Odd dummy subject markers

[Arnold M. Zwicky]

On Apr 3, 2006, at 11:10 AM, RonButters at AOL.COM wrote:

> A graduate student writes (in a paper on the semiotics of ethnic
> restaurant
> design):
>
> "On the sign it features a checkerboard pattern implying a tablecloth,
> whereas on the building it is simply a solid green line surrounded
> by yellow."
>
> I'm used to Southern substitution of "it" for "there," as in the
> second
> clause, but the "it" in the first clause seems weird to me.

others -- Charlie Doyle and Larry Horn -- have now suggested a
connection with the use of "it says" in quoting from books.  OED2 has
this usage, in subsection 3f for nominative uses of "it":
   In quoting from books, in the phrases _it says_, _it tells_, etc.
_Now arch. or colloq_.

it gives examples with "says" (from c1175 on), "tells", "reads", and
"follows well after" (in modern spellings), but only those with
"says" sound ok to me -- e.g. 1932 (from Chesterton) "...the world
teemed with quiet fun, as it says in the comic opera" and 1955 "In
Isaias it says that they fled into the land of Ararat."

i find this usage entirely natural; it hadn't occurred to me that it
was colloquial.  i find it especially natural in questions: "What
does it say on the sign?", "What does it say in your dictionary?"  i
can also get it in the past tense: "It said on the door that the
office was closed for lunch" etc.

what the OED doesn't say is what makes this usage so reasonable.  you
could argue that in "in the Bible it says..." etc., the "it" is in
fact referential; it denotes the text.  when the location is itself
conceptualized as a text (as the Bible is), the construction may seem
pleonastic, since alternatives like "the Bible says" will be
available.  nevertheless, it makes sense.

in contexts where the location is not naturally construed as a text
-- as in "On the door it says that the office is closed for lunch" --
the shorter alternative ("The door says that the office is closed for
lunch") has to be interpreted as metonymic, while the construction
with "it" is more transparent and more easily interpreted.

and with "is" + a quotation, only the longer alternative is possible:
"In Shakespeare it is 'all that glisters is not gold', not 'all that
glitters is not gold'", not of course "Shakespeare is 'all that
glisters...' ". (instead, "Shakespeare says..." or of course "In
Shakespeare it says...")

now, in Ron Butters's original example, the "it" can be understood
and referring to the depiction on the sign.  this would be an
extension of the "it" construction from text to a wider domain of
visual representation.  (or possibly a survival of an older wider
usage of "it".)  in addition, this way of looking at the example
allows the second "it" to be another instance of this
"representational" use -- what is depicted is simply a solid green
line surrounded by yellow -- rather than Southern-existential "it".

for me, representational "it" can refer only to text, not to other
representations (though "On the sign it shows a checkerboard patten"
doesn't sound half-bad), and the choice of verb is severely
restricted: for me, only SAY.  given the OED's cites, my pattern is
clearly a remnant of a construction that was formerly more
widespread.  (i wonder if other non-textual examples can be found in
older texts, or whether modern examples like ron's are an innovation,
an extension from the textual examples.)

arnold (zwicky at csli.stanford.edu)

------------------------------------------------------------
The American Dialect Society - http://www.americandialect.org