NetNews Usenet Archive 1993 #1/NN_1993_1.iso/spool/gnu/misc/discuss/4368

Internet Message Format | 1993-01-10 | 17.7 KB

Xref: sparky gnu.misc.discuss:4368 talk.philosophy.misc:3254
Path: sparky!uunet!charon.amdahl.com!pacbell.com!att-out!rutgers!newsserver.jvnc.net!darwin.sura.net!zaphod.mps.ohio-state.edu!caen!kuhub.cc.ukans.edu!husc-news.harvard.edu!husc10.harvard.edu!zeleny
From: [e-mail] (Mikhail Zeleny)
Newsgroups: gnu.misc.discuss,talk.philosophy.misc
Subject: Re: Copyleft vs Public Domain
Message-ID: <[e-mail]>
Date: 10 Jan 93 19:08:22 GMT
References: <[e-mail]> <[e-mail]>
Organization: The Phallogocentric Cabal
Lines: 300
Nntp-Posting-Host: husc10.harvard.edu

In article <[e-mail]>
[e-mail] (Noah Friedman) writes:

>Richard Stallman has asked me to post this to talk.philosophy.misc on his
>behalf since he doesn't read news and no mail<->news gateway exists (that
>we know of).  Followups are directed to gnu.misc.discuss, where this has
>already been posted.

Thank you, Noah.  Stallman has more courage than I gave him credit for.
Even though he will lose this argument, as he must, I salute him for
daring to enter it.  Since he does not access the net, I am sending him
a copy of this reply.  However, since its subject is said to deal with
philosophical issues, I ask that followups be directed to both of the
above newsgroups.

In article <[e-mail]>
[e-mail] (Richard Stallman) writes:

>             Copyleft vs Public Domain
>
>Some people ask why the FSF uses copyleft (the General Public License
>or GPL) to specify conditions for copying GNU software.  Why not just
>put the software in the public domain?
>
>The purpose of the GNU project is to give users in general the freedom
>to use software in many ways.  We want them to be free to study and
>change programs, and to cooperate with each other by sharing
>programs.  This is what we mean by "free software".  The GPL achieves
>this better than the public domain, because (1) it prevents the
>freedom from being stripped off when the program is distributed, and
>(2) it takes away the incentive to be uncooperative by refusing to
>share an improvement.

Observe the key restriction, of freedom of "users in general",
_simpliciter_, to their freedom "in many ways".  The latter, insofar as
the distribution practices of the FSF enter into consideration, may not
be construed as financial freedom, or the ability to obtain GNU gratis,
and consequently can only be interpreted as absence of constraint or
coercion.  But inasmuch as the use of GNU places the user under a
reciprocal obligation specific to it, which in particular requires him
to attach the GPL to any program he derives therefrom, -- and hence
limits his potential usufruct (in this instance, his right to utilize
and enjoy the profits and advantages of public goods, so long as the
right of the public to the original goods remains inviolate, on which
see below), quite regardless of whether or not he ever exercises his
limited right to copy, modify, or redistribute it, -- there certainly
exists a very rigid constraint associated with the program.  And
inasmuch as this freedom "to cooperate with each other by sharing
programs" is both mandated and regulated by the GPL, and enforced by the
FSF, its very definition is rooted in an act of coercion, and
constitutes an exemplary instance of doublethink.  Surely whatever can
be said about the users of GNU being free to study the program or
cooperate with each other, can, _a fortiori_, be said about the users of
public domain software.  However, Stallman somewhat disingenuously
claims practical superiority of his approach, on the grounds that "(1)
it prevents the freedom from being stripped off when the program is
distributed, and (2) it takes away the incentive to be uncooperative by
refusing to share an improvement."  Let us examine the merits of his
claim.

When taken at face value, the first statement appears utterly
groundless.  No text belonging to the public domain can be withdrawn
therefrom in virtue of being embellished with addenda or corrigenda.
Only the newly introduced additions and amendments may be so
appropriated; the original irrevocably belongs to everybody.  However,
we have observed that Stallman's definition of freedom "to study and
change programs, and to cooperate with each other by sharing programs"
both depends on and involves, its enforcement by the terms of the GPL.
In this peculiar sense, it becomes trivially true that the GPL prevents
the "freedom" from being stripped off when the program is distributed,
by preventing itself from being stripped off when the program is
distributed.  But of course, this efficiency of self-serving does not
yield a justification for its superiority to public domain software.
Thus, the first reason fails to withstand rational scrutiny.

Does the GPL really take away the incentive to be uncooperative by
refusing to share an improvement to a program licensed under it?
Certainly not, for the financial incentive to come up with a competitive
advantage continues to exist, quite independently of the legal means to
do so.  Perhaps a more felicitous way to express Stallman's intended
point might have been to say that the GPL takes away the user's legal
means to be uncooperative with the FSF by refusing to share with it an
improvement to any program licensed under its provisions.  Thus, in
accordance with his peculiar notion of freedom as coercion to share, the
GPL can indeed be said to render GNU more "free" than public domain
software, by limiting the real freedom of its users.  However, the
validity of this point depends on the merits of Stallman's peculiar
conception of freedom, which will be discussed anon.

