Advanced Issues on Cognitive Science and Semioticsedited by Priscila Farias and Joo Queiroz

2006 All rights reserved. No part of this book may be reproduced in any form by any electronic or mechanical means (including photocopying, recording, or information storage and retrieval) without permission in writing from the publisher. Advanced Issues on Cognitive Science and Semiotics Edited by Priscila Farias and Joo Queiroz ISBN

Contents

Preface

1

1

A-life, organism and body: the semiotics of emergent levels Claus Emmeche 5

2

Semiosis and living membranes Jesper Hoffmeyer

19

3

Biosemiotics and the foundation of cybersemiotics Sren Brier

37

4

Information and direct perception: a new approach Anthony Chemero 59

5

Revisiting the dynamical hypothesis Tim van Gelder

73

6

The dynamical approach to cognition: inferences from language Robert F. Port 93

7

Models of abduction Paul Bourgine Contributors

123

139

PrefaceJoo Queiroz & Priscila Farias

T

here seems to be a consensus that many of the classic problems in cognitive science are strongly connected to the fundamental issues of information, meaning and representation. There is, indeed, no domain of research, interested in cognitive processes, that has not been concerned, at some point, with these notions. At the same time, as many authors have noted, these seminal elements of investigation are frequently obscured by terminological conflicts, uncertainties and vagueness. Problematic as they may be, notions of information and representation, or its models, even if only implicitly, are always present in studies on cognitive systems, urging for reliable and sound theoretical basis. North-American pragmatist Charles Sanders Peirce, founder of the modern theory of signs, defined semiotics as a science of the essential and fundamental nature of all possible varieties of meaning processes (semiosis). Peirces concept of semiotics as the formal science of signs, and the pragmatic notion of meaning as the action of signs, have had a deep impact in philosophy, psychology, theoretical biology, and cognitive science. The reader will find here a collection of papers that present, from different perspectives, an attempt to relate semiotics and cognitive science with linguistics, logic, and philosophy of biology. As a first broad account of those subjects, it does not specifically focus on or privilege any of the different approaches that have been proposed up to now, but instead gives the reader the opportunity to consider the various directions and topics of research that emerge from such relations. Several chapters focus on the logic of semiosis, and propose a range of different analysis of the nature of mediation as an action of interpretation. Emmeche and Hoffmeyer explore the explanatory power of the biosemiotic approach as an alternative (or complementary) theoretical frame to the physicalist point of view. Emmeches chapter comments upon some of the open problems in artificial life from the perspective of qualitative organicism and the emergent field of Peirceanoriented biosemiotics. Hoffmeyer is interested in the rise and evolution of qualia. According to the author, a scientifically consistent theory might be developed on the basis of what he called semiotic materialism. Semiotic materialism considers qualia as an evolved instantiation of a semiotic freedom that was latently present in our universe from the beginning. It claims that our universe has a built-in tendency to produce organized systems possessing increasingly more semiotic freedom in the sense that the semiotic aspect of the systems activity becomes more and more autonomous relative to its material basis.

1

2

Brier proposes a transdisciplinary approach able to postulate a unified theory of information, cognition and communication. His proposal involves the setting of an epistemological framework that takes into account recent developments from ethology, second order cybernetics, cognitive semantics and pragmatic linguistics, and that builds on basic concepts from biosemiotics. According to his view, all living organisms are immersed in a web of signs, participating in what he defines as sign games, which eventually lead to wittgenteinian human language games. Chemero challenges the notion of information in Gibsonian-oriented approaches to perception. He proposes an understanding of direct perception and information that differs from the ecological psychology orthodoxy, the Turvey-Shaw-Mace view. According to his view, perception is direct when the perceiver and perceived are coupled and their relationship is unmediated by mental representations. Dynamical hypothesis on cognition, by rejecting the idea that cognition is to be explained exclusively in terms of internal representations, are close to Chemeros approach. Two papers discuss the use of dynamical system theory (DST) as a strategy for modeling cognition. Van Gelder is interested in exploring the essence of dynamical cognitive science, and to show how it differs from traditional computational cognitive science. His strategy is to oppose the traditionalist slogan that cognitive agents are digital computers to the dynamicist claim that cognitive agents are dynamical systems. Ports argument is about an essentially temporal (continuous, historical) feature of verbal language generation and processing on a phonological level of description. He suggests that, to constitute a general dynamic theory of language, from phonology to semantics, we should understand linguistic events and structures as temporal processes. One of Peirces most original contributions to the studies on cognitive inference is his development of the concept of abduction. However, Peirces abductive inference received, so far, relatively little attention. In the last chapter of this book, Bourgine proposes axiomatic and geometric models for abduction, close to the framework of belief revision, in an effort to formally capture what differentiates it from induction and deduction, while preserving a coherent relation between the three types of reasoning envisaged by Peirce. Bourgines models are exemplified in the context of three perspectives from the field of cognitive science: cognitivism, connectionism and constructivism. The various strategies presented here may be considered non-standard, and therefore remain peripheral in their fields. It is still too early to properly evaluate all the perspectives opened up by the research frontiers presented in this book. Indeed, it is premature to assert that they may one day constitute new scientific paradigms. What all these perspectives suggest, however, are alternative

