[LONG] Gweydr Raising and Control Verbs (Was: HELP: Subject Raising Verbs, ECM Verbs, etc.)

As promised, now that I'm back in San Diego and have my
materials with me, I'll post Gweydr's method of dealing with
raising and control verbs.

A very small percentage of people may remember the Gweydr
language.   It was the second I ever attempted, and I posted
something on it almost--wow, three years ago.   I decided to
revisit that project, though I only kept (most of) the orthography,
and (most of) the vocabulary.   Everything else has changed.

Gweydr is an SVO language, and each verb inflects for person
and number, as well as tense.   There are four different classes of
verbs, and these directly affect how raising and control issues
work.   I'll start with raising verbs then move onto control verbs.

***

(1) Subject Raising Verbs:

The most common example of a raising verb in English is "seem".
While such a verb may or may not exist in Gweydr, I do know
how it would work, so for the present purposes, I'm just going
to pretend that there's a verb /sovaS/, which means "seem", just
like in English. [Note: /-aS/ is an "infinitive" ending.]

(1a) sof iziS
/seem-1st.pres. redden-inf./
"I seem to be turning red."

This is pretty much like English.   One thing to note about subject
raising verbs is that since they assign a single theta role, and since
this theta role is assigned to the complement, they can often take
an expletive, or dummy subject.   This can be seen in English with
"It seems (that) I'm turning red."   There are no expletives (at least
those kind) in Gweydr, overtly, but the same effect can be seen,
where the verb is followed by an appositive conjugated in the
third singular:

(1b) is soft
/redden-1st.pres. seem-3rd.pres./

What you don't get that you do get in English is the phenomenon
shown in a sentence like, "There seems to be a fly in my soup".   In
Gweydr, this would simply be "There's a fly in my soup, it seems."

***

(2) Exceptional Case Marking (ECM) Verbs:

ECM verbs are verbs that seemingly mark the argument directly
following them with a particular case.   The case marking for Gweydr
ECM verbs doesn't seem very exceptional, though, because all ECM
verbs are experiencer verbs, and all verbs of experience assign the
nominative case to their objects.

Keeping that in mind, ECM verbs (and the rest, in fact) have two modes.
The first mode is for when the exceptionally marked part is an entire
clause, and the second mode is for when the exceptionally marked
part is simply an argument.   Here are some examples:

(2a) T&l mataZal ragol.
/want-1st.pres. nom.-see-inf.-2nd.sg.possessive part.-maple tree/
"I want you to see the maple tree."

In the above, the infinitive "to see" is treated as the direct object of
the verb "I want".   Any objects of a verb in the infinitive are put
into the partitive, regardless of class.   Additionally, the subject of the
embedded verb is the possessor of the embedded verb.   This strategy
works fine when you know who the subject of the embedded verb is,
and when you don't have a long embedded clause.   When the subject
of the embedded verb isn't known, though, and thus can't be easily
expressed pronominally, or when the embedded clause is too long,
there is a second strategy shown below:

(2b) T&l palinu mat gol.
/want-1st.pres. nom.-animal see-3rd.pres. nom.-maple tree/
"I want the animal to see the maple tree."

Above, even though the verb /mataS/, "to see", assigns the partitive
case to its subject, the semantic subject, the animal, is marked with the
nominative case, which is the case that the matrix verb /T&l/, "I want",
assigns to its object.   Additionally, the verbs must agree in tense, even
if the semantics don't match up (different tenses do different things for
different types of verbs).   So, if (2b) were to be put into the past tense,
it would be:

(2c) T&li palinu m&ti gol.
/want-1st.past nom.-animal see-3rd.past nom.-maple tree/
"I wanted the animal to see the maple tree."

Another way to do it is similar to the raising verbs.   The "I want" part
can simply be added at the end as an appositive statement.   In this way,
the tenses don't have to agree:

(2d) rapalinu mattaza gol, T&li.
/part.-animal see-3rd.fut. nom.-maple tree, want-1st.pres./
"I want for the animal to see the maple tree (some time in the future)."

***

(3) Subject Control Verbs:

Subject control verbs are like ECM verbs, in that a good many of them
are verbs of experience.   Some are not, though.   For those that are, the
relationship is as follows:

ECM:
(3a) T&l mataZal ragol.
/want-1st.pres. nom.-see-inf.-2nd.sg.possessive part.-maple tree/
"I want you to see the maple tree."

Subject Control:
(3b) T&l mataZam ragol.
/want-1st.pres. nom.-see-inf.-1st.sg.possessive part.-maple tree/
"I want to see the maple tree."

The subject of both the matrix and embedded clause are identical.
This works out very well with the first and second person.   When
you get to the third person, the distinction between an ECM verb
of experience and a subject control verb of experience becomes
clear:

ECM:
(3c) T&lt mataZak ragol.
/want-3rd.pres. nom.-see-inf.-3rd.sg.obv.possessive part.-maple tree/
"She wants her (different person) to see the maple tree."

Subject Control:
(3d) T&lt mataZaS ragol.
/want-3rd.pres. nom.-see-inf.-3rd.sg.possessive part.-maple tree/
"She wants to see the maple tree."

An ECM verb requires there to be two different subjects, whereas a
subject control verb requires there to be two identical subjects.   In the
third person, there is an obviative pronoun (what I'll call it for now) which
introduces a new third person argument.   This pronoun can never be
coindexed with a previous third person argument that mumble mumble
properly governs mumble...   Ahem.   Suffice it to say, the non-obviative
pronoun can be bound; the obviative one can't.

