Journal Papers vs. Books: The Direct/Indirect Income Trade-off

From: Stevan Harnad <>
Date: Sun, 11 Jul 1999 15:33:07 +0100

An intrinsically interesting side-issue is emerging from the exchanges
with Hal Varian. Nothing substantive hinges on it for the purposes
of the strategy and outcome I happen to advocate -- what I have taken to
calling the "optimal and inevitable" one for the refereed journal

    freeing the literature online for everyone, everywhere, forever,
    through (1) universal public self-archiving of all refereed journal
    papers, and, when the time comes, (2) a downsizing of journals to
    providers of the service of quality control and certification, paid
    for on the author-institution out of S/L/P savings rather than at
    the reader-institution end via access-blocking S/L/P.

Our disagreement is only about whether I am exaggerating the difference
between the motivations of refereed-journal papers-authors (regarding
THOSE papers) and authors in general -- of books, magazine articles,
etc. -- in suggesting that the former are, and have always been,
interested ONLY in giving those papers away, whereas other authors (as
well as themselves, when wearing other hats) have at least the hope of
some direct revenue from the sale of their texts.

If I AM exaggerating this difference, then Hal's view would be that
the "optimal and inevitable" for the refereed journal literature may wel
prove to be the optimal and inevitable for much of the rest of the
literature too.

Let me declare my cards at this point: I consider myself quite expert
on many aspects of the refereed journal literature (having edited one
major refereed paper journal for over two decades and one of the first
refereed online-only journals for nearly a decade, having provided an
online public archive for the self-archiving of unrefereed and refereed
papers over two years now, and having written a number of articles on
peer review and online journal publication), but I do NOT consider
myself expert on any other forms of publication (nor on ANY kind of

So I am taking up the gauntlet here for one reason only: That if I
should happen to be right and Hal should happen to be wrong about the
reality of the fundamental motivational difference between
refereed-journal authors and (just about) all others, then it is
important to sort this out, because it is only too easy otherwise to
dismiss self-archiving initiatives as naive and Quixotic, tilting at
the very Windmills of Capitalism, rather than as bringing out a
little-known home-truth about a small, anomalous subset of the
published corpus, a home-truth with radical implications (but for that
small literature only).

There have been several repeated points of misunderstanding. One has
been that I am clearly not claiming that refereed-paper authors don't
care about money! Nor that if there were some hypothetical,
counterfactual way that they could have their cake and eat it too --
i.e., if everyone everywhere could and would have free access to their
esoteric papers in perpetuo AND some entity were also prepared to pay
them (the authors) for it -- then they would quite happily accept the
eyes/minds AND the cash!

The trouble is, that in reality there is a trade-off: If money must
change hands in exchange for access, this entails that access is
reduced. Moreover (again, in reality), none of the money that changes
hands in exchange for access to the refereed journal corpus actually
ever reaches its authors! So they get the worst of both worlds from the
present system: Fewer (much fewer) eyes/minds, and no cash! Nor would
the cash that IS made from the toll-booths be nearly enough to
compensate them for the trade-off, even if it WERE passed on to the

Now what I contend is that this nonoptimal paradigm simply does not fit
book publishing in general (nor even scholarly/academic book publishing
in general -- although I agree that it may fit the special case of
esoteric scholarly monographs with only the hope of a succes d'estime, or
perhaps no hope of publication at all).

So it is not saintliness about potential direct revenue from these
papers, if revenue could drop like Manna from heaven, that sets these
authors (when wearing these hats) apart. It is the reality of the
trade-off, which means revenue here could only come at the expense of
access barriers, and there isn't faintly enough potential revenue to
make it even thinkable to embrace that trade-off.

So much for potential misunderstandings about DIRECT revenue, from the
sale of the papers themselves. There is no contest on the subject of
INDIRECT revenue, from the IMPACT of the papers, and the way that can
enhance the incomes of their authors and their authors' institutions.
Here too, refereed-paper authors are anything but saintly.

But there IS a persisting misunderstanding about the CONTINGENCY (or
causal relation) between the direct and indirect revenues: My point is
that (as we have established that the direct S/L/P revenue (a) does not
go into the author's pocket in any case -- and (b) it would not be
nearly large enough to compensate for the trade-off of eyes/minds for
cash even if it did) it follows that for this special literature the
fact that ANY DIRECT INCOME AT ALL is being made by ANYONE from
charging S/L/P tolls for access to the papers is actually NEGATIVELY
correlated with the INDIRECT-revenue benefits to the authors.

