Biology of Cognition1] Maturana (1970).pdf · Biology of Cognition Humberto R. Maturana (1970)...

Biology of Cognition

Humberto R. Maturana

(1970)

Biological Computer Laboratory Research Report BCL 9.0. Urbana IL: University of Illinois, 1970. As Reprinted in: Autopoiesisand Cognition: The Realization of the Living Dordecht: D. Reidel Publishing Co., 1980, pp. 5–58.

I. Introduction

Man knows and his capacity to know depends on his bi-ological integrity; furthermore, he knows that he knows.As a basic psychological and, hence, biological func-tion cognition guides his handling of the universe andknowledge gives certainty to his acts; objective knowl-edge seems possible and through objective knowledge theuniverse appears systematic and predictable. Yet knowl-edge as an experience is something personal and privatethat cannot be transferred, and that which one believesto be transferable, objective knowledge, must always becreated by the listener: the listener understands, and ob-jective knowledge appears transferred, only if he is pre-pared to understand. Thus cognition as a biological func-tion is such that the answer to the question, “What is cog-nition?” must arise from understanding knowledge andthe knower through the latter’s capacity to know.

Such is my endeavor.

Epistemology

The basic claim of science is objectivity: it attempts,through the application of a well defined methodology,to make statements about the universe. At the very rootof this claim, however, lies its weakness: the a prioriassumption that objective knowledge constitutes a de-scription of that which is known. Such assumption begsthe questions, “What is it to know?” and “How do weknow?”.

Biology

(a) The greatest hindrance in the understanding of theliving organization lies in the impossibility of account-ing for it by the enumeration of its properties; it must be

understood as a unity. But if the organism is a unity, inwhat sense are its component properties it parts? Theorganismic approach does not answer this question, itmerely restates it by insisting that there are elements oforganization that subordinate each part to the whole andmake the organism a unity [Cf. Bertalanffy, 1960]. Thequestions “How does this unity arise?” and “To whatextent must it be considered a property of the organiza-tion of the organism, as opposed to a property emergingfrom its mode of life?” remain open. A similar difficultyexists for the understanding of the functional organiza-tion of the nervous system, particularly if one considersthe higher functions of man. Enumeration of the transferfunctions of all nerve cells would leave us with a list, butnot with a system capable of abstract thinking, descrip-tion, and self-description. Such an approach would begthe question, “How does the living organization give riseto cognition in general and to self-cognition in particu-lar?”

(b) Organisms are adapted to their environments, and ithas appeared adequate to say of them that their organiza-tion represents the ’environment’ in which they live, andthat through evolution they have accumulated informa-tion about it, coded in their nervous systems. Similarlyit has been said that the sense organs gather informationabout the ’environment’, and through learning this infor-mation is coded in the nervous system [Cf. Young, 1967].Yet this general view begs the questions, “What does itmean to “gather information”?” and “What is coded inthe genetic and nervous systems?”.

A successful theory of cognition would answer both theepistemological and the biological questions. This I pro-pose to do, and the purpose of this essay is to put for-ward a theory of cognition that should provide an episte-mological insight into the phenomenon of cognition, andan adequate view of the functional organization of thecognizant organism that gives rise to such phenomena as

HUMBERTO R. MATURANA 1 Biology of Cognition

conceptual thinking, language, and self-consciousness.

In what follows I shall not offer any formal definitionsfor the various terms used, such as “cognition”, “life”, or“interaction”, but I shall let their meaning appear throughtheir usage. This I shall do because I am confident thatthe internal consistency of the theory will show that theseterms indeed adequately refer to the phenomena I am try-ing to account for, and because I speak as an observer,and the validity of what I say at any moment has its foun-dation in the validity of the whole theory, which, I assert,explains why I can say it. Accordingly, I expect the com-plete work to give foundation to each of its parts, whichthus appear justified only in the perspective of the whole.

Note: I shall be speaking of the organism as a unity, butwhen I wrote this essay I was not aware that the wordunit did not always quite mean unity. Since I cannot nowcorrect this. I beg the reader to bear this in mind.

II. The Problem

(1) Cognition is a biological phenomenon and can onlybe understood as such; any epistemological insight intothe domain of knowledge requires this understanding.

(2) If such an insight is to be attained, two questionsmust be considered:

What is cognition as a function?

What is cognition as a process?

III. Cognitive Function in General

The Observer

(1) Anything said is said by an observer. In his discoursethe observer speaks to another observer, who could behimself; whatever applies to the one applies to the otheras well. The observer is a human being, that is, a livingsystem, and whatever applies to living systems appliesalso to him.

(2) The observer beholds simultaneously the entity thathe considers (an organism, in our case) and the universein which it lies (the organism’s environment). This al-lows him to interact independently with both and to haveinteractions that are necessarily outside the domain of in-teractions of the observed entity.

(3) It is an attribute of the observer to be able to inter-act independently with the observed entity and with its

relations; for him both are units of interaction (entities).

(4) For the observer an entity is an entity when he candescribe it. To describe is to enumerate the actual or po-tential interactions and relations of the described entity.Accordingly, the observer can describe an entity only ifthere is at least one other entity from which he can dis-tinguish it and with which he can observe it to interact orrelate. This second entity that serves as a reference forthe description can be any entity, but the ultimate refer-ence for any description is the observer himself.

(5) The set of all interactions into which an entity canenter is its domain of interactions. The set of all relations(interactions through the observer) in which an entity canbe observed is its domain of relations. This latter domainlies within the cognitive domain of the observer. An en-tity is an entity if it has a domain of interactions, and ifthis domain includes interactions with the observer whocan specify for it a domain of relations. The observer candefine an entity by specifying its domain of interactions;thus part of an entity, a group of entities, or their rela-tions, can be made units of interactions (entities) by theobserver.

(6) The observer can define himself as an entity by spec-ifying his own domain of interactions; he can always re-main an observer of these interactions, which he can treatas independent entities.

(7) The observer is a living system and an understandingof cognition as a biological phenomenon must accountfor the observer and his role in it.

The Living System

(1) Living systems are units of interactions; they existin an ambience. From a purely biological point of viewthey cannot be understood independently of that part ofthe ambience with which they interact: the niche; nor canthe niche be defined independently of the living systemthat specifies it.

(2) Living systems as they exist on earth today are char-acterized by exergonic metabolism, growth and internalmolecular reproduction, all organized in a closed causalcircular process that allows for evolutionary change inthe way the circularity is maintained, but not for the lossof the circularity itself. Exergonic metabolism is requiredto provide energy for the endergonic synthesis of spe-cific polymers (proteins, nucleic acids, lipids, polysac-charides) from the corresponding monomers, that is, forgrowth and replication; special replication procedures


secure that the polymers synthesized be specific, thatthey should have the monomer sequence proper to theirclass; specific polymers (enzymes) are required for theexergonic metabolism and the synthesis of specific poly-mers (proteins, nucleic acids, lipids, polysaccharides)[Cf. Commoner, 19651.]

This circular organization constitutes a homeostatic sys-tem whose function is to produce and maintain this verysame circular organization by determining that the com-ponents that specify it be those whose synthesis or main-tenance it secures. Furthermore, this circular organiza-tion defines a living system as a unit of interactions andis essential for its maintenance as a unit; that which isnot in it is external to it or does not exist. The circu-lar organization in which the components that specify itare those whose synthesis or maintenance it secures in amanner such that the product of their functioning is thesame functioning organization that produces them, is theliving organization.

(3) It is the circularity of its organization that makes aliving system a unit of interactions, and it is this circu-larity that it must maintain in order to remain a livingsystem and to retain its identity through different inter-actions. All the peculiar aspects of the different kindsof organisms are superimposed on this basic circularityand are subsequent to it, securing its continuance throughsuccessive interactions in an always changing environ-ment. A living system defines through its organizationthe domain of all interactions into which it can possiblyenter without losing its identity, and it maintains its iden-tity only as long as the basic circularity that defines itas a unit of interactions remains unbroken. Strictly, theidentity of a unit of interactions that otherwise changescontinuously is maintained only with respect to the ob-server, for whom its character as a unit of interactionsremains unchanged.

(4) Due to the circular nature of its organization a liv-ing system has a self-referring domain of interactions (itis a self-referring system), and its condition of being aunit of interactions is maintained because its organiza-tion has functional significance only in relation to themaintenance of its circularity and defines its domain ofinteractions accordingly.

(5) Living systems as units of interactions specified bytheir condition of being living systems cannot enter intointeractions that are not specified by their organization.The circularity of their organization continuously bringsthem back to the same internal state (same with respectto the cyclic process). Each internal state requires thatcertain conditions (interactions with the environment be

satisfied in order to proceed to the next state. Thus, thecircular organization implies the prediction that an in-teraction that took place once will take place again. Ifthis does not happen the system disintegrates; if the pre-dicted interaction does take place, the system maintainsits integrity (identity with respect to the observer) andenters into a new prediction. In a continuously chang-ing environment these predictions can only be success-ful if the environment does not change in that which ispredicted. Accordingly, the predictions implied in theorganization of the living system are not predictions ofparticular events, but of classes of interactions. Every in-teraction is a particular interaction, but every predictionis a prediction of a class of interactions that is defined bythose features of its elements that will allow the livingsystem to retain its circular organization after the inter-action, and thus, to interact again. This makes living sys-tems inferential systems, and their domain of interactionsa cognitive domain.

(6) The niche is defined by the classes of interactionsinto which an organism cam enter. The environment isdefined by the classes of interactions into which the ob-server can enter and which he treats as a context for hisinteractions with the observed organism. The observerbeholds organism and environment simultaneously andhe considers as the niche of the organism that part of theenvironment which he observes to lie in its domain ofinteractions. Accordingly, as for the observer the nicheappears as part of the environment, for the observed or-ganism the niche constitutes its entire domain of interac-tions, and as such it cannot be part of the environmentthat lies exclusively in the cognitive domain of the ob-server. Niche and environment, then, intersect only to theextent that the observer (including instruments) and theorganism have comparable organizations, but even thenthere are always parts of the environment that lie beyondany possibility of intersection with the domain of inter-actions of the organism, and there are parts of the nichethat lie beyond any possibility of intersection with thedomain of interactions of the observer. Thus for everyliving system its organization implies a prediction of aniche, and the niche thus predicted as a domain of classesof interactions constitutes its entire cognitive reality. Ifan organism interacts in a manner not prescribed by itsorganization, it does so as something different from theunit of interactions defined by its basic circularity, andthis interaction remains outside its cognitive domain, al-though it may well lie within the cognitive domain of theobserver.

(7) Every unit of interactions can participate in interac-tions relevant to other, more encompassing units of in-


teractions. If in doing this a living system does not loseits identity, its niche may evolve to be contained by thelarger unit of interactions and thus be subservient to it. Ifthis larger unit of interactions is (or becomes) in turn alsoa self-referring system in which its components (them-selves self-referring systems) are subservient to its main-tenance as a unit of interactions, then it must itself be (orbecome) subservient to the maintenance of the circularorganization of its components. Thus, a particular self-referring system may have the circular organization of aliving system or partake functionally of the circular orga-nization of its components, or both. The society of bees(the honey producing bees) is an example of a third orderself-referring system of this kind; it has a circular orga-nization superimposed on the second order self-referringsystems that are the bees, which in turn have a circular or-ganization superimposed on the first order living systemsthat are the cells; all three systems with their domains ofinteractions are subordinated both to the maintenance ofthemselves and to the maintenance of the others.

Evolution

(1) Evolutionary change in living systems is the result ofthat aspect of their circular organization which securesthe maintenance of their basic circularity, allowing ineach reproductive step for changes in the way this circu-larity is maintained. Reproduction and evolution are notessential for the living organization, but they have beenessential for the historical transformation of the cognitivedomains of the living systems on earth.

(2) For a change to occur in the domain of interactionsof a unit of interactions without its losing its identitywith respect to the observer it must suffer an internalchange. Conversely, if an internal change occurs in aunit of interactions, without its losing its identity, its do-main of interactions changes. A living system suffersan internal change without loss of identity if the predic-tions brought forth by the internal change are predictionswhich do not interfere with its fundamental circular or-ganization. A system changes only if its domain of inter-actions changes.

(3) After reproduction the new unit of interactions hasthe same domain of interactions as the parental one onlyif it has the same organization. Conversely, the new unitof interactions has a different domain of interactions onlyif its organization is different, and hence, implies differ-ent predictions about the niche.

(4) Predictions about the niche are inferences about

classes of interactions. Consequently, particular interac-tions which are indistinguishable for an organism maybe different for an observer if he has a different cognitivedomain and can describe them as different elements of aclass defined by the conduct of the organism. The sameapplies to interactions that are identical for the organ-ism but different for (have different effects) its differentinternal parts. Such interactions may result in differentmodifications of the internal states of the organism and,hence, determine different paths of change in its domainof interactions without loss of identity. These changesmay bring about the production of offspring having do-mains of interactions different from the parental ones. Ifthis is the case and a new system thus produced predictsa niche that cannot be actualized, it disintegrates; other-wise it maintains its identity and a new cycle begins.

(5) What changes from generation to generation in theevolution of living systems are those aspects of their or-ganization which are subservient to the maintenance oftheir basic circularity but do not determine it, and whichallow them to retain their identity through interactions;that is, what changes is the way in which the basic circu-larity is maintained, and not this basic circularity in itself.The manner in which a living system is compounded asa unit of interactions, whether by a single basic unit, orthrough the aggregation of numerous such units (them-selves living systems) that together constitute a largerone (multicellular organisms), or still through the aggre-gation of their compound units that form self-referringsystems of even higher order (insect societies, nations) isof no significance; what evolves is always a unit of inter-actions defined by the way in which it maintains its iden-tity. The evolution of the living systems is the evolutionof the niches of the units of interactions defined by theirself-referring circular organization, hence, the evolutionof the cognitive domains.

The Cognitive Process

(1) A cognitive system is a system whose organizationdefines a domain of interactions in which it can act withrelevance to the maintenance of itself, and the process ofcognition is the actual (inductive) acting or behaving inthis domain. Living systems are cognitive systems, andliving as a process is a process of cognition. This state-ment is valid for all organisms, with and without a ner-vous system.

(2) If a living system enters into a cognitive interaction,its internal state is changed in a manner relevant to itsmaintenance, and it enters into a new interaction with-


out loss of its identity. In an organism without a nervoussystem (or its functional equivalent) its interactions areof a chemical or physical nature (a molecule is absorbedand an enzymatic process is initiated; a photon is cap-tured and a step in photosynthesis is carried out). Forsuch an organism the relations holding between the phys-ical events remain outside its domain of interactions. Thenervous system enlarges the domain of interactions of theorganism by making its internal states also modifiable ina relevant manner by “pure relations”, not only by physi-cal events; the observer sees that the sensors of an animal(say, a cat) are modified by light, and that the animal (thecat) is modified by a visible entity (say, a bird). The sen-sors change through physical interactions: the absorptionof light quanta; the animal is modified through its inter-actions with the relations that hold between the activatedsensors that absorbed the light quanta at the sensory sur-face. The nervous system expands the cognitive domainof the living system by making possible interactions with“pure relations”; it does not create cognition.

(3) Although the nervous system expands the domain ofinteractions of the organism by bringing into this domaininteractions with “pure relations”, the function of the ner-vous system is subservient to the necessary circularity ofthe living organization.

(4) The nervous system, by expanding the domain of in-teractions of the organism, has transformed the unit of in-teractions and has subjected acting and interacting in thedomain of “pure relations” to the process of evolution.As a consequence, there are organisms that include asa subset of their possible interactions, interactions withtheir own internal states (as states resulting from externaland internal interactions) as if they were independent en-tities, generating the apparent paradox of including theircognitive domain within their cognitive domain. In usthis paradox is resolved by what we call “abstract think-ing”, another expansion of the cognitive domain.

(5) Furthermore, the expansion of the cognitive domaininto the domain of “pure relations” by means of a ner-vous system allows for non-physical interactions be-tween organisms such that the interacting organisms ori-ent each other toward interactions within their respectivecognitive domains. Herein lies the basis for communi-cation: the orienting behavior becomes a representationof the interactions toward which it orients, and a unitof interactions in its own terms. But this very processgenerates another apparent paradox: there are organismsthat generate representations of their own interactions byspecifying entities with which they interact as if they be-longed to an independent domain, while as representa-

tions they only map their own interactions. In us thisparadox is resolved simultaneously in two ways:

(a) We become observers through recursively gener-ating representations of our interactions, and byinteracting with several representations simultane-ously we generate relations with the representa-tions of which we can then interact and repeat thisprocess recursively, thus remaining in a domain ofinteractions always larger than that of the repre-sentations.

(b) We become self-conscious through self-observation; by making descriptions of ourselves(representations), and by interacting with our de-scriptions we can describe ourselves describingourselves, in an endless recursive process.

IV. Cognitive Function in Particular

Nerve Cells

(1) The neuron is the anatomical unit of the nervous sys-tem because it is a cell, and as such it is an independentintegrated self-referring metabolic and genetic unit (a liv-ing system indeed).

(2) Anatomically and functionally a neuron is formed bya collector area (dendrites, and in some cases, also thecell body and part of the axon) united via a distributiveelement (the axon, and in some cases, also the cell bodyand main dendrites), capable of conducting propagatedspikes to an effector area formed by the terminal branch-ing of the axon. The functional state of the collector areadepends on both its internal state (reference state) and onthe state of activity of the effector areas synapsing on it.Correspondingly, the state of activity of the effector areadepends on both the train of impulses generated at thecorresponding collector area and on the pre-synaptic andnon-synaptic interactions with distributive elements andother effector areas that may take place in the neuropiland in the immediate vicinity of the next collector areas.This is true even in the case of amacrine cells, in whichthe collector and effector areas may be intermingled. Thedistributive element determines where the effector exertsits influence.

(3) Whether one or two branches of a bifurcating axonare invaded by a nerve impulse propagating along it de-pends on their relative diameter and on the state of po-larization of their membranes at their origin in the bifur-cation zone. As a result, the pattern of effector activity,


that is, the pattern of branch invasion which a train of im-pulses determines in the branches of the distribution ele-ment and effector area of a neuron, depends on the spikeinternal distribution of the train of impulses, which deter-mines the time that the axonal membrane at the branch-ing zone has for recovery before the arrival of the nextspike, and (ii) on the non synaptic influences which, inthe form of local water and ion movements caused bythe electric activity of neighboring elements, may pro-duce diameter and polarization changes at the branchingzones, and thus modify the invisibility of the branches bythe arriving spikes.

(4) At any moment the state of activity of a nerve cell, asrepresented by the pattern of impulses traveling along itsdistributive element, is a function of the spatio-temporalconfiguration of its input, as determined by the relativeactivity holding between the afferent neurons, that mod-ulates the reference state proper of the collector area. Itis known that in many neurons the recurrence of a givenafferent spatio-temporal configuration results in the re-currence of the same state of activity, independently ofthe way in which such a spatio-temporal configurationis generated [Cf. Maturana and Frenk, 1963; Morrell,1967]. [This is so in the understanding that two states ofactivity in a given cell are the “same” (equivalent) if theybelong to the same class, as defined by the pattern of im-pulses that they generate, and not because they are a one-to-one mapping of each other.] Also, the spatio-temporalconfiguration of the input to a neuron that causes in itthe recurrence of a given state of activity is a class ofafferent influences defined by a pattern in the relationsholding between the active afferents and the collector; agiven class of responses is elicited by a given class ofafferent influences.

(5) For every nerve cell, at any moment, its transferfunction at its collector area is a well-defined determin-istic process [Cf. Segundo and Perkel, 1969]. Many neu-rons have several transfer functions, and different classesof afferent influences change their activity differently,causing them to generate different classes of activity intheir effector areas. Because every nerve cell participatesin the generation of the spatio-temporal configuration ofafferent influences on the other nerve cells, all their statesof activity must be considered as significant for their nextstates of activity. Thus there are two aspects to considerwith respect to the activity of any given neuron: its gene-sis, which must be considered in reference to the neuronitself and to the afferents to it; (ii) its participation in thegeneration of activity in other neurons for which it is anafferent influence, which must be considered in referenceto those other neurons. In both cases the interactions be-

tween the neurons involved are strictly deterministic, al-though what is cause in one is not necessarily cause inanother.

(6) The nerve impulses that travel along the distribu-tive element originate at the point where this elementemerges from the collector area. Each nerve impulse isthe result of the state of excitation of the collector areaat a given moment (as determined by the spatio-temporalconfiguration of the afferent excitatory and inhibitory in-fluences acting upon it, and on its own internal gener-ating mechanisms, if any) that spreads reaching a giventhreshold at the point of emergence of the distributor. Ex-citatory and inhibitory influences, however, do not super-impose linearly; their relative participation in determin-ing the production of nerve impulses, and hence, the stateof activity of the neuron, depends on their relative spatialdistribution on the collector area. Inhibition works byshunting off the spreading excitatory processes; as a re-sult the relative contributions of a point of excitation anda point of inhibition in the generation of a nerve impulsedepend on where, on the collector, they stand with re-spect to each other and with respect to the point of emer-gence of the distributive element. Excitation and inhibi-tion must be seen as integral parts in the definition of thespatio-temporal configuration of afferent influences, notas independent processes. The shape of the collector area(its geometry) determines the class or classes of spatio-temporal configurations of afferent influences to whichthe cell responds.

(7) The neuropil is the site where the distributive ele-ments and effector areas of many different neurons in-termingle with each other and with the collector areasof the post-synaptic cells. Here non-synaptic interac-tions take place between neighboring elements whichmay cause in each other, as a result of the local move-ments of water and ions produced by their independentelectrical activity, changes in diameter and polarizationat their branching points. Depending on the time con-stant of these local changes, and on the capacity of theaxons to homeostatically maintain their diameter at thenew values, the pattern of branch invasion produced in agiven effector area by a given train of impulses may bemodified in a more or less permanent manner by thesenon-synaptic interactions. Something singular may hap-pen during synaptic concomitances at the collector areasif synapses also affect each other non-synaptically, due totheir spatial contiguity, causing each other more or lesspermanent changes in size (increase or decrease) and po-larization (with the corresponding changes in effective-ness) as a result of their independent electrical activities.Thus, the neuropil may have to be seen as constituting


a plastic system through which acquired self-addressingstates of activity attain their functional significance asthey become specified by the non-synaptic and synap-tic concomitances generated by the interactions of theorganism. It is not the repetition of the same state ofactivity which can cause neuronal changes of behavioralsignificance subordinated to the evolving domain of in-teractions of an organism, but rather it is the occurrenceof local concomitant states of activity produced by seem-ingly unrelated interactions which can cause such subor-dinated changes in the reactive capacity of neurons.

(8) It follows that one should expect in a significantnumber of neurons, which may vary in different classesof animals according to the organization of their differentneuropils, a continuous change in their transfer functions(from collector to effector area), or in the circumstancesunder which they are activated, as a result of the pasthistory of the organism. However, for the understandingof the functional organization of the nervous system it isnecessary to consider that nerve cells respond at any mo-ment with definite transfer functions to classes of afferentspatio-temporal configurations in their input, generatingdefinite states of effector activity, and not to particularafferent states. Furthermore:

(a) Any interaction is represented in the nervous sys-tem by the sequence of states of relative neuronalactivity leading to the conduct which it generates;this conduct should be repeatable to the extent thatthe interaction (sequence of states of relative ac-tivity) is reproducible, that is, as long as the histor-ical transformation of the nervous system (learn-ing) does not make it impossible.

(b) The nervous system always functions in thepresent, and it can only be understood as a systemfunctioning in the present. The present is the timeinterval necessary for an interaction to take place;past, future and time exist only for the observer.Although many nerve cells may change continu-ously, their mode of operation and their past his-tory can explain to the observer how their presentmode of operation was reached, but not how it isrealized now, or what their present participation inthe determination of behavior is.

(c) Any behavior is defined through a sequence ofstates in the receptor surfaces (external and inter-nal) that satisfy its direct or indirect subordinationto the maintenance of the basic circularity of theliving system. Since the nervous system is contin-uously changing through experience, what occurswhen the observer sees a given behavior reenacted

is a sequence of interactions that satisfy this sub-ordination independently of the neuronal processwhich generated them. The more complex the do-main of interactions of an organism, the more in-direct is this subordination (an adequate mode ofbehavior subordinated to another), but not the lessstrict.

(d) An organism is a unit to the extent that its conductresults in the maintenance of its basic circularity(and hence identity), and two modes of conductare equivalent if they satisfy the same class of re-quirements for this maintenance. For this reasonan organism, as a self-regulated homeostatic orga-nization, does not require a constant behavior inits deterministic component elements (in this case,neurons) if their changes become specified throughthe generation of conduct, and sameness of con-duct is defined with respect to an observer or afunction that must be satisfied.

Thus although at any moment every neuron functions de-terministically with a definite transfer function, and gen-erates a definite pattern of activity in its effector area, thetransfer functions and the patterns of effector activity inmany of them may change from one moment to anotherand the organism still will give rise to what the observerwould call “the same behavior”. The converse is also thecase, and through what the observer would call “differentbehaviors” the organism may satisfy its subordination tothe same aspect of the maintenance of its basic circular-ity.

