LINGUIST List 7.1243

Sat Sep 7 1996

Sum: unsubscribe TO or FROM

Editor for this issue: T. Daniel Seely <>


  1. Charlie Rowe, unsubscribe TO or FROM

Message 1: unsubscribe TO or FROM

Date: Sun, 01 Sep 1996 14:46:33 EDT
From: Charlie Rowe <>
Subject: unsubscribe TO or FROM
Many thanks for the responses to my query about the choice of preposition with
unsubscribe (TO vs.FROM). I received submissions from:

Frances Bond
Sarah Ogilvie
Karen Stanley
Dawn MacLaughin
Karen Ward
Tom McClive
Andy Kehler
Tracy Mansfield
Suzanne Hilgendorf
KD Caldwell
Tom Cravens
Steven Schaufele
Jane Edwards
Clodagh Lynam
L Hillman
Lynne Murphy
Norvin Richards
Donna Lillian
Lynn Messing
Stan Dubinsky
Alex Eulenberg
Chris Culy
Margaret Winters
Diana Maynard
Peter Daniels
Marc Picard
Barbara Silas
Dan Loehr
Ori Pomerantz
George Huttar
Glynis Baguley
Gordon Owen
Michael Robertson
Joe Foster
Andrew McCullough
Marie Egan
Mike Tugwell
Carsten Preust
Sharon Flank (special thanks to her for the very helpful reference she
 provided: Blejer, Hatte
R., Sharon Flank and Andrew Kehler, 1989. "On Representing Governed
 Prepositions and Handling
"Incorrect" and Novel Prepositions," in Proceedings of the 27th Annual Meeting
 of the Association
for Computational Linguistics, 26-29 June 1989, Vancouver, British Columbia,
 Canada, pp. 110-117.)

The overwhelming majority of respondents chose FROM, on the basis of

(1) semantics (here: transparent directionality) . Others mentioned that
 idiosyncratic, i.e.
nontransparent semantic variation might result in a different marker (TO):

Andy Kehler: "Since I don't know of any reason to think "unsubscribe" would
have an idiosyncratic marking, I would expect people to use the
semantically motivated one."

Mike Tugwell: "I think that you may get both FROM and TO as being
 "grammatical". Part of the
reason is that while people unsubscribe FROM a list, they send the
command "unsubscribe" TO the list(server)."

Michael Robertson: "What M.Lewis said in his book "The English Verb" (LTP 1986)
 comes to mind in
cases like this.
He contrasted (pp 41 - 43) the examples (1) "We met at London" and (2) "We
met in London", saying that although (2) is this less marked usage, (1) is
also correct if the speaker is thinking of London as a point on a voyage.
That is, it depends on the speaker's interpretation of the facts. London as
a place you enter ('in') or London as a point on a longer journey ('at')
However, an example like the one given can no doubt be conceptualised in
different ways, and so "unsubscribe from" may equally validly reflect the
speaker's perception.
"Unsubscribe to" could reflect the idea that (just as when you subscribed)
you were sending data "to" the list - regardless of the fact that the
result of this data will be that you are taken off/from the list. So, you
"subscribe to" and "unsubscribe to".
Perhaps, further doubt arises from variations in usage of words like
'bring', and 'take' (US/UK), as well as the 'newness' factor you referred

(2) syntactic-semantic identity with phrases like *desubscribe from* (JF),
 *resign from* (LH),
*delete from* (DL), *remove from* (AK), *exit/detach/disassociate from* (DM),
 *come from* (MR),
*depart from* (JE), and *withdraw from* (AM).

(3) the canonical lexical-morphosemantic pattern of *un-* pairs of this sort,
 like * load
onto/unload from, plug into/unplug from, pin to/unpin from, leash to/unleash
 from* (KD Caldwell);
*unfasten, unhook* (G Baguley).

 Other views under (3) included:

Norvin Richards: "The only cases in which I can get TO with some negative
 morpheme are cases where
the morpheme really means negation and not "undoing", as in "He's a
 non-subscriber to that list"
(bad, but better than with _from_)."

Similarly, Gordon Owen: "If forced to choose between the two, I'd go with
 "unsubscribe to" a list
or publication Rationale here is that the neologism "unsubscribe" carries its
 own negative so
that changing the positive "to" (toward) to the more negative "from" causes
 "unsubscribe from" to
smack of the double negative."

(4) empirical data. Two respondents offered empirical data from listserver

Alex Eulenberg, mailing list owner, unsubscribes FROM.

Glynis Baguley offers these comments: "My news-reader tells me that I have
unsubscribed TO, but I put that down to lazy programming."

Two other comments indicated for me that the use of TO with *unsubscribe* is
 not altogether

Steven Schaufele: "I'm pretty sure I would definitely say (how's that for a
speech act?) `unsubscribe FROM'."

George Huttar: "My immediate reaction was "unsubscribe from a list"--but that
 doesn't mean I won't
catch myself (or fail to catch myself) using "unsubscribe
 to" some day."

Further observations:

Many objected to the verb *unsubscribe*, preferring *desubscribe* or *cancel*.

 Interestingly, one "canceler" mentioned the choice of TO or FOR (with
(Marie Egan);
some " unsubscribers" preferred FOR over FROM.

Unsurprisingly, "desubscribers" chose FROM.

Carsten Peust argues for FROM on the basis of German (which is another squib

I submit that my intuition demands TO, which I would like to explain by invokin
 factors (1), (3),
 (4) in some form, and an additional factor (5); even so, I believe that
 invoking (2) would not
necessarily have to lead to a judgement for FROM. My thinking is summarized as

(1) The meaning of unsubscribe PREP is not completely transparent, because of
 (3) and (5).

(3) Given the two allomorphs of *un-*, it is not immediately clear whether the
 prefix *un-* in
*unsubscribe* has the meaning of [[de-]] or [[non-]]; i.e., whether
 directionality is invoked.
Typically, the [[de-]] meaning attaches to verbs and the [[non-]] meaning
 attaches to nouns; but
there are clear exceptions (4), some from colloquial speech (a), some from
 neologisms (b), (c):

(a) to uninvite someone TO/*FROM a party
(b) to undelete something *TO/FROM a document
(c) to unsend an email TO/*FROM someone

If we invoke (2) here, *uninvite* would require FROM (on the basis of *reject*
 or *expel*, etc.);
*undelete* would require TO (on the basis of *restore*); * unsend* would requir
 FROM (on the
basis of *recall*, i.e., to request that the Mail Service recall a letter from
 the addressee).
So, while (2) can perhaps adequately explain the motivation for an established
 and unequivocal
choice (i.e., as recorded in the lexicon), it cannot with complete success be
 invoked to proscribe
unequivocally the usage of FROM over TO, in my view.

(5) TO may befor speakers who accept it--grammaticalized to the notion of
 *subscription* (hence,
a noun), rather than to the process of *subscribing* (i.e., the verbal form),
 via the attachment
criteria in (3). This semantic attachment could also be reflected in the form
 of a
(morphological) bracketing paradox:
(d) [[un][subscribe]][FROM] versus
(e) [un][[subscribe][TO]]

The bracketing in (d) would seem to obliterate the semantics of *subscribe*.
 I.e., the rule is so
general that most any [[un][VERB]] would take FROM, productivelywhich of course
validation for invoking this rule, at least optionally. In (e) the semantics o
 *subscribe* are
crucial to choice of the preposition.

 This would explain the phenomenon that Marie Egan noted with *cancel a
 subscription TO* (where
*cancel* cannot subcategorize for a preposition (TO); it selects a direct objec
 rather than a PP
argument). She and others also noted FOR as a possiblility here. The engagemen
 of a semantically
less "loaded" preposition (FOR) here could very well signal a "neutralized"
 interpretation for
speakers who prefer it. It seems that FOR effectively removes the directionalit
 factor, and
substitutes a general "association" factor, which should be less likely to
 trigger a
counterintuitive reading (as some speakers felt with TO), or a
illformed reading (with FROM).

Ultimately, if *subscribe* subcategorizes for TO, it does so perhaps
 semantically; in my view,
*unsubscribe* may do so lexically (on the model of the base verb *subscribe*)
 for those that
accept TO. Those speakers that accept only FROM do so on semantic grounds.
 Speakers that prefer
FOR seem to take a "middle road" between lexical and semantic based
 determination. This would make
sense, given that morphologization procedes from semantic designation to lexica
Thus, seen in stages or levels of grammaticalization:

1st stage: unsubscribe FROM
2nd stage: unsubscribe FROM/FOR
3rd stage: unsubscribe TO

In particular, George Huttars comment (if I understand him correctly) would
 seem to support the
notion that *unsubscribe TO* may reflect a 3rd=-stage designation. Whether or
 not this usage will
"stick," or fossilize for some linguistic system (e.g., colloquial, jargon,
 standard etc.)
remains, of course, to be seen.

Thank you all again for your input. I welcome any further comments (by public
 or private
response) that you may have on the subject.

Charlie Rowe
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue