LINGUIST List 8.1809

Fri Dec 19 1997

Disc: Prescriptivism

Editor for this issue: Elaine Halleck <>


  1. manaster, Re: 8.1807, Disc: Prescriptivism
  2. Dick Hudson, prescriptivism
  3. bwald, Re: 8.1796, Disc: Prescriptivism

Message 1: Re: 8.1807, Disc: Prescriptivism

Date: Fri, 19 Dec 1997 09:30:17 -0500 (EST)
From: manaster <>
Subject: Re: 8.1807, Disc: Prescriptivism

David Harris says:

 Someone made the comment the other day that linguists who are
involved in dictionary projects are by definition prescriptivists
because they are contributing toward the production of a text which is
meant to serve as a linguistic "authority." Only not all dictionaries
do that. A dictionary is first and foremost a source of

That was me, except I did not quite say that. I myself have been
associated with the OED, which is as descriptive a dictionary as
there is for English I guess. Actually, I have some doubts even
here, but I think I tried to be clear that I meant those dictionaires
like Webester's, Random House etc. which clearly are prescriptive.

I should add that I do not buy the idea that one can replace
presecriptivism by merely "describing" the usage of "educated
speakers", as has sometimes been argued. That's just a cop-out.

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

Message 2: prescriptivism

Date: Fri, 19 Dec 1997 08:39:54 +0000
From: Dick Hudson <>
Subject: prescriptivism

Alexis Manaster-Ramer seems to object to mother-tongue teaching at school
because it's prescriptive. Have I misunderstood you, Alexis? How do you feel
about your own ability to use standard spellings and word-meanings, and to
write coherent sentences? Maybe you learned it outside school, but most
people depend on school teachers for that kind of life skill.

Here's what Alexis says in reaction to William Morris:

AM-R: William Morris says he thinks there is little if any objection to
prescriptivism of the sort described as:

WM: Elementary and secondary language studies: reading, writing, and (with
luck) some sort of speech arts. Every child that goes through school
is subjected to 8 to 20+ years of prescriptivism. Learning to spell
is prescriptivism, learning the (standard) definitions of words is
prescriptivism, and learning to write a coherent sentence is

AM-R: I dont agree. There are many cultures in which children are spared
this kind of (mis)treatment, and I would think that anyone who in
principle objects to prescriptivism would object here too. 

Richard (=Dick) Hudson
Department of Phonetics and Linguistics,
University College London,
Gower Street,
London WC1E 6BT
work phone: +171 419 3152; work fax: +171 383 4108
 home page =
 unpublished papers available by ftp =
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 3: Re: 8.1796, Disc: Prescriptivism

Date: Fri, 19 Dec 1997 02:51:59 -0800 (PST)
From: bwald <bwaldHUMnet.UCLA.EDU>
Subject: Re: 8.1796, Disc: Prescriptivism

I continue to follow the discussion of prescriptivism with interest, and
some degree of confusion about what the boundaries of prescriptivism are.
Finally, though, something quite concrete came up, including something I
had been thinking about. I like examples. Thus, Oesten Dahl notes the
following passage from an earlier message by Alexis Manaster-Ramer
(LINGUIST List 8-1780):

>"(4) It might be that certain aspects of prescriptivism do have some
>real basis.

It's me, Benji. I haven't finished the quote yet, but A means something
like "linguistic" by "real" here (probably something like "intrinsic to the
design or logic of language/s"). This passage as it continues was also of
great interest to me for the concreteness of the example and the conclusion
that A draws from it.

I have over the years collected some examples that seem
>to suggest this. For example, it is a well-known fact that in many
>European languages constructions like
 'Having read the Joy of Cooking, it was easy for him to make a
perfect souffle'
are (a) commonly used
>and (b) condemend as incorrect by the prescriptivsists. Now, since
>the grammarians and mavens of European languages are all connected,
>this is not necessarily remarkable. However, whenwe find that such
>constructions were common in Sanskrit but are not allowed by Panini's
>grammar (Peter Hook wrote a paper on this some years ago), we might
>well wonder if there is not something beyond mere arbitrary convention
>in what the prescriptive grammarians teach (at least some times)."

Now I had some ideas about this example -- the "dangling participle" it's
calledin English, and how it relates to "style" in a sense that goes
beyond "grammar" (as what is "possible" -- uh, "grammatical" -- in a
language) to preferred ways of saying things, meaning more"effective" than
some other way which is also "grammatical" but not as effective for some
reason, maybe cognitive in this case (I emphasize "maybe"). A is
distinguishing this kind of case from such cases of blatant discrimination
as attacks against multiple negation (earlier the only choice in English),
split infinitives (earlier not a possible choice in English -- any more
than in German?), and other things where the attacks are not made on a
"linguistic" basis (again, what A meant by a "real basis").

