Heinz von Foerster's Self Organization,

the Progenitor of Conversation and Interaction Theories

Contents
Dear Heinz and Mai, 351
INTRODUCTION 352
SELF ORGANIZATION 352
SOME DETAILS 355
AXIOMS AND OTHER POSTULATES 357
INTERACTION OF ACTORS 361
A COMMENTARY 361
A DISCUSSION 362
ACKNOWLEDGEMENT 362
REFERENCES 363
INDEX 364
Index, Table of Contents added by Nick Green
Last corrected 22
nd.
July, 2004
Queries to nick_green@blueyonder.co.uk
Systems Research RLSLARCH PAPLR
351 G. Pask
Systems Research Vol. 13 No. 3, pp. 349-362 1996
- Research Paper
Hcinz von Focrsicr's Sclf Organizaiion,
iIc Progcniior of Convcrsaiion and
Inicraciion TIcorics
Gordon Pask
System Research Ltd, Old Town, London, UK
Dear leinz and Mai,
In the late 1950s we worked on sel organization together. And since then we hae worked in
sympathy.
My own theories-o conersations and o the interaction o actors-come rom that work. In act, it
seems to me now, they come as much rom your work as rom mine.
\hat I present beore you in this lestschrit is the story o how my own ideas irst burgeoned rom
understandings you irst brought to lie, and grew to ully ledged theories. I expand these into mature
orm to present the results o this generation which, as you know, I call 'conersation theory'. Again,
taking some o the constraints o conersation, I de-limit them-as you hae shown me how to do-to
gie the basis or my 'interaction o actors theory'. In the middle I place an axiomatic presentation
which I hae made or you. It conjoins and extends the theories. It is intended both as a bridge and
a uniication.
I end with something decidedly personal: 'conersation and generation'-or you are not an ideas
machine. \hat I get rom you I get because you made the ideas: to adopt ,and adapt, you
'objectiism is the delusion that ideas can be made without a maker ,and can be enacted without an
actor,'. In this case it is the interaction o actors.
1his oering is abrupt and ragmented or it is limited to be but a paper. It may be that it is not
understood. I it is not understood I shall be sad. But what is important is that you understand it.
And I am sure you will.
Dear leinz, this paper comes to you with loe, empathy and .,vatie.
Gordon.
1his paper is the last paper written by Gordon Pask beore his death on the 29th March 1996. It was ery important to him to write this paper, both as
the tribute he had longed to pay to his riend and mentor leinz on loerster, and because writing this tribute gae him the chance to put together the
main elements o his work so that some o the coherence he saw could be indicated to others, with a new meta-theoretic position emerging ,see the
section 'Axioms and Other Postulates' set in a dierent typeace,. It went through ery many ersions as Pask struggled against illness and the
tremendous demand he was placing on himsel, to complete the text, and I beliee that his need to make this statement was one o the things that kept
him alie. le posted me the last ersion on March llth, and this is the ersion that appears here. I hae decided not to modiy the paper in any way
except by spell-checking and punctuating. It stands as it was sent to me. I beliee it is as clear as Pask could make it, een it i is somewhat ragmentary.
Indeed, I think the ragmentation is part o its charm and is appropriate to what is being said ,in as ar as I understand it,. 1he lack o a glossary o
terms was Pask's choice. 1he paper in the orm it takes here is the result o many hours o humble relection by Pask: humble in that he sought and
listened to much tough adice, and humble in that it was written through loe in the ace o death. It is a remarkable document and is, I beliee a itting
testament o how Gordon elt about leinz and to Gordon's lie o creatie struggle, also.
Guest Lditor
CCC 0731-7239/96/030349-14
1996 John Wiley & Sons Ltd Copyright fee paid at British Library
RLSLARCH PAPLR Systems Research
The Progenitor of Conversation and Interaction Theories
352
Hcinz von Focrsicr's Sclf Organizaiion,
iIc Progcniior of Convcrsaiion and
Inicraciion TIcorics
Gordon Pask
Sstcn HcscucI Ltd, Od Toun, London, UK
Oer more than three decades leinz on loerster and I hae collaborated and worked together as well as in
separate laboratories. 1his contribution gies a terse account o work which we hae done together and which is
releant to leinz' prescient notion o sel organization and its many arborizations. In the course o doing so it
spells out some o the history associated with cybernetics to which both leinz and I adhere.
Keywords concept, conversation (theory), interaction (of actors theory), observer, P-individual, self-
organization, spin.
IN1RODUC1ION
1his contribution might be called alpha to omega to ininity. \hy is that Alpha marks the beginning o actie
co-operation and liaison between leinz on loerster, my immediate colleagues and mysel. Omega marks the
current situation and since neither o us has ceased to create or to think, it is an indication o what is to come,
since all theories worth their salt eole. It is possible to demarcate seeral eras. lor example, the era o adaptie
machinery at point alpha and the notion o sel organization as proposed by leinz on loerster. During that
period a great deal o work took place and although leinz remained the same in spirit, his repute, or lag, i you
like, has undergone certain transitions. lor example at alpha he was known as a physicist and at point omega he
is known as a group or amily psychotherapist. 1hat is the popular image. 1he transition was not continuous and
not all o the discontinuities are charted in this contribution. lor example, there was a period when our interests
eered towards epistemology, anthropology, group interaction and consequently conersation and the interaction
o actors. 1hat is why there is such a close liaison with sel oganization attributable to leinz and conersation
theory ,C1, and inter-action o actors theory ,IA, to me.
In the course o the discussion it is possible to trace the eolution o cybernetics, to which both o us adhere,
and to obsere a transition between irst-order cybernetics and second-order cybernetics as they are currently
known. Also, it is possible to show the resilience o cybernetics, although it permeates all sciences, and its
identity may seem, to some people, lost. It has suried the zany and puerile epochs o cyber-punk the tedious
and repetitie taxonomies which accompany any science and the politically necessary changes in title. lor
example, there was a period when one could not speak o cybernetics but only o bionics, inormation science or
general systems theory. 1hese changes hae been related to our joint or mutualistic work.
