Re: Degrees of volition in active languages (was Re: Chevraqis: asketch)

> From: Jörg Rhiemeier
> Subject: Re: Degrees of volition in active languages (was Re: Chevraqis:
> asketch)
>
> The Gray Wizard wrote:
>
> > > [Degrees of volition in Nur-ellen]
> > >
> > > If the action is something the subject does in error or accidentally
> > > without intending it, the dative preposition _na_ is used, as in:
> > >
> > > Na  Turin dagnent   Veleg     mjeln.
> > > DAT Turin kill-PAST OBJ.Beleg OBJ.friend
> > >
> > > (Turin and Beleg are characters from _The Silmarillion_;
> > >  Turin kills his friend Beleg, mistaking him for an enemy.)
> >
> > amman iar (coincidently also derived from JRRT's elvish
> languages) has the
> > best of both worlds here.  Its nominal case system is syntactically
> > ergative, but it also has what I call a predicate inflection
> that is active.
>
> How did this double marking evolve in your fictional history?
> This is highly interesting.  Most languages use one and only one system
> of marking semantic relations or cases, but Amman-iar uses two different
> of them.
Not enough is known about the protolanguage Vulanayal to be certain, but
linguists speculate that it occurred through a combination of language
mixing from as many as four different sources at very widely separated
times.  The ur-language of amman seems to have been primarily active.  A
very early contact with Grey Elves (long before the coming of the
Numenoreans) apparently added a nominative/accusative case marking to
possibly all pronouns, but at minimum to the speech-act pronouns.  With the
coming of the Numenoreans, two new sources of language mixing occurred.
Since the escaping Numenoreans were of the court of Amandil of Andunie, they
still spoke Sindarin and influenced the language of amman primarily
lexically.  However, as Numenoreans they also spoke Adunaic (possible some
Sindarin/Adunaic) Creole) and it was the latter that is thought to have
added the ergative influence.

> > While the case system does not differentiate between degrees of
> volition,
> > the predicate inflections do.  Thus,
> >
> > i durnanne eleth en i velgan ernurgoraen
> >
> > \t i   durnanne         eleth             en         i
> > \m i   turin =an   -e   el-        -eth   en         i
> > \g the Turin =masc -[A] assertive- -past  agt.to.pat the
> > \p det nam   =gnd  -erg mood-      -tense ptp        det
> > \x the Turin            did               agt.to.pat the
> >
> > \t velgan           ernurgoraen
> > \m beleg =an   -0   er-  en-    ur-  coiro -ae      -n
> > \g Beleg =masc -[P] do-  cause- not- live  -agt/pat -actn/proc
> > \p nam   =gnd  -abs agt- caus-  neg- v     -val     -vc
> > \x Beleg            slay
>
> I note an interesting feature here, namely, the use of definite articles
> with proper names.  This strikes me as odd, because it is so redundant.
> But then, there all sorts of odd redundant things done in languages.
Actually, this is a borrowing from German taken to its extreme.  My wife is
German and we spend a lot of time there.  Germans are fond of referring to
Der Hartmut and Die Marlies.  Although this form is not mandatory in German,
it became so in amman iar, particularly in literature and at the court which
are the primary usages of the language.  Haven't ever come up with a good
concultural explanation for this, however.

> > Here Turin, in A-function, is marked with the ergative case
> while Beleg, in
> > P-function, is in the absolutive.  Further, the auxillary
> _eleth en_ marks
> > Beleg as patientive to Turin's agentive role and the agentive predicate
> > inflection _er-_ on the lexical verb denotes the intentional
> action of the
> > agent.  Finally, the valence _-ae_ on the lexical verb is an active
> > indication of the agentive/patientive semantic roles played by the
> > arguments.  So here, Turin has intentionally killed Beleg
> (contrary to the
> > Silmarillion's claim).
>
> I really like this marking system!  It is so absurdly complex, but it
> makes sense.
Were I to choose a single distinguishing characteristic of amman iar, this
would be it.  Though seemingly complex, it becomes quite natural with use.

> Nur-ellen, as can be seen from the examples I have given, does things in
> a simpler manner, it does active-type case marking on the noun.  In
> Nur-ellen, case marking is semantic, while the counterpart of the
> syntactic case marking of Amman-iar is word order.  Basic word order in
> Nur-ellen is SVO, but can be handled freely;
> for example, while Nur-ellen does not have passive voice, the object can
> be put first.  A sentence like _Veleg dagnent Turin_ could be translated
> as "Beleg was killed by Turin"; the case marking makes it clear that it
> cannot mean "Beleg killed Turin", which would be _Beleg dagnent Durin_.
> Sentences with inanimate "subjects", as in "A computer computed the
> orbit of Mars", are formed by using an instrumental phrase which in such
> cases is usually put in front, e.g.
amman iar uses word order for pragmatic distinctions, giving preference to
TFV (topic, focus, verb) order.

