First, a principal distinction between two different kinds of semiotic investigations is introduced, both required in the study of living signs and signs of life. Then, the attempt within the new field of Artificial Life to model and synthesise computationally based living systems is discussed with special attention paid to the possible emergence of genuine life-like behaviour in such models of for instance self-reproduction. Remarks will be made on a seemingly odd aspect of the biological concept of life; that it is not as coherent as normally conceived of. In general, biosemiotic emergence of new sign functions is distinguished from other kinds of emergence that pertain to the domain of the observer and the modeling relation.
Attempts to understand life as a semiotic phenomenon are motivated by an interest in bridging the gab between mind and matter, not by a purely philosophical intervention, but through the insights gained from scientific and transdisciplinary areas of investigation. A more unified and basic understanding of signs in living nature must involve a natural history of signs, i.e. an evolutionary account of sign functions not exclusively bound to the life of the signs in human society (as F. de Saussure originally conceived as the subject of his semiology), but sign and communication processes at several levels of organization beneath conscious human semiosis. It encompasses but cannot be reduced to an account of the evolutionary origins of Homo sapiens' cognitive functions and the evolution of science, religion and other specific cultural institutions. As an evolutionary epistemology it must investigate the evolution of the ontological characteristics of the universe that allowed sign functions to emerge. The attempt to construct biosemiotic theories of the crucial steps in a natural history of signs is faced with origin problems within the specific disciplines, such as the origin of language, the origin of life, and the origin of the sign in its most primitive form. As a study of such fundamental questions, biosemiotics must involve both conceptual, empirical, experimental and philosophical approaches. Origin problems are often framed within the concept of emergence, known for a long time in philosophy (e.g., Pepper 1926, Ablowitz 1939, Collingwood 1945, Klee 1984), debated among biologists (Morgan 1923, Woodger 1929, Mayr 1982, to name a few), and recently winning respect within physics and the study of complex systems (Forrest 1990).
The origin problems must be tackled both from a scientific point of view, by constructing specific theories (e.g., biogenesis as the emergence of cells by self-organizing metabolic processes in specific solutions of simple compounds) and from a more conceptual point of view, by analyzing the relations between, on the one hand, the descriptions of the simple components and laws that govern their behavior, and on the other hand, our concepts of integrated wholes that emerge in self-organizing processes. This is true for the origin of life as well as for the origin of cognition, language or `the spontaneous self-assembly' of rules of conduct in human behaviour within the private or public spheres. We have to use very different games of scientific language to describe the origin of cells from a prebiotic soup of macromolecules and the origin of language among a social group of hominid ancestors living together in a highly coordinated way. It is not likely that all social phenomena can be described (in nonreductionistic ways) by purely scientific means, but this can be left as an open question here; the point is that one should reflect upon the relationship between descriptions and the phenomena described belonging to different levels of organization. To remember this double character of the origin problem, I will suggest a distinction between two fields of semiotics, both needed in our inquiry into the questions of life, cognition and semiosis:
Biosemiotics proper deals with sign processes in nature in all dimensions, including (1) the emergence of semiosis in nature, which may coincide with or anticipate the emergence of living cells; (2) the natural history of signs; (3) the "horizontal" aspects of semiosis in the ontogeny of organisms, in plant and animal communication, and the inner sign functions in the immune and nervous systems; and (4) the semiotics of cognition and language (in itself an enormous field, so its subsumption under the mark of `biosemiotics' may appear a little misleading). Biosemiotics and its sub-fields are a topic of growing concern among many biologists and semioticians (e.g., Anderson et al. 1984, Sebeok 1977, Hoffmeyer in press, von Uexküll 1986). Biosemiotics can be seen as a contribution to a general theory of evolution, involving a synthesis of different disciplines. It is a branch of general semiotics but the existence of signs in its subject matter is not necessarily presupposed as far as the origin of semiosis in the universe is one of the riddles to be solved.
The other field is the system-observer semiotics, by which I mean the critical inquiry into the nature of the modeling relation to the various systems we can observe, describe, conceptualize and construct theories about. It is a semiotic of scientific experiment, observation, interpretation, operation upon and measurement of various systems. These issues have been studied by classical traditions of philosophy of science, but their special purport in relation to living systems has been developed recently by theoretical biologists and systems scientists (e.g., Rosen 1991, Kampis 1991, Cariani 1991, Pattee 1989; see also von Uexküll 1984, 1989). A simple example of the need for semiotic reflections when interpreting observations of other living systems is the well-known problem of ascription of anthropomorphic qualities to other creatures. When constructing theories about semiotic aspects of living nature, there is a powerful compulsion to project specific human properties onto the object of study -- for instance, high level properties and functions such as `rational inference', conscious intentionality, or language capability are often projectedwhen describing animal behaviour (cf. Sebeok 1987) -- even when a closer inspection might reveal that much simpler mechanisms govern the behaviour of the system.
The two kinds of semiotics should be thought of as complementary. They involve two different points of view of living systems: one holds that the rational reconstruction of different descriptions contributes to a coherent evolutionary perspective; the other offers a critical outlook at theories as symbolic systems of conceptual signs that should never be confused with the complexity and semiosis of life itself. When for instance a general view of the evolution of biological coding systems is outlined (see Hoffmeyer and Emmeche 1991), one must remember that informational descriptions of living systems are tied to the use of problematic metaphors, as far as nobody seems to agree upon a single clear-cut and operational definition of the semantic or `biologically meaningful' aspects of information in real biosystems (Emmeche and Hoffmeyer 1991). Our focus here will be on the system-observer semiotics of `emergent' lifelike phenomena of the so-called Artificial Life (Alife) systems; a subsequent paper will deal with the projection of computational properties into the thing being modeled.
