LINGUIST List 7.777

Tue May 28 1996

Disc: Ungrammaticality vs. Semantic Anomaly

Editor for this issue: Annemarie Valdez <>


  1. benji wald, ungrammaticality vs. semantic anomaly

Message 1: ungrammaticality vs. semantic anomaly

Date: Tue, 28 May 1996 02:46:00 PDT
From: benji wald <IBENAWJMVS.OAC.UCLA.EDU>
Subject: ungrammaticality vs. semantic anomaly

Guy Modica raised some important issues in his reraising of the issue
about the problem of how to evaluate, or analyse:

 the books sells well to raise money

I, and I assume, most other readers understand what he means when he
asks whether the problem with this type of sentence is
"grammaticality" or "semantics". However, I think we understand that
because we understand that he is opposing "syntax" to "semantics" in
this context. Furthermore, we are familiar with theories which
distinguish syntax from semantics. However, if we keep thinking about
that, we recognise that different theories distinguish syntax from
semantics in different ways, so that what is "syntax" for one theory
may be "semantics" for another ("and vice-versa" obviously goes
without saying). That's one problem, and I think the particular
example here is an excellent illustration of that problem. I myself
am not sufficiently committed to any particular theory, or have a
sufficient grasp of the implications of the example, to come down on
any side, i.e., that it's syntax and not semantics, or vice-versa, or
that it splits into semantic and syntactic aspects. That, I thought,
was where the late list discussion of "ungrammaticality" might have
led, esp with various respondents proposing that they had various
theories which would shed light on the problem/concept of
"ungrammaticality". But things did not go very far in that direction.

What I just said is related to what I said in agreeing with Karl and
Alexis in an earlier posting in saying that "grammaticality" is a
theoretical notion, and the theory defines what is "grammatical" and
what is not -- as opposed to how speakers "feel" about certain strings
of words, and whether or not the theory is also supposed to account
for that.

Meanwhile, there is another point, which is terminological. "Grammar"
often refers to BOTH syntax and semantics (and even phonology in some
uses of the term "grammar"), so that to say something is
"ungrammatical" does not yet identify what's "ungrammatical" about a
particular type, e.g., syntax or semantics, or both, and then more
specifically what IN syntax or semantics or both causes the impression
that there is something "wrong" with the example. So, it is futile to
argue about whether the source of an ungrammaticality impression (or
"judgment") is syntactic or semantic until we understand what that
means in a particular theory.

Does that mean we can't discuss the problem without agreeing on a
theory first? No. I hope not. And maybe the reason for that is that
we agree on enough, regardless of particular theory, to consider what
other facts might be relevant to the problem. (Maybe we even agree on
MORE when some particular theory is not at stake.)

I don't really want to pursue that unless pressed. I'd rather just
suggest that we can still bring to bear whatever considerations we
think are necessary to see if we can get some insight into the
problem. All we have to do is agree there's a problem. That's
sometimes difficult in itself, especially if we don't think that
there's anything wrong with the example, just somebody else does.
Then is it their problem but not ours? I leave that up for grabs. I
have an opinion, but I'll defer it.

To try to illustrate what I'm saying I'll take an initial stab at the
example. But, as I already said, I don't understand what's going on it
well enough to reach a conclusion. I just want to make a start (and
there is a body of literature to take it much further):

To begin with, then, we have:

They (can easily) sell the books
The books sell (well)

The example seems to be predicated on some kind of relation between
the above two sentences (taken as sentence types). There are both
syntactic and semantic differences between those two sentences -- by
anybody's theory (I expect -- except that some theories might deny
there is such a thing as syntax, which probably means what some
theories call syntax such theories would somehow present as some kind
of semantics. I thought, for example, Shaumjan's theory might be
leading to that type of position -- but I'm not sure. He recognises a
syntactic level, according to his diagram, but it seems to be a
special kind of semantics.)

In any case, by "analogy" we have

They (can easily) sell the books to raise money
The books sell (well) to raise money

The mechanical operation implied by "analogy" suggests that the
problem is "syntactic". But we already acknowledged that whether the
relation between the first two sentences was semantic or syntactic or
both is problematic. And, furthermore, regardless of terms, we were
not precise about just what the problem is. The best we can do so
far, if we can do anything, is agree that the last sentence is "weird"
("uncomfortable?"). From that point we can go in all kinds of
directions. For example, we can test if it's only purpose clauses
(phrases) that contain the problem, or whether there are other kinds
of "adjuncts" that also have it. (The more of those we find the more
we can veer toward syntactic aspects of the problem.) Or we can fix
on the problem of "well", which doesn't seem (to me) so good for the
transitive sentences, e.g.,

they can sell the books ??well (to raise money)

(Even without "to raise money" "well" doesn't seem "felicitous", to
say the least, in the transitive sentence. On the other hand, "these
books sell *easily*" doesn't seem so bad to me. -- And I am aware that
there are geographical differences in the acceptability of this
"syntactic" pattern, in terms of productivity across verbs.)

I'm not going much further in considering the example here, except to
note that, though there can be disagreement about the following, one
of my questions about the problematic sentence is what "can" it mean?
And why/how do we construe it to mean "that"? So if it means
something but that's not the "right" way to say that "something", what
is? EG

the books have proven/ successful vehicles in/for our fund-raising
 would prove

OK. That's long-winded. Is there a simpler way of saying that, that
doesn't upset anyone's linguistic sensitivities (in English)? EG

the books are/can be money-makers//good(/**well) sellers to raise $.

selling the books is/would be a good way to raise money

In fact, I don't think any of the above are successful paraphrases of
the "entirety" of what the problematic example "might" mean, but I
think attempted paraphrase is something that has to be considered. My
own impression, and it may be strictly my own, is that the sense of
the example would have to be something like:

they might sell well to raise money (for some charity), but if
you don't attach them to some appealing fund-raising effort they
won't/don't sell well at all.

CF (?) These cookies sell well to raise money for the Girl Scouts
 (but who wants to eat them? Give 'em to the dog)

I wouldn't be surprised if some other English speakers bridle at this
suggestion, but, for those who don't, it seems to me that there is a
"scope" issue involved, such that "to raise money" qualifies "well",
or mitigates it as a condition on "sell". If you do agree, it still
seems to me problematic how much such a construal owes to semantics
and how much to syntax.

There are, then, to my way of thinking quite a few pretheoretic
questions we can ask about the problematic sentence. Does it mean
anything (in its entirety)? What? How do we get that interpretation
(those interpretations) from the example? What's a "better" way to
say it (if the example's no good, but its intent is clear)? All these
seem to me to be valuable data toward deciding how to analyse the
problem and what's wrong with the example (if anything). That may
even help us figure out, once various theories are applied to
accounting for the example, and whether it's odd or not, whether the
various theories are, in fact, what used to be called NOTATIONAL
VARIANTS -- or whether there are genuinely substantive differences
among the theories in terms of further predictions they make and what
data they incorporate into the theory as relevant to explaining the
effect of the problematic example. What else can you ask for? --
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue