Re: Inverse marking (was: Kijeb text uploaded)

Eldin wrote:

 > These languages are a subset of the alignment-type called
 > "hierarchical alignment". In such languages, the "word"-
 > order of the clause and the agreement marking on the verb
 > always puts that participant which is highest in the
 > hierarchy first, regardless of whether it is the agent or
 > the patient.  Thus some kind of "voice" marking on the verb
 > is necessary to indicate whether this agreed-with
 > participant is the agent or the patient.

That makes sense.  I'll introduce hierarchical word order in
Kijeb, lest the inverse marking seem somewhat redundant.
However Kijeb also has nom/acc case marking; I'm considering
to restrict the marked accusative to animates.

 > (The hierarchy in question is usually one of agent-potency
 > (that is, potentiality to be an agent), as opposed to topic-
 > worthiness, according to M.H.Klaiman.)

That also makes sense, according to how I understand
the whole animacy hierarchy business.  That's also
why I'ld consider not to mark the accusative on
inanimates: they'ld be patients by default, so to speak.
It also would help to make the introduction of an
ergative system in later stages of the language more
realistic.

 > ----------
 >
 > Most languages that have both an "inverse voice and direct
 > voice" system of grammatical voices, and also an "obviative"
 > person system, have inverse voice whenever the agent is
 > animate obviative third person and the patient is animate
 > proximative third person.

AFA I understand from Blake's book this is typical of
Algonquinian languages.  The question is if it is *so*
typically Algonquinian that it is unrealistic in a
non-Algonquinian language?  Though Kijeb is spoken in
another universe/timeline, so perhaps it doesn't matter.

BTW would it be unrealistic for the direct voice to
be unmarked?

 > That is, if a clause has two animate "third person"
 > participants, one of which is the "protagonist" and the
 > other a "bit player", Direct voice indicates that the
 > "protagonist" is the agent, while Inverse voice indicates
 > that the "protagonist" is the patient.

What about two *in*animate "third person" participants,
unlikely as it may seem, though you may say "the stone
hit the rock"...?

 > In such languages, Inverse vs Direct is about persons as
 > well as about genders (noun classes) or animacy.
 >
 > Languages with obviatives almost (?) all have an
 > Inverse/Direct Voice system, and languages with an
 > Inverse/Direct Voice system almost (?) all have a
 > Hierarchical "alignment" (rather than Accusative/Nominative
 > or Ergative/Absolutive or Split-S or Active/Stative or
 > Split-Ergative or Tripartite.)
 >
 > The "Obviative" is necessary in two-participant clauses in
 > which both participants are "third persons" -- that is,
 > _not_ speech-act participants - - and, especially, if also
 > both participants are animate.

Kijeb as it now is has no obviative pronouns, but it has a
distinction between proximate--medial--distal pronouns and
local adverbs which could be used in that function.

 > Grammarians of these languages basically divide clauses into ten types;
 > * I. One (or fewer) participant.
 > * I.A. The participant is a speech-act-participant (speaker
 > or addressee)
 > * I.B. The clause-participant is an animate "third person".
 > * I.C. The clause-participant is an inanimate "third
 > person".
 > * I.D. There aren't any participants in the clause.
 > * II. Two (or more) participants.
 > * II.A. Every participant is a speech-act-participant
 > (speaker or addressee)
 > * II.B. Every clause participant is an animate "third
 > person".
 > * II.C. Every clause participant is an inanimate "third
 > person".
 > * II.D. At least one clause-participant is a speech-act-
 > participant, and at  least one clause-participant is an
 > animate "third person".
 > * II.E. At least one clause-participant is a speech-act-
 > participant, and at  least one clause-participant is an
 > inanimate "third person".
 > * II.F. At least one clause-participant is an animate "third
 > person", and at  least one clause-participant is an
 > inanimate "third person".
 >
 > ---
 >
 > It is type II.C. in which obviation is especially important.
 >
 > --------------------
 >
 > eldin

Eldin also wrote in reply to Carsten:

 >
 > This is both very interesting for its own sake, and, quite
 > natlang-like in its split between the acc/nom and the
 > erb/abs parts of the system.  (BTW you are aware, aren't
 > you, that Split-S is a different thing from Split- Ergative,
 > and they occur independently of each other?)
 >
 > However, Split-S systems and Split-Ergative systems are not
 > the same as Hierarchical systems, even though the
 > "hierarchy" is intimately involved in deciding exactly where
 > the "split" in the Split-Ergative system will be.
 >
 > (In a Split-Ergative system, participants in some top
 > segment of the hierarchy are treated
 > Accusative/Nominatively, while participants in some
 > (possibly overlapping) bottom segment of the hierarchy are
 > treated Ergative/Absolutively.)
 >
 > (In a Split-S (Active/Stative) system, the single
 > participant of an Intransitive clause may be marked like an
 > Agent (ergative?), or may be marked like a Patient
 > (accusative?), depending. For instance, it may be marked
 > "Ergative" for "Active Verbs" and "Accusative" for "Stative
 > Verbs"; or it may depend on something else -- what it
 > depends on is language- specific.  If the same participant
 > can be marked either way (depending) for the same verb, it
 > is a Fluid-S system; some Fluid-S systems have _three_
 > possible "cases" or "voices" for intransitive clauses.)
 >
 > In a Hierarchical system, the verb always agrees with
 > whichever participant is highest in the Hierarchy,
 > _regardless_ of the case-role played by that participant.

I envisage Kijeb as something of a mixture, with both
Hierarchical word order and verb marking, as well as nom/acc
marking for animates, as well as Split-S/Fluid-S, and the
daughter languages (perhaps not all of them) developing
split ergative marking.  Perhaps it is altogether
unrealistic, or at least highly redundant, to have it all in
the same bag!

 > "I hit it" and "It hit me" will _both_ have 1st-person
 > agreement on the verb.
 >
 > They'll be something like:
 >
 > Hit-1-DIR I it
 > "I hit it"
 >
 > Hit-1-INV I it
 > "It hit me"
 >
 >
 > (BTW Hierarchical languages are probably not APV nor PAV,
 > nor VAP nor VPA, nor AVP nor PVA in "word-order type";
 > instead of mentioning Agents before Patients or vice-versa,
 > they always mention participants in the order of the
 > hierarchy, (or, I imagine for some languages, always in the
 > opposite order).  My examples here assume a verb-initial
 > language in which the highest-ranked participant is always
 > mentioned first.)

That makes sense.  I'll adopt that (although Kijeb is and
remains verb final).

<snip>

 > And I imagine there are other techniques as well; including,
 > perhaps, using some other alignment system (e.g. accusative
 > or ergative) just exactly in cases where both speech-act-
 > participants have roles -- but not the same role -- in the
 > clause.

I'm thinking along those lines.

 > ----
 >
 > The similarity between Direct Voice and Active Voice, and
 > between Inverse Voice and Passive Voice, is this;
 >
 > In an Inverse Voice system, the Direct Voice indicates that
 > the participant highest in the hierarchy is the Agent; while
 > in a Derived Voice system, the Active Voice indicates that
 > the participant in the Subject grammatical relation is the
 > Agent.
 >
 > In an Inverse Voice system, the Inverse Voice indicates that
 > the participant highest in the hierarchy is the Patient;
 > while in a Derived Voice system, the Passive Voice indicates
 > that the participant in the Subject grammatical relation is
 > the Patient.
 >
 > It is not clear to me whether it is reasonable to assume
 > that "the participant highest in the hierarchy" is "the
 > Subject".

According to Blake it is reasonable.  Subjects are likely
to be animate actors.

 > BTW I made (at least) two mistakes in my previous post on
 > this thread (quoted below).
 >
 > 1. There are 11, not 10, possible clause types; add:
 >
 > "II.G. At least one clause-participant is a speech-act-
 > participant, and at least one clause-participant is an
 > animate third person, and at least one clause-participant is
 > an inanimate third person."
 >
 > 2. It is "type II.B.", not "type II.C.", for which obviation
 > is most necessary.
 >
 > ----
 >
 > Also note;
 >
 > There won't be any type I.D. clauses in languages that don't
 > have "impersonal verbs";
 > and,
 > There won't be any type II.G. clauses in languages that
 > don't have "trivalent verbs".
 >

--


/BP 8^)>
--
Benct Philip Jonsson -- melroch at melroch dot se

    a shprakh iz a dialekt mit an armey un flot

                                 (Max Weinreich)

Reply