SLLI ORGANIZA1ION
In 1958 or thereabouts the small town o Namur, a proincial capital in Belgium standing on a large hill at the
biurcation o the riers Meuse and Sambre, was a meeting place or the International Cybernetics
Association. Right at the top o Namur is the hotel Chateau de Namur. It is connected, either by a teleerique
which met with a mishap ,nobody was hurt but the teleerique was dysunctional, or by a tortuous road
around the mountain, which I alluded to as a 'hill', which is the Citadel o Namur. It is a citadel since it is
penetrated with a labyrinth o passages and the road around is precipitous. 1he passages are like Gibraltar,
with a military ballroom, een, hewn out o the rock. lrom the top you come down to the lotel Comte
D'larscombe or, lower still, to the railway station, the Place de la Gare where there is the lotel de llandres
and in those days, at least, the lotel Coq D'Or in the cae o which I was sitting. I had just had a beautiul
argument with Marin Minsky in the Chamber o Commerce, as a result o which we became irm and lasting
riends. leinz on loerster appeared at our table as a magic igure. I hae written a short tribute to this
Systems Research RLSLARCH PAPLR
353 G. Pask
aspect o his character on his retirement or the American Cybernetics Society lorum called '1he Importance
o Being Magic' ,Pask, 1980, and shall not stress it in this contribution.
But he is magic. \es, he is a conjuror and knows the tricks o legerdemain and also the skills ,or example, he
can ride a monocycle around a casino or cabaret loor while putting on six hats, one on top o the other,. le
is also a radio announcer. le is an eminent physicist, haing written papers both biological and physical on,
or example, memory not being stockage but being vevoir.
1he magic I intend to deal with lies amongst many curious and enigmatic subjects o his attention: it is that
o sel organization. I reer you to 'Obsering Systems' which is a sensible, sensitie collection o his works
edited by lrancisco Varela. Although I intend to deal with the chapter on sel organizing systems, I inite you
to read the lot.
leinz described sel organization in a seductiely respectable way. le used the entropic inormation measure
o Shannon, namely , and the entropic inormation which a system could possibly hae, that is va. le
pointed out that a ratio measure, redundancy, R ~ 1 ,va, is an appropriate measure o organization and
that a system is sel organizing i and only i dR,dt 0. Now H is clearly related to one o the measures o
the ractal dimension o a point set and R is one o seeral ratio measures o the sort that are used in
determining the eiciency o heat engines. 1aking this measure he proceeded to dierentiate it with respect
to a quantity called time. 1his sounds terribly respectable until we realize that and va are, a priori,
independent ariables.
It is deductiely true that in order to be sel organizing the entropy l multiplied by the rate o change o
maximum entropy dva,dt should exceed a alue greater than the maximum entropy multiplied by the rate
o change o entropy dl,dt. It is a simple condition and one that went unquestioned at that stage ,rather
disappointingly so,. It naturally brings to mind adaptie systems becoming more or less ordered or more or
less disordered. In act we may appeal to Schrodinger's delightul '\hat is Lie' ,Schrodinger, 194,, his
inaugural lectures at Dublin. lere he talks about order rom order and order rom disorder, as does leinz.
le thinks o lie or, i you preer it, sel organization, as being something o this kind.
In addition, howeer, leinz introduces a urther commodity i.e. 'noise' in the sense o Shannon. 'Noise'
stands or undierentiated energy ,although the point is made more clearly in later papers in the same
reerence olume such as 'Molecular Bionics',. le urther exempliies the limiting condition or sel
organization by shaking ,an undierentiated energy, a box o magnetised cubes or other geometrical orms
and o dierent species with respect to their magnetic aces. I these annul or i the cubes are non magnetic
there will be disorder. I they are oriented in such a way that they may stick together like the droguli o
Penrose and O. D. \ells, the system becomes almost completely ordered and crystalline. 1hese are the
limiting cases o sel organization ,R ~ 0, dR,dt ~ 0,.
It is ery conenient to think, as on loerster does, o the undamental ariables o and Hmax as being
regulated by demons and, or example, being determined by the prior alues o a,at and va by the
prior alues o ava ,at, the demons pulling the necessary leers or knobs. loweer, that would be an
uninteresting resolution and could be no more than an adaptation. 1he reality is that what has to be
maintained is the preiously stated balance between ratios: the demons goerning and va and their
change are asynchronous because H and Hmax

are independent, a priori, and so are the demons.
Really, interesting sel organization comes about so ar as the demons are coupled and learn about each other
and each other's enironment, this coupling being accomplished by a Petri-type inormation transer. lor me,
at least, the interaction o the on loerster demons is paradigmatic o a conersation between two
participants.
Amongst other consequences, an obserer who comes to know the system must be a participant in the
system. 1he boundaries o the system, ar rom being pre-abricated, are created by the actiities o the
system. 1his is a prescient notion o autopoiesis ,as deeloped by Maturana and Varela,, or organizational
closure, as we called it at Brunel Uniersity and in my own laboratory, System Research.
Sel organization is an important and ery useul discoery. It was particularly useul as I shall now indicate:
in the context o our research programme which sought to ind subjectiely hard data, in particular in the
ield o adaptie systems or regulation o perception, learning and innoatie processes on the part o people
and groups. But our sights were set urther, we wished to adumbrate organizations o all kinds: industrial,
RLSLARCH PAPLR Systems Research
The Progenitor of Conversation and Interaction Theories
354
biological, societal, and so on, which is the rationale ,later on, or a discussion o P-indiiduals ,psychological
indiiduals, and M-indiiduals ,machine indiiduals rather than necessarily biological ones,. Sadly in that era
,the late 1950s and early 1960s, attempts to maintain a status or this work were apt to be rejected as anciul
or not objectie ,they certainly are not objectie, and unscientiic ,I am inclined to deny this,. It was
necessary in this situation to demonstrate in a way acceptable to the established order that conersations did
occur and that consciousness was aroused, to address the notion o consciousness in an inormatie though
not explicatory manner and at least hint upon the notion o awareness. 1o do so was maniestly a large
project and the labour inoled could hardly hae been justiied had we concerned ourseles with education,
the particular ield o inquiry, alone. Or, indeed, o experimental psychology, which is inclined to misapply
statistical methods in the name o science or to neglect the presuppositions truly inoled.
lence we embarked upon an unashamed search or and to some extent discoery o what consciousness is
and how it may be manipulated other than by giing an anaesthetic or bashing people on the head with a club
and making them concussed. \e set up a system o interaction between participants through a mechanized
interace.
It is possible and more eicient to perorm this operation using human participants and the techniques o
hypnosis or o Piagetian interiew or the like. loweer, this would not hae pleased the establishment who
wished to hae behaiours recorded and statisticised in some way. 1o accomplish this we resorted to adaptie
machines which were not digital computers ,this was beore the computer generation,. Our programs were
probably more elaborate than current programs on computers and they were used widely both in the ield and
in the laboratory to regulate learning, innoation, perception, repetition, sel-consistency, oerload and
training. In this we recognized throughout that our interaction was a model o the interaction between on
loerster demons in a sel-organizing system and that the phenomena we obsered were those supporting the
eriication o this principle.
Apart rom the empirical work, the theoretical load was considerable-though the idea o sel organization as a
primary issue in mental or social actiity was consistently airmed. 1he type o experiment we did, or
example in the context o learning, was to set up an interaction through an interace ,as beore, it would hae
been more eicient to use a participant experimenter such as the Piagetian interiewer,. ,1he paraphernalia o
the learning and looking system such as CAS1L-Course Assembly System and 1utorial Lnironment-and
other systems we deeloped is completely unnecessary, in practice, except or proessionals anxious to do
away with your conceptions.,
1he results o these early experiments and those carried out throughout the 60s and early 0s were generally
positie. 1hey were the precursors or mesh and interace experiments and inoled adaptie machines.
1hus the adaptie system achieed, in concert with the participant, a higher rate o learning, a greater eiciency o
learning, a sel-consistency which i ed back on the mechanical enironment, increased its diersity. 1he participant
was obliged to signiy this act and the diersity was reduced. loweer these positie results-positie in the sense o
the establishment-were accompanied by a negatie and much more interesting result. As you increase the elaboration
o the adaptie machinery enabling it to adapt hierarchically, or example, to seeral leels and to hae a ector o
reactiity, which is extensie up to say eight or nine ariables that are related in the hierarchy, you appreciate that
something odd is happening. 1he participant either gets bored ,which is ery natural, and opts out o the
experiment, or preerably opts out by playing with the adaptie system. 1he more elaborate this system is, the more
it can be 'programmed' by the participant, rather than 'programming' the participant.
lor example, in an adaptie training system it is the correctness and rapidity o response which is ed back to the
adaptie system, in order to increase or decrease task diiculty. In a system constructed to maintain attention an
index o sel-consistency replaces correctness and can be used to introduce or reduce any learnable rules in a
mechanical enironment ,remoal o rules increases the diersity and ice ersa as well,. Some model o how the
participant learns, mentates and arouses attention must exist in each case. 1he adaptiity o the system can be and is
improed by searching among a collection o models and selecting the one ,i any, best suited to the participant and
his or her mental repertoire.
1he elaboration in the machine o his or her mental model is, howeer, limited by the initary constraints o the
system. 1he orm o constraints is gien away by the use o terms like 'hierarchy' or 'input' and 'output' ,assumed to
remain inariant and consequently ixed by the design o the system,. 1hey are allowed to ary only as a designer
Systems Research RLSLARCH PAPLR
355 G. Pask
prescribes. 1he participant is simply thwarted and his or her mental repertoire cannot become maniest in such a
system.
It was necessary or us to show that such strict interactie conditions must indeed preail. 1hat sel organization is
not adaptation or habituation or some tawdry orm o learning in the sense o mathematical learning cures. It is in
act a undamental paradigm which merges and orms a bridge between the old cybernetics and the new.
1he old cybernetics has been elaborated ad ivfivitvv.
1he new cybernetics ,some call it second-order cybernetics in contrast to the irst order o classical black boxes and
negatie eedback, is burgeoning well beyond the bounds o respectability which were imposed by the establishment.
I interaction, albeit interrupted by a phone call or a business trip, is unettered in intellectual low, I call it a
conersation which leads to the exchange o concepts, not necessarily o the topic, but o the participants. It may or
may not be about whateer is the alleged topic o conersation. Later we consider interaction o a broader kind, the
interaction o actors. Conersations are eents which hae a beginning and hae an end and may be subdiided into
kinematic portions, in contrast to interaction o actors which cannot. Comparable notions hae been mooted by
Petri, Glanille and others.
SOME DETAILS
People like leinz do not in essence change. loweer, their interests and reputations do luctuate in a material way.
lor example, leinz explored at one stage biological and brain neuro-physiology. 1he particular transition in
reputation and orientation o thought which I intend to deal with is a dierent one. It did not occur continuously, as
I am writing it down, but in small jumps with small groups and it went on rom small groups regulated through ,not
by, a machine. Soon this interest, which started with group interaction through ,not with, a machine, matured into
an interest in small groups held together by normatie constraints-paradigmatically, amilies. lence it was not
surprising that leinz became concerned with group psychology and anthropology ,along with Gregory Bateson, and
amilies. In particular, in the atmosphere at that stage, he became known or amily psychotherapy. 1his led to an
orientation toward such phenomena as the 'double bind' and 'alienation'. Let me emphasize that in no way did leinz
discard his other interests and origin as a physicist. But it is ery reasonable to notice that his undamental ocus o
attention was a conersation. At which point we united as social psychologists and cyberneticians, since conersation
is the principle medium entailed in the interaction o members o such groups. A serious study o this matter
requires a undamental re-orientation towards concepts, their character and exchange. In order to explicate this I
intend to introduce some deinitions which will be o alue in interpreting the transition in the reputed position o
leinz on loerster and to relate sel organization ery closely to conersation theory ,C1, and interaction o actors
theory ,IA,. 1he deinitions are as ollows.
Algorithms or programs are labeled, what-eer they are, by what they achiee, possibly a set or collection o entities
thus:
,a, PROC;1) where 1 may be anything, e.g. a tower, dog, etc. 1hat is anything that may be pointed at or
ostended.
