J Matthew Pearson <[log in to unmask]> writes:

> My characterization of Tokana as active is based on the fact that the
> case-marking of the subject (insofar as it is possible to characterise a notion
> "subject" for this language) correlates with things like degree of
> agentivity/animacy and degree of involvement:

I have the impression from the current and earlier discussions that
Marcus Smith's definition of an active language contains the criterium
"Does not mark nouns for case" ;-)  At least, he insisted on the
non-existence of case-marking active natlangs in a private-mail
discussion and refuted my Georgian counterexample.

But earlier, he *did* say that Nur-ellen was active even though it marks
nouns for case!  In fact, the case marking in Tokana is quite similar to
that in Nur-ellen, as the following comparison shows:

> Absolutive subjects (undergoers):
>
>   Ne      Tsion    itskane
>   the-Abs John-Abs arrived
>   "John arrived"
>
>   Ne      Tsion    tioke
>   the-Abs John-Abs died
>   "John died"

Nur-ellen uses the objective case here, which corresponds to the
absolutive of Tokana:

Jan`n    natelent.
John-OBJ arrive-PAST

Jan`n    firent.
John-OBJ die-PAST

The first example is "fluid", it could also be formed with the agentive
to express that John's arrival is a volitional action; but the second
one can hardly be formed so (unless one wants to say that John committed
suicide).

> Nominative subjects (volitional agents):
>
>   Na      Tsion    hostane
>   the-Nom John-Nom danced
>   "John danced"
>
>   Na      Tsion    tsitspit    kopo
>   the-Nom John-Nom smashed-the pot-Abs
>   "John smashed the pot"

Nur-ellen uses the agentive here, which corresponds to the Tokana
nominative:

Jan      limpent.
John-AGT dance-PAST

Jan      dringent   i   drom.
John-AGT smash-PAST the pot-OBJ

Only animate agents can be volitional, hence only animate nouns have an
agentive case.  Is this the same in Tokana (it seems so; how can
inanimate objects have volition)?

> Ablative subjects (non-volitional animate agents):
>
>   Inaul   Tsionu   tsitspit    kopo
>   the-Abl John-Abl smashed-the pot-Abs
>   "John accidentally smashed the pot"

Nur-ellen uses the dative for agents acting accidentally, as in the
example above:

Na  Jan  dringent   i   drom.
DAT John smash-PAST the pot-OBJ

The dative agent must be animate; inanimate nouns cannot be used with
the dative preposition which governs the agentive form of the noun.

To express that someone acts under external force, one would use the
instrumental.

> Instrumental subjects (inanimate agents):
>
>   Itan     suhoua    tsitspit    kopo
>   the-Inst wind-Inst smashed-the pot-Abs
>   "The wind smashed the pot" (by blowing it off the table)

Same in Nur-ellen:

Ni   i   sul  dringent   i   drom.
INST the wind smash-PAST the pot-OBJ

> Dative subjects (experiencers):
>
>   Ine'    Tsione   hilin   ikei
>   the-Dat John-Dat saw-the dog-Abs
>   "John saw the dog"
>
>   Ine'    Tsione   kesta
>   the-Dat John-Dat happy
>   "John is happy"

Same again in Nur-ellen:

Na  Jan  tirent   i   hu.
DAT John see-PAST the dog-OBJ

Na  Jan  gelent.
DAT John be-happy-PAST

Use of the agentive instead of the dative implies deliberate
observation:

Jan      tirent   i   hu.
John-AGT see-PAST the dog-OBJ
"John watched the dog."

> Consider also the following triplet of sentences, each translated "John cut his
> finger":
>
>   Na      Tsion    hane silh
>   the-Nom John-Nom cut  finger
>
>   Inaul   Tsionu   hane silh
>   the-Abl John-Abl cut  finger
>
>   Ine'    Tsione   hane silh
>   the-Dat John-Dat cut  finger
>
> The choice of which sentence to use depends on the degree of volitional
> involvement exhibited by John:  The first sentence would be used if John cut his
> finger deliberately (e.g., in preparation for a blood-sibling ritual).  The
> second sentence would be used if John performed an action which caused his
> finger
> to get cut, but did so without intent (e.g., he was cutting food and the knife
> slipped).  The third sentence would be used if someone or something else cut
> John's finger (e.g., he was running along and accidentally grazed his hand
> against something sharp).