The dangling participle is usually illustrated in cases where ambiguity
results -- and, as we all know, "ambiguity" is one of the great
rationalizations used by prescriptivists, since "clarity" is a value tyoted
by prescriptivism. "Clarity" is somehow related to "(ease of)
communication" (from the listener's POV) and "ambiguity" interferes with
"clarity" (at least when the ambiguity is noticed). (Cushing's book "Fatal
Words" shows us that sometimes ambiguity is not noticed and the plane
crashes because the pilot only noticed the wrong interpretation. Thus, the
matter is far from trivial in some cases.)

Thus, "walking down the street, the trees looked more beautiful than ever
to me" is a typical example of a proscribed dangling participle. Actually,
the principle, which is a stylistic one, not necessarily one of
grammaticality, seems to be a more general one of order or precedence (cf.
Langacker's old "command" relationship which claims that in "he was only a
child when John last saw me" the pronoun "he" and "John" canNOT refer to
the same entity -- but by changing the order OR which clause has the
pronominalisation they can be coreferential). Thus, also discouraged is:
"on the beach I saw a whale", if the idea is that the whale is on the
beach, and most certainly if I am NOT on the beach?, "in the tree I spied a
barrel-chested snipe", etc etc

Now A's point is that this may not be a language-specific principle, though
he does not go so far as to suggest it is a "universal" (or even a
"language-independent" principle). Thus, he does not make as rash/bold a
claim as was made about Langacker's command phenomenon (by many
generativists soon after Langacker made his observation, and perhaps, I
don't know, is still a big point to many formal grammatical theories).
However, the point seems to be that some kind of proximity relation is
prescribed for the sake of "clarity". The question could then be whether
in languages in general, or at least in a certain type of language,
proximity is *preferred* (stylistically) where two things "go together"
(like "me" and "walking down the street", or "whale" and "on the beach").
(If universal, and maybe even if not, one might suggest a "cognitive" basis
to the *preference* for having constituents related in a certain way occur
NEXT to each other -- or at least as close together as possible, a
qualification important to A's actual example, as noted further below.)

In any case, it seems clear that the proximity principle is not grammatical
in the sense that it cannot be regularly violated by speakers of a
language. Violations of the "dangling participle", among similar
constructions, are too common to assume that such a constraint is
Furthermore, if the basis for prescribing them is avoidance of ambiguity,
then where speakers assume that their interlocutors have sufficient
pragmatic knowledge to disambiguate (or not to be distracted by the
ambiguity) there is no need for avoidance. (A question then remains why it
is convenient for the speaker to allow the ambiguity, but that's not our
concern here.) So, I'm suggesting that the prescription proscribing
dangling participles and similar violations of proximity is about "style"
amd "preferred constructions", not about "grammaticality".

Also indicating the relationship between proximity as a preferred strategy
for semantic transparency and the possibility of alternative *grammatical*
constructions of greater opacity is the classical Latin commonplace of
separating a noun and its qualifying adjective by any number of words in a
sentence (sometimes in verse for prosodic reasons, but also in prose for
reasons I don't understand, and may simply have to do with distinguishing
written style from colloquial style). The separation was obviously
"grammatical" in Latin, and we can figure out which adjective qualifies
which noun, though sometimes with a little preliminary difficulty
(??because we're not used to processing our own languages in the same way).
But can we doubt that this was a somewhat artificial device, in the sense
that it was far less common in colloquial speech (if at all used) than in
elegant and learned writing? And would we expect a language with the
characteristics of Classical Latin (more precisely, its closest spoken
equivalent) to evolve such that the adjective might be REQUIRED to be
distant from the verb, i.e., the fancy literate construction becomes
obligatory (in spontaneous speech)? That is, could a language evolve a
(more or less) fixed word order which contradicts whatever proximity
principle is involved in head/modifier relations? I doubt it. And that
indicates that the proximity principle does play some kind of role in
languages (no doubt in all languages, but not necessarily the same role in
all languages, so maybe there is not one but a number of proximity
principles among languages, and that some preclude others).

So far so good. There is a relationship between the dangling
participle/etc and the construction that A singled out. But, no doubt on
purpose, he chose an example where ambiguity is not at issue, only
distance, because, unlike the examples I gave, there is no intervening noun
to which the participle might be attributed:
 'Having read the Joy of Cooking, it was easy for him to make a perfect

As I noted above, I am not sure in what way Sanskrit grammarians proscribed
whatever the equivalent Sanskrit construction was. Is it that they insist
on a nominative for the participle "head", so that "for him" is not good
style because it is in some other case (I guess without a preposition)?
Would proximity make it better? e.g.,
 having read the Joy of Cooking, for him it was easy to make a
perfect souffle
For that matter, what's "wrong" with the English version, and does the
changed word order make it better? I don't know. My sense of prescribed
STYLE is not that deep. However, I would suppose that by fiddling with
both case and proximity I can come out with an unexceptionable:
 having read the Joy of Cooking he was able to make a perfect
souffle with ease
 ... had no
difficulty making ...