Some roads to Artificial Life
Artificial Life or Alife is a new, multi-focused field of research. Like theoretical biology, it is concerned with formal, computational models of life, but it has expanded its scope by claiming the possibility of synthesizing lifelike behaviours within computers and other artificial media. Traditional biology has been primarily analytic and reductionist in outlook, taking an organic whole apart and analyzing its components. As Chris Langton introduces the field, Artificial Life complements traditional biology: `By expanding the empirical foundation upon which biology is based beyond the carbon-chain life that has evolved on Earth, Artificial Life can contribute to theoretical biology by locating life-as-we-know-it within the larger picture of life-as-it-could-be.' (Langton 1989, p.1). So hitherto, we have only known a special instance of life: biological life on Earth, or B-life if you wish; now Alife can set the real context for the study of life as a general phenomenon. If this idea seem a little surprising, provokative or even offensive -- I remember my own bewilderment when I first heard artificial life proposed as a serious hypothesis -- it should be remembered that this reaction may be partly intended, and that one should not let ones metaphysical or scientific prejudices block the way of inquiry into new areas of reality, even if these may seem rather virtual.
One of the virtues of Alife is that it brings together people from a lot of disciplines and evokes new questions and approaches to the study of complex phenomena. In its present form, the Alife research programme is not coherent in scope or purpose: some researchers merely want to develop the field to provide a tool-box of computational and engineering techniques in order to simulate biological systems or develop adaptive robots (or `animats', cf. Meyer and Wilson 1991), other proponents, such as Langton, advocate a `stronger version' of the program. I will give a short list of various versions of Alife as a research program, to make explicit the dissimilarities in the ways people conceive of the subject (for a more methodologically oriented classification, see Taylor 1991):
1. Computational Alife
(1a) Weak version: The purpose of Alife is to make computationally based models of natural biological systems.
(1b) Strong version: The ultimate goal is the realization of life in another medium -- i.e., the computational medium of a computer (`computer life'). This can be achieved, because life is a medium-independent phenomenon, a question of form or processes, not a specific material that constitutes `aliveness'.
2. Robotic Alife
(2a) Weak version: Robots and animats built for technical purposes may behave in a `lifelike' manner. This, however, is a by-product of our interpretation, the behaviour of these systems represents a category quite distinct from the behaviour of the carbon-based biological cells and organisms.
(2b) Strong version: The ultimate goal is the creation of autonomous, self-reproducing animats, capable of living a life of their own, adapting to a changing environment and eventually evolving into new species if located in an appropriate environment. A more proximate goal is the creation of robots or animats with the full behavioral capacities of living organisms. If any differences are at all recognized between systems based on biochemical mechanisms and `animat' devices based on human design, eventually in very small scale (`nano-technology'), these differences are not judged to effect the principal possibility of achieving the goal.
(2c) Alife as a royal road to AI: Life came before intelligence; so should real Alife come before the achievement of intelligent artificial systems; whether you think of them as being of `the same kind' as us or not. Alife is seen as an important source of spin-off techniques to speed up AI.
3. Chemical Alife. This view comprises attempts to make real material systems with lifelike characteristics, eventually as in vitro models of prebiotic processes; primitive metabolic systems; the so-called Eigen hypercycle systems, etc. The goal is to turn in vitro experiments into life in vivo de novo.
Version (1a) is a fairly trivial version of the program, and epistemologically a modest and sober position: Everybody seems to agree on the possibility of modeling living phenomena, these models can be computational or not. However, even here, a closer look reveals disagreement with respect to the relevance and adequacy of the formal, computational approach, even conceived of as just models (not the real thing). If we choose computational (or informational) models, is that just for representational convenience, or is it because the thing we model is also in some sense computational, or informational? This is a highly non-trivial question about the semiotics of the modeling relation that should be clarified. To give an example, there is a whole field of research into the simplest physical and biological elements to compute with, and it is often proposed that a cell is in fact information processing or doing `molecular computations' (for a recent entrance to the literature, see Marijuán 1991), without exactly substantiating in what sense one could apply such concepts, and without considering the implications for such a view of the non-algorithmic aspects of biochemical systems.
Within this position, some unease is felt with the very name Artificial Life; `nonlinear dynamical biosystems simulations' or a similar term might be preferred. Nevertheless, the term has come to stay; Artificial Life is fit for the political context of research, similar to the advertisement of other `outstanding ventures of humankind' such as putting `a man on the moon' or mapping the human genome, and of course, building `intelligent' machines. As these historical examples of other scientific programs indicate, the name of the general idea is more than simply a question of rhetoric and should be considered seriously.
Version (1b), computer life, is the focal interest of this paper. It embodies the same kind of claim made in Artificial Intelligence: namely, that the natural phenomenon under study (life, intelligence -- what about physics and chemistry?) reduces to its formal (or functional) structure, so that if this structure can be clarified (i.e., described by an `equivalent' algorithm), the execution of this algorithm in another programmable mechanistic system (a Turing machine such as a table computer) will realize the same functions, and thus realize the very phenomenon itself. The phenomena of intelligent or living behaviour are seen as abstract forms of movement; the logic of these forms can be abstracted from the material basis of the systems in which they are manifested. An organism's `aliveness', its capacity to live and reproduce, is a property of the form, not the matter (Langton 1989). Therefore it is possible to synthesize life (genuine living behaviour) on the basis of computational principles. This is partly based on the idea of abstracting the logic of self-reproduction developed in von Neumann's theory of self-reproducing automata from the late 1940s. It should be remarked that von Neumann himself considered the problem of the material basis which was by this very procedure abstracted away in the automata models, to be more critical than do his modern successors within Alife (von Neumann 1966, cf. his remark on p. 67; also discussed by Langton 1989, Pattee 1989, Cariani 1989, Laing 1989, Kampis 1991).
Strong version Alife seems to contradict an immediate intuition of molecular biologists, that `form' and `matter' do not represent separate realms (at least on the intracellular level), and that information is intimately bound to the properties of the material substrate (Emmeche 1992). These intuitions, however, are claimed by the `alifers' or strong Alife proponents to be `carbon-chauvinistic' prejudices, that cannot be justified from a more general view of life. I am sympathetic to the biological intuition, but I think that the discrepancy in views can be clarified and partly dissolved if one distinguishes between two theses of life as an `medium-independent' phenomenon:
The first thesis, according to which life is held to be possible to realize in media other than the specific kind of protein and nucleic acid based biochemistry (itself based on `CHNOPS': carbon, hydrogen, nitrogen, oxygen, phosphor and sulphate) that we know is universal to life on Earth. Life may have evolved on other planets with a remarkably different composition of chemicals, and thus another environment that allowed different forms of life to evolve. Other kinds of life, however, must be every bit as `material' and physically constrained by these specific other boundary conditions as life on Earth is constrained by specific material properties of their chemical components and physical environment -- e.g., the properties of water; the magnitude of gravity (compare Vogel 1988). Even if biochemists and molecular biologists can hardly imagine, say life without water, the general idea -- which should be called `the thesis of medium-dependent life in multiple possible media' -- when posed in an appropriate form is open to empirical investigation, eventually in the form of Alife version (3), eventually (though less probable) through the search for extra-terrestrial life and intelligence.
The second thesis of life as a truly medium-independent phenomenon is radically form-oriented and Platonic in character. It states that life is a phenomenon of being organized in a specific way, such that the parts of a system and their states are related to each other in a way that permits the relational structure to be maintained as a coherent, stable pattern in space and time, this pattern being quite independent of the specific medium in which it is realized. Different realizations in different media are equivalent with respect to the dynamic quality of being alive. Both formal computational and natural material systems are held to be possible supports of this quality. The thesis is either connected with notions of self-organization and emergence (A), or with a set of qualitative criteria of what constitutes lifelike behavior (B).
(A) Thus Langton argues for `genuine life in artificial systems' (1989, p. 32) by proposing an ontological distinction between 1) low level individual behavors of the simulated components (e.g., simulated birds; or simulated cells of an organism) that act according to local rules but only constitute informational processes within a computer, and 2) the emergent behaviors of a higher level of the system (e.g., flocking behaviour of birds; or morphogenesis of the organism) that are not in the same sense `simulated', but realize the same kind of thing (ibid., p. 33): `The crucial point is that we have captured -- within an aggregate of artificial entities -- a bona-fide lifelike behavior, and that the behavior emerges within the artificial system in the same way that it emerges in the natural system'; `... a properly organized set of artificial primitives carrying out the same functional roles as the biomolecules in natural living systems will support a process that will be "alive" in the same way that natural organisms are alive. Artificial Life will therefore be genuine life -- it will simply be made of stuff that is different from the life that has evolved here on Earth.' (ibid.).
Several criticisms have been raised against this thesis (Pattee 1989, Cariani 1991, Sober 1991, Emmeche 1991a, b, 1992). One deep problem is, that it is not at all well substantiated that the behaviour in the artificial computational system emerges in the same way as it does in Nature. First, the concept of emergence is not well defined (see below). Second, a finite state determined system computing its future states acts on the basis of a set of predefined primitives and state transition rules that are unchanged during the run of the simulation. This may seemingly allow for the support of `higher level phenomena' like coherent lifelike patterns of behavior (and indeed anyone who have observed a computer running cellular automata of the Wolfram type, the `Game of Life', the `Boid' flocking simulator or similar kinds of simulations, have been struck by the peppy activity on the screen); however, these need an external interpretation to be classified and understood as being alive. To the extent that life is not intrinsically computational (which is not necessarily implied by the thesis of life as an emergent phenomenon of medium-independent form), it is not at all clear that the small programs that by mutations, `cross-over' and other `genetic operators' develop in the information processing structures of the computer are alive. The `genetic algorithm' approach for instance, originally invented by the computer scientist John Holland, shuffles the instructions of the programmes like sexual reproduction in organisms recombines the content of the genomes every generation. This computational technique, clearly inspired by a biological process, have proven to be a strong tool in various simulations. Thus, in a study of the evolution of behaviour, Collins and Jefferson (1991) simulated an evolving population of ant colonies by combining genetic algorithms with neural networks, but the authors still conceived of their work as a `computational model', and only spoke loosely about `artificial evolution'. Such investigations depend heavily on the external ascription of labels with a biological semantic content to the input as well as the output of the simulation. There is nothing suspicious in that so long as the semiotics of the model is kept clear, and is distinguished from the biosemiotics of the real subject. No real ant communication, no new pheromones, and no genuine autonomous living processes emerge from the computational setup.
Even if we speak loosely about lifelike behaviour of the computational model, such a model can be shown to be equivalent to a formal system, which does not have the creative causal power of natural systems (the `component systems' of Kampis 1991) to generate new properties and thus potentially new signs during its evolution. That does not mean that there is not a high degree of `logical depth' in the model in the sense of a huge combinatorial space of bit-structures based on the algorithmic processing of information, and that interesting structures -- that can be interpreted as stable patterns of behavior, attractor phenomena, `co-evolution of programs' etc. -- cannot appear with the execution of the basic simulation program. But one should not confuse this computational complexity of the model (and the biological interpretation of it) with the intrinsic complexity seen in the autonomous creation of new structures in the evolution of living beings. The latter process is independent of any external observer's knowledge of it, and independent of computational support in the form of a computing device.
This is especially clear when the life-property modeled is self-reproduction, one of the few crucial criteria of life. Natural self-reproduction is complete in the sense that the information needed for guiding the process is fully contained in and integrated with the cell or organism being reproduced. In a computational (machine) model of self-reproduction (whether it is based on `universal construction' as in von Neumann 1966, or more elegant designs such as the loop structure of Langton, see Langton 1986, 1989) the `reproduced' entities -- visualised on a screen as specific configurations of automata states duplicating themselves -- do not really as intended contain all the information needed for determining the process of reproduction. From a purely formal view this might be the case, but the physical machine (that realizes the process and which is not reproduced) supports the embedding universe of the reproducing automata and acts as a co-determiner of the process, but is not itself determined by it. One could try to overcome this difficulty by simulating the physical machine in a higher level model of a self-reproducing system (as suggested by von Neumann), but in such a model, there is still an additional external machine whose determination does not depend on the process of reproduction. The information responsible for `self-reproduction' is not completely localised within the configuration, and the external additional specification (by the embedding `universe' and supporting machine) is equally important for the process. In an autonomous living system, we cannot make the distinction between the entity being reproduced and an ultimate machine whose properties do not depend on the process of reproduction and which is not reproduced itself. DNA is not an external `knowledge' or `description' of the cell, but forms an integral part of the very system; the informational determination is a tacit activity that expresses this information causally (as emphasized by Kampis and Csányi 1991). It is the intrinsic and causal property of the biosemiotics of the cell that explains why real self-reproduction is complete, while modeled self-reproduction involves external sign-relations between observer and the system modeled.
(B) Often a set of qualitative criteria is given as a `check-list' to support the claim of medium-independence and the view that computer viruses or other computer programs with the ability to `evolve' may be an instance of genuine life. Farmer and Belin (1991) give a list of eight criteria, the first of which is a restatement of the idea of life as a pure phenomenon of form. This -- as the authors confess -- incomplete and imprecise list of properties associated with life reads as follows: 1. Life is a pattern in spacetime (rather than a specific material object). 2. Self-reproduction, in itself or in a related organism. 3. Information-storage of a self-representation. 4. Metabolism that converts matter/energy. 5. Functional interactions with the environment. 6. Interdependence of parts within the organism (preserve the identity of the organism; ability to die). 7. Stability under perturbations of the environment. 8. The ability to evolve.
The authors want the list to be specific for life and do not include growth. Growth is not a specific property; `there are many inanimate structures such as mountains, crystals, clouds, rust, or garbage dumps that have the ability to grow. Many mature organisms do not grow' (ibid., p. 818). But exactly the same sort of difficulty applies to other properties. Farmer and Belin of course know that the list is `far from adequate'; nevertheless they use it as an argument that one should consider computer viruses as being alive. Though `not as fully "alive" as their biological counterparts', but as computers become more prevalent, complex, and interconnected, the authors guess that they will `progress far beyond its initial form' (ibid., p. 821). They interpret the conceptual problems with the various criteria and numerous exceptions as an indication that life is a `continuum property of organizational patterns, with some more or less alive than others' (p. 819).
It is reasonable to suggest that life when it appeared on Earth and in the first phases of its evolution was organized in more simple ways, and therefore a continuum existed between protobiological systems of macromolecules with a primitive metabolism and the first genuine cells. We can also imagine the origin of life as a sudden transition from quasi-stable disordered population of molecules to an ordered state of catalytic active `cooperating' molecules corresponding to a metabolism proper, i.e., a discontinuous transition (cf. Dyson 1985). In any event, today there exists a bigger gab in degree of organization and specificity between the known examples of physical `self-organizing' dissipative structures (studied in thermodynamics and chemistry) and even the simplest forms of single celled organisms with highly specific biochemical, genetical, physiological and ecological properties -- life is exactly complex in the observer-system semiotic sense of requiring many different types of descriptions (Wimsatt 1976, Rosen 1977). Even simple viruses (not much more than a long macromolecule) are complex forms of life in the functional sense, because their own replication and evolution, as well as our descriptive determination of them, presupposes the full-fledged biological cell as a phenomenon much more complex than a single macromolecule. So the problems with the criteria as logical denominations of `life as an abstract phenomenon' have more to do with their inadequateness for capturing the fundamental aspects of life as distinct from other dynamical systems of physics. What is needed is a more adequate understanding of the creative, the historical and the autonomous aspects of living systems.
Looking at the list above from a semiotic point of view, one might think that what is missing is obviously the most fundamental criterion of life: semiosis. Indeed, there is a point in emphasizing the connection between biological and semiotic processes (e.g., Emmeche 1991b). As mentioned above, the history of life can to a large extent be understood as a natural history of signs. We know that the genetic code is a universal sign of the importance of semiosis in evolution. However, within semiotics as a discipline or as a more general point of view, it is not quite clear how low levels of physical organization to which we can apply the concept of sign. Sometimes it is suggested that `only living things and their inanimate extensions undergo semiosis' and that the cell is the `minimal semiosic unit' (Sebeok 1986, p. 346 f). Another and more metaphysical suggestion is that `the universe originated with the sign' (Anderson et al., p. 31). As it is not evident how one should use concepts of sign (sensu Peirce) or information (in the sense of Gregory Bateson, as `a difference that makes a difference') if there is no interpretant or (which is not the same) any interpreting organism to whom this information could make a difference, the first suggestion seems to be the best (see also Hoffmeyer and Emmeche 1991). So semiosis should be included in a more complete list of criteria.
