Re: Free Access vs. Open Access

From: Stevan Harnad <>
Date: Wed, 31 Dec 2003 23:46:40 +0000

On Tue, 30 Dec 2003, Michael Eisen wrote:

> First, for the sake of clarity, can we just agree that, whatever relative
> value you place on the two, free access and open access are not equivalent
> and that it does no one any good to confuse the two.

We can agree for the sake of clarity that an attempt is being made to
formulate a distinction between "open access" and "free access." But
it is my contention that it is the distinction itself that causes the
confusion, because it is empty and divisive, leading to what should be
open-access allies saying to one another: "You are not promoting 'open
access' but merely 'free access.'"

No, they are not equivalent. But let us be very clear on what the putative
differences between them are, so we can weigh whether they have any substance.

> Free access gives all potential users immediate and permanent toll-free
> access to the text at a single fixed point on the internet (e.g. a
> self-archive or a journal website).

At a single fixed point on the internet? How many points does toll-free
full-text access have to be provided *from* if everyone, anywhere,
any time, can get it from that "single fixed point"? (Is this just
about mirroring, caching, band-width and redundancy?)

> Copyright would (in general) reside with
> authors or their assignee, and users would have fair use rights, such as the
> right to read, print, crawl and mine this copy of the article, but in
> general would not have any further rights, such as the right to redistribute
> or make derivative works.

Correct. They can't *republish* the work on paper (without
permissions). But they can redistribute the URL of the "single fixed
point" all they like -- which is what online-age open access means!

> Open access grants all of the rights inherent to free access

Free access is not a *right,* it is a *capability* -- toll-free online
access to and use of the full-text -- a capability provided either by the
publisher (BOAI-2, open-access journal publishing) or by the author (BOAI-1,
open-access self-archiving of toll-access articles). No new rights need
to be granted to provide toll-free full-text access (though it is nice
if -- when it is done via self-archiving -- the toll-access journal is
"green," i.e., gives its official blessing to author self-archiving,
as 55% already do).

> but the
> copyright holders grant (by signing some form of license) all users
> additional rights, especially the right to redistribute and make derivative
> works, in general asking only that the original work be properly cited. (The
> different definitions of open access are not identical, but all essentially
> say this).

No, I do not think granting the right to *republish* (i.e., include the
text in someone else's subsequent publication, or to publish it on paper
in a collected work) is at all inherent in the meaning of open access;
nor do I think it is important or necessary for someone else to be able
to republish my text in his online article, or to republish my text in his
online collection of papers (though I, for one, would almost always
grant anyone the permission to do so).

The only thing that is urgent, and important, is that any would-be user
of my text, webwide, should be able to read, download, store, print out,
computer-process, use, build-upon, apply and cite my text without his
institution's having to pay an access-toll to do it. *That* is open access.

(If you want to redistribute my text, send out the URL! no need to waste
bandwidth or paper sending out the text itself.)

> I hope we can agree that these are not equivalent, so that we can get onto
> the more important question, which is is this a meaningful difference - that
> is, are the additional rights given to users under open access meaningful,
> and does granting them benefit authors, readers and the research community
> in general. Obviously, you think the answer is no, while I think it is yes.

I am not sure how I, as author, am benefited if I grant a blanket right
to *republish* my text, either as part of someone else's published text,
online or on-paper, or as part of a collection someone is putting together
on paper. I think I'd rather be consulted, case by case. Otherwise they
can just insert the URL.

So much for *republishing* my text. But anyone is certainly free to *use*
my text in any way they wish, in their research or teaching. That's
open access. It pertains to use. The right to republish should be
called something else. It's not an open-access matter (though perhaps
an open-access *publication* matter: and that's the point, open-access
provision is not the same thing as open-access publication!)

> I address your 6 points below, although I think that you are conflating two
> different points.

That may be -- but it's sure a relief to see someone else using the word
"conflate" for a change... ;>)

> 1) Open access is unnecessary - everything that a user would want to do with
> a paper they can do with free access.
> 2) Open access is an obstacle to free access - demanding that publishers
> provide open access delays or obstructs their providing free access.
> I hope my answers below address why I think these are both incorrect.
<sh> So here is my list, again:
<sh> Once the full-text is immediately, permanently, and ubiquitously
<sh> (i.e., webwide) accessible toll-free, so any user anywhere, any time,
<sh> can read the full-text on-screen, download it, store it, print it off,
<sh> search/grep it, computationally process it, etc. -- which any user can
<sh> do if the author self-archives it -- the further rights and uses that
<sh> distinguish "free" from "open" become either moot or supererogatory:
> If all you are concerned about is getting toll free access to papers - in
> the form that they exist in self-archives or on journal websites - then the
> distinction between free and open is superfluous.

Well I wouldn't say that's *all* I'm concerned about. (I still have a
soft spot for goodness, truth and beauty, for example, and it would be
nice to do something about world hunger!) But I would certainly say this:
If the 2,500,000 annual articles in the 24,000 peer-reviewed journals
were all accessible toll-free as of tonight (as they could be!), I would
not be spending my new year's eve archivangelizing!

(And if that had been the blissful state of the research literature of
its own accord, I certainly would not have written my subversive proposal
10 years ago or spent all this time promoting -- what? XML? Republication

But given that we are nowhere near that blissful state at this time,
it seems ironic that we should be contemplating what *more* we might
want! It's rather like conducting a crusade to make the world hunger-free,
but not being ready to call the (unattained) state in which everyone
can eat their fill "hunger-free," but only "hunger-quenched," reserving
"hunger-free" only for a state in which all the food is also organic.

It isn't that I don't prefer organic food to supermarket food; it's
just that, in the starving world we are in, insisting on a free/quenched
distinction seems rather beside the point...

> However many readers
> (myself included) like to read articles in a familiar and user-friendly
> format that is often very different from the deposited version

But Mike, you are not asking yourself the right questions! You should be asking:

(a) Do I prefer no access (as now), to access in something other than
my preferred format? (That's the *real* divide between the starving and
the fed, not the organic tofu!)

(b) Doesn't it make more sense to work to free the full-text from tolls,
such as it is, and then apply ways to reformat it to your liking, rather
than to insist that the preferred format must be the one that is freed,
otherwise it's not worthy of the name "open access"?

> (in my
> opinion, this is one of the reasons that self-archiving isn't as popular as
> it should be).

(i) There are at least three times as many articles a year made
open-access via open-access self-archiving as via open-access
journal-publishing, so if self-archiving is not as popular as it should
be, open-access journal-publishing, with the "right" format is even less
popular! So that can't be it!

(ii) The reason *both* OA self-archiving and OA journal-publishing are less
"popular" than they should be is that the research community still knows
far less about OA and its benefits and its feasibility than it could
and should (and will).

(iii) Even so, OA self-archiving is more underexploited than OA
journal-publishing because the 1000 OA journals (out of the total
24,000 journals) are much closer to their immediate peak capacity today
(which is about 5% of the total refereed literature) than self-archiving
(whose capacity is 100%, 55% already wearing the publishers' "green"
"welcome" sign, yet still only about 15% full).

(iv) The remedy for this is certainly not to (mis)inform the uninformed
research community that self-archiving really only provides "free access,"
whereas what they really want and need is "open access" -- which only
OA journals can provide!

(v) It is certainly not true that the self-archived literature is not
getting heavy use, even if some of it is not in some users' preferred
format! Steve Lawrence's 336% estimate of the increased impact was based
on arbitrary self-archived formats, not optimal XML!

(vi) Self-archiving is hence plenty popular with users, as well as with
self-archiving authors. It is only those (the majority of the research
community) who don't yet know that it is possible (and legal) that it
is not yet popular. But that can and will be remedied.

> It is possible to take all of the free access articles and
> convert them into a more flexible format (e.g. the publishing XML being used
> now by PubMed Central) where users could control the way in which the text
> is rendered. I believe many readers would value this option. However, under
> free-access this is not possible without getting permission from every
> author or copyright holder, while under open access this is not only
> allowed, it is encouraged.

Given software for converting a toll-free full-text into their preferred
format for use, what on earth do you imagine would prevent users from
applying it to the self-archived articles they download?

> Ditto for someone who wants to translate into
> another language (by machine or by hand) a body of free access works.

Local automatic translation can be done on the downloads by precisely
the same kind of means. And I'm certain that a harvester-service that
wanted to do the same sort of thing with the full-texts in order to
make the whole corpus full-text searchable in another language could
get permission to do so -- supplying a converter to apply locally,
if the user decides to download the full-text (from its home-URL).

