Simplicity Theory
|
by Jean-Louis Dessalles
(created
2008.12.31)
(updated 2010.02.17)
The encounter problem is
spectacular, as it provides the best evidence that the human mind is sensitive to
description complexity.
Interest grows with the complexity of the place and with the simplicity of the encountered person.
Fortuitous encounters are
all the more unexpected that the place is
complex and the encountered person is
simple.
Example: In 2008, I was
travelling in
Fortuitous encounters seem
to be an exception to the rule of closeness, as the
interest grows this time with the remoteness of the place! A proper application
of the notion of unexpectedness, however, restores the prediction.
Interest grows with the complexity of the place and with the simplicity of the encountered person.
The complexity of the place l is the
relevant factor, not the distance: a big distant airport may be less complex
than the backyard of an obscure building of a lost suburb a few kilometres
away. The simplicity of the encountered person P is the relevant factor, not her closeness. Running into a
celebrity may be as unexpected as running into a close colleague. These
phenomena are correctly predicted by the fact that unexpectedness varies,
as we will show, as:
U = C(l) – C(P)
By definition, unexpectedness is the difference
between generation complexity and description complexity: Cw – C. Let’s
compute both terms.
Let us compute the
unexpectedness of the sequence ego*P*l*l(ego)*l(P). Here, l(ego) and l(P)
designate the presence of self and of the encountered person P at location l. We may write:
Cw(ego*P*l*l(ego)*l(P)) = Cw(ego)
+ Cw(P)
+ Cw(l)
+ Cw(l(ego)|l)
+ Cw(l(P)|l&l(ego)&P)
Note: Irrelevant elements in
conditional complexity expressions are omitted for the sake of clarity. For
instance, Cw(P|ego) = Cw(P).
The complexity of ego is supposed to be zero. For the
W-machine, Cw(P) = 0 and Cw(l) = 0, as P and l
are supposed to exist: the "world-machine"
has no work to do to produce them. In the ‘world’, P’s and ego’s common
presence at l are supposed to be independent: Cw(l(P)|l&l(ego)&P) = Cw(l(P)|l&P) (note that this is a crucial assumption). If ego and P play symmetrical
roles, the W-machine requires Cw(ego*P*l*l(ego)*l(P))
= 2 Cw(l(ego)|l) to generate the encounter situation.
One way to assess the term Cw(l(ego)|l) is to consider the complexity of the
decisions that ego had to take to end
up in location l. If l is not itself unexpected, then in most
cases, Cw(l(ego)|l) = C(l) and amounts to the minimum size of a
set of directions to reach l
(exception: if l is materially
difficult to reach). Finally:
Cw(ego*P*l*l(ego)*l(P)) =
The O-machine demands
complexity C(P) to determine P and C(l)
to determine l. On the other hand, C(l(ego)|l)
< C(l|l) + C(ego|l) =
C(ego*P*l*l(ego)*l(P)) = C(P) + C(l)
and thus as announced:
U(ego*P*l*l(ego)*l(P)) = C(l) – C(P)
An alternative computation
of Cw goes over the
sequence ego*l*l(ego)*l(P)*P. This time, P is first determined by her position l(P) = l(ego). Now Cw(l(P)|l(ego))
is zero, but Cw(P|l(P))
is not. To instantiate P, the W-machine
has to distinguish among all individuals that may happen to be in l. One procedure to do so consists in
checking local people first, by delimiting an area of radius R around l and considering all people living within it. Then R is increased until the actual P is reached. In this computation, P’s home and l play symmetrical roles, so that Cw(P|l(P)) = C(l) + c (l is
supposed to be as complex for P to
reach as for ego). The constant c depends on the spatial density of
people. The resulting unexpectedness C(l) + c – C(P) is similar to what we previously obtained.
Note that this second computation
still holds when the encounter occurs in the vicinity of ego’s home. In this case, Cw(l(ego)) and C(l(ego)) are negligible, but Cw(l(P))
is close to C(h(P)) (the complexity of P’s home) and the unexpectedness amounts to C(h(P)) +
c – C(P). It is indeed quite a coincidence
to meet a celebrity in front of one’s home.
Dessalles, J-L. (2008). Coincidences
and the encounter problem: A formal account. In B. C. Love, K. McRae &
V. M. Sloutsky (Eds.), Proceedings of the
30th Annual Conference of the Cognitive Science Society, 2134-2139. Austin, TX: Cognitive
Science Society.
Dessalles, J-L.
(2008). La pertinence et ses origines cognitives - Nouvelles théories.
![]()
Back to the Simplicity Theory page