Multiple intuitive conceptual models of life.
But this correction does not solve a more fundamental problem with criteria-checking as a way of deciding if alleged computational organisms is really `alive'. The problem appears when one reflects upon the attempt to catch the essentials of life in a coherent set of criteria. Alife research reveals that our concept of life is not a single one as we wish to think, and that no simple set of fundamental criteria can decide the status of our models and constructs when these are already embedded in specific preconceptions of what constitutes the aliveness of natural and artificial creatures. This `deconstructive' move (see below) may be a consequence of the inquiry of fundamental Alife research. It may help one to realize that at least the following different `conceptual models' of life as a biological phenomenon exist: An old idea about the organism as living animal; the scientific idea of the cell as the simplest living thing; life as an abstract phenomenon; and life in the cybernetic sense as a machine process that can be made by natural selection or by an engineer. Other ideas of life exist as well, but the focus is here restricted to the more-or-less rational ones.
Life in the biological sense has something to do with good old-fashioned organisms, made up of cells, these having a metabolism, constituted by specific macromolecules, etc. But this is already `too much'; we don't need all these details if we are just interested in drawing a boundary between the living and nonliving worlds. First and foremost, we have the old prototypic concept of life as a cluster of characteristics, more or less motivated by our knowledge of biology. There are many versions of this concept (e.g., Mayr 1982, Farmer and Belin 1991). Life has irritability and metabolism; it can self-reproduce, feed, and develop; it is vulnerable to illness and dead. A given instance may not have all of these properties, but it will have many.
We can then go on to see what is implied by these properties such as self-reproduction or metabolism. We have to consult different disciplines of biology and acquire knowledge of cells, membranes, enzyme systems, genetic coded information and other vital structures and processes. Thus the next step is a more specific conception of life described at a lower level (the cell and its macromolecular organization). Molecular biologists seldom care to define life -- they know it when they have it, and they know that its complexity is immense when compared with ordinary organic chemistry.
Logically, the two conceptual models are not equal, there is no simple connection between the prototypic old concept of life and the modern concept that explicitly or implicitly involves the existence of cells. Already, we can see here, that in fact the biology of our time has not one, but two conceptions of life. In fact, Carl Sagan (1973) drew attention to five different definitions of life, based on physiology, metabolism, biochemistry, genetics and thermodynamics. Now, if we want to construct life artificially by abstracting its `logical form' (to use von Neumann's and Chris Langton's words) and subsequently realize this form in other media, we get two different things depending on our point of departure.
From the first concept, we get neo-cybernetic life, animats, robots and other lifelike devices (Meyer and Wilson 1991). That is, machines that are living in the sense of being seemingly autonomous in their movements, with sensors and effecters and an internal structure that coordinates input information and output behaviour. (This is, of course not autonomy in the strict sense of a cell as a `component system' (Kampis 1991) or an `autopoietic system' (Maturana and Varela 1980)).
What do we get from the conteptual model of life taken from molecular biology? It is not quite clear. Biologically this definition is more fundamental. Attempts to `realize its content artificially' would lead to a material copy of a cell. Maybe we first have to create `chemical artificial life' before we can make a complete artificial cell. But if an artificial material cell -- if we really want it to be living in the sense of being an autonomous self-reproducing metabolising unit -- has to be made up of the same kind of biochemical compounds as a natural cell, i.e. the same types of molecules such as DNA, proteins, carbohydrates etc., then it will probably be just another instance of the same kind. The point in making it will in some sense disappear, provided that it will have the same level of complexity. It will be merely a replica. Of course, there will be a point in trying to create a more primitive version of a self-maintaining metabolizing cell-system (Dyson 1985), if possible.
One could say, that one should make a formal version of the cell, eventually realized in a cellular automaton model (Langton 1986) in two or three dimensions, realizing all the formal properties of the physical cell in the cellular automata space (whatever we mean by formal properties in this context). As mentioned, there have been attempts to formalize the property of self-reproduction (though only with success in trivial instances of self-reproduction, cf. Kampis 1991), and this approach could be extended to the other properties of life. This is precisely the idea of strong Alife: Any lifelike phenomenon can be realized in other media, because life is a question of form, not constituent materials; it is an abstract phenomenon, a coherent process-structure, an informational structure emerging from lower-level local interactions. But what is revealed here, is that this notion functions as a separate intuitive conceptual model of life.
So life seems to be a multiple of phenomena if we list these conceptual models of life, or lifelike processes, some of which are directly connected with the ideas of the Alife research programme: (1) Good-Old-Fashioned Biological Organisms (GOFBO), i.e., life as a list of properties known partly from the common sense of daily life, partly from the sciences of biology, physiology, genetics, and so on. Most often this is life conceived of as animals -- thus GOFBA: Good-Old-Fashioned Biological Animals. (2) Modern Macromolecular-based Cells (MOMACE) as characterized by molecular biology: You know it when you have it in your test tube. (3) Abstract Life (ABLI), i.e., life as a space-time pattern that `realizes' some formal properties of biosystems either within a biochemical medium, or in a symbolic formal space. (4) Robotic Life (ROLI), i.e., neo-cybernetic life, animats, nano-robotic life and so on. One could even add (5) CYBERlife; the idea of creating lifelike structures in a Virtual Reality to which we can relate through hyper-media; and (6) other non-scientific definitions and conceptions of life.
