"Tad bit"

Tue Aug 15 15:34:18 UTC 2006

On Aug 10, 2006, at 4:00 PM, Jerry Cohen wrote:

> Or more specifically, it started as a blend: from "a tad" + "a
> little bit."  Evidently there are several examples of blends
> involving the expression of something very small, e.g. "teeny" from
> "tiny" + "wee"; "tinetsy" from "tiny" + "teeny".
>
> Gerald Cohen
>
> ________________________________
>
> From: American Dialect Society on behalf of Arnold M. Zwicky
> Sent: Thu 8/10/2006 5:25 PM
> To: ADS-L at LISTSERV.UGA.EDU
> Subject: Re: "Tad bit"
>
>
>
> On Aug 10, 2006, at 7:10 AM, Charles Doyle wrote:
>
>> For several years I've been noticing, both in real life and on TV,
>> locutions like "I'm going to add a tad bit more ginger." I don't
>> believe OED or other dictionaries record that use of "tad"--or
>> maybe the phrase is an idiom.  A Google search for "a tad bit"
>> yielded 1,200,000 hits; for "a tad bit of" 102,000 hits; for "a tad
>> bit more" 103,000 hits.
>
> i'd guess that "tad bit" is just piling on two minimizing modifiers
> -- like "little small" or "tiny little" -- to get the meaning 'not
> just a little bit, but a *tiny* little bit'.
>
> arnold

i realize that this is an old complaint on my part, but i think that
jerry's use of "blend" here stretches the word beyond all
usefulness.  it's just wrong to use a single term for all expressions
that can be analyzed as a combination of two expressions.

to start with, some of these combinations are clearly inadvertent --
things that the person who produced them would (and sometimes does)
disavow.  most of the examples in jerry's 1987 book on syntactic
blends are like this.  the interpretation usually placed  on these is
that they arise from competition between two plans (one for each of
the contributing expressions) in production (either speech or
writing).  there are morphological blends of this sort too; we
occasionally cite some here, but here are a couple of others:

[musician, in a stanford music class:] "take a pitchfork".  pitch
pipe + tuning fork; the participants used both.  the speaker noticed
the error, and its comic character, and corrected herself.

[noted linguist, giving a talk at stanford:] "one horrifle..."
horrible + awful.  a strange look of unease crossed the speaker's
face at this point.

these are not only inadvertent, they also involve two *specific*
contributing expressions.

i'd like to reserve the term "blend" for such examples, and only
those.  well, i want *some* technical term with this meaning, and
i've been using "blend", in the belief that that usage was pretty
much standard.  if you want to use "blend" for some broader class of
phenomena, ok, but i then want a word for this specific meaning --
because this class of phenomena are united by the psychological
process that gives rise to them.

in contrast to the morphological blends, there are piles of
*advertent* combinations of words -- coinings like "smog" and
"Brangelina".  these i call "portmanteau words".  it might, or might
not, be that portmanteau words and morphological blends are similar
in their structures: both depend on speakers' knowledge of the
phonological and morphological structure of words, but it's not clear
that this knowledge is used in the same way in the two cases.  that's
an empirical question, as they say, and we shouldn't decide the
answer to that question in advance by using the same label for both
sets of phenomena.

(in the specific cases of "teeny" and "tinetsy", it's not at all
clear to me that a blend analysis is the right one.  this supposes
that their first creations were inadvertent, and then spread; this
kind of spread of inadvertent error does occur, but not, i think,
very frequently, so we should be cautious in positing such a
history.  here, portmanteau coining looks like a plausible analysis
to me.  in fact, at least in the case of "teeny", even portmanteauing
isn't necessarily involved: this could just be a sound-symbolic
diminution of "tiny" by a shift of the vowel to high front, the
"smallest" vowel available.)

on to syntax.  there are plenty of inadvertent combinations of
specific expressions (for which i want to reserve the label
"syntactic blends").  but there are also several types of advertent
innovations, which moreover don't necessarily involve specific
expressions.  one of these is what i called "piling on" in my early
posting; "reinforcement" would be another good name.  in piling on,
you use two constructions (with similar meaning) in combination, to
get an enhancement or reinforcement of this meaning: "tiny little",
"continue on", "return back", and hundreds of other types (including,
i think, "tad bit").  in my experience, people who produce these
"pleonastic" expressions *never* see them as errors to be disavowed;
they are saying just what they mean to say, with each of the
contributing constructions supplying its own bit of meaning.  if we
called these blends, who shouldn't we call other combinations of
constructions blends?  would we call a perfect double-object VP like
"have given them money" a blend of a perfect VP ("have given money")
and a double-object VP ("gave them money")?  no.

