LINGUIST List 4.737

Tue 21 Sep 1993

Disc: Reciprocals, Metanalysis

Editor for this issue: <>

Directory

1. , Each other
2. , 4.725 Metathesis
3. "Leslie Z. Morgan", More on Metathesis

So far as I can remember, the syntactic literature that I've read on
"each other" assumes that this phrase is a true reciprocal. That is,
(1) should mean that John saw Mary and Mary saw John.

 (1) John and Mary saw each other.

This is often the case, and certainly is true of (1). But I've been
noticing that many people use "each other" to mean something very
different, as in (2) and (3).

 (2) Not everybody knows how to get to the restaurant, so we'll
 all have to follow each other.
 (3) Those two boxes were stacked on top of each other.

If "each other" were a true reciprocal, following each other would not
get anyone anywhere, and two boxes being on top of each other would be
logically impossible.

What the phrase is intended to mean is something more like "each
member of the group is in a transitive relation with one other member
to form a linear sequence, except, of course, the first member, which
does not participate in the relation on one side, and the last member,
which does not participate in it on the other side."

Has this been discussed in print? This seems like an interesting and
difficult reading to capture with formal semantics. Or do this
reading and the `true reciprocal' reading both fall out of some more
general characterization, with pragmatics determining which is
appropriate for a given context?

I'm assuming that my understanding of `true reciprocal' is that which
is generally shared -- that is, that all members of the group
participate in the transitive relation in both directions with all
other members. Certainly (4) would not normally be used to mean that
John followed Mary, but Mary didn't follow John. And (5) would be
considered false if Chris loves Pat, Pat loves Sandy, Sandy loves Kim,
but Kim hates Chris.

 (4) John and Mary followed each other.
 (5) Chris, Pat, Sandy and Kim love each other.

So what's different about (3) and (4) vs. (4) and (5), aside from the
implausibility of the reciprocal readings in (3) and (4)?

I'd be interested to hear what anyone has thought, observed or written
about this, in English or other languages.

 Dale Russell

The phenomenon illustrated by the change from a nadder to an adder is called
metanalysis.
Bernard Rochet
Department of Romance Languages
University of Alberta
Alberta, Canada T6G 2E6

Since numerous folks are bringing in other languages and their
terminology for the "an adder" phenomenon, I might point out that
Rohlf's "Grammatica Storica" in its Italian translation speaks
of "concrezione dell'articolo" and "discrezione dell'articolo"
which is where I borrowed my terms. I too was unaware of English
terminology. Rohlfs has some nice examples too if anyone needs
a few more from other languages, including French as well as
Italian dialects.

Leslie Morgan
Dept. of Modern Langs. and Lits.
Loyola College in Md.