approaches to basic principles of cognition, information and representation. By offering innovative and consistent propositions, we hope that the ideas presented in this book may constitute a fresh breath, and point out important new directions to be followed in the future.Acknowledgments The authors would like to acknowledge the support received, in the form of research grants, from FAPESP - The State of So Paulo Research Foundation.

3

4

A-life, organism and body: the semiotics of emergent levelsClaus Emmeche

Introduction: Organicist philosophies Artificial Life research raises philosophical questions, just as cognitive science involves philosophy of mind. No clear demarcation line can be drawn between science and philosophy; every scientific research programme involves metaphysical assumptions and decisions about how to interpret the relations between experiment, observation, theoretical concepts and models (this was also evident when Artificial Life originally was formulated by C.G. Langton in the late 1980s; cf. Emmeche 1994). Yet we should not conflate questions that may be answered by science with questions that by their very nature are conceptual and metaphysical. The aim of this paper is to address from the perspective of biosemiotics a subset of the open problems (as described by Bedeau et al. 2000) raised by Artificial Life research, including wet Alife, about the general characteristics of life; the role and nature of information; how life and mind are related; and their relations again to culture and machines. Biosemiotics as the study of communication and information in living systems may provide some inspiration and conceptual tools for inquiry into these theoretical and philosophical issues. Firstly it is apt briefly to introduce organicism as a mainstream position in philosophy of biology, and also a variant called qualitative organicism, and then introduce biosemiotics as a non-standard philosophy of biology. Neither qualitative nor mainstream organicism is specific research paradigms; they are more like general and partly intuitive stances on how to understand living systems in the context of theoretical biology. Organicism. In its mainstream form (cf. Emmeche 2001) organicism endorses these theses: (a) non-vitalism (no non-physical occult powers should be invoked to explain living phenomena); (b) non-mechanicism (living phenomena cannot be completely described merely by mechanical principles, whether classical or quantum); (c) emergentism (genuine new properties are characteristic of life as compared with purely physical non-living systems) implying ontological irreducibility of at least some processes of life (though methodological reductionism is fully legitimate); (d) the teleology of living phenomena (their purpose-like character) is real, but at least in principle explainable as resulting from the forces of blind variation and natural selection, plus eventually some additional order for free (physico-chemical self-organization). What is studied within an organicist perspective as emergent properties are seen as material structures and processes within several levels of living systems (developmental

5

systems, evolution, genetic and biochemical networks, etc.), all of which are treated as objects with no intrinsic experiential properties. Mayr (1997) acknowledged his position as organicist, and mainstream organicism is widely accepted among biologists, even though the position was often mixed-up with vitalism (see also El-Hani & Emmeche 2000, Gilbert & Sarkar 2000). Accordingly, there are no principled obstacles to the scientific construction of life and mind as emergent phenomena by evolutionary or bottom-up methods. Qualitative Organicism. This is a more radical position differing from mainstream organicism in its appraisal of teleology and phenomenal qualities. It emphasizes not only the ontological reality of biological higher-level entities (such as selfreproducing organisms being parts of historical lineages) but also the existence of qualitative experiential aspects of cognitive behavior. When sensing light or colors, an organism is not merely performing a detection of external signals which then get processed internally (described in terms of neurochemistry or information processing); something more is to be told if we want the full story, namely about the organisms own experience of the light. This experience is seen as real. It may be said to have a subjective mode of existence, yet it is an objectively real phenomenon (Searle 1992 emphasized the ontological reality of subjective experience; yet, most of the time only in a human context). As a scientific stance qualitative organicism is concerned not only with the category of primary measurable qualities (like shape, magnitude, and number) but also with inquiry into the nature of secondary qualities like color, taste, sound, feeling, and the basic kinesthetic consciousness of animal movement. A seminal example of qualitative organicism is Sheets-Johnstone 1999. The teleology of living beings is seen as an irreducible and essential aspect of living movement, in contrast to mere physical change of position. This teleology is often attributed to a genuine form of causality (final causation, cf. Van de Vijver et al. 1998), and qualitative organicisms assessment of the reality of an instance of artificial life will partly depend on how to interpret the causality of the artificial living system. Biosemiotics. The study of living systems from the point of view of semiotics, the theory of signs (and their production, transfer, and interpretation), mainly in the tradition of the philosopher and scientist Charles Sanders Peirce (18391914) but also inspired by the ethologist Jakob von Uexkll (1864-1944), has a long and partly neglected history in 20th Century science (Kull 1999 for the history, Hoffmeyer 1996 for an introduction). It re-emerged in the 1990s and is establishing itself as a cross-disciplinary field attempting to offer alternatives to a gene-focused reductionist biology (much like one of the aims of Artificial Life, and indeed inspired by it), by gathering researchers for a new approaches to biology, or a new philosophy of biology, or ultimately with the hope to bridge the gab between science and the humanities. The semiotic approach means that cells and organisms are not seen primarily as complex assemblies of molecules,

