|
|
|
Mainstream 161.) Mindware - Andy Clark 2.) Identity theory 1.) Mindware Andy Clark Oxford University Press (2001) INTRODUCTION: Mindware represents a useful and readable summary of much mainstream thinking on neural processing and consciousness, although the latter is relegated to two appendices. Surprising aspects of this book are the abstract nature of much of the discussion, with only limited attempts to relate to the ongoing development of physical neuroscience. In particular, attention is almost exclusively focused on rational cognition as opposed to emotional processing, although neuroscience clearly shows that behaviour is impaired if the physical link between these two processes is damaged. The discussion in this book is surprisingly science-light in the sense of seldom relating the very abstract or computer-related discussion to the actual physical processes of the brain. This is excused at one point by the functionalist-type argument that it is the organisation rather than the actual material that is important. However, there is a problem with this approach. It only really works well if we can be sure that we have a full understanding of the physical structure on which the organisation is based. In looking at a machine, the system or process may be more important than the exact alloy out of which it is made or the type of screw used. This is because we have a full understanding of these features and provided they are adequately strong, the design or organisation may be more important. The problem here is that we do not have the same degree of knowledge in dealing with the brain, at least in respect of consciousness. It is difficult to have confidence in an almost entirely abstracted discussion when the physical underpinnings are so uncertain, and new research is continually emerging. The author himself criticises consciousness studies and artificial intelligence for taking the view that they can ignore the biological brain. However, this criticism is not really followed through in the actual development of the book, which has a largely non-physical emphasis. A further omission is the lack of attention to emotions. Evidence indicates that damage to the connection between the cognitive processing of the prefrontal and the areas responsible for emotional processing has serious consequences in terms of behaviour, but this book never touches on this. Consciousness meanwhile is pushed into appendices, although these are arguably the best part of the book, providing a concise and useful summary of mainstream ideas on consciousness. The author compares information system based theories of the mind with those that relate more to the biochemical aspects. In defence of the former, he quotes a science fiction story about electrical/information based aliens, who are baffled to discover that the inhabitants of Earth have brains made of meat. What this story really seems to do however is highlight the whole problem of how consciousness could ever arise from classical physics. We have a very good understanding of classical physics, and nothing we know about it produces consciousness. This applies not only to meat, but to information and to electrical activity, which like meat is instantiated in energy and matter. That said, it is arguable that we should not be so dismissive of meat. Meat is mainly protein associated with water, and as such it is organised by a dense mass of quantum bonds, only marginally stable, and possibly suitable for information processing. The same lack of grasp of the fundamental nature of the consciousness problem is apparent in the discussion of Searle's Chinese Room, a thought experiment in which a concealed person, possessing no knowledge of the Chinese language, uses a look-up system for Chinese symbols in order to conduct a convincing conversation in Chinese. From this Searle argues that mere manipulation of symbols, as proposed when the brain is viewed as a conventional computer, cannot explain consciousness. The author thinks that the appeal of Searle's argument to readers is merely intuitive, and argues that there is no reason in principle why a symbol system should not be conscious. However, I think this argument misses the whole point about the problem of consciousness, which is that there is no reason in terms of classical physics why anything should be conscious. In the real world, symbols are used the whole time without any suggestion that they are conscious, and non-conscious machines routinely process symbols. For consciousness to arise in a symbol system, it has to be in some physical way different from these everyday non-conscious symbol systems, but no attempt is made to suggest what this difference might be. The discussion of Ned Blocks idea of having the whole population of China pass round symbols in the manner of the human brain is similar. The author argues that it is only an intuition that makes readers think that consciousness or understanding will not arise from this. But again there is more to it. If the collective consciousness surmised doesn't exist at the beginning of the process, it requires some physical change to come over the Chinese population, for it to exist by the end, but no such physical change is suggested or explained. In a way, this is characteristic of the book as a whole, where there is relatively little attempt to relate hypothetical and abstract systems to the physical world. Artificial Intelligence: The author does highlight some of the problems of artificial intelligence. He discusses the question of depth of understanding. It is possible to provide a robot with a script for a simple process such as ordering food in a restaurant. But what if something unexpected happens, not catered for in the script. In principle the only solution to this problem is to add more programmes, but it is never practically feasible to deal with all the possible situations in this way. The author admits that the fact that there is a problem with robotics in this respect is confirmed by the lack of common sense understanding manifested by actual robots. The author sees this problem as a valid criticism of the idea of the brain as a straightforward symbol processor. Further, he does not think that the connectionist approach, where computers mimic the parallel processing seen in the brain, helps to solve this problem. He favours the notion that the brain is what he calls a bag of tricks accumulated over evolution with different neural processes adapted to different tasks. Dennett: Dennett's position in respect of neural processing is relevant because it is effectively part of his general attempt to explain away consciousness. His view is that non-human animals, plants and even machines such as a household appliance like a dishwasher have a quality which he calls 'intentional stance'. His view is that an animal, plant or machine has been designed either by evolution or human inventors not to behave randomly, but to behave in a way that makes sense and is effective. The organism or machine possessed of this intentional stance can be described as rational, in the sense of receiving information and directing its actions in response to that information. Dennett goes on to argue that humans are no different from plants and machines in being designed to process information and act in certain sensible and effective ways. The idea that humans have beliefs stored in their brains seems to be viewed as some form of illusion based on the fact that there are certain patterns in behaviour that Dennett deems, for no very apparent reason, to be perceived as beliefs. From an evolutionary point of view, it is not at all clear what would be the adaptive advantage of investing energy-intensive brain activity in supporting cognitive objects that have no apparent function. The Dennett approach appears to leave out two factors. Firstly, he ignores the much greater autonomy enjoyed by humans relative to dishwashers and plants. I leave aside for a moment, as something of a red herring, the difficult question of whether non-human animals are more like humans or more like dishwashers. The properly functioning dishwasher has no option but to act in a particular way in response to particular information, such as if button three has been pushed, run for 87 minutes. Even if we do not support the idea of freewill as such, it is clear that the brain usually has a whole range of options in response to informational input, and that the choice of option derives from internal processing, deterministic or otherwise. Furthermore, recent neuroscience emphasises that the brain involves an interplay between rational processing and subjective emotional experience, with serious deficits apparent, when brain damage disrupts the exchange between the emotional processing areas and some prefrontal areas. This important aspect of the physical brain seems to be entirely left out of Dennett's account. Amongst mainstream thinkers, Fodor is the main supporter of what is often called folk psychology, the idea that the brain works more or less the way that it feels as if it works. He takes the view that there are mental states that have meaning and are causal for actions. The author argues that Fodor's approach has good explanatory value in according with predictions as to how people might behave, and allowing for counterfactuals (scenarios of what might have happened) and also being consistent with what we know of politics, economics and sociology. Connectionism, which is where parallel computing mimics the parallel processing of the brain is discussed at length in this book. The author himself cautions about the gulf between connectionist computers and brains. The failure of the connectionist idea to gather much more momentum in recent years, or of artificial intelligence/robotics to use it to make the decisive break through to intelligent autonomy suggests that this is not the route that will explain the brain's functioning, let alone consciousness. Enthusiasts of connectionism have argued that connectionism disproves the 'folk' idea of causal beliefs, but the author suggests that this argument rests on unjustified assumptions about how such beliefs would be distributed in the brain. The author also takes up the idea that has emerged strongly in this century of embodiment or the interaction of the body with the brain's functioning, and also the idea that perception and action interact with one another, rather than perceptions emerging fully formed and then being acted on. However, in looking at the question of embodiment and the body's interaction with the brain, the author appears to favour the idea that more abstract cognition, such as planning next year's holiday, is separate from bodily inputs, while a more determinist view has been that even these areas of cognition were governed in much the same way as walking or reaching out to grasp something. There is only one place in this book where much attention is given to modern neuroscientific research. This relates to the work of Milner and Goodale (1995) on the different visual systems that relate to 'blindsight' where patients have no consciousness of vision, but can guess the location of objects at well above chance. Milner and Goodale's conclusion was that a dorsal processing stream controls visually guided action, such as reaching, and is unconscious, while a ventral stream dealing with categorisation of objects is conscious. Milner and Goodale used the Tichener circles illusion, where subjects regularly misjudge the size of discs, to show that when subjects reached to grasp the discs they in fact correctly judged the size of the discs, despite being consciously subject to the illusion of an incorrect size. This argued for the separate functioning of the ventral/category/conscious and dorsal/location/unconscious streams. Similarly, patients with ventral stream damage cannot identify objects, but can reach to grasp them. From the point of view of future consciousness studies there is an important point here. If the ventral stream is conscious but the dorsal stream is not, an understanding of the physical difference between the two might yield an understanding of the physical basis of consciousness. Consciousness – getting to it at last: The appendices of this book provide a useful summary of mainstream ideas on consciousness. The author takes the view that in looking at the question of consciousness, it is only the existence of qualia or raw subjective experience that cannot be explained in terms of cognitive science and information processing. This might be seen as a part of the normal mainstream process of trying to squeeze consciousness out of the picture. However, I think it is in fact a helpful approach, in that it can prevent consciousness studies from coming bogged down in a lot of areas that are not specifically to do with consciousness, and worse still the possibility of consciousness apparently being explained away as a result of dealing with one of these peripheral areas. Jeffrey Alan Gray, in his book, 'Creeping Up On The Hard Problem', came to much the same conclusion that we need only address qualia, and hence much of the baggage train of consciousness studies can be dispensed with. In a way, this is not that far from the ideas of Ned Block who splits consciousness into an information related access-consciousness and phenomenal consciousness. He thinks the mistake of much of consciousness studies is to deal only with the explicable access consciousness. The author also looks at 'representationalist' approaches to consciousness. These fall into 'simple' and 'higher order' forms. The simple form seems to be demonstrably wrong, claiming for instance that tissue damage is the same as pain. Even in the most reductionist account, we know that pain relates to electrical and chemical processes in the somatosensory cortex, and therefore it is not actually the same as the remote tissue damage. We could say that pain is the same as the signals in the cortex, but then we are back to the hard problem, as to why these are more than warning signals in a computing device, but come instead with the added subjective quality of the experience of pain. The 'higher order' variety is also known as the 'higher-order thought theory'. The idea is that the neural state, such as the cortical response to tissue damage, becomes conscious when it is itself the subject of thought. This idea looks to have the same quality as arguing that if two film cameras are pointed at one another, they should become conscious. Two unconscious processors somehow sum together to a conscious process. If we point non-conscious thought at non-conscious processing of tissue damage, we require some additional physical process to be generated by the interchange of the two, with that additional process being conscious, but no such explanation of a physical process is advanced with this theory. Dennett again: Finally, the author appears to favour the ideas of Dennett more than most others with respect to consciousness. The discussion here focuses on what is described as the 'narrative twist.' Consciousness is suggested by Dennett to be an illusion arising from immersion in human culture. This begs the question of animal consciousness, where at least in mammals most of the evidence of brain structure and behaviour argues in favour of conscious experience. Dennett's approach appears similar in some respects to a version of the 'higher order' theory. The human brain is suggested to be a 'virtual machine' meaning that humans can make 'cognitive objects' of their own thoughts, and from that create a narrative about the self and our freewill reasons for doing things. These are deemed to be the qualia. In fact, it is something of an evasion to define qualia in such a complex manner. Qualia are normally related to individual basic experiences such as the colour red, the taste of chocolate etc. and although actual conscious experience usually involves a combination of such qualia, the qualia are not usually discussed as being something complex as a life narrative, even though as a subject of experience it does look fair to also view these as qualia, albeit ones that are more confusing for the reader to think about. Dennett's approach to qualia is subtle in the way that it leads the reader away from the simpler experiences that are undeniable subjective happenings towards areas that are relatively difficult to comprehend and where subjectivity starts to get mixed with information that could be expressed non-consciously. However, the more complex Dennett form of qualia, and the only type of qualia discussed here are deemed to be something extraneous to the actual functioning of the brain, and to be illusory, and produced by immersion in culture. I find it hard to grant much credibility to these ideas. Even if our experience is a convoluted illusion of this kind, we still need a subjective consciousness to support the illusion, and Dennett in fact offers no suggestions to the physical basis of the supporting consciousness. His system presents a further problem in terms of conflict with evolutionary theory. The brain is energy-intensive and it is likely to require quite an amount of energy to run the Dennett virtual machine, and yet we get no indication here of what would be the adaptive advantage of such a machine. 2.) IDENTITY THEORY A few thoughts on identity theory: Firstly, I think that mind/brain should be replaced by consciousness/brain when discussing anything in this area that requires a certain amount of rigour. Mind is a word that seems peculiar to the English language and appears to cover, the brain, mental processing such as thoughts and response to stimuli and consciousness. It's a useful word, but it does allow a certain amount of fudging where you want to specifically talk about consciousness. A good example of identity theory is to say that because we accept that H20 is the same thing as water, so that there is an H20/water identity, we should accept that there is identity between consciousness of for instance pain and the firing of particular neurons. I don't think this works, although it can look superficially attractive. In the first place, why are we convinced by the H20/water identity? This is because what we know about the electromagnetic bonds between the hydrogen and oxygen atoms and between the individual H20 molecules gives a full explanation of the way the substance known as water behaves, with for instance liquidity at ambient temperature and the ability to interact with biomolecules. The key thing is that the H20 understanding explains the behaviour of the water we see both in everyday life and neuroscience. If water regularly did other things than are generally observed, such as brandishing the sword Excalibur all of itself, or defying gravity, the identity explanation would fail because the electromagnetic bonds in H20 do not allow for that behaviour. I think mind/brain identity or more specifically in the example above, consciousness of pain/neuronal activity identity fails on this basis. Identity theorists may invites us to think of the neuronal activity at the detailed level of electrical potentials and molecules. So we ought to be able to see how electrical potentials, molecules and any other details produce the observed characteristics of consciousness in the way that electromagnetic bonds produce the observed characteristics of water. However, it is not possible to do this within classical physics. Kelvin was right, in terms of classical physics, when more than a century ago he said that it had more or less all been done. We know how particular atoms, molecules and electrical potentials interact and what these interactions produce, such as forming compounds or transmitting electrical signals, but conscious experience is nowhere on the list of classical physics. It is no use in this context appealing to special conditions in the brain, because the whole thrust of modern science is to the effect that there is no dualism or vitalism, and that brain obeys exactly the same laws as the rest of the universe. Identity theory also looks to have a problem with the distinction between conscious and unconscious processing in the brain. The two forms of processing suggest some physical difference, but identity theory as usually described is identical to electrical potentials and molecules that exist in all parts of the brain. The failure of most identity theory seems to stem from concentrating explanation at the classical level, which can only produce a known limited range of behaviours. If we do eventually find a physical explanation for consciousness in the brain, it may in fact be possible to speak of an identity between consciousness and some particular brain states. |
| Email: quantum-mind@blueyonder.co.uk Copyright © Quantum Mind 2012. Web design by 7Soft.co.uk. |