International Journal of Psychophysiology, 14 (1993) 179-187 0 1993 Elsevier Science Publishers B.V. All rights reserved
Consciousness: criteria and possible mechanisms A.M. Ivanitsky Institute of Higher Nervous Activity and Neurophysiology, Russian Academy of Science, Moscow (Russia) (Accepted
Key words: Consciousness;
There are two sorts of criteria of consciousness - objective and subjective ones. They are the ability for operating with the knowledge which could be addressed to other people in the abstract form and the feeling of ‘ego’ as a ‘host’ of these actions, respectively. These two signs of consciousness are provided with the mechanism based on the synthesis in the brain structures of different kinds of information: sensory information, the data extracted from the memory and the signals from the centers of emotions and motivation. As a result of this synthesis, the sense of ‘ego’ arises and the message designated for others is determined. A significant role in the informational synthesis is played by dynamic cortical structures - foci of interaction. In perception they are localized predominantly in the projectional cortex, in thinking and in the associative areas. Realization is closely connected with communication and appearance of the interaction foci in the verbal zones of the left hemisphere. Pavlov (19511, in his program lecture in Madrid early in the present century, said that he saw the final aim of his study in the revealing of the mechanism and the inner vital sense of human consciousness. It is important, that Pavlov placed the words ‘mechanism’ and ‘vital sense’ near each other, i.e., he considered that the sense of consciousness would be realized through revealing its mechanism. This insight attracts our attention now, when the problem on consciousness mechanism is in scientific plans and the search for the meaning of life by the end of the twentieth centry, one full of dramatic events, acquires a special value.
OBJECTIVE AND SUBJECTIVE OF CONSCIOUSNESS
While studying objective and subjective criteria of consciousness, it is necessary to determine as accurately as possible the subject of study, i.e., the signs which discriminate the consciousness from other psychic manifestations. The difficulty is in the fact that the problem of consciousness is interdisciplinary by its nature, it stands at the very junction of humanitarian and natural knowledge. The state of consciousness can be estimated from the positions of objective observer, as
Correspondence to: A.M. Ivanitsky, Institute of Higher Nervous Activity and Neurophysiology, Russian Academy of Science, Butlerov Str. 5a, Moscow 117865, Russia.
if from the ‘outside’, and introspectively, from the ‘inside’. Only by coordination of these two approaches may we consider that a sufficiently complete description of consciousness signs has been found. Objective sign of consciousness was the most accurately formulated by Simonov (1990). His definition of consciousness involves the operating, mastering of knowledge and. the ability to direct information from one person to another in the form of abstract symbols (words, artistic images, etc.). It should be added that these symbols must be learned or discovered by the subject, they are not innate. This addition provides the opportunity to eliminate relatively simple kinds of communication characteristic, for example, of bees and ants from consideration as conscious
behaviour. It is also noteworthy that application of abstract symbols, in comparison to the inborn ones, allows a dialogue form of communication, where the information, which was not immediately comprehended, can be comprehended and added in the course of mutual communication. This objective sign of consciousness is considered to be sufficiently accurate. Despite the relative simplicity, it actually embraces all forms of conscious behavior. It is of interest that this objective sign of consciousness may be compared with the so-called ‘Turing criterion’ (19.50). The latter was introduced to determine the ability of a machine, i.e., a computer, ‘to think’. According to this criterion, a machine could think if it had the ability to be engaged in dialogue with the man, who, having no opportunity to see his collocutor, could not judge whether he contacted another person or the machine. The main difference of the ‘Simonov criterion’ from the ‘Turing criterion’ is that if a man, in contrast to machine, not only may but wants to say something to his collocutor, he possesses the definite motivation, connected with satisfaction of his need. This need in the most common sense is directed to life maintenance, i.e., to self-development, with consciousness being the higher manifestation of this process. The second sign of consciousness as it has been already said, is based on the introspective approach to consciousness. Thus, if in the first case we speak about ‘another’s’ consciousness, in the second one, it is about our own. In general, introspectively, consciousness is the sensation, awareness of our life, our being in the environment. It is formed from realization of our ability to apprehend, understand the surrounding, conform it to our needs and find the ways for their satisfaction by performing definite actions. Thus, consciousness comprises perception, thinking and emotions. All these psychic manifestations are unified by the feeling of our own ‘ego’ as a subject of our own experiences (feeling) and, in a sense, a ‘host’ of our deeds. For Leontyev (19811, consciousness differs from lower organized forms of psychic activity by, namely, distinction of the ‘ego’ itself from the surrounding world, a state, in which the outside world is as if presented to the
subject. Leontyev figuratively wrote that during this process ‘thought about the book interflows neither with the book nor the experience of this thought’.
One may suppose that these two approaches to the problem of consciousness - objective and subjective - are not very similar. However, a detailed consideration of this problem brings us to the conclusion that the connection between them is rather close. It is expedient to begin such consideration with the problem of correlation between consciousness and speech, as the objective sign of consciousness - ability to direct information transmission in the form of signs - is first of all associated with speech, which is the most highly-developed method of interpersonal contact. There is also direct evidence that return from a state of coma and restoration of verbal contact with a patient coincides with the moment of the establishment of interaction between the motor verbal cortical zones and the left low-temporal area, as it is revealed in the coherence in their electrical potentials (Grindel, 1985). Sidorova and Tsyganok (19891 in patients with local lesions of the cerebral cortex demonstrated that damage to the frontal area of the left hemisphere led to a disturbance of mental reproduction of emotions and lesions of the left temporal lobe - to a disturbance of emotions identification. In both cases, voluntarily-controlled components of these functions were disturbed. Their involuntarily components in the form of vegetative shifts in damage of the zones of the left hemisphere were not disturbed and even became more expressed. Lesion of the frontal and temporal parts of the right hemisphere, in contrast, decreased the involuntary and increased the voluntary components of reproduction and perception of the emotions. It could be seen that even evolutionary more simple, in comparison to speech, forms of communication (emotional expression and impression) in their voluntary part were connected with the verbal zones. It should be noted that later this conclusion was also con-
firmed on healthy people by using a new method of brain biopotentials mapping - mapping of the intracortical interaction, which will be considered below. At the same time, the above data demonstrate only a close connection between consciousness and speech. It is not yet a ground for suggesting their identity. Some of the clinical data, obtained in patients with the lesions of the cortical verbal zones provide evidence against such identification. It is known from study in neurological patients (as was also clearly revealed in Sidorova’s studies) that loss of the verbal functions does not invariably lead to the disturbance of consciousness. Such patients could maintain their placeand time-orientation and, in the case of recovery from the disease process as frequently takes place in lesions of vascular origin, can completely reproduce their feelings and process of events during the period of aphasia (expressive aphasia). Conclusion about the lack of the direct connection between consciousness and speech is confirmed also by other clinical observations, giving an opposite example: consciousness may be disturbed while the verbal function is preserved. Thus, in senile dementia (Alzheimer’s disease) the so-called ‘mirror symptom’ occurs. The patient, having seen his reflection in the mirror, thinks that it is another person and begins a prolonged ‘conversation’ with his imaginary partner. In this case, the patient could not be considered as being in full consciousness, though his speech is formally not disturbed. It should be noted that self-recognition in the mirror is one of the most intimate and accurate signs of consciousness. It is practically absent, even in higher mammals other than man, but is revealed early in childhood. This significant sign of conscious behavior, closely connected with the ‘ego’ sensation was noted also by Eccles (1980). Important data on the relationship of consciousness and speech were obtained during examination of the patients with ‘split brain’. Those experiments demonstrated that patients with dissection of the corpus callosum could accurately identify and select one object from a number of others. However, they could verbalize their actions only if this signal was directly sent to the
left hemisphere (right visual field). When the information came to the right hemisphere only, the patients could not say why and what they had done. Analyzing these data, the experimentators came to the conclusion that in the latter case one may speak not about a disturbance of the consciousness but only about impossibility for the patient to verbalize his actions due to the fact that the information did not reach the verbal centers of the left hemisphere (Nass and Gazzaniga, 1987). According to these data, the conclusion can be drawn that the most significant sign of consciousness (self-consciousness) is not directly connected with speech. Speech is one of the main instruments of consciousness, but consciousness itself is based on some other mechanism providing the possibility for the detaching of information from the subject. Such consideration goes together with Simonov’s definition of consciousness as mastering, operating of knowledge which implies the functional structure (‘the master’) able to operate with knowledge and to determine what message, to whom and with what purpose should be transmitted. It is easy to see, that such an approach to the problem of the objective and subjective criteria of consciousness practically suggests a high degree of similarity between the two. Probably, they are connected with the functioning of one mechanism (or different components of the same mechanism). Only in one case the work of this mechanism is estimated by external observer and in the other case it is experienced introspectively. Search for this mechanism is the basic task of the physiology of human higher nervous activity.
INFORMATIONAL SYNTHESIS MECHANISM OF SENSATION
The aim of the physiologist is to find the scheme of nervous processes organization, which could explain the appearance of the basic consciousness properties and which discriminate it from simpler forms of mental activity. It is assumed that the search for this scheme uses the methodological principle: more complicated func-
tions are accompanied by a higher degree of integration of the nervous processes. This principle was confirmed by our studies of the sensation mechanism, based on the parallel recording of the brain evoked potentials and psychophysical indices of the signal detection (Ivanitsky et al., 1984; 1989). It was established that the key link in the mechanism underlying sensation was a synthesis of two kinds of information on the stimulus: its physical parameters and its significance for the subject. The first ones are estimated by the sensory systems. The second component of the information synthesis - stimulus significance _ is determined on the basis of comparing physical characteristics with memory, providing information about the relation of this stimulus in past experience to the definite motivation and need. Informational synthesis is provided by a special mechanism of backward excitation: the recurrence of nervous impulses from the associative cortex and centers of emotions and motivations into the projection cortex, where they interact with the traces of the sensory afterdischarge. It is important that the moment of informational synthesis, determined by peak latency of the appropriate wave of the evoked potential coincides precisely with the time of the sensation as measured in psychophysical experiments. It is of interest that this mechanism correlates rather well with neurocybernetic construction by Edelman (1978) on excitation re-recurrence as the brain basis of mental functions. The analysis of perceptual mechanisms also provided some data on the brain mechanism associated with the ‘ego’ feeling. As far as the sensation appears as a result of amalgamation of the stimulus information and memory, it may be suggested that this memory, probably, presents in consciousness the ‘ego’ particle, in relation to which the obtained signal is perceived as something external. Actually, the perception of our own ‘ego’ is nothing else than recollection of the past events, our attitude towards them and our actions (Ivanitsky, 1990). In perceptual mechanisms, one may outline some main features of the brain organization of the mind: integrative character of the brain processes, presence of the key structure, realizing
the synthesis of information and participation learning and memory in the psychic origin.
METHOD OF INTRACORTICAL TION MAPPING
This approach to the brain mechanism of psyche was confirmed in our studies of the brain base of thinking. In this study, we used a new method of brain mapping - intracortical interaction mapping. The method develops one of the fundamental ideas of Russian neurophysiology: neural connections are promoted with the synchronization of the electrical potentials in the interacting brain areas (Livanov, 1986). The idea was also successfully developed in other countries (Gevins et al., 19901. In our study, we considered separate elements of the EEG power spectra as a manifestation of the activity of the neuronal groups characterized with definite rhythmic activity. The precise frequency coincidence of such peaks in different EEG traces suggested that these cortical areas were functionally connected because they included the neuronal groups precisely adjusted to each other due to the synchronization of their activity. Thus, the program of the intracortical interaction mapping included the following computer operations. After the Fast Fourier Transform (FFT) of the EEG traces, the computer determined three of the most prominent peaks in each of the main EEG ranges (delta, theta, alpha and beta). Then it searched for the peaks in the EEG spectra of different cortical areas that coincided in frequency with the accuracy of 0.24 c/s. Afterwards, the computer calculated the number of such coincidences for each of the cortical regions and for every EEG range. This value was normalized to eliminate the influence of the number of electrodes on the interaction characteristics. On the basis of these data, the brain maps were built by means of the interpolation method. In this map, the cortical areas differing by the number of such coincidences, i.e., the number of the cortical connections coming to these areas, were designated in different colors or type of shading.
The interaction maps in two kinds of mental operation were studied: the first one modelling the processes of the imaginative thinking and the other one the verbal-abstract thinking. 20 and 9 healthy subjects took part in this study, respectively. In the imaginative thinking task the subject had to recognize the emotion expressed by the actor on the face photo presented on the monitor
screen. The set of photos included 13 pictures expressing one of the four basic emotions: fear, joy, anger and sorrow or the mixed states, for example fear and anger. The subject should recognize the emotion. His verbal response was the trigger for the disappearance of the picture. Maximal time for the task solving was 6 s. After this time interval the picture was switched out even in absence of the verbal response. The anagram-solving task modelled the prothinking. The subject cess of the verbal-abstract
Rl-tn 0.15 .:.
@II ~, R.Jo _:;..:<-;..
8.R Fig. 1. The intracortical interaction maps in alpha range frequencies in the emotion recognition task (left) and the anagram-solving task (right). Above, when the decision was not found; below, at successful decision of the task. At emotion recognition, the foci are situated in the temporo-parietal cortex, at the word guessing in the frontal areas. The decision is accompanied by the foci appearance in the verbal zones of the left hemisphere. The scales indicate the normalized number of the connections.
had to guess and pronounce the word consisting of five letters which were presented on the monitor screen in a random order. The maximal time for the task solution was also 6 s, then the letters disappeared from the screen. The earlier disappearance of the anagram could be triggered by the subject’s verbal response. To catch the moment of the response, and task solution as well as precisely as possible, the EMG of the mouth angle muscles was recorded. The computer monitored the moment when the peak-to-peak EMG amplitude increased to the threshold of 35 PV in time interval from 150 ms after the stimulus presentation to the disappearance of stimulus from the monitor screen. Mean reaction time for both tasks was about 4 s. The control recordings were also made in the resting state and when the subject was presented with the word correctly written on the screen. The EEG from 10 electrodes placed according to the lo/20 scheme was recorded in left and right frontal, central, temporal, parietal and occipital regions. The amplification was made with a Medicor 32-channel EEG machine in the frequency band from 0.5 to 70 c/s. Then the signals via an analog-digital converter went to a IBM AT computer performing the above mentioned processing. The computer program also included the double-step artefact rejection. Firstly, the high amplitude traces were eliminated and then the correction of eyes movement artefact was made with the special program analogous to described by Gratton et al. (1983). The 2-s EEG traces preceding the verbal response or the picture disappearance were averaged. The averaging procedure was made for about 50 picture presentations in each subject and then across the subjects groups. The preliminary averaging procedure increased the reliability of the FFT transform because it accumulated in the averaged potential to be analyzed the frequency components involved in the decision process and time locked with it. The averaged event-related potentials (actually the kind of the readiness potential recorded in the situation where the active choice of reaction was made) were then mapped with the intracortical interaction mapping method.
The interaction map in the resting brain was usually characterized by a rather simple and symmetrical picture. During the mental task it sharply changed, and the definite, more or less complicated, picture was revealed. The main elements of this picture were the dynamic cortical structures designated as interaction foci. These foci were the cortical regions with the higher number of coincidences between the frequency peaks in this and other EEG traces, i.e., with the higher number of the cortical connections coming to this region. The topography of the interaction foci was quite typical for the definite mental operations. The most demonstrative results were obtained while studying the cortical connections in alpha diapason frequencies (Fig. 1). During emotion identification, the task solution was characterized with the interaction foci appearance in the occipital and temporo-parietal zones in left and right hemispheres, i.e., posterior cortical associative region. The higher connection level was in left hemisphere including the sensory verbal zone. These results corresponded rather well to above mentioned Sidorova’s data. When the decision was not found, the interaction foci were localized in the frontal zones of both hemispheres (Sidorova and Kostyunina, 1991). In the anagram task, the architecture of the cortical connections was quite different. At successful task-solving the predominant interaction foci were localized in the frontal zones of the both hemispheres (anterior associative cortical zone) and in the left temporal area. When the solution was not achieved the foci were in right temporo-parietal and left temporal zones (Ivanitsky and Ilyuchenok, 1992). Thus, the topography of the interaction foci in imaginative and verbal thinking was quite opposite. The picture of the cortical interaction in emotion recognition task was close to those in failure in anagram solving. These data allow us to suggest the existence of two hypothetical cognitive brain systems, which may be designated as the systems of sensual and reasonable cognition. The first one is connected with imagination and
sensual element of thinking and the second with logic operations using the abstract symbols. These systems are connected predominantly with the cortical associative zones - correspondingly posterior and anterior ones - and can interact and compete as well. So, as was mentioned above, the formation of interaction foci in the frontal cortex prevented the correct identification of emotions, which was of intuitive and not of a rational character. Both systems are connected with the verbal communication-interaction foci at successful tasks solving appeared in the verbal zones of the left hemisphere: sensory at perception of emotions and the motor at the words guessing. The difference between these cognitive systems is also as follows. The sensual system relates predominantly to the afferent processes, such as recognition, being the final stage of perceptive act. The aim of the reasonable cognition system includes the active operation (such as letter transposition) and relates mostly not to perceptual but behavioral decision. These ideas correspond also rather well with the some results obtained in PET experiments (Posner et al., 1988).
INTERACTION FOCI: DYNAMICAL CORTICAL STRUCTURES FOR INFORMATIONAL SYNTHESIS Thus, the cortical topography of the interaction foci could reflect the specificity of the mental operation rather well. But, what really is the purpose and the inner structure of this kind of the dynamic cortical structures? It was mentioned above that the interaction foci are the cortical areas characterized with the higher number of frequency peaks coinciding precisely with the frequency components in the EEG of the other cortical regions. According to the main idea, the interaction foci could be considered as the centers of the cortical connections, and their function could be as well the synthesis of the informational messages coming to the center from other cortical regions and, perhaps, some subcortical nuclei. Such information can include sensory inflow, data extracted from memory, signals from
Fig. 2. The scheme of the interaction foci. The focus consists of the neuronal groups Cl), connected with the hard-wired synaptic connections (2). The focus receives and synthesizes the information coming from the other cortical areas via the functional connections based on the principle of the coincidence of the rhythmic characteristics.
the center of emotions and motivations and, finally, genetic experience. But to perform this important role the interaction focus should have a specific structure (Fig. 2). Hypothetically, it consists of neuronal groups with definite firing characteristics which are tuned to the neuronal groups in the periphery. The connection between the central and the distant groups is the functional one and is based on the principle of the coincidence of the rhythmic characteristics. But the informational messages from other brain regions can not interact inside the focus on the base of the same principle. The neuronal groups inside the focus differ in their rhythmic properties, therefore, their interaction may be realized only on the basis of hard-wired connections, based on the structural changes in the synapses. Such connections could be formed in ontogenesis during the process of developing of the given mental function forming its central core. The dynamic system of connections appears
around this core, its peripheral areas being connected with the nucleus by labile functional connections. The general configuration of such connections, probably, can be changed in each definite case, stipulating the unique and inimitable character of the experienced mental act. It is also probable that in the process of consolidation of the function, the part of functional connections may be transferred into the structural ones, the same process as during conditioned reflex formation. At the early stages of the conditioned reflex elaboration, as has been shown (Livanov, 19861, synchronization takes place in the interacting cortical areas. After elaboration of conditioned reflex this phenomenon disappears, what may be explained by the reflex consolidation with appropriate synaptic changes. Thus, both the solid (hard-wired) and labile connections take part in ensuring the brain mechanism of thinking. These data can be compared with Bechtereva’s theory (1978) on solid and labile links in subcortical nuclei involved in the mental processes. The specific topography of the interaction foci in mental operations and rather high reliability of the data obtained in the representative groups of the subjects can indicate that the process of the informational synthesis performing by this dynamic structure really plays an important role in the brain mechanisms of the mind. It is of interest that this mechanism of synthesis studied in the thinking process has some resemblance with the described above sensation mechanism where the center of integration in the projection cortex was also singled out. The difference between perception and thinking is that in the first case integration takes part in the projection cortex on the neurons with the unique property of eliciting specific sensations, characteristic for each projection zone, such as feeling of light, sound, touch, etc., whereas in the second case the integration takes place outside the projection zones and, thus, it is not followed by such sensitive manifestations. At the same time, the thinking process is also accompanied by another significant sign of highly-organized neural processes: the sense of ‘ego’. In comparison to sensations where, as it
was mentioned above, the ‘ego’ feeling appeared at the junction of sensory signal and memory, at thinking the ‘ego’ image could appear at comparison of the operative and extracted from the longterm memory information. It is of interest that the close idea on the ‘ego’ image origin was also expressed by Edelman (1978). Thus, the comparison and the processing of the qualitatively different information in cortical centers of integration are resulted in ‘ego’ feeling and in the information message designated to the other people. This final act of the conscious process involves into functioning the verbal centers of the left hemisphere.
CONCLUSIONS cl), Consciousness is characterized by the objective and subjective criteria. Correspondingly, an objective criterion is an ability for operating with the knowledge, which could be directly transferred to the other people in an abstract form. A subjective criterion is characterized by the feeling of ‘ego’ as a ‘host’ of these actions. (2), Objective and subjective signs of the consciousness are provided by the mechanism based on the synthesis in the brain structures of different kinds of information: sensory information, the messages extracted from the memory and impulsation from the centers of emotion and motivation, As a result of this synthesis the sense of ‘ego’ appears and the message designated for others is determined. (31, A significant role in the informational synthesis is played by dynamic cortical structures foci of interaction. In perception, they are localized predominantly in the projectional cortical zones, at thinking in the associative areas. (41, Realization is precisely connected with communication and formation of interaction foci in the verbal zones of the left hemisphere.
REFERENCES Bechtereva, N.P. (1978) The Neurophysiological Aspects of Human Mental Actiuity, 2nd Edn. Oxford University Press, New York, 181 pp.
187 Eccles, J. (1980) The Human Psyche, Springer International, Heidelberg. Edelman, G.M. and Mountcastle, V. (1978) Mindful/ Brain, MIT Press, Cambridge. Gevins, AS., Bressler, S.A., Cutillo, B.A., Illes, J., Miller, J.C., Stern, J. and Jex, H.R. (1990) Effects of prolonged mental work on functional brain topography. Electroenceph. Clin. Neurophysiol., 76: 339-350. Gratton, G., Coles, M.G.H and Donchin, E. (1983) A new method for off-line removal of ocular artifact. Electroenceph. Clin. Neurophysiol., 55: 468-484. Grindel, O.M. (1985) Intercentral relations in the cerebral cortex by indices of the EEG coherence at rehabilitation of the consciousness and speech after prolonged coma. Zh. Vysh. Nervn. Deyat., 35: 60-67. (Russian). Ivanitsky, A.M. (1990) The ‘ego’ and the brain. The Man, 4: 16-23. (Russian). Ivanitsky, A.M. and Ilytchenok, I.R. (1992) The mapping of brain potentials in verbal task decision. Zh. Vyssh. Nervn. Deyat., 42: 625-633. (Russian). Ivanitsky, A.M., Strelets, V.B., Korsakov, I.A. (1984) Informational Brain Processes and Mental Activity, Nauka, Moscow, 200 pp. (Russian). Ivanitsky, A.M., Strelets, V.B. and Korsakov, I.A. (1989) Perception as a result of the synthesis of the sensory and emotional. In Sudakov, K.V. (Ed.) Systems research in physiology, Vol. 2, Emotions and behavior: a systems approach, Gordon and Breach, New York-London, pp. 4765. Leontyev, A.N. (1981) Problems of Mental Development,
Moscow University Publishing House, Moscow, 584 pp. (Russian). Livanov, M.N. (1986) Rhythms of the electroencephalogram and their functional significance. Neurosci. Behav. Physiol., 16: 173-186. Nass, R.D. and Gazzaniga, M.S. (1987) Cerebral lateralization and specialization in human nervous system. In Handbook of Physiology, Section 1; Nervous system, VT5, Higher Functions of the Brain. Part 2, Am. Physiol. Sot., Bethesda, pp. 701-761. Pavlov, I.P. (1951) Experimental psychology and pathopsychology in animals, Complete works, USSR Academy of Science Publishing House, V.2, Book 1, pp. 23-39. (Russian). Posner, M.I., Petersen, S.E., Fox, P.T. and Raichle, M.E. (1988) Localization of cognitive operations in the human brain. Science, 240: 1627-1631. Sidorova, O.A. and Kostyunina, M.B. (1991) The role of cortical zones in the process of the perception and reproduction of the human emotional states. Zh. Vyssh. Nervn. Deyat., 41: 1094-1102. (Russian). Sidorova, O.A. and Tsyganok, A.A. (1989) Studies of the ability to perception and realization of the emotional states in patients with local brain lesions. In Problems of Neurocybemetics, Proc. IX All-Union Conf., Rostov-onDon, p. 254. Simonov, P.V. (1990) The Motivated Brain, Gordon and Breach, New York. Turing, A.M. (1950) Computing Machinery and Intelligence. Mind, 5: 433-460.