LINGUIST List 5.1380

Fri 02 Dec 1994

Disc: Native speaker intuitions

Editor for this issue: <>


  1. Michael Kac, Intuition and Grammaticality
  2. "Margaret E. Winters, OF as a verb

Message 1: Intuition and Grammaticality

Date: Mon, 28 Nov 1994 19:50:58 Intuition and Grammaticality
From: Michael Kac <>
Subject: Intuition and Grammaticality

I was off the list when Marilyn Silva solicited the comments reported
in her recent summary; I had intended to respond to some of them,
but then lost the message through a blunder too embarrassing to
make public. Since then a couple of other postings have come in
which are relevant to what I wanted to say, so here it is.

One thought is relevant specifically to the control properties of *ask*.
I am among those for which this is a rigid object control verb if an
object is present (that is, for me, in *X asked Y to Z* only Y can be
construed as the complement subject). I have no doubts, however,
that there are genuine cross-speaker differences on this point; I can
remember, for example, a conversation with a fellow graduate
student more years ago than I care to divulge having to do with
exactly this point. Although I *am* a syntactician, nothing that I hold
dear hinges on this particular example, so I think my judgement can
be taken at face value; and I share Wayne Lawrence's skepticism
about the effects of 'intuition fatigue'. My judgements are no less
firm for having thought a lot about examples like these.

I am also apparently immune to the alleged influence of pragmatic
factors, the force of which may, I think, have been overstated.. The
following pair of examples has been cited as evidence for the claim that such
 factors may influence interpretation:

 [1] The teacher asked the child to leave the room.
 [2] The child asked the teacher to leave the room.

To me, the only pragmatic effect discernible in regard to (2) is that
the sentence reports a somewhat improbable occurrence -- it doesn't
cause me to suddenly get a reading of a kind I was unable to get

Indeed, I wonder if there isn't a danger here of confusing two
different phenomena. One is the existence, for some speakers, of a
genuine indeterminacy in regard to the control properties of *ask* --
an indeterminacy which would be clearly evident in sentences like
*Sandy asked Pat to leave the room*, where no power differential
between the individuals mentioned is implied; the other is the
possibility that a hearer might, given an example like (2), infer from
the presumptive oddity of the situation reported that what the
speaker *intended* to say was that the child sought permission to
leave the room, even though such an interpretation is not actually
possible for the hearer. I wonder if that's not what lies behind Ted
Harding's comment -- it would be interesting to know his response.

The sort of thing I've just described *does* happen; here is an
example from my own experience, though I suspect that there will
be LINGUIST readers who respond as I did.

In Josephine Tey's novel *Brat Farrar*, a young man -- the Brat
Farrar of the title -- who has grown up in an orphanage and closely
resembles the long missing twin brother of the heir to a substantial
fortune -- agrees to participate a swindle in which he is to
impersonate the missing twin (who, by virtue of being slightly older,
is the true heir). When Brat first meets Simon, the surviving twin,
Simon's immediate response is a look of relief -- a source of some
puzzlement to Brat at the time. Later, Brat learns that the
disappearance of Simon's brother is due to Simon's having murdered
him. At which point the following sentence -- or something very
much like it -- occurs (I can't vouch for the exact wording, but the
following will nonetheless make the point): "Brat understood the
look of relief that had crossed Simon's face as soon as he had gotten a
good look at himself".

Now, I cannot -- even when reading as masterful a writer as
Josephine Tey -- get anything but *he* (and, derivatively, *Simon*)
as the antecedent of *himself* in that sentence; at the same time, it's
completely obvious in the context of the story as a whole that she
intends the antecedent to be *Brat*. The only conclusion I can draw
is that reflexive anaphora in this type of case is clause-bound for me
but not for her, marvel at this curious fact, and read on.

Another reason I'm skeptical about claims having to do with effects
of pragmatic factors comes from the fact that it's possible, in
ambiguous sentences, for the sense that's recognized first to be the
one which is *inconsistent* with pragmatic factors such as world
knowledge. Anyone who, like me, is a lifelong reader of The New
Yorker, knows that this phenomenon is quite real. For example, I
recall a New Yorker newsbreak which quotes an article about
Princess Anne's showing in an equestrian event. "The daughter of
Queen Elizabeth and her horse came in third", the article dutifully

A more general concern of mine has to do with the presumption that
the data of syntax are intuitions or judgements. They are
unquestionably among the data of the study of *syntactic
processing*; but that hardly exhausts the study of syntax. Indeed,
you can't even study syntactic processing intelligibly without
recognizing that there's a difference between what's *actually*
(un)grammatical and what speakers *judge* to be (un)grammatical --
sometimes erroneously. It's common, for example, for speakers to
wrongly judge sentences like *The rat that the cat that the dog bit
chased squeaked* as ungrammatical; similarly for certain types of
garden path sentence (though I'll go on record as of the persuasion
that the well known center-embedding cases involve garden pathing
too -- but that's another story for another time).

A parting shot: I recall from Silva's summary a reference to
'introspection' as a source of data about grammaticality.
Introspection is not what's involved here -- a point, I might add, on
which there's persistent confusion but on which the early generative
literature is very clear. It's precisely because the content and
structure of the grammars of natural languages are *not* accessible
to introspection that we are forced to rely on indirect, inferential
methods; admittedly, these methods are typically (and in my view,
perfectly appropriately) applied to data based on the intuitions or
judgements of the linguist; but when you make grammaticality
judgements, you're not introspecting -- that is, reflectively examining
the contents of your mind -- any more than you are when you judge
a person who runs a red light to have done something against the
law, or judge a person who is not a native speaker of your language
to have a foreign accent. You're simply noticing whether something
does or doesn't match the norms you've internalized.

Michael Kac
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: OF as a verb

Date: Mon, 28 Nov 94 22:25:46 CSOF as a verb
From: "Margaret E. Winters <>
Subject: OF as a verb

Tony Bex raises some very interesting questions in the posting
about native speaker intuitions. However, there are at least
two other possible explanations for `could of' spellings which
plausibly might be offered by naive and not very literate
native speakers of British and American English:
 1. the `of' segment may be a pure homophone which is both
a verb and a preposition without any link other than pronunciation
between them. I think most people would hesitate, for example,
to claim that
 a) He might of said that and
 b) The queen of hearts
use the string pronounced `of' in the same way.
 2. `of' in `could of' might be some kind of preposition
rather than a verb, along the lines of `up' in `wake up the baby'
and so on.

Either of the above might be the native speaker's understanding
of `of' - it would be interesting to know if anyone has ever
queried this by asking native speakers of English in some fashion
or other.

 Margaret Winters
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue