This paper attempts to position the development of consciousness as intentionality within a semiotic perspective. The theoretical grounding is in Peircean semiotics but work in neurobiology by Gerald Edelman, and in systemic functional linguistics by Michael Halliday is also drawn upon. The intent of this paper is to open for some constructive discussions between researchers working on more technologically oriented aspects of AI research, and those in the humanities and natural sciences.
The interactional, intersubjective nature of the evolutionary development of consciousness and intentionality in human beings is one of the most difficult and often subverted aspects of human experience for experts working with modelling systems addressing such problems as natural language processing, development of inferencing and knowledge systems, interface design and indeed, human-computer interaction in general. A semiotic perspective has been chosen for this paper because I believe that semiotics provides an ontological framework that allows the evolutionary development of consciousness as intentionality to be examined within a broad cross-disciplinary framework, and because I believe that working within a semiotic framework is one of the most profitable ways for humanist scholars to approach this particular field of study.
Recently, an increasing number of attempts have been made to situate consciousness within a framework of evolutionary and ontogenetic change with reference to empirical biological, neurophysiological and biochemical data.
It is common in the human sciences to make a distinction between consciousness and enlightenment or meaning. With consciousness we understand the subjective experience human beings have of being situated in the world or in reality. Understanding or meaning is taken to refer to an experientially and rationally founded understanding, developed over time, of the ethical, moral and ontological consequences of this situatedness in the real world of external objects and processes.
According to Immanuel Kant [1] human thought and enlightenment (Erkännung) are dependent on a limited number of a priori perceptual forms and ideational categories, such as causality, quality, time and space. Philosophers must retire to a meditative space and think their way systematically and reductively back to these prerequisites by means of the intellect (Verstand). In Charles Sanders Peirce's pragmaticism, the central idea is one of a "community of interpreters" [2, 3]. Enlightenment is synonymous with an evolving consensus concerning the true meaning of the "real". Truth or enlightenment is contingent on an infinite process of scientific enquiry within a community of interpreters vis-a-vis phenomena of the real world, rather than being a transcendental state of mind arrived at by one particular individual (or group of individuals) at some particular period of time.
Peirce outlined four presuppositions of Cartesianism which he meant needed further investigation. They are that:
1. we have no power of introspection, but at the same time all knowledge of our internal world is derived by hypothetical reasoning from our knowledge of external facts.
2. we have no power of intuition, but yet every cognition is determined logically by previous cognitions
3. we have no power of thinking without signs.
4. we have no conception of the absolutely incognizable
The self as we know it is constituted through signs and belongs to the realm of nature and takes part in the sign processes of the natural environment and sign processes within the community of interpreters. What is real reveals itself to us as known or knowable external signs that inhabit or populate the domain of thought. The conscious self intersects with these semiotic chains and becomes open to the rational structures of the universe. Insofar as the self attains self-awareness or self-knowledge it must do so through acts of comparison that work their way through the series of external signs and their fields of meaning [4].
In the study of human behaviour we do not in fact generally attempt to address the question of whether having consciousness or mind or not is possible for human beings. The phenomenon of mind is simply taken as a basic presupposition or "given". RenéDescartes [5] for instance took mind for granted when he postulated "cogito ergo sum". Marvin Minsky [6] has even gone so far as to say the "mind is what brains do" - which although being true in one kind of sense, is a somewhat simplistic description.
Peirce defined "belief" as some state of mind which is so stable and "habitual" that we are prepared to act on the basis of this. He also maintained that the unconscious is the place where most of these beliefs are located.
"Belief is not a momentary mode of consciousness; it is a habit of mind essentially enduring for some time, and mostly (at least) unconscious; and like other habits, it is (until it meets with some surprise that begins its dissolution) perfectly self-satisfied. Doubt is of an altogether contrary genus. It is not a habit, but the privation of habit. Now, a privation of habit, in order to be anything at all, must be a condition of erratic activity that in some way must be superseded by a habit. [CP 5.417][1]
Knowledge is constituted as habits of mind capable of effecting intentionally directed actions vis-á-vis the real world.
"A practical belief may, therefore, be described as a habit of deliberate behaviour. The word "deliberate" is hardly completely defined by saying that it implies attention to memories of past experience and to one's present purpose, together with self control. The acquisition of habits of the nervous system and of the mind is governed by the principle that any special character of a reaction to a given kind of stimulus is (unless fatigue intervenes) more likely to belong to a subsequent reaction to a second stimulus of that kind, than it would be if it had not happened to belong to the former reaction." [CP 5.538]
Peirce makes a distinction between practical and theoretical belief, on the basis of the way in which habits are associated with these.
"But habits are sometimes acquired without any previous reactions that are externally manifest. A mere imagination of reacting in a particular way seems to be capable after numerous repetitions of causing the imagined kind of reaction really to take place upon subsequent occurrences of the stimulus. In the formation of habits of deliberate action, we may imagine the occurrence of the stimulus, and think out what the results of different actions will be. One of these will appear particularly satisfactory; and then an action of the soul will take place which is well described by saying that that mode of reaction "receives a deliberate stamp of approval." The result will be that when a similar occasion actually arises for the first time it will be found that the habit of really reacting that way is already established. [CP 5.539]
This implied relationship between belief, thought and action in interaction with the real world formed the basis for Pierce's pragmatic maxim, which states that:
"In order to explain the meaning of an intellectual conception one should consider what practical consequences might conceivably result by necessity from the truth of that conception; and the sum of those consequences will constitute the entire meaning of the conception." [CP 5.9]
Even though the number of possible signs and interpretations or understandings of these signs by interpreters is in principle (and for Peirce, de facto) unlimited or infinite and not in any way bounded by temporal or spatial considerations, it is still possible for some of these sign relations to become so habitual (that is: so "stabilised" within our consciousness or unconscious that we come to accept them as "truths", standing for conventionalised causal relations having independent "meanings"), that they result in highly complex patterns of intentional behaviour, which have a lasting effect on the real world.
Gerald Edelman [7, 8, 9, 10] operates with two central theses about consciousness:
1. Consciousness (mind) must be understood on the basis of insights from empirical biological research, particularly research into the evolution and development of the human brain
2. Consciousness is intentionality
The laws of physics as they are currently formulated cannot illuminate the relationship between consciousness and the evolution and ontogenesis of the human brain, says Edelman. Physical laws attempt to describe nature as symmetrical - as following certain rules with discrete causal relationships between its most elementary units. There are huge differences in the ways individual brains are structured at microlevel, and a complete lack of symmetry in the developmental trajectories of neural pathways and complexes of synaptic networks in individual brains. There is also the problem of symbolic representation and intentionality (the capability of referring to things or beings outside of oneself) to take account of. Physical laws must be formulated symbolically (by means of a semiotic system such as language) on the basis of interpretations made by human beings.
Heisenberg's Uncertainty Principle has taught us that precise measurement of the position of subatomic particles at any given point in time is not possible. The degree of precision of the instruments set up by the observer causes the wave-function to "collapse" at the very moment the measuring instrument and the particle in question interact to give a definite measurement. The wave function ¥ in the Schrödinger wave equation is a linear combination of functions describing all possible outcomes of the measurement. When a specific measurement is made, the wave-function collapses or "projects" onto one of the possible outcomes. When a dynamic process is projected to only one state ("frozen" in time and space to a discrete value), it ceases to be dynamic, and therefore changes character. Wigner [11] has claimed that the wave particle collapses as a result of the intervention of the consciousness of the observer (a view I would tend to support), since the observable phenomenon (the measurement) is only actualised at the time when the observer becomes conscious of it.
Niels Bohr [12] solved this problem by saying that there is in any case no ultimate or deep reality. Measurements belong to a context of observation, so the measurement one gets is at any time is merely the one would get given the context of measurement, particle state, apparatus and observer. Bohr's Copenhagen interpretation is the position taken by most physicists today, since it gives a formula describing what one can observe with an apparatus that itself is made up of the same quantum particles one is measuring.
Roger Penrose [13] has claimed that the problem of consciousness will be solved whenever one constructs a satisfactory enough quantum theory. This however is not sufficient to explain intentionality, since it ignores phenomenological and biological aspects of the development of consciousness. Physics is able to provide the necessary bases for biology, but does not concern itself with biological structures, processes and principles in the human organism, especially those of the brain. What is lacking is a scientific, biologically and evolutionary based description of the structures and functions of the organism tied to the development of consciousness. We need a theory that describes ontogenetic and evolutionary processes at the morphological level.
Logical and functional models in cognitive science which compare the brain and human thought processes with computers, are inadequate, since formal operations (manipulations of abstract logical symbol systems) may be performed without reference to the meaning of these symbols. The human genome (the complete set of genes belonging to a human being) is not sufficient to specify explicitly (algorithmically, syntactically or procedurally) the synaptic structure of the developing human brain. There is a staggering degree of structural variation at all levels of organisation of the brain. These variations are unique and tied to the personal history and development of each individual organism, and to its interactions with the natural environment. The behaviour of living organisms is enormously diversified at the biological level. This is independent of whether or not they can have subjective experiences and report them to others, as human beings do by means of language and other semiotic systems.
It seems highly unlikely that the physical and social world would be able to function as a tape for a Turing machine (a machine that can execute any kind of algorithm or mathematically formulated procedure without being aware of doing so). As Edelman puts it: "The world, though constrained by physics, is an unlabelled place."
Philosophers of language such as Hilary Putnam [14] and John Searle [15] have criticised theories that compare the human mind with computers. They point out that psychological states of mind including propositional attitudes (like believing that p, desiring that p and so on) cannot be described by computational models (Putnam) and that purely syntactical models are insufficient for dealing with semantic content, which is a basic characteristic of human minds (Searle). Semantic contents involve meanings, and a syntax does not itself deal with meanings. We cannot individuate concepts and beliefs without reference to the environment. No organism can have intentional states if it lacks subjective experiences. Consciousness as intentionality (the ability or tendency to refer to things of the world, which cannot themselves refer) is irrevocably tied to subjective and intersubjective experiences of individual organisms as they position themselves in the environment. Computers lack such experiences which are by nature unpredictable and cannot be fully programmed or predicted by means of computational models.
Meaning is interactional. Interactions between human organisms and the signs of the physical and biological environment play a central role in determining what the meanings of words refer to for speakers or communities of speakers. A priori inclusive descriptions of how meaning is constituted through experience are not possible by means of procedural models. The biological environment, like the human organism, is open ended. It is comprised of many dynamic open systems with a non-determinate tendency to self-regulated change towards greater levels of complexity. As each individual brain is structured through the interactions of the organism with its environment, the body plays an important role in determining meanings. To learn what things mean, we must develop language while moving among and communicating with other living organisms about our experiences of the world.
Habituated patterns of nervous system mappings of the environment are tied to individual life-histories. Adequate response patterns for each organism are selected over time through its interactions with the environment. This is true at both individual (ontogenetic) and evolutionary level. There is variation not only between different nervous systems, but also within individual systems across time. Mapping systems represented at the anatomical level of organisation in living organisms vary enormously. Not only are there fluctuations in the borders of such maps over time, but the variability of mapping in adult animals depends on the signal input at any given time. This is a strong argument against extreme functionalist models of internal representations which claim that representations have meanings independent of the ways in which they are instantiated.
This problem touches so-called neural network models which are not self-regulating at the physical and structural level. Someone - for instance a programmer - must construct by programming on a computer new configurations of these networks; they cannot regulate this process experientially by themselves. Learning in neural networks is then instructional, not selective, to use Edelman's terminology.
Cognitive and linguistic theories of meaning that attempt to deal more or less exclusively with issues of semantics and mental representation are inadequate for coping with the problem of consciousness as intentionality. For us to refer to something outside ourselves formal representations must become intentional. For this to happen, the system that performs such a transformation or translation must have both "awareness" and a "self"; a biologically based, individuated, subjective experience of its own situatedness in the world. No computer (at least not those we have today) can have this kind of experience.
Michael Halliday is concerned with the role of language in the development of human consciousness, and how the language of science carries on an ongoing dialogue with nature. Language is an essential condition of higher-order consciousness. When we want to exchange meanings with physical or biological nature we have to process information coded in very different ways, and need to go through two or three stages of translation before we can apprehend it [16]. Language and grammar are as much a product of evolution as we are ourselves; the lexicogrammars of living languages are evolving, and are evolved systems, not designed systems. Language is not separate from nature, but an essential part of being human. Natural languages enable us to interpret both the social order and the natural order which are construed by generalising and abstracting from the micro-environments in which language evolves. Language as a social semiotic contributes to the construction of self-consciousness by construing and enacting interpersonal relationships.
Without the interpersonal metafunction enacted in language there would be no culture. The lexicogrammar of language not only construes; it also enacts the life-world. Through grammar, human beings enact their "interpersonal" (intra-species, inter-organism) relationships. Lexicogrammars constitute societies, and through societies, the individual self. Instantially, they enact dialogic roles and the ongoing personification of "I" and "you"." [17]. From early infancy language impinges in a number of ways on the construction of the social order, mediating the dialogue built upon the intimacy that develops between a child and its caregivers. The efficacy of this "language for loving and caring rather than for knowing and thinking" is not judged referentially. Language creates society; and it does so without referring to the processes and the structures which it is creating. Both Halliday and Edelman refute attempts to reduce language to a communication code, or the brain to a computer. The brain is more comparable to a rapidly growing jungle than to a computer.
Neither Edelman nor Halliday are averse to the idea of using computer simulations to investigate naturally occurring systems and emergent phenomena, such as language and consciousness. They are merely respectful of the enormity of the task of modelling artificial environments and systems that represent accurately the multivariate levels of complexity of organisation, and the continuous, dynamic nature or diachronicity [18] of self-regulating systems. Computers are not appropriate models of brains, but they are powerful heuristic tools with which to try and understand the matter of the mind. Computational artefacts with higher order consciousness would need language and the equivalent of behaviour as a member of an interpretational community. So far, human beings remain the only known systems with linguistically based higher-order consciousness, and competing artefacts are still a long way off.
Elsewhere [19] I have argued that one way of learning more about how the mediation of human communication by technology is situated within the evolution of language and the development of interpretative scientific communities is through phenomenologically oriented studies of how textual norm systems evolve in the virtual environment of distributed text-based multi-user dialogues. Where these technologically mediated social fields seem to differ from real life processes of the same kind is with regard to the range and depth of significational devices available to participants as they instantiate and construe the norms of text-cultures emerging in these virtual environment.
The relative lack of richness of the language code (written language simulating oral forms of communication) and the unavailability of other semiotic systems such as non-verbal communication, intonation, olfaction and tactile communication is attributable first and foremost to the "closed" or "limiting" nature of the mediating medium offered by the technology of the system, not to the communicative potential of language or the sociocultural habitus of the participants. One of the most interesting aspects of these mediated social worlds is, aside from their obvious potential for scientific collaboration and discussion in the longer term, that the ongoing process of meaning-making is not only being produced, but also capable of being documented (logged in real time) by the participants while they are instantiating it. Researchers taking part in the development of distributed virtual environments, while at the same time being interested in studying their ontogenesis and evolution, enjoy a privileged position, having a unique opportunity to document and study the dymanics of developmental processes within an emergent social semiotic.
Studies of this kind may serve to throw light on what, in my opinion is one of the most difficult problems of all to explain scientifically, and presumably also to model and create artificially, even taking account of the leaps and bounds that we are witnessing within the fields of evolutionary neurobiology and systemic functional linguistics, namely the very human experience of intimacy and closeness which we have come to label as intersubjectivity; that sudden shock of recognition of the "other as other" that human beings still seem able to experience from time to time when two gazes meet. I believe we still have very much more to learn with regard to the ways in which computer-mediated and simulated forms of communication impinge upon the establishment, growth and maintenance of interpersonal and social relationships that, when all comes to all, are the very basis for the establishment and growth of meaning in any kind of interpretative community, and thus also a prerequisite for the further evolution, not only of science and technology, but also of the human race.
[1] Kant, Immanuel 1781. Kritik der reinen Vernunft.
[2] Corrington, Robert S., 1993. An Introduction to C.S. Peirce, Philosopher, Semiotician and Ecstatic Naturalist. Lanham, Maryland: Rowman and Littlefield.
[3] Peirce, Charles S. Collected Papers. Vol. V-VI. 1960. Edited by Charles Hartshorne and Paul Weiss. Cambridge: Harvard University Press.
[4] Corrington, Robert S. 1994. Ecstatic Naturalism. Bloomington and Indianapolis: Indiana University Press.
[5] Descartes, René1637. Discours de la Méthode.
[6] Minsky, Marvin 1985. The Society of Mind. MIT Press.
[7] Edelman, Gerald M., 1987. Neural Darwinism: the theory of neuronal group selection. New York: Basic Books.
[8] Edelman, Gerald M., 1988. Topobiology: An Introduction to Molecular Embryology. New York: Basic Books
[9] Edelman, Gerald M. 1989. The Remembered Present: A Biological Theory of Consciousness. New York: Basic Books.
[10] Edelman, Gerald M., 1992. Bright Air, Brilliant Fire. On the Matter of the Mind. New York: Basic Books.
[11] Wigner, E.P. 1979. 'Remarks on the Mind-Body Question.' In I.J. Good (ed.): The Scientist Speculates. London: Heinemann.
[12) Bohr, Niels, 1964. Atomic theory and the description of nature. Cambridge : Cambridge University Press, 1964
[13] Penrose, Roger 1989. The emperor's new mind : concerning computers, minds, and the laws of physics. Oxford: Oxford University Press
[14] Putnam, Hilary 1975. Mind, Language and Reality (Philosophical Papers II). Cambridge University Press.
[15] Searle, John 1979. 'What is an intentional state?' Mind, 88 (1979) pp 72-84.
[16] Halliday, Michael A.K., 1987. "Language and the Order of Nature". In: N. Fabb & al. (Eds.), The Linguistics of Writing. Manchester: Manchester University Press, 1987, chapter 9.
[17] Halliday, Michael A.K. 1992. On language in relation to the evolution of human consciousness. Paper prepared for the Nobel Symposium 92 "The relation between language and mind", Stockholm, 8-12 August 1994. Unpublished draft copy.
[18] Coppock, Patrick J. in press (a). Ascribing continuity to the diachronicity of textual norms in virtual environments. Paper presented at the Nordic Association of Semiotic Studies Research Congress, Trondheim, October 21-23 1994. To appear in a special number of Semio Nordica in the course of 1995.
[19] Coppock, Patrick J. 1997.
"The
semiotics of a phenomenological research paradigm for investigating
the evolution and ontogenesis of cultural norm systems in distributed
virtual environments". Semiotica 115-3/4 (1997),
235-262.