6

as far as these molecules rightly described by chemistry and molecular biology are sign vehicles for information and interpretation processes, briefly, sign action or semiosis. A sign is anything that can stand for something (an object) to some interpreting system (e.g., a cell, an animal, a legal court), where standing for means mediating a significant effect (called the interpretant) upon that system. Thus, semiosis always involves an irreducibly triadic process between sign, object and interpretant. Just as in chemistry we see the world from the perspective of molecules, in semiotics (as a general logic of sign action) we see the world from the perspective of sign action, process, mediation, purposefulness, interpretation, generality. Those are not reducible to a dyadic mode of mechanical actionreaction, or merely efficient causality. The form of causality governing triadic processes is final causation. Organisms are certainly composed of molecules, but these should be seen as sign vehicles having functional roles in mediating sign action across several levels of complexity, e.g., between single signs in the genotype, the environment, and the emerging phenotype. Biosemiotics is a species of qualitative organicism for these reasons: (i) It holds a realist position regarding sign processes of living systems, i.e., signs and interpretation processes are not merely epistemological properties of a human observer but exists as well in nature, e.g., in the genetic information system (ElHani et al. 2006). (ii) Biosemiotics interprets the teleology of sign action as related to final causation (Hulswit 2002). (iii) The qualitative and species-specific subject of an animal (i.e., its Umwelt understood here as a dynamic functional circle of an internal representation system interactively cohering in action-perception cycles with an environmental niche) can to some extent be studied scientifically by the methods of cognitive ethology, neurobiology and experimental psychology, even though the experiential feeling of the animal is closed to the human Umwelt (on Umwelt research, see papers in Kull 2001). (iv) Signs have extrinsic publicly observable as well as intrinsic phenomenal aspects. We can only access the meaning of a sign from its observable effects, a good pragmaticist principle indeed, but observation of the phenomenal experiences of another organism may be either impossible or highly mediated. However, reality exceeds what exists actually as observable. (v) Even though sign activity generally can be approached by formal and logic methods, sign action has a qualitative aspect as well. Due to the principle of inclusion (Liszka 1996) every sign of a higher category (such as a legisign, i.e., a sign of a type) includes a sign of a lower category (e.g., a sinsign, i.e., a sign of a token. A type has somehow to be instantiated by a token of it, just like any sign must be embodied). A symbol is not an index, but includes an indexical aspect, which again involves an icon. All signs must ultimately include (even though this might not be pertinent for their phenomenology) qualisigns, which are of the simplest possible sign category, hardly functioning as mediating any definite information, yet being signs of quality and thus having a

