16 D INTRODUCTION: RHIZOME
gists distinguish between long-term memory and short-term memory (on
the order of a minute). The difference between
them is not simply quantita-
tive: short-term memory is of the rhizome or diagram type, and long-term
memory is arborescent and centralized (imprint, engram, tracing, or pho-
tograph). Short-term memory is in no way subject to a law of contiguity or
immediacy to its object; it can act at a distance, come or return a long time
after, but always under conditions of discontinuity, rupture, and multipli-
city. Furthermore, the difference between the two kinds of memory is not
that of two temporal modes of apprehending the same thing; they do not
grasp the same thing, memory, or idea. The splendor of the short-term
Idea: one writes using short-term memory, and thus short-term ideas, even
if one reads or rereads using long-term memory of long-term concepts.
Short-term memory includes forgetting as a process; it merges not with the
instant but instead with the nervous, temporal, and collective rhizome.
Long-term memory (family, race, society, or civilization) traces and trans-
lates, but what it translates continues to act in it, from a distance, offbeat,
in an "untimely" way, not instantaneously.
The tree and root inspire a sad image of thought that is forever imitating
the multiple on the basis of a centered or segmented higher unity. If we con-
sider the set, branches-roots, the trunk plays the role of opposed segment
for one of the subsets running from bottom to top: this kind of segment is a
"link dipole," in contrast to the "unit dipoles" formed by spokes radiating
from a single center.
13
Even if the links themselves proliferate, as in the
radicle system, one can never get beyond the One-Two, and fake multiplici-
ties. Regenerations, reproductions, returns, hydras, and medusas do not
get us any further. Arborescent systems are hierarchical systems with cen-
ters of signifiance and subjectification, central automata like organized
memories. In the corresponding models, an element only receives infor-
mation from a higher unit, and only receives a subjective affection along
preestablished paths. This is evident in current problems in information
science and computer science, which still cling to the oldest modes of
thought in that they grant all power to a memory or central organ. Pierre
Rosenstiehl and Jean Petitot, in a fine article denouncing "the imagery of
command trees" (centered systems or hierarchical structures), note that
"accepting the primacy of hierarchical structures amounts to giving
arborescent structures privileged status.... The arborescent form admits of
topological explanation.... In a hierarchical system, an individual has only
one active neighbor, his or her hierarchical superior.... The channels of
transmission are preestablished: the arborescent system preexists the
individual, who is integrated into it at an allotted place" (signifiance and
subjectification). The authors point out that even when one thinks one has
reached a multiplicity, it may be a false one—of what we call the radicle
I
NTRODUCTION: RHIZOME □ 17
type—because its ostensibly nonhierarchical presentation or statement in
fact only admits of a totally hierarchical solution. An example is the
famous friendship theorem: "If any two given individuals in a society have
precisely one mutual friend, then there exists an individual who is the
friend of all the others." (Rosenstiehl and Petitot ask who that mutual
friend is. Who is "the universal friend in this society of couples: the master,
the confessor, the doctor? These ideas are curiously far removed from the
initial axioms." Who is this friend of humankind? Is it the .pMosopher as
he appears in classical thought, even if he is an aborted unity that makes
itself felt only through its absence or subjectivity, saying all the while, I
know nothing, I am nothing?) Thus the authors speak of dictatorship theo-
rems. Such is indeed the principle of roots-trees, or their outcome: the
radicle solution, the structure of Power.
14
To these centered systems, the authors contrast acentered systems,
finite networks of automata in which communication runs from any neigh-
bor to any other, the stems or channels do not preexist, and all individuals
are interchangeable, defined only by their state at a given moment—such
that the local operations are coordinated and the final, global result syn-
chronized without a central agency. Transduction of intensive states
replaces topology, and "the graph regulating the circulation of information
is in a way the opposite of the hierarchical graph.. . . There is no reason for
the graph to be a tree" (we have been calling this kind of graph a map). The
problem of the war machine, or the firing squad: is a general necessary for n
individuals to manage to fire in unison? The solution without a General is
to be found in an acentered multiplicity possessing a finite number of
states with signals to indicate corresponding speeds, from a war rhizome or
guerrilla logic point of view, without any tracing, without any copying of a
central order. The authors even demonstrate that this kind of machinic
multiplicity, assemblage, or society rejects any centralizing or unifying
automaton as an "asocial intrusion."
15
Under these conditions, n is in fact
always
n - 1. Rosenstiehl and Petitot emphasize that the opposition,
centered-acentered, is valid less as a designation for things than as a mode
of calculation applied to things. Trees may correspond to the rhizome, or
they may burgeon into a rhizome. It is true that the same thing is generally
susceptible to both modes of calculation or both types of regulation, but
not without undergoing a change in state. Take psychoanalysis as an exam-
ple again: it subjects the unconscious to arborescent structures, hierarchi-
cal graphs, recapitulatory memories, central organs, the phallus, the
phallus-tree—not only in its theory but also in its practice of calculation
and treatment. Psychoanalysis cannot change its method in this regard: it
bases its own dictatorial power upon a dictatorial conception of the uncon-
scious. Psychoanalysis's margin of maneuverability is therefore very