And causality is indeed the issue, rather than mere correlation, for
Hal is merely pointing to the overall positive correlation between
publication/sales and impact/income: {more publication, more sales} --
{more impact, more income}. Sure! But if there were no sales, the
impact/income would be still higher (MUCH higher, by my lights)!

Now here comes my ignorance about the rest of the literature: I feel,
intuitively and anecdotally, that this paradigm does not fit it, as it
does the refereed journal literature. It's certainly true that authors'
(and especially academic authors') careers are likewise enhanced, the
more they are read. And that if their books were given away for free,
they would certainly be read more. I'm even ready to believe that their
indirect income (from being read) outweighs their direct income (from
the gate-receipts) in some cases (especially for beginning authors,
vanity authors, and authors who, for whatever reason, are not destined
to have many readers -- I have already conceded the authors of esoteric
academic monographs as a special case).

But (and it is hard to put this in formal terms), I still feel that
that this is not the PARADIGMATIC case for books, as it IS for refereed
papers; with books, is not the "name of the game," and hence it would be
misleading to treat it as such.

But that's my best shot. I have no stronger arguments. So I now proceed
to quote/comment mode with someone who is clearly far more expert in
such matters than myself:

On Sat, 10 Jul 1999, Hal Varian wrote:

>sh> Yes, but the question was whether they would trade it off for their
>sh> refereed journal articles, not for their books and other activities.
> But of course they do---the time spent writing textbooks, consulting,
> and so on takes away from time spent writing and publishing academic
> articles.

This is a misunderstanding I have tried twice to correct! I am talking
about trading off (a) the collection of gate receipts for a given
article, and even the receipt by the author of all revenue therefrom, in
exchange for (b) the loss of eyes/minds therefrom.

In other words, it's about whether they would rather have the eyes or
the cash (even if the cash were forthcoming -- which in this literature
it is not!), NOT about how they trade off the rest of their timetable of

I can't understand why Hal's alternative, irrelevant and arbitrary
trade-off keeps being swapped for the relevant one I had singled out
from the outset. We are talking about the trade-off between selling it
and giving it away; and in this special literature (papers) -- as
opposed to the rest (books) -- there is simply NO CONTEST!

Perhaps the direct/indirect trade-off would be resolved the same way
for some no- or low-market books too, or at some early embryonic stages
of their authors' careers, but those would be case-by-case decisions,
whereas for the refereed literature it is, as they say, an open and
shut case, paradigmatic.

(Or am I the only one who continues to see a profound disanalogy here,
where direct revenue could at least in principle, if not always in
practise, offset the trade-off in the one case, but not even in
principle in the other?)

>sh> Perhaps the meaning of "make money on their texts" was not antirely
>sh> clear the first time. I hope that now it is: It is the blockage of free
>sh> access to THOSE texts (and not other texts they may write, or other
>sh> things they may do with their time, that we are discussing here).
> This is a very, very limited sense of "make money on their texts".
> Obviously academics make money on their works since they are paid,
> in part, to publish them. Perhaps you meant to say that academic
> authors would prefer to maximize the readership of their works.
> I agree that this is true. But it is also true of non-academic
> authors.

Let's just talk in terms of the direct/indirect income trade-off
(direct = sales royalty income; indirect = income from impact); it
encapsulates it all. And let's even throw out the red herring of the
nonacademic author. So henceforth we are only talking about academic
authors. And let's drop the polysemous trade/nontrade vocabulary
(because academics write both "trade" and "scholarly" books).

So it's just about journal-papers vs. books (by academics). And my
contention is that, because of the direct/indirect trade-off, the
PARADIGM for journal-papers is give-away and the paradigm for the rest
is not (although there are exceptions, such as esoteric monographs).

> > I said something stronger, and I repeat it: The authors of refereed
> > journal articles are not interested in making money on THOSE texts at
> > all.
> I submit they would be happy to make money on those texts, as long as it
> didn't dramatically reduce the readership. You have implicitly assumed
> that the only way to "make money on those texts" is to charge people to
> read them, thereby limiting access. I think this is a very narrow view,
> and limits the kinds of business models you might consider for information
> industries.

Insofar as my refereed journal articles are concerned, For-Fee or
For-Free ARE the only pertinent options. (I will get to the putative
alternative "business models" presently; to let the cat out of the bag,
though: the story does not change one bit if one introduces a "delay"
variable -- "fee now, free later": the trade-off is the same, and so is
the outcome; besides, the author doesn't get the fee anyway, and even if
he did, it wouldn't be worth his while, now or later; and research
authors want immediacy as well as impact for their work!)

>sh> The conclusion is that in the case of the refereed journal literature
>sh> only, the author has an interest in there being NO INCOME (to anyone)
>sh> from sale of his texts, because income means access-blockage. This is
>sh> also in the interest of the author's institution, which likewise gains
>sh> kudos (and grant income) from the impact of his research (not to
>sh> mention their savings from S/L/P cancellations!)
> Here is where your confusion lies. It is false (as an economic
> proposition) that there is a direct relationship between "no income" and
> "access blockage". Consider a field where there is an economic demand for
> the research (engineering, medicine, finance, etc.) It is common to see
> research produced in these fields where publication is delayed. It is
> first made available for a high price via consultancies, newsletters,
> affiliates, etc., and later made available to the academic world at large.
> This way the authors have their cake and eat it too---make money, and
> still make the research available to all. Obviously this practice can be
> abused, but that is beside the point. The point is that income can be
> made even while providing broad access.

Good case in point: The reply, insofar as my refereed journal papers are
concerned, is loud and clear: Forget about the pennies I could make from
the gate by denying or delaying access. Let everyone have it for free
NOW. There aren't pennies enough in the (realistic) world to make me
change my mind.

No, offering the author a cut of the gate receipts will NOT cut it in
this peculiar literature, and if that's the alternative "business
model," it's a nonstarter here. (And hence the analogies where it may
work are simply irrelevant, again.)

> A similar example is JSTOR, where back issues journal articles are
> available at zero marginal price to subscribing institutions.

Yes, IF (equals where, and when) the subscription is paid. Tell that to
all my potential readers who are NOT lucky enough to be at an
institution that can afford the subscription!

(The weasel-word was "marginal price," because it glosses over the
S/L/P barrier itself. Nor is an author happy about having his
retrospective journal publications held hostage to S/L/P either: why
should he be?)

Remember, the alternative to each of these cases is self-archiving (which
authors can do retrospectively too!).

> Indeed, this model is becoming somewhat common in the for-profit trade
> literature. An author first publishes a book, then 2 or 3 years puts the
> text up on his or her web site.

Interesting, but its relevance to the journal literature is unclear,
for the reasons mentioned above.

(Moreover, short-half-life print runs and the actuarial stats behind
them are probably behind the practise you mention, as is the current
inconvenience of on-screen discursive reading for book-length
materials: Immediacy, and constant on-line accessibility and
navigability are assets peculiar to the journal literature.)

> This is "second degree temporal price discrimination" and presents one
> interesting business model for information goods.

Might be useful to a journal publisher looking for add-ons to help
preserve an S/L/P market, but the author wants immediate visibility, so
it is a minus, not a plus for him, with or without the pennies.

>sh> So with refereed journals it makes sense (and with books it does not)
>sh> that costs should be paid for at the author-institution end (yielding
>sh> free access for all) rather than the reader-institution end (blocking
>sh> access, impact income).
> The conclusion has some merit, but I disagree with the analysis. I claim
> that the fundamental force is not the difference in motivation of the
> authors (since trade authors and academic authors like both money and
> readers) but in the nature of the production process of research, and, in
> particular, that the producers and consumers are the same people.

I'm afraid I can't entirely agree. First, the trade/academic distinction
is a red herring here; it's journal-papers by academics vs. books by
academics that we are focusing on now.

It is true that both of these have a smaller and more inbred
constituency than the nonacademic book authorship/readership, but even
in the former it is not literally true that the producers and consumers
of any given work are the same (not even to the extent that it is true
that contemporary composers compose only for one another!). It may very
well be that a large proportion of the readers of my refereed journal
articles are themselves authors of refereed journal articles, but so
what? And so what if (when) it is true that most of the buyers of a
book by an academic author are themselves authors of refereed journal
articles? (The buyers can't be mostly book-authors -- not enough to go
'round except in the special case of those esoteric monographs, for a
narrow circle of peers only, which are again nonrepresentative.)

I don't think any principles follow from this inbredness -- except for
that last special case where fellow-specialists are the only ones
willing to shell out for an esoteric volume in their own specialty. The
academic community is simply too large to be treated as just a peer
market for one another's work.

I think a potentially much more important side-effect of this
inbredness, though, is peer review itself, and here again, the
implications are different, and much more profound, for journals:
For we are not only the authors and the consumers of the refereed
literature, we are also its referees: WE do the quality controlling,
and we do it for free, motivated, when all is said and done, by
the Golden Rule, intrinsic interest (and perhaps some superstition and
sometimes worse) more than anything else.

