Date: Tue, 7 Apr 1998 15:42:18 -0400
From: Larry Horn
Subject: Re: more on non-negative "un"

At 1:38 PM -0500 4/7/98, Mark Mandel wrote:
>I found this in the "Selling It" section of the March, 1995 issue of _Consumer
>Reports_ while clearing out (p. 211 = inside back cover):
> ===============
> Oh. So you'll refreeze our pipes?
> ---------------------------------
>Perhaps Brown's Certified Welding should have thawed out this wording a little
>(Photo of ad evidently clipped from newspaper):
> PIPE UNTHAWING <----[largest lettering in the whole ad]
> ===================
>Presumably this was meant, but not read, as the same non-negative
>"un-" as in "unravel", "unloose", and "unpeel", with the sense of
>'release, un[either sense!]bind'.

Not just presumably; this is a fairly well-attested sense; the OED
citations date back four centuries, and the meaning is always the redundant
one, in which UNTHAW = 'thaw' or 'unfreeze'. I discussed this in a paper
on un-verbs ten years ago, from which I can provide this excerpt for anyone
curious about why "unthaw" is a synonym of "thaw", while "unfreeze" is an
antonym of "freeze", which I originally presented in the course of a
general analysis in which all un-verbs basically involve "helping entropy
along", returning their theme arguments to their original or unmarked
state. (Hope no Dutch readers take offense.)

Why is it that while "unthaw", which looks like an antonym of "thaw" or
"unfreeze", is in fact their synonym, while "unmelt", which looks like it
ought to mean whatever unthaw means, in fact does not occur at all? In OED
citations dating back four centuries, including those in (16),

(16) The ponds were almost unthawed.
The men [in Holland] are cold to such a degree that neither Love nor
wine can unthaw them.

"unthaw" is invariably equivalent simply to "thaw". Of course, "unfreeze"
too always equates to "thaw". This conforms to our hypothesis, assuming
that the natural state of H2O is liquid, that is, water. Hence, freezing
can be un-done, restoring the state of nature, while thawing--which itself
creates a natural state--cannot. The systematic character of this
asymmetry is reinforced by the existence of dialectal forms eave and
uneave, both glossed as 'thaw' (Wright 1961)...

[Discussion of relation of unV to 'V out' as opposed to 'V up' omitted.]

Predictably, neither "unfill" nor "unempty" occurs freely, each being
normally pre-empted by its underived synonym (cf. Clark & Clark 1979). But
the former does occasionally surface unblocked as the entropic antonym of
"fill", while the latter, typically in the dialectal version "unempt",
occupies that same semantic slot, duplicating its unprefixed entropic
counterpart. To "unfill" or "unempt(y)" a bowl or a lake is to empty it.
And of course we fill UP a bowl of water but empty it OUT. We are now
prepared to recognize that it is no accident that a lake freezes up, but
thaws out.
So much for thawing and freezing. Now what of the missing
"unmelt"? If melting is akin to thawing, the transmutation from a solid
(or frozen) to a liquid (or unfrozen) state, why don't we have redundant
"unmelt" alongside "unthaw"? The answer is that melting is in fact
crucially unlike thawing: while a thawed object, say a turkey, is a
temporarily frozen theme or patient which returns to its unmarked state,
retaining its physical integrity, the integrity of a melted theme or
patient may be affected or even destroyed by the change of state. A thawed
turkey is still a turkey (indeed, a turkey par excellence), a thawed lake
still a lake, but a melted ice cube, snowman, or wicked witch is not simply
a different form of the same object, but essentially a puddle of one kind
or another. Notice too that where things freeze up and thaw out, they melt
away. If I come to your door complaining that I'm frozen stiff, my
reaction might differ considerably depending on whether you offered me a
brandy to help me thaw out or a potion to help me melt away.
(If you want more, the article is "Morphology, Pragmatics, and the
Un-Verb", and it's in ESCOL '88, pp. 210-33.)


P.S. The last word on these redundant un-verbs from a card-carrying
prescriptivist (to be read while banging dictionary loudly on the lectern):

The verb "to unloose" should analogically signify "to tie", in like manner as
"to untie" signifies "to loose". To what purpose is it, then, to retain a
term, without any necessity, in a signification the reverse of that to
which its etymology manifestly suggests?...All considerations of analogy,
propriety, perspicuity, unite in persuading us to repudiate this
preposterous application altogether.
(George Campbell, Philosophy of Rhetoric, 1776--
thanks to Dennis Baron for the citation!)