rs should contain q and quote

[ebeshero]

As raised on the TEI list on Oct. 30, 2016, we can't nest a <q> (or a <quote> element inside a <rs> element . The examples from Scott Vanderbilt's archaeology project posted on the list look like unusual but plausible use-cases fpr <rs> to contain quoted material, and might be useful in the Guidelines to broaden diversity of examples in the Guidelines, should we decide this ticket is "Go." This example looks especially useful:

   <rs type="findspot">in excavation, re-used in a fourth-century
   <q>chalet</q></rs>

and perhaps this:

<rs>The town <quote>formerly known as Princeton</quote></rs>

[peterstadler]

As I just wrote on TEI-L, I'd support this FR but was down voted once, see http://tei-l.970651.n3.nabble.com/quotes-in-names-tp4026585.html for some historic discussion.

[lb42]

Firstly, as I said on the thread on TEI-L, it seems wrong to me to use <rs> for a descriptive phrase like this. It's meant to behave like a name : an alternative to <placeName> or <persName>. You would never say "I found this axe at 'in excavation reused in a 4th century chalet', nor expect that description to turn up in a list of places. Hence my suggestion of using a more neutral block level (and typable) element such as <ab> to record these find-place descriptions. Secondly, if you insist on using a phrase level element such as <rs>, you only have phrase level content at your disposal. So you could use <mentioned> or <soCalled> (either of which seems to suit the "chalet" example, or (if you just want to record the presence of the quote marks) <hi>. Thirdly, the phrase "formerly known as Princeton" is not a quote, so far as I can tell, except from me! Introducing <q> into the content of a phrase level element would disrupt the current model quite severely. One prototypical use for <q> is "stuff in quotes" (it's almost a synonym for <hi rend="quoted">). The other is to encode direct speech without the bother of having to wrap each part in <p> elements. That's why it's in model.inter rather than model.phrase.

[ebeshero]

That is just the problem that people tend to have with <rs>, though, isn't it? It's a phrase that's kind of like a name, but isn't. It's a "string" that "refers": that makes it some snippet of text, and we shouldn't have to make workarounds when a string that refers happens to have quotation marks in it for whatever reason it happens to have quotation marks. A "referring string" might well contain spoken words, or words that are quoted from another source. If we can use <soCalled> why can't we also use <q> as well as <hi rend="quote">? Shouldn't all these options be available to the encoder? What is so disruptive about permitting <q> and <quote> if we're permitted <mentioned> and <soCalled>?

I think it's a little premature to close this discussion.

[lb42]

Sorry, didn't mean to close the ticket. But I do think the proposal is disruptive of the current hierarchy of content models. mentioned and soCalled and hi are phrase level elements; q is not. If it were you wouldn't be able to markup passages of direct speech without enclosing each one within a <p> for example.

[ebeshero]

Okay, so the issue, if I understand this right, is that <q> isn't a "phrase-level" element, but that <rs> is phrase-level. Even so, we use <q> above and below the paragraph level, and that makes it difficult to explain what we mean by a "phrase level" element. Well, of course <rs> is a member of model.nameLike, and that is a member of model.phrase. I understand the content-model problem, but I'm still trying to wrap my head around why we disallow <q> and <quote> from that which is model.phrase.

The issue of direct speech seems to be one of scope, but <q> is used all the time inside what most of us think of conventionally as "just a phrase." I'm studying model.phrase and wondering if the idea was to exclude quoteLike and qLike` because these are always (by the Guidelines) considered to be above phrase-level by definition.

Frankly, I'm confused by the dynamic applicability of model.inter which permits the use of qLike and quoteLike above the paragraph level AND below it. The application below the paragraph level seems to be the question here. Okay, obviously we don't want <rs> to be so big and expansive that it can contain whole paragraphs the way <q> does--I get that that would be the disruptive thing. rs is a member of model.phrase for good reason.

