LINGUIST List 5.758

Wed 29 Jun 1994

Disc: Popularization of linguistics

Editor for this issue: <>


  1. Logical Language Group, Re: 5.735 Popularization of linguistics
  2. Mark Douglas Arnold, Speaker intuitions

Message 1: Re: 5.735 Popularization of linguistics

Date: Mon, 27 Jun 1994 10:53:29 Re: 5.735 Popularization of linguistics
From: Logical Language Group <>
Subject: Re: 5.735 Popularization of linguistics (George Oliver) writes:

> First, the issue of the descriptive /prescriptive positions can
> easily be put into a larger rhetorical perspective. We are arguing
> rhetorically questions [stases] of "Definition" and "Value". I note to my
> students that these positions can be translated as the difference between
> "what is it that exists?" versus "is what exists desirable?"

Yes, this is the IS/OUGHT distinction from my earlier posting.

> It seems to me that many linguists are
> arguing that the existence of some grammatical form is enough to justify
> it, whereas non-linguists argue that that is not enough. What we as
> linguists need to show is that the existence of some form may be desirable
> since it pushes the language in some natural, desirable direction or
> because it reveals something about the nature of language or because it's
> more logical in some linguistically patterned sense. I realize that this
> grossly oversimplifies the argument and assumes agreement on certain
> premises, but the point is that we may have to do more than argue existence
> as an end.

But why is it the business of linguists as such (as opposed to linguists as
citizens, or linguists as generalist thinkers, or ...) to "show ... that
the existence of some form may be desirable ...?" This is OUGHT, not IS.
What light does the science of language have to shed on any such matter?

If the schools were teaching people to believe that all languages are
descended from Hebrew (or Arabic), because that was Adam and Eve's original
language, linguistic science would have something to say in rebuttal, just as
biologists are concerned to rebut anti-Darwinian biology (and at one time,
physicists were concerned to rebut anti-Galilean physics).

Likewise, linguists as such have expertise on the subject of "some languages
have only 300 words" or "some languages don't have grammar" or "Eskimos have
100+ words for snow"; and rebutting these myths is a worthy activity.

But all of these are (false) beliefs ABOUT language, i.e. bad linguistics.
They aren't directly connected with the use of language.

> Second, to contrast these two positions clearly, I make a couple of
> analogies to my students. "What," I ask them, "is a weed?" It's amazing
> how many students believe that plants can be objectively described as
> weeds. Further questioning: Would a plant biologist classify any plants
> as weeds? It becomes clear in such a discussion that prescriptive grammar
> is akin to gardening: whatever does not "belong" according to some
> arbitrary notion of acceptable plants (grammatical form) in a certain
> context (sentence and situation) is a weed.

I think this analogy is excellent, save for one point that undermines the
whole thing. On what evidence is the definition of "acceptable ... grammatical
form[s]" characterized as "arbitrary"? My suspicion is that there is no
reason to think it arbitrary, and plenty of reason to think otherwise.
It is arbitrary only in the technical sense that lexical items are called
arbitrary; as Samuel Johnson put it with his usual perspicacity: "My name
might have been Nicholson as well as Johnson originally, but if you were
now to call me Nicholson, you would call me very absurdly." The name
"Johnson" is an arbitrary name, but there exists an explanatory structure,
having reference to the origin and use of English surnames, that justifies
it (provides reasons for its existence and use).

> While this analogy seems to
> make prescriptive grammar appear benign, it is also clear that if such weed
> pulling is done, it must be done with the understanding that the plant
> being pulled is not bad, just unwanted for some reason. And it's the
> reason that may *not* be benign.

>May< not be benign. If linguists (qua scientists) value clarity,
cross-cultural accessibility, and other Good Things in their scientific
discourses, it's no wonder their papers tend to be published (if in English)
in a standard version of English (either Commonwealth or American).
These, surely, are benign reasons for pulling weeds?

> At the end of my course, many students understand that a real
> perspective on language takes the drive out of their prescriptivism because
> they no longer need to contain/control what they no longer misunderstand
> and thus fear. In the end, botanist-gardeners may still do some judicious
> weeding, but it will be clear that such actions are based on criteria that
> have to do with a human sense of order, aesthetics, pragmatics, and that
> such criteria may not match some other "natural" order. Now, at least, we
> have a sense of what we're really arguing about.

Agreed. The question to me is whether those criteria you describe
are themselves subject to rational discussion, or whether they represent
merely the taste of a dominant social order. I believe the former.