>Copyleft does prevent certain people from doing what they would like
>to do.  Those who would like to take GNU software, make some changes,
>and call the result their property are not free to do so.  We think
>this is a good thing.
>
>To understand why, first note that it is not possible for society to
>permit "all possible freedom," because some freedoms are incompatible
>with others.  This is often stated as, "Your freedom to swing your
>fist ends where my face begins."
>
>We always resolve the conflicts between freedoms by prioritizing
>them.  For example, the quotation above implicitly assumes that the
>freedom not to be punched is more important than the freedom to swing
>a fist.
>
>There is more than one way to apply a concept such as "free" to the
>area of software, because there are different choices of priority.
>The question is not, which is the true meaning of "free software", but
>rather, which of the valid meanings is best.

There is very little to disagree with in the above passage, until one
comes to apply its principles to its own reasoning.  Let us grant for
the sake of argument, that Stallman's peculiar notion of freedom is a
legitimate contender for the title of the best provision for so
qualifying software.  But what is the salient notion of merit, to which
we appeal in ordering our valid meanings of "free software"?  If we are
to subject the meaning of "free" to the tribunal of reason, surely our
intellectual integrity calls for doing likewise with the meaning of
"best".

For the sake of simplicity, let us limit ourselves to considering two
notions of the latter: merit as fitness to one's purpose, and merit as
moral excellence.  (I shall be happy to consider any reasonable
alternative suggested by the audience.)  It is obvious, and obviously
irrelevant, that the GPL fits the purpose of the FSF, as exemplified by
its own provisions; on the other hand, the degree of its fitness to the
purposes of the foundation's customers, depends on their interest in
deviating from its terms.  Since the latter question, which amounts to a
conflict of will, cannot be resolved on pragmatic grounds alone, we are
ineluctably led to the moral conception of merit.  In short, if Stallman
merely intends to claim that his idea of "free software" is practically
superior to the alternative conceptions, his argument is either
irrelevant, because self-serving, or incomplete, because implicitly
dependent on hidden moral assumptions.  Consequently, the best idea of
"best" must be grounded in moral considerations.  Let us bear this in
mind, as we evaluate the following.

>The GNU project is based on the idea that the freedom to decide your
>own actions with the programs you use--for example, whether to copy
>them or change them--is more important than occasional power over
>other people's actions.

This is the crux of Stallman's argument, and it is not surprising that
he made it so elliptic, as to totally obscure its true meaning.  Let us
resolve to elaborate it in light of his own examples.

>Making a program proprietary means interfering with the important
>freedoms--other people's freedom to study, share and change the
>program.  This is the software analogue of swinging the fist through a
>user's face.  Preventing this may bother those who want to swing the
>fist.  But don't sympathize too much; you might be one of the users
>who would get it in the face.

Turn this claim upside down, and consider its possible counterpart.
Licensing a program under the GPL admittedly interferes with one
important freedom, the freedom to own and control the product of one's
labor, insofar as the terms and conditions of the GPL automatically
attach to all programs derivative of GNU.  Let us assume for the sake of
argument that demanding exclusive ownership and absolute control would
indeed constitute a moral equivalent of "swinging the fist through a
user's face", as it is arguably the case in the instance when the
derivative program is no more than an adaptation of GNU to a specific
hardware platform.  But on the other hand, by Stallman's own lights, the
means employed by the FSF in preventing this appropriation, amount to a
similar sort of violence aimed at the most typical user of his product.

Consider the most likely beneficiary of the FSF's munificence, an
anonymous UNIX hacker, who is likely to be employed in other people's
proprietary projects.  The greatest benefit he could derive from GNU,
namely the ability to use it in his work, is expressly and prohibitively
regulated by the provisions of the GPL.  Inasmuch as the rhetoric of
"The GNU Manifesto" may be taken as a serious guide to the intentions
which underlie these provisions, one may assume that these regulations
are caused by the desire of the FSF to permanently change the terms and
conditions of distribution of all extant software, thereby forcing our
hero to change his occupation from the selling of goods to the selling
of services.  Naturally, this coercive program is bound to alienate many
of potential supporters, who might otherwise be inclined to contribute
funds or effort in exchange for using the programs in their work.  But
such heuristic considerations belong solely to the pragmatic realm; and,
for the moment, we need not concern ourselves here with the questions of
the best means, the optimal strategy, or even the likelihood of the FSF
attaining this goal in the foreseeable future, as our argument depends
primarily on the intrinsic merits of its aims.  Thus we must reformulate
Stallman's above statement, by noting that the GNU project appears to be
based on the idea that the freedom of the FSF to control and regulate
the actions of its customers with respect to the programs they use (for
example, by mandating the terms and conditions they must observe in
order to to copy them or change them), or, in other words, their power
over other people's actions, is more important to him than the freedom
of men to do as they see fit, which is absolutely unrestricted in case
of public domain software.

In short, by choosing to use the GPL over the public domain, the FSF
chooses to favor their proprietary conception of freedom of *things*,
over the universal conception of freedom of *people*.  And by choosing
to demand that the users share their enhancements and extensions of GNU,
rather than request that they do so of their own free will, the FSF
chooses *coercion* over *persuasion*.  And insofar as the first choice
is imbued with the venal values of commodity fetishism, whilst the
second one is definitively characteristic of the most reprehensible
business practices of their "software hoarder" competitors, they render
Stallman's notion of software freedom into a moral equivalent of a
tawdry, meretricious trick of semantic legerdemain.

>If not for the GPL, most users of our software would not have the
>freedom to redistribute and change it.  That is not just speculation;
>the examples of X Windows, TeX, and Berkeley's Unix extensions show
>that most users of these programs have only proprietary versions and
>do not have the freedom to share or change them.  The first authors of
>these programs did not themselves take away those freedoms, but did
>not defend them either.  Where that path leads was clear when the GNU
>project was started, and therefore we chose another path.

This is a straw man argument, which establishes its conclusion only by
glossing over legitimate alternatives to legal restrictions.  It is very
likely that, were the GNU programs placed into the public domain, with
but its current resources, the FSF could easily match, and even surpass,
any proprietary enhancements to their programs, in the same way as the
same is done by commercial companies striving to keep ahead of their
competitors.  Furthermore, in doing so, the FSF would attract a great
deal of additional support from those, who are currently put off by the
strictures of the GPL.  By contrast, in choosing its present strategy,
the FSF chooses convenience over principle.

>The GPL also encourages companies which make improved versions to
>return their improvements for inclusion in the standard version.  If
>not for this, GCC and Emacs would not be nearly as good as they are.

Were this the case, I would have had no reason to complain.  Alas, the
GPL does not encourage; it *demands*.

>But is this enough justification?  That is a fundamental philosophical
>question.  Some people believe it wrong to place any restrictions on
>anyone, ever--even restrictions against making any other
>restrictions.  Those readers who believe in pacifism and condemn use
>of force even to protect innocent victims would naturally disagree
>with our approach.

I have no objections against the use of force; however I also recognize
the moral need for a viable alternative to permanent strife.  The
marketplace is already chock full of small-time Napoleons, motivated
solely by the urge to stick it to their competitors, before they can
stick it to them.  The FSF stands uniquely able to offer an example of
free and voluntary cooperation, as distinct from opportunistic alliances
or coercive bonds.  Instead, it chooses to rely on the very same
reprehensible techniques of legal intimidation, as are used by their
openly profitmaking counterparts and alleged moral antipodes.  However,
no one could legitimately begrudge them this choice, had they not also
chosen to promote themselves as the altruistic champions of selfless
charity and universal freedom, instead of presenting themselves as what
they are -- advocates of a paternalistic personal vision of universal
obedience to centrally decreed and regimented sharing of intellectual
goods.

>That is not the philosophy of the GNU project, however.  We are not
>pacifists, and being passive and never saying "No" to anyone is not
>our goal.  Our aim is positive--to give the users the freedom to
>cooperate, which is distinguished from the freedom to obstruct.  That
>has been the goal ever since the beginning.

On the contrary, your aim is wholly negative, -- to force the users to
cooperate under the terms set by you, by obstructing a more or less
well-defined part of their freedom to use your product as they see fit.
Once again, public domain works do not in any way lend themselves to
obstruction or appropriation, nor do they in any way restrict their
users' freedom to cooperate in public or private projects dedicated to
their study and improvement.

Everyone who has spent any time in textual research, is aware of the
immense boon to scholarship, precipitated by each occasion of passing of
the body of work of an important artist or writer into the public
domain.  The same could be expected of software.  Why not give yourself
a chance to find out?

>If we put our software in the public domain, then we would have a
>great excuse to make.  We could say, "Don't blame us if you have no
>freedom to share and change this program--it was that other guy who
>redistributed it with a nondisclosure license and no source."  But we
>want to succeed in giving users that freedom, not prepare excuses for
>failure.  We use the GPL because it succeeds.

You use the GPL because it succeeds in furthering your agenda.  By
favoring your freedom to control and regulate the actions of your
customers over their freedom to do as they see fit; by choosing your
peculiar proprietary conception of freedom of things over the universal
conception of freedom of people; by adopting coercion over persuasion,
and convenience over principle; and worst of all, by arbitrarily
electing to deceptively promote yourself as a lone, altruistic champion
of selfless charity and universal freedom, instead of honestly
presenting yourself as an advocate of a paternalistic personal vision of
universal obedience to centrally decreed and regimented sharing of
intellectual goods, you undermine the moral foundation on which your
program is alleged to rest.  I entreat you to choose between forswearing
your falsely idealized self-image as the staunch supporter of freedom,
or living up to your own liberationist rhetoric.

cordially,
mikhail [e-mail]
"Les beaulx bastisseurs nouveaulx de pierres mortes ne sont escriptz
en mon livre de vie.  Je ne bastis que pierres vives: ce sont hommes."