But what about permitting <q> and <quote> within <rs> and other phrase level elements because they are every bit as applicable as other things we treat as phrases in Real Life? You're going to tell me this is something people can do for themselves at ODD level, and I won't disagree, except they'd be doing something that explicitly (now) cuts against our hierarchy of content models. Sigh.

[ebeshero]

After all that, here's my question: Would it be an awful thing to permit a member of model.inter inside a member of model.phrase?

[ebeshero]

F2F subgroup recommends the content of rs should include model.qLike .

[sydb]

@sydb will compare lists of elements that have macro.phraseSeq as content and those that have macro.paraContent. Also to take a look at how big a deal it might be to move <q> out of model.qLike.

[sydb]

Have %macro.phraseSeq; in their content model: <abbr>, <actor>, <addName>, <addrLine>, <affiliation>, <author>, <biblScope>, <birth>, <bloc>, <catchwords>, <citedRange>, <cl>, <colophon>, <country>, <death>, <distinct>, <distributor>, <district>, <docAuthor>, <docDate>, <edition>, <editor>, <education>, <eg>, <email>, <expan>, <explicit>, <extent>, <faith>, <finalRubric>, <floruit>, <foreign>, <forename>, <fw>, <genName>, <geoDecl>, <geogFeat>, <geogName>, <gloss>, <headItem>, <headLabel>, <heraldry>, <incipit>, <label>, <material>, <measure>, <mentioned>, <name>, <nameLink>, <nationality>, <num>, <objectType>, <offset>, <orgName>, <origPlace>, <persName>, <phr>, <placeName>, <pubPlace>, <publisher>, <region>, <residence>, <role>, <roleDesc>, <roleName>, <rs>, <rubric>, <s>, <secFol>, <settlement>, <sex>, <soCalled>, <socecStatus>, <speaker>, <stamp>, <street>, <surname>, <term>, <textLang>, <watermark>, <wit>, and <witDetail>.

Have %macro.paraContent; in their content model: <ab>, <add>, <camera>, <caption>, <case>, <colloc>, <corr>, <damage>, <def>, <del>, <docEdition>, <emph>, <gen>, <gram>, <hi>, <hyph>, <iType>, <imprimatur>, <lang>, <lbl>, <mod>, <mood>, <number>, <orig>, <orth>, <p>, <per>, <pos>, <pron>, <ref>, <reg>, <restore>, <retrace>, <rhyme>, <salute>, <secl>, <seg>, <sic>, <signed>, <sound>, <stress>, <subc>, <supplied>, <surplus>, <syll>, <tech>, <title>, <titlePart>, <tns>, <u>, <unclear>, <usg>, and <writing>.

