Back from house-buying on the Virginia Peninsula (success!), let's chat a bit
more about the notion of verb valence.

Robert Einarsson wrote on 3/12:

>The post from Rebecca Wheeler is very interesting  -- I gather from
>what she says that "valence" is a potentially very sophisticated
>system of classifying verbs according to the syntactical constraints
>that they carry.
>

Robert asks whether:

>the concept of "valence" simply a more developed, sophisticated
>application of  the concept of "transitive or intransitive;"  or is
>"valence" a radically different concept of verb syntax?


There are some real commonalities between verbs being categorized as
intransitive/transitive on the one hand, and as being of a certain "valence"
on the other.

Obviously, the transitive/intransitive distinction has to do with whether  or
not there is an NP following the verb.(though that muddies the terrain some
when you have linking verbs, as in 'she became president').

Similarly, a valence description of a verb lays out the phrasal configuration
that that verb characteristically evokes.

So, both the transitive/intransitive distinction and a 'valence model' look at
what other structures are in orbit around the verb, as it were.

Notice, however, that the intransitive/transitive distinction seems to focus
simply on whether or not we find an  NP or two following the verb. And yet, NP
or no NP seems a rather blunt instrument given regularities of pattern we can
readily observe.

Again, quoting from Morenberg (Doing Grammar) as a handy, accessible work (p.
14), we see 6 verb types. Notice that 'intransitive' and 'transitive' do
figure in here,as two of the broad 6.

        1.  NP:Subj + Intransitive Verb (+/- adv)
        2.  NP:Subj + Linking Verb + {Adjective Phrase:Pred.adj}
                                     {NP:PredN}
                "Jami seems happy"
                "Dakota became top dog"

        NOTE: NP:PredN translates
                as "Noun Phrase functioning as/doing the job of a
                Predicate Noun"

        3.  NP:Subj + Transitive Verb + NP:DObj
                "Bill bakes cakes"

        4.  NP:Subj + Vg {NP:IObj + NP:DObj}
                (with variation in order of IO/DO)
                GIVE-VERBS
                "Mika sent/gave Jen a package"

        5.  NP:Subj + Vc + {AdjPh:Object Complement} or
                           {NP:Object Complement}  or
                           {InfinPh:Object Complement}
                "John considered the mountains amazing"
                "They painted their house aubergine"

        6.  NP:Subj + BE {AdjP: PredAdj} or
                         {NP: PredN} or
                         {Adv of Place:PredAdv}
            (note -- Morenberg argues for BE as a class
             of verbs distinct from linking verbs)


Quite aside from whether one would put BE in the category of linking verbs or
not, the basic point remains that we can say a great deal about the behavior
of verbs depending on which type they are instantiating at a particular
moment. (notice, that verbs can belong to multiple categories, depending on
the syntactic environment in which they occur as below:)

        intransitive:           Miki painted energetically.
        transitive:             Miki painted the house.
        "give-type" verb:       Miki painted Joni a special picture.
        "consider-type" verb:   Miki painted the house black.

Of course, different verbs will have different capacities for appearing in
this or that "syntactic frame" (NP V NP AdjP    etc.). so, 'sleep' will not
tend to occur in frames like 'give' verbs or like 'consider' verbs. Hmmm....
so much does 'sleep' not occur in these environments, that I can't even think
up a relevant "bad/ungrammatical/nonoccuring" example.

But the reason is that the MEANING of "sleep" invites us to conceptualize
someone doing it -- the being denoted by the NP functioning as subject. And
then you might say how someone is sleeping with some optional adverb, but
that's about it for 'sleep'.

The key insight is that the MEANING of the verb compels/invites/specifies a
range of different associated players (NP/AdjP/AdvP). So, the meaning leads
the syntactic way.

Inversely, given a syntactic environment  (NP V NP2 NP3: Jami __  Sue a book),
we KNOW that the verb will be of some conveyance meaning -- sold, gave, lent,
read, etc.

Interestingly, while that syntactic frame COULD key us into 'consider' type
verbs (NP V NP2 NP3/Adj:  John considered Bill president), the meaning of
"sue" and "a book" tells us that that verb slot COULD NOT be filled by a
'consider' verb.  That is, we would not have "Jami considered Sue a book",
because there is a very specific relationship between NP2/NP3 in that frame:
NP2 = NP3 -- to put it simplistically.  And while it could be that Bill =
President, it can't be that Sue = book.