(9) From these notions it is apparent that the neuron can-not be considered as the functional unit of the nervoussystem; no neuron can have a fixed functional role in thegeneration of conduct if it must be continuously chang-ing its participation in it. For the same reason a fixedcollection of cells also cannot be considered as a func-tional unit of the nervous system. Only conduct itselfcan be considered as the functional unit of the nervoussystem.

(10) If nerve cells respond to classes of afferent config-urations and not to particular afferent states, they mustnecessarily treat as equivalent particular afferent config-urations that arise through interactions which for the ob-server are otherwise unrelated.

Architecture

(1) In any given nervous system the great majority (andperhaps the totality) of its neurons can be assigned to


well-defined morphological classes, each characterizedby a given pattern of distribution of the collector and ef-fector areas of its elements. As a result, the elements ofthe same class hold similar relations with each other andwith other classes of neurons; the shapes of the nervecells (collector area, distributive element, and effectorarea) specify their connectivity. These shapes are genet-ically determined and have been attained through evo-lution; the whole architecture of the brain is geneticallydetermined and has been attained through evolution. Thefollowing implications are significant for the understand-ing of the nervous system:

(a) There is a necessary genetic variability in theshape of nerve cells as well as a variability that re-sults from interactions of the organism with inde-pendent events during its development. The func-tional organization of the nervous system must besuch as to tolerate this double variability.

(b) Due to the genetic and somatic variability no twonervous systems of animals of the same species(particularly if they have many cells) are identi-cal, and they resemble each other only to the ex-tent that they are organized according to the samegeneral pattern. It is the organization defining theclass, and not any particular connectivity, whichdetermines the mode of functioning of any givenkind of nervous system.

(2) The shapes of nerve cells and their packing are suchthat there is in general a great overlapping in the collectorand effector areas of neurons of the same class. Also, thespatial distribution and the interconnections between dif-ferent classes of neurons is such that any particular partof the nervous system is in general simultaneously re-lated to many other parts; the parts interconnected, how-ever, differ in different species, and as a result these havedifferent interacting capabilities.

(3) The organism ends at the boundary that its self-referring organization defines in the maintenance of itsidentity. At this boundary there are sensors (the sen-sory surfaces) through which the organism interacts inthe domain of relations and effectors (the effector sur-faces) through which the nervous system modifies theposture of the organism in this domain. The sensorysurfaces are in general constituted by collections of sen-sory elements (cells) with similar, though not identical,properties (classes of properties) which in their mode ofinteraction with the nervous system share the character-istics of neurons in general. As a result whenever theorganism enters into an interaction within the physical

domain of interactions of the sensors, as a rule not onebut many sensory elements are excited. The effectors arealso multifarious and differ from each other in the man-ner in which they change the receptor surfaces of the or-ganism during the interactions: action always leads to achange in the state of activity of the receptor surfaces.

(4) The architectural organization of the nervous systemis subordinated to the order of the sensory and effectorsurfaces. This subordination has two aspects: the recep-tor and effector surfaces project to the central nervoussystem retaining their proper topological relations; (ii)the topological relations specified by the receptor andeffector surfaces in their projection constitute the basisfor all the architectural order of the central nervous sys-tem. As a consequence, this architectural organizationconstitutes a system that interconnects these surfaces ina manner that permits the occurrence of certain concomi-tances of activity and not others in the different neu-ropils, and thus secures well-defined functional relationsbetween these surfaces, specifying how they modify eachother. Truism: the nervous system cannot give rise to aconduct that implies the concomitance of states of activ-ity for which there is no anatomical basis. As a resultof its architectural organization every point in the cen-tral nervous system constitutes an anatomical localiza-tion with respect to the possibility of establishing cer-tain functional concomitances. From this it follows thatany localized lesion in the nervous system must neces-sarily interfere in a localized manner with the possibilityof synthesizing some specific conduct (state of neural ac-tivity).

Function

(1) The way the nervous system functions is bound to itsanatomical organization. The functioning of the nervoussystem has two aspects: one which refers to the domainof interactions defined by the nervous system (relationsin general); the other which refers to the particular partof that domain used by a given species (particular classesof relations): Different species interact with different setsof relations (have different niches).

(2) The nervous system only interacts with relations.However, since the functioning of the nervous system isanatomy bound, these interactions are necessarily medi-ated by physical interactions; for an animal to discrim-inate objects visually the receptors in its eyes must ab-sorb light quanta and be activated; yet, the objects thatthe animal sees are determined not by the quantity oflight absorbed, but by the relations holding between the


receptor-induced states of activity within the retina, ina manner determined by the connectivity of its varioustypes of cells. Therefore, the nervous system definesthrough the relative weights of the patterns of interac-tions of its various components, both innate and acquiredthrough experience, which relations will modify it at anygiven interaction [Cf. Maturana, 1965]. Or, in general,the organization and structure of a living system (its ner-vous system included) define in it a “point of view”, abias or posture from the perspective of which it interactsdetermining at any instant the possible relations accessi-ble to its nervous system. Moreover, since the domainof interactions of the organism is defined by its structure,and since this structure implies a prediction of a niche,the relations with which the nervous system interacts aredefined by this prediction and arise in the domain of in-teractions of the organism.

(3) Due to the properties of neurons, and due to the ar-chitecture of the nervous system, interactions within thenervous system give rise to activity in aggregates of cells.Also, for the same reasons, any given cell may assumethe same state of activity under many different circum-stances of interactions of the organism. Thus, underno circumstances is it possible to associate the activityof any particular cell with any particular interaction ofthe living system. When any particular interaction takesplace at the level of the sensors, the relations accessibleto the nervous system are given at this level in a certainstate of relative activity of the sensing elements and notin the state of activity of any particular one [Cf. Mat-urana, Uribe, and Frenk, 1968]. At the same time, al-though operational localizations can be established in thenervous system [Cf. Geschwind, 1965], these localiza-tions are to be viewed in terms of areas where certainmodalities of interactions converge, and not as localiza-tions of faculties or functions. As a result of the modeof organization of the nervous system that I have em-phasized, localized lesions should produce discrete func-tional deficiencies by impeding the convergence of ac-tivities necessary for the synthesis of a particular con-duct (state of activity). The anatomical and functionalorganization of the nervous system secures the synthesisof behavior, not a representation of the world; hence, itis only with the synthesis of behavior that one cam in-terfere. The nervous system is localized in terms of theorganism’s surfaces of interaction but not in terms of rep-resentations of the interactions it can generate.

Representation

(1) The fundamental anatomical and functional organi-zation of the nervous system is basically uniform; thesame functions and operations (excitation, inhibition, lat-eral interaction, recursive inhibition, etc.) are performedin its various parts, although in different contexts, andintegrated in different manners. A partial destruction ofthe nervous system does not alter this basic uniformity,and, although the parts left untouched cannot do the samethings that the whole did, they appear in their mode ofoperations identical to the untouched whole. To the ob-server, once the boundary of the sensors is passed, thenervous system, as a mode of organization, seems to be-gin at any arbitrary point that he may choose to consider;the answer to the question, “What is the input to the ner-vous system?” depends entirely on the chosen point ofobservation. This basic uniformity of organization canbest be expressed by saying: all that is accessible to thenervous system at any point are states of relative activityholding between nerve cells, and all that to which anygiven state of relative activity can give rise are furtherstates of relative activity in other nerve cells by formingthose states of relative activity to which they respond.The effector neurons are not an exception to this sincethey, by causing an effector activity and generating aninteraction, cause a change in the state of relative activ-ity of the receptor elements at the receptor surfaces. Thishas a fundamental consequence: unless they imply theirorigin (through concomitant events, their locations, orthrough the consequences of the new interactions whichthey originate) there is no possible distinction betweeninternally and externally generated states of nervous ac-tivity.

(2) The relations with which the nervous system inter-acts are relations given by the physical interactions ofthe organism, and, hence, depend on its anatomical orga-nization. For the observer the organism interacts with agiven entity that he can describe in his cognitive domain.Yet, what modifies the nervous system of the observedorganism are the changes in activity of the nerve cells as-sociated with the sensing elements, changes that hence-forth constitute an embodiment of the relations that arisethrough the interaction. These relations are not those thatthe observer can describe as holding between componentproperties of the entity in his cognitive domain; they arerelations generated in the interaction itself and depend onboth the structural organization of the organism and theproperties of the universe that match the domain of inter-actions that this organization defines. Whenever such arelation recurs at the sensory surface, the same state of


relative activity arises among the neurons in contact withthe sensing elements. Two interactions that produce thesame state of relative activity are identical for the ner-vous system, no matter how different they may be in thecognitive domain of the observer.

(3) Every relation is embodied in a state of relative activ-ity of nerve cells, but also every state of relative activityacts to modify the relative activity of other nerve cells.Thus, relations through their embodiment in states of rel-ative activity become units of internal interactions andgenerate additional relations, again embodied in states ofrelative activity which in turn may also become units ofinternal interactions, and so on, recursively.

(4) If an external interaction takes place, the state of ac-tivity of the nervous system is modified by the changein relative activity of the neurons, which in close asso-ciation with the sensing elements embody the relationsgiven in the interaction. Accordingly, that which the dif-ferent states of activity thus generated can be said to rep-resent are the relations given at the sensory surfaces bythe interaction of the organism, and not an independentmedium, least of all a description of an environment nec-essarily made in terms of entities that lie exclusively inthe cognitive domain of the observer.

If an internal interaction takes place, the state of activ-ity of the nervous system is modified by one of its ownsubstates of relative activity that embodies one set of re-lations. However, that which the new state of relativeactivity represents is the relations given in the internal in-teraction and not an independent set of relations or theirdescription, in terms of some kind of entities, such asthoughts, that lie only within the cognitive domain of theobserver.

(5) The classes of relations that can be embodied havebeen defined through the evolution of the general struc-tural organization of the organism, and particularly, ofthe sensors, that has defined the classes of relation thatare accessible to the nervous system; and (ii) through theevolution of a particular organization of the nervous sys-tem that defines for each class of animals (species) thespecific mode of how these relations generate a behaviorrelevant to their maintenance.

(6) For any class of relations, the particular relationsgiven as a result of a present interaction are embodiedin a set of particular states of activity occurring in thepresent. This is the case independently of the history ofthe system. However, the relevance of the behavior gen-erated by those states of activity for the maintenance ofthe living system is a function of history, and may depend

both on the evolutionary history of the species and on thepast experiences of the organism as an individual. In thefirst case I would speak of instinctive behavior, and inthe second case of learned behavior. The description oflearning in terms of past and present behavior lies in thecognitive domain of the observer; the organism alwaysbehaves in the present. The observer, however, by inter-acting with descriptions that he generates can treat inter-actions which do not recur as if they were in the present.This apparent paradox is resolved by generating the no-tion of time, past, present, and future, as a new expansionof the domain of interactions. Whenever an interactiontakes place which is an element of a class experiencedfor the first time, it is sufficient that the state of activ-ity which it generates be followed by the suppression ofa peculiar concomitant internal state of activity (that isapparent in what the observer calls the emotion of anxi-ety or uncertainty) for the organism to experience the re-currence of an interaction of the same class, which takesplace without such a concomitant state, as not new (in thesense that it can generate an established conduct as is ap-parent in the absence of anxiety) and, hence known. Anyexperience without anxiety can be described as known,and thus serve as a basis for the functional notion of time.

(7) There is no difference in the nature of the embodi-ment of the relations generated through either external orinternal interactions; both are sets of states of neuronalactivity that can be said to represent the interactions. Ina nervous system capable of interacting with some of itsown internal states as if they were independent entities,there are two consequences:

(a) The distinction between externally and internallygenerated inter-actions can only arise through aconcomitance of events that indicates the source(sensory surface or not) of the state of activitycaused by them, or through the outcome of newinteractions which they initiate. A nervous systemthat is capable of treating its internally generatedstates of activity as different from its externallygenerated states, that is, of distinguishing their ori-gin, is capable of abstract thinking.

(b) The nervous system can interact with the represen-tations of its interactions (and hence, of the organ-ism) in an endless recursive manner.

(8) Four comments:

(a) Notions such as embodiment of representation ex-press the correspondence that the observer sees be-tween relations, or sets of relations, and differ-


ent states of activity of the nervous system, and,as such, lie in his cognitive domain. They de-scribe the functional organization of the nervoussystem in the cognitive domain of the observer,and point to the ability of the nervous system totreat some of its own states as independent entitieswith which it can interact, but they do not charac-terize the nature of the functional subordination ofthe nervous system to its own states. This subor-dination is that of a functionally closed, state de-termined, ultrastable system, modulated by inter-actions [Cf. Ashby, 1960].