GI	phraseSeq	paraContent
ab		macro.paraContent
abbr	macro.phraseSeq
actor	macro.phraseSeq
add		macro.paraContent
addName	macro.phraseSeq
addrLine	macro.phraseSeq
affiliation	macro.phraseSeq
author	macro.phraseSeq
biblScope	macro.phraseSeq
birth	macro.phraseSeq
bloc	macro.phraseSeq
camera		macro.paraContent
caption		macro.paraContent
case		macro.paraContent
catchwords	macro.phraseSeq
citedRange	macro.phraseSeq
cl	macro.phraseSeq
colloc		macro.paraContent
colophon	macro.phraseSeq
corr		macro.paraContent
country	macro.phraseSeq
damage		macro.paraContent
death	macro.phraseSeq
def		macro.paraContent
del		macro.paraContent
distinct	macro.phraseSeq
distributor	macro.phraseSeq
district	macro.phraseSeq
docAuthor	macro.phraseSeq
docDate	macro.phraseSeq
docEdition		macro.paraContent
edition	macro.phraseSeq
editor	macro.phraseSeq
education	macro.phraseSeq
eg	macro.phraseSeq
email	macro.phraseSeq
emph		macro.paraContent
expan	macro.phraseSeq
explicit	macro.phraseSeq
extent	macro.phraseSeq
faith	macro.phraseSeq
finalRubric	macro.phraseSeq
floruit	macro.phraseSeq
foreign	macro.phraseSeq
forename	macro.phraseSeq
fw	macro.phraseSeq
gen		macro.paraContent
genName	macro.phraseSeq
geoDecl	macro.phraseSeq
geogFeat	macro.phraseSeq
geogName	macro.phraseSeq
gloss	macro.phraseSeq
gram		macro.paraContent
headItem	macro.phraseSeq
headLabel	macro.phraseSeq
heraldry	macro.phraseSeq
hi		macro.paraContent
hyph		macro.paraContent
iType		macro.paraContent
imprimatur		macro.paraContent
incipit	macro.phraseSeq
label	macro.phraseSeq
lang		macro.paraContent
lbl		macro.paraContent
material	macro.phraseSeq
measure	macro.phraseSeq
mentioned	macro.phraseSeq
mod		macro.paraContent
mood		macro.paraContent
name	macro.phraseSeq
nameLink	macro.phraseSeq
nationality	macro.phraseSeq
num	macro.phraseSeq
number		macro.paraContent
objectType	macro.phraseSeq
offset	macro.phraseSeq
orgName	macro.phraseSeq
orig		macro.paraContent
origPlace	macro.phraseSeq
orth		macro.paraContent
p		macro.paraContent
per		macro.paraContent
persName	macro.phraseSeq
phr	macro.phraseSeq
placeName	macro.phraseSeq
pos		macro.paraContent
pron		macro.paraContent
pubPlace	macro.phraseSeq
publisher	macro.phraseSeq
ref		macro.paraContent
reg		macro.paraContent
region	macro.phraseSeq
residence	macro.phraseSeq
restore		macro.paraContent
retrace		macro.paraContent
rhyme		macro.paraContent
role	macro.phraseSeq
roleDesc	macro.phraseSeq
roleName	macro.phraseSeq
rs	macro.phraseSeq
rubric	macro.phraseSeq
s	macro.phraseSeq
salute		macro.paraContent
secFol	macro.phraseSeq
secl		macro.paraContent
seg		macro.paraContent
settlement	macro.phraseSeq
sex	macro.phraseSeq
sic		macro.paraContent
signed		macro.paraContent
soCalled	macro.phraseSeq
socecStatus	macro.phraseSeq
sound		macro.paraContent
speaker	macro.phraseSeq
stamp	macro.phraseSeq
street	macro.phraseSeq
stress		macro.paraContent
subc		macro.paraContent
supplied		macro.paraContent
surname	macro.phraseSeq
surplus		macro.paraContent
syll		macro.paraContent
tech		macro.paraContent
term	macro.phraseSeq
textLang	macro.phraseSeq
title		macro.paraContent
titlePart		macro.paraContent
tns		macro.paraContent
u		macro.paraContent
unclear		macro.paraContent
usg		macro.paraContent
watermark	macro.phraseSeq
wit	macro.phraseSeq
witDetail	macro.phraseSeq
writing		macro.paraContent

