LINGUIST List 3.466

Sat 06 Jun 1992

Disc: Rules

Editor for this issue: <>

Directory

1. , 3.464 Innateness and Rules
2. David Stampe, 3.459 Rules
3. benji wald, Re: 3.464 Innateness and Rules

Vicki Fromkin gives a list of ten ill-formed constructions and comments:

"The knowledge that these are ill-formed in English has nothing to do
with normative rules -- just grammatical constraints."

The fact that I immediately read this as a contradiction in terms indicates
that some clarification of our concepts is in order. If 'normative rules'
in this discussion is taken to mean 'explicitly formulated statements by
prescriptive grammarians', Fromkin's statement makes sense - but that is
not what I have taken the discussion to be about. Perhaps we should use the
term 'prescriptive rules' for this phenomenon, and reserve the term 'norms'
for the implicit (i.e., not explicitly formulated) social conventions that
we have to assume in order to account for people's ability to distinguish
between correct and incorrect behaviour. We should also distinguish between
questions concerning these norms and questions concerning the way knowledge
of them comes about and is structured in the individual.

Even strictly descriptive linguists are 'normative' in the sense that they
are describing a norm-based phenomenon. This does not make their task less
empirical (in spite of Itkonen) - as opposed to that of the prescriptive
grammarians, who try to impose would-be norms of their own or someone
else's creation on us.

Helge Dyvik

This whole discussion of normal and abnormal data seems abnormal.
Is "dog the" a noun phrase? It might happen, if you were talking
about noun phrases. But who would talk about noun phrases?

Niemann & Noone's _Generative description of an abnormal variety of
English_ provides a formal description of the syntax of an isolated
community where, due to Scandinavian influence, English sentences like
"Beware of the dog" have been replaced by sentences like "Beware of
dog the". Although older speakers only change the order of the
definite article, Niemann & Noone note that young speakers change the
order of the indefinite article as well, e.g. "Take hike a". They
argue that this provides strong support for recent work in syntactic
theory. They admit that their grammar occasionally fails to conform
to their corpus (1)
 ----------------------------------------------------------------
(1) The entire corpus is available in machine readable form for a
nominal license fee from the Oxford Text Hoard, who kindly granted
permission to quote these examples for educational purposes, provided
that no one else reads them.
 ----------------------------------------------------------------
e.g.,
	EXPECTED			OBSERVED
	cat the in hat the		cat in hat the the
	Hague The			The Hague
	another thing			nother thing a
	apple an day a			apple day an a
	B A C's				A B C's
	Arbor Ann			Ann Arbor

but they convincingly dismiss these cases as performance abnormalities.

Message 3: Re: 3.464 Innateness and Rules

I missed the discussion leading up to the exchange of June 3,4 re: subject
3.459. Therefore, I am not sure exactly what is meant by 'normative' rules.
However, it seems clear that it means either rules that are overtly taught, or
more relevantly to the discussion, rules which discriminate between what
(some) people say (not clearly distinguishing speech errors from nonstandard
stigmatised forms) from notions of correctness or grammaticality (in the
nonlinguistic popular sense) -- however the latter are obtained. In any case,
I have to agree with Alexis' message of 4 June, that crucial decisions are
made in sifting the data, and justification is not always as easy as many
linguists seem to think. Vicky's suggestions of 3 June are a case in point.
Without even questioning whether or not her list consist exclusively of
bona fide speech errors, the METHOD itself will not allow discrimination of
speech errors from stigmatised, commonly used, and grammatically constrained
rule-governed behavior which result in the following in some dialects of
English.
1) hold your cards where can't nobody see them
2) it don't be dark yet
3) I wouldn't did that (which can only have a habitual reading in the
 dialect considered here, i.e., it means "I didn't used to do that",
 or would you prefer "I used not to do that"? , and not "I wouldn't
 have done that".
not to mention banalities like:
4) he ain't got none
Do the same experiment as Vicky suggests with the above, and you'll
get lots of agreement that there is something "wrong", "funny" and/or
ungrammatical. This being the case, what does Vicky's suggestion prove?
It does not distinguish speech errors from stigmatised nonstandard
normative rules/grammatical constraints.
The issue here is not whether or not there are speech errors, but how to
develop a scientific method to distinguish them from stigmatised rule-
governed forms. To show how difficult this may be, even Labov, who may
have been the first to make the argument that I am calling attention to
here, made a mistake when he held up the following structure as non-
English, and thus he would presumably have considered it a speech error
if he had encountered it:
5) anybody doesn't know that (meaning "nobody knows that")
It didn't take long before he got jumped on by Irish English speakers.
It turns out to be perfectly acceptable in the English spoken in Ireland.
Thus, the distinction between speech errors and stigmatised norms/rules
is far from obvious IN PRACTICE. The interesting questions have to do
with what is a speech error when said by one speaker but not when said by
another. The implication is that a speech error may anticipate linguistic
change (and eventually prescriptive change). That is, change which either
relaxes or adds further constraints on grammatical rules. I think Vicky
would agree with this, because her most common use of speech errors is to
show what they reveal about the nature of linguistic rules. Where someone
may anticipate linguistic change in making a speech error, maybe we have an
indication that the rule is "fragile". That is, regardless of how a
particular speaker feels about the "error", the rule which makes him/her
feel that way is not supported by other grammatical rules which would
help it resist change. CF. the old controversy about the status of
quantifiers like "every" under negation, e.g., whether "everybody doesn't
know that" can mean "nobody knows that" vs. the other meaning. There are
speakers who insist that only the other interpretation is acceptable, and
yet have actually used the "nobody knows" in speech. Error? Who knows?
Is there really a constraint, or is this an example of the limits of the
depth to which "constraints" penetrate into a grammar, beyond which we just
have preferred and non-preferred STRATEGIES, which might result in such
reactions as "Yeah, maybe when somebody else says everybody doesn't know
that, they mean nobody knows that, but when I say it I mean some do, some
don't. If I meant "nobody..." I would say "nobody..." But, in the heat of
conversation, would this speaker do so? And if not, is it a grammatical
error for the speaker, or "just" a strategic error, in terms of the speaker's
(overtly?) preferred strategies?
Finally, with regard to the question of whether linguistic capacity is
something specific or a mainfestation of something else, the errors or
anticipations of change discussed above may be quite different from other
kinds of errors which "noone" would dispute or expect to become "norms" in
some community, e.g., long-range metatheses like "a cuff of coppee" -- but
the blends depending on lexical constraints are more problematic. Most
of Vicky's examples either are, or are intended to be, of this type.
It is not clear to me that long-distance metatheses are motivated by any more
of an innate linguistic capacity than typing letters in the wrong sequence or
putting the jelly on before the peanut butter. The line between strategies
and grammar is possibly the best place to look for innate linguistic
capacities. Blends may be unclear because they have different possible
motivations, both linguistic and non-linguistic. Does this make any sense?