(b) The closed nature of the functional organization ofthe nervous system is a consequence of the self-referring domain of interactions of the living or-ganization; every change of state of the organismmust bring forth another change of state, and soon, recursively, always maintaining its basic cir-cularity. Anatomically and functionally the ner-vous system is organized to maintain constant cer-tain relations between the receptor and effectorsurfaces of the organism, which can only in thatway retain its identity as it moves through its do-main of interactions. Thus all conduct, as con-trolled through the nervous system, must (neces-sarily, due to the latter’s architectural organiza-tion) lead through changes in the effector surfacesto specific changes in the receptor surfaces that inturn must generate changes in the effector surfacesthat again . . . and so on, recursively. Conduct isthus a functional continuum that gives unity to thelife of the organism through its transformations inthe latter’s self-referring domain of interactions.The evolutionary subordination of the architectureof the central nervous system to the topology ofthe sensory and effector surfaces appears as an ob-vious necessity.

(c) The ability of the nervous system to interact withits own internal states, as if these were indepen-dent entities, enters these internal states as modu-lating factors in the continuum of behavior. Thisrequires an anatomical and functional internal re-flection so that the internal organization of the ner-vous system can project itself onto itself retainingits morphological and functional topological rela-tions, as the receptor and effector surfaces do intheir own projection. This seems to have acquiredan autonomous evolutionary course with the de-velopment of the neo-cortex in mammals, whicharises as a center of internal anatomical projection,and whose evolution in this line is accompanied

by an increased dependency of the organism on itsown states of nervous activity.

(d) The closed nature of the functional organizationof the nervous system (open only to modulationsthrough interactions) is particularly evident in sys-tematic observations that explicitly show the sub-ordination of conduct to the correlation of ac-tivity between the receptor and effector surfaces[Cf. Held and Hein, 1963] . Experiments such asthose of Held and Hein show that a cat does notlearn to control its environment visually if raisedin darkness and carried about only passively, byanother cat, when under light. From these obser-vations, it is apparent that the “visual handling”of an environment is no handling of an environ-ment, but the establishment of a set of correlationsbetween effector (muscular) and receptor (propri-oceptor and visual) surfaces, such that a particularstate in the receptor surfaces may cause a partic-ular state in the effector surfaces that brings fortha new state in the receptor surfaces . . . and so on.Behavior is like an instrumental flight in which theeffectors (engines, flaps, etc.) vary their state tomaintain constant, or to change, the readings ofthe sensing instruments according to a specifiedsequence of variations, which either is fixed (spec-ified through evolution) or can be varied during theflight as a result of the state of the flight (learning).The same is apparent in the experiments with in-nate perception of depth [Cf. Gibson, 1950] thatshow that there is an innate system of correlationsbetween certain states of the receptor and effectorsurfaces. The reference to a pre-established per-ception of depth is a description that lies in thecognitive domain of the observer, and as such onlyalludes to relations, through the observer, betweenelements that lie in his cognitive domain; but, asa process, this innate behavior obviously corre-sponds to one of optimization of sensory states.

Description

(1) A living system, due to its circular organization, isan inductive system and functions always in a predictivemanner: what happened once will occur again. Its orga-nization, (genetic and otherwise) is conservative and re-peats only that which works. For this same reason livingsystems are historical systems; the relevance of a givenconduct or mode of behavior is always determined in thepast. The goal state (in the language of the observer)that controls the development of an organism is, except


for mutations, determined by the genome of the parentorganism. The same is true for behavior in general; thepresent state is always specified from the previous statethat restricts the field of possible modulations by inde-pendent concomitances. If a given state of relative activ-ity in the nerve cells originates a given behavior, a recur-rence of the “same state” of relative activity should giverise to the “same behavior” no matter how the recurrenceoriginates. The relevance of such a behavior is deter-mined by the significance that it has for the maintenanceof the living organization, and it is in relation to this rel-evance that any subsequent behaviors are the same. Withthe expansion of the cognitive domain during evolution,the types of behavior have changed as well as how theirrelevance is implemented; different kinds of behavior arerelevant to the maintenance of the basic circularity of theliving organization through different domains of interac-tions, and hence, different fields of causal relations.

(2) Since the niche of an organism is the set of all classesof interactions into which it can enter, and the observerbeholds the organism in an environment that he defines,for him any one of the organism’s behaviors appears asan actualization of the niche, that is, as a first order de-scription of the environment (henceforth denoted by acapital D: Description) This Description, however, is adescription in terms of behavior (interactions) of the ob-served organism, not of representations of environmentalstates, and the relation between behavior and niche liesexclusively in the cognitive domain of the observer.

(3) An organism can modify the behavior of another or-ganism in two basic ways:

(a) By interaction with it in a manner that directsboth organisms toward each other in such a waythat the ensuing behavior of each of them dependsstrictly on the following behavior of the other, e.g.:courtship and fight. A chain of interlocked behav-ior can thus be generated by the two organisms.

(b) By orienting the behavior of the other organism tosome part of its domain of interactions differentfrom the present interaction, but comparable to theorientation of that of the orienting organism. Thiscan take place only if the domains of interactionsof the two organisms are widely coincident; in thiscase no interlocked chain of behavior is elicitedbecause the subsequent conduct of the two organ-isms depends on the outcome of independent, al-though parallel, interactions.

In the first case it can be said that the two organisms in-teract; in the second case that they communicate. The

second case is the basis for any linguistic behavior; thefirst organism generates (as is apparent for the observer)a Description of its niche that, in addition to its own sig-nificance as a behavior (within the cognitive domain ofthe first organism, and independently of it), orients thesecond organism within its cognitive domain to an inter-action from which ensues a conduct parallel to that of thefirst one, but unrelated to it. The conduct thus elicited bythe orienting behavior is denotative: it points to a featureof the environment that the second organism encountersin its niche and Describes by the appropriate conduct,and that he can treat as an independent entity. The ori-enting behavior is, for the observer, a second order de-scription (henceforth denoted by italics: description) thatrepresents that which he considers it to denote. By con-trast, the orienting behavior of the first organism is con-notative for the second one, and implies for it an inter-action within its cognitive domain which, if actualized,originates a behavior that Describes a particular aspectof its niche; that which an orienting behavior connotesis a function of the cognitive domain of the orientee, notthe orienter.

(4) In an orienting interaction the behavior of the firstorganism, as a communicative description causes in thenervous system of the second one a specific state of activ-ity; this state of activity embodies the relations generatedin the interaction and represents the behavior of the sec-ond organism (Description of its niche) connoted by theorienting behavior of the first one. This representation,as a state of neuronal activity, can in principle be treatedby the nervous system as a unit of interactions, and thesecond organism, if capable of doing so, can thus interactwith representations of its own Descriptions of its nicheas if these were independent entities. This generates yetanother domain of interactions (and hence, another di-mension in the cognitive domain), the domain of interac-tions with representations of behavior (interactions), ori-enting interactions included, as if these representationswere independent entities within the niche: the linguisticdomain.

(5) If an organism can generate a communicative de-scription and then interact with its own state of activitythat represents this description, generating another suchdescription that orients towards this representation . . . ,the process can in principle be carried on in a potentiallyinfinite recursive manner, and the organism becomes anobserver: it generates discourse as a domain of interac-tions with representations of communicative descriptions(orienting behaviors).

Furthermore: if such an observer through orienting be-havior can orient himself towards himself, and then gen-


erate communicative descriptions that orient him to-wards his description of this self-orientation, he can, bydoing so recursively, describe himself describing him-self . . . endlessly. Thus discourse through communica-tive description originates the apparent paradox of self-description: self-consciousness, a new domain of inter-actions.

(6) A nervous system capable of recursively interactingwith its own states as if these were independent entitiescan do so regardless of how these states are generated,and in principle can repeat these recursive interactionsendlessly. Its only limitation lies in the need that the pro-gressive transformation of its actual and potential behav-ior, which in such a system is a necessary concomitantto behavior itself, be directly or indirectly subservient tothe basic circularity of the living organization. The lin-guistic domain, the observer and self-consciousness areeach possible because they result as different domains ofinteractions of the nervous system with its own states incircumstances in which these states represent differentmodalities of interactions of the organism.

Thinking

(1) I consider that in a state-determined nervous system,the neurophysiological process that consist in its interact-ing with some of its own internal states as if these wereindependent entities corresponds to what we call think-ing. Such internal states of nervous activity, otherwisesimilar to other states of nervous activity that participatein the specification of behavior, as in reflex mechanisms,cause conduct by determining specific changes of statein the nervous system. Thinking thus conceived, and re-flex mechanisms, are both neurophysiological processesthrough which behavior emerges in a deterministic man-ner; they differ, however, in that in a reflex action wecan, in our description trace a chain of nervous interac-tions that begins with a specific state of activity at thesensory surfaces; while in thinking, the chain of nervousinteractions that leads to a given conduct (change in theeffector surfaces) begins with a distinguishable state ofactivity of the nervous system itself, whichever way itmay have originated. Accordingly, thinking is a modeof operation of the nervous system that reflects function-ally its internal anatomical projection (possibly multiply)onto itself.

(2) The process of thinking as characterized above isnecessarily independent of language. That this is so evenfor what we call “abstract thinking” in man is appar-ent from the observations of humans with split brains

[Cf. Gazzaniga, Bogen and Sperry, 1965]. These ob-servations show that the inability of the non-speakinghemisphere to speak does not preclude in it operationsthat the observer would call abstract thinking, and thatthe lack of language only implies that it cannot gener-ate discourse. When we talk about concepts or ideas wedescribe our interactions with representations of our de-scriptions, and we think through our operation in the lin-guistic domain. The difficulty arises from our consid-ering thinking through our description of it in terms orconcepts as if it were something peculiar to man, and insome way isomorphic with the notions embodied in thedescriptions, instead of attending to the functional pro-cess that makes these descriptions possible.

Natural Language

(1) Linguistic behavior is orienting behavior; it orientsthe orientee within his cognitive domain to interactionsthat are independent of the nature of the orienting in-teractions themselves. To the extent that the part of itscognitive domain toward which the orientee is thus ori-ented is not genetically determined and becomes spec-ified through interactions, one organism can in princi-ple orient another to any part of its cognitive domainby means of arbitrary modes of conduct also specifiedthrough interactions. However, only if the domains of in-teractions of the two organisms are to some extent com-parable, are such consensual orienting interactions pos-sible and are the two organisms able to develop someconventional, but specific, system of communicative de-scriptions to orient each other to cooperative classes ofinteractions that are relevant for both.

(2) The understanding of the evolutionary origin of nat-ural languages requires the recognition in them of a basicbiological function which, properly selected, could orig-inate them. So far this understanding has been impossi-ble because language has been considered as a denotativesymbolic system for the transmission of information. Infact, if such were the biological function of language,its evolutionary origin would demand the pre-existenceof the function of denotation as necessary to developthe symbolic system for the transmission of information,but this function is the very one whose evolutionary ori-gin should be explained. Conversely, if it is recognizedthat language is connotative and not denotative and thatits function is to orient the orientee within his cognitivedomain, and not to point to independent entities, it be-comes apparent that learned orienting interactions em-body a function of non-linguistic origin that, under a se-lective pressure for recursive application, can originate


through evolution the system of cooperative consensualinteractions between organisms that is natural language.Particular orienting interactions, like any other learnedconduct, arise from the substitution of one type of inter-action for another as a cause for a given behavior, andtheir origin as a function of the general learning capacityof the nervous system is completely independent of thecomplexities of the system of cooperative interactions towhich their recursive application gives rise. Widespreadamong animals other than man-orienting interactions areparticularly evident in primates, in which it is easy tosee how the audible and visible behavior of one individ-ual orients others within their respective cognitive do-mains [Cf. Jay, 1968], and in dolphins which seem tohave evolved a rich and efficient system of auditive co-operative interactions [Cf. Lilly, 1967]. In accordancewith all this I maintain that learned orienting interactions,coupled with some mode of behavior that allowed for anindependent recursive expansion of the domain of inter-actions of the organism, such as social life [Cf. Gardnerand Gardner, 1969] and/or tool making and use, musthave offered a selective basis for the evolution of the ori-enting behavior that in hominids led to our present-daylanguages.