(I really think these objections are next to trivial. The best thing
one could do now would be to set aside these far-fetched organic-food
scenarios on a near-nonexistent toll-free corpus, and set about freeing
it from the tolls! The existence of that toll-free corpus itself will
then be by far the strongest impetus and incentive to generating these
further refinements and services. But insist instead on their first
becoming organic now, a priori, and you simply strengthen the tolls'

> It seems to me that there are two issues here. 1) Rights description. Many
> authors who make their works freely available through self-archiving would
> be happy to allow these uses, but don't currently have a way to say this.

Sure they do: If someone wants to *republish* my texts (as opposed to merely
accessing and using them), they need only ask.

> This is why it is important to PLoS that we use the creative commons license
> that describes user rights in advance.

That's just fine, if PLoS wishes to do that. But what has that to do with
the free/open distinction, and with wishing to deny that self-archived
articles are open-access articles?

Licensing and license modifications are publishing matters. But
open-access publishing is not synonymous with open access! That is not
how or why BOAI came into existence: As I noted in my reply to Peter
Suber, the BOAI is not the "BOAPI" ("Budapest Open Access Publishing
Initiative"); it consists of *two* strategies for providing open
access: BOAI-1, self-archiving toll-access journal articles and BOAI-2,
publishing articles in open-access journals. And BOAI-1 has no need to
use the creative commons license! (Not even BOAI-2 *needs* to use it to
provide open access, though they can if they prefer.)

> I see that the open archives group
> has launched an effort to embed rights descriptions with text, so I think
> this case will be covered.

Yes, but the "open" in the Open Archives Initiative (OAI) is not the
"open" of open access in the Budapest Open Access Initiative! The OAI is
concerned only with *metadata*, not full-text, and it is concerned with
the interoperability of the entire digital literature, both toll-access
and open-access (and not just journals but books, multimedia, software
-- all digital contents). So *of course* OAI will be interested in
describing, in its metadata, the rights for its full-texts (e.g., its
toll-access full-texts)!

> 2) Rights restrictions by publishers. I think you
> feel that many publishers are/would be willing to allow self-archiving, but
> only on the condition that they retain all other rights to the text, and
> that if they are allowed to do this, then they would be willing to provide
> free access.

Something much simpler than that: Open-access provision via self-archiving
need not concern itself with rights. The only relevant right is the right
to self-archive! The rest comes with the territory. All the rest of the
toll-access journal publisher's copyright transfer or licensing agreement
can remain essentially as it is -- as long as the author can self-archive
(and the preprint + corrigenda strategy ensures that he always can
self-archive, even under the most restrictive of rights agreements).

> The problem, in my mind, is that this would prevent the kinds
> of uses that I describe above (which are, in my opinion, only the tip of the
> iceberg).

Please itemize the iceberg, as I have already taken care of all the tips
you've mentioned so far:

(i) Preferred format conversion? Go ahead, do it on the locally downloaded

(ii) Translation? Ditto.

(iii) Republication? Nolo contendere. This is not an open-access matter. Please
take it to the organic-food counter!

> You apparently think that getting free-access is the most
> important thing, even if the manner of achieving free access precludes
> optimal use of the material,

No, I think toll-free full-text access provides optimal use and the rest
is superfluous frills, relative to the real and urgent problem at hand
(that 85% of the literature is *not* yet accessible toll-free). Moreover,
once we achieve the toll-free full-text access, the frills will soon
come too.

> while I think this is a needless compromise of
> the interests of the scientific community and the public to the narrow
> commercial interest of publishers.

Needless? What alternative are you offering? Waiting for Gold? (23,000
journals to go!) I'd much rather have the "compromise" of immediate open
access! And the narrow commercial interests of publishers have absolutely
nothing to do with it. They are not the ones holding us back from open
access! It is the research community's own unawareness of the benefits of
open access -- and the means of attaining them, immediately. But that
unawareness will be remedied.

> Again, this is an open argument, but it's
> important not to glaze over this argument by pretending that free and open
> access are equivalent.

Toll-free full-text access *is* open access (and it is something we do
not have!). Republication rights are not the same thing as open access
-- and we don't have them either, but they are *far* less important or
urgent or within-reach than open access -- and they will come soon after
open access anyway.