7

phenomenal character of feeling (Peirce preferred a type-token-tone trichotomy for the type-token dichotomy; a tone is like a simple feeling). The argument (v) may strike a reader not acquainted with Peirce as obscure, but it is a logical implication of the ontological-phenomenological basis of Peirces semiotics and points to an interesting continuity between matter, life and mind, or, to phrase it more precise, between sign vehicles as material possibilities for life, sign action as actual information processing, and the experiential nature of any interpretant of a sign, i.e., the effects of the sign upon a wider mind-like system (Emmeche 2004). To recapitulate, the biosemiotic notion of life is a notion of a complex web of sign and interpretation processes, typically with the single cell seen as the simplest possible autonomous semiotic system. Synthetic biosemiosis? Computers are semiotic machines (Nth 2003) and computers or any other adequate medium, such as a complex chemistry, can in principle function as a medium of genuine sign processes. Not all sign processes need to be biological, although all signs seem to involve at some point in their semiosis interpreters who typically would be biological organisms. Remember the distinction above between the interpretant as the effect or meaning of a sign and the interpreting system (or interpreter) as the wider system in which semiosis is taking place. So, what then is the biosemiotic stance regarding true synthetic life or wet artificial life? To answer this question, we have to consider, though more carefully than can be done here, (a) three non-exclusive notions of life; (b) the relation between the notions of organism, animal, body, and the general embodiment of various levels of signs processes; and (c) the semiotics of scientific models. This necessity of a precaution in assessment of the degree of genuineness of synthetic life in other media is related to another organicist theme: The thesis of irreducibility of levels of organization, or, as we shall interpret that thesis here, levels of embodied sign action. Life in Lebenswelt, biology and ontology Synthetic life provokes, of course, the general question what is life?; partly because of an intuition we (or some people) have from our ordinary life, as the German philosophers would say, from the Lebenswelt (life world) of human beings, that life (like death) is a basic condition we as humans hardly can control, know completely, or create. Now science seems to teach us otherwise. A contribution to clarify the issue is to be aware of the fact of the existence of at least three, non-exclusive notions of life. I will briefly sketch these: Lebenswelt life: A set of diverse, non-identical, culture-specific notions (determined by intuitive, practical, ideological, or social factors) of what it is to be alive, what life and dead is, why being alive and flexible is more fun than being dead and rigid, and so on. Science is distinct from, but not independent of, forms of the human Lebenswelt (just as scientific concepts can be seen as presupposing

8

and being a refinement of ordinary language). For biological relevant notions of life, we can talk about folk and experiential biology (Emmeche 2000). Biological life: The so-called life sciences are not interested in the life of the Lebenswelt as a normative phenomenon, but in the general physical, chemical and biological properties of life processes, as conceived of within separate paradigms of biology. This leads to several distinct ontodefinitions of life (Emmeche 1997) such as life as evolutionary replicators, life as autopoiesis, or life as sign systems (and probably many more). However, advances in biotechnology and biomedicine will tend to mix up, hybridize or create new boundary objects (sensu Star & Griesemer 1989) between the domains of bioscience and a life-world deeply embedded in technoscience. Ontological life: Depending upon the ontology chosen, an ontological notion of life is marked by distinctions to other, similarly general and essentials domains of reality. Take the ontology of Peirce, for instance. Here life is of the category of Firstness, it does not only include life in organisms, evolution or habit taking (which are of the category of Thirdness); life is seen as an all-inclusive aspect of the developing cosmos, on par with spontaneity and feeling: insofar as matter does exhibit spontaneous random activity (think of measurement error or Brownian motion), it still has an element of life left in it (Reynolds 2002, 151). Biosemiotics typically does not use Peirces broad ontological notion of life, but construes a notion of life derived from contemporary biology, as mentioned, life as organic sign-interpreting systems. But biosemiotics entails a thesis of the reality of ideal objects, including possibilities like a fitness space, virtuality in nature, or tendencies in evolution and development, and so the possibilities for final causes to prefer one tendency over another. Thus biosemiotics entails an ontological revolution admitting the indispensable role of ideality in this strict sense in the sciences (Stjernfelt 2002, 342). The invention of synthetic wet life may affect all the three non-exclusive preoccupations with understanding life, that is, life within a cultural context, life as studied by science, and life as a metaphysical general aspect of reality. To approach how this may come about, we must analyze some levels of embodied life from the perspective of an emergentist ontology. Level-specific forms of life and embodiment As a species of organicism, biosemiotics is an emergentist position. However, it is not so often that emergent levels of sign processes have been explicitly discussed (von Uexkll 1986, Kull 2000, El-Hani et al. 2004). The account given here should be seen as preliminary; the important point is not the number of levels (more fine-grained analyses may be done) but the very existence of separate levels of embodied sign action. See also Table 1.

9

The body of physics

The body of biology

The body of evo-devoresearch Vegetative swarm of cells coordinating multi-cellular communication with multiple organic codes

The body of zoology

The body of anthropology

The body of sociology

Self-moving, actionperception cycles, animation, kinesthetics

Physiologic-homeostatic units with a genetic code-plurality, and irritability

Table 1. Ordering relations between forms of embodiment. The epistemic dimension (top row) is shown by organizing those forms according to different domains of science each constituting its own objects; the ontic dimension (bottom row) is implied by an underlying ontology of levels of organization in Nature. Increasing specificity from left to right; for each new level the previous one is presupposed.

