The Problem with Bergsonian Dualism
If you find this essay useful why not visit my website -
http://homepages.tesco.net/greengate Where, if you feel so inclined, you may make small pp donation via the site which would be most welcome in assisting me in paying off my student debt.
Thanks and hope you enjoy my efforts. Paul King portcullis@tesco.net
Introduction
Both Bergson and Churchland believe that they have overcome the traditional philosophical mind-body problem which has its roots in Descartes’ dualism. Bergsonian dualism attempts to circumvent the difficulties of ordinary dualism by offering a modified dualistic account. Churchland’s eliminative materialism position does not confront the mind-body problem in its traditional formulation, that is, it does not acknowledges any domain of properties that are ‘...metaphysically distinct from the objective physical properties addressed by orthodox science.’1 As such, in the language of the mind-body problem, what we term ‘mind’ is essentially reducible to body. It maintains, contra-Bergson, ‘...that consciousness, with all its functions, is born out of the interplay of material elements.’(Matter and Memory p.72)
This essay will argue that Bergson’s dualistic solution to the mind-body problem is problematic. It will do so on the basis that many of his examples, and much of his argumentation, do not necessarily lead to a dualistic solution but are equally well catered for form the perspective of eliminative materialism.
It might be asked, why, in order to establish the objective of the essay, bring the work of Churchland to bear upon Bergson’s dualistic hypothesis, rather than that of any other materialist? The answer to this revolves around the issue of the brain, and the severe demolition of its functional importance it receives at the hands of Bergson in order to establish a workable dualistic solution. Bergson’s attempt to circumvent the problems of original dualism has drastic consequences for the ‘office of the brain’ which is reduced to the position of being purely ‘...an intermediary between sensation and movement...’(MM177). I will argue that this relegation of the brain is unacceptable to the twentieth-century philosophy of mind debate.
This is precisely the point at which the work of Churchland is so forceful since in its impact on Bergson. Churchland’s theory of the brain not only demonstrates how completely inadequate Bergson’s brain-model is, but it smoothly integrates with his eliminative materialistic position making much progress in providing a naturalistic understanding of the ‘mind’. In re-establishing the functional importance of the brain Churchland’s account goes much of the way in explaining perception, representation sensory qualia and, implicitly, memory.
The organisation of this essay will be as follows: chapter one will provide a critical explication of Bergson’s dualistic account in his attempts to circumvent the problems inherent in original dualism. In doing so the key issues will be clearly drawn out in order to allow chapter two to engage with them in the most fruitful way. Chapter two will then examine how Churchland’s eliminative materialist position, with its importing of the latest findings from neuroscience (empirical study of the brain) and artificial intelligence, provides a materialistic account which is able to absorb and answer the arguments and proofs Bergson uses to establish his dualistic hypothesis.
Chapter 1 : Bergsonian Dualism - A circumvention the problematic nature of original dualism?
What is the nature of Bergsonian dualism? At first glance Bergson appears to tackle the traditional Cartesian formulation of the mind-body problem head on, in that he tries to establish how two radically different substances - matter and spirit - interact. As Bergson writes in the introduction to Matter and Memory he affirms the reality of both matter and of spirit. Naturally this implies the irreducibility of one to the other, or the irreducibility of spirit.
However, a closer comparison between Cartesian and Bergsonian dualism reveals differences. Descartes’ substance dualism posits a total cleavage between what a human essentially is, a thinking thing or a mind, and the body: ‘I am a being whose whole essence or nature is to think, and whose being requires no place and depends on no material thing’2. The term substance is usefully defined by Armstrong as ‘...something which is logically capable of independent existence...’3. In the Cartesian frame-work mind and body have this independence. Descartes, however, was unable to resolve how the two components of dualism combine other than by notoriously unsatisfactory recourse to the pineal gland as the point of ‘...psychophysical transactions...’4. But, in Bergson we seemingly have a dualistic account which surmounts the difficulties
Bergson selects for his starting point, a different dualistic dichotomy; one which he believes will allow a circumventing of the problems of the Cartesian formulation. Memory, he says, ‘...is ...the intersection of mind and matter.’(Matter and Memory p.13) and the ‘ ...the classical problem of the relations of soul and body ...[centre] upon the subject of memory.’(MM13). Memory then is one aspect of Bergsonian dualism. As Pilkington notes ‘The independence of mind [has been] narrowed down to ... the independence of memory...’5. This move of ‘narrowing down’ cannot be passed over without comment. The significance becomes apparent in what constitutes our real experience: in our real experience, ‘...there is no perception which is not full of memories.’(MM 33) - memory is posited as one component of our real experience. Essentially, the second part of the dualism is matter, but as can be seen from the previous quote, perception has somehow been shifted to equate with matter; something which Bergson achieves through his unusual definition of matter, which is a new ‘.. way of looking at matter’(MM11).
It could be said then, that Bergson attempts to forge a workable dualism by utilising two theories; one of memory which was against the notion of the ‘...nineteenth-century orthodoxy...[of]... associationism’6, and one of matter which is mid-way between idealism and realism (MM9). Furthermore, Bergson writes that in a dualism starting from pure perception, where ‘...subject and object coincide...’ (MM221), the difficulties for understanding the link between mind and body are less formidable. How so? As mentioned, our real experience is a composite of perception and memory - this is Bergson’s choice of dualistic dichotomy. He understands perception and memory to be separated by a difference in kind and that to confuse and mix pure perception with memory is a metaphysical error (MM48). He thus posits the notions of pure perception and pure memory which exist ‘only in theory’ in order to see how pure perception might arise from matter; if this is established it would neatly clarify Bergson’s notion of subject and object coinciding (MM221), since perception would be one half of our composite of experience and, at the same time, essentially matter i.e. at the level of pure perception the subject-object relation way of looking at things is dissolved - they coincide.
The difference between Cartesian substance dualism and Bergsonian dualism now starts to resolve. In Bergson’s framework there is no absolute cleavage between both halves of his dichotomy of perception and memory. The composite is the reality and there is nothing to join back together as in Cartesian dualism. This is part of the Bergsonian method, to divide composites according to their natural articulations of difference in kind7. Pure perception exists only ‘in theory’ - effectively, not at all; pure memory as described by Bergson does exist, but it is ‘...preserved in a latent state...’ (MM141) and thus inextended and powerless. Memory, although different in kind, is not an independent substance from matter: it could not exist in isolation as in Cartesian dualism. This difference-in-kind form of dualism can be understood to be ‘...a toning down of the popular hard contrast of “matter with spirit”...’8 This further implies that there is more to being human that the pure Cartesian ‘thinking thing or substance’. But we are still left with a difference in kind, which still implies an irreducibility of memory to matter.
What is it that Bergson has to prove then, in order for his dualism to be tenable? First it is necessary to show how perception is so closely related to matter as not to be different in kind from it; then, since there is a difference in kind between perception and memory their mechanisms of interaction need to be exposed and explained.
To summarise the import of this we can say that, along with his two new theories of matter and memory, the notion of pure perception becomes the key to the reconciliation of mind and body: ‘We can understand that spirit can rest upon matter, unite with it, in the act of pure perception.’ (MM220)
It is important to consider the Bergsonian material universe, since from this he derives an account of the act of perceiving. This universe is supposedly one of common sense where matter is an ‘...aggregate of images’ (MM9). The term image here shuns the philosophical notions of reality in favour of so called common sense. It rejects both the idea that matter is an idealist construct of the mind, and also that there is anything more to it, or standing behind it, than meets the eye - as in Kant’s phenomenon/noumenon division. The common-sense image exists ‘...as we perceive it...’ (MM 10) with its primary and secondary qualities undifferentiated and equally real (although this definition of image slips to one of vibrations/movements; an issue which will be tackled later). This holds for all matter including the body, brain and nervous system etc.: they are all images. All these images act and react upon one another in a predictable law like manner as in, for instance, two billiard balls striking one another: if their mass, velocity and angle of incidence are known the collision is a predictable out come. Living matter, however, at the level of the macro-unity of the entity, is different in this regard. A stimulus does not necessarily lead to an instantaneous and predictable output, movement or action, or any response at all; there is no necessity of reaction as in the billiard ball case. Bergson, here introduces the notion of indetermination of response or action as ‘...a true principle...’ or, as Dewey states, ‘...indeterminateness is introduced as a specifying feature...’9. I would propose that, in the context of the notion of a zone of indetermination, we could say that inanimate matter and living matter are differentiated by the former having a zone of indetermination of radius zero, and the latter having a radius dependent upon its complexity (MM32). The indetermination of the being, which is itself ‘...suggested by the structure of the nervous system...’ (MM33) allows for an ever greater sophisticated response the more complex it is. There is a sliding scale of difference in degree only between inanimate and animate (living) matter. As the centre of indetermination of a being is diminished, as in the lower orders of life for example, the reaction to a stimulus becomes more immediate, and ‘...the more immediate the reaction is...the more perception becomes a mere contact...the process of perception ... [approaching] ...mechanical impulsion followed by a necessary movement’ (MM32). We can now see why a billiard ball doesn’t perceive - it has no centre of indetermination! Here, Bergson is equating the nervous system of a being with action; between perception and reaction there exists only a difference in degree and not kind. This is further borne out in Bergson’s distinction between the reflexive function of the spinal cord and perceptive function of the brain, where he posits only a difference in degree and not kind.
What is Bergson’s pure perception, how is it equated with matter and how does it arise? Briefly outlining Bergson’s overall suggestion first. He states that the material universe is awash with already existing representations, but they are virtual, not actual i.e. unactualised (MM36) . Virtual would seem to suggest that, in terms of a visual representation, the light rays which can potentially form a picture or actualised representation, have not yet struck a screen so to speak, and formed an image. This has some credibility if we consider that Bergson talks of a zone of indetermination as playing in some part a screen (MM39). Perception itself is a kind of cut-out or detached ‘picture’ (MM36); a reduced component part of the virtual representation which has been isolated and halted i.e. made actual (struck the screen of indetermination); effectively turned into a perception. So, as Harward notes ‘...perception itself, in so far as it is an image, [is] posited ... to begin with.’10
Perception arises ‘... when a stimulation received by matter is not prolonged into a necessary action’(MM32), and this is achieved by living matter having a zone of indetermination. This, Bergson explains, is to do with utility. Perception has a pure utilitarian origin (MM158), it is not speculative (MM28); it arises because the body performs actions and feels affections or, as Dewey states, a living body has special interests11. This interest causes a centre of indetermination to filter external influences. Those in which its ‘..functions find no interest...’ (MM36) are allowed to pass through, the rest isolated into a perception. We could say that there is an attunement to the environment, which is utilitarian in an evolutionary sense, with connotations of survival being ensured by the gearing or adaptation (MM84) of the organism to the environment; this being the ‘...general aim of life’ (MM84). Perception is the isolation of that which interests the living being in terms of the action it can effect on external objects or external objects on itself i.e. possible or virtual action. Perception shows, in the image world, the eventual or possible action of my body (MM22). Thus the filtering phenomenon is effected by a being’s centre of indetermination, which is itself a measure of the flexibility, in terms of sophistication, of what stimulus can be put into contact with what motor apparatus.
So nothing is added to perception by the brain, rather, the reverse happens: Perception is ‘...a problem of selection and elimination...’12 or, as already mentioned, a filtering. To clarify this point still further, we can say that perception is not a cognitive creation of the brain. Bergson believes this is an important point, since in passing from the unactualised representation (image/presence) to the actualised (perception) we have a diminution (MM35) which allows perception to be linked with matter and made external. It is external in the sense that Bergson understands perception as no longer being a mental creation occupying an inner realm. Perception is not within us or in the brain; that is, it is not a cognitive creation of the brain. Although the brain is obviously a material thing and, as such, it forms a link in the chain of perception as Bergson describes it. He states that the whole of the mechanism of perception can be described as follows: ‘external images reach organs of sense, modify nerves, propagating their influence to brain...The movement will pass through cerebral substance and expand into voluntary action.’ (MM40). Perceptions do not depend on the molecular movements of the cerebral mass (MM25) and in this sense only are they external: ‘ Perception, in its pure state ... is a part of things.’(MM64). This makes sense of the idea that ‘...subject and object combine...’ (MM221) in a dualism starting from pure perception. Perception is effectively shifted to be equated with matter, with no difference in kind existing between them. Dewey neatly sums up Bergson’s position on perception. Perception, he writes, ‘...is concerned directly with physical things...[it has]...no mental states intervening...Above all, perception is primarily a fact of action, not cognition.’13.
The idea of diminution being the pivot to allow perception to be made ‘external’ is a hard one to swallow. In making perception external and part of matter, Bergson is opposing all tradition: of theories of mind-body in materialism and original dualism; and theories of reality in idealism and realism - all of which situate perception in the inner realm or associate it with the mental. It has important consequences for what we are to take as the function of the brain, which will provide an important point of contention for the validity of his dualistic solution of the mind-body problem, from the perspective of eliminative materialism. The reason being that, the function of the brain is relegated to ‘central telephone exchange’ the purpose of which is to ‘... allow stimulation to choose its effect...to allow communication or delay it... its office is limited to the transmission and division of movement.’(MM30). It is the brain that allows a wide centre of indetermination of movements of the body image. Perception does not come from the brain, it only appears like it (MM41), i.e. the brain is not an organ of representation; a stimulus may travel to the brain but, once there, it does not change itself into a representation/perception (MM31) but it is routed to a motor mechanism by law of utility. So, on this understanding, we have no inner representations of the outer world upon which processing work of any kind is done.
Having seemingly established within his own theoretical frame-work, that perception ‘... is really a part of matter...’ (MM222) and not in the brain, in the form of a cognitive construction or mental creation, and therefore, that the brain is not an organ of representation, the brain suffers further marginalisation in Bergson’s theory of memory: as memory is banished from the brain along with perception. Obviously this is a crucial step for Bergson’s dualistic hypothesis; he has to show that memories cannot be stored in the brain, since the brain is part of the material universe - an image - and ‘...images do not create images...’ (MM23). As Harward states, he has to show that ‘...our experience is the meeting of two [independent] reals, spirit in the form of memory and matter resolved into motion.’14.
Bergson differentiates between two different ways in which the past is preserved i.e. two different forms of memory: in motor mechanisms and independent recollections (MM78). These are the two extreme forms of memory in their pure state, which in our real experience are normally combined. Bergson goes so far as to aver that these two forms of memory are different in kind (MM80/81). Motor memory is constructed and held the in nervous system, (MM94) whereas independent recollections or pure memory is not deposited in the cortical cells (MM119): ‘...there is not, there cannot be in the brain a region in which memories congeal and accumulate.’(MM126). In this instance our concern is mainly with the latter since this is the realm distinct from matter, or to be more precise, Bergson posits for this type of memory a difference in kind from matter. Pure memory, now isolated by Bergson, is posited as interacting with its dualistic counterpart - perception - in two ways: memory covers ‘...with a cloak of recollections a core of immediate perceptions...’ and that it ‘...contracts a number of external moments into a single internal moment...’ ; the latter leading to the ‘...subjectivity of sensible qualities...’(MM34). Deleuze’s neat terming of these as recollection-memory and contraction-memory respectively, indicates the bi-directional flow or tension between perception and memory15
It will be useful to examine what proofs Bergson uses to establish the differences between these two pure forms of memory, especially true memory in its independence/difference in kind. Also, a tracing of the full evolution of the two ways in which this latter from of memory interacts with perception, which is different in kind, will be attempted. In proceeding in this way the nature of true memory as Bergson understands it can be drawn out; why it is posited as being different in kind from matter and, given this difference in kind, how it is supposed to interact with matter/perception to constitute our experience. Also, the following analysis will assist in ascertaining if Bergson’s whole project cannot also be interpreted and fitted into a materialistic position; one which has the advantage of unifying contemporary understanding of the brain. This will be attempted in chapter two.
Bergson instances the idea of learning a lesson (MM79), such as playing the guitar, to draw a distinction between the two forms of memory. I can have ten lessons in playing a particular melody. Each time I become more proficient, until after the final one I have mastered it. I now have two things: the acquired skill of playing the melody and a memory of each of the successive lessons. The former he suggests as being stored in the body in the form of a mechanism, perhaps by a particular connection or arrangement of neurones; but the point is it has been materially registered and stored. It is a habit or action lived and acted not represented, (MM81) and one could forget that its was an acquired skill since it seems so innate. The individual memories of the lesson however, are representations, and as such are independent of the brain; they are ‘...absolutely independent of matter.’(MM177).
This being the case it immediately quashes the obviously tempting question, ‘where are they?’, as this is a question that only makes sense in spatial terms. It might be more appropriate to ask ‘when were they?’. This is precisely one aspect of Bergson’s method which is to ‘state problems in terms of time rather than of space.’16 As Pilkington notes, ‘Bergson considers all the events of one’s life to be stored up [preserved] and it is essential to his theory that he regards the totality of one’s past as being preserved...’17. Indeed Bergson regards the past which we are unaware of as being just as real as the space beyond our immediate visual vicinity which we can not see: ‘[There is no] reason to say the past effaces itself as soon as perceived than to suppose material objects cease to be when we cease to perceive them’ (MM142). Within this context the idea that ‘Because it has been shown that one thing is within another, the preservation is not thereby made any clearer.’(MM149) makes sense. Deleuze notes18 that Bergson sees the past or pure recollection as preserved in itself. Perhaps, at the risk of confusing things, it could be likened to ‘space’ preserving itself in itself; that is space is not ‘in’ anything else.
Apart from the necessity of making pure recollections/memory independent of the brain for Bergson’s hypothesis, what proof does he offer that they do not reside in the brain? Firstly, he writes that ‘...the physiology of the brain was more and more successful for localising sensations and movements in the brain, but never ideas...’(MM123). So, sensation and movement can be traced back directly to a specific part of the brain (i.e. have a spatial location); any damage to these areas uniquely and consistently affects a particular sensation or movement. But Bergson points out that this is not so with pure memories/recollections. This obviously defeats the simplistic idea that memory images are ‘ready made and have an abiding place in the brain’ (MM125). Secondly, although Bergson opens up a difference in kind between the two forms of memory, securing the independence of pure memory, the latter is dependent upon the former in terms of its actualisation: ‘..memories need, for their actualisation, a motor ally...’(MM120). This allows Bergson to maintain that, in instances of psychic blindness, deafness etc.(defects of recognition) caused by brain damage, no memories have been destroyed, rather, their actualisation has been interfered with due to damaged motor mechanisms. Finally, in auditory recognition, (MM117) if there is an associated image for the recognition of each word ‘...you must assume that there are as many auditory images of the same sound as there are pitches of sound and qualities of each voice.’(MM117), in other words an almost infinite amount.
The two ways in which Bergson forges an operational link between pure memory and pure perception - which are different in kind - resulting in concrete perception, will now be traced and pieced together by way of an example. (i) recollection memory covers ‘...with a cloak of recollections a core of immediate perceptions...’.(ii) contraction memory ‘...contracts a number of external moments into a single internal moment...’ or, it is a ‘...synthesising act of absorbing data into consciousness and binding them together with memories...’19; the latter leading to the ‘...subjectivity of sensible qualities...’(MM34). These will now both be considered separately.
Taking point (i) first: recollection memory covers ‘...with a cloak of recollections a core of immediate perceptions...’. Let us assume I have a perception which is diluted to the point of being pure. This fulfils this criterion of pure memory needing a motor ally to attain actualisation, since perception is virtual action (nascent/sketched out). Certain pure recollections or pure memories which exist in the virtual state and are attached to the past, therefore unextended and powerless, spontaneously go out to meet (MM99) the perception. Pure memory ‘...attains to a realised image as it expands...’ (MM134); expansion indicating there is no abrupt transition in terms of unextended/extended (MM213). Even at this point of actualisation from pure memory to memory image there is a profound difference i.e. a difference in kind (MM140). By the term memory image Bergson seems to mean a literal image (if we are considering visual images), one which has become conscious. This is obviously the case when we note that memory images ‘...go out to meet the perception, and feeding on its substance, acquire sufficient vigour and life to abide with it in space’(MM103); and that ‘...a memory image can interpret out perception so thoroughly that we can’t discern what is perception and what is memory.’(MM103): in other words a memory image could stand in for a perception, as in a hallucination. These memory images become ‘...more and more capable of inserting themselves into the motor diagram...’ (MM126) allowing concrete perception which is ‘... only defined and distinguished by its coalescence with a memory image...’ (MM127). So that is the full process, but what is it that is actually happening? Pure memory is virtual - a genuinely existing psychological state which is unconscious and unextended; the perception is extended. Bergson appears to be positing a smooth transition from the inextensive virtual state to the actualised extensive state whereby the two exist compounded together. This is confirmed by the case of sensation (different in kind to perception); the dawning memory of a sensation (unextended) is itself the sensation coming to be, and sensation is extended and localised, being of the body.
What about point (ii): contraction memory ‘...contracts a number of external moments into a single internal moment...’ or is a ‘...synthesising act of absorbing data into consciousness and binding them together with memories...’20 which supposedly leads to our subjectivity of sensible qualities? What is meant here by an external moment or, more vaguely, an item of data? It is referred to variously as ‘...a plurality of moments...’(MM34), ‘...the real..’(MM34) - as in contraction of the real, ‘...billions of vibrations...’ (MM203), ‘...the continuous flow of things...’(MM210). All these are certainly suggestive of vibrations. But Harward draws attention to the fact that an external moment is also a pure perception21 and, Bergson does indeed refer to concrete perception as ‘...a synthesis, made by memory, of an infinity of pure perceptions...’ (MM182). Harward presses this point to its conclusion: it means that Bergson treats matter in two ways (a) matter as an ‘aggregate of images’ whereupon a pure perception is then ‘..a vision of matter both immediate and instantaneous..’ and (b) matter resolved into the numberless vibrations of physical science where a pure perception equates to a single vibration. It would seem Bergson needs to treat pure perception as (b) - vibrations, in order for him to produce a sensible explanation for contraction memory; but as Harward rightly points out, Bergson leaves the matter ambiguous, allowing for himself the double meaning of (a) and (b) oscillating between them as and when his exposition requires. However, if we select (b) and return to the idea that our sensible qualities are effected by this contraction, (and it is certainly easiest to think of contractions operating upon vibrations) let us see what sense can be made out of it. Bergson gives a clear example of this with a musical note (MM203). An audio tone is perceived in concrete perception as a quality. But if its frequency is lowered ‘...to tally with the habits of our consciousness...’ (MM203) the quality resolves itself into repeated vibrations. So it appears Bergson is suggesting that pure memory holds the successive, individual pure perceptions, or the individual single vibrations, of sound, light or whatever; and it is the holding together which actually constitute their qualities.
There is no consciousness without memory, and no continuation of a state without the addition, to the present feeling, of the memory of past moments. It is this which constitutes duration. Inner duration is the continuous life of the memory which prolongs the past into the present22
If this is the case there seems to be an odd lacuna in the theory. Successive vibrations from what ever source, and collected by what ever sense, go through an electrochemical reduction or encoding. That is, memory does not effect contraction on, say, ‘light’ vibrations directly in order to form the quality of any particular colour; the vibrations upon reaching the eye are encoded in the form suitable for the nervous system: sets of electrochemical spiking frequencies23. But perhaps Bergson posits contractions in terms of this natural medium of communication of the nervous system? Just what contraction memory is contracting and how it gets a handle on it, since pure memory is different in kind, is very unclear. If we allow vibrations then the problem is, are these vibrations the vibrations of the physicists - the light vibrations, sound vibrations etc.? As already said, they cannot be. If the vibrations being contracted are the electrochemical signals of the nervous system then we could say the original vibrations of the physicist have already been contracted by the body’s nervous system before pure memory can get its hands on them, so to speak. This ambiguity in
the contraction mode of memory is critical for Bergson dualistic hypothesis, since it is the meeting between matter and memory, and it does not appear to be satisfactorily resolved. Harward, in reference to Bergson’s analysis of perception, describes it as being more of a metaphorical description than analysis, ‘...intended rather to stimulate the readers imagination than convey precise ideas.’24. Perhaps the same could be said of the way theoretically pure memory supposedly interacts with theoretically pure perception.
In the preceding paragraphs we have seen how Bergson tries to ‘outer’ perception from the brain, to make it part of the material universe, already present but in a virtual state. The brain is said not to create it, as such, but make it actual by virtue of the fact that the brain constitutes a centre of indetermination of action. In this way the functionality of the brain is greatly reduced. With memory Bergson attempts a similar move: he differentiates between two forms of memory - motor and pure, to the extent he avers they are different in kind. Then through a series of apparent proofs and arguments he seemingly establishes that pure memory cannot be attributed to a cerebral condition of the brain. In a consideration of the two ways in which this pure memory is supposed to come together with perception a serious difficulty became apparent in contraction memory, in terms of what was supposed to be contracted and the exact mechanism of the contraction. It was seen how this difficulty also drew another one with it, in that Bergson duplicitously oscillates between two alternative definitions of matter to suit his thesis.
In the following chapter I will consider if Bergson’s ‘proofs’ for establishing pure memory as being independent from matter (and which consequently lead him to adopt a dualistic stance) cannot actually be explained away by Churchland’s eliminative materialist account with its sophisticated theory of the material brain. From this it naturally follows to consider if perception also might be able to be absorbed back into the brain, that is, to make the brain wholly responsible for in terms of a mental/cognitive construction. And, since Churchland’s position incorporates a empirically detailed theory of the brain, I will consider how justified Bergson has been in minimising the functional importance of the brain. Following on the reinstation of the functional importance of the material brain by Churchland, his explanation of our sensory qualia will be considered as a model superior to the apparently flawed contraction memory suggested by Bergson.
Chapter 2 - The Eliminative Materialist response to Bergsonian Dualism
The following questions, all of which arise naturally from the termination of chapter one, will be set as a guide in probing Bergsonian dualism from the eliminative materialist’s perspective; the ultimate objective being to establish firm answers: (i) Can we really allow Bergson to limit the office of the brain as being nothing more than a ‘central telephone exchange’ the purpose of which is to ‘... allow stimulation to choose its effect...to allow communication or delay it... its office [being] limited to the transmission and division of movement’? (ii) Is Bergson’s explanation of our sensible qualities in terms of the contraction of pure perceptions or vibrations credible? (iii) Is the banishment of both perception and memory from the material brain tenable? With this question we have to be careful not to turn Bergson into a Cartesian dualist in terms of implying separate domains; as discussed in chapter 1, Bergson maintains that perception is not ‘...in the brain-matter...’ (MM38) in the form of a cognitive construction since perception is part of matter and the brain is not an organ of representation but only a link in the chain of perception. Also pure memory/recollection is virtual/different in kind to matter, and is not stored in the brain. Bergson’s interpretative ‘proofs’ based on clinical empirical data (chapter 1) for this latter point will also be considered. All of these questions are types of issue which can be tackled satisfactorily from the eliminative materialist perspective and particularly so through the work of Churchland, who introduces much empirical information from the arena of neuroscience in order to construct new theories of the mind/brain.
Before attempting to answer these questions, and by way of introduction of the eliminative materialist perspective, a brief overview of the role of language in the context of the mind-body problem will be considered. Both Bergson and Churchland display a startlingly similar discontent with the misleading functionality of language employed in divining the nature of the inner-self, or ‘mental’ life. For Churchland ‘Language appears as a biologically idiosyncratic mode of social interaction...it is mastered thanks to the ‘...versatility and power of a more basic mode of activity...’(Neurocomputational Perspective p.16 - Italics added). Bergson’s discontent with language is remarkably similar: he clearly brings out the utilitarian aspect of language in suggesting that the juxtaposing or spatialising inherent in language is fine and necessary for normal daily life25, but is inadequate and distorting in discoursing upon mental life in general.
However, this is where the similarity ends, since, for Bergson language is completely inadequate but Churchland is implying that we need a different language. Churchland, from his eliminative materialist position, holds that the everyday language we use - what he refers to as folk psychology (FP) - for explaining psychological/mental phenomena constitutes a false theory/conceptual framework, and that this language, will be eliminated or superseded by an alternative conceptual frame-work (NCP1) allowing a profoundly deeper, more accurate rendition in mutual understanding and introspection. This view is extremely radical since it means that ‘...mental processes as traditionally conceived do not exist.’26; they do not exist in the sense that, FP talk about mental states is similar to talk about the sun’s rising and setting, when the real scientific explanation is the rotation of the earth.27 The new framework, he argues, will be realised by empirical science - neuroscience in particular; essentially a materialistic account of the functioning of the brain. From this perspective, the traditional mind-body problem is an illusory construct due to the misleading conceptual framework of FP; thus it does not deserve a direct answer since it is a badly formulated question. Rather, what does need to be provided is a materialist account of the working of the ‘mind’. This eliminative materialist account turns out to be essentially quantitative and scientific; it is therefore in complete opposition to Bergson’s methodology of holding the mind to be qualitative and not addressable in quantitative terms. It could be said that Bergson’s problem with language is that it is quantitative in nature and for that reason inappropriate to discourse upon mind, whereas for Churchland, FP is too vague in its quantitative terms and needs to be more rigorously quantitative.
As we have seen already, Bergson’s attempt to affirm the reality of both matter and spirit has, from the contemporary point of view, resulted in an unpalatable marginalisation of the brain. Churchland holds no such affirmation of spirit and matter - quite the opposite; the idea that there can be any domain of properties that are ‘...metaphysically distinct from the objective physical properties addressed by orthodox science.’28 is anathema to him. Churchland’s methodology in providing an account of mind is to operate with a bottom-up, reductive approach. Importing into the eliminative materialist position the empirical findings of the neurobiological branches of science, which themselves smoothly integrate with the rest of the physical sciences such as physics, chemistry and biology, he tries to establish an explanation for the many features of mind. The conviction being that all these features can be reduced to the neurobiological level thus sweeping aside FP; the assumption is that there will no smooth unification between FP and the sciences.
Returning to the questions set at the beginning of this chapter, and first considering the problem of sensorimotor coordination, we find a note of agreement between Bergson and Churchland in acknowledging that sensorimotor coordination is a basic problem for life in general to solve. Bergson terms it thus:
The afferent nerves bring to the brain a disturbance, which, after having intelligently chosen its path, transmits itself to motor mechanisms created by repetition. Thus is ensured the appropriate reaction, the correspondence to environment - adaptation, in a word - which is the general aim of life. (MM84 - Italics added)
Churchland writes that,
...sensorimotor coordination is the most fundamental problem that any animal must solve, a means of solution...must surely arouse our curiosity... different creatures will have different means of locating objects, and different motor systems to effect contact with them, but all of them will face the same problem of coordinating positions... (NCP78/91)
Bergson’s phrase ‘intelligently chosen its path’ is odd in two ways: it suggests that it is the stimulation that chooses rather than the organism, making the brain passive rather than active; but, more importantly, it glosses over the how of the ‘...appropriate reaction, the correspondence to environment...’. But, Churchland supplies a credible and empirical how by introducing into philosophical debate on the nature of mind the recent findings of neuroscience. These findings are immediately at odds and undermining of the notion of the brain being nothing more than a ‘...a central telephone exchange...’; since the brain is portrayed as a active, powerful representational and computational organ. Representation being a capacity Bergson consistently denies the brain of being capable.
Within Churchland’s schema, sensorimotor coordination or the Bergsonian ‘...appropriate reaction...’ is basically a computational problem, such as may be found in moving a robotic arm in a paint spraying plant for instance. Problems of this computational nature obviously require a representation on which to compute - no computation can take place without data, and the data in this instance is some form of representation (NCP102) of the external environment. However, when we talk about representation here, this does not imply a kind of subject-object stance, whereby the subject regards the object which is the data/representation. Rather, the subject or self is constituted by the data/representation and the processing performed on it: i.e. representation + computation = self. What kind of representation and computation does Churchland suggest for a biological life form though? He proposes (NCP92) neural state-space representation and coordinate transformation computation; and this on the basis of hard empirical evidence derived from studying the small scale architecture of the brain, and subsequent successful modelling of the neural network structures found, in non-biological substrates such as electronic and software.
State-space representation and coordinate transformation need a little explanation.
State-space representation is the method of modelling ‘..various aspects of reality..’(NCP79). In Churchland’s sensorimotor two-dimensional crab example (NCP82-90) the position of the crab’s prey (input) is represented by its eye angles (one aspect of an external reality) in a two-dimensional state space or coordinate space, which would be implemented in a connection of neurones. The same goes for the motor output of the simple two-dimensional arm. This constitutes the state space representation. What about the coordinate transformation? The coordinate transformation puts these two state space representations into a functional relationship with one another which, mathematically, is relatively complex. The transformation is achieved by the way the two state spaces are effectively wired together. This provides a simplified example of how sensorimotor coordination is achieved.
It might be argued that this amounts to nothing more than the ‘...motor mechanisms constructed by repetition...’ that Bergson makes reference to, but it can be seen that we are already talking about an internal representation that is approaching what constitutes a cognitive construction of perception. Furthermore, Churchland pushes the idea of this form of representation and computation way beyond the case of simple two-dimensional sensorimotor coordination. This is made clear in Churchland’s consideration of the various peripheral input transducers of the senses which are highly suggestive of utilising state space representation. These transducers such as the eye, tongue, olfactory bulb etc. respond to the various aspects of reality, in a way which can be seen as a quantitative breaking up of the input continuum in question, rather like a kind of spectrum analysis; this is followed by a recombination in a suitable n-dimensional state space, thus constituting our qualitative sense of a particular input. For instance, in terms of the colour aspect of vision, the eye has three sets of colour receptors that respond to three different key wavelengths of light. These three input channels, so to speak, can be, Churchland suggests, internally represented in a three-dimensional state space. What bears this idea out is that, assuming similar discrimination along each axis of the different sensory state spaces, then for every extra channel the input transducer in question has - such as the four channels for taste in comparison to the three for colour - ‘...the variety of different taste sensation will be greater that the variety of different colour sensations by roughly an order of magnitude...’ (NCP105) and Churchland notes that this is the case. Thus a ‘...genuinely reductive account of one domain of sensory qualia...’ has been provided. (NCP105)
How does this account of sensory qualia for colour compare to Bergson’s? As discussed in chapter 1, Bergson talks of sensory qualities in vague metaphorical terms as being a contraction of the real; and when Bergson slips into his secondary mode of treating matter as vibrations this becomes a contraction of vibrations, where vibrations are the ‘...matter resolved into the numberless vibrations of physical science...’ - as referred to by Harward in chapter 1. This has a strange, perhaps coincidental, resonance with Churchland’s state space schema, in that we experience the three-dimensional colour state space qualitatively: the state space is a complex three-dimensional quantitative representation, but our experience of it is qualitative. The qualitative here could perhaps be seen as a ‘contraction’ or compression of the complex quantitative representation. But what is important is that Churchland has effectively put a quantitative and reductive explanation behind the mental state of the sensation of colour - something Bergson is generally against, that is, talking about the inner qualitative mental states in quantitative terms. However, when Bergson talks about contractions of vibrations, is he too, not really putting a quantitative explanation behind a qualitative inner mental state? It would seem so.
It will be noted that in the state space examples given so far, there is an important difference. In the sensorimotor example two state spaces were connected by a functional relationship and their inputs and outputs were derived from, and in, the external world respectively. In the state space explanation for colour the input is from the external world and the output effectively ends in the brain. So we can say that the inputs to and outputs from functionally interrelated state spaces can take three modes: external/external (sensorimotor example); external/internal (colour vision); and, importantly internal/internal. The importance of the latter arises from Churchland’s extrapolation that this form of representation and computation may be responsible for,
‘...the higher cognitive activities...’(NCP91). In positing this Churchland is now pushing his processing schema even further, extending it beyond the sensorimotor coordination and sensory qualia examples already given. As he writes, in any ‘...creature of complexity, we can expect a long chain or hierarchy of internal systems interacting with one another, systems that are the maps of other internal systems and whose outputs drive the activities of further internal systems...’.(NCP96). The higher cognitive activities are listed as language use and propositional knowledge; should the representational and computational modes suggested by Churchland be responsible for these areas then the quest for a fully reductive account of mind in neurobiological terms would be achieved.
Arguably this account of Churchland’s, in terms of extending state space representation and coordinate transformation to the heights of all higher cognitive activity is admirably plausible and awaits only the progress of empirical science to confirm it. The plausibility can be based upon evolutionary grounds. Such an elegant biological representational and computational solution has ample evidence for being an evolutionary solution for the sensorimotor coordination problem. Furthermore, as the biological implementation, has no dimensional limitations (NCP100) in its ‘...mathematical operation or physical realisation...’ - which suggests potentially phenomenal representational and processing power - it is difficult to see how evolutionary forces of natural selection would not seize upon this schema of sensorimotor coordination and develop it into something which accounts for human intelligence and possibly consciousness; a high level of intelligence being an excellent faculty for survival, in terms of the evolutionary notion of survival of the fittest. Another pertinent factor conspires to give this reading: the ‘...massively parallel nature...’ (NCP90) of the representational/computational model suggested. This gives rise to two evolutionary favourable factors in terms of survival. Firstly, great speed of operation is achieved, even in a biological substrate, which is far from ideal in this respect (NCP90); and secondly, great redundancy is effectively built into the system, thus allowing for the failure of a great many neurones that implement the structures with little loss in performance (NCP90). This last point could be said to be a necessity for any biological life which is to be evolutionary successful.
When we consider the issue of memory, so important in Bergson’s work, we find little mention of it by Churchland. For Churchland it is not a focus of concern, but it is obviously implicit in any schema of processing data - the data has to be ‘held’ whilst processing; also, those state spaces - implemented by connections of neurones - which effectively process it have been organically grown, and must in some manner count as a form of memory. We have with Churchland, contrary to Bergson, a model of memory which is materialistic and suggestive of being distributed throughout the entire brain.
Now, Bergson is adamant that memory is not in the brain and that it does not have a cerebral condition to it, (MM43) instancing various ‘proofs’ and arguments, (chapter1 p.7/8) to support this. However, if his objections can be answered by utilising Churchland’s materialistic account of mind we can safely return memory and memories to the material brain; and it would seem that they can be met. Briefly restated, Bergson’s proofs are: (1) individual memories cannot be located in the brain and do no appear to be totally destroyed by limited degrees of brain damage; (2) since he maintains that memories for their actualisation need a motor ally then and memories apparently lost can be accounted for in terms of damaged motor mechanisms interfering with their actualisation; (3) in recognition, (auditory being considered) if there is a memory image for each word then the qualities of pitch and sound of each voice would require an almost infinite amount of storage.
As far as auditory recognition is concerned, Churchland too is against the idea that this could be achieved by matching against a simple list or set of store memory images. The range of acoustic variation among acceptable and recognisable vowel sounds, for instance, (NCP163) is enormous and defies analysis/recognition by a simple list. But after having said this Churchland does provides us with a physical system, based on the brain architecture, in the form of a multi-layered neural network which can ‘...recognise such intricacies.’(NCP164) An example being the trained neural network for the discrimination of echoes between mines and rocks. This network is effectively using state space representation and coordinate transformation as discussed above, and once trained can perform a discriminatory/recognition task whose complexity is on a par with word recognition, without recourse to huge lists or previously stored memory images. This network, once trained, effectively embodies knowledge about certain aspects of its environment, and this knowledge is stored as ‘...a carefully orchestrated set of connection weights...’ between the synapses of the neurones that make up the network; and, it might be added, it is also distributed throughout the network and not to be found located in one place. This effectively answers Bergson’s point (3). For recognition to take place we do not have to have a store of all possible images, auditory, visual or otherwise, and yet recognition or memory can still be implemented by a physical or material system.
What about point 1: individual memories not being able to be located in the brain and being resistant to brain damage. Could we perhaps say that memories are effectively stored materially in the brain, but in a distributed sense, i.e. materially stored but not spatially located? That is, they are not to be located anywhere as such, but are effectively embodied in the ‘...connection weights...’of large populations of neurones. This would account for memories being resistant to destruction of areas of the brain. A suitable analogy of storage is suggested by idea of holographic photography: if we take a plate-glass holographic representation of, say, an orange, and break it into numerous fragments, each fragment retains the whole image but in reduced quality. However this analogy must be immediately discounted since we are back in the position of suggesting ready made and stored memory images, and both Bergson and Churchland are against this. Churchland has said of his trained neural network that it only embodies knowledge not memory images.
A distinction between ‘memory’ and ‘memories’ needs to be opened up. In the above network memory has a material foundation but memories do not. Bergson’s use of the virtual and actual in relation to memories or memory images seems to have some useful explanatory power. In the preceding paragraph I said that memories are effectively stored, materially, in the brain; there is no getting beyond this for a materialist position. For effectively we could say virtually or potentially, that is, they are stored as Churchland’s information or knowledge in brain like networks, but not as ready made things.
What then enables this materially stored information or knowledge (potential or virtual memories) to actualise itself from this virtual state into memory images in consciousness? First of all we could say that it does take a ‘consciousness’; for a Bergsonian virtual recollection to change form virtual to actual it needs consciousness to become actualised in. Unfortunately consciousness would count as one of the higher cognitive functions Churchland makes reference to, and in this area his schema is only speculative. Also we are considering only a simple model of a neural network which cannot be said to be conscious.
But bearing this limitation in mind the following passage is helpful:
...the whole problem of how to retrieve relevant information is transformed by the realisation that it does not need to be ‘retrieved’. Information is stored in brain-like networks in the global pattern of synaptic weights. An incoming vector activates the relevant portions.... of the trained network by virtue of its own vectorial make up... (NCP195-Italics added)
Now we are still trying to answer point 1: individual memories not being able to be located in the brain and being resistant to brain damage but we must also consider point 2 at the same time: memories, for their actualisation, need a motor ally, and that memories apparently lost can be accounted for in terms of damaged motor mechanisms interfering with their actualisation. Churchland’s notion of an incoming vector activating relevant portions of a trained network is akin to Bergsonian actualisation via a motor ally; in Bergson’s schema the incoming vector would be a motor stimulus/ally but in Churchland’s schema this can be a motor ally or an internal stimulus - as in the stimulus of state spaces already mentioned, either external or internal. Continuing with the analysis, we could suggest that various areas of the brain are activated by incoming vectors - motor, external, internal or otherwise - and the result of all the component parts of the stimulation is responsible for forming a memory image in consciousness. In other words individual memories are actively created or recreated out of the information or knowledge stored in Churchland’s ‘...brain like networks...’ as a result of the particular pattern of stimulation.
To sum up the import of this idea it can be said that Bergson’s notions of virtual and actual as applied to memory images or recollection can be left standing, since they are quite insightful; and that memory does have a material basis or cerebral condition to it in the form of information or knowledge, but that memories, as such, have only a very tenuous material foundation: they could not exist apart from the material brain but would seem to be a creation of its dynamic activity. This answers point 1 and point 2: since within Churchland’s schema memories do not necessarily need a motor ally for the actualisation - they could be actualised internally - but in some instances they may well have.
It would seem that Bergson latches onto the simplistic and essentially false notion of memories being stored as ready made images, only to easily knock it down and then use this to suggest that memory and memories are independent from the brain, or not stored in the brain, or do not have a cerebral basis. It has been clearly shown that a purely materialistic account with a little more sophistication than a naive simplicity can amply deal with Bergson’s objections when it comes to memory. Memories however must be considered as outlined above, i.e. as active creations of a functioning human brain through stimulation of knowledge containing networks, and in this respect Bergson’s virtual/actual schema is metaphorically insightful; but this stimulation does not need to be considered as only a ‘motor ally’ when we consider Churchland’s schema.
Conclusion
I have attempted to show how Bergson’s dualism of ‘difference in kinds’ attempts to divorce both perception and memory from the material brain - it is not the brain that is responsible for them - in a way which is contemporarily unacceptable, especially in the professional community associated with the philosophy of mind. We have seen how he needed to fabricate the idea of pure perception in order to collapse the distinction between perception and reaction/mechanical impulsion; thus siding perception with matter in terms of there being no difference in kind between perception and matter. This he does in order to demonstrate how pure memory - posited as different in kind from matter - then combines with perception in order to produce our experience. This is his dualistic hypothesis.
The objective of this essay was expose Bergson’s dualistic solution to the mind-body problem as problematic in that much of his argumentation and evidence can equally well be interpreted from the eliminative materialistic position. Churchland’s brand of eliminative materialism also has the added advantage in that his theory of the brain, which fully integrates with his eliminative materialist position, is superior to Bergson’s inadequate model. It incorporates the latest findings from the research of neuroscienc - the empirical study of the brain - and feed back from AI research which has successfully modelled many of the structures found by neuroscience in the material brain. Churchland effectively reinstates the importance of the material brain enlarging its office from the mere ‘transmission and division of movement’ which Bergson’s dualistic thesis limits it to.
Within Churchland’s schema we have seen that perception and memory can be relocated in the brain in the sense that it is the material brain that is responsible for them. Perception, contra-Bergson, can be understood to be ‘in’ the brain in the sense of a cognitive construction and representation. The basis for this is that the brain is effectively a powerful parallel and distributed processing organ; that it uses states space representation and coordinate transformation in its processing feats.
When it comes to memory, much of Bergson’s dualist argumentation for suggesting it is independent of the material brain can be adequately met by Churchland’s eliminative materialism. Memory, as distinct from memories, has been shown to be the responsibility of the brain. This is in the sense of knowledge or information being embodied within the connection patterns and weights of large populations of neurones. However, in considering recollections, memories, or memory images, it has been shown that there is what could only be described as a very tenuous material basis, since these could be said to be an active creation of the working brain based on the stimulation of its information containing networks.
Notes and References
2 Flew, A. A Dictionary of Philosophy 1984, Pan Books, p 91.
Armstrong, D.M. A Materialist Theory of The Mind Routledge & Kegan Paul, 1976, p.7.
Flew, op. cit., p.92.
Pilkington, A.E. Bergson And His Influence - A Reassessment
Cambridge University Press, 1976, p.12.
ibid., p.7.
Deleuze, G. Bergsonism Zone Books, 1991, p.22.
Fawcett, E.D. “ IV- “Matter and Memory” ” Mind, 21 no.82 (1912): 213.
Dewey, J. “Perception and organic action.” The Journal of Philosophy, 9, no.24 (November,1912): 667.
Harward, J. “ VI - Discussion. What does Bergson Mean by Pure Perception?”
Mind, 28, no.112 (1919): 467.
Dewey, op. cit., p.651.
ibid., p.650.
ibid., p.651.
Harward, J. Op. Cit., p.469.
Deleuze, G. Bergsonism Zone Books, 1991, p.52.
ibid., p.31.
Pilkington, op.cit., p.9.
Deleuze, op. cit., p.54.
Pilkington, op. cit., p 8.
ibid., p 8.
Bergson, H. An Introduction To Metaphysics Prentice Hall, 1949, p.40.
Churchland, P.M. A Neurocomputational Perspective. The Nature of Mind and the Structure of Science MITT Press, 1993, p.104.
Harward, J. op. cit., p.207.
Bergson, H. op.cit., p.29.
Brown, S. et al. Biographical Dictionary of Twentieth-Century Philosophers
Routledge, 1996, p.148.
ibid., p.147
Churchland, op.cit., p.74.
If you found this essay useful why not visit my website -
http://homepages.tesco.net/greengate Where, if you feel so inclined, you may make small pp donation via the site which would be most welcome in assisting me in paying off my student debt.
Thanks and hope you enjoy my efforts. Paul King portcullis@tesco.net