,b, Let a participant be the ordered pair P-indiidual, M-indiidual indexed by a sub-script z with alues
such as ., , ..., standing or dierent participants.
,c, Participants are capable o interpreting an otherwise inapplicable PROG so that we may write PROCZ;1)
~ PROC;1), ^1RZ > where ~ ., , ...,
,d, 1he ^1RZ is the interpretation on an occasion by a participant z.
Consider now the ormation o a concept in greater detail because that, ater all, is what is shared wholly or partially
in a conersation.
Let PROC;1) be a program or i you like an algorithm or achieing 1, whateer that may be. Let PROC;1) be not
equal to PROCZ;1) which, howeer, exist in ordered pairs PROC and its interpretation or its compilation in some
thing, some kind o M-indiidual, biological or not. Because to talk about running a program is utter atuity unless
you know the computer and the compiler and surrounding conditions or application ,Lars Logren at the Baden-
Baden meeting and his comments in my lestschrit-edited by Glanille, 1993,.
Gien these conditions a PROC is applicable and i applied yields a product. 1here is a process product,
complementarity. 1hat is, there is no such thing as a product without a process to create it and no such thing as a
process that creates no product. Let that product be written as ollows .P;CO^Z;1)) ~> DZ;1) ,which is a
description or Spencer Brown distinction o, say, the characteristic behaiour o the dier-entailed growth phase o
RLSLARCH PAPLR Systems Research
The Progenitor of Conversation and Interaction Theories
356
a mushroom,. 1his recursion gies us the nature o CO^Z;1). Now 1, I admit, is undeined, except that it may be
regarded as a locus or target o a conersation between participants A and which are as yet undeined. \e shall,
howeer, proceed to deine them. It is eident that one may image, as one possible representation, a PROC as a
directed line which i it exists eats its own tail, hence becoming a directed circle ,directed either clock-wise or
anticlockwise,. It has hermeneutic or coherence truth, namely, a progressie reinement o meaning ,1aylor,
Rescher, on \right and others,.
It is the case that eery CO^ contains an ininite number o reinements called PROC ,with their indexes, subscripts
or superscripts diering-either, both or multiple-which can become obscure-where appropriate, the limitations are
typographical only, no doubt, but any o which ,singly, or all ,together, are capable o application. 1he application
should be more careully indicated to read PROC or CO^ ,because PROC is a CO^), produces amongst other
things, DZ;1). In particular, it also pro-duces, since distinctions can be taken to 'repel' each other by irtue o being
distinct, either it or any other application in the mental repertoire, an operator, CO^Z, acting upon a domain o
concepts and o the type 'concept' haing the eect o producing and, incidentally, reproducing the concepts. 1hus
the system is both organizationally closed and inormationally open.
But what is the scope o this miraculous operator CO^Z. Its scope is what I hae come to call a P-indiidual.
lence A and are deined where its domain is some other concept or concepts, its scope is the entire domain o
concepts under consideration.
1he circularity induced when a PROC eats its own tail gies rise to a product, which may be represented as a cylinder
surrounding it and is a torus o genus 1. loweer, there is a restriction that CONs are produced and incidentally
reproduced in order that they be stable, not only exist transiently. lence it ollows that there must be more than
one, in act, to make it simple and easy to expound there must be three o them at least. Since all the distinctions are
pushing each other apart towards ininity ,they 'repel' each other,, what holds them together I we hae chosen
clockwise as the direction or our concept, CON, then an anticlockwise lux encloses them and holds them together
as a stable concept and moreoer it produces a boundary or distinction around it. 1hat is the name we gie to the
concept also justiying our assumption about the ininite number o procedures or reinements within any concept.
In addition to those circular clockwise or anticlockwise motions there is an orthogonal orce directed outwards
repelling each rom each other so that we must hae, in order to orm a stable concept, a contrary orce-the
conseration o parity, in act, surrounding the whole lot. lurther, it is possible to link concepts together under
certain conditions which hae been speciied in detail in many other papers. 1he requirement is, o course, that in
linking them together one can enisage an operator acting orthogonally, with respect to the circular clock-wise or
anticlockwise operators, which penetrate the boundaries that are created as products and unoldments on the mesh
or the nexus o concepts, i.e. the mental repertoire o z, where z ~ A or B ,as beore,, in such a manner that it orms
a heterarchical or een hierarchical structure. As it happens, one can extract ,by order presering union, and
coalesce the unoldments which look like trees or quasi-trees. 1his is thought or many thoughts, some o which may
be plans. Unique lines o thought are delineated by ollowing one tree and noting, amongst other things, that
percepts being concepts is a special case where the application o a concept traerses the enironment. 1he domain
o a percept is a loop in the enironment which also determines the order in which thought o actions may be
executed.
lurther, the domain o operation o the concept is one or all o the remaining concepts. 1he domain o some
concepts, howeer, is qualiied in quality and directionality. I shall call this a prepositional operator appended to
those concepts which are responsible or diering sensory modalities and their directionalities.
In order to ormalize these concepts we adjoin a so-called prepositional operator, q. 1his oercomes the diiculties
such as, or example, you can't get to the airport until you hae taken a taxi ,when thought is realized as action,. 1hat
is, or example, in eeryday lie, trael may be thought about in many ways, but in action it must be done as ordered:
taxi, airport, aircrat, unless a light is delayed in which case wait. 1he statements are conditional. More orceully in
the domain o geography we may think o let and right as we please, but we cannot simultaneously turn let and
right. ,Although in the domain o politics we oten do!,
In this we must keep in mind the stability o the concepts ,not merely their transient state, and the domain upon
which CO^Z operates:
,i, 1he domain o all concepts and percepts is the mental repertoire o a participant ,or participants,.
,ii, 1he domain o CO^, is the domain o all concepts existing. Its range is all the concepts that may or
may not exist. Its domain is what it operates on. Notably, the literal transer o tokens takes place in the
Systems Research RLSLARCH PAPLR
357 G. Pask
set theoretic |i.e. the D;1)| part o the concepts and it is possible, in act usual, or the description D to
include a listing o PROC;1).
