The knowledge argument. Is it successful?
Is an emotion a passion?
Could there be artificial intelligence (AI) in what Searle calls strong sense?
Consciousness appears to be essentially intrinsic and ineffable. No amount of investigation into neurological states and no amount of questioning can reveal how it feels for a given person to have experiences he or she is having. Although it can be assumed that perceptions, sensations, thoughts can be explained with help of exhaustive scientific research into the nature of those phenomena, it seams obvious that it is not possible that the feel of conscious experience can be understood using the same methods. Explaining why physicalistic account is inappropriate faces difficulties due to the same private nature of analyzed phenomena. If it is not even possible to describe the subject of investigations then how one can start argue about it? Nevertheless, Frank Jackson devices a thought experiment to build an argument demonstrating that physicalistic approach is ill suited to explain conscious experiences.
Briefly, though experiment is constructed as follows. Imagine Mary, she knows all possible facts about colour vision, not only facts accessible to today’s scientists, but everything – all there is to know about colour vision and all related issues. Although she is not colour blind, she lives in the environment where all colours, from outside colour full world, are accessible only after they are converted to black and white spectrum. Those are the only colours she sees. At some point she gets out to normal environment and sees, let us say, a yellow banana. At this point it is no surprise to hear from her something like “so that’s what it is like to see yellow colour”. It is hard not to agree with Jackson that it is simply obvious that she has learned something new about colour despite all that complete physical knowledge she was familiar with. Now she knows a new fact – the feel of yellowness.
Based on this thought experiment Jackson builds the following ‘knowledge’ argument. He starts with a premise that Mary knows all physical facts and if physicalist is right there is nothing possibly more to know. Then he demonstrates that under certain conditions she will obviously gain some extra facts. Therefore her knowledge was not complete. Since she knows everything about colour physics, the new knowledge she gains must be of non-physical nature. This way knowledge argument aims to establish two claims. Firstly, that conscious experience can only be explained if we accept that there are facts that are explainable in terms of non-physical properties. This necessarily leads to second statement that physicalism must be wrong since it postulates that all possible facts are explainable in terms of basic physical properties. Based on knowledge argument, Jackson puts forward explanation postulating property dualism. It does not deny that physical properties have no explanatory power, but it introduces additional non-physical properties that are needed to form complete picture of consciousness.
Before analysing possible responses to this though experiment it is worth to note that its validity is doubtful. For how one can imagine someone having all the possible knowledge about colour vision. It is simply not conceivable. Imagining what present day scientists know about this topic is hard, let alone knowing what kind of information will be available with completed physics. Dualist can respond that we know what kind of facts physical enquiry can reveal. We can imagine them and we know that no mater how many facts of that kind will be found they will simply be of the wrong kind. But this is not necessary true. There is plenty of historical evidence discrediting such generalizations. Take explanation of emergence of life. It was claimed that it can not be explained in terms of scientific research, but it turns out that science is capable of providing satisfactory explanation. Therefore, by analogy, science might reveal some new facts explaining at least a fraction of conscious experiences (Frankish 2005, 77). Although the very validity of thought experiment is doubtful, it can be assumed that it is not critical to the argument and demonstrate further problems it faces.
Physicalist can accept Jackson’s premises and demonstrate that conclusions that are reached at the end of the argument are incorrect. There are few different responses following that strategy. It can be claimed that Mary learns no new facts, but gains some other type of knowledge describing how the facts can be used. That is Mary gains ability to use her facts about yellow colour in a new ways like imagining, remembering and recognizing (Lewis 1988 in Frankish 2005). Another line of response again states that she learns no new facts, but claims that she acquires new perspective on the facts she knows. That can be presented under different concepts that do not necessarily have the same meaning. Before her release she had no concept of ‘yellowness’ to use to interpret facts she knew about colour yellow. After her new experience she is able to conceptualize knowledge from previously non-existent perspective (Tye 1995 in Frankish 2005). Both responses accept that Mary will learn something new, but claim that this knowledge will not come in a form of propositional facts. They accept that knowledge argument is successful in demonstrating that there is something special about conscious experience, but deny that it can not be physicalistically explained.