(3) Behavior (function) depends on the anatomical orga-nization (structure) of the living system, hence anatomyand conduct cannot legitimately be separated and theevolution of behavior is the evolution of anatomy andvice versa; anatomy provides the basis for behavior andhence for its variability; behavior provides the ground forthe action of natural selection and hence for the historicalanatomical transformations of the organism. Structureand function are, however, both relative to the perspec-tive of interactions of the system and cannot be consid-ered independently of the conditions that define it as aunit of interactions, for what is from one perspective aunit of interactions, from another may only be a compo-nent of a larger one, or may be several independent units.It is the dynamics of this process of individuation, as anhistorical process in which every state of a changing sys-tem can become a unit of interactions if the proper cir-cumstances are given, what makes the evolution of livingsystems a deterministic process of necessarily increasingcomplication. Thus, in the evolution of language, naturalselection, by acting upon orienting behavior as a functionthat if enhanced strongly increases the cooperation be-tween social animals, has led to anatomical transforma-tions which provide the basis for the increased complex-ity of the orienting conduct and the diversity of the in-teractions toward which man can be oriented in his cog-nitive domain. The complexity of the orienting conducthas increased through an increase in the complexity and

variety of motor behavior, particularly through vocaliza-tion and tool making. The diversity of the interactionstoward which man can be oriented has increased througha concomitant expansion of the internal projection of thebrain onto itself, by means of new interconnections be-tween different cortical areas (as compared with otherprimates), between cortical areas and subcortical nuclei[Cf. Geschwind, 1964], and possibly also between differ-ent cortical layers and cellular systems within the cortexitself.

(4) So long as language is considered to be denotative itwill be necessary to look at it as a means for the trans-mission of information, as if something were transmittedfrom organism to organism, in a manner such that the do-main of uncertainties of the “receiver” should be reducedaccording to the specifications of the “sender”. However,when it is recognized that language is connotative andnot denotative, and that its function is to orient the ori-entee within his cognitive domain without regard for thecognitive domain of the orienter, it becomes apparent thatthere is no transmission of information through language.It behooves the orientee, as a result of an independent in-ternal operation upon his own state, to choose where toorient his cognitive domain; the choice is caused by the“message”, but the orientation thus produced is indepen-dent of what the “message” represents for the orienter. Ina strict sense then, there is no transfer of thought from thespeaker to his interlocutor; the listener creates informa-tion by reducing his uncertainty through his interactionsin his cognitive domain. Consensus arises only throughcooperative interactions in which the resulting behaviorof each organism becomes subservient to the mainte-nance of both. An observer beholding a communicativeinteraction between two organisms who have already de-veloped a consensual linguistic domain, can describe theinteraction as denotative; for him, a message (sign) ap-pears as denoting the object which the conduct of theorientee Describes (specifies), and the conduct of the ori-entee appears determined by the message. However, be-cause the outcome of the interaction is determined in thecognitive domain of the orientee regardless of the sig-nificance of the message in the cognitive domain of theorienter, the denotative function of the message lies onlyin the cognitive domain of the observer and not in theoperative effectiveness of the communicative interaction.The cooperative conduct that may develop between theinteracting organisms from these communicative inter-actions is a secondary process independent of their oper-ative effectiveness. If it appears acceptable to talk abouttransmission of information in ordinary parlance, this isso because the speaker tacitly assumes the listener to beidentical with him and hence as having the same cog-


nitive domain which he has (which never is the case),marveling when a “misunderstanding” arises. Such anapproach is valid, for man created systems of communi-cation where the identity of sender and receiver is implic-itly or explicitly specified by the designer, and a message,unless disturbed during transmission, necessarily selectsat the reception the same set of states that it represents atthe emission, but not for natural languages.

(5) It behooves the interlocutor to choose where to ori-ent in his cognitive domain as a result of a linguistic in-teraction. Since the mechanism of choice, as in everyneuronal process, is state-dependent, the state of activ-ity from which the choice (new state of neuronal activ-ity) must arise restricts the possible choices and consti-tutes a reference background in the orientee. The same isvalid for the speaker; the state of activity from which hiscommunicative description (linguistic utterance) arisesconstitutes the reference background that specifies hischoice. All the interactions that independently specifythe reference background of each interlocutor constitutethe context in which a given linguistic interaction takesplace. Every linguistic interaction is thus necessarilycontext-dependent, and this dependency is strictly deter-ministic for both orienter and orientee, notwithstandingthe different backgrounds of the two processes. It is onlyfor the observer that there is any ambiguity in a linguis-tic interaction that he observes; this is because he has noaccess to the context in which it occurs. The sentence,“They are flying planes,” is unambiguous for both inter-locutors, regardless of the subsequent behavior which itoriginates in each of them; for the observer, however,who wants to predict the course of the ensuing interac-tions, it is ambiguous.

(6) If one considers linguistic interactions as orientinginteractions it is apparent that it is not possible to sepa-rate, functionally, semantics and syntax, however sepa-rable they may seem in their description by the observer.This is true for two reasons:

(a) A sequence of communicative descriptions (wordsin our case) must be expected to cause in the ori-entee a sequence of successive orientations in hiscognitive domain, each arising from the state leftby the previous one. “They are flying planes”clearly illustrates this; each successive word ori-ents the listener to a particular interaction in hiscognitive domain that is relevant in a particularmanner (apparent in the conduct it generates) thatdepends on the previous orientation. The fact thatit seems that the observer can more easily describethe word are (or any word) by referring to its gram-

matical and lexical functions, rather than by spec-ifying the nature of the orientation that it causes(in terms of conduct or interactions), should notobscure the problem. The observer speaks, andany explanation of the word are that he may givelies in the descriptive domain, while the orienta-tion caused by the word itself, as a change of stateof the listener, is an internal interaction in his cog-nitive domain.

(b) An entire series of communicative descriptionscan itself be a communicative description: thewhole sequence once completed may orient the lis-tener from the perspective of the state to which thesequence itself has led him. The limit to such com-plications lies exclusively in the capacity of thenervous system to discriminate between its owndiscriminable internal states, and to interact withthem as if with independent entities.

(7) Linguistic behavior is an historical process of con-tinuous orientation. As such, the new state in which thesystem finds itself after a linguistic interaction emergesfrom the linguistic behavior. The rules of syntax andgenerative grammar [Cf. Chomsky, 1968] refer to reg-ularities that the observer sees in the linguistic behavior(as he would see in any behavior) which, arising fromthe functional organization of the system, specify the in-teractions that are possible at any given moment. Suchrules, as rules, lie exclusively in the cognitive domain ofthe observer, in the realm of descriptions, because thetransitions from state to state as internal processes in anysystem are unrelated to the nature of the interactions towhich they give rise. Any correlation between differentdomains of interactions lies exclusively in the cognitivedomain of the observer, as relations emerging from hissimultaneous interactions with both.

(8) The coordinated states of neuronal activity whichspecify a conduct as a series of effector and receptorstates whose significance arises in a consensual domain,does not differ in its neurophysiological generation fromother coordinated states of neuronal activity which spec-ify other conducts of innate or acquired significance(walking, flying, playing a musical instrument). Thus,however complex the motor and sensory coordinations ofspeech may be, the peculiarity of linguistic behavior doesnot lie in the complexity or nature of the series of effectorand receptor states that constitute it, but in the relevancethat such behavior acquires for the maintenance of thebasic circularity of the interacting organisms through thedevelopment of the consensual domain of orienting in-teractions. Speaking, walking, or music-making do not


differ in the nature of the coordinated neuronal processeswhich specify them but in the sub-domains of interac-tions in which they acquire their relevance.

(9) Orienting behavior in an organism with a nervoussystem capable of interacting recursively with its ownstates expands its cognitive domain by enabling it to in-teract recursively with descriptions of its interactions. Asa result:

(a) Natural language has emerged as a new domainof interaction in which the organism is modifiedby its descriptions of its interactions, as they be-come embodied in states of activity of its ner-vous system, subjecting its evolution to its inter-actions in the domains of observation and self-consciousness.

(b) Natural language is necessarily generative becauseit results from the recursive application of the sameoperation (as a neurophysiological process) on theresults of this application.

(c) New sequences of orienting interactions (new sen-tences) within the consensual domain are necessar-ily understandable by the interlocutor (orient him),because each one of their components has definiteorienting functions as a member of the consensualdomain that it contributes to define.

Memory and Learning

(1) Learning as a process consists in the transformationthrough experience of the behavior of an organism in amanner that is directly or indirectly subservient to themaintenance of its basic circularity. Due to the state de-termined organization of the living system in general,and of the nervous system in particular, this transforma-tion is an historical process such that each mode of be-havior constitutes the basis over which a new behaviordevelops, either through changes in the possible statesthat may arise in it as a result of an interaction, or throughchanges in the transition rules from state to state. The or-ganism is thus in a continuous process of becoming thatis specified through an endless sequence of interactionswith independent entities that select its changes of statebut do not specify them.

(2) Learning occurs in a manner such that, for the ob-server, the learned behavior of the organism appears jus-tified from the past, through the incorporation of a rep-resentation of the environment that acts, modifying its

present behavior by recall; notwithstanding this, the sys-tem itself functions in the present, and for it learning oc-curs as an atemporal process of transformation. An or-ganism cannot determine in advance when to change andwhen not to change during its flow of experience, nor canit determine in advance which is the optimal functionalstate that it must reach; both the advantage of any partic-ular behavior and the mode of behavior itself can only bedetermined a posteriori, as a result of the actual behav-ing of the organism subservient to the maintenance of itsbasic circularity.

(3) The learning nervous system is a deterministic sys-tem with a relativistic self-regulating organization thatdefines its domain of interactions in terms of the statesof neuronal activity that it maintains constant, both in-ternally and at its sensory surfaces, and that specifiesthese states at any moment through its functioning, andthrough the learning (historical transformation) itself.Consequently, it must be able to undergo a continuoustransformation, both in the states it maintains constant,and in the way it attains them, so that every interactionin which new classes of concomitances occur effectivelymodifies it (learning curves) in one direction or the other.Since this transformation must occur as a continuous pro-cess of becoming without the previous specification ofan end state, the final specification and optimization of anew behavior can only arise through the cumulative ef-fect of many equally directed interactions, each of whichselects, from the domain of structural changes possible tothe nervous system in its structural dynamism, that whichat that moment is congruent with its continued operationsubservient to the basic circularity of the organism. Oth-erwise the organism disintegrates.

(4) The analysis of the nervous system made earlier in-dicated that the states of neuronal activity that arise init through each interaction embody the relations givenin the interaction, and not representations of the nicheor the environment as the observer would describe them.This analysis also indicated that functionally such em-bodiments constitute changes in the reactivity of the ner-vous system, as a system closed on itself, to the mod-ulating influences of further interactions. Consequentlywhat the observer calls “recall” and “memory” cannotbe a process through which the organism confronts eachnew experience with a stored representation of the nichebefore making a decision, but the expression of a modi-fied system capable of synthesizing a new behavior rele-vant to its present state of activity.

(5) It is known that many neurons change their trans-fer functions as a result of the different concomitances


of activity that occur in the neuropils of their collectorand effector areas. Although it is not known what thesechanges are (development of new synapses or changes intheir size, membrane changes, or changes in the patternof spike invasion at the branching points of the axons),it can be expected from the relativistic organization ofthe nervous system that they should result in local mor-phological and functional changes that do not representany particular interaction, but which permanently alterthe reactivity of the system. This anatomical and func-tional transformation of the nervous system must neces-sarily be occurring continuously as changes that the cellsare able to stabilize with a permanency that lasts until thenext modification, which can occur in any direction withrespect to the previous one, or that subside by themselvesafter a certain number of interactions, but which are be-ing locally triggered and selected through the actual con-comitances of activity taking place in the neuropil itself.

(6) All changes in the nervous system during learningmust occur without interference with its continued func-tioning as a self-regulating system; the unity that the ob-server sees in a living system throughout its continuoustransformation is a strictly functional one. Accordingly,what appears constant for the observer when he ascer-tains that the same behavior is reenacted on a differentoccasion, is a set of relations that he defines as charac-terizing it, regardless of any change in the neurophysio-logical process through which it is attained, or any otherunconsidered aspect of the conduct itself. Learning, as arelation between successive different modes of conductof an organism such that the present conduct appears asa transformation of a past conduct arising from the recallof a specifiable past event, lies in the cognitive domain ofthe observer as a description of his ordered experiences.Likewise, memory as an allusion to a representation inthe learning organism of its past experiences, is also adescription by the observer of his ordered interactionswith the observed organism; memory as a storage of rep-resentations of the environment to be used on differentoccasions in recall does not exist as a neurophysiologicalfunction.

(7) It is sufficient for a system to change its state after aninteraction in a manner such that whenever a similar in-teraction recurs some internally determined concomitantstate does not recur, although the same overt behavior isreenacted for it to treat two otherwise equivalent interac-tions as different elements of the same class. Such a pe-culiar state could be described as representing the emo-tional connotation of uncertainty which, present when-ever a class of interactions is experienced for the firsttime, is suppressed after such an experience; the absence

of such a concomitant state would suffice henceforth totreat differently (as known) all recurrent interactions ofthe same class. I maintain that modifications of this sortin the reactivity of the nervous system constitute the basisfor the unidirectional ordering of experiences in a livingsystem through “recognition” without any storage of rep-resentations of the niche. First interactions that by errorof the system are not accompanied by the above men-tioned concomitant internal state (emotional connotationof uncertainty) would be treated as if known, as occursin the deja vu. Conversely, interference with the suppres-sion of the concomitant state of activity correspondingto this emotional connotation would result in the treat-ment of any recurrent interaction as if new (loss of recentmemory).

If such a system is capable of discourse, it will generatethe temporal domain through the ascription of a unidi-rectional order to its experiences as they differ in theiremotional connotations, and although it will continue tofunction in the present as an atemporal system, it willinteract through its descriptions in the temporal domain.Past, present, and future, and time in general belong ex-clusively to the cognitive domain of the observer.

The Observer

Epistemological and Ontological Implications

(1) The cognitive domain is the entire domain of inter-actions of the organism. The cognitive domain can beenlarged if new modes of interactions are generated. In-struments enlarge our cognitive domain.

(2) The possibility of enlargement of the cognitive do-main is unlimited; it is a historical process. Our brain, thebrain of the observer, has specialized during evolution asan instrument for the discrimination of relations, both in-ternally and externally generated relations, but relationsgiven through and by interactions and embodied in thestates of relative activity of its neurons. Furthermore,this occurs under circumstances in which the discrimi-nations between states of relative activity—that for anobserver represent the interactions of the organism, forthe nervous system, that operate as a closed network—constitute only changes of relations of activity that arisebetween its components while it generates the internaland the sensory motor correlations that the states of theorganism select. This has two aspects: one refers to thefunctional organization of the nerve cells which, withtheir responses, discriminate between different states ofrelative activity impinging upon them; the other refers tothe ability of the nervous system, as a neuronal organiza-


tion, to discriminate between its own states as these aredistinguished and specified by the further states of activ-ity that they generate. From this capacity of the nervoussystem to interact discriminately with its own states ina continuous process of self transformation, regardlessof how these states are generated, behavior emerges asa continuum of self-referred functional transformation.We cannot say in absolute terms what constitutes an in-put to our nervous system (the nervous system of the ob-server), because every one of its states can be its inputand can modify it as an interacting unit. We can say thatevery internal interaction changes us because it modifiesour internal state, changing our posture or perspective (asa functional state) from which we enter into a new inter-action. As a result new relations are necessarily createdin each interaction and, embodied in new states of activ-ity, we interact with them in a process that repeats itselfas a historical and unlimited transformation.

(3) The observer generates a spoken description of hiscognitive domain (which includes his interactions withand through instruments). Whatever description hemakes, however, that description corresponds to a set ofpermitted states of relative activity in his nervous systemembodying the relations given in his interactions. Thesepermitted states of relative activity and those recursivelygenerated by them are made possible by the anatomi-cal and functional organization of the nervous systemthrough its capacity to interact with its own states. Thenervous system in turn has evolved as a system struc-turally and functionally subservient to the basic circular-ity of the living organization, and hence, embodies aninescapable logic: that logic which allows for a matchbetween the organization of the living system and the in-teractions into which it can enter without losing its iden-tity.

(4) The observer can describe a system that gives riseto a system that can describe, hence, to an observer. Aspoken explanation is a paraphrase, a description of thesynthesis of that which is to be explained; the observerexplains the observer. A spoken explanation, however,lies in the domain of discourse. Only a full reproductionis a full explanation.

(5) The domain of discourse is a closed domain, and itis not possible to step outside of it through discourse.Because the domain of discourse is a closed domain itis possible to make the following ontological statement:the logic of thedescription is the logic of the describing(living) system (and his cognitive domain).

(6) This logic demands a substratum for the occurrenceof the discourse. We cannot talk about this substratum

in absolute terms, however, because we would have todescribe it, and a description is a set of interactions intowhich the describer and the listener can enter, and theirdiscourse about these interactions will be another set ofdescriptive interactions that will remain in the same do-main. Thus, although this substratum is required for epis-temological reasons, nothing can be said about it otherthan what is meant in the ontological statement above.

(7) We as observers live in a domain of discourse inter-acting with descriptions of our descriptions in a recursivemanner, and thus continuously generate new elements ofinteraction. As living systems, however, we are closedsystems modulated by interactions through which we de-fine independent entities whose only reality lies in theinteractions that specify them (their Description).

(8) For epistemological reasons we can say: thereare properties which are manifold and remain con-stant through interactions. The invariance of propertiesthrough interactions provides a functional origin to enti-ties or units of interactions; since entities are generatedthrough the interactions that define them (properties), en-tities with different classes of properties generate inde-pendent domains of interactions: no reductionism is pos-sible.

V. Problems in the Neurophysiologyof Cognition

(1) The observer can always remain in a domain of inter-actions encompassing his own interactions; he has a ner-vous system capable of interacting with its own states,which, by doing so in a functional context that definesthese states as representations of the interactions fromwhich they arise, allows him to interact recursively withrepresentations of his interactions. This is possible be-cause due to the general mode of organization of the ner-vous system there is no intrinsic difference between itsinternally and externally generated states of activity, andbecause each one of its specific states of activity is speci-fiable only in reference to other states of activity of thesystem itself.

(2) An organism with a nervous system capable of in-teracting with its own states is capable of descriptionsand of being an observer if its states arise from learnedorienting interactions in a consensual domain: it can de-scribe its describing [Cf. Gardner and Gardner, 1969].Through describing itself in a recursive manner, such anorganism becomes a self-observing system that generates


the domain of self-consciousness as a domain of self-observation. Self-consciousness then is not a neurophys-iological phenomenon, it is a consensual phenomenonemerging in an independent domain of interactions fromself-orienting behavior and lies entirely in the linguisticdomain. The implications are twofold:

(a) The linguistic domain as a domain of orienting be-havior requires at least two interacting organismswith comparable domains of interactions, so that acooperative system of consensual interactions maybe developed in which the emerging conduct ofthe two organisms is relevant for both. The speci-fiability through learning of the orienting interac-tions allows for a purely consensual (cultural) evo-lution in this domain, without it necessarily involv-ing any further evolution of the nervous system;for this reason the linguistic domain in general,and the domain of self-consciousness in particu-lar, are, in principle, independent of the biologicalsubstratum that generates them. However, in theactual becoming of the living system this indepen-dence is incomplete, on the one hand because theanatomical and neurophysiological organization ofthe brain, by determining the actual possibilitiesof confluence of different states of activity in it,specifies both the domain of possible interactionsof the organism with relations and the complexityof the patterns of orienting interactions that it candistinguish, and on the other hand because the nec-essary subservience of the linguistic domain to themaintenance of the basic circularity of the organ-ism through the generation of modes of behaviorthat directly or indirectly satisfy it limits the typeof conduct that the organism can have without animmediate or eventual disintegration, or, of course,reduced rate of reproduction. Consequently, then,although the purely consensual aspects of the cul-tural evolution are independent of a simultaneousevolution of the nervous system, those aspects ofthe cultural evolution which depend on the possi-bility of establishing new classes of concomitancesof activity in the nervous system, and generate newrelations between otherwise independent domains,are not thus independent. Accordingly, once a cul-tural domain is established, the subsequent evolu-tion of the nervous system is necessarily subordi-nated to it in the measure that it determines thefunctional validity of the new kinds of concomi-tances of activity that may arise in the nervous sys-tem through genetic variability.

(b) Since self-consciousness and the linguistic domain

in general are not neurophysiological phenomena,it is impossible to account for them in terms of ex-citation, inhibition, networks, coding, or whateverelse is the stuff of neurophysiology. In fact, thelinguistic domain is fully explained only by show-ing how it emerges from the recursive applicationof orienting interactions on the results of their ap-plications without being restricted as a domain bythe neurophysiological substratum; what indeed isthe problem is the need to account in purely phys-iological terms, without reference to meaning, forthe synthesis of behavior in general, and for thesynthesis of orienting behavior in particular. Ac-cordingly, the fundamental quest in this respectshould be to understand and explain

(i) how does the nervous system interact with itsown states and is modified by them as if theywere independent entities?

(ii) how are these states specified neurophysio-logically if they are defined by their own ef-fectiveness in bringing forth certain internalor sensory states in the system?

(iii) ) how is a given effector performance synthe-sized that is defined by the relative states ofactivity that it generates in the sensory sur-faces and in the system itself?; and

(iv) how do the double or triple internal anatomi-cal projections of the nervous system onto it-self determine its capacity to single out someof its own states and interact with them inde-pendently?.

(3) At any moment each nerve cell responds in a deter-ministic manner, and according to well defined transferfunctions to classes of spatio-temporal activity causedat its collector area by the afferent influences imping-ing upon it; this occurs independently of how these af-ferent influences arise. This mode of cellular operationconstitutes the basis for an associative process in which,whenever a given state of activity is produced in the ner-vous system, all neurons for which this state generatesthe proper classes of afferent influences enter into activ-ity. Association thus conceived neurophysiologically isan inevitable process that calls into activity all cells thatcan be activated at any moment by a given state of thenervous system. No consideration of meaning enters intosuch a notion, since meaning, as a description by the ob-server, refers to the relevance that a mode of behaviorhas in the maintenance of the basic circularity of the or-ganism as a consequence of self-regulation, and not in


the mechanisms of the genesis of conduct. Associationin terms of representations related by meaning lies in thecognitive domain of the observer exclusively. The ner-vous system is a system that functions maintaining con-stant certain states of relative activity, both internally andat the sensory surfaces, with reference only to some of itsother states of relative activity. In this context the follow-ing considerations about its functional organization aresignificant:

(a) The nervous system can be described as a systemthat has evolved to specialize in the discriminationbetween states of neuronal relative activity (par-ticularly in man) each of which is defined by thebehavior it generates. This is valid for innate andlearned behavior in circumstances in which everybehavior is defined either by a set of states of ac-tivity maintained constant, or by their path of vari-ation, both internally and at the sensory surfaces.

(b) The basic connectivity of the nervous system, andthe original reactive capacity of the nerve cells,with which any animal is endowed by develop-ment, secures a basic pattern of flow for the ner-vous activity originating at any point in it. Thus,development specifies and determines both an ini-tial repertoire of behavior over which all new con-duct is built in a historical process of transfor-mation, and an initial structurally specified set ofpossible associations that changes in an integratedmanner with the historical transformation of be-havior [Cf. Lorenz, 1966].

(c) Any modification of the transfer function of anerve cell, resulting from new concomitances ofactivity, occurs modifying a preexisting behaviorin a system that operates through maintaining in-variant its definitory internal relations. In fact, anylocal change that would lead to the synthesis ofa modified conduct by the organism, must be im-mediately accompanied by other changes arisingthrough the adjustments that this must undergo inthe process of maintaining constant its internal re-lations under its changed behavior. This is whyit is the immediate relevance of a conduct for themaintenance of the organism in the present whichat any moment selects the changes that take placeduring learning, and not the possible value of theconduct for future action.

(d) It is apparent that the nervous system cannot de-termine in advance the concomitances of activityunder which it should change in a permanent man-ner; for it to satisfy future needs of the organism,

it must operate under non-predictive changes con-tinuously selected by the concomitances of activityarising in it. For this the nervous system must becapable of successful operation under the contin-uous transformation of its capacity to synthesizebehavior, which necessarily results from a contin-uous change of the neurophysiological concomi-tances that determine the effective spatio-temporalconfiguration of activity impinging on the collec-tor areas of its component neurons. Accordingly,it would seem of fundamental importance for thefunctional transformation of the system that manyof its neurons should be able to change their rel-ative participation in the synthesis of behavior aselements of different states of relative neuronal ac-tivity, independently of whether or not this is ac-companied by any change in their transfer func-tions. In these circumstances the actual problemfor the successful operation of the nervous sys-tem is the generation at any moment of the opti-mal configuration of activity necessary to synthe-size a given behavior. However, since this contin-uous transformation of the functional capacity ofthe nervous system necessarily occurs under con-tinuously successful behavior, such optimizationrequires no other specification than its attainmentthrough the converging transformation of behavioritself.

(e) Since the nervous system is an inferential system,that is, since it functions as if any state that oc-curred once will occur again, a significant featureof its organization must be its necessary and con-tinuous transformation as a function of the newconcomitances of activity occurring in it. Thisfunctional requirement could be satisfied, for ex-ample, if any new local concomitance of activityin the neuropils changes the nerve cells in a deter-ministic and specific manner which does not rep-resent any entity or event, but which modifies theneurophysiological circumstances under which thecorresponding post-synaptic neurons are activated.Such can occur if the probability of spike invasionat the branching points of the afferent axons in theneuropils is permanently modified in one direc-tion or another by the coincident novel activity inthe neighboring structures, which, in the absenceof synaptic interactions, cause, through local cur-rents, local processes of growth or ungrowth in thebranching zones of these axons. If this were thecase four things would occur:

(i) The state of the nervous system would


change, and hence, also its conduct, accord-ing to the new concomitances of activity pro-duced in the neuropils through its differentinteractions.

(ii) Each state of activity of the system (as a stateof relative neuronal activity) would be de-fined by the concomitances of activity in theneuropil that generate it, such that if they re-cur, it recurs.

(iii) Each new functional state of the neuropilswould necessarily constitute the basis fortheir further modification, in such a mannerthat their morphological and functional orga-nization would be under continuous histori-cal transformation.

(iv) These changes in the neuropils would changethe participation of the different neurons inthe synthesis of behavior, independently ofwhether or not there are also changes in theirtransfer functions, by changing the circum-stances of their activation. Accordingly, if aninteraction (as described by the observer) re-curs, no past conduct could be strictly reen-acted by the organism, but this would haveto synthesize a new adequate behavior thatgenerates, in the context of its present inter-action and in a manner that became specifiedthrough its structural transformation alongits history of interactions, the internal andsensory motor correlations that maintain itsidentity.

(4) Learning is not a process of accumulation of repre-sentations of the environment; it is a continuous processof transformation of behavior through continuous changein the capacity of the nervous system to synthesize it. Re-call does not depend on the indefinite retention of a struc-tural invariant that represents an entity (an idea, image,or symbol), but on the functional ability of the systemto create, when certain recurrent conditions are given, abehavior that satisfies the recurrent demands or that theobserver would class as a reenacting of a previous one.As a consequence, the quest in the study of the learningprocess must answer two basic questions:

“What changes can a neuron undergo (in any of its com-ponent parts) which it can maintain constant for a cer-tain time, and which modify in a definite manner its pos-sible participation in different configurations of relativeneuronal activity?”; and

“What organization of the nervous system would permit

continuous changes in the relative activity of its anatom-ical components, as a result of different concomitancesin their activity, and still permit the synthesis of a con-duct that is defined only by the states of relative neu-ronal activity that it generates, and not by the compo-nents used?”.

(5) The nervous system is a strictly deterministic systemwhose structure specifies the possible modes of conductthat may emerge (be synthesized) from its functioning ina manner that varies according to the species, and the re-active perspective from which these modes of conductmay emerge. The reactive perspective, which the ob-server would call the emotional tone, does not specifya particular conduct, but determines the nature (aggres-sive, fearful, timid, etc.) of the course of the interac-tion [Cf. Kilmer, McCulloch and Blum, 1968]. Changesduring development, maturation, hormonal action, drugs,or learning, do not modify the deterministic character ofthis organization but change the capacity that the sys-tem has at any moment to synthesize behavior. Further-more, although any conduct or functional state alwaysarises through a process of historical transformation frompre-existing modes of conduct or functional states, thenervous system functions in the present, and past his-tory does not participate as an operant neurophysiolog-ical factor in the synthesis of conduct; nor does mean-ing, the relevance that a particular mode of conduct has,participate in it either. Time and meaning are effectivefactors in the linguistic domain, but as relational entitiesdo not have neurophysiological correlates in the opera-tion of the nervous system. Nor is the functional unityof the nervous system attained through a specific fea-ture of its organization, but emerges from the function-ing of its components (whatever these may be), each oneto its own accord, under circumstances that define theensemble as a unit of interactions in a particular domain[Cf. Lindauer, 1967, as an example in a social organism],and has no reality independent of these circumstances.Thus there is no peculiar neurophysiological process thatcould be shown to be responsible for this unity and toexplain it. Furthermore, in a strict sense, although thenervous system has anatomical components it does nothave functional parts since any mutilation leaves a func-tioning unit, with different properties as expressed by itspossible interactions, but a unit in the corresponding do-main. It appears incomplete only for the observer whobeholds it as an entity from the perspective of what hethinks it should be. Each component of the nervous sys-tem that the observer describes is defined in the domainof interactions of his observations, and as such is aliento the system which it is supposed to integrate. Everyfunction has a structure which embodies it and makes it


possible, but this structure is defined by the function inthe domain of its operation as a set of relations betweenelements also defined in this domain. Neurons are theanatomical units of the nervous system, but are not thestructural elements of its functioning. The structural el-ements of the functioning nervous system have not yetbeen defined, and it will probably be apparent when theyare defined that they must be expressed in terms of invari-ants of relative activities between neurons, in some man-ner embodied in invariants of relations of interconnec-tions, and not in terms of separate anatomical entities. Inmanmade systems this conceptual difficulty has not beenso apparent because the system of relations (the theory)that integrates the parts that the describer (the observer)defines is provided by him, and is specified in his domainof interactions; as a consequence, these relations appearso obvious to the observer that he treats them as arisingfrom the observation of the parts, and deludes himself,denying that he provides the unformulated theory thatembodies the structure of the system which he projectsonto them. In a self-referring system like a living systemthe situation is different: the observer can only make adescription of his interactions with parts that he definesthrough interactions, but these parts lie in his cognitivedomain only. Unless he explicitly or implicitly providesa theory that embodies the relational structure of the sys-tem, and conceptually supersedes his description of thecomponents, he can never understand it. Accordingly,the full explanation of the organization of the nervoussystem (and of the organism) will not arise from any par-ticular observation or detailed description and enumera-tion of its parts, but rather like any explanation, from thesynthesis, conceptual or concrete, of a system that doeswhat the nervous system (or the organism) does.

VI. Conclusions

The aim set forth in the introduction has been accom-plished. Through the description of the self-referringcircular organization of the living system, and throughthe analysis of the domains of interactions that such anorganization specifies, I have shown the emergence of aself-referring system capable of making descriptions andof generating, through orienting interactions with other,similar, systems and with itself, both a consensual lin-guistic domain and a domain of self-consciousness, thatis: I have shown the emergence of the observer. Thisresult alone satisfies the fundamental demand put forthat the outset: “The observer is a living system and anyunderstanding of cognition as a biological phenomenonmust account for the observer and his role in it”, and

proves the validity of this analysis.

Although the answers to the various questions posed inthe introduction and the fundamental implications of theanalysis are to be found in the text itself to the extentthat the theory adequately founds its whole development,there are several conclusions that I would like to state ex-plicitly:

i) The living organization is a circular organizationwhich secures the production or maintenance of the com-ponents that specify it in such a manner that the productof their functioning is the very same organization thatproduces them. Accordingly, a living system is an home-ostatic system whose homeostatic organization has itsown organization as the variable that it maintains con-stant through the production and functioning of the com-ponents that specify it, and is defined as a unit of inter-actions by this very organization. It follows that livingsystems are a subclass of the class of circular and home-ostatic systems. Also, it is apparent that the componentsreferred to above cannot be specified as parts of the liv-ing system by the observer who can only subdivide asystem in parts that he defines through his interactions,and which, necessarily, lie exclusively in his cognitivedomain and are operationally determined by his modeof analysis. Furthermore, the relations through whichthe observer claims that these parts constitute a unitarysystem are relations that arise only through him by hissimultaneous interactions with the parts and the intactsystem, and, hence, belong exclusively to his cognitivedomain. Thus, although the observer can decompose aliving system into parts that he defines, the descriptionof these parts does not and cannot represent a living sys-tem. In principle a part should be definable through itsrelations within the unit that it contributes to form by itsoperation and interactions with other parts; this, how-ever, cannot be attained because the analysis of a unitinto parts by the observer destroys the very relations thatwould be significant for their characterization as effec-tive components of the unit. Furthermore, these relationscannot be recovered through a description which lies inthe cognitive domain of the observer and reflects only hisinteractions with the new units that he creates through hisanalysis. Accordingly, in a strict sense a unit does nothave parts, and a unit is a unit only to the extent that ithas a domain of interactions that defines it as differentfrom that with respect to which it is a unit, and can bereferred to only, as done above with the living system,by characterizing its organization through the domain ofinteractions which specify this distinction. In this con-text, the notion of component is necessary only for epis-temological reasons in order to refer to the genesis of the


organization of the unit through our description, but thisuse does not reflect the nature of its composition.

ii) For every living system its particular case of self re-ferring circular organization specifies a closed domain ofinteractions that is its cognitive domain, and no interac-tion is possible for it which is not prescribed by this or-ganization. Accordingly, for every living system the pro-cess of cognition consists in the creation of a field of be-havior through its actual conduct in its dosed domain ofinteractions, and not in the apprehension or the descrip-tion of an independent universe. Our cognitive process(the cognitive process of the observer) differs from thecognitive processes of other organisms only in the kindsof interactions into which we can enter, such as linguis-tic interactions, and not in the nature of the cognitive pro-cess itself. In this strictly subject-dependent creative pro-cess, inductive inference is a necessary function (mode ofconduct) that emerges as a result of the self-referring cir-cular organization which treats every interaction and theinternal state that it generates as if it were to be repeated,and as if an element of a class. Hence, functionally, fora living system every experience is the experience of ageneral case, and it is the particular case, not the gen-eral one, which requires many independent experiencesin order that it be specified through the intersection ofvarious classes of interactions. Consequently, althoughdue to the historical transformation they have caused inorganisms, or in their nervous systems, past interactionsdetermine the inductive inferences that these make in thepresent, they do not participate in the inductive processitself. Inductive inference as a structural property of theliving organization and of the thinking process, is inde-pendent of history, or of the relations between past andpresent that belong only to the domain of the observer.

iii) Linguistic interactions orient the listener within hiscognitive domain, but do not specify the course of hisensuing conduct. The basic function of language as asystem of orienting behavior is not the transmission of in-formation or the description of an independent universeabout which we can talk, but the creation of a consen-sual domain of behavior between linguistically interact-ing systems through the development of a cooperativedomain of interactions.

iv) Through language we interact in a domain of de-scriptions within which we necessarily remain evenwhen we make assertions about the universe or about ourknowledge of it. This domain is both bounded and infi-nite; bounded because everything we say is a descrip-tion, and infinite because every description constitutes inus the basis for new orienting interactions, and hence, for

new descriptions. From this process of recursive applica-tion of descriptions self-consciousness emerges as a newphenomenon in a domain of self-description, with noother neurophysiological substratum than the neurophys-iological substratum of orienting behavior itself. The do-main of self-consciousness as a domain of recursive self-descriptions is thus also bounded and infinite.

v) A living system is not a goal-directed system; it is,like the nervous system, a stable state-determined andstrictly deterministic system closed on itself and modu-lated by interactions not specified through its conduct.These modulations, however, are apparent as modula-tions only for the observer who beholds the organism orthe nervous system externally, from his own conceptual(descriptive) perspective, as lying in an environment andas elements in his domain of interactions. Contrariwise,for the functioning of the self-referring system itself allthat there is is the sequence of its own self-subservientstates. If this distinction is not made, one is liable tofail by including in the explanation of the organism andthe nervous system features of interactions (descriptions)that belong exclusively to the cognitive domain of the ob-server.

vi) It is tempting to talk about the nervous system as onewould talk about a stable system with input. This I rejectbecause it misses entirely the point by introducing thedistortion of our participation as observers into the ex-planation of systems whose organization must be under-stood as entirely self-referring. What occurs in a livingsystem is analogous to what occurs in an instrumentalnight where the pilot does not have access to the outsideworld and must function only as a controller of the val-ues shown in his flight instruments. His task is to securea path of variations in the readings of his instruments,either according to a prescribed plan, or to one that be-comes specified by these readings. When the pilot stepsout of the plane he is bewildered by the congratulationsof his friends on account of the perfect flight and landingthat he performed in absolute darkness. He is perplexedbecause to his knowledge all that he did at any momentwas to maintain the readings of his instruments withincertain specified limits, a task which is in no way rep-resented by the description that his friends (observers)make of his conduct.

