LINGUIST List 5.1227

Thu 03 Nov 1994

Disc: Corpus analysis of -BODY/-ONE

Editor for this issue: <>

Directory

1. , Re: 5.1196 Corpus analysis of -BODY/-ONE
2. Michael Newman, Re: 5.1196 Corpus analysis of -BODY/-ONE
3. Lou Burnard, more data on -BODY/-ONE

Message 1: Re: 5.1196 Corpus analysis of -BODY/-ONE

It seems to me altogether possible, even likely, that there is the
following interaction, namely, that the particular lexical form <everyBODY>
has come to acquire, for many speakers certainly (me, for instance, a quasi
native speaker, and I share Ellen Prince's intuition concerning her
sentence #1) a distinctly collective sense, whereas <every-ONE> maintains,
again for these speakers at least, a distributive sense, at least as a
living option. Thus

 1'. Everybody came, bringing their respective wives.

seems especially good, though, oddly enough maybe,this instantiates the
special claim that, for this particular lexical -BODY item, the
formal/informal register distinction has at least contextually?) collapsed.
 Now, it seems plausible, but no more than that in the present state
of knowledge (or of ignorance) that there is a more underlying interaction,
namely, that there may be something more naturally colloquial about using a
collective rather than a distributed plural when an option presents itself.
Actually, this follows from an algebraic, non-Fregean view of
quantification, which I have written on but cannot go into ion this
message, save to indicate summarily that it is a capital mistakme to
imagine that, for instance, the same semantical reopresentation ought to
apply, even for logico-deductive applications, to say, <each/every> and
<all>, respectively.Thus, invoking a correspndence that I shall not unpack
here,

 i. EVERYBODY brought their wives/*Everybody brought his wife
 i'.They ALL brought their (respective) wives/Each man brought his
wife.
 ii.*Each man brought their wives/ They each brought their
(respective)

wives.
This surely suggests, does it not, that the interaction under review is
distinctly asymmetrical (or, I think, rather, non-symmetrical), for, the
<each> form of quantification, by logical entailment of a fairly obvious
kind (from the distributed, that is, serial, meaning, to the
collective/mass, can take a collective sense, whereas the reverse (you
cannot non-arbitrarily unpack a mass b ecause there is no algorithm in this
direction correspnding to the seial addition inherent in the other
direction) is not available.
 ALL, then, (and the other collective-sense quantifiers and the
nominals that condense such quantifiers (as in the BODY cases) quantiify,
i.e., select as a choice function, over a Power Set, choosing the element
equivalent to the improper 'subset of the whole'. EACH-type quantification
is over the elements of a Set, taking the index 'i' ranging, algebraically,
successively from 1 (or, in some cases, 0) to N, the cardinality of the set
in question.

F. K. L. Chit Hlaing

Message 2: Re: 5.1196 Corpus analysis of -BODY/-ONE

In my dissertation (1993: A theoretical and descriptive study of epicene
pronouns) I analyzed a corpus of 24 US TV talkshows. I got 16 examples of
NPs with "every" coreferent with a pronoun. The distribution of the nominal
element can be found below:

 10 -body
 3 -one
 3 lexical nominal(s)

The lexical nominals were "member," "grand jury and trial witness," and
"sales person in the store." In addition one of the cases with -body was
"everybody involved" and another was "everybody else."

So far this seems to be along the lines discussed by Ellen Prince and Jane
Edwards. Where my data varies is that in every case the coreferent pronoun
was THEY, as was incidently the case with EACH and NO. As far as the
analysis is concerned, my results, not just from quantifiers but from NPs
with other specifiers, indicated that there are various factors at work in
the selection of the pronoun, as indeed Edwards argues.

The following seemed important:

NOTIONAL PLURALITY:
Obviously the case with EVERY, NO, and EACH. In some cases a singular
pronoun is impossible when there is plural semantics and formal
singularity, as is well known from studies of quantifier scope. Even
however when it is structurally possible, the use of singular pronoun can
be awkward as Prince intuited, though I suspect more than just the nominal
element of the compound is involved. In any case, in two out of the three
EVERYONE cases was clearly outside the scope, so the plural pronoun was
inevitable. In the third case it was in a parenthetical (from Donahue) and
I'm not sure of the fine points regarding the formal semantics:

"Since I don't know how everyone develops, what their sexual attraction is,
you could easily have been into infantilism, or purple handkerchiefs, or
same sex experiences"

GENDER STEREOTYPES:
This issue didn't arise with EVERY cases, since all were with THEY, but in
others I found that stereotypically male generic referents (e.g participant
in congressional sex scandals, and lumberjacks) more frequently received
the pronoun HE with other types of antecedents.

INDIVIDUATION:
This is a rather fuzzy, but I think inevitable, semantic category that
refers to the degree with which interlocutors are using a referent as an
individual versus how that referent is being used as a generic. The
fuzziness is necessary because logical real existence does not map into our
everyday reference. A personification (discussed with regards to pronouns
by McConnell-Ginet) is a generic referent treated as if it were an
individual. The opposite case, a real person or thing treated as a generic
also exists and is discussed with regard to definiteness by Talmy Givon. I
found a couple of cases like this.
from a Geraldo, the reference is to Princess Grace:

"How do you fall in love with somebody at a photo opportunity, not see them
for eight months and the next time you see them you're engaged."

from a different Geraldo:
"I heard, -saw someone run by and they were screaming."

Individuation is clearly the same thing that Jane Edwards noticed in her
point about how singular pronouns are associated with specific people. And
similar point was also made by Pinker in THE LANGUAGE INSTINCT in the
chapter on prescription.

I found the best analysis of the whole issue can be made by adopting the
agreement theory developed by Michael Barlow in his '88 dissertation A
SITUATED THEORY OF AGREEMENT published last year in the Outstanding
Dissertations Series. Anybody else have any data?
Michael Newman
Dept. of Educational Theory & Practice
The Ohio State University

MNEWMANMAGNUS.ACS.OHIO-STATE.EDU

I can't refuse Jane anything. Here is some more data, just in from the British
National Corpus. (Well, actually only from a very small 1% sample of it)

Corpus size: 625,000 written and 182,000 spoken words (as defined by CLAWS)

Proportional raw frequencies for some common -one/-body words:

 written spoken total
 anyone:anybody 57:31 17:32 74:63
 someone:somebody 72:38 36:64 108:101
 no-one:nobody 18:57 2:18 20:75
 everyone:everybody 81:43 12:23 93:66

What I find interesting here is the *absence* of any strong evidence for
preferences for -one or -body form (excepting the pathological case of
"no-one") when you look at the whole corpus, rather than separating out the
spoken and written component parts. I'm optimistically interpreting that as
evidence that in the corpus as a whole we've got a good spread of textual
variety.

I'll repeat the analysis on the whole BNC when we've finished building its
index...

Lou Burnard

p.s. I looked for "No one" too -- it appears only once, in a sentence beginning
"No one person would ..."