From: Stevan Harnad <[log in to unmask]>
Date: Thu, 9 Feb 2012 01:42:01 -0500


On 2012-02-08, at 9:41 PM,  Brian Harrington wrote:

> Maybe I'm just pedantic, but shouldn't the version accessed be the
> version cited?  Admittedly, in the humanities fields I'm most familiar
> with, citation is often accompanied by quotation, or at least
> references to page numbers, so the need to consult the
> version-of-record seems self-evident.  But even granting different
> citation practices in other fields, the idea of citing something that
> you haven't actually seen strikes me as going too far.  If the
> refereed final draft is the only version that the researcher needs to
> consult, why not cite it?  This seems especially true if the draft is
> the version that the reader is more likely to have access to.

(1) The OA version is for would-be users who do not have access to the
version-of-record because their universities cannot afford access.

(2) For such users, OA means the difference between being able to
read, use, apply and cite an article, or being unable to do so.

(3) The reason the refereed final draft is the version that OA
mandates specify is that it has fewer publisher restrictions
(copyright, embargoes) on it than does the version-of-record..

(4) One can quote from the OA version (which is usually identical to
the version-of-record, apart from minor copy-editing).

(5) The locus of quotes can beindicated by section headings and
paragraph numbers when page numbers are unavailable.

(6) One reads and quotes from the version one has accessed (whether
the paper journal on the shelf, the publisher's archival version, a
photocopy, or the author's refereed final draft), but one cites the
canonical version-of-record in all cases.

(7) One can (and should) add to the canonical citation the URL of the
OA version.

Stevan Harnad