[Citizendium-editors] What should the dispute resolution process be?

Larry Sanger sanger-lists at citizendium.org
Wed Nov 8 03:25:42 EST 2006

Final "big issue."  This is another one that we discussed at some length on
Citizendium-L, here:


(Look for "Editorial dispute resolution.")

I reproduce my kick-off post from that discussion here.



Dear all,

I want to start a discussion about how to resolve disputes among editors, a
topic we haven't really hashed through yet.  It is one of the most important
topics that I *haven't* yet written about in the CZ policy doc (it's taking
a long time because I keep getting distracted by other stuff!), and the
reason I haven't is that I'm not sure what to say yet.  It's not that I have
no ideas, either (since when did I ever lack for ideas?).  It's because we
absolutely must get this right; it's really deeply important.  I'm asking
for your help.

The problem is this.  Unlike, say, the Stanford Encyclopedia of Philosophy
or the Encyclopedia of Earth, which both have lead editors assigned to
particular articles or subject areas, which editors make the decisions with
regard to those articles/areas, *we* are proposing to approach editorial
responsibilities in "the wiki way."  That is, for any given article or set
of articles, there is no particular editor who is *assigned* to make
decisions about that/those article(s).  Rather, experts--who meet some
objectively verifiable qualifications, arrive on the scene, and choose to
become editors--must work together collaboratively.  So naturally they'll
have disagreements, some of which can't be resolved by discussion and
compromise.  *Decisions*, binding decisions about content, will need to be
made--particularly considering that we will be publishing particular
versions of articles as "approved."

Before actually forwarding any proposals for a process to make these
editorial content decisions, I think it's a good idea to consider some
constraints, or conditions, that a good proposal should satisfy.  What
should a well-designed decision process look like?

(?) The process should elicit the truth.  This seems like an obvious
constraint.  The outcome of these decisions will be to make a CZ article
read one way rather than another.  Presumably, whether it reads one way or
another, if it comes down to a formal process, is going to seem important to
some very well-informed people.  It is crucial, then, that the process
outputs something satisfactory.  Now we'd like to say that the satisfactory
outcome is precisely expressible as the truth.  The trouble, however, is
that if we're using this process to settle a question, then clearly,
well-informed people have different ideas about what the correct answer to
the question is.  In that case, do we declare what is, in fact, the truth?
No, because that will put off the losing party and bias the resource.  The
better requirement is this:

(1) The process should be neutral.  It should elicit expert knowledge (or
opinion) faithfully and neutrally--or, when an issue is primarily
socio-political, then according to the preponderance of opinion in the
English-speaking world.  But it should *accurately* describe the full range
of such knowledge, or opinion.  On Wikipedia, there are many people who make
all sorts of claims about what is or isn't "neutral" or "biased" and clearly
haven't got the first clue about what the requirements of the neutrality
policy (never, ever to be referred to as "NPOV" on the Citizendium please
:-) ) is.

(2) The process should not put off editors; in particular, it should be
regarded as fair.  This has a raft of subrequirements:

	(a) People should get a fair hearing.  They should have an
opportunity to express themselves fully and to be heard.  There should be a
full examination of the relevant facts, i.e., a dialectic.

	(b) The process itself should not be biased in favor of any
particular party to the dispute.

	(c) The person(s) making the decision should be as unbiased as
possible, and should be publicly committed and responsible to the ideal of
unprejudiced judgment.

	(d) The process should not depend on reading ancillary judgments or
debates, or on the outcome of such judgments or debates, that happen apart
from the process; in other words, it must be an *independent* process.
(Otherwise, the process can easily be corrupted and rendered unfair by that
ancillary debate.)

	(e) An exception to the latter is that the decision should be
broadly consistent with "precedent."

	(f) Matters of bad behavior should be strictly separated from the
merits of content questions.  A credible position should not be saddled with
the poor behavior of a rogue editor.  Therefore, probably, the "history" of
the question (on the wiki) might be ruled as "inadmissible" in the

(3) The process should not drag out endlessly.  It should have an end-point
and an established way to reach it.  A spectacularly dynamic content
creation system depends on the smooth and rapid operation of this process.
Therefore, it should not have multiple bottlenecks and steps a la the old
Nupedia system.

(4) The process must be a "last resort."  Since it is apt to be
time-consuming, most controversies over decisions should be made more
informally, in keeping with the "wikiwiki" (i.e., quick) way.

(5) On the one hand, decisions, once made, should not be able to be unmade
quickly, and should be unmade only if there is some new fact or
consideration of which the decisionmakers were previously unaware.

(6) On the other hand, a decision made "long ago" (however long that might
be) should be able to be revisited when the relevant personnel have changed
sufficiently (whatever that might be).

(7) In the interests of efficiency, it is best not to create new roles, but
to employ people from a pre-existing group (e.g., editors in a given
workgroup), unless absolutely necessary.

(8) Note that the plan of record is to employ, **somehow**, groups of
editors to make these decisions, in the form of "editorial workgroups." This
is only a weak constraint.  The basic and stronger constraint is that
*editors* are the ones who make these decisions.

(9) What else?

I am tempted to go on and offer some proposals for decisionmaking processes
that satisfy these conditions.  But I'd rather let you do that, because this
is already long (but not as long as some other of my posts!).


More information about the Citizendium-editors mailing list