Beyond Cognitivism: Mutualism and Postmodern Psychology.
Abstract. Postmodern science is shifting from mechanism toward organicism, from the metaphysics of Descartes toward those of Whitehead. However, cognitivism, which has until recently been the predominant paradigm of Western psychology, retains an implicitly Cartesian view. This is restricting the investigation of consciousness.
The paper sketches an alternative paradigm, termed mutualism, that attempts to ease this restriction. Psychological inquiry needs to take fuller account of the reciprocally evolved levels of physical, biological and psychological order which support mental life and from which it has emerged. These levels interact through organic modes of causation which are essentially the interchange of meaning. Mental life and consciousness are integral parts of this interchange. Psychological inquiry into consciousness needs to recognise the special character of organic causation and to consider meaning in an ecologically realistic way.
Recent critiques of cognitivism are reviewed to show a substantial overlap with mutualism. The paper concludes by considering how the postmodern condition of psychology promotes a concern with values as well as facts.
1. Psychology at the postmodern turn.
Postmodernism is said to offer psychology: " ... new ways of conceptualising itself and its potentials." [Gergen, 1992, page 25]. This paper will explore how psychology's treatment of consciousness might change as a result. Of particular concern is that postmodernism promotes Whitehead's philosophy of process over the implicit Cartesian basis for contemporary psychology (Griffin, 1988). The intention, echoing Toulmin's project for postmodern science more generally, is to promote the treatment of consciousness as lying within rather than apart from the natural order (Toulmin, 1982). It is also the intention to use the term 'postmodern' in the constructive sense employed by Toulmin, in distinction from its use in a deconstructive, even destructive, by writers on broader cultural and political issues (Jameson, 1991).
A postmodern appraisal of science supports the following critique: nature is an organic whole, but modernist science reduces our image of it to mechanistic fragments. Physical, biological and psychological levels of order are studied separately. Their integration, it seems, must wait until the fundamental workings of these different levels are properly understood. This analytic world view is part of modernism in its broader cultural sense (Havel, 1994; Gergen, 1992, page 19). The fragmentation central to what Weber called the 'disenchantment' of the world, marginalises consciousness and the human condition. A state of affairs that alienates scientists and non scientists alike (Holbrook, 1987; Berman, 1981; Bookchin, 1988).
Now, however, the limits of the modernist agenda are clearly in view. In physics the search for ultimate particles has been abandoned since at the present limits to investigation, reality appears more process-like than particle-like. Even were a unified theory of matter and energy to appear, it now seems unlikely that it would have any significant bearing on the more complex levels of organic or social order to which consciousness is so intimately attached (Barrow, 1991). Biology too has moved away from a modernist agenda on which genes, the biological equivalent of ultimate particles, are taken as a sufficient specification of organic organisation. Instead, genes are seen as necessary but not sufficient participants in a process of organic self-assembly which produces and stabilises biological order (Goodwin, 1989).
As science addresses the limitations of modernism, there is a renewal of interest in Whitehead (Griffin, 1988; Jencks, 1992). For Whitehead, nature is essentially an organic process. A process, moreover, that is in continuous creative advance. Consciousness is a participant in this advance, not a follower or observer. The advance itself cannot be adequately understood if science retains a modernist agenda. Attempting to explain complex organic events by reduction to minute details of isolated systems will not do. The natural order is not merely a collection of discrete objects and cannot be understood by observing disconnected systems. The material particulars revealed by such observation are not reality. If the world is held still and dissected, its structure will indeed be disclosed, but in only a limited, static way. Its dynamics will be obscured because, according to Whitehead, occasions, that is events in which organisms act, are the basic ground of the natural order. Science needs to reflect this: "Science is taking on a new aspect which is neither purely physical nor purely biological. It is becoming the study of organisms. Biology is the study of the larger organisms whereas physics is the study of the smaller organisms." (Whitehead, 1925, page 150).
Modernist science has been successful in disclosing the material vehicle for organic action. What is emerging with the postmodern turn is an agenda that addresses the action itself and with it the dynamics of Whitehead's creative advance. To echo Bateson, Prigogine and Bergson, what is in the making is a science of becoming and relations as opposed to one of being and absolutes (Bateson, 1980; Prigogine and Stengers, 1984; Jones, 1987).
A shift away from mechanistic being and towards organic becoming will have significant implications for where and how science looks into nature in order in order to understand consciousness. A science of being has absolute, ultimate constituents of reality as its subject and a final and complete system of laws as its goal. Upward causation is primary, with theories and methods assuming relative simplicity and regularity to be lying beneath natural diversity. By contrast, a science of becoming is concerned with stability in dynamic relations; its goal is to understand the emergence of autonomous organisation from simpler precursors. Under this latter aspect, consciousness is assumed to participate in and be supported by levels of organisation which can exert downward causal influence on the systems from which they have emerged (Sperry, 1994, page 4). A science of becoming takes the productivity of consciousness as a dialectical product of both downward and upward causation. Process thinking is central to a science of becoming. It is organic rather than mechanistic and thus concerns evolving relations rather than timeless absolutes. Nature, under this aspect, stands revealed as essentially rather than accidentally historical. It is this that opens the way to a new psychological inquiry into consciousness whereby it is treated as participating in physical and biological events rather than as something which follows them, like an acausal shadow.
During the modern era, while Western science has been almost totally a science of being, mechanism and timelessness were accepted as the fundamental nature of reality. Timelessness makes being the essence and becoming an illusion. Becoming, lived experience, Bergson's dureé, was thereby excluded from science. Husserl protested that this, part of the legacy of positivism, would remove the heart from psychology since dureé, the essence of consciousness, is its central phenomenon.
Indeed, psychology does appear to be the last resting place of the modernist agenda. Psychology is presently dominated by cognitive science or, more broadly, cognitivism. Metal life is taken to depend on processing cognitive representations of the external world, where the processing and the representations are treated as a species of symbolic computation. This computational view is presented both as the consistent theoretical and methodological basis necessary for a scientific psychology (Anderson, 1983, ch. 1), and hence as the basis of a unified theory of the mind (Newell, 1991). Computation, if that is what the brain is doing, can be studied formally and in isolation from cultural, historical or even biological context (Gardner, 1985, page 6). The success of cognitivism has led to its more radical pronouncements being presented as more or less fait accompli. It influences most other branches of psychology and is probably the major contemporary influence on the philosophy of mind and consciousness (e.g.. Dennett, 1991).
The commitment to mechanism is partly an inheritance from psychology's struggle to be accepted as properly scientific. The triumphs of nineteenth century classical science left such a legacy of respect that psychology took it as an ideal, something that one historian of psychology has termed 'physics envy' (Leahey, 1987). The irony is, that as psychology seeks to emulate classical practices and views, so the sciences from which they originated are moving beyond them. This has produced the paradox where the ideal of reductive mechanism is more likely to be found in psychology than in the sciences which it takes as a model (Morowitz, 1985).
A mechanistic view hinders inquiry into the organic basis of consciousness since mechanisms, by definition, lack interiority and necessary historicity, as is the case with organic systems. If psychology shifts towards an organic perspective, this hinderance will be lessened. The next section sketches such a perspective. It will be refered to as mutualism, although it should be noted that this term is used elsewhere is different senses, (e.g. Still and Costall, 1991). It emphasises that consciousness cannot be realistically considered apart from the evolutionary process.
2. Mutualism: Taking Evolution Seriously.
The eclipse of modernism has created a productive metaphysical vacuum. Within it the conundrum of mind and experience creates a natural focus. For modernist science, timelessness and mechanism were the essence of reality. On the postmodern agenda timelessness and mechanism are special cases. Indeed, some physicists now suggest that the appropriate level at which to search for fundamental physical principles are the irreversible, that is, historical, events that occur within organic systems (Rae, 1986, chapter 10). Such a proposal makes it easier to inquire into consciousness which, being an essentially historical processes, cannot be properly understood by reduction to timeless mechanisms.
Consciousness and processes that support it are parts of an organic system. Consciousness is one level, preconscious psychological processes another, brain another, body another and so on (Bohm, 1990). Parts of this system bear both external and internal relationships to each other because they share an evolutionary history. Organic systems are nested within each other and integrated by historically structured forms of reciprocal causation. The flow of action within such systems is open and generative. This flow cannot usefully be reduced to some closed, timeless set of elementary physical laws, causes or objects, and for that reason modernist science does not provide a suitable framework in which to understand it.
Mutualism is a proposal for how postmodern psychology can inquire into this flow in an appropriate and sensitive way. Three related aspects of mutualism will be examined in this section: formal causation, evolutionary emergence and an ecological realistic perspective on meaning.
Firstly, let us look at formal causation. Brains, minds and consciousness exist by virtue of a mutually evolved system of internal and external relations. Within such systems there exist organic forms of causation which are qualitative distinct from and which cannot be reduced to inorganic forms (Sperry, 1985, 1994). This promotes attention to the somewhat neglected Aristotelian category of formal causation. This concerns the structured effects of structured causes and deals with the recurrent patterns of causation by which one part of an organic system influences another (cf. Rosch, 1994). In this sense it easily escapes Hume's deconstructive analysis of causation since effects are patently similar to and hence are phenomenologically prefigured in their causes. Formal causes are carried by patterns of relations within organic systems and cannot be understood on the basis of knowledge of isolated systems of matter and energy alone. These are the material vehicle for formal causation, but not the thing itself.
The contemporary treatment of formal causation is relatively underdeveloped, even in biology, since modernist science deals primarily with material and efficient causation. However, formal causation has had an enduring, if minor, influence on biological science until relatively recently. For example, the rational biology of Goethe and d'Arcy Thompson, who studied morphology as the product of dynamic formative processes, remains a persistent presence in contemporary biological science (Goodwin, 1989).
Recent work on the origins of life suggests that it is a matter of formal rather than mechanistic causation (Eigen & Schuster, 1979; Wicken, 1987). Likewise, but in a much more fundamental way, the work of David Bohm and his colleagues suggests that all causation may depend on structured internal relations, that is, formal causation (Bohm and Peate, 1989; Bohm & Hiley, 1993, chapter 15). It is no surprise that in this work, consciousness is a central issue (e.g. Bohm & Hiley, 1993, pages 381 - 390)
Mechanism limits the understanding of both consciousness and of formal causation, on which it depends. Causation involves transfer. A mechanistic view, dealing primarily with material and efficient causation, is at its strongest when dealing with transfers of matter or momentum. Formal causation however, although material and efficient causation is its vehicle, is more appropriately seen as the transfer of information or meaning. It is this that is the basis of organic action and of consciousness.
Next, let us examine emergence, which is often used in a contrastive sense to reduction. Reduction is useful where isolated inorganic situations are concerned, but when organic action is in question, it is simply unproductive. Organic action does not break physical laws; indeed, they cannot be broken. But they can be irrelevant, and organic events cannot be reduced to simpler underlying physical events. The latter are not causes but effects, entrained in mutually evolved patterns of organic action. These patterns are self-organising and integrate physical, biological and psychological levels of order into functional wholes (Wicken, 1987, page 4). Such patterns are dynamically stabilised by the interchange of formal causes which generate and are generated by organic action (Serres, 1982). These patterns are neither predictable from nor reducible to simpler precursors; that is, organic order is emergent.
Now emergence is occasionally suspected, sometimes with justification, as a surreptitious attempt to re-introduce mystery into science. However, mutualism aims for emergence without mystery. Just as it is legitimate to talk of the emergence of social and political systems, so biologists and psychologists may talk about the emergence of biological and psychological systems (Popper, 1990). Under this view, mental life and consciousness does not require a special substance or force somehow intervening in the world of matter. Brain, mind and consciousness are different aspects of an emergent evolutionary production, a form of life. Within this mutualist world picture, no part of a system is privileged. Action within mutually evolved systems, the support for consciousness, is essentially composed of organic events. It is not reducible to simple physical causes. Whitehead's concern with the occasion as the fundamental spatio-temporal aspect of reality is a direct response to this reductive fallacy (Hartshorne, 1965). Likewise, to take emergence seriously is to recognise that consciousness must be considered within a framework rich enough to address the historical process that has produced it.
Lastly let us examine the ecologically realistic treatment of meaning. Meaning is most commonly treated as having to do with the semiotic systems of human culture. However, the interchange between animals and between animals and their surroundings is a semiotic process, and human communication has many characterisitics in common with this level of interchange (Sebeok, 1977). In the mutualist view, the very existence of organic order, the basis of ecological relations, depends on an interplay of meaning. The significance of this for organic sciences like psychology and biology is that action is also part of this ecological interplay. Bohm, in describing organic action, holds that ecological integration between physical, biological and psychological levels of organisation is an evolutionary product based on meaning:
"Most of the material environment ... can be described as the somatic result of the meanings that material objects have had for human beings over the ages. These meanings fundamentally affect our actions towards nature and the action of nature back on us ... some of the simpler kinds ... are just reflexes that are built into the nervous system , or instincts that reflect the accumulated experience of a species ... With higher animals this operation of meaning becomes more evident and in man it is possible to develop conscious awareness, and meaning is then most central and vital." (Bohm, 1987, page 79)
These "actions towards nature and the action of nature back on us" are intimately involved in evolutionary change itself (Bateson, 1988; Piaget, 1979). It is in this way that evolution is open, generative and reflects a history of ecologically situated actions. Bohm, acknowledges some influence from Whitehead in developing his ecological perspective on meaning. In particular we may note here Whitehead's discussion of perception as a symbolic rather then merely physical interaction between perceiver and perceived (Whitehead, 1928, pages 3 - 6). It is in this sense that formal causation was described above as the transfer of meaning. James Gibson's theory of direct perception and in particular his notion of affordance also takes an ecologically realistic perspective on meaning (Michaels & Carello, 1981; Lombardo, 1987). Jacob von Uexküll too, as a framework for his work on animal senses, takes organic relations to be based on meaning rather then on mechanics. He suggested that to understand them would require a science that:
"... embraces natural systems of signs alongside and before the human system ... one that breaks down the division between natural and human sciences. The basis of this science is a reversal of the picture of reality offered by traditional science ... true reality lies behind the "nature" that physicists, chemists and biologists conceive of ... and reveals itself through signs ... the rules and laws to which these signs are subject are the only real laws of nature." (von Uexküll, 1982, pages 2 - 3)
These different aspects of an ecologically realistic perspective on meaning are components of a natural semiotics, an integral part of a postmodern science of mental life. The subject matter of this science is neither culturally relative conventions nor supposedly absolute laws of nature but evolutionary systems which combine both. Shifting psychology in this mutualist direction helps to free it from the modernist restrictions by which it is presently hampered.
Mutualism in drawing attention to formal causation, emergence and an ecologically realistic perspective on meaning, promotes a shift away from the modernist agenda of cognitivism with its effort after a unified mechanistic theory. Postmodern science is shifting beyond this agenda, and psychology might well do likewise. The next section briefly reviews some recent critiques of cognitivism which suggest that something like this shift is in fact underway, and that there is significant common ground shared by these critiques and mutualism.
3. Critiques of cognitivism.
Cognitivism has been a major, if not the predominant, theoretical and methodological orientation of Western psychology for about four decades. It assumes that mental life depends on internal cognitive structures that represent or symbolise external reality, constituting a language of thought, as it were (Fodor, 1975). A further assumption is that these structures are manipulated in a rule like way to produce the psychological dynamics of action, thought and consciousness. The analogy with the computation is clear, and computational models are used as psychological theories in cognitivism's flagship discipline, Artificial Intelligence (AI). Furthermore Turing's demonstration that there exist universal principles of computation holds out the glittering prize that computation may provide a unified theory for mental life. To the extent that AI is taken to be psychologically plausible, it is presented as just such a unified theory, at least of cognition (Newell, 1991). There are even claims that with sufficiently powerful computers and sophisticated programs it will be possible to create centres of consciousness and subjectivity outside the biological realm. This idea, known as 'strong AI', is the culmination of the project initiated by Hobbes and Leibniz to treat reasoning as some form of manipulation of symbols, as Haugeland points out (Haugeland, 1985).
However Haugeland also notes that there is a problem with this project, which he calls 'the mystery of original meaning' (Haugeland, 1985, page 27). Until quite recently, AI has relied on programming languages designed for symbol manipulation. Now symbols carry meaning by denoting something else. But in AI systems the denotational meaning of a symbol is other symbols within the system, albeit that some of these refer to input or output conditions. In this rather narrow view, meaning lies predominantly within the system. The generally limited performance of robots that rely on symbol manipulation is an index of this unrealistic isolation of meaning.
This decontextualisation of meaning is an important part of the objections that Bruner directs at cognitivism (Bruner, 1990). His view is that a commitment to empty mechanism has hijacked the original impulse behind cognitivism's overthrow of behaviorism in the 1950's. In particular, he feels that the development of the mind within a local system of cultural meanings is a fundamental desideratum for any proper psychological theory. In practice, this inevitable relativism has been neglected in favour of the effort to describe the mature mind in non-developmental, that is, ahistorical, universal terms. The assumption is made that there exist culturally neutral universals of mental life, perhaps similar to those proposed by Turing for computation. Once these are discovered it will then be possible to return to development and understand it properly. But this, objects Bruner, is getting things in the wrong order. The mind, and particularly the human mind, develops by assimilating the culturally supported meanings that surround the developing individual. These meanings create the capacities of the mature mind, not the other way around. Computational principles on their own are not enough, the cultural context of the mind must be taken into prior account.
Edelman broadens this line in suggesting that cognitivism not only fails to take account of the cultural context, but also fails to take sufficient prior account of the organic nature of the brain and of the evolutionary processes from which its structure emerges (Edelman, 1992). He proposes that the brain develops by a blend of ontogenetic and phylogenetic selection. While initial genetic specification is important, the mature structure of the brain also reflects the actions of the organism which in turn are bound up with the reward system geared to what the organism values. Development, in Edelman's view is by selection, that is, it involves a dynamic and selective loss of neurones, rather than the growth of neural structures under genetic control. He believes that this rules out the idea that the brain carries out computations in the formalised Turing sense. Computation in this latter sense is value free and involves permanent functional relations between components of a simple system for handling symbols. These functional relations depend in turn on the persistent internal structure of the components. If a significant part of the functional architecture of the brain develops its in response to action and to values that reflect and depend on what is encountered in an environment, then the Turning machine seem an unlikely model for what it is doing.
Edelman also finds unsatisfactory in the way cognitivism approaches the encounter between organism and environment. The naturally occuring objects and events of the world lack the sorts of regular properties that might be captured in symbolic descriptions. Moreover, the 'properties' that an object or event has for one organism are different from those it may have for other organisms. This organism-relative specification of the world has a great deal in common with ecological psychology and with what Varela et al. term an 'enactive' view of cognition (Varela et al., 1991).
Edelman's approach to brain function is evolutionary, in both the phylogenetic and ontogentic timescales. The mature structure of the brain and the way it works cannot be understood in absolute, ahistorical terms. To understand how it functions requires that we understand how it develops as part of a form of life. That is, how as part of a mutually evolved, organic system, it supports organic patterns of causation. Searle objects to the strong AI position on somewhat similar grounds (Searle, 1990). Symbol manipulation alone, he claims, cannot be the basis of subjectivity, beliefs, intentions and so on. While computation in this sense may capture what Searle terms the syntax of intelligence, its content or semantics can only be found in systems with appropriate biological causal powers. According to Searle, brains possess these powers by virtue of being organic systems and for the same reason, computers do not. The problem is how to say what these powers might be without lapsing into some sort of vitalist regress. What is required is an evolutionary approach that takes more realistically into account the biological and cultural contexts of the mind.
In fact, this attention to the biol-cultural foundations of human cognition is found in other efforts to broaden the scope of psychology beyond mere cognitivism. For example, Maturana and Varela's notion of 'structural coupling' between organism and environment takes the structure of the nervous system to reflect the history of action of the species (Maturana & Varela, 1987). This blend of ontogenetic and phylogenetic learning, as well being reflected in the internal structure of the organism, also depends on the accumulation of the effects of past actions in the environment: " ... structural coupling ... becomes embodied in both the structure of the living system and the structure of the medium ... " (Maturana, 1978, page 39). This account of cognition is part of a theory of organic self-building which Maturana and Varela refer to a autopoiesis. Like Gibson's theory of direct perception, autopoiesis assumes a continuity of ontogenetic and phylogenetic learning and an active assimilative role for the organism (Michaels & Carrello, 1981, page 76). Maturana and Varela reject the idea that cognition depends primarily upon images or on computation-like manipulation of representations. To approach the mind in this way, that is, in terms of a formal and hence freely transportable 'knowledge level', is neo-dualism. The problematic causal powers of the Cartesian res cogitans are simply transfered to something like a computer program.
Instead, Maturana and Varela propose that the internal structure of an organism, especially that of the brain, is a self maintaining system that is perturbed by the encounter with an environment in which the organism acts. The results of the perturbation is a change of both the external and the internal structures which carry the system. These changes develop and may persist over various time scales, some evolutionary, others shorter, some transient. These changes, which reflect the history of interaction between an organism and it environment, have the effect that future interactions occur in a new and adaptive way. From an external point of view it appears that 'something has been learned about an environment' and 'stored in the nervous system'. Maturana and Varela, however, present this as a developmental process whereby a self-stabilising system preserves its integrity by more effectively coupling the internal and external structures that comprise the system. From this point of view, intelligent action, rather than relying on symbolic representations, emerges from structures which hold a history of interaction. An organism is a self maintaining system of relations. The internal relations reflect the environment to which the organism is coupled. But the environment too is structured by the history of interaction. Organism and environment thus mutually constitute and support each other. What contributes to the unique character of the human case is that this mutuality is strongly influenced by the objects and practices of cultural systems, especially language. This leads to a special case of self-organisation, namely, consciousness.
In their critique cognitivism Winograd and Flores, combine autopoeisis and Heidegger's action-based view of meaning, to conclude that conscious meaning, whether carried in language or in action, "always derives from an interpretation that is rooted in a situation" (Winograd & Flores, 1986, page 111). Decontextualised treatments of consciousness will thus be necessarily incomplete. Dreyfus also concludes that a computational metaphor is unable to capture much that is important about natural intelligence in general and human intelligence in particular (Dreyfus, 1979). While less concerned with biology and evolution per se, Dreyfus also combines a phenomenological view with a consideration of the situated nature of cognition and action. His fundamental objection, again following Heidegger and also Merleau-Ponty, is that perceiving and acting does not depend on internal representations. While these may exist and may even be important in some cases, the great majority of natural intelligent action does not involve internal symbolic processing. Rather, in a given situation, appropriate adaptive action arises from the activation of traces of an indefinite number of occasions where similar situations have been encountered. These occasions were, in turn, created by actions, that is, by: " ... purposes in the organism and goals picked up by the organism from an ongoing culture ..." (Dreyfus & Dreyfus, 1988, page 39). Thus the traces of occasions are not only held within the individual but are also distributed over the objects, the situations and the practices to which that individual becomes attuned by engaging with them and assimilating them. The implication here is that psychological inquiry has to be extended well beyond cognitive structures, that any structures that are considered should not be too strictly localised in space and time and that action is central.
Recent work in robotics has moved sharply in this direction Indeed, the shift is so marked that the work resulting from it is now tends to be called 'New AI' or artificial life (AL), in conscious distinction from the symbol based work of the last few decades. The most influential variant of this new work is connectionism, the attempt to produce biologically plausible models of the brain by building networks of neurone-like units arranged in input, output and intermediate layers. Connections between units carry patterns of excitation or inhibition which determine how the network as a whole behaves. The connections change according to various rules, some of which have to do with regularities on the input layers, others have to do with the reinforcement of desired patterns on the output layers. Some systems bring about change on the basis of outside intervention, others are more autonomous.
Connectionism has challenged the symbol based approach in a fundamental way, particularly in the fields of development and phenomenology. As concerns development, networks that structure themselves on the basis of experience are able, as a result, to generalise, recognise and classify inputs. As well as these demonstrations of, broadly speaking, perceptual learning, there are also cases of learning to act as well. For example, research in robotics has demonstrated that neural networks autonomously learn to solve control problems in a naturalistic environment with unknown dynamics and unpredictable perturbances. There is a striking biological plausibility to this research since the developmental change in the control system depends upon action-related flows of information. In fact, the network is designed using the development of a natural organism, a lion-fish, as a model (Fujii & Ura, 1994). As concerns phenomenology, connectionism provides a fresh and productive perspective on intentionality. By 'intentionality' here is meant the 'aboutness' of human consciousness that is the fundamental and problematic connection between mental states and the non-mental world. For Husserl and Merleau-Ponty intentionality in this sense is a prime object of phenomenological inquiry. The essential point made in recent work is that the historicity of connectionsim creates a distinctive perspective on intentionality. Rather than the difficult direct inquiry into what intentionality might be, it offers the more grounded and progressive inquiry of how it might have arisen (Valentine, 1995; Bechtel & Abrahamsen, 1991)
Thus, one of the fundamental points of interest about connectionism is that it is intrinsically historical. That is, the behaviour of a network cannot be separated from the history of action from which its structure derives. However, it is still very much in the mechanistic mode and concentrates attention within the brain, rather than across the system that includes both the brain and the environment to which it has become structurally coupled. More in line with this latter theme, is the robotics research of Brooks, which, although distinct from connectionsim, is also a fundamental element in the 'New AI'. Brooks attempts, as he puts it, "to build complete creatures" (Brooks, 1991, a). His robots have nothing like the computational resources of earlier generations of robots. Their relatively simple internal workings which consist of processing modules responsible for functions such as simple visual skills, elementary motor acts, internal monitoring and the setting of simple goals. These modules communicate via what Brooks terms 'subsumption architecture' where control and information passes adventitiously between modules, with no particular module having a co-ordinating or superordinate role. The actions of such robots in can appear natural, organised and even purposive. This is not programmed into them. It emerges from interaction, once their internal structure is appropriately coupled to what their surroundings permit. Brooks appears to feel that his work on robotics has, eventually, have profound implications for psychology in general: "My feeling is that thought and consciousness are epiphenomena of being in the world .... we will see the same evidence for thought and consciousness in our systems as we see in other people now." (Brooks, 1991, b, page 591). While the position being advanced here would not present consciousness as epiphenomenal, and might also question the strong AI line that Brooks appears to hold, there is clearly overlap with his fundamental point: psychology is mislead if it concentrates too exclusively on the rational manipulation of internal symbolic representations (Brooks, 1991, c).
Varela, Thompson & Rosch use Brooks' work to illustrate what they call their 'enactive' approach to the mind (Varela, Thompson & Rosch, 1991, page 208). The enactive approach emphasises that workings of the mind need to be understood as embodied in the situations, biological structures and and historical processes that support it. The enactive approach assumes that mental life cannot be usefully separated from flow of organic action of which consciousness is a part. Varela et al. present this view a stage in the progressive development of mind science from cognitivism through connectionism to enaction. They note how cognitive science has " .... drifted away from the idea of the mind as an input-output device that processes information toward the idea of mind as an emergent autonomous network." (ibid., page 151). The parallel progression in theory is from mechanism to organicism.
The enactive approach combines biological, psychololgical modes of inquiry. This pluralism of both theory and method helps repair the fragmentation that has been the result of psychology's emulation of modernist science. Varela et al.'s enactive approach calls on Western philosophers such as Heidegger and Merleau-Ponty as well as on Buddhist philosophers such as Nagarjuna, of the Madhyamika tradition. This eclectic style of mind science is a good illustration of the postmodern tendency to combine traditions distant in space and time to create new and productive perspectives. Indeed, the critiques of cognitivism sketched here are much in line with postmodern developments, which are leading to a science of becoming and relation. The search for a unified theory and the inquiry into the mind to one particular level is being replaced by a pluralist program in which cognitivism is blended with other traditions, both scientific and not.
These critiques of cognitivism share a number of related concerns. These are that proper account be taken of the biological basis of cognition, of both the evolutionary history and cultural production of mind and the of distributed system of biological and cultural meaning on which the mind and consciousness depend. Mutualism, as it was sketched above, also shares these concerns. Instead of internal symbolic processing, psychological inquiry takes as its starting point the process by which cognition and action have emerged from an evolutionary history of ecological habitation. It approaches the dynamics of mental life as depending on the exchange of structured patterns of causation between mutually evolved systems. Formal causation is thus central to the biological basis of cognition. Emergence is central to psychological inquiry into the evolutionary history and cultural production of mind. An ecologically realistic perspective on meaning is central in any account of how consciousness is situated in biological and cultural processes.
As the critiques of cognitivism reviewed here share a significant amount of common ground with the mutualist perspective, it may be the means to develop some of them further. One example is Searle's claim that the brain has biological causal powers. The problem here is that this concentrates too strongly on the just brain itself. Mutualism, in considering the brain within a larger system of ecological relations, seeks to account for these powers in way that is less localised, both in space and time. These relations are in turn only to be understood historically, that is, as situated in an evolutionary process. Structure within the brain is only part of the story. To be appropriately understood, that structure has to be considered together with structures distributed over space and over an evolutionary history of ecological and cultural habitation.
Another example is what Haugeland calls the 'mystery of original meaning', which may likewise be less of a mystery in the mutualist view. Unlike AI, natural intelligence deals in structured exchanges of causation within the mutually evolved structures that make up organic systems. Consciousness is clearly concerned with meanings that link an organism with the environment within which it has evolved and hence in which it is able to act. It thus seems unlikely that it could be understood in any ecologically realistic way merely by analogy with isolated symbol manipulation. If the meaning of symbols has to be grounded outside such systems, then, given the contingent nature of what lies outside the system, the ideal of a unified psychological theory based on the necessary workings of an ideal mechanism has to be abandoned. Mechanisms may be considered apart from their environment, but organisms can not.
What these critiques have in common and what they share with mutualism, is that cognitivism cannot take proper account of the organic vehicle for consciousness. It is necessary to contextualise consciousness as a precondition to understanding it rather than to decontextualise it by reduction to formal computational principles. Indeed, to hold that the nature of consciousness can be captured by computational principles may simply be a mistake, as Penrose suggests (Penrose, 1989, chapter 10; Penrose, 1994). To do so is to import Whitehead's 'fallacy of misplaced concreteness' into psychology. Clearly, there may be mental operations that can be so captured, for instance in reasoning, language, mathematics and game-playing. But equally clearly, there are aspects of mental life that lie outside such formalisation. Such aspects are the hard-wired epistemology of the brain, that is, its pre-symbolic capacity for supporting mental life in general and consciousness in particular.
Such aspects of mental life as may be understood in computational terms, will tend to be the products of relatively recent stages of cultural evolution. Predating this is an enormous evolutionary history which provides the organic foundation for mental life. Indeed, Popper and others have proposed that evolution is essentially the accumulation of information in biological and cultural structures, and thus fundamentally an epistemic or cognitive process. However, the great majority of this accumulation is not the explicit, formalisable information that is involved in computation. Rather, as Bohm proposes echoing Whitehead's view that both mind and body participate in the creative advance of nature, it is implicit, enfolded within organic systems and the flow of formal causation within them, (Bohm, 1986). Brain, mind and consciousness are parts of this advance. It is this sense that mutualism encourages psychology to recognise more fully that the biological nature of the brain, the vehicle for the mind, is not irrelevant.
Whitehead, in a more fundamental sense, holds that organic action creates occasions which, accumulating, form the ground of all phenomena, including consciousness. In following Whitehead, mutualism would look beyond human culture to the history of ecological habitation to link human consciousness with non-human forms. Mutualism promotes an ecologically realistic vocabulary and, as a variant of Whitehead's process view of nature, belongs on the agenda of postmodern psychology. The research program of this discipline will be similar in many respects Schneirla's efforts to contextualise psychology more fully within an evolutionary framework (Piel, 1987).
Recasting psychology as a postmodern science will also raise the issue of whether it should deal only with facts or whether it also should deal in values. As Bohm proposes: "A postmodern science should not separate matter and consciousness and should not therefore separate facts, meaning and value." (Bohm, 1988). Raising such issues is timely, since one of the principal problems facing human culture at present is managing the ecological impact of technocratic culture. This, clearly, involves matters of value as well as matters of fact. The last section of this paper briefly examines how mutualism may promote a closer integration of knowledge and value within psychological science.
4. Mutualism, Facts and Values.
Consciousness, as Langer points out, is essentially a matter of value and feeling (Langer, 1988). But since contemporary psychology treats the brain a physical object and since physical objects are feeling free, how then may feeling, and hence value, be part of psychological inquiry? Sperry, however, believes the study of human consciousness is an arena within which scientific knowledge and human values may be considered together in a particularly productive way (Sperry, 1985, chapter 6; Sperry, 1994).
But Sperry's position retains much of the anthropocentrism of the modernist agenda. In broad contrast, the mutualist paradigm sketched here would enlarge the value base of psychology still further to take fuller account of both the evolutionary origins as well as the cultural context of consciousness. That layer of human consciousness that is culturally conditioned, although the most accessible and salient, is not the whole. Indeed, it may only be the surface, a layer of paint, so to speak, covering a far larger volume of pre-cultural consciousness. Approaching consciousness in this way emphasises the continuity between the human and non-human conditions. It also serves as a reminder that human interests are not the only standard by which questions of value are to be judged. Psychological inquiry into consciousness needs to to concern itself with the ecological support for all forms of mental life. This will help redress the alienation of human action and values from the rest of the natural order that is so characteristic of modernist science (Berman, 1981; Toulmin, 1990).
The values expressed in actions reflect the modes of ecological habitation possible for the actor (Pickering, 1993). Values, rather than exclusively attached to human affairs, are immanent in the natural order. This is an echo here of the presocratic concept of isonomia - the principle that all parts of the natural order have value (Bookchin, 1988). The effects of human action, and especially the impact of technocracy, should be considered within this framework of natural values. Modernist anthropocentrism has obscured this framework to such an extent that moral philosophy has until recently concerned itself almost exclusively with the domain of human action.
This dissociation of the human phenomenon from the rest of nature is reflected in the split between facts and values in science, including psychology. However, once the mutual evolution of the human domain and its organic support are more closely considered, it is clear that facts and values are not dissociable in this way. Evolution is the accumulation of knowledge that is value-laden at its very core. The values involved do not have to do with struggle so much as with integration:
" To generalise that natural selection is over-all and even in a figurative sense the outcome of struggle is quite unjustified ... Advantage in ... reproduction is usually a peaceful process in which the concept of struggle is really irrelevant. It more often involves such things as better integration into the ecological situation ... exploitation of environmental possibilities that are not the object of competition or are less effectively exploited by others." (Simpson, 1949, pages 221 - 222).
Changes to any part of a mutually evolving system are immediately subject to evaluation. Those incompatible with the system as a whole do not persist while those that promote more open, assimilative ecological habitation will tend to persist and to be amplified. As Piaget points out, assimilative action drives and directs evolution (Piaget, 1979). Mutualism promotes psychological inquiry into this assimilative drive. As action expresses value, it is perhaps in something like the mutualist sense that Waddington, following Whitehead, described values as being "within nature" (Waddington, 1977, chapter 1).
The last four centuries of Western science have created an arsenal of techniques which operate on the world powerfully but not sensitively. Instead of wisdom, science is providing mere knowledge; a deep technology of anthropocentric power and control. Technocracy, running far in advance of ecologically grounded values, has relinquished the goals of understanding and harmonisation, characteristic of premodern science. The science that has resulted has transformed our experience of the world. Instead of the support for a balanced existence, the world is now seen as what Heidegger has termed the mere 'standing resource' for unconstrained human action (Heidegger, 1977).
The effects of this alienated perception are manifesting themselves all too clearly. As a response, part of the project of constructive postmodernism is an effort to rediscover and incorporate traditions distant in space and time from those of recent Western culture. Here it is easy to find less anthropocentric world views which strike a different balance between the intuitive and the rational. Taoism, for example, holds that the natural order arises from the spontaneous cooperation of all things, living and non-living, according to their diverse natures (Po-keung, 1986). Buddhism and Hinduism take all living beings to exist within a causal mesh whose every component has value, not just those attending human action (Rosch, 1994; Wicken, 1984; Naess, 1978; Pickering, 1995). Here, human action is contextualised within a broader and less anthropocentric scheme of values than that of Western scientific technocracy.
Mutualism is a proposal to shift psychology towards this perspective, in line with the broader postmodern shift towards organicism. Whitehead is naturally an important figure in this shift. Laszlo, for example, takes Whitehead as the basis for his view that values are to be found in:
" ... all self-maintaining and self-evolving organisations brought forth on this good earth and, if not perturbed by man, existing here in complex but supremely balanced hierarchical interdependencies." (Laszlo, 1972, chapter 14).
Mutualism treats value as an integral part of an ecologically grounded, postmodern science of the mind. This, blending traditions, is also a religious step in the sense of re-ligation, the linking back of psychological science to a context of evolutionary value and meaning from which modernism has detached it. This returns responsibility to human beings without the damaging anthropocentrism of either Western science or religion.
However, to adopt a mutualist approach is not to import religion into science nor is it to make some bland rapprochement between the two. Rather, it is part of what Whitehead described as a search for " .. wider truths and finer perspectives within which a reconcilliation of a deeper religion and a more subtle science will be found." (Whitehead, 1926, page 264; Hartshorne, 1965). Of course, Whitehead is not alone in finding little conflict between science and religion:
" ... science can only be created by those who are thoroughly imbued with the aspiration towards truth and understanding. This source of feeling, however, springs from the sphere of religion ... science without religion is lame, religion without science is blind." (Einstein, 1950)
This "aspiration towards truth and understanding" will be helped by a deeper recognition of the evolutionary production of consciousness. Mutualism is directed at the reciprocally evolved patterns of organic causality that support consciousness. Its aims lie beyond the discovery of yet more ways to intervene in these patterns to achieve human ends. Mutualism, as part of postmodern science, aims to understand how consciousness arises and is sustainable within such patterns. Following Whitehead, we may expect that such a science will be more naturally brought together with the meanings and values of other traditions.
Anderson, J. (1983) The Architecture of Cognition, Cambridge, Mass: Harvard University Press.
Barrow, J. (1991) Theories of Everything: The Quest for Ultimate Explanation. Oxford University Press.
Bateson, G. (1980) Mind in Nature: A Neccessary Unity. Fontana, London.
Bateson, P. (1988) The Active Role of Behaviour in Evolution, in Ho, M-W. & Fox, S. (Ed.s) Evolutionary Processes and Metaphors, London:Wiley.
Bechtel, W. & Abrahamsen, A. (1991) Connectionism and the Mind. Blackwell, Oxford.
Berman, M. (1981) The Reenchantment of the World, Cornell University Press.
Bohm, D. (1987) Unfolding Meaning. London: RKP.
Bohm, D. (1990) A New Theory of the Relationship between Mind and Matter. Philosophical Psychology, 3(2): 271 - 286.
Bohm, D. & Peat, D. (1989) Science, Order & Creativity, London: Routledge.
Bohm, D. & Hiley, B. (1993) The Undivided Universe. Routledge, London.
Bookchin, M. (1988) Toward a philosophy of Nature - The Bases for an Ecological Ethics, in Tobias, M. (Ed.), Deep Ecology, San Marcos, California: Avant Books.
Brooks (1991,a) How to Build Complete Creatures Rather than Isolated Cognitive Simulators. In Architectures for Intelligence, edited by Van Lehn, K.. Erlbaum, New Jersey.
Brooks (1991,b) Intelligence Without Representation. Artificial Intelligence, 47: 139 - 160.
Brooks (1991,c) Intelligence Without Reason. Proceedings of the 12th. International Joint Conference on Artificial Intelligence.
Bruner, J. (1990) Acts of Meaning. Harvard University Press.
Dennett, D. (1991) Consciousness Explained. Little, Brown & Co., NY.
Dreyfus, H. (1979) What Computers Can't Do. Revised Edition, Harper Row, New York.
Dreyfus, H. & Dreyfus, S. (1988) Making a Mind Versus Modeling the Brain: Artificial Intelligence Back at a Branchpoint, in The Artificial Intelligence Debate, edited by Graubard, S., MIT Press.
Edelman, G. (1992) Bright Air, Brilliant Fire. Basic Books, New York.
Eigen, M. & Schuster, P. (1979) The Hypercycle: A Principle of Natural Self-organisation. New York: Springer.
Einstein, A. (1950) Out of my Later Years, London: Thames & Hudson.
Fodor, J. (1975) The Language of Thought. Harvard University Press, Cambridge.
Fujii, T. & Ura, T. (1994) Self-Generation of Neural-Net Controller by Training in Natural Environment. In Computational Intelligence: Imitating Life, edited by Zurada, J. et al., Institute of Electrical and Electronic Engineers Publications.
Gardner, H. (1985) The Mind's New Science: a History of the Cognitive Revolution. Basic Books, New York.
Goodwin, B. (1989) Organisms and Minds as Dynamic Forms, Leonardo, Vol. 22 (1): 27 - 31.
Edelman, G. (1992) Bright Air, Brilliant Fire. Basic Books, New York.
Griffin, D. R. (1988) Introduction to The Reenchantment of Science: Postmodern Proposals, edited by Griffin, D.R., State University of Ney York Press, Albany, NY.
Havel, V. (1994) Address on receiving the Liberty Medal in Philadelphia on July 4th. 1994. Available from the author.
Harsthorne, C. (1965) The Development of Process Philosophy, in Philosophers of Process, edited by Browning, D., New York: Random House.
Haugeland, J. (1985) AI the Very Idea, MIT Press.
Heidegger, M. (1977) The Question Concerning Technology, (trans. Lovitt, W.) New York: Harper Row.
Holbrook, D. (1987) Evolution and the Humanities, Gower Press, Aldershot, UK.
Jameson, F. (1991) Postmodernism, or the Cultural Logic of Late Capitalism, Durham, Duke University Press.
Jencks, C. (1992) The Postmodern Agenda. In The Postmodern Reader, edited by Jencks, C., Academy Editions, London.
Jones, M-R. (1987) Some Thoughts on the Relevance of Bergson to Contemporary Psychology, in Bergson and Modern Thought, edited by Papanicolaou, A. and Gunter, P., London: Harwood Academic Publishers.
Langer, S. (1988) Mind: an Essay on Human Feeling. Johns Hopkins University Press, Baltimore.
Laszlo, E. (1972) Introduction to Systems Theory. New York: Gordon & Breach.
Leahey, T.H. (1987) A History of Psychology. 2nd. Edition, London: Prentice-Hall International.
Lombardo, T.J. (1987) The Reciprocity of Perceiver and Environment: the Evolution of James Gibson's Ecological Psychology, Hillsdale, New Jersey: Erlbaum.
Maturana, U. & Varela, F. (1987) The Tree of Knowledge: the Biological Roots of Human Understanding. Shambala, Boston.
Maturana, U. (1978) The Biology of Language: the Epistemology of Reality. In Psychology and Biology of Language and Thought: Essays in Honour of Eric Lenneberg, edited by Miller, G. A. & Lenneberg, E., New York: Academic Press.
Michaels, C. & Carello, C. (1981) Direct Perception. Prentice-Hall.
Morowitz, H.J. (1985) Rediscovering the Mind. In The Mind's I, edited by Hofstadter, D. and Dennett, D., London: Penguin.
Naess, A. (1978) Through Spinoza to Mahayana Buddhism or through Mahayana Buddhism to Spinoza? In Spinoza's Philosophy of Man, edited by Wetlesen, J., University of Oslo Press.
Newell, A. (1980) Physical Symbol Systems, Cognitive Science, 135 - 183.
Newell, A. (1982) The Knowledge Level, Artificial Intelligence, 18: 87 - 127.
Newell, A. (1991) Unified Theories of Cognition. Harvard University Press.
Penrose, R. (1989) The Emperor's New Mind. Oxford University Press.
Penrose, R. (1994) Shadows of the Mind: On Consciousness, Computation and the New Physics of the Mind. Oxford University Press.
Piaget, J. (1979) Behaviour and Evolution. Translated from the French by Donald Nicholson-Smith. London, Routledge and Kegan Paul.
Piel, G. (1987) Each Animal in its Own Psychological Setting. In Cognition, Language and Consciousness: Integrative Levels, edited by Greenberg, G, & Tobach, E., Hillsdale, New Jersey: LEA.
Pickering, J. (1995) Buddhism and Cognitivism: a Postmodern Appraisal. Asian Philosophy, vol. 5, no. 1.
Pickering, J. (1993) Signs, Actions and Values. In Life World: Sign World. Festschrift für Martin Krampen, edited by Dreyer, C. et al.. Jansen Verlag, Lüneberg.
Po-keung, I. (1986) Taoism and Environmental Ethics. In Religion and the Environmental Crisis, edited by Hargrove, E., University of Georgia Press.
Polanyi, M. (1962) Personal Knowledge. Routledge & Keegan Paul, London.
Popper, K. (1990) Towards an Evolutionary Theory of Knowledge, in Popper, K., A World of Propensities, Bristol, UK: Thommes.
Prigogine, I. & Stengers, I. (1984) Order out of Chaos. London: Flamingo.
Rea, A. (1986) Quantum Physics: Illusion or Reality? London: Cambridge University Press, 1986.
Rosch, E. (1994) Is Causality Circular? Event Structure in Folk Psychology, Cognitive Science and Buddhist Logic, Journal of Consciousness Studies, 1(1), 50 - 61.
Searle, J. (1990) Could the Brain's Mind be a Computer Program? Scientific American, January, pages 19 - 25.
Sebeok, T. (1977) Zoosemiotic Components of Human Communication. Chapter 38 in How Animals Communicate, edited by Sebeok, T., Indiana University Press, London and Bloomington.
Serres, M. (1982) The Origins of Language: Biology, Information Theory and Philosophy. In Hermes: Literature, Science, Philosophy, edited by Harari, J. & Bell, D., Johns Hopkins University Press, Baltimore.
Shrödinger, E. (1945) What is Life? Cambridge University Press.
Simpson, G.G. (1949) The Meaning of Evolution. New Haven: Yale University Press.
Sperry, R. (1994) Consciousness and the Cognitive Revolution: A True Worldview Paradigm Shift. Anthropology of Consciousness, vol. 5, no. 3, pages 3 - 7.
Sperry, R. (1985) Science and Moral Priority: Merging Mind, Brain and Human Values. New York: Praeger.
Still, A. & Costall, A. (1991) (Ed.s) Against Cognitivism: Alternative Foundations for Cognitive Psychology. Harvester, London.
Swenson, R. & Turvey, M. (1991) Thermodynamic Reasons for Perception-Action Cycles. Ecological Psychology, 3(4), 317 - 348.
Toulmin, S. (1990) Cosmopolis: the Hidden Agenda of Modernity. University of Chicago Press.
Toulmin, S. (1982) The Return to Cosmology: Postmodern Science and the Theology of Nature. University of California Press, Berkeley.
Uxeküll, T. von (1982) Jacob von Uexküll's Theory of Meaning. Semiotica, 42: 1 - 88.
Valentine, E. R. (1995) Orthodox Artificial Intelligence, Connectionism & Phenomenology In Phenomenology and Cognitive Science: A Handbook. Edited by Baumgartner, E. et al.. Josef Röll, Würtzburg.
Varela, F., Thompson, E. & Rosch, E. (1991) The Embodied Mind. MIT Press, Boston.
Waddington, C. (1977) Tools for Thought, Paladin.
Whitehead, A. (1925) Science and the Modern World. MacMillan, New York.
Whitehead, A. (1928) Symbolism: its Meaning and Effect. Cambridge University Press.
Wicken, J. (1984) The Cosmic Breath: Reflections on the Thermodynamics of Creation. Zygon, 19: 487 - 505.
Wicken, J. (1987) Evolution, Thermodynamics and Information: Extending the Darwinian Program, New York: Oxford University Press.
Winograd, T. & Flores, F. (1986) Understanding Computers and Cognition, NJ: Ablex.