From a point of view of traditional biology -- as the study of general principles of life -- this is rather surprising. Artificial Life may help us to see that the idea of universality of the fundamental principles of life may be a presupposition, a metaphysical prejudice with a questionable basis. Traditional biology has been haunted by a lot of conceptual dualisms and metaphysical contradictions (as discussed by Woodger (1929) and Oyama (1985)) pertaining to the methods of investigation as well as the subject matter: The dualism between structure and process, form and function, part and whole, inheritance and environment, contingency and necessity, holism and reductionism, vitalism and mechanism, energy and information, concept and metaphor. The construction of Artificial Life may help to dissolve some of these dualisms, or maybe combine or re-invent them in more fruitful ways, and inseminate new ideas about the nature of living beings. In this perspective Artificial Life can be seen as a new way of `reading' the science of biology. Using a metaphor from literary criticism we may call it a deconstructive reading: Alife actuates a deconstruction of the Good-Old-Fashioned-Biological Life. The very opposition between living and dead nature -- the organismic and the inorganic domain -- that has been constitutive for the whole science of biology since its definition in the beginning of the nineteenth century -- may be reconstructed in a new framework, drawing on insights gained from disciplines outside Artificial Life (e.g., the thermodynamic study of self-organization and dissipative structures in physical systems, or the mathematical study of complexity in computational systems) that could represent common principles for living as well as non-living phenomena.
The system-observer semiotic criticism of naive conceptions of life in abstract and natural systems should, however, not go too far and surmise any concept of life to be no more than a `social construct' or similar relativistic speculation. Here, the importance of the complementarity between the two kinds of semiotics becomes clear. Biosemiotics as a field of research should provide us with realistic accounts of the evolution of sign functions from the primitive to the more and more sophisticated levels of organization. If one wants to retain an evolutionary world view, one must realize that the specific form of cultural semiosis called science in fact can account for real aspects of the genesis of communication systems throughout the history of the world.
If molecular biology is mainly reductionist in approach, one of the suggested promises of Alife research is, that by allowing for emergence of lifelike processes in the computer and ultimately for `synthesising life', Alife is a way to overcome the gab between holism and reductionism in theoretical biology. However, the idea that life as a medium-independent phenomenon may emerge in a sufficiently complex computational setup was said to be based on an insufficiently defined concept of emergence. How could this concept be clarified? This was discussed at several occasions during the First European Conference on Artificial Life (December 11-13, 1991, in Paris); there were several suggestions but none came up with a final answer. I will close this note by a few hints and point to some relevant literature. The concept of emergence had been seen as a quasi-vitalistic concept, but now it is widely recognized that vitalism is not the issue and many think the term is only meant to imply emergence in the `innocent' sense, that there exist specific (emergent) properties of a composite entity not possessed by any of its parts. The existence of such properties is trivial. The key issue involves questions of observation of emergents; of the relationship between emergence and the evolution of new levels of hierarchically organized systems; and also of reduction -- i.e., whether emergent properties can be reduced to properties of the components so that a theory of the simple parts (e.g., the molecules of a cell) could in principle predict emergent properties at the higher level.
Better definitions of emergence should help us to decide whether the `emergent phenomenon' of a model also pertains to the thing modeled. With respect to models of self-reproduction, a preliminary answer is that this property indeed seems to be an emergent (or systems) property of real organisms relative to their macromolecular parts, while the representation of this property in the models seen so far involves quite different schemes of formal description-guided production, as discussed above.
The work of Cariani (1989, 1991) is significant, because it clearly separates three distinct concepts of emergence and advocates for the importance of specifying the frame of reference of the observer in deciding whether a set of computations is really emergent. The concepts are:
(1) Computational emergence. This is what the Alife society speaks loosely about; it is not well defined but the idea is, as we have seen, that local micro-deterministic computational interactions together determine the emergence of coherent behavior at the macro-level of the system. But according to Cariani, this is fundamentally based on ascriptions by the interpreter and hence is not intrinsic to the formal system: `The interesting emergent events that involve artificial life simulations reside not in the simulations themselves, but in the ways that they change the way we think and interact with the world. Rather than emergent devices in their own right, these computer simulations are catalysts for emergent processes in our minds; they help us to create new ways of seeing the world.' (1991, p. 790).
(2) Thermodynamic emergence. This term alludes to various studies of the physics and chemistry of self-organizing systems in order to describe how stable structures can arise far from thermodynamic equilibrium. One problem here is that the theory of thermodynamics of these systems is not yet connected with a biological theory of the appearance of new functions in evolution.
(3) Emergence relative to a model. This is Cariani's preferred concept. As I interpret it, it has two aspects, a system-observer semiotical and a potentially biosemiotic one. The first one is the best developed in Cariani: (a) It sees emergence as the deviation of the behaviour of a physical system from an observer's model of it. If one is observing a system which changes its internal structure and behavior in such a radical way that one needs to change ones model to `track' the system's behavior in order to continue to predict its actions, then the change is truly emergent relative to the model. The crucial point in case of Alife simulations is that the observer can in principle always choose his frame of reference so as to be able to fully predict the track of the system. The observables can be chosen to coinside with the computable states of a finite state automaton equivalent with the simulation. Therefore, simulations will always be non-emergent. (However, in the case of Alife models with chaotic and algorithmically complex behaviour this form of `prediction' will be the somewhat pathological form of predictive model duplicating the very system -- cf. the `unsimulatable complexity' of Pagels (1989, p.101). This may therefore be seen as a quasi-form of genuine emergence, rather than non-existent.) (b) Emergence relative to a model is concerned with the formation of global structures (emerging from the action of local rules or entities) which subsequently constrain and alter local interactions (Cariani 1991, p. 778). It is tempting to see exactly this kind of non-reducible `downward causation' in an evolutionary perspective to account for the action of genetic information as a biosemiotic boundary condition constraining the generation of form in phylogeny and ontogeny. This aspect of the theory remains to be developed. A major challenge is to develop a coherent theory of emergent structures which unite the formal, the physical, the functional, and the evolutionary aspects of emergent properties.
In a workshop organized by Jesper Hoffmeyer on `Biosemiotics and biotechnology: The psycho-neuro-immunological case' (Oct. 29 - Nov. 2, 1991, in Denmark, Tisvilde) the Norwegian mathematician Nils A. Baas presented an informal version of a general theory of complexity, hierarchies, emergence and evolution. According to Baas, these four interrelated phenomena (the `hyperstructure' of his theory) are always realized in biological systems, and the systems of Alife research are interpreted as supporting this claim. Whenever we encounter life, it must be hierarchically organized, and hierarchies are the things that have had the time to evolve from simple to complex structures; complex in the algorithmic sense of needing a long `programme' for their specification or a long route of computational development (`logical depth'). Hierarchies make complexity manageable through several levels of organization; life cannot do with just one macro-level/micro-level distinction. Evolution by natural selection is the process which gives rise to new levels; in a certain sense by this process the environment acts as an observer that `sees' and `acts upon' higher level properties, thereby establishing recurrent forms of interactions within and between the different levels.
Emergent properties must be observable, but they appear because of the system of interactions among the lower level objects (and not because of observation). Baas too distinguishes several concepts of emergence, among them are (a) deducible/computable emergents (cf. Cariani), and (b) observable emergents, which is the more profound type. A genuine example of an observable emergent relative to a formal system is the observation of the truth function value (true) of a string of symbols which is not deducible from the axioms and rules of inference of the formal system itself, i.e., non-computable but observable theorems (cf. Gödel's theorem on the existence of formally undecidable propositions in any formal system of arithmetic). Baas also includes nonlinear dynamical systems where the superposition principle fails as an example (i.e., where properties of the parts do not `add up', so to speak, so that the whole is more and more is different). In a sense, this may be deduced, but in another sense, this can eventually only be `observed' by computational methods. Thus Baas would not by a priori arguments dismiss emergent properties from Alife simulations.
In Baas's general theory, both the traditional micro-deterministic upward causation and the more controversial downward causation (that may even occur across several levels) are allowed for; whether these forms of causation are actualised in real systems is an empirical question (Baas 1992).
Artificial Life models are putting new fuel to discussions of old riddles in philosophy and theoretical biology. Semiotics should intervene this field. The study of the semiotics of emergence and computation is in its formative stage. For more general treatments of emergent properties, see the papers by Klee (1984) and Cariani (1989, 1991). On computational emergence in general, see Langton (1989) and Farmer and Belin (1991) and other papers in these two proceeding volumes. References to the critical literature is given above. A special area addresses the possible emergence of `computations' (not to be confused with the general notion of computational emergence) in a model of phase-transitions of nonlinear dynamical systems, see the volume edited by Forrest (1990). This area too needs semiotic reflections. A nice introduction to the problem of emergence in theoretical biology is the paper by Fernandez, Moreno and Etxeberria (1991). These authors argue, among other things, that the `evolution of organisms implies the appearance of states and/or transitions for which no complete pre-definition is possible'. Modeling life, one should bear this in mind.
1939: The Theory of Emergence. Philosophy of Science 6 (1), 1-16.
Anderson, Myrdene; Deely, John; Krampen, Martin; Ransdell, Joseph; Sebeok,
Thomas A., and Uexküll, Thure von
1984: A semiotic perspective on the sciences: Steps toward a new paradigm. Semiotica 52 (1/2), 7-47.
Baas, Nils A.
1992: Hyperstructures -- a framework for emergents, hierarchies, evolution and complexity. (Preprint from The Mathematical Institute of Trondheim University, to be published).
1989: On the Design of Devices with Emergent Semantic Functions. Ph.D. dissertation, Department of Systems Science, State University of New York at Binghamton. Ann Arbor: University Microfilms.
1991: Emergence and Artificial Life. In Artificial Life II. Santa Fe Institute Studies in the Sciences of Complexity Proceedings Vol. X, Christopher G. Langton, Charles Taylor, J. Doyne Farmer and Steen Rasmussen (eds.), 775-797. Redwood City, Calif.: Addison-Wesley.
Collingwood, R. G.
1945: The Idea of Nature. London: Clarendon Press. (paperback: Oxford University Press, 1960)
Collins, Robert J. and Jefferson, David R.
1991: AntFarm: Towards Simulated Evolution. In Artificial Life II. Santa Fe Institute Studies in the Sciences of Complexity Proceedings Vol. X, Christopher G. Langton, Charles Taylor, J. Doyne Farmer and Steen Rasmussen (eds.), 579-601. Redwood City, Calif.: Addison-Wesley.
1985: Origins of Life. Cambridge, Mass.: Cambridge University Press.
1990: Det Biologiske Informationsbegreb. Århus: Kimære.
1991a: Det levende spil. Biologisk form og kunstigt liv. København: Munksgaard.
1991b: A Semiotical Reflection on Biology, Living Signs and Artificial Life. Biology and Philosophy 6 (3): 325-340.
1992: Life as an abstract phenomenon: is Artificial Life possible? In: Proceedings of the 1st European Conference on Artificial Life (to be published by the MIT Press/Bradford Books).
Emmeche, Claus and Hoffmeyer, Jesper
1991: From language to nature - the semiotic metaphor in biology. Semiotica 84 (1/2), 1-42.
Farmer, J. Doyne and Belin, Aletta d'A.
1991: Artificial Life: The coming Evolution. In Artificial Life II. Santa Fe Institute Studies in the Sciences of Complexity Proceedings Vol. X, Christopher G. Langton, Charles Taylor, J. Doyne Farmer and Steen Rasmussen (eds.), 815-838. Redwood City, Calif.: Addison-Wesley.
Fernandez, Julio, Moreno, Alvaro and Etxeberria, Arantza
1991: Life as emergence: the roots of a new paradigm in theoretical biology. World Futures 32, 133-149.
Forrest, Stephanie, ed.
1990: Emergent Computation. (= Physica D vol.42). Amsterdam: North Holland.
(in press): Semiotic aspects of biology: Biosemiotics. In Semiotics: A Handbook of the Sign-Theoretic Foundations of Nature and Culture, Robert Posner, Klaus Robering and Thomas A. Sebeok (eds.). Berlin & New York: Mouton de Gruyter.
Hoffmeyer, Jesper and Emmeche, Claus
1991: Code-duality and the semiotics of nature. In On Semiotic Modeling. Myrdene Anderson and Floyd Merrell (eds.), 117-166. Berlin & New York: Mouton de Gruyter.
1991: Self-modifying Systems in Biology and Cognitive Science. New York: Pergamon Press.
Kampis, G. and Csányi, V.
1991: Life, Self-reproduction and Information: Beyond the Machine Metaphor. Journal of theoretical Biology 148, 17-32.
Klee, R. L.
1984: Micro-determinism and concepts of emergence. Philosophy of Science 51, 44-63.
1989: Artificial Organisms. In Artificial Life. Santa Fe Institute Studies in the Sciences of Complexity Proceedings Vol. VI, Christopher G. Langton (ed.), 49-77. Redwood City, Calif.: Addison-Wesley.
Langton, Christopher G.
1986: Studying Artificial Life with Cellular Automata. Physica D 22, 120-149.
1989: Artificial Life. In Artificial Life. Santa Fe Institute Studies in the Sciences of Complexity Proceedings Vol. VI, Christopher G. Langton (ed.), 1-47. Redwood City, Calif.: Addison-Wesley.
Marijuán, Pedro C.
1991: Enzymes and theoretical biology: stetch of an informational perspective of the cell. Biosystems 25, 259-273.
Maturana, H.R. and Varela, F.J.
1980: Autopoiesis and Cognition. The Realization of the Living, Dordrecht: Reidel.
1982: The Growth of Biological Thought. Cambridge, Massachusetts: The Belknap Press of Harvard University Press.
Meyer, Jean-Arcady and Wilson, Stewart W. (ed.)
1991: From Animals to Animats. Proceedings of the First International Conference on Simulation of Adaptive Behavior. Cambridge, Mass.: The MIT Press.
Morgan, C. Lloyd
1923: Emergent Evolution. London: Northgate and Williams.
1985: The Ontogeny of Information. Developmental Systems and Evolution. Cambridge: Cambridge University Press.
Pagels, Heinz R.
1989: The Dreams of Reason. New York: Bantam Books.
1926: Emergence. Journal of Philosophy 23, 241-245.
Pattee, Howard H.
1989: Simulations, realizations, and theories of life. In Artificial Life. Santa Fe Institute Studies in the Sciences of Complexity Proceedings Vol. VI, Christopher G. Langton (ed.), 63-77. Redwood City, Calif.: Addison-Wesley.
1977: Complexity as a system property. Int. J. General Systems 3, 227-232.
1991: Life Itself. A Comprehensive Inquiry Into the Nature, Origin, and Fabrication of Life. New York: Columbia University Press.
1973: Life. In The New Encyclopædia Britannica, 15th ed., Macropædia vol. 10. London: William Benton Publ.
Sebeok, Thomas A.
1977: How Animals Communicate. Bloomington: Indiana University Press.
1986: The doctrine of signs. Journal of Social and Biological Structures 9, 345-352.
1987: Toward a natural history of language. Semiotica 65 (3/4), 343-358.
1991: Learning from Functionalism -- Prospects for Strong Artificial Life. In Artificial Life II. Santa Fe Institute Studies in the Sciences of Complexity Proceedings Vol. X, Christopher G. Langton, Charles Taylor, J. Doyne Farmer and Steen Rasmussen (eds.), 749-765. Redwood City, Calif.: Addison-Wesley.
Taylor, Charles E.
1991: "Fleshing Out" Artificial Life II. In Artificial Life II. Santa Fe Institute Studies in the Sciences of Complexity Proceedings Vol. X, Christopher G. Langton, Charles Taylor, J. Doyne Farmer and Steen Rasmussen (eds.), 25-38. Redwood City, Calif.: Addison-Wesley.
1988: Life's Devices, The Physical World of Animals and Plants. Princeton: Princeton University Press.
von Neumann, John
1966: The Theory of Self-Reproducing Automata. A. W. Burks, (ed.). Urbana, IL: University of Illinois Press.
von Uexküll, Thure von
1984: Semiotics and the problem of the observer. Semiotica 48 (3/4), 187-195.
1986: Medicine and semiotics. Semiotica 61 (3/4), 201-217.
1989: Naturwissenschaft als Zeichenlehre. Merkur 43 (3), 225-234.
Wimsatt, William C.
1976: Complexity and organization. Boston Studies in the Philosophy of Science 27, 174-193.
Woodger, J. H.
1929: Biological Principles. A Critical Study. London: Kegan Paul.