Have neither in their content model: <TEI>, <abstract>, <accMat>, <acquisition>, <activity>, <addSpan>, <additional>, <additions>, <address>, <adminInfo>, <age>, <alt>, <altGrp>, <altIdent>, <altIdentifier>, <alternate>, <am>, <analytic>, <anchor>, <annotationBlock>, <anyElement>, <app>, <appInfo>, <application>, <arc>, <argument>, <att>, <attDef>, <attList>, <attRef>, <authority>, <availability>, <back>, <bibl>, <biblFull>, <biblStruct>, <bicond>, <binary>, <binaryObject>, <binding>, <bindingDesc>, <body>, <broadcast>, <byline>, <c>, <cRefPattern>, <caesura>, <calendar>, <calendarDesc>, <castGroup>, <castItem>, <castList>, <catDesc>, <catRef>, <category>, <cb>, <cell>, <certainty>, <change>, <channel>, <char>, <charDecl>, <charName>, <charProp>, <choice>, <cit>, <classCode>, <classDecl>, <classRef>, <classSpec>, <classes>, <climate>, <closer>, <code>, <collation>, <collection>, <cond>, <condition>, <constitution>, <constraint>, <constraintSpec>, <content>, <correction>, <correspAction>, <correspContext>, <correspDesc>, <creation>, <custEvent>, <custodialHist>, <damageSpan>, <dataFacet>, <dataRef>, <dataSpec>, <datatype>, <date>, <dateline>, <decoDesc>, <decoNote>, <default>, <defaultVal>, <delSpan>, <depth>, <derivation>, <desc>, <dictScrap>, <dim>, <dimensions>, <div>, <div1>, <div2>, <div3>, <div4>, <div5>, <div6>, <div7>, <divGen>, <docImprint>, <docTitle>, <domain>, <eLeaf>, <eTree>, <editionStmt>, <editorialDecl>, <egXML>, <elementRef>, <elementSpec>, <empty>, <encodingDesc>, <entry>, <entryFree>, <epigraph>, <epilogue>, <equipment>, <equiv>, <etym>, <event>, <ex>, <exemplum>, <f>, <fDecl>, <fDescr>, <fLib>, <facsimile>, <factuality>, <figDesc>, <figure>, <fileDesc>, <filiation>, <floatingText>, <foliation>, <forest>, <form>, <formula>, <front>, <fs>, <fsConstraints>, <fsDecl>, <fsDescr>, <fsdDecl>, <fsdLink>, <funder>, <fvLib>, <g>, <gap>, <gb>, <geo>, <gi>, <glyph>, <glyphName>, <gramGrp>, <graph>, <graphic>, <group>, <handDesc>, <handNote>, <handNotes>, <handShift>, <head>, <height>, <history>, <hom>, <hyphenation>, <iNode>, <ident>, <idno>, <if>, <iff>, <imprint>, <incident>, <index>, <institution>, <interaction>, <interp>, <interpGrp>, <interpretation>, <item>, <join>, <joinGrp>, <keywords>, <kinesic>, <l>, <lacunaEnd>, <lacunaStart>, <langKnowledge>, <langKnown>, <langUsage>, <language>, <layout>, <layoutDesc>, <lb>, <leaf>, <lem>, <lg>, <licence>, <line>, <link>, <linkGrp>, <list>, <listApp>, <listBibl>, <listChange>, <listEvent>, <listForest>, <listNym>, <listOrg>, <listPerson>, <listPlace>, <listPrefixDef>, <listRef>, <listRelation>, <listTranspose>, <listWit>, <localName>, <locale>, <location>, <locus>, <locusGrp>, <m>, <macroRef>, <macroSpec>, <mapping>, <measureGrp>, <media>, <meeting>, <memberOf>, <metDecl>, <metSym>, <metamark>, <milestone>, <model>, <modelGrp>, <modelSequence>, <moduleRef>, <moduleSpec>, <monogr>, <move>, <msContents>, <msDesc>, <msFrag>, <msIdentifier>, <msItem>, <msItemStruct>, <msName>, <msPart>, <musicNotation>, <namespace>, <node>, <normalization>, <notatedMusic>, <note>, <notesStmt>, <numeric>, <nym>, <oRef>, <objectDesc>, <occupation>, <opener>, <org>, <origDate>, <origin>, <outputRendition>, <pRef>, <param>, <paramList>, <paramSpec>, <particDesc>, <pause>, <pb>, <pc>, <performance>, <person>, <personGrp>, <persona>, <physDesc>, <place>, <population>, <postBox>, <postCode>, <postscript>, <precision>, <prefixDef>, <preparedness>, <principal>, <profileDesc>, <projectDesc>, <prologue>, <provenance>, <ptr>, <publicationStmt>, <punctuation>, <purpose>, <q>, <quotation>, <quote>, <rdg>, <rdgGrp>, <re>, <recordHist>, <recording>, <recordingStmt>, <redo>, <refState>, <refsDecl>, <relatedItem>, <relation>, <remarks>, <rendition>, <repository>, <resp>, <respStmt>, <respons>, <revisionDesc>, <root>, <row>, <said>, <samplingDecl>, <schemaRef>, <schemaSpec>, <scriptDesc>, <scriptNote>, <scriptStmt>, <seal>, <sealDesc>, <segmentation>, <sense>, <sequence>, <series>, <seriesStmt>, <set>, <setting>, <settingDesc>, <shift>, <signatures>, <source>, <sourceDesc>, <sourceDoc>, <sp>, <spGrp>, <space>, <span>, <spanGrp>, <specDesc>, <specGrp>, <specGrpRef>, <specList>, <sponsor>, <stage>, <state>, <stdVals>, <string>, <styleDefDecl>, <subst>, <substJoin>, <summary>, <superEntry>, <support>, <supportDesc>, <surface>, <surfaceGrp>, <surrogates>, <symbol>, <table>, <tag>, <tagUsage>, <tagsDecl>, <taxonomy>, <teiCorpus>, <teiHeader>, <terrain>, <text>, <textClass>, <textDesc>, <textNode>, <then>, <time>, <timeline>, <titlePage>, <titleStmt>, <trailer>, <trait>, <transcriptionDesc>, <transpose>, <tree>, <triangle>, <typeDesc>, <typeNote>, <undo>, <unicodeName>, <vAlt>, <vColl>, <vDefault>, <vLabel>, <vMerge>, <vNot>, <vRange>, <val>, <valDesc>, <valItem>, <valList>, <value>, <variantEncoding>, <view>, <vocal>, <w>, <when>, <width>, <witEnd>, <witStart>, <witness>, <xenoData>, <xr>, and <zone>.

