Re: THEORY: "Finite Verbs" vs "Non-Finite Verbs" in Languages with Poly-Personal Agreement

Thanks for your reply, taliesin!

On Sun, 16 Jul 2006 20:28:02 +0200, taliesin the storyteller <taliesin-
conlang@...> wrote:
>*Eldin Raigmore said on 2006-07-15 23:24:44 +0200
>>(I _think_ this is a "THEORY:" post; maybe it's a "USAGE:" post.)
>>(Which tag belongs on questions about linguistic terminology?)
>
>I think it sorts under THEORY. USAGE is for YAE?Ts and the like, IIRC.
Thanks for the information.

>>Question 1: In languages in which no verb is ever required to agree with
>>anything, are _all_ the verbs "non-finite"?  Or, in such languages, does
>>the "finite" vs "non-finite" distinction not exist?
>
>In Norwegian, verbs do not agree. Period. Non-finite verb-forms are then
>things like the infinitive and the participles, basically *verbforms
>that are not marked for tense* or *need an auxiliary to show tense*.
I'm used to a distinction between "untensed" forms of verbs and "non-
finite" forms of verbs.

An "untensed" form is one which is _not_ marked for tense (and/or possibly
for aspect/aktionsart, modality/mode/mood, and/or voice), when the nuclear
verbs of declarative main clauses, _including_ some other form of _this_
verb, usually _are_ so marked.

In some languages "non-finite" forms can be tensed.  "Infinitives", in
particular, can be past or present or future in some languages.  In
English, we are quite used to participles having voice and/or aspect and/or
tense, although they are "non-finite" since they needn't even have a
subject, much less need they agree with their subject.

So what I was angling for was the following question.
My best understanding to-date is as follows; is it correct?
In Norwegian there is no point in talking about "finite" vs "non-finite"
verb forms, since _all_ Norwegian verbs are non-finite.
However in Norwegian there _is_ a distinction between tensed and untensed
verb-forms.

>>Question 2:  In a language in which most bivalent or higher-valency verbs
>>usually must agree with more than one participant (i.e. a language with
>>polypersonal agreement); if a form of such a verb occurs which agrees with
>>one participant only, is that a "finite" form of that verb, or a "non-
>>finite" form of the verb?
>
>I can't remember ever having seen that finite-ness has anything to do
>with agreement so I'm at a loss to answer these... Somebody with a good
>knowledge of Basque would be the ideal person to ask though.
Google on "finite untensed" and you will see that the first three hits do
not distinguished between finite and tensed, nor between non-finite and
untensed.  Except for the fourth hit I can't be sure any of the first ten
hits do so distinguish.  So obviously you shouldn't feel alone in thinking
as you have thought up 'til now.

But the fourth hit
http://faculty.washington.edu/bcitko/LING462/Chapter%207%20slides.pdf
does claim there is a difference.
I'm not sure how well it _explains_ the difference, since that difference
is not really the topic of the document.
Look especially at its page 15; also at its pages 3, 5, 7, 16, and 17.
(It's a slide show with two slides per page.)

OTOH
http://tinyurl.com/nk52j
says:
"A finite verb is a verb form that occurs in an independent clause, and is
fully inflected according to the inflectional categories marked on verbs in
the language."

This would imply that the verb of a subordinate clause could never be
considered finite, no matter how completely inflected it is.  Also, that if
the language has poly-personal agreement, a form of a bi-or-more-valent
verb, which (the form) did not agree with more than one participant, would
be called non-finite.  Also, that when a language has no agreement at all,
whether or not a verb form counts as "finite" would in fact involve its
tense/aspect/mode/voice marking (if the language has such).

http://tinyurl.com/r6f8d
says:
"A nonfinite verb is a verb that is not fully inflected for categories that
are marked inflectionally in a language, such as the following: Tense,
Aspect, Modality, Number, Person."

I suppose that means that according to modern cross-linguistic terminology,
the "finite vs non-finite" distinction incorporates the "tensed vs
untensed" distinction if the latter exists; and doesn't involve number and
person in languages in which verbs don't inflect for number and person.

I was basing my questions on the etymology of, or historical meaning of,
the terms "finite" and "non-finite".  Presumably these were first used to
describe what happens in languages (such as Latin and Greek) in which non-
finite verb forms don't agree in number and person with their subjects, and
finite verbs do; and verbs which agree in number and person with their
subjects are finite, and verbs which don't are non-finite.

Right now I can't find that "historical meaning" on-line.  Sorry.

>>So why the heck am I asking?
>>I am conning a clause-chaining lang.
>       ^^^^^^^
>Ooh, widening of meaning or more likely
>backformation.
Yup.  Back-formation.