In Nur-ellen, the first situation would be expressed by agentive, the
second by dative, the third also by dative if it is an accident; if
someone else cuts John's finger, it would be:

Ristent  lebed      e    Jan`n.
cut-PAST finger-OBJ PART John
"[Someone] cut John's finger."

An example of the impersonal construction Nur-ellen uses instead of a
passive; more on this see below.

> I acknowledge that languages which are usually called "active" don't really work
> like this, but I don't know what else to call the Tokana case-marking pattern.

Well, up to a few minor differences, it works the same way as
Nur-ellen.  So either both languages are active, or both are not.  And
the last time I checked, there was a consensus that Nur-ellen was
active...

In a second post, Matt addresses the question of passive and
antipassive:

> Tokana does not have passives or antipassives, in the sense that you mean.

Nor does Nur-ellen.

> Viz.,
> there are no constructions which serve to promote or demote arguments, changing
> their case in the process.  Perhaps the closest thing to a relation-changing
> operation in Tokana is the reflexive, formed by adding the prefix "uma(k)-" to a
> transitive verb.  The subjects of reflexive verbs are always in the absolutive
> case, regardless of the case used for the corresponding transitive subject:
>
>    Ma    kahtin
>    I-Nom hit-Pst-him(Abs)
>    "I hit him"
>
>    Me    umakahte
>    I-Abs Refl-hit-Pst
>    "I hit myself"

Nur-ellen uses a reflexive pronoun, which does not change case marking.

> Otherwise, passive/antipassive-type meanings are expressed by simply leaving out
> arguments, as you guessed.

Nur-ellen does the same.

>       To express "John was killed", simply omit all mention
> of the agent.  There are actually two possible constructions here:  If the focus
> is on the event ("John got killed"), the regular past tense form of the verb is
> used; if the focus is on the resultant state ("John has been killed, John is
> dead"), the present tense stative form is used:
>
>    Ne      Tsion    kaihe
>    the-Abs John-Abs kill-Pst
>    "John was killed"

Nur-ellen:

Jan`n    dagnent.
John-OBJ kill-PAST

>    Ne      Tsion    kaieha
>    the-Abs John-Abs kill-Stat
>    "John is (in a state of having been) killed"

Jan`n    dag`n.
John-OBJ kill-PRES

> Eliminating the patient requires the addition of the prefix "u(k)-" to the verb
> stem.  I call this prefix an antipassive marker, but since the addition of this
> prefix does not cause any change in case-marking/grammatical relations, the term
> is not entirely appropriate.  Instead, adding "u(k)-" indicates that the
> participant which *would be* expressed with absolutive case has an
> unknown/indeterminate referent:
>
>    Na      Tsion    kaihin            Taniel
>    the-Nom John-Nom kill-Pst-the(Abs) Daniel-Abs
>    "John killed Daniel"
>
>    Na      Tsion    ukaihe
>    the-Nom John-Nom Antipass-kill-Pst
>    "John killed someone"

In Nur-ellen, the patient is simply omitted, without an antipassive
marker or anything like that:

Jan      dagnent.
John-AGT kill-PAST

Notice the difference in case to the "John was killed" example.  To
recall:

Jan`n    dagnent.
John-OBJ kill-PAST

> Notice that the agent is still marked with nominative case in the second
> sentence.  Eliminating this agent, we get:
>
>    Ukaihe
>    Antipass-kill-Pst
>    "Someone was killed"

In Nur-ellen, this is:

Dagnent.
kill-PAST

That's all!

So is Tokana active or not?  I'd say it is.

Jörg.