\hat is communicated in a conersation is, o course, this set theoretic part. \ou can speak ery reasonably o
PROC being the member o a set o PROC., but you cannot speak o PROC as being a member o a set. It is a
member o a collection, a category, perhaps, but not o a set. A set theoretic part may and usually does contain a
listing o the program PROC and this may be communicated. lurthermore, it may be inscribed at an interace which
I call an entailment mesh. 1his has been sadly maligned, simpliied and idiotically reduced to the cruder notion I
started with which I call an entailment structure.
,\hat I hae continued to call an entailment mesh can be actiated so that the entities residing in it become dynamic
concepts. Perhaps not the concepts o you or I, but o someone or some liing thing. \hereas in the original
entailment structure these dynamic entities were represented by models or processes attached to the names, this is
no longer the case. An entailment mesh is no longer a stupid nominal thing o topic networks or inormation maps
or the like.,
Suppose ., one participant, has his or her meaning attached to a concept, 1, o let us say, a dog and the other
participant has his or her meaning attached to a dog. And .. and . meaning een though they ostend the
'same' thing must be entirely dierent. lor example . considers an aectionately well-clipped poodle and
considers a erocious quadruped like the lound o the Baskerilles. 1his is the really ascinating and important part
about conersations.
1he main purpose o conersation is not communication about 1, whateer that may be, een though 1 is the ocus
o the conersation. But about . and , about A's iew o , about . iew o A, about getting to know each
other, about their coalescences and their dierences, and the society they orm. 1hat is why I was so particular about
speciying a P-indiidual. 1he M-indiidual ,which goes with it, may be any interpretatie medium, human, animal,
organizational, possibly mechanical, possibly een cosmological. loweer, it is expedient to notice that the M-
indiidual in which a P-indiidual is incarnated should bear that index o the P-indiidual, namely z. I you like, the
particular moiety o the interpretatie medium, the ^1R, assumes the name o the P-indiidual which is
incarnated in it.
Let us restrict the term 'conersation' to interactions which hae a beginning and an end. 1hough there are
interruptions-a tele-phone call, an absence on a trip or whateer-the conersation continues as though it had not
been broken o.
AXIOMS AND OTHER POSTULATES
(1) Let PROG (T) be an algorithm open to application if it is interpreted in some
(any dynamic) medium to yield T.
(2) PROC
Z
( T ) = <PROG ( T ), INTER
Z
> such that AP(PROC
Z
(T) => d
Z
which is a
member of or a subset of D
Z
(T )
(3) Let us suppose for convenience that expressions of form
(AP(CON
Z
(T)) => D
Z
(T)) = T
Z
that is T
Z
=<AP(CON
Z
(T)), D(T) >
Lacking the subscript z the process exists, if at all, in an interface.
(4) For z = A or B CON
A
(T) and CON
B
(T) are such that
T
A
T
B
(amongst other things) in a particular CON*
z
, where '' stands for 'is
not equal to' .
(5) AP (CON
Z
(T) ) => D
Z
(T); D
Z
(T) is a cylindrical carapace, the products of the
motion.
Under AP, of T
Z
, in particular the form under AP of
< AP(CON
Z
(T)), ); D
Z
(T)>= T
Z
is a convenient rewriting of the lengthy expression and is valid under the
conservation of AP. It does for special values of the index z such as T
A
and
RLSLARCH PAPLR Systems Research
The Progenitor of Conversation and Interaction Theories
358
T
B
indicate a pair of participants and if they happen to be embodied in
distinct parts or M-individuals of the interpretative medium, they are
separate participants. (Notice both T
Z
and the ensemble continually evolve;
they are not static).
(6) We may sketch in the creative, productive and reproductive, processes
(production and reproduction under AP). The productive and reproductive
operators are of a clockwise or anti-clockwise form and 'eat their own tails' .
Using the previously introduced notation a participant is defined as <P-
individual, M-individual>; since the extent of the medium or fabric in which
a P-individual is incarnated is indefinitely specified (it can be anything) we
may leave it unindexed except for that portion of it INTER, in which the
defining subscript is inherited from the P-individual therein incarnated or
embodied.
(7) Note that PROG (T) or in particular cases PROG
i
(T) is in D
Z
(T).
(8) i is not equal to j.
(9) => implies goes into (amongst other entities), notably for some productions in
z's repertoire CON*
z
.
(10) CON*
z
(T) = [PROC
Z
i
(T)] (which is parallel by token of '[and]').
(11) CON
Z
(T) = {PROC
Z
j
(T), [PROC
Z
(T)]} (which is concurrency, an essential
feature of life).
The essence of this is captured by noting that this expression is not as trivial as it
seems. It does not mean that the activity of CON*z is started and stopped
while AP (CON
Z
(T)) or any other CON is started and stopped. There is no
such shuttling. The processes are simultaneous. In fact they are concurrent
and that is the form of life. This is z's consciousness obtained between the P-
individuals constituting the participants, z, that is z's awareness. The P-
individuals may reside in one medium such as one brain in which the
conversation is between parts of the mind. On the other hand they may exist
in different M-individuations in which case the conversation is between
participants. This is A's consciousness with B.
(12) For at least some productions in the z-ensemble the term => (amongst other
products) means produces CON
Z
*.
(13) AP-individual, z = A, B, . . . , is specified as the scope of CON* (That is a
collection of concepts T
A
, T
B
, ... for z = A, B, . . ) .
(14) Whereas it is possible to out-line the productive and reproductive character of
concepts it is also possible to illustrate the transfer of set theoretic tokens,
elements of or sub-sets of D through an interface, I.
(15) Due to the fact that a conversation is defined as having a start (s) and a finish
(f) we have the following inequalities.
(16) (a) T
s
A T
f
B, (b) T
s
B T
f
B, (c) T
s
A T
s
B.
(17) (a) T
f
A

T
s
B, (b) T
f
A T
s
A, (c) T
f
A T
s
B.
Systems Research RLSLARCH PAPLR
359 G. Pask
(18) A B.