[sydb]

OK, sorry, but almost a year later someone is going to have to remind me of what we meant by “Also to take a look at how big a deal it might be to move <q> out of model.qLike”.

[sarcanon]

I feel like something drastic needs to happen to get this ticket moved along. Googles self-immolation...

[lb42]

I don't see that any of the doubts raised at the start of this thread have been addressed in subsequent discussion. So probably self immolation is your best course.:grinning:

[ebeshero]

Well, short of such violent metaphorical drastic measures, we could put a Guidelines 3.5.0 milestone marker on this. Apparently we were intending to implement this after all...

[ebeshero]

Reviewing the history of the ticket, what was marked as "go" on 6 February 2017 was simply this:" F2F subgroup recommends the content of rs should include model.qLike . " How did that become so complicated afterwards?

[jamescummings]

I'm still of the opinion that this is a mis-use of <rs>. I still think @lb42 has it right in October 2016 https://github.com/TEIC/TEI/issues/1531#issuecomment-257180878

Can someone provide me an example of something which is truly a general purpose name or string referencing a name (like 'he', 'she', etc.) that needs quotes inside it? (i.e. that shouldn't be just marked with <name>?)

[hcayless]

I guess the larger issue is that <q> is meant to mark "quoted" text, and that can happen at any level. Not thinking so much about <rs> here, but any name: what about Constantino "Big" Paul Castellano, e.g.? I know there are other ways to do it, but it seems weird that I can't use the particular element designed to mark quoted text.

[ebeshero]

How about this for an example, making a point of nesting the <q>:

<rs>the one who goes by <q>he who must not be named</q></rs>

[ebeshero]

Of course you could use <soCalled> in the previous one, so here's another:

<rs>the one who coined the phrase <quote>the buck stops here</quote></rs>

[jamescummings]

@hcayless Again, that isn't an rs, that is a proper name. So use name or persName.