<sh> don't need a further specified right to download, store, process or
<sh> print off any of the other material that they can download, store and
<sh> print off from the web -- as long as the material is itself not pirated
<sh> by another consumer, but provided by its own author, as is the case with
<sh> one's own self-archived journal articles.
> What you are saying, in essence, is that fair use gives all users all the
> rights they need, and that there is no need to specify any additional
> rights.

Fair use! We *have* fair use! What we *don't* have is toll-free access
(also known as *open access*). Have we gotten so caught up in the desert
that we've forgotten that we don't have the main course?

> This is simply incorrect, a good example of a use that is not
> allowed under fair use is the inclusion of text in course readers, which (at
> least in the US) is not covered by fair use.

Please include the URL of the toll-free version in the text. Students
are *real good* at using the web these days.

And then get back to the hard job of getting researchers to provide the
full-texts behind the URLs...

> Nor are other forms of
> aggregation.

Online aggregators: Ditto.

On paper? Nolo contendere. Open online access never promised open on-paper access.
Open online access was a new possibility opened up by the online medium. If we're
back to paper publication, all bets are off.

> For example, I might want to print a series of virtual
> journals that contain the best works in given fields, and send them (by
> subscription) to scientists in that field.

Send them the contents and the URLs.

> I think a lot of people would
> greatly value such a service, but it is not possible under narrow
> free-access.

It is fully possible with open-access (sic) self-archiving of the full-texts of
the toll-access articles in question.

<sh> (3) NO NEED OR RIGHT TO RE-PUBLISH: There is no need or justification
<sh> for demanding the further right to re-publish a full-text in further
<sh> *print-on-paper* publications ("derivative works") when it is already
<sh> ubiquitously accessible toll-free. That was never part of the rationale
<sh> or justification for demanding free/open access in the first place. What
<sh> ushered in the open-access era was the newfound possibility of providing
<sh> all would-be users with free, ubiquitous *online* access to texts,
<sh> thereby maximizing their research impact. This newfound possibility,
<sh> created by the Web, had nothing whatsoever to do with the right to
<sh> re-publish those texts on paper!
> This may never have been your rational for demanding FREE access, but it was
> a key part of my and many other users demands for OPEN access. My reason for
> getting into this in the first place was my desire to better link
> experimental data (in my case genome sequence and experimental genomics
> data) with the primary literature.

Get the research community to provide open access to that primary
literature -- either by publishing it in a suitable OA journal, if
there is one, or else by publishing it in a suitable TA journal and also
self-archiving it. Then link it.

> I wanted to build a database that would
> hold our experimental data along with a large collection of relevant papers
> that would be marked up in a unique way that allowed for integrated browsing
> of the data and literature. The only practical way to do this is to have
> local copies of the papers in our database, something that is needlessly
> precluded by strict free-access.

The still more practical way is to link to their URLs, wherever the
papers reside. (For local use, you could download the full-texts of the
articles in question locally. But if you are providing a secondary service
to other users, it will have to be set up so they download their own
versions directly from the source URL, rather than yours. Silly detour,
but transitional. Once it's all open-access, no one will care where the
papers sit.)

<sh> It may be that (some) open-access journals do not need or want to
<sh> have exclusive publication or republication rights. But open-access
<sh> journal-publication is not the only form of open-access provision.
<sh> Author/institution self-archiving of one's own toll-access journal
<sh> articles is another way to provide open access, and a much more
<sh> immediate and powerful way than to wait for toll-access journals
<sh> to become open-access journals.
> Yes, it is true that open access publishing is not the only path to open
> access. However, it is also true that self-archiving does not provide open
> access - it provides free access. While self-archiving may be more
> immediate, unless it ultimately leads to open access it is not more
> powerful.

What you are saying is that (in your view) self-archiving is only a path
to open access if it leads to open access *publishing*.

I am afraid this is our most fundamental point of
disagreement. Self-archiving provides toll-free full-text access,
something we lack and urgently need today, yet you are not prepared to
call what it provides open access. The difference between the access
that self-archiving provides and the access that open-access publishing
provides is (based on the above) republication rights. The difference
between the access that self-archiving provides and the access we have
now is tolls, access-denial, and research impact-loss.