(19) If t
A
and t
B
are the duration times of A and B then t
A
t
B
. This is so for all values
of the participant, index z. Clearly because of the beginning and end certain
events become evident with the transfer of set theoretic tokens (syntactically
unique symbols) from A to B. The notion of understanding between A and B is
a punctuation mark which must be specified. How is it specified? Clearly
because of the beginning and end certain events give rise to what I call, in a
technical sense, an understanding. Briefly, A asks B 'What?' questions by
ostending some intellectual or concrete entity. B ostends T, that similar
looking entity, intellectual or concrete, as it may be. Having asked that, there
is an exchange of questions, this one being the basic 'What?' question. The
exchange of questions continues to 'How? ' . How do you make e.g. a dog?
How do you use a dog? How do you do what ever you wish with a dog or
whatever T stands for. These explanations are exchanged. Finally one asks
'Why?' . Why is it that you chose this derivation of dog from other concepts in
your repertoire and perhaps concepts you invented. If this were firmed up it
would complete the recursion. But firmed up it is not.
The durations for A and B, . . . , differ since t
A
and t
B
are distinct and hence
time is vectorial but commensurable. Commensurable in terms of duration,
where a mapping is possible at each punctuation point of understanding on to
Newtonian time.
Clearly it is possible to approximate, because of the beginning and end of certain
events, by understandings, where it becomes evident that transfer of tokens
from A to B is set theoretic only. But recall that PROG
i
(T) is usually a member
of D
Z
(T) from AP(PROC
Z
(T)) => D
Z
is usually a member of D
Z
(T) .
(20) By process-product complementarity each circular process gives a distinction
as a product. Hence the snapshot (kinematic image) is that of a many-holed
torus of genus 1 to count-able infinity (for each hole). Many onion-like
layers are distributed, alternately marked by their circular generating
processes as being clock-wise or anticlockwise 'orientation' (that is of their
'spins ').
(21) Due to the repulsions produced by the clockwise and anticlockwise motions
of mentation and coalescence to form conceptual clusters (that is, coherent
clusters), there must be a process orthogonal to the clockwise or
anticlockwise processes of mentation. This process, already noted, has been
called unfoldment,and unfoldment is conserved throughout the ensemble.
There are certain exceptions under certain types of anaesthesia and mental
pathology.
AP (if you like, undifferentiated energy) is, however, conserved over the
clockwise and anticlockwise processes of mentation and over the process of
unfoldment which cuts across the distinctions of clockwise and anticlockwise
mentation processes and thereby reveals them as distinctions or
descriptions-a slightly extended use of which we have adopted throughout.
This applies, of course, to the clockwise and anticlockwise processes of
production and reproduction as well as the application of concepts.
(22) If distinction is conserved it follows that the mesh of conceptual clusters is
transformed into a tree-or heterarchy, perhaps. Further, that it is possible to
extract certain branches of this tree as being trains of thought whereas the
heterarchy or hierarchy are all possible thoughts that one might think, for
RLSLARCH PAPLR Systems Research
The Progenitor of Conversation and Interaction Theories
360
example those one might think in moments of great hazard as for example
when mountaineering.
Unfolding from T a particular free or arbitrary train of thought is dissected
out.
(23) These thoughts may be in any direction. For example in this world (of
geography) I think of left and right knowing full well that I cannot turn left
and right simultaneously whereas in the world of politics I can entertain
both left and right points of view at one and the same time. This is the
freedom of thought.
However not everything can be done in every world. To repeat the point in
the world of geography, it is impossible to simultaneously turn left and right.
Whereas in the world of politics it is perfectly natural to subscribe to left and
right views. The dilemma is resolved by attaching a prepositional operator q
which characterizes the domain of concepts. A concept with this attachment
adjoined is called a percept. The application of a percept gives rise to action,
but action appropriate to the universe and domain it describes and
appropriate to the order of operations in such a domain.
(24) A conversation is an inter-action between at least two P-individuals who if
embodied in different media are distinct participants.
(25) The truth value of a conversation is a metaphor or an allegory designating
an analogy (difference A, B; similarity, generalization of coherent content of
A and B ) .
(26) To sum up the gist of the argument, there are clockwise and anti-clockwise
processes of augmentation; the application of concepts in one circular
direction and their productions and reproductions in the opposite direction.
Call this mentation (AP being conserved). Each process gives rise to a
distinction.
(27) There is (also conserved) UN or unfoldment for one or more concepts. UN
cuts its way through the distinctions made by mentations and is more or
less specific thought. UN is a category the-oretic unfoldment of one or more
concepts in a direction orthogonal to the clockwise or anticlockwise motion
previously discussed.
(28) Thought gives rise to action if a concept is qualified by q. If so qualified the
concept is a percept.
(29) Interaction in a conversation is restricted essentially by an s (start) and f
(finish) which admits a punctuation of symbolic set theoretic transfer
between P-individuals.
(30) If these P-individuals are incarnated in the same M-individual then we have
the dialectics of mind.
(31) If embedded in distinct M-individuals we have, in the usual sense,
conversation between two participants.
(32) In each case the truth-existence-value of a conversation, or mental dialectic,
is a metaphor or allegory designating the analogy which is the
conversational act. With reference to clause (11) awareness and
Systems Research RLSLARCH PAPLR
361 G. Pask
consciousness are carriers of affective meaning, i.e. the real meaning:
emotion.
IN1LRAC1ION OI AC1ORS
1he interaction o actors is a matter that calls or ar more undamental assumptions. I shall sketch some
o these only, the main intention being to compare and contrast conersation theory with interaction o
actors theory.
,i, An actor is a participant enacting a role. 1his role may be deined ,as in Lliot Jacques' work e.g.
Jacques, 1964, or outlined as in lenton Robb ,Robb, 1980,, or may emerge rom a situation and
the actions taking place. lor example, in Robin-son's research ,Robinson 19, into the Brighton
Rent Group ,an orderly organization o squatters, where roles o 'public relations manager' and
'legal expert' emerged in order to maintain an organizationally closed and inormationally open
sel-organized ,and iable, system.
,ii, Once established a role may be played ,oten in dierent ways, by the same actor or dierent
actors. lence there is mutual actor and role interdependence which inluences the eolution o
the system.
