154

12

Systems and Networks

(such as an organism considered as a whole) it may be a great handicap. A random

extraneous disturbance (i.e., noise) of sufficient amplitude may suffice to place the

system in a different basin (dehabituation).

12.4

Self-organization

The concept of self-organization appears to have originated with Immanuel Kant. In

Sect. 65 of his Kritik der Urteilskraft (1790), ²³ we read “In einem solchen Produkte

der Natur wird ein jeder Teil, so wie er nur durch alle übrigen da ist, auch als um

der anderen und das Ganzen willen existierend, d.i. als Werkzeug (Organ) gedacht;

welches aber nicht genug ist (denn er könnte auch Werkzeug der Kunst sein und so nur

als Zweck überhaupt möglich vorgestellt werden), sondern als ein die anderen Teile

(folglich jeder den anderen wechselseitig) hervorbringendes Organ, dergleichen kein

Werkzeug der Kunst, sondern nur der allen Stoff zu Werkzeugen (selbst denen der

Kunst) liefernden Natur sein kann; und nur dann und darum wird ein solches Produkt

als organisiertes und sich selbst organisierendes Wesen ein Naturzweck genannt

werden können”. The emphases, given by the author, seem to indicate his own feeling

of the importance of this statement. As for the idea that a living organism is both

cause and effect of itself, that is to be found in the preceding Sect. 64: “Ich würde

vorläufig sagen: ein Ding existiert als Naturzweck, wenn es von sich selbst Ursache

und Wirkung ist ellipsis. . .”. Note Kant’s caution in putting this forward as a provisional

idea. Little, if anything, seems to have been added by the latterly often cited work

of Maturana and Varela who introduced the term “autopoiesis”; they seem rather to

have rendered a clear enough conception recondite. More constructive was the term

“homeostasis” introduced by Cannon in the 1920s, and which became incorporated

into Ashby’s cybernetics (Sect. 3.2). Nevertheless, let us be mindful of Ashby’s and

von Foerster’s critiques of self-organization (see Footnote 30 in Sect. 6.4).

Consider a universe U comprising a system S and its environment E; i.e., U = Sunion∪

E. Self-organization of the system implies that its entropy spontaneously diminishes;

that is,

delta upper S Subscript normal upper S Baseline divided by delta t less than 0 periodδSS/δt < 0 .

(12.31)

According to the second law of thermodynamics, such a spontaneous change can

only occur if, concomitantly,

delta upper S Subscript normal upper E Baseline divided by delta t greater than 0 commaδSE/δt > 0 ,

(12.32)

with some kind of coupling to ensure that the overall change of entropy is greater

than or equal to zero. If all processes involved were reversible, the changes could in

principle exactly balance each other, but since, inevitably, some of the processes will

be irreversible, overall

delta upper S Subscript normal upper U Baseline divided by delta t greater than 0 periodδSU/δt > 0 .

(12.33)

23 Quotations are taken from the third edition (Kant 1799).