The fact that even the quality controllers are us means that if journal
publishers should prove (as I hope they will not) to be inclined to try
to hold the literature hostage to access-tolls by trying to prevent us,
through regressive copyright agreements, from self-archiving, then
editorial boards are certainly in a position to bolt (as they have
already done -- or threatened to do -- in a few cases)...

> To see
> why, look at another genre with this characteristic---fanzines for
> hobbiests (science fiction, collectors, etc). In the print fanzine world
> authors were not paid and the readers paid for zine via subscription, just
> as in the academic genre. Now most of the hardcopy zines have moved to
> the Web where they are freely available. I think you are right that this
> will happen with academic journals, but it isn't because academic authors
> wouldn't like to get paid for their texts.

This is the general Net freeware/vanity-press analogy, and I'm afraid I
still don't find it helpful: Yes, refereed journal articles are
give-aways, like fanzines, but they are quality-controlled, certified
give-aways, hence not vanity press (and they still require an economic
model for covering the essential costs of

> Well, we could quarrel about the metaphysics of "cause", I guess, but I
> would say that there is a pretty direct causal relationship between the
> number of readers of one's refereed articles and one's academic salary in
> the US.

Please see above for a more fine-grained causal analysis of the causal
structure of that correlation!

>sh> The trade/nontrade divide is fairly represented by desire and
>sh> expectation of revenue directly from the sale of the text.
> This is simply wrong. Most trade books make very little money for the
> author directly from royalty revenues. You only have to look at the
> statistics on authors' incomes. The National Writers Union publishes
> such a survey and the numbers are strikingly small.

Are those figures adjusted for time-lines? At any given time, there must
be a large number of authors at the earliest embryonic stages, where
their output is not matched by income. And perhaps the profession also
contains a large proportion of failures. But surely no writer's desires
and expectations are based on the actuarial facts about starters and

There is a counterpart in refereed-journal writing too, though. Most
articles are read by few and cited by none (except the author himself).
Does this restore the analogy? Hardly, for the dream of the researcher
is of making an important finding and thereby making an important impact
on further research (and, yes, its indirect income-enhancing
consequences); what is the nonacademic writer's dream, if not being
widely read?

(I made a mistake when I wrote earlier that reaching a maximum number
of eyes/minds was an end in itself for the researcher, rather than just
a means of making money from the sales of the texts: the end is
actually reaching the RIGHT minds, the ones that will understand, build
upon, and, yes, cite their work. What is the nonacademic author's
counterpart for this?)

We had agreed we would only speak of academic authors' though: DO
academic authors write books with motivations so different from
those of nonacademic authors? Are they really primarily after the
indirect benefits rather than the direct ones from sales? It would be
interesting to see the stats for that specific literature (and
controlled for cumulative life-cycle effects -- and perhaps even culled
for esoteric "failures")... But remember that "desire and expectation"
does refer to motivation (dreams), and not actuarial realism...

>sh> this reader-end market is the wrong one for
>sh> refereed journal articles, which authors have always wanted to give
>sh> away, but have been prevented from doing so by the economics of paper.
> Again you are making the leap from "maximize the readership of" to
> "give away". I agree with the former motivation (at least to the first
> order) but I deny that it immediately follows that the content must
> be given away. This is one tool, but not the only one.

Would you not say that the long-standing practise of distributing free
(paper) reprints to reprint-requesters tends in my direction? As does
the absence of a corresponding tendency with books (except in the case
of Vanity Press)? (Publishers are always happy to pay royalties to an
author in the form of free copies of his book: Why is this not taken up
more by authors?)

I would be interested to know what would be a BETTER "tool" for maximizing
access to my papers than making them available online for free for all
forever. For self-archiving is within reach of one and all. Paul
Ginsparg has led us to the water, and all we need do now is drink...

> I claim that a smaller
> and smaller fraction of total written works will be sold in in the trade
> model. The majority of written work will be given away on the Web
> (in fact, that is probably true today).

Would you say that the rising success of Amazon Books was evidence for or
against your prediction?

> In my experience, most publication is a mixture of economic and
> non-economic motivations. My textbooks and trade book were successful
> because my academic papers were widely read. The visibiilty resulting
> from the academic, text, and trade books leads to more offers for speeches
> and consulting. I don't see the stark division that you describe.

Uncontested. And still the papers are give-aways whereas the rest are

> I agree that the filtering is what need to be paid for. (A quibble:
> some referees are paid, though typically not a sum that covers the
> time cost.)