,iii, Conersation is deined as haing a beginning and an end and time is ectorial. 1he components
o the ector are commensurable ,in duration,. On the other hand actor interaction time is
ectorial with components that are incommensurable. In the general case there is no well-deined
beginning and interaction goes on indeinitely. As a result the time ector has incommensurable
components. Both the quantity and quality dier.
,i, In both cases the generalization o distinctions and their resolution by analogy creation is a matter
o conlict and conlict resolution. 1his could be represented by orms o competing and co-
operating ,'chaotic' or 'strange', attractors i it were possible to construct a respectable phase
space. loweer, in iew o the ectorial nature o time this is no simple matter.
,, 1here is some similarity between the many holed tori and their onion-like layers o distinctions
|clause ,20, o the last section| and dynamic structures in the general interaction ,o actors,
theory, howeer in the general case a 'snapshot' is inadmissible |it is at most a useul approxi-
mation in clause ,20,|, we are in a kinetic world and the tori and their skins are deeloping.
,i, Interaction o actors has no speciic beginning or end. It goes on oreer. Since it does so it has
ery peculiar properties. \hereas a conersation may be mapped ,due to a possibility o obtaining
a ague kinematic, picture-rame image, o it, onto Newtonian time, precisely because it has a
beginning, and an end,, an interaction, in general, cannot be treated in this manner. Kinematics
are inadequate to deal with lie: we need kinetics. Len so as in the minimal case o a strict
conersation we can construct the truth alue, metaphor or analogy o A and B 1he .,
dierences are generalizations about a coalescence o concepts on the part o . and ; their
commonality and coherence is the similarity. 1he dierence ,reiterated, is the dierentiation o .
and B ,their agreements to disagree, their incoherences,. 1ruth alue in this case meaning the
coherence between all o the interacting actors.
It is essential to postulate ectorial times ,where the components o the ector are
incommensurate, and urthermore times which interact with each other in the manner o Louis
Kauman's knots and tangles ,Kauman,1995,. 1his has been noted, or example, by Petri ,1963,
and by the editor, Glanille ,195, 1988,. \e see it in any airport. 1he times o the booking
clerks, the captains o the crat, the baggage controllers and so on are all dierent times. Quite
dierent, quite dierent in kind and quality. \ou cannot map them on to the simple co-ordination Newton
deised. 1his, I beliee, is well known but requently hidden as not supporting the status quo.
A COMMLN1ARY
1he orm not the content o the theories ,conersation theory and interaction o actors theory, returns to and is
congruent with the orms o physical theories, such as wae particle duality ,the set theoretic orthogonal unoldment
part o conersation theory is a radiation and its reception is the interpretation by the recipient o the descriptions so
exchanged, and ice ersa,. 1he particle aspect is the recompilation by the listener o what a speaker is saying.
RLSLARCH PAPLR Systems Research
The Progenitor of Conversation and Interaction Theories
362
1heories o many unierses one at least or each participant-one to participant A and one to participant B--are
bridged by an analogy. As beore this is tbe trvtb ratve of av, ivteractiov; tbe vetabor for rbicb i. cvttvre it.etf.
A DISCUSSION
During the preparation o this paper I had the great adantage o a discussion with Graham Barnes and he asked
me, amongst other things, what a conersation with leinz on loerster, himsel, was like. I replied that it could be
terribly ormal, like the last ew paragraphs.
'Is that all' said Graham. 'Oh certainly not.' I replied. '\ell tell me what was so special in a conersation with
leinz'
I pondered or a moment. '\ou need a context to talk about a specialty o some kind. I am going to choose courage
as the attribute concerned.' '\ell,' said Graham 'Go ahead'.
A little bit o background is needed. leinz and Mai had three sons, o whom Johnny was the eldest. Johnny had just
been killed, a ew days past only, in a motorcycle accident while sering with the International Peace Corps. leinz
was grieed and although he had agreed to speak at and in a sense to co-chair an AGARD conerence in Lurope, I
didn't or a moment expect that he would make it. But in act he did, een though
distressed. \e met with a kindly reception in lrance and in many other places, with what leinz called our 'circus'
consisting o eminent people, proessors and locally a lot o students, research associates and assistants o proessors
o great merit. ,In a sense I belonged to their generation., Only at one point was any oence created and that was by
a bureaucrat in a certain town in Germany, and it was because they entertained us as a prestigious group and part o
NA1O so badly they had the erontery to inite us to dinner in a railway waiting room and sel-serice buet. At
the end o this eeryone was obiously upset. I asked leinz i as a special aour he would gie the ote o thanks.
le did so in perect German. I hae neer heard a speech so humorous, so sympathetic and so damning o the
bureaucrats. \e walked out and seeral o the graduate students, assistants and so on ollowed us because they had
become ery ond o leinz. I said to him 'Shall we go as usual to a restaurant and ask them to a glass o wine' le
said. 'Oh, yes I suppose so but I really eel in no mood or jolliication.' I was ashamed but I sort o lead him on and
we went to a small place and I had a word with one o the local participants in the conerence and I said '\ell let us
go as soon as possible,' relating why. Namely leinz' recent bereaement. \et leinz came, chatted to the students,
satisied them. le got to know them because there is no conersation with leinz where you did not get to know
more about leinz than you did about the subject, ormal or otherwise, being discussed. And true to his word my
colleague amongst the participants airly soon terminated the eent. leinz and I came out distressed but happy.
\e walked back to our lotel which the bureaucrats had selected. A ery comortable one. I think it was only the
local participants they didn't like. I became increasingly anxious in case I had oended leinz and wondered whether
to tell him the truth or whether it would be more polite not to do so. 1he truth is, as I outed it in the hotel once we
were inside the oyer, '\ou hae just reincarnated Johnny, Johnny on loerster.' And I alluded, o course, to all
those graduate students, research assistants and so on who participated and were so glad to hae their own iews
understood and authorized.
'Did leinz reply to that' said Graham.
'le hugged me as a ather might do his son. As my ather would hae done to me in a moment o danger or
distress. 1hat hug encompassed all the students, proessors and so on.'