Do you really think the motivation for open access is republication rights, rather
than access/impact-denying tolls?

I freely admit that for me the meaning of "open-access" (and its un-named
precursor, before the BOAI) has always been in relation to *toll-access*:
non-toll (i.e. toll-free) access = open access. And non-open access =
toll access.

It's too late for me to switch terms now. We're going to have to fight
it out for the referent of the term. I think I have argued here for why
the free/open distinction is empty, amounting to on-paper republication
rights plus potential secondary service rights on a still glaringly
absent primary toll-free corpus. (And the rest of the putative free/open
differences are all nullified by linking the URL).

I shall continue to use open access to refer to what I think it naturally
means: toll-free access, and I shall continue to refer to BOAI-1 and
BOAI-2 as the two means of providing it.

I hope that the confusion that will result from others using the term
in another way will not work too much to the detriment of open access
and the research community's at last providing it, as soon as possible.

<sh> the open-access journal's republication policy on the definition of
<sh> what counts as open access itself would be to impose an arbitrary and
<sh> unnecessary constraint on the second (and larger) of the two means of
<sh> providing open access. It is one thing to ask toll-access publishers to
<sh> support author/institution self-archiving, so as to maximize the impact
<sh> (usage, application, citation) of a text by maximizing access to it
<sh> online; it is quite another thing to demand that toll-access publishers
<sh> agree to put anyone and everyone on a par with themselves, in having the
<sh> right to publish that text in print. That would only serve to provoke
<sh> (justifiable) toll-access publisher opposition to self-archiving --
<sh> and hence to open-access provision by that means.
> Now this is a completely different point. First, there seems to be
> conflation of open access and free access here again.

Not conflation. The continuing use of open access to mean toll-free
access and the bracketing of republication and secondary-service rights
as either irrelevant to open-access provision or premature and likely to
come with the territory (for the reasons I have given).

> While it is possible
> to provide open access by self-archiving, without the redistribution rights
> it is only free access.

This sentence sounds self-contradictory: While it is possible to provide
X via Y, without Z, X is not X!

What you mean is that (for you) open-access = toll-free access plus
republication/secondary-reprocessing-service rights. I am afraid I have
to disagree.

> This is not an imposition of the "redistribution
> policy" (sic - this is an oversimplication - it is really about
> redistribution and, more importantly, reuse).

No, it is your preference for a rival definition, one that I have given
my reasons for not accepting. What we need, urgently, is toll-free access,
not republication/service rights. Unless adding republication/service
rights as a further requirement for open-access can be shown to *hasten*
and facilitate open-access rather than to handicap and slow it, I think
we are far better off without that extra requirement. Toll-free access
is requirement enough, and face-valid.

> It seems like what you are really saying is that open access is the enemy of
> free access,

Open access is not the enemy of free access: Adding the further
requirement of republication/service rights to the definition of
"open-access" would mean BOAI-1 lost any momentum that has come from
the coining and promotion of this neologism, just as BOAI-2 would lose
momentum if it had to use "reinforced access" instead of "open access"
for what it was providing!

I just want to keep calling a spade a spade.

> because toll-access publishers would allow free-access through
> self-archiving, but are unwilling to go to real open access, and that
> demanding open access delays the implementation of free access.

No, it's not that at all, and I think the reason you keep seeing it
like that is that you *have* indeed made "open access" identical, in
your mind, with "open access publishing" -- but you for some reason
don't want to come right out and say so!

What I want to avoid is suggesting (falsely) to the research community
(the only constituency that really matter in all this) that the only
way they can provide open access (sic) to their articles is if their
publishers agree to become open access publishers! Nor do I want to give
publishers that false impression. Publishers can support open-access
provision without having to become open-access ("gold") publishers
simply by supporting open-access self-archiving by their authors
("green"). Going gold, in other words, is not the only way a publisher
can help open access: going green helps too.

> I understand
> your argument, but I think you are wrong. I believe that the it is an
> illusion to imagine that it is possible to have universal free access
> through self-archiving AND to support journals through subscriptions.