@ebeshero Again, that isn't an rs that is a [PHRASE] that contains an rs. I'd say <rs rend="quoted">he who must not be named</rs> In the second example it is <rs>the one</rs>.

I just feel all of these are abuse of rs, I might be in the minority but just trying to be honest.

I should state that I'm not adverse to rs (and *name) having q in its content model... just still haven't seen an example that really looks like an rs with a quote mark inside. I just want a really good example that no one can argue with. ;-) Or since a F2F subgroup already decided to allow it.... go ahead but we should be clear what we want to use it for and not encourage the kind of tag abuse (IMHO) that started this ticket.

[ebeshero]

@jamescummings I beg your pardon, but neither of my examples was a sentence! Both were phrases intended to indicate a person who isn't properly named, and restricting the encoding to surround <rs>the one</rs> would be incomplete and insufficient to indicate the individual indicated.

[jamescummings]

Apologies, I should have said 'phrase', corrected. How is surrounding <rs>The one</rs> less sufficient than surrounding <rs>The fighter</rs> (an example in the guidelines) or similar? If you want to relate it to a person then you use the @ref attribute or give it an @xml:id and point to it.

rs should, IMHO, be a word or phrase standing in for a name. 'his lady' and such. The first example given by the original poster:

<rs type="findspot">in excavation, re-used in a fourth-century
   <q>chalet</q></rs>

Just is not an rs in my reading of the guidelines.

[ebeshero]

@jamescummings Perhaps you don't yet know to whom "the one" refers and aren't in a position to give it a @ref attribute on first encoding. We don't always start projects with full-blown personographies.

[jamescummings]

@ebeshero Sure... then you don't give it a ref... But that doesn't mean you should be wrapping whole phrases and sentences in rs. IMHO.

[ebeshero]

@jamescummings My point is that it does sometimes make a difference to capture a full "referencing string" and abbreviating it to "the one" strikes me as manifestly improper because it's semantically incomplete.

[jamescummings]

How is 'the one' semantically incomplete but 'the fighter' not?

[ebeshero]

@jamescummings Imagine capturing every referencing string used for a particular person, even when you do have them filed in a personography and can use @ref. Perhaps the project is to capture and extract a list of all phrases used to refer to a particular person (and not to other persons). I'd want to pan outward a little bit and having <rs> capture a phrase with a quote inside it doesn't seem far-fetched to me.

[ebeshero]

@jamescummings "the fighter" in a project full of warriors might well be semantically incomplete without the fuller context supplied by the thing s/he'd say.

[ebeshero]

@jamescummings It's simply a matter of precision.

[jamescummings]

The example is:

<s n="2">
 <rs xml:id="R4">The fighter</rs> is serving 30-40 years
 for a 1975 armed robbery conviction in <rs xml:id="R5">the penitentiary</rs>.
</s>

Do you think the findspot example that started this a valid use?

[ebeshero]

@jamescummings Sure, there's nothing invalid about this use. And this example is not relevant to the discussion of whether <q> or <quote> should be used inside <rs>.

[jamescummings]

But it is the example which the original poster cited as the reason for this ticket?!!

   <rs type="findspot">in excavation, re-used in a fourth-century
   <q>chalet</q></rs>

Is just pure abuse to me.

[ebeshero]

Ah, I thought you were referring to the example you just posted in your message--I didn't scroll up to check the original case. I'm not sure about that "findspot" example and I wouldn't want to use it in the Guidelines--there's something very project specific about this that isn't easy to understand. But we can (and have) come up with better (because generally accessible) examples.

[hcayless]

Again, that isn't an rs, that is a proper name. So use name or persName.

@jamescummings Can't. Their content models don't permit it. <q>'s note even says that it may be thought of as syntactic sugar for <hi rend="quote"> in some circumstances, but it can't be used in the same places as <hi>.

[jamescummings]

No, the example I posted (the fighter) is one from the Guidelines.. what we currently recommend.

