COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	histor[w85,jmc]		Notes for a science of history
C00017 ENDMK
C⊗;
histor[w85,jmc]		Notes for a science of history

I have been reading English history, first Hume, now Churchill, also
Macaulay.  It is a dismal affair, the struggles of gangs of thugs.
As these writers put it, the first tragedy was the fall of the
Roman Empire, which had established an orderly society capable
of defending itself against marauders until its internal weaknesses,
manifesting themselves most obviously in unprofessional emperors,
made it incapable of defense.  This is to be checked by reading
Gibbon.

I propose to proceed by asking the question:  How could European
society have done better?  This is deliberately vague, because
it doesn't say, "Who could have done better?"

I propose to begin with a science fiction approach.  What if I
were king of this or that country?  What if I were dropped in
by time machine with my spotty knowledge of modern science
and technology and history?  On a more aggregate level, we can
ask what should England have done?  More specifically what should
Henry the nth have done?  Suppose I brought a substantial quantity
of gold or guns?

This methodology is obviously subject to attack by historians as
the most idle speculation fit only for entertainment.  However,
I shall justify it as follows.

1. History is interested in causality not merely in chronicle.
According to the doctrine of causality I promote, in AI and
philosophy, causality is concerned with counterfactuals, with
might-have-beens.

2. As the philosophers point out, counterfactuals are murky.
My doctrine is the Cartesian product theory used to make
approximations to reality.

3. In dealing with causal systems, we approximate autonomous
systems by simpler systems with inputs from the outside.  In
these simpler systems, might-have-beens can have definite
answers.  For example, we may ask whether Harold would have
recovered from the loss of the battle of Hastings if he hadn't
been killed.  We regard an arrow hitting him in the eye as
an accidental event, and it doesn't do too much violence to
the wholeness of our concepts to imagine it not happening.

A more iffy question is whether Harold could have recovered
from the loss of the battle of Hastings.  We are on more shakey
ground here, because we need to limit the space of Harold's
possible actions.  It isn't much connected with the way we
think about history to ask what would have happened if Harold
had retreated to the North and set his men to manufacturing
gunpowder, although thinking about this might illuminate some
interesting questions.

It is still rather iffy to ask what would have happened if
Harold had had the military experience of a Roman general.
Would that experience have been relevant and useful?  The
answer we get depends on to what extent we take into account
the actual character and knowledge of Harold.  Of course, we
don't really know much about that character and knowledge, so
any speculation about his alternatives must ignore much of it.

Churchill speculates on whether Harold could have done better
by avoiding an immediate pitched battle with the Normans.
Apparently many historians thought he could.  Churchill doubts
the comparison with the British tactics against the Romans on
the grounds that Harold had infantry and the Normans had
cavalry, and therefore harassing tactics would have been
relatively difficult for him.

4. My methodology is to use all levels of might-have-beens.
Each gives useful information.

5. It is common to imagine that some historical character
or institution could have done very much better than he or it did.
If only he had learned the lesson of some event X in the past,
he wouldn't have repeated Y's mistake.  This is a reasonable,
but one shouldn't be too confident about a particular speculation
of this kind.  Particularly if the history was known, we should
look for additional constraints on the action.

Now to a speculation.

	Suppose I ruled a small European kingdom in the year 900.
How could I establish and maintain a society reasonable from
my present point of view?  We want prosperity, as much internal
justice and equality of opportunity as possible.  It also must
be able to defend itself against the Danes, Normans and other
raiders.

	1. If we could have modern military technology and
maintain this advantage, the task would be much easier.  We
won't ignore this speculation, but clearly it is more interesting
to do without it.  However, there is no clear line where technology
leaves off and mere organization begins.

	2. Let's not forget Mark Twain's Connecticut Yankee.
We can perhaps speculate about how the Connecticut Yankee
could have done better or couldn't have done what Mark Twain
depicts him as doing.  Also Mark Twain decided to finish off
the story with his defeat, and perhaps he stacked the deck
against him starting at page 200 or so.

	3. A common speculation is an a democratic society of
yeomen with military training who would alertly fight off raiders.
Consider it as a kind of 17th century Switzerland transported
to the earlier period and dropping the 17th century technology.
That might have worked, but there are difficulties.  We are considering
professional farmers as part time warriors.  The Israeli's
manage this, but they are extremely capable and have a modern
information technology, and it remains to be seen whether they
could maintain this for hundreds of years.  After all, the Britons
defended themselves successfully for 150 years after the Romans
left, but failed in the longer run.

The major difficulty is that we are asking amateurs to defend themselves
against professional pirates and raiders with professional leaders.
Also the piratical society can suffer huge losses, and there will
be a new batch of pirates later.

The best chance is that our organized society is large enough to
maintain a professional army capable of defeating the maximum
size raid with the aid of mobilization of a militia.  Perhaps Alfred
the Great managed that, and in this case, the only problem that
has to be solved is how to manage the succession of power so as
to avoid too much internal struggle and have capable leaders.

The success of capable kings whose kingdoms collapsed after their
deaths suggests that this is the key problem.

Here we come to the problem of ``original sin''.  If only the uncles
of the heirs had been more loyal, feudalism would have worked better.
More specifically, if only humanity had evolved more co-operative
people.  This line is worth following, but we can also follow
an institutional line, either educational or political.  The educational
line is to ask how princes could have been brought up better so
as to be more co-operative.  The political line is to ask how the
institutions could have been better.

The obvious political speculation is to ask whether modern democratic
institutions would have worked with the technology of that day and
perhaps some other social factors of that day.  There is reason to
doubt it.  Perhaps the accumulation of resources necessary for
defense required a level of hardship for the majority of the population
that could only be maintained by oppression.  To put it sharply, suppose
that it requires an average lifetime of 25 years for the majority in
order to permit  a part of the population to live to their fifties with
sufficient leisure to think about strategy of survival of the
country.  This too is a speculation.  Maybe there wasn't that much
efficiency gain from oppression.

***
Reprise on methodology

Our methodology involves the enthusiastic use of might-have-beens.
Historians who discuss the use of might-have-beens confront the
question of what level of might-have-been to use.  Is it only
legitimate to ask what would have happened if Harold hadn't been
killed at Hastings or is it also legitimate to ask if he could
have salvaged the situation if he had been Alfred the Great?
Is it legitimate to ask if the Britons by better social organization
could have defended themselves indefinitely.

The problem of what level to admit is so worrisome that slippery-slope
arguments lead some historians to claim to reject might-have-beens altogether
and also causality and claim to be just recounting the facts.  This
approach is incompatible with our goal of ``understanding'' history,
whether understanding is considered for its own sake or for permitting
present society to draw useful lessons from it.

Our solution is to cut the Gordian knot.  We allow all levels of
might-have-beens and will discuss a given issue, e.g. what Harold
might have done, at different levels of relaxation of what he
actually did, what actually happened, what kind of person he was,
and even what social environment he operated in.

This methodology is suggested by studying the formalization for
AI purposes of commmon sense notions of causality.