why not?  first, because these combinations are advertent.  second,
because what are combined here are two *patterns*, two constructions,
not two specific constructions.  and third, because the combination
doesn't arise from *competition*, even between patterns, but from
*composition* of two things.

once again, you could claim that the piling-on examples originated in
error and then spread to become part of the grammars of many
speakers.  but this is just gratuitous, and often involves
considerable uncertainty as to the specific expressions involved.  is
"return back" based on "go back" or "come back" or "move back" or
what?  is "return home" (or some other expression involving "return")
a contributor?  how could we tell?

we can get a satisfying analysis without having to answer these
imponderable questions: english has a construction in which motion
verbs (including "go", "come", "move", and "return") can be used
intransitively, and it has a construction in which verbs (with the
appropriate semantics) can be modified by following directional
adverbials (including "to school", "home", "here", "out", and
"back"), and these two constructions can be composed, even when the
verb and the adverbial overlap significantly in their semantics.

one more thing that shouldn't be lumped in with inadvertent
combinations of specific expressions: pattern extension to new
lexical items.  suppose you come across someone producing "donate him
money" (just google on the string, and you'll get some real-life
examples).  do we say, ah, a syntactic blend, of "give him money"
with "donate money to him"?  i think not.  to start with, people who
use "donate" with "him" and "money" as its objects also use it with a
variety of other indirect and direct objects; one of the contributors
to the phenomenon is a *pattern*, or construction -- the double-
object construction -- not a specific expression.  and, again,
occurrences of such examples are almost always advertent; most of the
time, people who  produce them see no anomaly in them.  (occasionally
they are clearly produced -- as evidenced in self-correction or
speaker distress or unsureness -- as inadvertent results of
competition between two plans, but most occurrences are not of this
sort.)

again, we can get a satisfying analysis without the involvement of
blending.  speakers of english know several things about the double-
object construction: what its form is, what its semantics is, and
what lexical items are eligible to serve as the head verb in it.  you
can learn the first two things and learn that there's a list of
lexical items occuring frequently as head, without knowing just how
big that list can get.  (you can't know what's in other people's
heads, and you can't get a panoptic view of "the whole language",
only what you've experienced.)  so you can either be conservative,
and stick with the ones you've heard, or you can strike out and
extend the list to items that are semantically like the ones you
already know about.  for some constructions, virtually everybody
strikes out, but for others, most people are conservative, and still
others lie somewhere in between.  (it's an extremely important
research question why different constructions behave differently.)
even for constructions where most people are pretty conservative --
the double-object construction is certainly one of these -- some
people extend the construction to at least a few heads they haven't
heard.  and then, of course, they're available to serve as models for
other people (though the effect they'll have will probably be  small,
because of the low frequencies of occurrence involved).

and again, you could claim that things like "donate him money"
originate in errors, and that's not entirely impossible, because
errors do occasionally occur, but it's hard to believe that errors
are the main mechanism, because they are so rare.  we can observe
pattern extension in action -- for the double-object construction,
linguists have reported becoming aware of when a new verb became ok
for them in the construction -- so we should assume that most of the
time this is a sufficient account of the phenomenon.

one place where pattern extension can be seen all over the place is
in the extension of idioms.  people are foreover playing with
(relatively transparent) idioms, trying out new things in old slots.
"throw s.o. to the wolves" gets varied with "bears", "alligators",
and so on.

an idiom case that came up here recently is that of "what
everything" (for the colloquial idiom "what all").  my suggestion was
that this extension was motivated by a desire to improve on "what
all", to moderate its colloquial character by trying out the perhaps
more formal "everything" in the "all" slot.  whether or not that's
the best analysis for this case, idiom extension happens a lot, and
these cases can rarely be satisfactorily explained in blend terms.
(note: there are, of course, idiom blends, though few of them catch on.)

a final point: i'm not denying the significance of errors in general,
and blends (in the narrow sense) in particular -- blends occur with
modest frequency, and errors do sometimes play a role in linguistic
change -- but i am cautioning against seeing blends everywhere two
things are in some sense combined.  i'm also also emphasizing the
role of creativity -- usually unconscious creativity, but creativity
nonetheless -- as a factor in language use, linguistic variation, and
linguistic change.

arnold (zwicky at csli.stanford.edu)

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