linguists could use this, as a rough
equivalent to MathML. i've seen tree diagrams
of utterances down to the word in <abbr
title="second language
acquisition">SLA</abbr>  theory. probly
doesn't warrant its own module though.

i've never written/used a custom XML
application, but i've imagine there'd be
problems with phrasal verbs. like split
infinitives they'll be interrupted by other
parts of speech.

(ignoring the myth of tense for a moment)

<sentence type="declarative">
<nounphrase><pronoun number="singular"
case="nominative">I</pronoun></nounphrase>
<verbphrase>
<phrasalverb tense="past simple"
number="singular">
woke up</phrasalverb>
<adverb type="time">late</adverb>.
</verbphrase>
</sentence>

make the verb transitive and it's split by
its object


<verbphrase><phrasalverb tense="present
simple">wake <object type="direct"><pronoun
number="singular"
case="accusative">him</pronoun></object>
up</phrasalverb> 

semantically the object should be nested
inside the phrasal verb when it's transitive
"wake [s.o.] up."

i don't know if that applies to modifiers tho.

e.g.

get medieval on [s.o.]

i'm gonna <phrasalverb>get medieval on
<object ><trope="synecdoche" alt="you">your
ass</trope></object>

works fine, but

i'm gonna GET wicked MEDIEVAL borgia-style ON
[YOUR] sorry [ASS].

has the transitive phrasal verb (including
the object) split into four sections, and the
modifiers truly are separate from the verb
(and therefore shouldnt be children of
<phrasalverb>).

and,

<contraction title="I
am">I<linkingverb>'m</contraction> going
to</linkingverb>

is gonna break something.

(i also imagine there'd be some polarization
between pro-descriptive and pro-ascriptive
camps, split infinitives and "this is he" and
whatnot...)

nate jarvis


"Ernest Cline" からのメール

> 
> 
> Ha, ha.
> A nice little joke, yet one with a grain of
> truth in it.
> 
> It certainly makes as much sense for there
to
> be an XHTML2 Grammar 
> Module as there does an XHTML2 Computing
> Module, Indeed, we've already 
> had several requested elements that would
fit
> in such a module, one of 
> which, nr, is still under consideration.
> 
> Here follows a semi-serious proposal for a
> Grammar Module
> 
> <sentence>
> <clause>
> <subject> or <subj>
> <predicate> or <pred>
> <nounp> or <np> (noun phrase)
> <verbp> or <vp> (verb phrase)
> <prepp> or <pp> (prepositional phrase)
> <noun> or <n>
> <pronoun> or <pron>
> <verb> or <v>
> <adjective> or <adj>
> <adverb> or <adv>
> <preposition> or <prep>
> <conjunction> or <conj>
> <interjection> or <interj>
> <determiner> or <det>
> 
> Each of the parts of the speech would of
> course also need to have 
> attributes defined so as to allow for even
> finer grained information to 
> be provided for those who really feel the
> need to include it. (The 
> proposed nr element could be handled as <adj
> role="number"> in such a 
> module.
> 
> <sentence>
>  <conj>However</conj>,
>  <clause>
>   <clause mood="subjunctive">
>    <conj>if</conj>
>    <subj><np><pron
> person="2">you</pron></np></subj>
>    <pred tense="present">
>     <v>think</v>
>     <np>
>      <pron>that</pron>
>      <clause>
>       <subj><np><pron person="1"
> number="sing">I</pron></np></subj>
>       <pred tense="present">
>        <v role="aux">would</v>
>        <v role="main">want</v>
>        <np>
>         <adj>every</adj>
>         <n>sentence</n>
>         <pp role="adj">
>          <p>in</p>
>          <n><abbr title="XML HTML Version
> 2.0">XHTML2</abbr></n>
>         </pp>
>         <vp role="adj">
>          <v>marked</v>
>          <np>
>           <adj>this</adj>
>           
>
<n>way</n></np></vp></np></pred></clause></np></pred></clause>,
>   <clause>
>    <conj>then</conj>
>    <subj><np><pron
> number="2">you</pron></np></subj>
>    <pred tense="present">
>     <v>are</v>
>    
<adj>crazy</adj></pred></clause></clause>!
> </sentence>
> 
> Rather, even if something like this were
> included in XHTML2, (which I 
> am not in favor of, just as I am not in
favor
> of the elements that 
> would be in the proposed Computing Module.)
I
> would expect it to be 
> used only when the author considered marking
> up the grammar to be 
> important in conveying his meaning.
>     
>