usually always

Laurence Horn laurence.horn at YALE.EDU
Wed May 23 04:04:22 UTC 2007

At 10:14 PM -0500 5/22/07, Cohen, Gerald Leonard wrote:
>I didn't have "usually almost always" in mind.  Rather: "usually"
>(e.g., "I usually park there") + "almost always" (e.g., "I almost
>always park there") blend to "usually always" (e.g., "I usually
>always park there").
>As for the few examples of "usually almost always" (mentioned in the
>message below), they too represent a blend.
>Gerald Cohen

But the same argument would apply equally well to "often", which
could blend with "almost always" to form "often always", yet this
rarely occurs.  (There are many irrelevant hits, involving multiple
choice options on questionnaires.)  Nor is "sometimes   As James
Harbeck notes, the adjacent scalar position of "usually" and
"always", given the scale <always, usually, often, sometimes>, seems
crucial here.  The same process occurs on the negative side with
"rarely ever" (or "seldom ever"), which is redundant (and hence
scorned by prescriptivists) in the same way that "usually always" is,
although an analysis associating this with "rarely if ever" (parallel
to "usually if not always") would be possible here.

The question I have about the blend analyses is in determining before
the fact which elements are likely to blend and which ones aren't;
after the fact, there's never a problem.


>From: American Dialect Society on behalf of James Harbeck
>Sent: Tue 5/22/2007 9:49 PM
>Subject: Re: usually always
>>Looks like a blend: "usually" + "almost always."
>FWIW, I've never heard "usually almost always." It does get 805 hits
>on Google, but a lot of them are actually "usually (almost always)"
>or "usually, almost always" or constructions to that effect.
>("Usually always" gets 304,000 hits, but many of them are from forms
>in which they are adjacent items on a scale -- I found 13 in the
>first 50 that were actually the construction in question.)
>James Harbeck.
>The American Dialect Society -

The American Dialect Society -

More information about the Ads-l mailing list