Re: CHAT: Telek nominalization

---Original Message----
Marcus Smith wrote:

> There are three main nominalizers in Telek now: subject-oriented,
> object-oriented, and oblique-oriented. A fourth nominalizer occassionally
> occurs, but it has a very specialized usage and is not productive (except
> perhaps through analogy).
>
> Subject-Oriented: -Vn
>
> This suffix is used when the entity refered to by the nominalized word
> would be the subject of the base verb, e.g., a dancer is someone who
> dances, a singer is someone who sings, etc.
>
> tele 'speak' -> tele-n 'speaker' (also the name of the people who speak Telek)
> hosy 'whisper' -> hosy-n 'whisperer'
> wifaana 'run' -> wifaana-n 'runner'
> na'ni 'cook' -> na'ni-n 'cook'
> axin 'be red' -> axin-in 'one that is red'
> igassi 'be short' -> igassi-n 'dwarf'
> ken-e 'give' -> ken-e-n 'giver'
>
> Object-Oriented: -Vm
>
> This suffix is used when the entity refered to by the nominalized word
> would be the object of the base verb, e.g., employee is the object of
> employ, gift is the object of give, etc. (Neither of these are perfect
> examples, but they are close enough.)
>
> na'ni 'cook with fire' -> na'ni-m 'that which is cooked with fire; meal
> (that was prepared with fire)'
> naali 'tell' -> naali-m 'that which is told; story'
> kene 'give' -> kene-m 'that which is given to someone; gift'
> ajlo 'follow' -> ajlo-m 'one that is followed; prey'
>
> Oblique-Oriented: -atap
>
> This suffix is used when the entity refered to by the nominalized word
> would be the argument of an applicative. Sorry, English does not have
> anything like this for use in examples.
>
> ken-e 'give' -> ken-e-tap 'recipient'
> ax-na'ni 'cook for' -> ax-na'ni-tap 'one who something is cooked for'
> ngo-wifaana 'run to' -> ngo-wifaana-tap 'place where one runs to'
> ba-na'ni 'cook with (tool)' -> ba-na'ni-tap 'something used in cooking;
> cooking utensil'
Based on what I know about other languages with applicatives--in which the
argument that is added by the applicative morphology takes on the properties of a
direct object--I would have expected "-m" to be used for these guys, giving
contrasts like:

  na'ni-m   "that which is cooked"
  ax-na'ni-m  "one who is cooked-for"

I should have realized you wouldn't let that slip by. :) This was, in fact, how
the system looked when I started working on this topic, but I quickly decided
it wasn't very interesting.

But then, I don't know how applicatives work in Telek (could we have a lesson?).

Sure, but it will have to wait for a little while.

Do applied arguments have any object properties in this language, or are they
treated as obliques?

They are treated like objects as far as verbal agreement and passivization is
concerned. However, applied arguments and regular objects have the same
possibilities in regards to incorporation (though you can't have two nouns
incorporated into one verb). On the other hand, you cannot possessor raise out
of an applied argument, but you can out of an object

Maybe your idea is that the choice of nominalizing suffix is based not on
grammatical relations (subject, object, oblique), but on semantic relations?  If
so, then you should probably rename the subject-oriented and object-oriented forms
"agent-oriented" and "patient-oriented", respectively.  Hence "-n" is added to a
verb X to form a noun denoting the agent of X, while "-m" forms nouns denoting the
patient of X, and "-atap" forms nouns denoting some non-agent non-patient
participant:

Yes and no.

 -n clearly does not refer to an agent, because it can be used with stative
verbs like _axin_ 'red' and _igassi_ 'short' which cannot have an agent. It can
also be used with unaccusatives like _bassi_ 'arrive' -> _bassin_ 'arriver'. It
can also be used with a causer (which I treat as distinct from agent).

One could probably get away with calling -m "patientive", but that is because most
direct objects are patients. I don't know of any way to distinguish between
"direct object" and "patient" in Telek at this point.

The case of -atap is where things get interesting. It is possible for an argument
licensed by an applicative to appear as the subject, especially in experiencer
verbs, where the subject is introduced by a benefactive applicative. In these
cases, the nominalizer is stil -atap. In other words, this morpheme is used for
arguments that lie between agent and patient on the thematic heirarchy.

The way I conceive of the system is that the nominalizer is chosen according to a
Paninian scale. -k is the most specific, -atap is the next, followed by -n, and
ending with -m. When deciding which nominalizer to add, you follow this
process:

1) Does this refer to a language? If Yes, add -k.
2) Does this refer to an applied argument? If Yes, add -atap.
3) Does this refer to the subject? If Yes, add -n.
4) Add -m.

One question:  Can you add these suffixes to entire verb phrases, or just verb
stems?

  I could imagine constructions like this (here I'm pretending that "maka"
means "meat", not knowing the Telek word):

  maka na'ni-n  "one who cooks meat"
  maka ax-na'ni-tap  "one for whom meat is cooked"

Entire verb phrases currently, though I'm still debating whether or not to allow
adverbs inside. To follow up on your examples (using the proper Telek word for
'meat' :)). Generally speaking, object will be incorporated into the verb,
though leaving it free is not unheard of. Not only that, but you can still get
"agreement" with an applied argument inside a nominalization.

  emfyy-na'ni-n "one who cooks meat"
  emfyy na'ni-n "one who cooks meat"
  l-ax-emfyy-na'ni-n "one who cooks meat for me"

  ax-emfyy-na'ni-tap "one for whom meat is cooked"
  emfyy ax-na'ni-tap "one for whom meat is cooked"


If constructions of this sort fit with the character of Telek, this could give you
a neat way of forming relative clauses--viz., by juxtaposing a head noun with the
appropriate nominalization:  "man one-who-cooks-meat" = "the man who cooks meat",
"man one-for-whom-meat-is-cooked" = "the man for whom the meat is cooked".

This is an interesting proposal. But I already have my internally-headed relative
clauses that I like far to much to abandon.

  You
could also use this construction as the basis for forming clefts ("It was John for
whom the meat was cooked", etc.).

I haven't thought about clefts at all yet. I will definitely consider this, but
I'm not sure how it will look in relation to the rest of the language.

Marcus

Replies