the thin line between error and mere variation

Arnold M. Zwicky zwicky at CSLI.STANFORD.EDU
Sat Jun 26 18:10:58 UTC 2004

In trying to write some commentary on Geoff Nunberg's discussion of
"nucular" (in a 10/2/02 Fresh Air commentary on NPR, now in his
collection Going Nucular), i've been reflecting on that thin line
between error and mere variation.

Nunberg begins this piece by drawing a distinction between "typos" and
"thinkos" -- in my terms, between inadvertent errors, things that are
"wrong" for the person who produces them, and advertent errors, things
that are ok so far as the producer is concerned but "wrong" from the
point of view of at least some other people.  (Faced with a typo, you
call in the psycholinguist; faced with a thinko, you call in the

The distinction is a familiar one in the literature on language errors.
  In the typo camp, you have, for instance, Fay/Cutler malapropisms (so
called from a 1977 article by David Fay and Anne Cutler), like my
(alas, only too frequent) productions of "verb" for "vowel", or vice
versa, in class lectures.  In the thinko camp, you have, for instance,
classical malapropisms (so labeled by me in a 1979 article), like
"behest [beset] with all these difficulties", written by someone who
*meant* to write "behest" (and was willing to defend this word choice).
  It might be hard to decide, in any particular instance, which kind of
malapropism you're looking at, but in principle, with more information
about the producer and their intentions, you can sort things out.

But matters are not so clear in the world of thinkos.  The deviance of
thinkos ranges from extremely high, as in clear examples of classical
malapropisms, to extremely low, as in violations of the more fanciful
proscriptivist pronouncements, like the one against possessive
antecedents of pronouns.

(A side issue: It would be a good thing to expunge the moral language
usually applied to thinkos, even by Nunberg, who should know better:
Typos "can make you look foolish, but they aren't really the signs of
an intellectual or ethical deficiency, the way thinkos are.  It's the
difference between a sentence that expresses an idea badly and a
sentence that expresses a bad idea." (p. 59 of GN).  Look at the most
extreme case...  Someone who writes "behest" for "beset" is certainly
wrong.  But they aren't morally defective, or evil, or stupid.
Technically, they are very specifically ignorant, of one  of the
zillion facts about the world one might be called on to marshal in
everyday life.  It's like getting Bjo"rk mixed up with Bork, or not
knowing at all who Hugo Wolf is.)

The "behest" thing is, yes, an extreme case.  But things don't get any
clearer as we work towards possessive antecedents.  They just get
messier and messier, in fact.  As soon as we leave the clear "behest"
zone (where almost everyone says the usage is wrong for them), we have
to confront a world in which usage is contested and variable.

We come first to the Retart Zone, a label I use to honor a poster to
the newsgroup sci.lang:
"D---" on sci.lang, 6/24/04, called by Peter Daniels on the voiceless
final consonant in his insult "What a retart":

And what's wrong with my use of "retart"'s a perfectly acceptable
word when describing
those who are SLOW.  A retart is a SLOW person.
In later discussion, the pugnacious D--- concedes that (some) other
people say, and write, "retard", but maintains that *his* version is
perfectly fine.  That is, he claims that this is a case of variation,
not error.  He is surely in a small minority in his pronunciation, but
probably not a loner; I have no doubt that some searching would turn up
others with his pronunciation.

Certainly, there *are* plenty of examples of variation.  Some English
speakers (I am one) have a voiceless final consonant in "with", some
have a voiced final (a fact that I did not appreciate until I gave an
exercise in phonetic transcription in an introductory linguistics
course); I believe that the voiced variant is statistically the
predominant one, by a considerable margin (some dictionaries list only
this pronunciation), but theta-speakers like me don't provoke dark
looks and snickers with our minority pronunciation.  Similarly, some
English speakers (including a great many South Africans) have edh
rather than theta in the "South" of "South Africa"; I believe that they
are definitely in the minority in the English-speaking world, but who
am I, an American theta-speaker, to tell South Africans how to
pronounce the name of their country?  Similarly, many New Yorkers stand
"on line" rather than "in line"; they're a small minority in the
English-speaking world, and they are aware (at some level) that other
people use "in" here, but everybody knows that people speak differently
in different places, so where do you get off telling them they're

On the other hand, we do tell "needs V-ed" speakers (again, a small
minority in the English-speaking world) that they're "wrong".  These
folks are aware (at some level) that other people say "needs V-ing",
but most of the people they know personally are "needs V-ed" speakers,
so from their point of view, they're talking appropriately, and the
dark looks and snickers from outsiders are just nastiness.

Even in the Retart Zone, we're in trouble.  What's unremarkable
variation, and what's an thinko-type error?

But then we get to the Nucular Zone, the Hone-In-On Zone, and the
Another-Thing-Coming Zone.  The percentage of people who use the
(historically) innovative variant steadily increases.  (Google web
searches have "home in on" somewhat above "hone in on", 64,200 to
35,200 in raw numbers, but "another thing coming" *way* over "another
think coming", 21,400 to 5,830.)  Those who use the innovative variants
are probably aware (at some level) that other people have other
variants, but for them this is just unremarkable variation, and their
version is, well, *their* version, and perfectly ok.

The argument from history isn't going to carry much weight for these
people, and anyway it's intellectually disreputable, since very few
current standard variants have a pedigree going back to Old English;
almost everything was an innovation at some point.  How to decide when
the ship of language change has sailed?

The argument from authority won't carry much weight, either.  I can
tell you that *I* (a noted linguist and writer) use "nuclear", "home in
on", and "another think coming" (and "too big a dog" rather than "too
big of a dog", but don't use positive "anymore", etc.), but you're
entitled to ask why I should be telling you how to talk and to note
that anyway you think I sound bookish and prissy.

If anything might work, it would be the appeal to the practice of those
who are noted for their abilities in writing and speaking -- there's a
reason AHD ended up with a Usage Panel, awkward though it turned out to
be -- but in fact these experts are quite often divided in their
practices and in their opinions, and in any case they're not
necessarily models for writing and speaking in other than formal

All of this is familiar territory for people on this list.  The fact
seems to be that the line between mere variation and error is largely a
matter of intellectual fashion  -- lord knows why speaker-oriented
"hopefully", restrictive relative "which", split infinitives, logical
"since" and "while", etc. get picked on while other variants thrive
without criticism -- rather than a result of observation and reasoning.
  In this context, the label "thinko" doesn't really seem much better
than "error" or "mistake".

arnold (zwicky at, who has more to say specifically
about "nucular"

More information about the Ads-l mailing list