[ebeshero]

@jamescummings Yep, and I see how you're using it to suggest a sort of minimum valid length for an <rs>, but I don't think it holds water...I do think we can do a better job to articulate a limit of applicability for <rs>, but I wouldn't be in favor of limiting it to no more than two words in the string.

[jamescummings]

@hcayless And that is indeed a good argument for it. As I said, I'm not adverse to the idea... just haven't had an example given that does it for me personally. If allowing in rs (which is meant to stand in for a name or string referring to a name) then it also needs to be allowed in name, persName, placeName, etc.

[ebeshero]

Very much agreed on that point. But I rather like my examples, and I think @jamescummings might well reconsider them... ;-)

[jamescummings]

@ebeshero It is nothing to do with its length, but its semantics. An rs IMHO is a word or phrase standing in for a name. People keep posting examples which don't do that.

[ebeshero]

How is this example not standing in for a name?

<rs>the one who coined the phrase <quote>the buck stops here</quote></rs>

[ebeshero]

Again, I'm going to say, in the example above, "the one" isn't enough to distinguish the individual precisely. We need the rest to make it clear who is being referred to.

[jamescummings]

@ebeshero If I reconsider, in your 'the one who goes by "he who must not be named"' example, to me, still, the bit that is standing in for the name is 'he who must not be named' I guess. So I'd still mark that be up as an rs, but not the whole phrase. Otherwise you'd have to mark up "The one who goes by James Cummings" as a persName rather than just James Cummings. (If you see what I mean?)

[ebeshero]

Hmmm... <rs>the one who goes by <q><persName>James Cummings</persName></q></rs> ?

[ebeshero]

Sorry--dang tick marks! ;-)

[ebeshero]

My problem with that example is that it's probably better to use <soCalled>...</soCalled> there, which is permitted. So I turned my attention to quotable aphorisms, on the formula of, <rs>the one who quipped <q>X cool statement</q></rs>.

[jamescummings]

That is closer, but again, in that instance the whole restrictive relative clause shouldn't be marked up as rs. The whole thing is not standing in place for a name, IMHO.

[ebeshero]

A reminder of the content model of <rs>:

<content>
 <macroRef key="macro.phraseSeq"/>
</content>

It's not an issue for an <rs> to be a phrase-level element.

[ebeshero]

@jamescummings I'm pretty serious about this: if it were my project (and sometimes it is), I'd want full-blown distinguishing phrases and dependent clauses when they are dedicated to identifying a particular named entity. I'd not be content with delimiting <rs> to the shortest possible definable markup. There are good reasons for wanting this, including the attribution of spoken expressions to an individual.

[jamescummings]

For spoken expressions, don't we already have 'said'? We clearly have differing understandings of what 'rs' is for... And I'm certainly not against it (and *Name) having 'q' in its content model. I just keep seeing things as examples that I really think 'rs' was intended for. Why would you be using rs instead of any of the other possibilities?

[hcayless]

And that is indeed a good argument for it. As I said, I'm not adverse to the idea... just haven't had an example given that does it for me personally. If allowing in rs (which is meant to stand in for a name or string referring to a name) then it also needs to be allowed in name, persName, placeName, etc.

@jamescummings That was my point. If it should be allowed in <name> then it should presumably be allowed in the other name-like elements. I suspect the problem is that <q> is trying to do too much.

[ebeshero]

Think about this in terms of supporting serious research questions, which is the spirit of the original post in this ticket. The phrase-level aspect of <rs> has made it flexible for people to apply to short or long referring phrases. To be sure, <rs> can be all to easily applied wrongly where people ought to be using things like <seg>, but as long as a string of words can be said to refer to a named entity or function in a naming way, I don't think we should be keeping <q> out of it. There's nothing wrong with using <q> for spoken expressions, but if we really want to prevent people from saying, "use a different element", then let's consider a distinguishing expression that was written rather than spoken.