>Yummy. Dangerous word to use outside of our circle though.
>A: "What, are you a con-man?" <grabs for phone to call police>
>B: "Yessir, I love conventions! I'm off to a rock-collector's next
month..."
>A: "Uh..." <looks for dictionary>
:-)

[snip]

>>I want my lang to have a switch-reference system, too.
>>In a switch-reference system, some verbs are obligatorily marked to show
>>whether some referent is the same as, or different from, a similar
>>referent of a reference clause.
>
>AFMCL, it does switch-reference, or basically has a prefix to show same
>subject as previous clause in a clause chain. If the subject is
>different, it needs its own NP. Also, it's the only time the verb agrees
>with anything... if we disregard type 4 object-incorporation as
>agreement.
Ooh!  I wanna see!  I wanna see!

[snip]

>Was this the end of your mail
Yes, it was.

>or was it cut somewhere for being too long?
Well, yes, it was cut, but I'm the one that cut it (before posting it).
I had thought of including material like the following in the original
post; but I decided it both would make it too long and didn't really belong
on that post.

However I think it might now belong _here_ as a reply to some things you
said in _your_ post.

A clause may be dependent on another clause without being embedded in it;
and a clause may be embedded in another clause without being dependent on
it.  In case a clause is both dependent on and embedded in another clause,
it is called _subordinate_ to that other clause, which is called its
_matrix_.

=== [[Relative Clauses]] ===

One common type of subordinate clause is the _relative_ clause.  These are
clauses used as adjectives to modify a shared referent (shared with the
matrix clause, that is); the shared referent is always the _topic_ of the
relative clause.

So if a switch-reference system applies to relative clauses, it makes sense
for;
o the marked clause to be the subordinate clause;
o the reference clause to be its matrix clause; and,
o for the switch-reference marker to indicate when the _topic_ of the
_marked_ clause is _any_ participant of the _reference_ clause, and which
participant it is.

=== [[Long-Distance Anaphora]] ===

Another thing which happens in many languages is "long-distance anaphora"
or "long-distance reflexives".  In these, a deeply-embedded subordinate
clause -- such as a subordinate clause of a subordinate clause, that is, a
clause whose matrix is itself subordinate to yet another clause -- can
refer (via a LDA or LDR) to the subject of _any_ superordinate clause, from
its matrix, to its "grandmatrix" (its matrix's matrix), etc. all the way up
to the main clause.

So if a switch-reference system applies to subordinate clauses that use
long-distance anaphora, it makes sense for;
o the marked clause to be the subordinate clause;
o the reference clause to be the main clause; and,
o for the switch-reference marker to indicate when the _subject_ of the
_reference_ clause is _any_ participant of the _marked_ clause, and which
participant it is.

=== [[Logophoric Pronouns]] ===

If a sentence is quoted as a constituent of another sentence, some
languages have "persons" and/or "person-markers" or "pronouns" for the
Illocutionary Source (the guy that said it first) and, often, also for the
Illocutionary Target (the guy he originally said it to).  (If such a
language also has obviatives, I wonder whether there might not also be an
Illocutionary Protagonist (the guy he was originally talking about)?)

These are what is usually meant by Logophoric Pronouns; and the one for the
Illocutionary Source is the one that occurs in more languages than the
others.

A quoted sentence is an embedded clause that is not necessarily a dependent
clause.  In some languages with logophoric pronouns, the use of these
pronouns spreads from embedded clauses which are complements of verbs of
saying or thinking, to nearly any and all embedded clauses.

In some such languages, there are pronouns by which an embedded clause can
refer to any participant of its matrix clause.  These are also called
logophoric pronouns, as near as I can tell.

So, if such a language has a switch-reference system that applies to
embedded clauses, it makes sense for;
o the marked clause to be the subordinate clause;
o the reference clause to be its matrix clause; and,
o for the switch-reference marker to indicate when _any_ participant of the
_marked_ clause is _any_ participant of the _reference_ clause, and which
participant it is.

===

Now, that last would be a bit much; I am not aware of any switch-reference
markers other than;
SameSubject vs DifferentSubject
SameObject vs DifferentObject
SameLocation vs DifferentLocation

And, as I said, the SameSubject vs DifferentSubject marker is usually
extended to mark one of the following;
Marked clause's Subject is same as which participant of reference clause
Or
Reference clause's Subject is same as which participant of marked clause.

====

>There was no sig(nature) in the mail I received.
Huh!?!  There _should_ have been -- I wonder why there wasn't?

>Thanks for the switch-reference overview as well!
Thanks for your interest!

-----
eldin

Reply