I have already suggested that there may be a language-independent angle to
what might be "stylistically awkward" about A's example in English.
Distance alone may somehow become considered "awkward". But since English
(and French), unlike Sanskrit and some of the other languages A may have in
mind, has developed an "expletive" (dummy) 'it' in 'IT is easy for him',
there is further awkwardness of an intervening "referent", IT, between the
"having read..." and "for him". In this case I do not know whether it
would be a prescriptivist or cognitivist trick to suggest that the
intervening IT interferes with connecting "having read..." and "him". The
argument would be something like, proximity (in English) prefers the
participle to refer to the closest (first following) referent. "it" is a
(potential?) referent (actually, it refers to the infinitival phrase "to
make a perfect souffle"). Therefore, "it" interferes with proximity and
defeats it as a preferred strategy of comprehension processing. Therefore
it should be avoided (and, amazingly or not, almost any grammatical
construction can be avoided in favor of another one, allowing much room for
prescriptive statements about stylistic uses of grammar -- uh, did I just
dangle a participle?)

To reiterate my basic point, in English the proximity principle is about
style. Prescription prescribes a lot of things, some have to do with what
is "allowed" or not in a prescriptive grammar and some are about style,
preferred and dispreferred *grammatical* choices. Prescription is not
shallow in what it covers, because language is not shallow. It is complex
and subtle ("subtle" in comparison with usual grammatical theory).
Neither "prescription" nor "language" is confined to what's "grammatical"
and what isn't (in either a prescriptive or descriptive sense of the
word). Prescription is only shallow in its feeble attempts to account for
(or, yes, "explain") its pronouncements, or to seek out the implications of
those pronouncements.

(NB. Of course, the authoritarianism underlying prescription does not have
to give an account for its choices. It only does so as a matter of
didactic convenience. The attitude is: don't challenge the explanations.
If they don't satisfy you, then just forget them and MEMORIZE the rules.
The descriptive linguist is more thoughtful and curious. The cases of
general interest for linguistic/grammatical theory are just such
"stylistic" rules as A has identified. The other types, such as
prohibitions against multiple negation and split infinitive, don't amount
to much in terms of general linguistic insight. Sure, the particular
historical circumstances concerning any proscribed construction are of
interest within the language affected, but that does not necessarily have
any interesting implications for any other language. -- well, actually I'm
not sure about split infinitives, because they have to do with responses to
grammatical innovations -- and grammatical innovations are usually
evaluated negatively when first noticed. Yet they somehow have a logic
which eventually supercedes the logic that the language managed to get
along with fine enough before the innovation.)

I almost forgot. I wouldn't have bothered to write the above (at least at
this time), except that Oesten Dahl also commented on what A said. His
basic comment was:

>This relates to something that I have been thinking about for some
>time. Back in 1969, Paul Postal introduced the notion of "Anaphoric
>Island" which, reformulated in theory-neutral terms, boils down to a
>constraint against examples such as

"*John is an orphan and Peter's ones are dead too",

in which an anaphoric device refers to an inferred
>entity rather than one introduced by an explicit antecedent. Postal
>argued that such sentences are just ungrammatical, but it seems to be
>an empirical fact that people do sometimes say similar things, which
>is of course might be taken as simply implying Postal was

The issue of "grammaticality" aside, the claim that people say things like
Postal's example astounded me. But Oesten went on immediately to say:

However, the classical Swedish prescriptivist grammar,
>Wellander 1939, contains a section on precisely this type of sentence,
>ruling it out as "bad Swedish". That Wellander finds it necessary to
>include this section in his book, in my opinion, shows two things:
>first, that the constraint is not just a figment of Postal's
>imagination, second, that people do tend to violate it.

Well, that does not establish as "empirical fact" that people say things
like Postal's example. Wellander could be under some kind of illusion or
have extremely *bad taste* in his pedagogical strategems (I underlined
because of the possible irony about "style"). In any case, I would like to
know just what Wellander did offer which was "like" Postal's example.

Oesten ended with:

>interesting question remains: what is its status?

I would indeed be interested in what Wellander said and what the facts are.
As I anticipated in my earliuer response to the issue of prescriptivism,
the issue of "ungrammaticality" in avowedly descriptive linguistics
sometimes has a seemingly, though unintentionally, "prescriptivist" taint.
In any case, familiar as I am with English colloquial speech, I am totally
unfamiliar with anything that strikes me as similar to Postal's example of
an island constraint violation of "weak anaphora" (as I said above, quite
apart from whether or not the construction should be considered
"ungrammatical"; NB. if people regularly say it it is grammatical, but if
people don't regularly say it, it does not automatically follow that it is
ungrammatical, e.g., it could be "stylistically awkward" for some reason or
other, maybe a cognitive one?) -- Benji
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue