This mailing list is no longer active and has been transitioned to Members of the I-coordination mailing list have been moved to the new mailing list. To learn more, visit

[I-coordination] New: How do we dissect Internet governance? [Was: Europe at a tipping point?]

Andrew Sullivan ajs at
Wed Dec 18 14:53:11 CET 2013

Hi Milton,

On Wed, Dec 18, 2013 at 04:33:01AM +0000, Milton L Mueller wrote:
> Brian's "rider" is a proposition that is patently false. Many
> governments have tried, and some have succeeded, in affecting "the
> way the internet is used by society" by regulating, influencing or
> controlling the administration of technical parameters.

Yes, of course.  But you (and many in this discussion) are committing
a fallacy of composition here.  "The Internet" does not have uniform
properties, including connectivity, throughout; that is not an
accidental property, but essential.

"The Internet" looks like the singular of a count noun, and the word
has the unfortunate property of actually _being_ a count noun in some
cases.  For instance, imagine two companies, each of which has two
large campuses in the world.  These campuses have internal networks
with routers and so on, and are linked together as well.  Thus, each
company has an internet -- a network of networks.  And if we suppose
they merge with one another, then they make their two internets into
one internet.  (This was more obvious back in the days when everyone
on earth didn't share the same TCP/IP+mostly-Ethernet-like
connectivity, but it's just as true conceptually today.)  None of
these "internets", however, is the Internet.

The Internet is the term for the colleciton of all of the above
internets of any sort when they voluntarily interconnect to one
another. It follows from this that there will be areas of the Internet
where not all of the functionality is working as it is everywhere
else; this is sometimes due to technical failure and sometimes due to
induced failure for policy reasons. If the latter policy reasons are
too invasive, it is correct to say that the node in question is not
really part of "the Internet", though it presumably has connectivity
to an internet. (In this sense, there is a good comparison between
half-functioning, web-only hotel networks, and networks subject to a
lot of interference by some nation-state.) There is no bright line
here, only degrees of "working". It is the claim quoted above -- that
governments have successfully messed with technical parameters in
order to close off parts of the Internet -- that I claim is the
fallacy of composition, because it suggests that one can interfere
that way without limitation and still be part of the Internet.  By
definition, you can't.

It also follows, by the way, that the reason a particular _technical_
interference is bad is not because of human rights, or bad adherence
to any governance model, or whatever; but rather, because it decreases
utility.  The Internet in the largest sense is simply a tool, and
attempting to interfere with the operation of the tool in order to
ensure that only particular utility functions are satisfied by it
undermines the tool itself.  For this reason, a country "leaking"
routing announcements and making a mess of global routing decisions,
a private firm that abruptly de-peers, thereby messing with global
routing decisions, and spammers that chip away at the utility of
messaging systems, are all bad precisely because they undermine the
overall utility of the system, even if only in a brief temporal span.
We don't need a metaphysics of morals to evaluate the wrongness of an
action, because we have an empirical test.
> in the public interest (a public policy judgment). In other words,
> he linked technical administration and use regulation. When he did
> that, he was refuting the arguments made here more effectively than
> anything I can add.

The mere fact of judging that two things are related to one another
nowise implies a judgement that they are the same thing, nor that they
should be treated the same way. I fail to see the force of the above

Best regards,


Andrew Sullivan
ajs at

More information about the I-coordination mailing list