"soon before"?

[Arnold M. Zwicky]

On May 18, 2005, at 1:20 PM, Alison Murie wrote:

> Seymour Hersh in  /Chain of Command/ on pp. 355-6 writes:
> "....security official said in an interview _soon before_ the
> transfer of
> sovereignty that..."  [my emphasis].
> This looks & sounds very odd to me;  I would expect _shortly
> before_here.

"shortly" is fine, but "soon" sounds perfectly good to me here.  in
wh questions of extent, with "how", "soon" is very common: "How soon
before the game did you arrive?"

outside of contexts with "how", "soon before"  is common, though not
nearly as connon as "shortly before" or "soon after".  in raw google
web hits:

   --> "soon before" -how: ca. 31,900
   "soon after" -how: ca. 2,670,000

   "shortly before": ca. 2,960,000
   "shortly after": ca. 11,400,000

here, "after" has the edge over "before".  modified by "just" or
"long", however, "before" has the edge, though not as big an edge:

   "just before": ca. 9,770,000
   "just after": ca. 4,440,000

   "long before": ca. 8,510,000
   "long after": ca. 5,260,000

(overall, "before" and "after" are roughly the same in frequency,
894m and 882m.)

admittedly, "soon before" (marked with an accusatory arrow above) is
very low compared to the other combinations.  so there does seem to
be some dispreference.  or even the possibility that some speakers
have an odd gap in the pattern of modifying "before"/"after" by
"soon"/"shortly"/"just"/"long".

in general, i have no idea where these asymmetries come from.

arnold (zwicky at csli.stanford.edu)