Shifting negative

[Laurence Horn]

At 9:00 AM -0400 7/11/10, Seán Fitzpatrick wrote:
>I recently found myself saying to someone who had arrived for a reception at
>1:30 and been told that nothing was happening:  “I don’t think we got back
>from the cemetery until a little after 2”.
>
>Of course, I meant just the opposite.  I DO in fact think that we got back
>after 2.  But any speaker of colloquial English would take my meaning, viz.,
><<I think we didn’t get back from the cemetery until a little after 2>>.
>
>
>
>Is there a name for this, and is it part of a more general pattern related
>to the shifty effect some words have on negatives, e.g., until, anymore?
>
Not sure what you mean exactly, but I suspect
there are two issues involved, one involving the
placement of the negative and the other involving
the interpretation of "(not) until".  To take the
former first:  There is a much described tendency
to "raise" negatives over verbs like "think",
"believe", "want", "ought to", and other verbs
(and adjectives, as in "it isn't likely
that..."), resulting in a higher clause negative
having lower clause meaning.  St. Anselm
described this tendency in the 11th century ("we
say 'non debet uxorem ducere', lit., he is not
obliged to marry, for 'debet non uxorem ducere',
he must not-marry), and more recently it was
rediscovered in the early years of
transformational grammar under the rubric of
"neg-raising" or "negative transportation",
applying to predicate of a certain semantic
class.  But this doesn't produce a reversal of
the meaning, only a strengthening of the negative
force, in the sense that "I don't think Spain
will win the finals" could be used to convey the
stronger (but not opposite) "I think Spain will
not win the finals".  This is a shifting, as you
say, and results in a contrary meaning for what
"ought to be" a contradictory (in the 19th
century, grammarians treated neg-raising as "a
logically unwarranted placement of the negative")
but the alternative, as in "I want not to see you
again" or "I want to not see you again", often
seem impossible or awkward compared to the
"illogical" meaning of "I don't want to see you
again".

Back to your sentence:  undoing the neg-raising would yield:

"I think we didn't get back until after 2"

Now let's get rid of the "I think" part, which
basically just acts as a qualifier or hedge.  We
have, then,

"We didn't get back until after 2".

This primarily says that we didn't get back
before 2 and secondarily implies, suggests,
presupposes, or entails (the literature is split
on this) that we *did* get back at or after 2.
If "before" had been used instead of "until", no
such implication obtains (or at least not as
strongly), so "He didn't say another word before
he died" is less likely to suggest a breaking of
the silence than "He didn't say another word
until he died".  But intuitions differ.

To the extent that "She doesn't live there
anymore" and "She didn't move away until 1980"
both have a negative main assertion (not living
there now, not moving away before 1980) and a
positive presupposition or implication (living
there at some point in the past, moving away in
or after 1980), "anymore" and "until" are indeed
parallel, as is "yet" ("She hasn't recovered
yet", implying the possibility of recovery in the
future.)  I actually published a paper about that
parallel exactly 40 years ago ("Ain't it hard
(anymore)", CLS 6, 1970).

LH

------------------------------------------------------------
The American Dialect Society - http://www.americandialect.org