Re-inventing the already punctured wheel: reflections on a seminar with Humberto Maturana
Birch J. (1991) Re-inventing the already punctured wheel: reflections on a seminar with Humberto Maturana. Journal of Family Therapy 13: 349–373. Available at http://cepa.info/2798
Table of Contents
Autopoiesis: the organization of the living
Autopoiesis: and a self-refuting paper
The autopoietic unity: and a circular definition
The autopoietic unity: an already refuted assertion
The biology of cognition
Structure and organization: a prior description
Arguments in favour of the autonomous nervous system
Further experimental rejection of autonomy
Transjunctional values: the cybernetic ontology of Gunther
The ontological map of Maturana: objectivity-in-parenthesis
The comparison with Gunther
So what is happening in family therapy?
Three central themes of Maturana’s work – autopoiesis, the biology of cognition, and cybernetic ontology – are examined. Evidence is offered that Maturana’s treatment of these themes is either unoriginal or flawed. The uncritical acceptance of Maturana’s work by family therapists raises questions about the maturity of their discipline, especially in so far as many practitioners claim an understanding of cybernetics.
I have for some years taken an interest in the connections between cybernetic theory and the practice of family therapy. As a consequence I have followed developments in Maturana’s work, both through his own writings, and through commentators such as Parry (1984), Dell (1985), Tomm (1986) and Leyland (1988). In 1988 I agreed to write a report on a seminar with Maturana, an event which prompted me into attempting this formal critique of his work.
Maturana’s work, and his work in collaboration with Varela, became generally available with the publication in 1980 of Autopoiesis and Cognition: The Realization of the Living. The volume contained two essays, ‘The Biology of Cognition’ by Maturana, which is dated 1970, and ‘Autopoiesis: The Organization of the Living’ by Maturana and Varela, which is dated 1973. It is clear from the Introduction to the 1980 edition that Maturana seeks to present his theories as an internally consistent world view. Autopoiesis supplies the framework for the theory of cognition, which in turn provides the starting point for the ontological theory.
Autopoiesis: the organization of the living
Autopoiesis is the answer to one of Maturana’s two fundamental questions – ‘What is the organization of the living?’
Autopoiesis: and a self-refuting paper
Maturana proposes that a fundamental understanding of the process of life can only come from a consideration of the organization of living systems – not from knowledge of the nature of their components. The joint paper with Varela and Uribe in 1974 makes it clear that the original project was to specify the network of interactions of the components of a living system, as opposed to defining necessary properties of those components. The point was made and re-made that it was organization and not materiality which was fundamental to this novel understanding of living systems. This was an exciting departure, since it fitted well with cybernetic theory, which was already claimed to be a theory of organization whose truths were independent of materiality (Ashby, 1956).
Maturana asserts (Varela et al., 1974) that autopoietic (literally ‘self-’making’) organization, in physical space, is the organization of the living. A hierarchy of conditions is offered, each of which has to be satisfied, before an organization can be deemed autopoietic:
1. Can a unity be discerned? (i.e. is there a ‘thing’ about which the observer can make further observations?)
2. Can constitutive elements of the unity be discerned? (i.e. can one identify ‘parts’ to the thing?)
3. Are there regularities of relationships between the components? (i.e. do the parts interact in a regular fashion – is the thing a machine?)[Note 1]other’-`making’) machines – machines which produce something other than themselves.
4. Are the boundaries constituted out of the relationships between boundary components? (i.e. does the mechanism include the observed boundary of the thing?)
5. Are the boundary components produced as a consequence of the internal relations of the entity? (i.e. are the boundary components generated from the mechanism?)
6. Are all other components either constituted or are they necessary permanent constitutive components?[Note 2]
An autopoietic system thus defined continuously generates and specifies its own organization through its operation as a system of production of its own components. Such a system is also living, since the equivalence of autopoietic systems and living systems is explicit in this theory – ‘… for a system to have the phenomenology of a living system it suffices that its organization be autopoietic’ (Varela et al., 1974, emphasis added).
A second part of the 1974 paper offers a ‘simple embodiment of the autopoietic organization’, which ‘may permit the development of formal tools for the analysis and synthesis of autopoietic systems’. As the authors imply by this comment, the computer model appears to fulfil all the requirements for an autopoietic system as laid out elsewhere in their paper.
What now would be needed to refute Maturana’s proposition that autopoietic organization is sufficient to produce the phenomenology of the living? Clearly the most direct refutation would be to call forth an example of physical autopoiesis which is not a living system. Consideration of the claims made in the 1974 paper yields a crucial point. Recalling that the authors regard their computer example as an ‘embodiment of the autopoietic organization’ (emphasis added), then the thesis that autopoiesis suffices as the organization of the living must be false, since the computer example has a physical autopoietic organization yet is manifestly non-living. This is the only example of a self-refuting paper of which I have knowledge. Students of reflexivity may know of others.[Note 3]
The present paper attends to autopoiesis as a description which leads to an understanding of cognition, and thus to an ontology. The status of autopoiesis as a description of living systems is peripheral to this author’s purpose. The critique of the 1974 paper is included to demonstrate how an argument which falsifies itself can nevertheless survive if not examined.
The autopoietic unity: and a circular definition
Maturana asserts (Maturana and Varela, 1980; p. 80) that autopoietic organization has four ‘paramount’ consequences. The fourth assertion is the basis for much that he has to say on cognition, but it does not hold under scrutiny (not that his first three assertions are robust). Maturana claims: ‘Autopoietic machines do not have inputs or outputs.’ He contrasts this with the condition of allopoietic machines. Allopoietic machines can have inputs and outputs, but an autopoietic machine can be ‘… perturbated by independent events and undergo structural changes which compensate these perturbations’.
In Maturana’s view a motor with a speed governor, an allopoietic machine, has its load as an input and its power as an output. An increase in load leads to an increase in power, and vice versa. By contrast, a light-sensitive protozoan which moves away from shadow or from over-bright conditions is to be seen as undergoing perturbations. The light level is not to be seen as ‘input’, nor the movements as ‘output’.
Although this is presented as a new distinction, Maturana offers no method to distinguish inputs and outputs from perturbations. The local rise in temperature after resetting a thermostat (allopoietic) remains indistinguishable from that which follows my contracting a fever (autopoietic): warm is warm. How is it that the handling of the thermostat control is an input, whereas my virus is a perturbation? There is no valid rule put forward, but Maturana’s position seems to be:
the thermostat control is an input, and the rise in temperature is an output, because the thermostat system is an allopoietic machine (and the thermostat system is an allopoietic machine because it has inputs and outputs);the febrile illness is a perturbation because I am an autopoietic machine, (and I am an autopoietic machine because I am perturbed).
These are circular definitions, and not satisfactory.
The autopoietic unity: an already refuted assertion
Even if one wished to allow the fourth assertion as a hypothesis there is another problem. Maturana (with Varela, 1980; p. 78) further asserts of autopoietic machines that they are ‘homeostatic machines and all feedback is internal to them’. He continues: ‘If one says that there is a machine M, in which there is a feedback loop through the environment so that the effects of its output affect its input, one is in fact talking about a larger machine M’ which includes the environment and the feedback loop in its defining organization.’
The machine M, entirely with internal feedback, is pivotal to the radically closed system which Maturana presents as the basis of cognition. Unfortunately for Maturana, Ashby (1962) has already demonstrated that a wholly self-referential system cannot exist as such, but will always require interaction with at least one external variable if it is to maintain its internal organization. (Ashby assumes that the external variable is a source of information, not energy.) So long as Ashby’s demonstration remains unchallenged, Maturana’s claim of the possibility of radical closedness (machine M’) remains void.
The biology of cognition
Maturana’s second principal question in the 1980 essay is: ‘What takes place in the phenomenon of perception?’, or, ‘What is the biology of cognition?’
To follow the arguments on cognition requires an understanding of the concept of structure-determinism, often treated by commentators (e.g. Dell, 1985; Leyland, 1988) as if it were an original concept.
Ashby (1952) has already covered the same ground in his general discussion of machines. He assumes that all matter – whether in the environment, or part of a living system, or in the nervous system – is strictly determinate. If on separate occasions a universe is brought to the same state then the same behaviour will follow. Ashby defines a system as an arbitrary set of variables – a set nominated by the observer, and thus a subset of all the available variables in the ‘real’ machine or organism. He defines the state of the system as the set of values of its variables at a given instant. He defines a line of behaviour as a succession of states over time. A system which always follows the same line of behaviour when started from the same initial state is an absolute system. An absolute system is thus state-determined. The structure-determined system of Maturana is the special case of the state-determined system where all variables are notionally included in the system (Ashby calls this ‘a real “machine” ‘). Following Ashby the given state determines what the following state will be; following Maturana the given structure determines what the following structure will be.
Structure and organization: a prior description
Maturana places stress on the distinction between structure and organization. Structure within a system can vary from instant to instant whilst the organization which the observer construes within the system is seen to maintain constancy, or constant relations between sub-systems of the system. For example, the organization which the observer identifies as the domestic cat relies for its continuing existence on certain essential variables remaining within certain limits – blood pressure, nutrient intake, body temperature, degree of trauma, levels of micro-organisms in its tissues, and so on: the structure is the set of the values of these variables at a given instant. The sub-systems which maintain these values and the interactions which constitute them as sub-systems are viewed as the organization of the cat. Should the essential variables fall outside their limits then the organization will lose its cat-like qualities for an observer. So, after a serious collision with a motor car (the cat’s organization would normally be seen as generating behaviour to avoid such a collision) we can predict the possibility of falling blood pressure, rising micro-organism levels, and so on, to a point where cat organization can no longer be discerned in the structure. Within limits, structure can be very variable, and yet the given organization can still be construed within it. Compare the cat asleep with the cat pouncing on prey. Blood pressure, heart rate, muscular activity, and so on, will reflect different structures all of which nevertheless invite identification as the same cat organization.
Maturana and Varela (1980; p. xix) comment that, at the time of their original papers, ‘we were just beginning to realize the fundamental distinction between organization and structure, and we do not separate the terms with complete rigor’. Although they present their exploration of the relationship between structure and organization as novel, they could have referred to Ashby (1962) where they would have found the relationship adequately examined with greater generality, and with rigour.
Ashby adds a new dimension to the discussion. To ‘elementary variables’ (i.e. structure) and organization he adds all the possibilities of elementary variables which might be imagined: this third domain he calls the ‘product-space’. This third domain I take to be the place where I locate what might be the structure (not the notion) of unicorns and dodos, as well as the structures of ‘real’ things. Organization can now be understood in terms of the observation of constraint:
::… the presence of ‘organization’ between variables is equivalent to the existence of a constraint in the product-space of possibilities. I stress this point because while, in the past, biologists have tended to think of organization as something extra, something added to the elementary variables, the modern theory, based on the logic of communication, regards organization as a restriction or constraint. [emphasis in original]
Ashby is entirely explicit on the relationship of the observer to the organization (cf. Maturana, ‘Everything said is said by an observer’):
The real world is the subset of what is; the product-space represents the uncertainty of the observer. The product-space may therefore change if the observer changes; and two observers may legitimately use different product spaces within which to record the same subset of actual events in some actual thing. The ‘constraint’ is thus a relation between observer and thing; the properties of any particular constraint will depend on both the real thing and on the observer. It follows that a substantial part of the theory of organization will be concerned with properties that are not intrinsic to the thing but are relational between observer and thing. [emphasis in original]
This brief resume of one part of Ashby’s paper shows that issues of structure and organization had been dealt with rigorously, and with greater complexity, well before the publication of Autopoiesis and Cognition.
Arguments in favour of the autonomous nervous system
Maturana argues that since the organism is a living system and hence autopoietic, it is radically isolated from its environment. The argument continues (in 1988) that cognition depends on the closed network of activity which is the nervous system. The theory is adamant that events in the environment cannot specify changes within the nervous system.
Maturana (1988) illustrates his claim of nervous system autonomy with the coloured shadow illusion; with experimental evidence concerning feature detectors in the ganglion cells of the frog’s retina; and with the experimental evidence on vision in the salamander.
(a) The coloured shadow illusion. When red and white lights illuminate a surface, the observer sees (recognizes, has an awareness of, notices, apprehends) the surface as pink. Shadows are cast opposite each light by an opaque projection from the surface. The observer sees (in the above sense) a red shadow opposite the white light and a green shadow opposite the red light (Figure 1). The red shadow is not a problem: the surface shaded from white light is illuminated only by red light, hence a red shadow. The green shadow is a puzzle though: the surface shaded from red light is illuminated only by white light, yet one sees green. This is the illusion. Maturana declares, ‘Our experience with a world of coloured objects is literally independent of the wavelength composition of the light coming from any scene we look at’ (Maturana and Varela, 1987). Hence, the argument goes, the nervous system is autonomous, and not specified by the environment. The illustration is vivid, but it does not confirm the claim. We can only speak of the illusion of the green shadow because we can reliably predict that this combination of lights and shadows will elicit precisely this illusion for the observer. It is not the case that sometimes the shadow is green, sometimes purple, or grey, or iridescent. It is the case that the shadow is undeniably green, always.
A different combination of lights and shadows will elicit different but equally reliable illusions. What is more, Maturana reports exactly this. He describes (Maturana and Varela, 1987) von Guericke’s demonstration in 1672 of the illusion of blueness in the shadows of a surface illuminated by a candle and the sun, as if it supports his thesis. But if one changes the colours of the illuminating lights one changes the colour of the shadow illusion. The illusion, then, has quite the opposite force to that attributed to it by Maturana – the illusion is seen to be a world of coloured objects dependent on the wavelength composition of the light coming from it – blue shadow opposite a yellow light, and green opposite red. The environment does enter into the specification of the experience. The specification does not have a simple one-to-one relationship, hence the illusion, but Maturana’s position is that there is no specification, which is manifestly wrong.
Whilst addressing colour illusions there is an additional point to be made. Maturana implies (with Varela, 1980; p. xiv) that the idea of attributing colour experience to the state of the nervous system, rather than to spectral qualities of light in the environment, was a new advance. It should be noted that this separation of the colour experience of the observer from the physical spectrum was first made by Goethe. Goethe had become intrigued by what he felt was a defect in Newton’s theory of light. He became convinced that Newton’s ideas could not fully explain the subjective experience of colour, and he began to explore the role of the eye in forming this experience. In 1796 he wrote, in a letter to Schiller, ‘If you wish, it becomes singularly the world of the eye which is conceived through structure and colour.’ If one replaces ‘eye’ with ‘visual system’ this statement becomes one with Maturana, e.g. ‘… we must concentrate on understanding that the experience of colour corresponds to a specific pattern of states of activity in the nervous system …’ (Maturana and Varela, 1987; p. 22). Goethe was to develop these ideas, proposing a distinction between ‘physiological’ and ‘physical’ colours, the former related to the functioning of the eye, the latter to the behaviour of light in the environment. It should be added that there is no implication in Goethe’s theory that the physical colours took no part in specifying the physiological colours.
(b) Edge detectors. Maturana’s second argument is based on findings from ‘What the Frog’s Eye Tells the Frog’s Brain’ (Lettvin et al., 1959). This original work demonstrated the detector cells in the frog’s visual system which discharge whenever an object moves across the animal’s field of view. Maturana (1988) argues that such moving edge detectors are not ‘informed’ by events in the visual field, but are merely ‘triggered’ (i.e. defined as sustaining ‘perturbation’ rather than receiving an ‘input’) by such events. It is suggested that this experiment demonstrates the closed nature of the nervous system which is responding autonomously to external triggers, whilst not being specified by them.
Here the elegance of the experimental work should not obscure the adequately established fact that sensory nerve cells discharge when irritated in specific ways. The organization of receptor and environment taken as a whole constrains the frequency of discharge, which is proportional (within physiological limits) to the intensity of whichever type of event the nerve ending responds to. The same constraints operate on the edge receptor. It is constrained by its nature as an edge receptor in a nervous system, hence it doesn’t respond to heat or pressure, or to an impossibly slow or impossibly fast retinal shadow. It is additionally constrained by the rate of travel of the shadow over the retina and this is clearly stated in the 1959 paper: ‘The frequency of discharge increases with the velocity of the object within certain limits’ (p. 1946). Nervous system and environment taken as a whole constrain, specify, or inform, the activity of the nerve cell. This is not ‘autonomy’ as that term is presented by Maturana, rather than its opposite.
There is a second problem here if we are to maintain the notion that the world is structure-determined, for, if so, a relevant ‘outside’ change must necessarily result in a specific change in the ‘closed’ nervous system. If not, then we have to fall back on some explanatory system, new to science, whereby natural laws can be construed as acting consistently within the organism or throughout the environment, but are apparently without their normal effects in the boundary region (which is a function of the observer’s act of distinction) between organism and environment. Maturana does not appear to propose this, nor would it be remotely plausible to do so, since it would require the abandonment of structure-determinism.
(c) The salamander. The third argument in favour of the closedness of the nervous system comes from Sperry’s experiments with the salamander.[Note 4] The salamander’s nervous system has nerve cells capable of regeneration and reconnection. Sperry avulsed the eyes of salamanders and replaced them, turned through 180°. Once the neuronal connections have been re-established, those areas of the retinae which once received light from the front of the animal now receive it from the rear, and vice versa. The unimpaired salamander, when food is placed within its view, strikes at the food with its tongue. The experimental animals, when food is placed in front of them, turn and strike behind them. That is to say, they appear to locate their food on the basis of which retinal region receives the image of the food. The inference offered for this finding by Maturana is that the presence of food only ‘triggers’ the activity of the salamander’s nervous system, but does not specify it – the specification being effected by the rearranged autonomous nervous system.
As with the coloured shadow illusion, the effect is quite predictable and replicable, so the experimental animals demonstrate that the behaviour of the salamander is specified, at least in part, by the environment. The salamander does not strike at ‘food absences’. Also the choice of illustration begs examination. Maturana does not draw attention to Held’s (1965) experiments with human subjects who were given eyeglasses with prisms which turned their visual input upside down – they saw the world as if standing on their heads. Despite this constraint, subjects were able to accommodate to the complete reversal of images on their retinae, and eventually stopped experiencing their environment as inverted. This experiment demonstrates that the nervous system is not necessarily closed to specification by the environment. Held’s volunteers were behaving contrarily to Sperry’s salamanders: their nervous systems had become re-specified through their interaction with their inverted world. These contrary findings do not address the debate over closedness versus openness to the environment, instead they only draw attention to different degrees of plasticity in neural systems. Presumably the feeding behaviour of the salamander is a ‘hard-wired’ feature of its nervous system – such as a reflex, or the consequence of imprinting. Such a feature could not be expected to be reversed later on in life. To expect the salamander’s feeding behaviour to adapt would be comparable to transposing the knee muscles of Held’s volunteers – reversing their knee-jerk response – and then expecting the subjects to reverse the neuronal architecture of their spinal cords in order to re-establish normal reflexes.
Further experimental rejection of autonomy
Recent work by Sejnowski and Lekhy (1987) in neural network studies provides a new view on ‘What the frog’s eye tells the frog’s brain’. Neural networks simulate, within a computer program, the interactions of ‘cells’ modelled to behave like simple neurones. The cells are arranged as if in interconnected layers, a pattern found in many natural nervous systems. The experimenter can ‘train’ the network by giving positive or negative feedback according to whether the network behaviour is desired or unwanted. Sejnowski and Lekhy constructed a network with three layers of cells – input, output, and an intervening middle layer. They then presented images of surfaces of varying convexity to the input layer. The network was trained to produce a particular output pattern for each different depth of curvature. The network behaved randomly initially, but after many training cycles it was able reliably to produce the required output for a given curvature.
This main result in itself is significant. If we assume that the network models the real behaviour of a visual system then it demonstrates one way in which visual stimuli could contribute to the specification of the structure of such a system. A subsidiary finding was more intriguing. Sejnowski and Lekhy investigated the behaviour of the middle layer cells in the trained system. When moving edge stimuli were offered to the input, the middle layer cells behaved like the moving edge detector cells of the 1959 frog’s eye experiment. This finding questions the notion that an edge detector cell is necessarily a neurophysiological ‘building block’, and allows the view that it may once have been a less specialized cell whose behaviour has been specified by interaction with an outside environment.[Note 5]
The necessity of comprehending animal behaviour as the emerging interaction between nervous system and environment has already been elegantly demonstrated by Held (1965). In his ‘kitten carousel’ experiment he contrived that two kittens, from the time their eyes were open, should have identical visual input to their (broadly identical) nervous systems: they differed only in that one kitten was able to move actively within its environment whilst the other was moved only passively. When finally removed from the constraints of the apparatus the active kitten showed normal visual behaviour. By contrast the passive kitten behaved initially as if it were blind, showing that passive exposure to retinal stimulation had not established a basis for visual perception.
Maturana (Maturana and Varela, 1980; p. 26) places a curious gloss on this experiment:
… it is apparent that the ‘visual handling’ of an environment is no handling of an environment, but the establishment of a set of correlations between effector (muscular) and receptor (proprioceptor and visual) surfaces, such that a particular state in the receptor surfaces may cause a particular state in the effector surfaces that brings forth a new state in the receptor surfaces … and so on.
Maturana is here claiming that the circuit of effector and receptor was interrupted for the ‘blind’ kitten. But this circuit remains intact. Nothing in the experiment prevents the passive kitten from moving or focusing its eyes: the autonomous circuits are clearly preserved. Instead it is the ‘external’ circuit which is broken – the circuit of scene, eye, nervous system, paw, back to scene – the circuit which includes the environment. The experiment establishes beyond doubt that the kitten’s vision can be specified neither by a ‘triggered’ autonomous nervous system alone, nor by a visual environment alone, for if either had been adequate then the passive kitten would have been a seeing kitten when released. Vision requires the interaction of both nervous system and environment.
In drawing attention to the defects in Maturana’s arguments there is no desire to devalue his wish to escape from simplistic stimulus and response models of neurophysiological functioning. This is entirely in accordance with cybernetic understandings pioneered by Bateson (1951) in his seminal work on information and codification. Ashby (1952) draws attention to an older strand of this tradition, where cognition is seen to require the interaction of organism and environment, when he quotes Jennings, from 1906:
In most cases the change which induces a reaction is brought about by the organism’s own movements. These cause a change in relation of the organism to the environment: to these changes the organism reacts. The whole behaviour of free-moving organisms is based on the principle that it is the movements of the organism that have brought about the situation.
On the basis of his understanding of autopoiesis and his theory of cognition, Maturana has extended his thesis into ontology (i.e. the theory of ‘what there is’, as opposed to epistemology, the theory of ‘how we know’). It is instructive to compare what Maturana has to say about ontology with the comments of previous authors in the field, especially since Dell (1985) implies that Maturana has pioneered the debate on ontology within cybernetics.
Cybernetics has come late to ontology: theories of ‘what there is’ have been debated for millennia prior to the establishment of cybernetics. In the sixth century BC the ‘Tao Te Ching’ of Lao Tzu devotes a quarter of its eighty-two chapters to issues of ontology, arguing that the act of discrimination brings into being the ‘ten thousand things’ out of the unknowable matrix of the eternal Tao.
Buddhist teachings also address ontology. The Lankavatara Sutra, dating back to the fifth century AD (Suzuki, 1932), comments: ‘There is nothing that is to be born, nor is there anything that has been born; even causation is not; it is because of worldly usage that things are talked of as existing’ (Chapter 2, verse 144). Nor are such notions exclusively part of eastern philosophy. The first- and second-century Gnostics regarded the world as divided into the pleroma and the creatura, equivalent respectively to the Tao and the ten thousand things. Such Gnostic ideas inspired Jung[Note 6] in writing ‘Septem Sermones ad Mortuos’ (c. 1925) under the pseudonym Basilides:
Differentiation is creation. The created world is indeed differentiated. Differentiation is the essence of the created world and for this reason the created also causes further differentiation. That is why man himself is a divider, inasmuch as his essence is also differentiation.
Transjunctional values: the cybernetic ontology of Gunther
Of contemporary authors coming from a cybernetic background, Gunther (1962) provides a basis for comparison with Maturana’s treatment of cybernetic ontology. Gunther introduces his paper thus:
The following investigation arrives at the result that our present (classical) ontology[Note 7] does not cover ‘everything’. It excludes certain phenomena of Being from scientific investigation declaring them to be of irrational or metaphysical nature. The ontologic situation of cybernetics is characterised by the fact that the very aspect of Being that the ontologic tradition excludes from scientific treatment is the thematic core and centre of this new discipline.
Gunther restricts his argument to the phenomenon of subjective self- reflection. He draws heavily on the philosophical discussions developed around quantum mechanics, arguing that this is the discipline which finally required that classical ontology be transcended. He quotes Heisenberg’s conclusion of 1931 that a radically objective system of physics with the separation of Reality into ‘thing’ and ‘thought’ was now impossible – ‘the radically isolated object has, on principle, no describable properties’.
Gunther observes that quantum mechanics, for the first time in science, required absolutely that the observer be included in any exhaustive description of phenomena. Such ideas have found general currency, for example in the notion that the observer cannot locate a particle in space without losing information about its momentum, and conversely cannot specify the momentum of a particle without losing information about its position in space. Gunther introduces a pair of dichotomies between subject and object which he claims are both required if one is to strive towards a more exact specification of the world. He draws the following schema:
(Classical ontology)D1: Object (Oo)Subject (Ss)D2: Object (Oo)Object (Os) s)D3: Object (Oo) o)Subject (Ss)(Transjunctional ontology)
Note: The indices should be read ‘as if; hence (Os) should be read as ‘Object as if subject’.
Dichotomy 1 (D1) represents the subject/object dichotomy of classical ontology. Both categories are distinct and admit no interpenetration – all mind or all matter, but never both. Material entities are deemed to have an independent existence and are represented in a non-physical, mental realm by thought entities. In this view the pursuit of a truthful reality is the striving for an even greater correspondence between the entities of mind with those of matter. This was the paradigm which could no longer be sustained as the new physics emerged in the early years of the century. Gunther introduces the paired dichotomies, D2 and D3, to restore the possibility of a specification of reality. Dichotomy 2 does allow for the specification of an object, provided that part of the objective world always remains within the subject world and is forbidden access to the objective world. This indeterminate world, the ‘object as if subject’, is the means of the observer’s observing. Whilst it is permitted to the observer to objectify the means of observing of another observer (examining behaviour, sense organs, brain discharges, etc.), the means of the observer’s observing as object are simultaneously the means of the observer’s observing as subject. The observer’s eye can never finally observe itself observing, for in observing the observing eye it fails to observe the self-observing eye. The attempt to retrieve the situation by now observing the self-observing eye necessarily fails to observe the eye which observes the self-observation, and so on ad infinitum. Objectification can only go so far – there is always an objective mystery which eludes the observer, yet which is undeniably object in terms of Dl: this is the object-as-if-subject.
Dichotomy 3 draws attention to the obverse phenomenon – that there must always be an element of the classical (D1) subject which is excluded from subjectivity (D3). Our knowing – seeing, hearing, and touching included – is immediate and apparently certain, but the roots of our knowing are inscrutable. How we punctuate the pleroma, what creatura we select out of it in terms of immediate experience, is rigorously constrained by what we already know; and what we already know we only knew because there was an earlier knowing, and a yet earlier knowing, and so on. These ulterior knowings which inform our experiencing are, broadly, ‘ideas’ or the stuff of subjectivity (D1), yet their influence on our subjective (D3) world is as if they were ‘objects’ on the margins of our consciousness through which we discover the world. We can take this subject-as-if-object (S°) as being Bateson’s (1951) ‘network of value’ which, he argues, partly determines the ‘network of perception’. These notions can be seen as the culmination of an earlier enquiry into learning where Bateson (1942) asks, ‘How does the dog acquire a habit of punctuating or apperceiving the infinitely complex stream of events (including his own behaviour) so that this stream appears to be made up of one type of short sequences rather than another?’ The theme of punctuation, which was developed further by Bateson (1964) and extended by Watzlawick et al. (1967), is essentially the theme of ulterior knowings which partially determine our current knowings.
Gunther argues that our attempt at a comprehensive specification of reality requires that we describe the world in terms of both D2 and D3. The wholly objective or wholly subjective worlds of DI are delusions, and our descriptions instead must find a new position. This is neither wholly subjective nor wholly objective, but ‘trans- junctional’, i.e. transcending the classical junction which supposedly divides subject from object.
The ontological map of Maturana: objectivity-in-parenthesis
Maturana’s ‘map’ or diagram (Figure 2) attends to the dichotomy between ‘objectivity’ (M)[Note 8] and ‘objectivity-in-parenthesis’. The dichotomy arises from the rejection or acceptance of the question, ‘How is it that I, the observer, am observing what I am observing?’ If we reject the question we are ‘in’ objectivity(M), if we accept it we are ‘in’ objectivity-in-parenthesis.
When Maturana speaks of objectivity he implies a concurrent subjectivity (Maturana, 1988). Since objectivity(M) is said to imply a concrete universal reality separate from the observer, we can be easily satisfied that objectivity(M) and its implicit subjectivity(M) constitute the classical object/subject dichotomy (D1). How then is one to understand the nature of the advance which Maturana and his commentators believe is generated by objectivityin-parenthesis? If that term must include all understandings which are not objectivity(M), then it must stand for everything under the line on Gunther’s ‘ontological map’. Has Gunther created, in D2 and D3, a pair of unnecessary entities (in William of Ockam’s sense) both of which Maturana subsumes under a single category, without loss of meaning?
For an examination of objectivity-in-parenthesis I will rely on Maturana’s (1986) written explication. The quotations below come from paragraph (ii), entitled ‘Objectivity in parenthesis’, of section (5), also entitled ‘Objectivity in parenthesis’. I reproduce Maturana’s argument in full (A), step by step, commenting on each statement (C)
(1A) ‘The assumption of objectivity is not needed for the generation of a scientific explanation.’
(1C) First, care is needed with the use of ‘objectivity’ here. Two or more people negotiating on how they are to explain their experience assumes objectivity in some sense. It is hard to imagine what kind of science does not involve debating a credible objectivity, for science is, by Maturana’s own definition (see below), just such an enterprise.
Second, Maturana argues in a preceding paragraph that his explanations require no recourse to a notion of objectivity, taking it as ‘constitutive that existence[Note 9] is dependent on the biology of the observer’. He apparently fails to notice that by calling on the biology of the observer he is undeniably introducing a version of objectivity. (In Gunther’s terms this ‘biology of the observer’ would appear to be the object-as-if-subject, Os). Maturana’s contradictory assertion that he makes no reality claims through the very device of making a reality claim is argued in full by Held and Pols (1987a, b) in a debate with Dell (1987).
(2A) ‘Therefore, in the process of being a scientist explaining cognition as a biological phenomenon I shall proceed without using the notion of objectivity to validate what I say, that is, I shall put objectivity in parenthesis.’ [Emphasis in original]
(2C) ‘Therefore’ shows that Maturana knows that the claim that he will ‘proceed without the notion of objectivity’ depends on the validity of assertion 1A. But 1A has been shown to be flawed, so Maturana’s claim that he has validly excluded objectivity from consideration is not substantiated.
(3A) ‘In other words, I shall go on using an object language _because this is
the only language we have (and can have) …’ _[Emphasis added]
(3C) Maturana confirms the points made in 1C and 2C – that he has failed to establish a new manner of talk which allows him to abandon objectivity. When he says, ‘… I shall go on using an object language …’ he is making a virtue of necessity – or perhaps he fails to notice that his new distinction has the contradictory qualities of being (to him) crucial and, simultaneously, un-realizable.
(4A) ‘… but although I shall use the experience of being in language as my starting point whilst I use language to explain cognition and language, …’ [Emphasis added]
(4C) I doubt whether Maturana aims to be confused here; it seems more likely that he is unaware of the importance of avoiding category- mistakes when philosophizing (Ryle, 1949).[Note 10] All three usages of ‘language’ in statement 4A require scrutiny.
(a) ‘… the experience of being in language …’ implies a subjectivity definition of language, requiring a subjective discussion, i.e. How do I find myself moved by this language? What is it like to know this plea as ‘weak’, or that as ‘strong’? How is hearing this as ‘poetry’, or that as ‘banality’, different for me? (Remembering that ‘find’, ‘know’ and ‘hearing’ should be taken neither to refer to my mechanism as an organism, nor to a quasi-mechanical mind substance.)
(b)‘… whilst I use language to explain …’ implies vocalizations, or silent rehearsals of vocalizations whilst interacting with an other. This calls for a discourse on syntax, semantics, or pragmatics – What are the redundancies observed in a listener’s behaviour in conjunction with given speech patterns? What is the variety of messages available to the speaker, and what variety is there in the discriminations the listener demonstrates? These questions come from the realm of objectivity, but are still ‘language’.
(c) ‘… to explain cognition and language …’ leaves ‘language’ unspecified and thus might be calling on ‘explain’ to function in an objective sense, or a subjective sense, or perhaps both (cf. Ryle’s example ‘She came home in a flood of tears and a sedan chair’, where ‘in’ is similarly overloaded). Given these ambiguities, one is left uncertain as to what force this dependent clause is to carry with the main clause 5A, which follows.
(5A) ‘… I shall not claim [emphasis in original] that what I say is valid because there is an objective reality independent that validates it.’ [Probably ‘… objective independent reality …’]
(5C) Disregarding objection 4C, conclusion 5A is internally flawed anyway. An objective independent reality can’t be argued to have validated anything for anyone, since all objective independent reality claims are themselves unverifiable (because verification requires a prior knowledge of an objective independent reality, itself unverifiable, and so on.) Maturana produces a straw man here; if a cybernetician exists who claims a world view validated by an objective independent reality then he or she is a relic from an epistemological Stone Age.
(6A) ‘I shall speak as a biologist, and as such I shall use the criterion of validation of scientific statements to validate what I say …’
(6C) Provided that there is no implication that Maturana the biologist is dispensing with objectivity (1A and 1C) there need be no problem with this statement. Maturana’s definition of the criteria of validation of scientific statements are:
(a) Describing the phenomenon to be explained in a way acceptable to a body of observers.
(b) Proposing a conceptual system capable of generating the phenomenon to be explained in a way acceptable to a body of observers (explanatory hypothesis).
(c) Obtaining from (b) other phenomena not explicitly considered in that proposition, as also describing its conditions for observation by a body of observers.
(d) Observing these other phenomena obtained from (b) [an arrow
leads back from here to (a)] (Maturana and Varela, 1987).
(7A) ‘…accepting that everything that takes place is brought forth by the observer in his or her praxis of living as a primary experiential condition …’
(7C) Cf. Spencer-Brown (1969), ‘… our understanding of [the] universe comes not from discovering its present appearance, but in remembering what we originally did to bring it about.’[Note 11]
(8A) and that any explanation is secondary.’
(8C) Cf. Ashby (1952), ‘… the fact of the existence of consciousness is prior to all other facts.’
This concludes Maturana’s definition of ‘objectivity-inparenthesis’. As well as being inconclusive it also appears to assimilate, without acknowledgement, ideas already available. When Maturana (1986) comments. objectivity-in-parenthesis entails accepting that existence is brought forth by the distinctions of the observer …’, the identity of this notion with Spencer-Brown’s brought-about universe is plain. And Spencer-Brown acknowledges the venerable origins of these ideas by prefacing his work with a quotation from Lao Tzu:
The nameless is the beginning of heaven and earth.
The comparison with Gunther
Objectivity-in-parenthesis, as defined and as used, attends to Gunther’s dichotomy 3 – the knower as divider, the knower as universe creator. Dichotomy 2 is not to be found in objectivity-inparenthesis. Maturana does allude to D2 in his reality claim of the biology of the observer. But even this view might be over-generous. The biology of the observer is presented with the certainty which belongs to classical (D1) objectivity. It is not presented as a partial objectivity subsuming the subjectivity of the observer. It is as if the particle physicist, with fingers crossed behind his back, has announced with certainty the position and velocity of an electron.
Maturana’s exposition fails to challenge Gunther’s requirement that a cybernetic ontology has to be transjunctional – encompassing D2 and D3, and identifying object-as-if-subject and subject-as-ifobject. Bateson’s epistemological ponderings, flawed though they might be (Dell, 1985), still stand as the ideas closest to Gunther’s ontological demands. ‘Form, substance, and difference’ (Bateson, 1970) is perhaps the clearest early statement of Bateson’s struggle with the notion of ‘immanent mind’ as a solution. Here Gunther’s cool logic is contrasted by Bateson’s intensity: ‘It is the attempt to separate intellect from emotion that is monstrous, and I suggest that it is equally monstrous – and dangerous – to separate the external mind from the internal. Or to separate mind from body’ (p. 438).
I have argued that, on autopoiesis,
1. An early paper on autopoiesis refuted its own claim that autopoietic organization in physical space is sufficient to produce the phenomenology of living systems.
2. The later theory, of molecular autopoiesis, restores the specification of the nature of components, thus reintroducing the requirement which the original theory celebrated excluding.
3. The assertion that autopoietic machines relate to the environment through ‘perturbations’ rather than inputs and outputs is based on a circular definition.
4. The notion of a self-regulating entity which is organizationally closed, is already refuted.
5. Structure-determinism is cognate with state-determinism, a concept which was already described in the cybernetic literature.
6. The presentation of the relationship between structure and organization is cursory compared with Ashby’s earlier treatment.
7. The notion of an autonomous entity, in Maturana’s radical sense, is incompatible with structure-determinism.
8. Experimental ‘evidence’ for autonomy can be used to refute autonomy.
9. Experimental evidence is available which shows that nervous systems are necessarily partly specified by their environment.
10. The definition of ‘objectivity-in-parenthesis’ is inadequate.
11. ‘Objectivity-in-parenthesis’, as used, is comparable with other theories of knowing and being going back twenty-six centuries.
12. Gunther’s system of cybernetic ontology was available prior to, and remains more comprehensive than, that of Maturana.
So what is happening in family therapy?
The above appraisal of three basic strands of Maturana’s work raises wider questions. Within family therapy, commentaries on Maturana have repeatedly accepted his work as if it were flawless, or original. The possibility that parts of the Maturana oeuvre might even be absurd seems to have been overlooked. Critiques are sparse. Beyond Held and Pols (1987a, b) there are few others. Simon (1985), in an interview, is necessarily short on detailed theoretical discussion, but shows his reluctance to be wooed. Stagoll (1986) focuses on the moral implications of the ideas from Santiago but without calling the theory behind the ideas into question, at least in a systematic way. Finally Colapinto (1985) argues that to analogize from the cellular to the social level must be open to suspicion. He points out the following paradox: Maturana concludes that the passion (as a therapist) to change others is misconceived, whereas in his writings and presentations Maturana does show a passion (as a philosopher) to change others.
Why so little enquiry, so few questions?[Note 12] Even Keeney’s (1979) report of Bateson, ‘… the centre for this study is now in Santiago, Chile under a man named Maturana’, seems generally to be taken at face value, without reference to Bateson’s fondness for irreverence and quirky humour. A little more irreverence and quirkiness might be the antidote to the surprising expectation that an experimental neurophysiologist can offer a world view which will encompass our discipline and answer our problems. Is it so easy to forget that each one of us, as therapists, still has the personal responsibility to find our own unique path through therapy? The Maturana phenomenon suggests that the longing for someone else to do it for us is not easy to give up. And if this is the moral relationship we set up with our teachers, will this not in turn define the moral relationship we set up between our clients and ourselves? And do we really want to become our clients’ own mini-superstars – revered and beyond criticism? Stagoll and Colapinto are surely right. The debate is never merely academic, but always moral and political too.
Ashby, W. R. (1952) Design for a Brain. New York: Wiley and Sons.
Ashby, W. R. (1956) Introduction to Cybernetics. London: Chapman and Hall.
Ashby, W. R. (1962) Principles of the self-organizing system. In: H. von Foerster and G. W. Zopf (eds), Principles of Self-Organization. New York: Pergamon. (Reprinted in W. Buckley (ed.) (1968), Modern Systems Research for the Behavioural Scientist. Chicago: Aldine.)
Atkins, P. W. (1981) The Creation. Oxford: Freeman.
Bateson, G. (1942) Social planning and the concept of deutero-learning. Ch. 4 of Science, Philosophy and Religion, Second Symposium. New York: Harper and Row.
Bateson, G. (1951) Information and codification: A philosophical approach. In: J. Ruesch and G. Bateson, Communication: The Social Matrix of Psychiatry. New York: W. W. Norton.
Bateson, G. (1964) The logical categories of learning and communication. In G. Bateson (1972), Steps to an Ecology of Mind. Frogmore, St Albans: Paladin.
Bateson, G. (1970) Form, substance and difference. In G. Bateson (1972), Steps to an Ecology of Mind. Frogmore, St Albans: Paladin.
Colapinto, J. (1985) Maturana and the ideology of conformity. Family Therapy Networker, 9: 29-30.
Dell, P. (1985) Understanding Bateson and Maturana: Towards a biological foundation for the social sciences. Journal of Marital and Family Therapy, 11: 1-20.
Dell, P. (1987) Maturana’s constitutive ontology of the observer. Psychotherapy, 24: 462-466.
von Goethe, J. W. (1796) Letter to Schiller, 15 November. In: R. Matthaei (1971), Goethe’s Colour Theory. London: Studio Vista.
Goldberg, D. and David, A. S. (1991) Family therapy and the glamour of science. Journal of Family Therapy, 13: 17-30.
Gunther, G. (1962) Cybernetic ontology and transjunctional operations. In: M. C. Yovits, G. T. Jacobi and C. D. Goldstein (eds), Self-Organizing Systems. Washington, DC: Spartan.
Held, B. S. and Pols, E. (1987a) Dell on Maturana: A real foundation for family therapy? Psychotherapy, 24: 455-461.
Held, B. S. and Pols, E. (1987b) The philosophy of Dell and Maturana. Psychotherapy, 24: 466-468.
Held, R. (1965) Plasticity in sensory-motor systems. Scientific American, 213: 84-94.
Jennings, H. S. (1906) Behaviour of the lower organisms. New York. (Reprinted in W. R.Ashby (1952), Design for a Brain. New York: Wiley and Sons.)
Jung, C. G. (c.1925) VII Sermones ad Mortuos. In: S. Hoeller (ed.) (1982), The Gnostic Jung and the Seven Sermons to the Dead. Wheaton, Illinois: Theosophical Publishing House.
Keeney, B. P. (1979) Glimpses of Gregory Bateson. Pilgrimage: The Journal of Existential Psychology, 7: 17-44.
Lao Tzu (1972) Tao Te Ching, trans. by G. Feng and J. English. New York: Random House.
Lettvin, J. Y., Maturana, H. R., McCulloch, W. S. and Pitts, W. H. (1959) What the frog’s eye tells the frog’s brain. Proceedings of the IRE, 47: 1940-1959.
Leyland, M. L. (1988) An introduction to some of the ideas of Humberto Maturana. Journal of Family Therapy, 10: 357-374.
Linsker, R. (1988) Development of feature-analysing cells and their columnar organization in a layered self-adaptive network. In: R. M. J. Cotterill (ed.), Computer Simulation in Brain Science. Cambridge: Cambridge University Press.
Maturana, H. R. (1986) The Biological Foundations of Self-Consciousness and the Physical Domains of Existence. Unpublished manuscript.
Maturana, H. R. (1988) Seminar. Adelaide.
Maturana, H. R. and Varela, F. J. (1980) Autopoiesis and Cognition: The Realization of the Living. Dordrecht: Reidel.
Maturana, H. R. and Varela, F. J. (1987) The Tree of Knowledge. Boston: New Science Library.
Parry, A. (1984) Maturanation in Milan: Recent developments in family therapy. Journal of Systemic and Strategic Therapies, 3: 35–43.
Ryle, G. (1949) The Concept of Mind. Reprinted (1973), Harmondsworth: Penguin.
Sejnowski, T. and Lekhy, S. (1987) Presentation at the Society for Neuroscience 17th Annual Meeting, New Orleans. Reported by A. Anderson (1988), Learning from a computer cat. Nature, 331: 657-658.
Simon, R. (1985) An interview with Humberto Maturana. Family Therapy Networker, 9: 36 43.
Spencer-Brown, G. (1969) Laws of Form. Reprinted (1979), New York: Dutton.
Stagoll, B. (1986) Epistemology and torture. Australian and New Zealand Journal of Family Therapy, 7: 97-101.
Suzuki, D. T. (1932) The Lankavatara Sutra: A Mahayana Text. London: Routledge and Kegan Paul.
Tomm, K. (1986) Seminar. Adelaide.
Varela, F. J. (1975) A calculus for self-reference. International Journal of General Systems, 2: 5-24.
Varela, F. J., Maturana, H. R. and Uribe, R. (1974) Autopoiesis: The organization of living systems, its characterization, and a model. Biosystems, 5: 187-196.
Watzlawick, P., Beavin, J. and Jackson, D. (1967) Pragmatics of Human Communication. New York: W. W. Norton.
’Machine’ is used technically here, as indicating a specific grouping of components which interact in a predictable fashion. If the whole is repeatedly set in motion with a given set of initial conditions it will always repeat the same set of consequent actions. In cybernetics, ’machine’ is used to describe both living and non-living systems, whereas in common usage the word generally refers to non-living artefacts. Maturana calls these latter commonplace machines allopoietic’ (i.e. other’-`making’) machines – machines which produce something other than themselves.
Mention of ’necessary permanent constitutive components’ foreshadows Maturana’s later reversion to a specification of life which requires (rather than rejects) that the nature of components be known.
Maturana (1988) has revised this view. Now living systems arc autopoictic molecular (i.e. not merely physical) systems. The theory now reintroduces the specification of components which it had originally celebrated excluding. On the issue of theories of living systems Maturana is back in step with those authors who specify molecules in their explanations, authors like Atkins (1981), who suggests: ’Once molecules have learnt to compete and to create other molecules in their own image, elephants and things resembling elephants, will in due course be found roaming through the countryside.’
I rely on Maturana’s account for the details of Sperry’s work.
Neural network studies should not be assumed to be promoting a tabula rasa explanation, where the nervous system is specified exclusively by the environment. Experiments by Linsker (1988) with multilayered systems have demonstrated that such networks can develop columns of orientation-specific cells (a feature of natural visual system architecture) even in the absence of environmental input.
And in turn Bateson (1970), although he used the terms inconsistently.
’Classical ontology’ is Gunther’s term for the notion that our logical or thought processes (Internal’ reality) necessarily mirror events in a physical world (`external’ reality).
I will add ‘M’ in parentheses to any terms which come from Maturana’s thesis and which might otherwise be ambiguous as to their origin.
How we are to understand ‘existence’ is unspecified. See later note on category-mistakes.
Readers unfamiliar with the term ’category-mistake’ are referred to 16ff. of The Concept of Mind. One of Ryle’s examples of the error is of the visitor to the university who, unfamiliar with the usage of ’university’, says to his host, ’Thank you for showing me the colleges, the senate building, the library, and the student’s union. Now could you please show me the university, for that is what I had particularly wanted to see.’ The visitor has failed to understand that `university’ is in a logical category distinct from ’college’, ’library’, etc. ’Seeing’ the university means seeing many things, not just one. Rylc argues that when dealing with concrete issues, as with this example, we can generally notice the absurdity of category-mistakes, but when we deal with abstractions the absurdity is not always evident. On ’existence’, for example, Ryle says, ’It is perfectly proper to say, in one logical tone of voice, that there exist minds and to say, in another logical tone of voice, that there exist bodies. But these expressions do not indicate two different species of existing, for “existence” is not a generic word like “coloured” or “sexed”. They indicate two different senses of “exist”, somewhat as “rising” has different senses in “the tide is rising”, “hopes arc rising”, and “the average age of death is rising”. A man would be thought to be making a poor joke who said that three things are now rising, namely the tide, hopes, and the average age of death.’
This work is Spencer-Brown’s exposition of his Calculus of Indications. Varela (1975) has made a study of the work and has published an extension of the calculus..
Since this paper was completed, Goldberg and David (1991) have published their explanation for this question. See especially the last paragraph of p. 19.
Found a mistake? Contact corrections/at/cepa.infoDownloaded from http://cepa.info/2798 on 2016-06-16 · Publication curated by Alexander Riegler