'\hat did ,ov say' asked Graham.
'I burst into tears. I was .o glad to see a demonstration o courage, enough to reitalise the world.'
ACKNOWLLDGLMLN1
I would like to thank Graham Barnes who came rom Sweden to help to interiew me in hospital and Nicholas
Green or discussions reealing a number o pathways well worthy o inestigation also while I was in hospital. I
thank the editor or his helpul criticisms and M. Martinez ,Portugal,, Pablo Naarro ,Spain,, my colleagues at
Montreal, Noolk Virginia and at System Research especially Brian Lewis and Bernard Scott or helping to
germinate this theory.
Systems Research RLSLARCH PAPLR
363 G. Pask
RLILRLNCLS
Barnes, G. ,1996,. Systems Research.
on loerster, l. ,1981,. Observing Systems, Varela, lrancisco ,ed., InterSystems, Seaside, CA.
on loerster, l. ,1995,. Cybernetics of Cybernetics, 2nd edn, luture Systems, Minneapolis, MN.
on loerster, l., and Pask, G. ,1960, A predictie eolutionary model, Cybernetica, 4, 258-300, 1-55. Glanille,
R. ,195,. A cybernetic deelopment o theories o epistemology and obseration, with reerence to space and
time as seen in architecture, unpublished PhD 1hesis, Brunel Uniersity, Depart-ment o Cybernetics.
Glanille, R. ,196,. \hat is memory that it can remember what it is in 1rappl, R., and
Pask, G. ,eds,. Recent Progress in Cybernetics and Systems Research, lemisphere, \ashington, DC, Vol. 2.
Glanille, R. ,1988,. Objekte, Mere, Berlin.
Glanille, R. ,ed., ,1993,, lestchrit or Gordon Pask, Systems Research, 10, (3).
Goguen, J. ,196,. Objects International Journal of General Systems 1, 23.
Jaques, L. ,1964,. Time Span Handbook, leineman, London.
Kauman, L. ,198,. Sel-reerence and recursie orms, Journal of Social and Biological Structures, 10, 53-2.
Kauman, L. ,1995,. Knot set theory, American Cybernetics Society Conerence, Chicago.
Pask, G. ,1980,. 1he importance o being magic, Journal of the American Society for Cybernetics, special edition.
Petri, C.-A. ,1963,. M.L1. thesis on Communicating automata. Also in Bateson, C. ,ed, Our Own Metaphor and
Annatole lolt's comments on C.-A. Petri s outstanding work at Gregory Bateson's Conerence at Borg
\artgenstein.
Rescher, N. ,193,. Coherence Theory of Truth, Oxord Uniersity Press, Oxord.
Robb, l. ,1980,. 1he dynamics o opinion change rom a systems theoretical iewpoint, unpublished PhD
1hesis, Brunel Uniersity, Department o Cybernetics.
Robinson, M. ,19,. luman social groups: a cyber-netic account o stability and instability, unpub-lished PhD
thesis, Brunel Uniersity, Department o Cybernetics.
Schrodinger, L. ,194,. What is Life?, Macmillan, New \ork.
1aylor, C. ,196,. The Explanation of Behaviour, Rout-ledge, Kegan and Paul, London.
1aylor, C. ,193,. Interpretation o the sciences o man, Review Metaphysics 25, 3-51.
\ells, O. ,1965-195,. Artorga, Journal for Research into Artificial Organisms.
on \right, G. ,1963,. Norm and Action, Routledge, Kegan and Paul, London.
RLSLARCH PAPLR Systems Research
The Progenitor of Conversation and Interaction Theories
364
INDLX
, 357, 358, 361
=, 353, 355, 356, 357, 358, 359
=>, 355, 357, 358, 359
A and B, 356, 357, 359, 360, 361
analogy, 360, 361, 362
anticlockwise, 356, 359, 360
AP, 355, 357, 358, 359, 360
application, 356, 357, 359, 360
beginning, 352, 355, 357, 359, 361
bionics, 352
CASTE, 354
clockwise, 356, 358, 359, 360
coherence, 356, 361
complementarity, 355, 359
CON, 356, 357, 358
concepts, 355, 356, 357, 358, 359, 360, 361
CONs, 356
Conversation, 350, 351, 352, 361
conversation theory, 351
conversations, 351, 354, 357
CON
Z
(T), 356
CON
Z
*, 356
CT, 352, 355
cybernetics, 352, 355
Cybernetics, 352, 363
description, 355, 357, 359
dH/dt, 353
difference, 360, 361
differentiation, 361
distinctions, 356, 359, 360, 361
dR/dt, 353
D
Z
(T), 355, 357, 359
end, 351, 355, 357, 359, 361, 362
general systems theory, 352
Glanville, 355, 361, 363
H, 353, 363
Heinz and Mai, 351, 362
hierarchy, 354, 360
Hmax, 353
IA, 352, 355
information science, 352
informationally open, 356, 361
input, 354
INTER, 357, 358
interacting actors, 361
interaction, 351, 352, 353, 354, 355, 361
interaction of actors, 351, 355, 361
kinematic, 355, 359, 361
mentation, 359, 360
metaphor, 360, 361, 362
output, 354
participant, 353, 354, 355, 356, 357, 358,
359, 361, 362
participants, 353, 354, 355, 356, 358, 360,
362
percept, 356, 360
Petri, 353, 355, 361, 363
prepositional operator, 356, 360
PROC, 355, 356, 357
process, 355, 357, 359, 360
product, 355, 356, 359
production, 358, 359
PROG(T), 355, 357
PROGs, 355, 357, 358
psychology, 354, 355
punctuation, 359, 360
q, 356, 360
R, 353, 363
reproduction, 358, 359
SELF ORGANISATION, 352
self organization, 351, 352, 353, 354, 355
set theoretic tokens, 358, 359
spins, 359
system, 353, 354, 356, 361
T, 355, 356, 357, 358, 359, 360
T
f
A, 358
T
f
B, 358
torus, 356, 359
T
s
A, 358
T
s
B, 358
understandings, 351, 359
unfoldment, 359, 360, 361
Varela, 353, 363
vectorial times, 361