# Comparing negatives: “she seems not to know” vs. “she doesn't seem to know”

What is the difference in style and meaning between the following two:

• She seems not to know.
• She doesn't seem to know.

Is there a name to this type of construction?

-
I would also like to know. – hermann Dec 2 '13 at 15:05
I see no difference in meaning. (I'll leave the style question to others.) – Andreas Blass Dec 2 '13 at 17:05

Yes, there is a name for this kind of alternation between constructions.
It's called Negative-Raising, or Neg-Raising (NR), among other things,
and it's governed by the predicate seem in this case;
there are a number of other predicates that govern it.

NR is a minor cyclic alternation rule.
That means that it is governed by the matrix predicate (all cyclic rules are governed)
and that the set of predicates governing it is small and specialized (that's the "minor" part)
and that it relates two different but synonymous sentence structures (that's the "alternation" part).

What happens is that, when you have a complement clause with a negative in it, like

• Bill wanted/seemed/intended/tried/managed not to be driving the truck.

with some predicates, but not others, this construction is equivalent to the same
sentence with the matrix predicate negated, but the complement not negated:

• Bill didn't want/seem/intend to be driving the truck. (equivalent with want, seem, intend)
• Bill didn't try/manage to drive the truck. (not equivalent with try or manage)

I.e, want, seem, and intend govern NR, and try and manage don't govern it.

Another way to look at it is that NR predicates are "transparent to negation", because
they don't really contribute much to meaning beyond individual perceptions and desires.
Whereas the vast majority of complement-taking predicates do contribute to meaning,
and are therefore "opaque to negation".

By the way, I used infinitives in the examples above for simplicity, but the phenomenon
is not limited to them. With the right predicates, tensed complements can undergo NR, too:

• I thought (that) you didn't want toast. = I didn't think (that) you wanted toast.
but ...
• He said (that) you didn't want toast. He didn't say (that) you wanted toast.
-
Helpful post in general. However, the answer presupposes that this is a syntactic phenomenon, as opposed to a pragmatic or semantic one. Even Postal's paper that you link to merely argues that the case against a syntactic account is not as terminal as had previously been thought, and that the case for a pragmatic account is not quite so decisive as one might think. (In other words they tacitly recognise the more compelling story offered by other accounts). – Araucaria Jun 14 '14 at 9:44
Feel free to ignore it if it compromises your faith. Like the faith that syntax, semantics, and pragmatics are independent. I tend to view them as three of many different kinds of phenomena, all of which interact. – John Lawler Jun 17 '14 at 23:31