Another line of response is quite different. It takes radical view, denying that knowledge argument demonstrates that Mary learns something. To the contrary it insists that she learns nothing new, at least if conditions set in thought experiment are strictly followed. Daniel Dennett, explaining his no-learning view, states that Mary will not be surprised by what she will see when stepping outside her house (Dennet 1991 in Frankish 2005). She will know what it is like to see yellow banana. Again – she knows all there is about colour vision. Including information about physical effects it will have on her behaviour and neurological states. This means that, if someone will try to play a trick on her showing blue banana, Mary will not be fooled. She will be able to tell the difference, because she is able to discriminate between yellowness and blueness.
Such mechanistic explanation of Mary’s abilities can be started from the fact that she can manipulate complete set of relevant information. Given this it is possible to assume that she is able to construct effects of seeing yellow colour, possibly based on connections with other related concepts. For example, she will know that experiences of yellow often happen together with experience of warmness. She will know what effects colour has on human behaviour and so on (Frankish 2005, 77, 162). All this combined with the fact that she is able to observe huge number of such connections and inferences gives at least some grounds for Dennett’s claim that she can discriminate between those special feels attached to her colour experiences.
Comparing no-learning view with dualist position, the latter has tremendous appeal of intuitiveness and clarity. Even unconceivable task of imaging someone knowing all the possible facts of completed physics, appears plausible compared to Dennett’s claim that Mary will be able to work out feels of experiences. This counter-intuitiveness of no learning view is a direct result of attempt to play according to the rules set out in the conditions of Jackson’s thought experiment (Dennett 1991 in Frankish 2005, 217). Knowledge argument relays on ‘obviousness’ of its claims and unrealistic requirements are easily overlooked: absolute knowledge; possibility to get all that information through black and white vision. Dennett accepts conditions of though experiment and extends its scenario demonstrating that issues connected to those conditions are not as easily dismissible as it might appear to be. Thus the counter-argument against no-learning view based on intuitiveness, if intuitiveness can be taken as a viable form of argument at all, is in fact argument against dualist position itself.
Dualists can accept that Dennett is right and it is possible to work out what colour experiences one is having. They will agree that Mary will not be fooled by blue banana. Nevertheless, they will argue that the issue of experiencing blueness as some kind of absolutely novel experience is not addressed. The problem with no-learning view is that it fails to explain the feel of experience. It is a theory about something else but not about consciousness. Mary learns new aspect of colour vision. She learns what it is like to have an experience of all those facts she knew before leaving her black and white room.
Now this leads to rather puzzling characterization of consciousness. Information form sensible world is gathered through our sense organs, which is later translated into form ready to be used in our cognitive activities. Vision, for example, works through translating waves reflected from outside surfaces into mental imagery. This in turn is reflected in our behaviour and thoughts. In this Mary is no different from any of us. While being in her black and white environment she is able to use colour names in phone conversations with her friends in colour full environment and nobody there can tell that there is something wrong with Mary. The feel part then requires some other extra layer which translates already translated information into the form of what-it-is-likeness (Dennett in OU CD5, 23). Further more translation must be made in centralized fashion for we are experiencing one single flow of consciousness at a given time. This concludes in the description of consciousness as some kind of centralized translating agency residing somewhere in the brain and playing no role in describing our behaviour and thoughts. Jet there is nothing in this theory that describes how and why this extra agency functions.
The goal of knowledge argument is to demonstrate that physicalist fails to explain the feeling of experience. One can have all physical information there is to have and still lack some extra knowledge about sensible world. Additional non physical properties are needed to close this gap. The argument appeals to common sense – “it seams just obvious she [Mary] will learn something about the world” (Jackson 1982 in Frankish 2005, 196), but its explanatory power does not go beyond this. It only creates ‘intuition pump’, in Dennett’s terminology, that forces this feel of obviousness, but is unsuccessful in demonstrating how and why introduction of extra non-physical properties can explain conscious experiences. Therefore knowledge argument does not succeed in its attempt to demonstrate that physicalism can not by right. Since, Jackson’s premises and assumptions are grounded solely in intuition, it makes the entire project more similar to “just a little story” (Dennett in OU CD5, 21), than to a strong philosophical argument.
Frankish, Keith (2005) Consciousness, Bath, Bath Press.
Dennett, C., Daniel (1991), Mary and the blue banana, in Frankish, K., ed. (2005) Consciousness, Bath, Bath Press, pp217-218
Dennett, C., Daniel in Open University AA308 CD5, track 21
Dennett, C., Daniel in Open University AA208 CD5, track 23
Jackson, Frank (1982), The knowledge argument, in Frankish, K., ed. (2005) Consciousness, Bath, Bath Press, pp193-196
Lewis, David (1988), The ability hypothesis, in Frankish, K., ed. (2005) Consciousness, Bath, Bath Press, pp219-221
Tye, M (1995), Mary’s room, in Frankish, K., ed. (2005) Consciousness, Bath, Bath Press, pp222-224
Žygimantas Medelis
It is certainly possible to claim that an emotion has no element of a passion or, in opposition, to assert that an emotion is a passion. Both views will lead to accepting a set of properties, which will have to be attributed to the mechanics of emotions. There I will demonstrate that it is not possible to provide a coherent account interpreting emotions only within one of those radical frameworks. The reasons why such accounts are not sufficient will also show in which sense emotions are passions.
Often, if not always, uncontrolled bodily changes follow emotional states. If ashamed, one cannot help blushing no mater the efforts to stop it. Often, when we are frightened we talk in lower voice, without noticing it. Those and number of other similar situations, build a common perception of emotions as something that has influence on our physiology and we have at best very limited control over those changes. That is, emotions ‘happen’ to us. This conception involves two separate claims about the way emotions are functioning. First, that we have no voluntary control over our emotional states and second, that emotions have influence on our physiology and possibly psychology. Then, following Roberts (Roberts 1988, in Price, 217), those can be treated as a pool of conditions that need to be accounted for if emotions are to be called passions.
Let us assume that emotions are passions. Meaning that we are not capable of controlling emotional states and/or we are involuntary influenced by them. Before evaluating implications of this view, it is necessary to understand what is meant by individual’s ability of self-control. Beliefs, thoughts and desires are all mental states that we are able to control. We can choose to think about certain set of problems instead of the other. We can form our beliefs and desires based on our experience, value and moral systems. In other words it is all the properties, over which we have full control, that constitute our cognitive system. If emotions are understood as involuntary occurrences that happen to us and we are not able to influence them, then it must be accepted that they are functioning outside cognition. Experiencing anger is then similar, but not equal, to perception of a loud bang somewhere outside my room. I can not help not to hear this bang and the same is with anger, or any other emotional state.
William James (James 1890, in Price 2005, 194) puts forward an argument that describes emotions as functioning in this manner. He assumes that perception of some elicit event, say a phone informing that I won in a lottery, automatically leads to corresponding physiological changes. The perception of those changes, a feeling, then is an emotion. In this scheme there is no room for rational action, bodily changes and perception of those happens involuntary, with no intervention from cognitive apparatus.
James theory supports view that emotions are passions, but it has some flaws. Firstly, all it states is that emotions are always accompanied by involuntary physiological changes. It does not eliminate possibility that controllable cognitive states are involved in emotional experiences through playing a role in causing bodily changes. Secondly, intentionality of emotional states cannot be fully explained using this account. Although they are intentional in a sense that emotions are about feelings, it is not a proper explanation of emotional intentionality. The set of bodily changes that happens to me when I receive the news about lottery is exactly the same, regardless of what I am happy about. I can be happy about me becoming rich or I can be happy about that charity organization and its cause for which I am going to give all the millions. This theory allows element of passion in to account of emotion, but the explanation itself fails to provide complete and coherent picture.
Yet it is not necessary to dismiss James idea altogether. The discussed physiological changes need to be understood in terms of autonomous modules that operate in the context that has no direct connection to conscious though. Paul Griffiths discusses emotions in this framework. He claims that emotions are complex events involving patterns of bodily changes, thoughts and coping behaviours (Griffiths 1990, in Price 2005, 222). The important bit in his theory, in the context of present discussion, is that emotional occurrences function within environment of informationally encapsulated modules. By encapsulation Griffiths means that those modules are not a part of cognitive system and cannot exchange information with it. Therefore everything that happens in that module cannot be caused or influenced by conscious thought. Further more this account solves problem of intentionality. Emotional responses – ‘affect programs’ are triggered by appraisal systems that are intentional that is they represent objects in surrounding environment (Price 2005, 57).
Griffiths provides necessary amendments to push forward James idea that bodily changes are involuntary. Further more he asserts that those changes are the necessary part of emotion. Therefore emotions are passions, but with important qualification, only part of emotional state, the one involving physiological changes, is a passion. Yet it is still not clear if this part of emotion, which seams to be outside the realm of control, truly belongs to notion of emotion or that there is no other important dimensions of emotional experience that function outside modules. Yes, I will involuntary step back if someone will imitate hitting me, although I know perfectly that he will not do that. Affect program will be triggered by appraisal system that has evaluated situation as dangerous. But am I truly afraid in this situation? To be afraid in this case I have to make an evaluative judgement that situation is dangerous and in this case I can not do that because I certainly know that I will not be wronged. It seams that it is not possible to account for emotions in terms of physiology only.
Robert Solomon’s (Solomon 1973, in Price 2005, 198) cognitive theory claims that emotions can not be called passions in any sense. Not only because he dismisses idea that bodily feelings are part of emotions, but mainly because he assumes that emotions are evaluative judgements and so are fully controllable. Consider the same example of winning in lottery. The one, who wins, will be happy only if he will be able to grasp the consequences of this event. Namely he will be able to judge that it is a big amount of money that allows acquiring needed things. Someone who is already rich will evaluate that amount of money, to be received, as making no significant influence to his purchasing abilities and therefore will remain indifferent to the news. In this view emotion is controlled by reason. Even if there are some bodily feelings, like initial increase of heart beat, after our rich man received the news it can not be called emotions. Stoics name it ‘first movement’, some initial physiological states that were triggered by perception, but they are not emotions.
Cognitivist view rejects that emotions are passions in the sense discussed within context of Griffiths and James theories. As a consequence, issues that are easily explained in those theories, pose serious problems for Solomon’s explanation. Griffiths can easily explain why emotions conflict with judgements, which is hard to achieve within the framework of cognitivist account. In the latter view emotions are judgements, thus my beliefs about environment I am in must match my emotional state. This is not always the case. Race car driver might judge that to make the next turn at such and such speed will be extremely dangerous, but feel no fear. Informationally encapsulated modules, in Griffiths theory, explain this situation very clearly. Functioning of cognitive system and that of module are not synchronized in this particular case. Next problem is the intensity of emotional states. Judgments do not vary in intensity, but emotions do. It is again easily explained within the theory that allows that emotions are passions in a sense of involuntary bodily changes that can be bigger or smaller and so trigger different intensity of emotional experiences.
Then it is clear that accounts attempting to explain emotion only in terms of involuntary physiological changes or in contrast explaining only in terms of cognition will yield unsatisfactory results. On the other hand it is also evident that both: ‘as judgement’ and ‘as passion’ dimensions must be provided in coherent account of emotions.
Observations made by C.T. Warner can provide needed connection between those two accounts. It allows explaining cognitive states and passions both within the same framework. He notices that emotions have self-perpetuating quality (Price 2005, 145). Take emotions of anger, fear, joy even emotional attitudes like jealousy or love. All of them are experienced in the increasing manner, if they stop to perpetuate then they disappear fast. Anger for example. From some kind of mistreatment to its resolution in violent actions, we can observe variety of phenomena that increases in intensity: faster hart beat, shouting, hitting something. Emotional occurrence then is a series of events each triggered by potentially different affect-program. Cognition pools those events into emotion. Then, following Seneca, it is possible to interpret affect-program, functioning as a passion and forcing the mind to focus on elicit event that triggered emotional response (Price 2005, 180). Under those conditions mind makes evaluative judgements, but constrained by the frame that limits possible choices only around current problem. This frame itself is not directly controllable by the mind as well. Solomon notes that judgments, he describes as emotions, are fast and therefore inaccurate (Solomon 1973, in Price 2005, 206), which supports this assumption. Actions motivated by rash decisions will either conflict with module or will not help to resolve emotional situation. In both cases it will lead to further perpetuation of emotional state, so repeating this cycle until it is resolved.
Emotions are complex events that are shaped both by affect-programs and judgements that are made concerning eliciting situation. Interaction between those components develops in self-perpetuating manner, which is not directly controllable as well. Therefore emotions are passions in two aspects. Firstly, in a sense of involuntary changes triggered by affect-programs and secondly in a sense of self-perpetuation which leads to partly uncontrolled perpetuation of emotional state. Although it is important to stress that cognition plays important role in the way emotions are functioning, through pooling together involuntary responses and actions into single emotional occurrence.
Price, Carolyn (2005) Emotion, Bath, Bath Press.
James, William (1890) ‘Emotion follows upon the Bodily Expression in the Coarser Emotions at least’, in Price, Carolyn, ed. (2005) Emotion, Bath, Bath Press, pp194-197
Solomon, Robert, C. (1980) ‘Emotions and Choice’, in Price, Carolyn, ed. (2005) Emotion, Bath, Bath Press, pp198-207
Griffiths, P.E. (1990) ‘The Affect-program theory’, in Price, Carolyn, ed. (2005) Emotion, Bath, Bath Press, pp222-229
Thought and Experience (2005) CD2, The Open University
Žygimantas Medelis
“A machine can think simply as a result of instantiating a computer program” (Wilkinson, p100) is a central claim made by strong AI. The problem of AI realization is not the matter of available resources or political and ethical will, as Tim Crane puts it. Rather it is the question of the very possibility of artificial intelligence. If the core concepts behind this project are fallacious then all attempts to build such thinking things are doomed even if backed by all available resources and political support. Number of assumptions are made to make the core arguments for strong AI to work. At least some of these are suspicious, since they involve conceptual confusions.
Strong AI argument outlined above rests on number of rather specific definitions, which are interconnected and hard to explain in separate meaningful blocks. Thus the rest of this paragraph deals with definition of machine, thought and algorithms as they appear in AI discussion. Only a certain type of machines can think in AI sense. Machine needs to instantiate an algorithm, which in turn generates its minds. Therefore only the machines that are capable of algorithmic calculations are discussed within AI arguments. Digital computers are the most widely spread type of such devices, but it is important to stress that those devices are only one possible implementation of 'computing machines'. Therefore the latter term is more appropriate and reflects the main requirements for the machine to be a body of AI. Computing machines are only needed to execute mind generating algorithms. The material those machines are made of, be it silicon transistors, human brains or beer cans, is not an essential feature of thought (Wilkinson, p103). This leads to the concept of mind or thought. As already mentioned, thoughts must be formulated in terms of algorithms. To think algorithmically would mean that each individual though, like “get the cup of water to drink”, can be split into finite number of simple steps. Upon successful completion of each of those steps, say finding a cup, there is no ambiguity what to do next, that is to pour in water. Note that the thought was not formulated as vaguely as “I want to drink”, since this involves a number of ambiguities: whether to get a cup or a bottle, to pour in water or milk, and so on. Further more, there are thoughts like “satisfy thirst”, which are hardly expressible in terms of algorithmic procedures. When the procedure involving “satisfaction” must end is dependable on number of subjective properties. Algorithms can be specified only in terms of symbols (or representations that stand for things) and strict true-or-false type rules. Therefore, Searle concludes that “computer programs are formal (syntactic)”. Important point is that according to strong AI, all there is to thinking is manipulation of formal symbols.
Therefore, if all the mental contents, namely, sensations, emotional life, imagination, thoughts, memories, and mental causation can be reduced to sequences of algorithmic procedures, then strong AI, as described above, can be produced. Such kind of “algorithmic” mind, though, does not account for certain properties that are found in human mental states. Latter are always about something. Minds always have some content. Meaning is the construct not to be accounted in terms of syntax only: “If a symbol system has meaning, if it is about something, then it is said to have semantics” (Wilkinson, p105). This lack of semantic properties in computer program renders it fundamentally different from human mind. Searle uses China Room thought experiment to illustrate this dissimilarity (Searle, p204). His experiment boils down to demonstration that successful symbol manipulation (syntax) does not involve actual understanding (semantics) of what this system of symbols mean. That is, syntax alone can not produce semantics. No matter how complex the calculations involved are, understanding can not emerge solely from them. But the truth of this Searle's assumption is not obvious. As complexity of the system grows, it is possible that it will start to demonstrate some new qualities (Wilkinson, p113). Let us take the regular cloud. Up to certain concentration of water molecules, it is just a white blob in the blue sky. As the system's complexity increases, i.e. more and more molecules are accumulated in the system, it produces a new phenomena – rain. By analogy, the same kind of generation of emerging properties can occur in China Room. That is understanding of Chinese can emerge from overall complexity of Chinese symbols manipulation. For this argument from analogy to work somehow it must be demonstrated that the issues discussed are truly analogous. Firstly, phenomenas similar to described above are observed in nature. The cause why system displays those new properties can be explained only if all the elements of the system and their inner workings are well known. That is, workings of gravitation between water molecules are explained and therefore rain is easily accounted for. However, in which ways the possibility of those emerging properties must be designed in to artificiality build systems is not clear at all. Especially when the phenomena of mind, that must emerge, is not yet explained. Secondly, the emergence of new properties, as the system gets more and more sophisticated, is a contingent fact. Empirical observations are all we have to back this claim. It does not follow that it is logically necessary for emerging properties to manifest themselves in complex systems.
Nevertheless, let us assume that hypothesis of emerging properties is true and the meters could be so that semantics can emerge from complex syntax. Still strong AI as formulated by Searle is problematic. From the Church-Turing hypothesis (Wilkinson, p102) follows that computer programs can be executed on various types of machines that satisfy requirements outlined above. From this it necessary follows that programs are something entirely different from the matter the machines are made of. Hence, the world is explained through two entirely different substances. Problem of mental causation, already found in Cartesian dualism, emerge. How those two systems interact, if they are two different substances? Most probably the attempts to explain that will result in having difficulties Descartes was faced with. He had to introduce God – concept of immense complexity – in order to keep his theory going. It does not mean that AI theorists will need God to build a thinking thing, but there is a possibility that construct, equally resistant for articulation, will be needed. To stress it again Cartesian dualism and AI dualism bear some similarities but they are not identical. This argument by analogy can be accused of the same weaknesses as the argument of emerging properties was, but the similarities between the latter positions are more evident, after all both are different theories dealing with the same mind-body problem. Even if latter argument can be refuted and it can be demonstrated that AI needs no sky hooks to work, there is more fundamental problem project under discussion is faced with. Introduction of dualism of any kind is at odds with the grounding principles of scientific enterprise of which AI claims to be a part (Wilkinson, p115). This demonstrates that there are some serious inconsistencies in the AI research program.
One possible reason those inconsistencies emerge, can be explained in terms of separation between 'knowledge how' and 'knowledge that'. Any theory explaining mind-body problem needs to include explanation of both types of knowledge. AI as one of those theories is not an exception, but it faces a problem that everything it needs to account for must be described in terms of calculable procedures or algorithms. 'Knowledge that' poses no problems, since it consists of propositional statements like “circle is a closed curve with all its points at an equal distance from the centre”. 'Knowledge how' can not be expressed in propositional statements. It constitutes our skills and abilities. Rule sets of propositional statements describing how to write poetry for everyone who wishes to be a poet are unimaginable. A poet, if asked, will not be able to produce such “Poetry manual”. Even if produced, it would not work for all “poets to be” if for anyone at all. “Knowing how to do something is at least in part a non-propositional aspect of human knowledge” (Wilkinson, p123). AI is a scientific theory and as such must use scientific method and nothing else. The core principle of scientific method is “discovery of objective, essentially spatiotemporal properties of the physical world which could be mathematically described and related by general laws” (Nagel, p220). The mental contents of 'knowing how' are simply outside the realm of application of scientific method. Human experience is grounded in meanings taken from rich body of culture and personal developments. All of which is classified in incomputable hierarchies and accompanied with unique ability to select from that wast knowledge base only the items that are relevant in particular context. All of those issues: the meaning problem, the ontology problem and the frame problem, representatively, are approached by AI using inappropriate methodology. The non-surprising outcome of the project is conceptual inconsistencies.
Showing that AI research program is inconsistent is only the one line of possible exposition of problems within strong AI. Namely the problems related to aspects of subjective human nature like: qualia, aboutness and consciousness were not examined in more detail, except in relation with analysis of semantics–syntax relations. Nevertheless, those issues pose serious difficulties for AI. The issue exposed by Searle in his China Room though experiment, that syntax can not generate semantics, outlines the reasons why realization of strong AI is problematic. Firstly, it is hard to demonstrate how meaning emerges from syntax only, unless emerging properties hypothesis is right. Secondly, the very attempts, made by AI to account for all there is in the world by syntactical constructions only, exposes its inconsistencies. Therefore, the existence of artificial intelligence, in what Searle calls strong case, is not possible. That is, we will not produce digital computers that will demonstrate mental contents as the ones possessed by human beings.
Bibliography
Wilkinson, R. (2002) Minds and Bodies, Alden Press Ltd
Searle, J.R. (2002) Is the Brain's Mind a Computer Program? in Minds and Bodies, Wilkinson, R.
Thomas, N. (2002) Consciousness and Objective Reality in Minds and Bodies, Wilkinson, R.
A211 Audio Cassette 5 side 1, Open University