(yikes! what did all THAT mean? that meant that verbs tell us what kind of
relationships hold between the verb and its satellites, and between the
satellite phrases. So, if we see two NPs, like "Sue" and "Book", as NP2/NP3,
we KNOW the verb can't be a 'consider' verb, because consider requires NP2 =
(in some sense) NP3, and Sue just can't BE a book)

These are some dividends that we get from the 'valence model' that we don't
get from the simple transitive/intransitive model. The dividends are ones of a
finer-grid delineation of patterns we find in our language.

As far as definitions go, here's a rough and ready one:

The valence treatment of verbs takes the verb as the semantic and syntactic
core of a sentence. To quote Morenberg again, "verbs determine the other
components in sentences and define the relationships among those
components"(5). -- just like we saw in the example above with 'paint',
'consider' and 'sleep'. This insight lets us identify a small number of
fundamental types of sentences based on this verb + satellite structuring.

On the grammar/writing front, I have found these insights helpful in my frosh-
comp classes.  Often, for example,  I will find that students have constructed
nominalization ('realization', 'inspiration', 'discovery', etc.) upon
nominalization in a sentence, sometimes ending up with 4 or 5 in a single
sentence. (don't have an example on me this sunday night). Then the student
will get lost in what they are trying to say,where they are trying to go.

So, I'll (without naming it), call on the verb + satellite insight of Valence
work. I'll ask them to circle all the words that are nouns that started out
life as verbs. (we'll talk about that a bit). One trait of nominalizing is
that you can take a verb and its satellite arguments, and blot out one or more
of the satellites. here,meaning gets lost, and students might get confused.
thus, in an example like

"The realization led to immediate action."

I will ask students to find the verbs that correlate to 'realization' and
'action'.

I write them on the board.

Then I ask the students how they would make a simple sentence with 'realize'
-- like what ELSE would be in the sentence besides that verb.  And they come
up with something like "We realized that we were early" or "we realized how
lucky we were" or "we realized our good fortune."

I point out that in our example sentnce (The realization led to immediate
action), we've somehow LOST all that other information that tends to revolve
around the verb 'realize'.... and so we look to put it back in, and create a
sentence that recaptures the lost info, or at least what is needed for the
passage.

In his email of 3/12 Robert continued:

>The basic concept is the same as "transitive or
>intransitive."  Valence would therefore be a valid and sophisticated
>extension of a traditional concept, which would be a conclusion
>satisfying to me personally.

I would say it is an extension of a traditional concept.

>
>I looked up "valence" in the dictionary:  it is a word from chemistry
>that means "the unique quality of an atom that determines how many
>other atoms that it can combine with."
>

Nice analogic grounding for its use in grammar!


>Why should the qualities of the _verb choice_ determine
>the structure of the _predicate_ as a whole?

It's really more that the qualities of the verb determine the structure of the
SENTENCE AS A WHOLE. While most of the divergence happens after the verb,
there are nonetheless consequences for the subject position too.

Thus, take the sentence,

        That interest rates will rise seems likely.

"Seem" and some other linking verbs are receptive to their subject slot being
filled by a clause of the form "that interest rates will rise"...  In turn, we
can flip an 'it' in there and get "it seems likely that interest rates will
rise.

But other verbs do not permit such a subject.  Thus, while we do have
"That interest rates will rise gave Bill the shivers." we don't have
"*That interest rates will rise considered John the best president."

And that's because, 'consider', at least, selects for an animate subject (here
we are operating at an even finer grain.).

But Robert's question is a deep one: Why should the nature of the verb
determine the syntax of the sentence as a whole? and further, he asks "Why
should _word_ qualities have an effect on _syntax_?"

These are huge, huge questions, and this email is too long already.



>
>This to me is a very interesting idea that possibly collapses the
>separation between the lexcial level and the syntactical level.

we don't need to go that far. there are clear structural differences between
what happens with words, and what happens within syntax.

Enuf is enuf here.

I hope that this email has laid out a bit about the 'valence model' of
analyzing sentences.  Actually, I like the image of the verb and its satellite
phrases.

ciao,

rebecca wheeler


(reply to [log in to unmask] -- my university has bombed out our ability
to SEND email out under any but a single communication package.... humph!)