In terms of their functional organization living systemsdo not have inputs and outputs, although under pertur-bations they maintain constant their set states, and it isonly in our descriptions, when we include them as partsof larger systems which we define, that we can say thatthey do. When we adopt this descriptive approach in ouranalysis of the living organization we cannot but subor-


dinate our understanding of it to notions valid only forman-made (allo-referring) systems, where indeed inputand output functions are all important through the pur-poseful design of their role in the larger systems in whichthey are included, and this is misleading. In the organiza-tion of the living systems the role of the effector surfacesis only to maintain constant the set states of the receptorsurfaces, not to act upon an environment, no matter howadequate such a description may seem to be for the anal-ysis of adaptation, or other processes; a grasp of this isfundamental for the understanding of the organization ofliving systems.

vii) The cognitive domain of the observer is boundedbut unlimited; he can in an endless recursive manner in-teract with representations of his interactions and gen-erate through himself relations between otherwise inde-pendent domains. These relations are novelties which,arising through the observer, have no other (and no less)effectiveness than that given to them by his behavior.Thus, he both creates (invents) relations and generates(specifies) the world (domain of interactions) in whichhe lives by continuously expanding his cognitive domainthrough recursive descriptions and representations of hisinteractions. The new, then, is a necessary result of thehistorical organization of the observer that makes of ev-ery attained state the starting point for the specificationof the next one, which thus cannot be a strict repetitionof any previous state; creativity is the cultural expressionof this unavoidable feature.

viii) The logic of the description and, hence, of behav-ior in general is, necessarily, the logic of the describingsystem; given behavior as a referential and deterministicsequence of states of nervous activity in which each statedetermines the next one within the same frame of refer-ence, no contradiction can possibly arise in it as long asthe latter remains unchanged by intercurrent interactions.If a change in the frame of reference takes place while agiven behavior develops, a new one appears, such thatthe states following the change are determined with re-spect to it. If the new sequence of states (behavior) ap-pears to an observer as contradicting the previous ones,this is so because he provides an independent and con-stant frame of reference in relation to which the suc-cessive sequences of states (behaviors) are contradictory.Such contradiction, however, lies exclusively in the cog-nitive domain of the observer, or of whatever providesthe independent constant frame of reference. Contradic-tions (inconsistencies) then, do not arise in the generationof behavior but pertain to a domain in which the differ-ent behaviors acquire their significance by confrontingan encompassing frame of reference through the interac-

tions of the organism. Accordingly, thinking and dis-course as modes of behavior are necessarily logicallyconsistent in their generation, and that which the ob-server calls rational in them because they appear as con-catenations of non-contradictory sequence dependent de-scriptions, is an expression of this necessary logical con-sistency. It follows that inconsistencies (irrationalities)in thinking and discourse as they appear to the observerarise from contextual changes in the circumstances thatgenerate them while the independent frame of referenceprovided by the observer remains unchanged.

ix) Due to the nature of the cognitive process and thefunction of the linguistic interactions, we cannot say any-thing about that which is independent of us and withwhich we cannot interact; to do that would imply a de-scription and a description as a mode of conduct repre-sents only relations given in interactions. Because thelogic of the description is the same as the logic of the de-scribing system we can assert the epistemological needfor a substratum for the interactions to occur, but we can-not characterize this substratum in terms of properties in-dependent of the observer. From this it follows that real-ity as a universe of independent entities about which wecan talk is, necessarily, a fiction of the purely descriptivedomain, and that we should in fact apply the notion ofreality to this very domain of descriptions in which we,the describing system, interact with our descriptions asif with independent entities. This change in the notionof reality must be properly understood. We are used totalking about reality orienting each other through linguis-tic interactions to what we deem are sensory experiencesof concrete entities, but which have turned out to be, asare thoughts and descriptions, states of relative activitybetween neurons that generate new descriptions. Thequestion, “What is the object of knowledge?” becomesmeaningless. There is no object of knowledge. To knowis to be able to operate adequately in an individual or co-operative situation. We cannot speak about the substra-tum in which our cognitive behavior is given, and aboutthat of which we cannot speak, we must remain silent, asindicated by Wittgenstein. This silence, however, doesnot mean that we fall into solipsism or any sort of meta-physical idealism. It means that we recognize that we,as thinking systems, live in a domain of descriptions, ashas already been indicated by Berkeley, and that throughdescriptions we can indefinitely increase the complex-ity of our cognitive domain. Our view of the universeand of the questions we ask must change accordingly.Furthermore, this re-emergence of reality as a domainof descriptions does not contradict determinism and pre-dictability in the different domains of interactions; onthe contrary, it gives them foundation by showing that


they are a necessary consequence of the isomorphism be-tween the logic of the description and the logic of thedescribing system. It also shows that determinism andpredictability are valid only within the field of this iso-morphism; that is, they are valid only for the interactionsthat define a domain.

x) The genetic and nervous systems are said to codeinformation about the environment and to represent itin their functional organization. This is untenable; thegenetic and nervous systems code processes that spec-ify series of transformations from initial states, whichcan be decoded only through their actual implementa-tion, not descriptions that the observer makes of an envi-ronment which lies exclusively in his cognitive domain[Cf. Bernal, 1965]. The following is an illustration of theproblem:

Let us suppose that we want to build two houses. Forsuch a purpose we hire two groups of thirteen workerseach. We name one of the workers of the first group asthe group leader and give him a book which contains allthe plans of the house showing in a standard way thelayout of walls, water pipes, electric connections, win-dows, etc., plus several views in perspective of the fin-ished house. The workers study the plans and under theguidance of the leader construct the house, approximat-ing continuously the final state prescribed by the descrip-tion. In the second group we do not name a leader, weonly arrange the workers in a starting line in the field andgive each of them a book, the same book for all, contain-ing only neighborhood instructions. These instructionsdo not contain words such as house, pipes, or windows,nor do they contain drawings or plans of the house tobe constructed; they contain only instructions of what aworker should do in the different positions and in the dif-ferent relations in which he finds himself as his positionand relations change.

Although these books are all identical the workers readand apply different instructions because they start fromdifferent positions and follow different paths of change.The end result in both cases is the same, namely, a house.The workers of the first group, however, construct some-thing whose final appearance they know all the time,while the workers of the second group have no viewsof what they are building, nor do they need to have ob-tained them even when they are finished. For the ob-server both groups are building a house, and he knowsit from the start, but the house that the second groupbuilds lies only in his cognitive domain; the house builtby the first group, however, is also in the cognitive do-mains of the workers. The coding is obviously differentin the two cases. In fact, the instructions contained in

the book given to the first group clearly code the houseas the observer would describe it, and the decoding taskof the workers consists in purposefully doing things thatwill approximate to the construction of the described fi-nal state; this is why the house must be in their cognitivedomain. In the second case, the instructions containedin each one of the thirteen identical books do not codea house. They code a process that constitutes a path ofchanging relationships which, if carried through undercertain conditions, results in a system with a domain ofinteractions which has no intrinsic relationship with thebeholding observer. That the observer should call thissystem a house is a feature of his cognitive domain, notof the system itself. In the first case the coding is iso-morphic with a description of the house by the observer,and in fact constitutes a representation of it; in the sec-ond case it is not. The first case is typical of the wayin which the observer codes the systems that he builds;the second corresponds to the way that the genome andnervous system constitute codes for the organism and forbehavior, respectively, and one would never find in thesecodes any isomorphism with the description that the ob-server would make of the resultant systems with whichhe interacts. In what sense could one then say that thegenetic and nervous systems code information about theenvironment? The notion of information refers to the ob-server’s degree of uncertainty in his behavior within adomain of alternatives defined by him, hence the notionof information only applies within his cognitive domain.Accordingly, what one could at most say is that the ge-netic and nervous systems generate information throughtheir self-specification when witnessed by the observeras if in their progressive self-decoding into growth andbehavior.

xi) There are different domains of interactions, and thesedifferent domains cannot explain each other because it isnot possible to generate the phenomena of one domainwith the elements of another; one remains in the same do-main. One domain may generate the elements of anotherdomain, but not its phenomenology, which in each do-main is specified by the interactions of its elements, andthe elements of a domain become defined only throughthe domain that they generate. Any nexus between dif-ferent domains is provided by the observer who can in-teract as if with a single entity with the conjoined statesof nervous activity generated in his brain by his concomi-tant interactions in several domains, or with independentdescriptions of these interactions. Through these con-comitant interactions in different domains (or with sev-eral descriptions within the descriptive domain) the ob-server generates relations between different domains (orbetween different descriptions) as states of neuronal ac-


tivity that in him lead to definite modes of conduct (de-scriptions) that represent these conjoined interactions assingular independent entities. The number and kinds ofrelations the observer can generate in this manner is po-tentially infinite due to his recursive interactions with de-scriptions. Thus, relations, as states of neuronal activityarising from the concurrent interactions of the observerin different domains (physical and relational) constitutethe elements of a new domain in which the observer in-teracts as a thinking system, but do not reduce one phe-nomenological domain into another. It is the simultane-ous logical isomorphism of the new element (relations)with their source systems through their mode of origin(class intersection) that gives the new domain thus gen-erated (descriptions) its explanatory capacity. An expla-nation is always a reproduction, either a concrete onethrough the synthesis of an equivalent physical system,or a conceptual one through a description from whichemerges a system logically isomorphic to the originalone, but never a reduction of one phenomenological do-main into another. An adequate understanding of thisirreducibility is essential for the comprehension of thebiological phenomena, the consensual domains that liv-ing systems generate, and their conjoined evolution.

Many conclusions about self-consciousness and knowl-edge which arise from this mode of analysis have beenproposed in one way or another by scientists and philoso-phers from their intuitive understanding, but never, to myknowledge, with an adequate biological and epistemo-logical foundation. This I have done through the dis-tinction between what pertains to the domain of the ob-server, and what pertains to the domain of the organism,and through carrying to their ultimate consequences theimplications of the circular self-referring organization ofthe living systems: the implications of the functionallyclosed nature of the relativistic organization of the ner-vous system as a system under continuous transforma-tion determined by relations of neuronal activity withoutthe system ever stepping outside itself; and the impli-cations of the non-informative orienting function of lin-guistic interactions. It is only after this has been donethat the functional complexity of the living and linguisti-cally interacting system can be properly grasped withoutits being concealed through such magic words as con-sciousness, symbolization, or information. Most of thedetailed work is yet to be done, of course, but the funda-mental first step of defining the perspective from whichto look has here been taken. As a final remark, one couldsay what appears to be another paradox, but which pointsto the conceptual problem:

Living systems in general, and their nervous system inparticular, are not made to handle a medium, although ithas been through the evolution of their handling of theirmedium that they have become what they are, such thatwe can say what we can say about them.

Post Scriptum

No scientific work should be done without recognizingits ethical implications; in the present case the followingdeserve special attention:

i) Man is a deterministic and relativistic self-referringautonomous system whose life acquires its peculiar di-mension through self-consciousness; ethic and moral-ity arise as commentaries that he makes on his behav-ior through self-observation. He lives in a continu-ously changing domain of descriptions that he gener-ates through recursive interactions within that domain,and which has no other constant element in its histori-cal transformation than his maintained identity as an in-teracting system. That is, man changes and lives in achanging frame of reference in a world continuously cre-ated and transformed by him. Successful interactions di-rectly or indirectly subservient to the maintenance of hisliving organization constitute his only final source of ref-erence for valid behavior within the domain of descrip-tions, and, hence, for truth; but, since living systems areself-referential systems, any final frame of reference is,necessarily, relative. Accordingly, no absolute system ofvalues is possible and all truth and falsehood in the cul-tural domain are necessarily relative.

ii) Language does not transmit information and its func-tional role is the creation of a cooperative domain of in-teractions between speakers through the development ofa common frame of reference, although each speaker actsexclusively within his cognitive domain where all ulti-mate truth is contingent to personal experience. Sincea frame of reference is defined by the classes of choiceswhich it specifies, linguistic behavior cannot but be ratio-nal, that is, determined by relations of necessity withinthe frame of reference within which it develops. Con-sequently, no one can ever be rationally convinced of atruth which he did not have already implicitly in his ulti-mate body of beliefs.

iii) Man is a rational animal that constructs his ratio-nal systems as all rational systems are constructed, thatis, based on arbitrarily accepted truths (premises); beinghimself a relativistic self-referring deterministic systemthis cannot be otherwise. But if only a relative, arbitrar-


ily chosen system of reference is possible, the unavoid-able task of man as a self-conscious animal that can bean observer of its own cognitive processes is to explic-itly choose a frame of reference for his system of values.This task he has always avoided by resorting to god as anabsolute source of truth, or to self-delusion through rea-son, which can be used to justify anything by confusingthe frames of reference and arguing in one domain withrelations valid in another. The ultimate truth on which

a man bases his rational conduct is necessarily subordi-nated to his personal experience and appears as an actof choice expressing a preference that cannot be trans-ferred rationally; accordingly, the alternative to reason,as a source for a universal system of values, is aestheticseduction in favor of a frame of reference specifically de-signed to comply with his desires (and not his needs) anddefining the functions to be satisfied by the world (cul-tural and material) in which he wants to live.


Biology of Cognition1] Maturana (1970).pdf · Biology of Cognition Humberto R. Maturana (1970)...

Documents

Transcript of Biology of Cognition1] Maturana (1970).pdf · Biology of Cognition Humberto R. Maturana (1970)...