A propos, from a previous posting:

    There is another Harnad, not myself, but a mathematical physicist
    by the name of John Harnad <> (and the one
    who first drew my attention to the Ginsparg Archive way back in the
    early '90's). J. Harnad has some further recommendations on the
    subject of referee answerability and compensation: He recommends
    that (a) all referees should be paid to referee papers; (b) payment
    for a rejected paper should be minimal (say, $200), but payment for
    an accepted paper should be commensurate with the effort of seeing
    it through the successive revisions (say, $2000) to successful
    publication; and, to avoid the potential abuse discussed above, (c)
    if a paper is accepted, the name of the accepting referee(s) should
    be co-published with it, to share the responsibility, praise or
    blame. He feels this would raise the quality of the refereeing and
    make the entire process much more answerable, hence effective, than
    it is now.

    Obviously this proposal is compatible with the transition from
    reader-institution-end payment to author-institution-end payment,
    but it is an as yet untested peer-review reform proposal; all such
    reform proposals need to be tested empirically and practically
    before being implemented on any scale. Hence it should not detain
    us on the road to freeing the CURRENT refereed literature, such as
    it is.

    (I also think that there is not enough money in the world to pay
    fairly for the precious time that referees steal from their own
    research in order to do the mostly thankless task of peer review;
    hence the Golden Rule is probably the only one we can continue to
    rely on! SUBMISSION charges, creditable toward PUBLICATION charges
    should the paper be accepted, may not be a bad idea to levy on
    authors, though, with or without referee payment, for it to might
    help raise submission standards and even revision
    conscientiousness, hence perhaps even lightening some of the burden
    on the work-horse referees; but this too would need pre-testing.)

>sh> It's not academic vs. trade authors! It's academics when they wear
>sh> their refereed-journal-author hats and every other author, including
>sh> themselves, when wearing their book-author hats (except in the case of
>sh> esoteric monographs).
> Hmm. So academic authors like money when they write trade books and don't
> like money when they write [journal papers]?

Correct. Vide supra.

> The real difference is not at the level of the author, but at the level of
> the industry. In the trade book industry the authors and the readers are
> quite distinct, whereas in the academic publishing industry the authors
> and readers are, by and large, the same people. The academic publications
> are both an input to and an output of the research process. Other
> publishing industries where authors and readers are the same people had
> similar business models, and, I suggest will have similar dynamics in
> their move to the Web.

This again conflates books/journals and separates academic/nonacademic,
which I think are the wrong fault lines. (I have tried to discuss the
inbred issue above.) Perhaps you could give another example (besides
esoteric monographs) that fits the give-away algorithm of refereed

> (The apparent necessity of paying for the
> refereeing overhead makes things a bit more complicated in the academic
> markets.)

Vide supra.

>sh> This is all no doubt true. But books, academic and nonacademic, can in
>sh> principle bring direct incomes to their authors, and journal articles
>sh> can't, and that's the point.
> I agree that journal articles don't bring direct income to the authors
> under the current business model, but that is a rather obvious
> observation. Whether they "can't" or not could be disputed.

I'd like to hear a realistic scheme for this! To my mind it founders on
the following: Journal papers average next to no readers/citers
(with the access barriers of both paper and S/L/P). Going on-line is one
step forward; bypassing S/L/P through self-archiving is a second. Can
you think of a better way?

> But the more
> interesting observation is that trade books generally don't bring much
> direct income to authors either. Authors publish trade books because of
> glory they get, for the most part, some of which can be translated into
> money.

So how do most trade authors make their fortunes? Is it all appearances
and movie rights now?

> You are simply making the same false claim over and over again. Authors
> would be happy to make money from their journal articles, it's just that
> few publishers offer them this option. So it is not the motivation of the
> authors that is relevant, but rather the business model for the industry.

By what business model would authors make money from their journal
articles? how much? and how do you think that would fare against
self-archiving? (Another way to put it -- I think I've already asked
this once: How much money would the publisher of my refereed paper have
to pay me so that I relinquish my self-archiving right? And where would
he get enough money for that?)

> I don't have substantive quarrels with the rest of your argument, or
> with your understanding of the academic publishing industry. I just
> think that you don't understand the economics of the trade publishing
> industry. The simple fact is that very, very few authors make any
> substantial sum of money in the trade industry, and their motivation
> for publishing trade books is more complex than you assert.

If none of my comments so far bears on how this is to be interpreted or
applied, then I acquiesce to your much more extensive knowledge on that

Stevan Harnad
Professor of Cognitive Science
Department of Electronics and phone: +44 2380 592-582
Computer Science fax: +44 2380 592-865
University of Southampton
Highfield, Southampton
Received on Wed Feb 10 1999 - 19:17:43 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:45:34 GMT