John Cowan sharing account <> for now
 e'osai ko sarji la lojban.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Speaker intuitions

Date: Tue, 28 Jun 1994 12:16:45 Speaker intuitions
From: Mark Douglas Arnold <>
Subject: Speaker intuitions

Wen-Choa Li posted the responses to her question regarding speaker
intuitions. As she noted, the responses varied in the way they
characterized the issue. In particular, comments from Hal Schiffman
(and Jakob Dempsey, to a lesser degree) prompted the following
comments. Initially I had planned to respond only to HS, though now
that I have thought about it, I think it would be worth opening the
discussion. The heart of the matter, it seems to me, is the tension
between the apparent arrogance of theoretical linguists and the view
of language undertaken by non-"asocial" linguists. My hope, (being
one who practices methodological idealization) is that we can use
each other's work, i.e. "raid" whenever possible, and ultimately
develop a relatively fine-tuned model of this thing we call language.
I apologize for the length, but there are many issues to address. I
turn to the comment from HS which prompted my reply:

>Theoretical Linguists speak an invariable idealized language,
thinking of themselves as the ideal speaker/hearer, for whom there
are no fuzzy judgements. Who can believe such people?<

I'm hoping that he has overgeneralized to make a point, and the
generalization is based on the fact that he has encountered specific
theoretical linguists who are extreme in their methodology concerning
the investigation of language, i.e. the working assumption that we
can imagine an idealized language user. Of course the ideal is not
real, but if one is interested in competence rather than performance,
then one has to start somewhere. My personal experience is that
theoretical linguists often end up saying things like, "I don't have
strong judgements on that data," or "My judgements don't align with
yours, but let's grant the data in order to follow the line of
reasoning and see what the implications are for the overall system".

The non-anecdotal evidence that there are non-extreme theoretical
linguists (for whom there are "fuzzy judgements") is that the
theoretical linguistic literature often contains discussions of
graded judgements. (In many cases the gradation can be seen to be
the result of the interaction of various abstract mechanisms, or
perhaps the difference between particular lexical specifications
interacting with universal principles. In other cases the variation
is noted for the problem it poses, but is left for future work;
nevertheless, the point is that theoretical linguists do acknowledge
variation, work to explain it when possible, and state the problems
the variation poses when current models can't offer insight
concerning the variation.)

There is a further wrinkle concerning the "fallibility" of speaker
intuitions and the self-serving practice of linguists using their own
intuitions to develop their theory. Take a "sentence" like

 >Red ran woman the up with hill sneakers the.
My impression is that judgements concerning this string are
essentially infallible (contra HS's comment: "... the notion that
native speakers are infallible on intuitive judgements of
grammaticality is unsupportable". Even given "context" issues, I
cannot imagine the above "sentence" being uttered by any
(non-aphasiac) English speaker. Certainly this is the extreme
example, but here's why I bring it up: even though most of us can
tell the difference between an x-ray of our teeth and an MRI of our
knee, it is often the case that the average sonogram of a fetus is
nothing more than a "fuzzy" image. (I've made an analogical leap
here without any rhetorical device previewing the leap, and I
apologize, but on to the point --->) Asking non-linguists for
judgements on data which is anything more than very basic is like
asking non-specialists to make sense of sonograms.

This is especially true when theoretical linguists _do_ consider the
context in which an utterance is/isn't acceptable. Any theoretical
discussion of pronominal/antecedent relations relies on the
difference between a sentence merely being "acceptable" versus being
"acceptable under the interpretation where X refers to Y". The clear
example is that while "Mary saw her" is an acceptable English
sentence, it is very difficult (i.e. one must imagine a very specific
(and probably contrived) scenario) for "Mary saw her" to mean "Mary
saw herself." Combine these types of reference phenomenon with
quantifier scope, and the problem gets more intricate:

 >Everyone's mother loves him.
 >He is loved by everyone's mother.
even though the first sentence has two readings, (1-there is one
person that everyone's mother loves, or 2-for every X, X's mother
loves X), the second one is "acceptable" only for the interpretation
that everyone's mother loves a certain person, and is "not
acceptable" under the interpretation that for every X, X's mother
loves X. (For some it is enough to OBSERVE these facts; others want
to know why the facts turn out the way they do.) Now add another
degree of comlexity:
 >Every MIG pilot who shot at some F16 hit it.
 >Every MIG pilot who shot at it hit some F16.
Do these two strings have all of the same interpretations, or are
some interpretations available for one but not the other? (If you're
interested in figuring this out, but don't know where to start, ask
yourself if in each sentence the number of F16s which got hit is
completely determined by, optionally determined by, or completely
independent of the number of MIG pilots who shot at the F16s.)

In much the same way that an orthopedist can probably tell if an
x-ray of a knee is a posterior view of a left knee or an anterior
view of a right knee, (while I can only tell that it's a knee,) so it
takes training to develop a sensibility about what sentences are
"acceptable" under what "contexts". Of course non-linguists "get
confused" when we ask them things about their language --- linguists
do to. Perhaps certain individual linguists are less likely to admit
it, or perhaps certain individuals have very strong intuitions, but
by and large, my personal experience is that theoretical linguists do
acknowledge "fuzziness" (gradation) of intuitions, and in certain
instances work to offer insight about how the system should be
modelled so that we can predict the gradation along the empirically
attested lines.

Just as we used to think that the sun moved in relation to a
stationary earth, (and certainly that is observational true,) it took
looking at less directly obvious evidence to come to a more accurate
understanding of the solar system. Theoretical linguists imagine an
ideal language user in order to try to gain access to the less
directly observable aspects of human language.

Again, my apologies for going on. Clearly there is much more to be
said. Hopefully we can say it as comarades.

Mark Arnold
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue