26
>> Vitricus et gladiis et acuta dimicet hasta,    27
>>      Et victor multa caede cruentus eat:       28
>>
>> However, the PHI edition upon which some CTS implementations rely to
>> assign numbers to verses includes vv. 25-26 in the text, so the next
>> verse "Vitricus et gladiis et acuta dimicet hasta" is numbered as 27,
>> and so on.
>>
>> If I'm not wrong, a possible CTS URN for verse 27 would be:
>> urn:cts:latinLit:phi0959.phi005:27
>> Where:
>> - phi0959 refers to Ovid in the PHI Canon;
>> - phi005 refers to the Remedia Amoris as the 5th work by Ovid in the
>> same Canon;
>> - 27 is verse number 27 of the notational text (i.e. not a specific
>> version of the text, but the 'abstract' notational text).
>>
>> To manufacture this URN, I'm mostly relying on
>> http://www.homermultitext.org/hmt-docs/cite/cts-urn-overview.html
>> Please do correct me here and anywhere I'm wrong!
>>
>> I think that the underlying assumption (valid, in fact, in most cases
>> for Greek and Latin classical texts) is that "versions of a text are
>> related to a notional text in a conceptual hierarchy" (quoted from
>> http://katahdin.cs.dartmouth.edu/reports/TR2009-649.pdf paragraph 2,
>> point 3)
>>
>> The protocol also provides you with the option of referring to a
>> specific version of the text.
>> For example, in the Perseus Digital Library "Text URI"
>> http://data.perseus.org/citations/urn:cts:latinLit:phi0959.phi005.perseus-lat1:Rem.
>>
>> refers to the specific version/edition of the text of the Remedia
>> Amoris identifed as "perseus-lat1".
>> I think this very good, as different editions have different texts.
>>
>> My question, instead, is about the option, included in the standard,
>> to have URNs like
>> urn:cts:latinLit:phi0959.phi005:27
>> where there is no reference to a specific version/edition of the text,
>> but the reference is to verse 27 of the canonical notational text of
>> the Remedia Amoris.
>>
>> I suspect that this option will be largely implemented, as it is a
>> really convenient and effective way to formalize semi-formal
>> references ubiquitous in classical scholarship such as "Ov. Rem. 27".
>>
>> The latter URN (urn:cts:latinLit:phi0959.phi005:27) looks OK today, as
>> the scholarly community currently agrees that verses 25-26 are part of
>> the text.
>>
>> However, what will happen if in the future new research creates a
>> consensus on the thesis that verses 25-26 are spurious? Or if editor
>> Joanne Brown publishes in 2016 her own scholarly edition where verses
>> 25-26 are considered spurious and expunged from the text? In these
>> cases, most new editions (or just Brown's 2016 edition) would have the
>> following text:
>>
>> puer es, nec te quicquam nisi ludere oportet: 23
>> Lude; decent annos mollia regna tuos.         24
>> Vitricus et gladiis et acuta dimicet hasta,   25 (formerly 27)
>> Et victor multa caede cruentus eat:           26 (formerly 28)
>>
>> My understanding is that a system implementing the CTS protocol would
>> refer to the verse "Vitricus et gladiis et acuta dimicet hasta" as:
>>
>> 1) urn:cts:latinLit:phi0959.phi005.brown2016:25
>> if referring to verse 25 in the text of the "brown2016" edition of the
>> text, assuming that J. Brown provides, like Perseus does, a formal way
>> to reference her own text by means of CTS URNs;
>>
>> 2) urn:cts:latinLit:phi0959.phi005:27
>> if referring to that verse without reference to any specific version
>> of the text, but simply referring to the abstract notational text.
>>
>> In other words, my understanding is that the reference URN (when _not_
>> referring to a specific version to the text, but to the abstract
>> notational text, i.e. in case 2 above) would remain bound forever to
>> the archetypal text based on which the original verse numbers have
>> been assigned -- that is TLG and PHI.
>>
>> Am I right? Or is there a way in which the protocol deals with this?
>>
>> I won't dwell any longer on this, so I'll just quickly list other
>> possible applications of this issue:
>>
>> - verse displacement (e.g. an editor thinks that verses 1-2 should be
>> displaced after verse 8);
>>
>> - poem divisions/boundaries (some think that Propertius' elegy 2.22 is
>> one poem, some split it into two poems Prop. 2.22a Prop. 2.22b);
>>
>> - suspected (or certainly) spurious works that PHI still attributes to
>> an author (urn:cts:latinLit:phi0959.phi015 would refer to the
>> Epicedion Drusi, so it is listed under author phi0959, Ovid, though
>> with the notation [sp.]);
>>
>> - works that PHI does not attribute to an author, but that might be
>> found to belong to that author in the future
>> (urn:cts:latinLit:phi0692.phi003 would refer to the Culex and is
>> listed under author 0692, the Appendix Vergiliana, instead of author
>> 0690, Vergilius);
>>
>> - all other philological/papyrological discoveries that may affect the
>> identification of authors (e.g. TLG distinguishes between two
>> different "Aeschylus Trag.", identified as tlg0031 and 0085, the
>> famous one), the attribution of works to authors and of sections
>> (verses/poems/chapters etc.) to works.
>>
>> Thank you,
>> Paolo
>>