> > However,
> >
> > eleth an i durnanne i velgan enurgorial
> >
> > \t eleth             an     i   durnanne         i
> > \m el-        -eth   an     i   turin =an   -e   i
> > \g assertive- -past  to.pat the Turin =masc -[A] the
> > \p mood-      -tense ptp    det nam   =gnd  -erg det
> > \x did               to.pat the Turin            the
> >
> > \t velgan           enurgorial
> > \m beleg =an   -0   en-    ur-  coiro -ia      -l
> > \g Beleg =masc -[P] cause- not- live  -pat/thm -actn
> > \p nam   =gnd  -abs caus-  neg- v     -val     -vc
> > \x Beleg            slay
> >
> > Here Turin and Beleg continue to be marked for their syntactic roles as
> > ergative and absolutive respectively, but now the auxillary
> _eleth an_ marks
> > Turin as semantically patientive.  The absence of an agentive
> inflection on
> > the lexical verb indicates a nonvolitional action while the valence
> > indicates the patientive/theme roles played by the arguments.  So here
> > (faithful to JRRT's intentions), Turin has accidentally killed Beleg.
>
> An interesting solution.  So the roles are shifted like this:
>
> Turin:  Agent   ---> Patient
> Beleg:  Patient ---> Theme
>
> Do I understand it correctly?
Precisely!

> In contrast, in Nur-ellen, the transformation is thus:
>
> Turin:  Agent   ---> Experiencer
> Beleg:  Patient ---> Patient
Yes, I noticed that.  amman iar doesn't really differentiate an Experiencer
semantic role, but is rather limited to the three, Agent, Patient and Theme.
The choice of shift was made here because, lacking volition, the A-function
argument is in some sense acted upon and is thus patientive.  This leaves
only the Theme role for the P-function argument.

> > > With verbs of perception, dative distinguished cursory perception
> > > from intentional observation, as in
> > >
> > > Martin lastent linnel e Daeron.
> > > Martin listened to the singing of Daeron.
> > >
> > > vs.
> > >
> > > Na Martin lastent dringel e neug.
> > > Martin heard the hammering of the dwarves.
> >
> > The same pattern as above applies to verbs of perception, thus
> >
> > vartanne eleth en i limper i dhaerannion erlauiel
> >
> > \t vartanne          eleth             en         i
> > \m martin =an   -e   el-        -eth   en         i
> > \g Martin =masc -[A] assertive- -past  agt.to.pat the
> > \p nam    =gnd  -erg mood-      -tense ptp        det
> > \x Martin            did               agt.to.pat the
> >
> > \t limper                i   dhaerannion
> > \m linno =per       -0   i   daeron =an   -ion
> > \g sing  =actn/proc -[P] the Daeron =masc -of
> > \p v     =nsfx      -abs det nam    =gnd  -gen
> > \x singing               the of_Daeron
> >
> > \t erlauiel
> > \m er-  lauo -ie      -l
> > \g do-  hear -agt/thm -actn
> > \p agt- v    -val     -vc
> > \x hear
> >
> > Would be translated as "Martin _listened to_ the singing of Daeron."
> >
> > while,
> >
> > eleth an i vartanne limper i dhaerannion lauial
> >
> > \t eleth             an     i   vartanne
> > \m el-        -eth   an     i   martin =an   -e
> > \g assertive- -past  to.pat the Martin =masc -[A]
> > \p mood-      -tense ptp    det nam    =gnd  -erg
> > \fp did               to.pat the Martin
> >
> > \t limper                i   dhaerannion     lauial
> > \m linno =per       -0   i   daeron =an   -ion lauo -ia      -l
> > \g sing  =actn/proc -[P] the Daeron =masc -of  hear -pat/thm -actn
> > \p v     =nsfx      -abs det nam    =gnd  -gen v    -val     -vc
> > \fp singing               the dhaeronannion     hear
> >
> > would be translated as "Martin _heard_ the singing of Daeron."
>
> While Amman-iar and Nur-ellen shift roles differently, each language
> uses
> the same pattern for both situations.
>
> In Amman-iar:
>
> Martin:                Agent   ---> Patient
> the singing of Daeron: Patient ---> Theme
>
> In Nur-ellen:
>
> Martin:                Agent   ---> Experiencer
> the singing of Daeron: Patient ---> Patient
Yes, and that is as I would have expected given the distribution of semantic
roles in the two languages.  Hmmm, I wonder if Nur-ellen had been yet a
fifth influence on amman iar in some forgotten past? :-)

> So Amman-iar has a Patient/Experiencer polysemy and Nur-ellen a
> Patient/Theme polysemy.  However, the two patients are still treated
> differently in Amman-iar as one is in ergative and the other in
> absolutive case.
Yes, because in amman iar nominal case marking is purely syntactic.

> > Nur-ellen certainly looks interesting.  Do you have a website?
>
> Unfortunately not yet.  I am still working on it, but I am going to set
> up a page about Nur-ellen soon.
Well, work a little harder!  Can't you see there are those of us anxious to
learn more about this interesting language? ;-)

David

David E. Bell
The Gray Wizard
    dbell@graywizard.net
    www.graywizard.net

"Wisdom begins in wonder." - Socrates