Subject control experience verbs can also appear in an appositive
construction,
like those above.   Here's an example:

(3e) mataza gol, T&li.
/see-1st.fut. nom.-maple tree, want-1st.pres./
"I want to see a tree (some time in the future)."

This is kind of like the seldom used construction, "I want that I should see
a tree."
(Well, seldom used round my parts.)

Another type of subject control verb can take an object as an
argument.   A quick example of one of these is with the verb
"promise", which I'll just kind of make up for the purposes of
this example: /grumaS/.

So, you can get the normal kind of subject control:

(3f) grum ramataZam ragol.
/promise-1st.pres. part.-see-inf.-1st.sg.possessive part.-maple tree/
"I promise to see the maple tree."

The only difference above is that "promise" is a performative verb,
and performative verbs assign the partitive case to their complements.
Thus, the embedded verb phrase is marked with the partitive case.
However, you can also promise someone to do something.   This is
what that looks like:

(3g) grum klij ramataZam ragol.
/promise-1st.pres. term.-2nd.sg. part.-see-inf.-1st.sg.possessive part.-maple
tree/
"I promise you to see the maple tree."

In (3g), a second person argument has been added.   Since it's the direct
object
of the verb "promise", it gets marked with the terminative case (a kind of
combined accusative and dative, plus some, well, terminative).

Again, these can be rendered with an appositive construction.   The only
difference from the previous ones is that the direct object tags along:

(3h) mataza gol, grum klij .
/see-1st.fut. nom.-maple tree, promise-1st.pres. term.-2nd.sg./
"I promise you (that) I will see the maple tree."

***

(4) Object Control Verbs:

Now for the grandaddy of them all.   Object control verbs actually
assign a theta role to the subject of the embedded clause.   A common
example is "persuade".   When you persuade something, you actually
affect it, as opposed to when you want someone to do something.
That's the semantic difference.   So, for "persuade", again, I'm just
creating a verb for the example: /prystyl/.   (Or perhaps it should be
/astenel/?)

Object control verbs are always transitive verbs in Gweydr.   Transitive
verbs assign the partitive case to indefinite direct objects and the
terminative case to definite direct objects.   Similarly, these two cases
are used to determine the degree of success of an object control verb.
So, for example:

(4a) prysty kimataZal ragol.
/persuade-1st.past term.-see-inf.-2nd.poss. part.-maple tree/
"I (successfully) persuaded you to see the maple tree."

(4b) prysty ramataZal ragol.
/persuade-1st.past part.-see-inf.-2nd.poss. part.-maple tree/
"I didn't manage to persuade you to see the maple tree."

One difference between subject control verbs and object control
verbs is that while you *can* have the same subject for both the
matrix and embedded clause for both sets of verbs, they yield
different results.   Subject control verbs render a meaning like,
"I want to see the tree", whereas object control verbs render a
meaning like, "I persuaded myself to see the tree".   Here are those
two sentences.

Subject Control:
(4c) T&l mataZam ragol.
/want-1st.pres. nom.-see-inf.-1st.sg.possessive part.-maple tree/
"I want to see the maple tree."

Object Control:
(4d) prysty kimataZam ragol.
/persuade-1st.past term.-see-inf.-1st.poss. part.-maple tree/
"I (successfully) persuaded myself to see the maple tree."

Additionally, if you look at third person arguments, you can now
have *both* an obviative and non-obviative pronoun as the subject
of the embedded clause.   If it's an obviative pronoun, the reading
must be that there's two subjects.   If it's non-obviative, the reading
must be that the two subjects are identical.   Here are the two
sentences in question:

Obviative:
(4e) prysty kimataZak ragol.
/persuade-3rd.past term.-see-inf.-3rd.poss.obv. part.-maple tree/
"She (successfully) persuaded him to see the maple tree."

Non-obviative:
(4f) prysty kimataZaS ragol.
/persuade-3rd.past term.-see-inf.-3rd.poss. part.-maple tree/
"She (successfully) persuaded herself to see the maple tree."

Like ECM verbs, if there's a non-pronominal subject, or if the
complement clause is too long, there's another way to use an
object control verb.   This is shown below:

(4g) prysty kipalinu m&ti gol.
/persuade-1st.past term.-animal see-3rd.past nom.-maple tree/
"I persuaded the animal to see the maple tree."

This method is dispreferred, though, because you lose the info
about the success of the persuasion.   What you gain, though, is
the ability to express the full argument, and the ability to indicate
whether that argument is definite or indefinite.   [Note: You can
emphasize the success of the venture with an emphatic particle,
though.]

Finally, though not common, an object control verb can appear
in an appositive construction.   It's not common because the object
must be introduced at the very end, and can't be omitted.   Here's
an example:

(4h) mattaza gol, prysty kipalinu.
/see-3rd.fut nom.-maple tree, persuade-1st.past term.-animal/
"I persuaded the animal that he should see the maple tree (in the future)."

***

Okie doke, that's it.   I'm sure this system is not without its flaws.   If
you
spot any, let me know.

-David
*******************************************************************
"sunly eleSkarez ygralleryf ydZZixelje je ox2mejze."
"No eternal reward will forgive us now for wasting the dawn."

-Jim Morrison

http://dedalvs.free.fr/