Bill,
I think you are right on. I have usually taken Householder’s “God’s
truth” approach as being like producing a table or a paradigm and looking up
the answer. The “hocus-pocus” approach is more like producing an algorithm
whereby the answer can be calculated. Sometimes one, sometimes the other is
more appropriate. I suspect that the algorithm approach would be great in
explaining the nesting or embedding structures we find in the syntax of clauses.
The table approach would probably be best in explaining the arbitrary idioms
that have been conventionalized. But that seems to leave a lot of territory in
between – “small clauses” the expressions under discussion which seem to show
attributes of both constructions and amenable to either (or neither) approach.
My book (Paraphrastic Grammar of English) has been a sad history
of me trying both approaches. I first did a relatively comprehensive analysis
of clause types. In that framework these constructions (small clauses) needed
to be analyzed as reduced clauses (never finished). Then the clause itself
(sentence) needed more careful analysis. The appropriate apparatus seemed to
be the semantic nets and their derivatives. After all, this is the direction recently
taken by Word Grammar development. Construction Grammar seems to be resting
between the two poles. Perhaps the best approach is to sit on the fence and
let the dust clear. I think the result will have to be an amalgamation of several
approaches. The interfaces between domains of Grammar abound with unresolved
issues due to framework differences and the kinds of differences between
languages mentioned by Herb. It seems that this issue may well be just that:
an interface between clausal syntax (patterns of phrases) and phrasal syntax
(patterns of words).
Bruce
From: Assembly for the Teaching of English
Grammar [mailto:[log in to unmask]] On Behalf Of Spruiell, William
C
Sent: Wednesday, December 03, 2008 1:59 PM
To: [log in to unmask]
Subject: Re: Correct?
Bruce,
It certainly ends up producing something like the same kind of
classification system that some construction grammars might produce – but not
necessarily for the same reason. A construction grammar (if I understand them
correctly) would essentially say, “These two constructions are different, and
the their difference inheres in the constructions themselves”; it’s as
if the construction, in some ways, is like a very complex version of a lexical
item. I happen to like that idea, but I know a lot of other linguists
wouldn’t. I’m trying to aim for something more like, “These constructions
are different, and for our current purposes, the fact that they’re different is
more important than getting in an argument about where the difference comes
from .” To use some dated, but still useful terms, construction grammar
operates as a kind of “god’s truth” theory; I’m just sticking to “hocus-pocus.”
The problem, of course, is that the means by which we validate
the claim that two constructions are different may themselves lock us into a
theoretical framework that’s under contention. As an extreme example, if I said
that two otherwise-identical constructions were different in that one of them
has a gap and the other has a zero-element, I’d be persuasive only to people
who agreed with me about the proposed gap and zero-element. The strategy I’m
toying with (“If I say that ‘X V-ed Y to Z’ am I saying that it’s Y who will be
doing the Z-ing?”) uses questions to check for consensus among speakers, and
then uses that consensus to validate the distinctions (unlike what happens with
the typical use of grammaticality judgments, I wouldn’t argue that the
consensus necessarily represents “what is actually true”; instead, it’s just to
achieve a workable compromise description. It’s the hocus-pocus position
again).
As another example, consider this pair that tends to come up
with some regularity on the list (well, not the Athelwulf part, but still….):
Athelwulf is easy to please.
Athelwulf is eager to please.
Both of these can be described as having the structure [S BE ADJ
INF] – but in the first one, the expectation is that other people are the ones
who please Athelwulf, while in the second, it’s Athelwulf who’s pleasing
others. There are different ways of accounting for this difference, and each of
them locks the analyst into a particular theoretical position. Regardless of that
variation, though, I think pretty much all native English-speakers would agree
that the expectations, as described above, are different. We can thus
productively (I think) talk about an “easy-to pattern” and an “eager-to
pattern” without necessarily adopting any particular statement about why
they’re different – just that they are different (again, I think a
construction grammar would say they’re different because they’re different
constructions, rather than that they’re different but we don’t really care why
right now).
And, of course, this is in practice very little different from
what traditional grammarians have done; the only change is that the methodology
is foregrounded, since it’s the methodology that’s designed to foster a
“politically viable” compromise.
Sincerely,
Bill Spruiell
Dept. of English
Central Michigan University
From: Assembly for the Teaching of English
Grammar [mailto:[log in to unmask]] On Behalf Of Bruce Despain
Sent: Wednesday, December 03, 2008 9:43 AM
To: [log in to unmask]
Subject: Re: Correct?
Your pattern, “If I say that ‘X V-ed Y to Z’ am I saying
that it’s Y who will be doing the Z-ing?” looks like what might be described in
a constructional grammar (CG). These folks are averse to describing
the relationships of constructions as built up of other constructions.
They like to contrast the usage construction meaning vs. the grammatical
construction meaning.
From: Assembly for the Teaching of English
Grammar [mailto:[log in to unmask]] On Behalf Of Spruiell, William
C
Sent: Tuesday, December 02, 2008 7:36 PM
To: [log in to unmask]
Subject: Re: Correct?
Dear All:
I suspect that one of the reasons that many modern grammars use
what seem to be simplistic structural pattern definitions (e.g. [S V DO INF]
for both “We wanted him to be hired” and “We wanted him to go home”) is that
the differences among those sentences are differences in what the various
participants are doing – the relationships among them – and we don’t really
have a theoretically agnostic way of talking about that. The minute a term like
“underlying subject” is used, the description is locked into a particular
model.
This is true of all descriptions, of course (simply by
using a label like “infinitive,” I’ve committed to a kind of model), but cases
like these bring up major points of contention among current models. Almost
everyone who works on English is happy with the term “infinitive,” but there is
nowhere near the same level of consensus about the idea that infinitives
are really, truly, made out of full sentences, etc. I have a knee-jerk reaction
the minute I see a phrase like “underlying subject,” and I’m sure I use phrases
that others on the list would have an immediate negative reaction to as well.
One way authors of grammar books can try to dodge the entire issue is
simply to omit any references to this type of material at all, and thus we end
up with [S V DO INF].
Older grammars, like the ones Herb mentions, did something that
I think we can still do: we can all agree that there are different patterns of
relationships among the participants, even if we don’t agree on why
those differences exist. To some extent, the differences among the patterns can
be “anchored” by relating them to native-speaker reactions to questions about
implications of the structure (e.g. “If I say that ‘X V-ed Y to Z’ am I saying
that it’s Y who will be doing the Z-ing?”). In other words, we can adopt
ways to probe for differences that there will be wide consensus on, even if
there is no such consensus on what the differences mean for a theory of
linguistic structure (this is what I’m trying to get at with the term
“theoretically agnostic”).
Bill Spruiell
Dept. of English
Central Michigan University
NOTICE: This email message is for the sole use of the intended recipient(s) and
may contain confidential and privileged information. Any unauthorized review,
use, disclosure or distribution is prohibited. If you are not the intended
recipient, please contact the sender by reply email and destroy all copies of
the original message.