I can speculate (and have speculated) about the ultimate hypothetical
outcome too
but what we are talking about now is something more concrete and urgent:
Providing open access *now*. There is nothing hypothetical about the
fact that authors *can* provide (and *are* providing) immediate open
access now (and 55% of them can even do it with the explicit blessing of their
publishers). Not enough of them are doing it yet, and our (or their)
conjectures about whether or not self-archived open-access versions of
every one of the 2.5 million articles in the 24,000 journals can co-exist
forever alongside the toll-access versions have nothing to do with it.

The reason the reason the research community is not yet providing open
access in sufficient numbers (via either BOAI-1 or BOAI-2) is that they
are not yet sufficiently informed of the benefits of open access (impact
maximization) or the ways to attain them. But it is only a matter of time
(and information).

> Self-archiving is, almost by definition, parasitic (and I say that in a good
> way!). And, like most parasites, the host has to be healthy for it to
> survive. If we imagine that all works are suddenly self-archived, who is
> going to subscribe to journals? I just don't see how self-archiving can
> provide universal free access without killing off toll-access journals in
> the process (do you really think selling print subscriptions will sustain
> them?).

I can speculate; you can speculate. But what we all really need to be
doing is providing immediate open access (by whichever of the two means
is suitable) for all of our research output, now.

The empirical evidence so far is (1) that TA journals in physics are
alive and well after 12 years of systematic self-archiving, which in
some areas (e.g., HEP) has long ago reached 100% and (2) in one case
(JHEP) born-OA journal (gold) has even successfully reverted to becoming
a TA journal (green), even though it is in an area (HEP) that was and
still is 100% self-archived.

    "JHEP will convert from toll-free-access to toll-based access"

(By the way, why are you so worried about the welfare of TA journals
when they are menaced by the self-archiving parasite, but not when they
are menaced by the OA journal competitor?)

> I feel that living under and promulgating the illusion that
> self-archiving and toll-access journals are mutually compatible does not
> hasten universal access, it delays it because it delays us facing up to the
> reality that we need a new economic model for scientific publishing.

Let me translate that into a formula for immediate action, if you don't mind:

You seem to be suggesting (to the perplexed researcher):

    "If you cannot find a suitable open-access journal in which to publish
    your article right now, and hence must publish it in a toll-access
    journal, do not self-archive it now, even though that will enhance its
    impact, because (a) that is just free access, not open access, and (b)
    it doesn't 'face up to the reality that we need a new economic model
    for scientific publishing.' So, instead of self-archiving it right now,
    you should _______ right now."

I would be very interested to hear what your recommended "_______" is...

I recommend immediate self-archiving of all toll-access journal articles,
right now, as so many others have done and are doing.

<sh> articles are not themselves research data (though they may contain
<sh> some research data), but they can be treated as computational data if
<sh> they are accessible toll-free online. Again, there is no need for any
<sh> further rights or computational capabilities to do be able to do this:
<sh> The full-text need merely be immediately, permanently, and ubiquitously
<sh> (i.e., webwide) accessible toll-free, so any user anywhere, any time,
<sh> can read the full-text on-screen, download it, store it, print it off,
<sh> search/grep it, computationally process it, etc.
> While there is a lot that can/could be done with self-archived free-access
> works, the inability to serve up cached, or more importantly, digested and
> reprocessed versions of works greatly and needlessly limits the types of
> computational analysis and data-mining that can be done on the literature.
> If all you want to do is search, then self-archiving is ok (although still
> suboptimal), but for any more sophisticated analyses it is not.

Could you please be more specific about exactly what you cannot do? The
fact that you have already cited so many things that one allegedly
cannot do with toll-free full-texts -- yet it turns out that in fact one
either can do them after all, or there is an obvious alternative way,
or they turn out to be marginal frills compared to the much more urgent
and central need and benefits of toll-free full-text itself -- emboldens
me to want to wrestle directly with the specifics.

Happy New Year,

Stevan Harnad

NOTE: A complete archive of the ongoing discussion of providing open
access to the peer-reviewed research literature online is available at
the American Scientist Open Access Forum (98 & 99 & 00 & 01 & 02 & 03):
    Post discussion to:

Unified Dual Open-Access-Provision Policy:
    BOAI-2 ("gold"): Publish your article in a suitable open-access
            journal whenever one exists.
    BOAI-1 ("green"): Otherwise, publish your article in a suitable
            toll-access journal and also self-archive it.
Received on Wed Dec 31 2003 - 23:46:40 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:47:14 GMT