10

The emerging forms of embodiment of life could be suspected merely to reflect a historically contingent division of sciences; an objection often raised against simple emergentists level ontologies. Thus one should pay respect to the fallibilist principle and never preclude that new discoveries will fundamentally change the way we partition the levels of nature. The point is that from the best of our present knowledge we can construct some major modes of embodiment in which life and sign action plays crucially different roles, and in which we can place such broad phenomena as life, mind, and machines. Reflexivity is allowed for, so even the scientific description of these phenomena can be placed in this overall scheme of processes. A consequence is that wet artificial life is seen as a hybrid phenomenon of the body of biology and the body of sociology, as will be explained below. The emergent modes of embodiment, increasing in specificity (Table 1) are one-way inclusive and transcending: The human body includes the animate organism, which again presupposes multicellularity and basic biologic autopoiesis (but not vice versa). A human body (e.g., the body of a child, a soccer player, or a diplomat) as studied by anthropology is something more specific (i.e., in need of more determinations) than its being as an animal, thus transcending the mere set of animate properties (as having an Umwelt) and organismic properties (like growth, metabolism, homeostasis, reproduction), just as an organism is a physical system, yet transcending the basic physics of that system. That an entity or process at an emergent level Z is transcending phenomena at level Y has two aspects. One is epistemic, i.e., Zs description cannot adequately be given in terms of a theory generally accounting for Y, even though this Z-description in no way contradicts a description of the Y-aspects of Z. The other is ontological,

The life in societal institutions, habit formation

Language, culture-specific Lebenswelt

Complex dissipative, selforganizing structures

i.e., crucial properties and processes of Z are of a different category than the ones of Y, even though they may presuppose and depend on Y. Thus, a Z-entity is a highly specific mode of realizing a Y-process, not explained by Y-theory. The organism is a physical processual entity with a form of movement so specific that physics (as a science) cannot completely account for that entity. The organism is a very special type of physical being, as it includes certain purposeful (functional) part-whole relations, based upon genuine sign systems of which the genetic code is the most well known but not the only example. Here is a brief characteristic of the levels. Life as self-organization far from equilibrium Physics deal with three kinds of objects; first, general forces in nature, particles, general bodies (matter in bulk), and the principles (laws) governing their action; second, more specifically the structural dynamics of self-organized bodies (galaxies, planets, solid matter clusters, etc.); third, physical aspects of machines (artifacts produced by human societies and thus only fully explainable also by use of social sciences, like history of technology). One has often seen attempts to reduce all of physics to a formalism equivalent to some formal model of a machine, but there are strong arguments against the completeness of this programme (Rosen 1991), i.e., mechanical aspects of the physical world are only in some respects analogous to a machine. Some of the general properties of bodies studied in physics have a teleomatic character (a kind of directedness or finality, cf. Wicken 1987), which may be called thermo-teleology, as this phenomenon of directedness is best known from the second law of thermodynamics (a directedness towards disorder), or from opposing self-organizing tendencies in far from equilibrium dissipative systems. Often when physicists talk about life in the universe the reference is to preconditions for biological life such as self-organizing nonequilibrium dissipative processes, rather than the following level. Life as biofunctionality - organismic embodiment A biological notion of function is not a part of physics, while it is crucial for all biology. Biofunctionality is not possible unless a living system is self-organizing in a very specific way, based upon a memory of how to make components of the system that meet the requirement of a functional (autopoietic and homeostatic) metabolism of high specificity. For Earthly creatures this principle is instantiated as a code-plurality between a digital genetic code of DNA, a dynamic regulatory code of RNA (and other factors as well), and a dynamic mode of metabolism involving molecular recognition networks of proteins and other components (see the semiotic analysis by El-Hani et al. 2004). Symbolic, indexical and iconic molecular sign processes are all involved in protein synthesis. The symbols (using DNA triplets as sign vehicles) seem to be a necessary kind of signs for a stable memory to pick out the right sequences for the right job of metabolism.

11

12

This establishes a basic form of living embodiment, the single cell (a simple organism) in its ecological niche. This presupposes the workings of the physical body as a thermodynamic non-equilibrium system, but transcends that general form by its systematic symbolic memory of organism components and organismenvironment relations. Biosemiotics posits that organismic embodiment is the first genuine form of embodiment in which a system becomes an autonomous agent acting on its own behalf (cf. Kauffman 2000), i.e., taking action to secure access to available resources necessary for continued living. It is often overlooked that the subject-object structure of this active agent is mediated not only energetically by a structured entropy difference between organism and environment, but also semiotically, by signs of this difference; signs of food, signs of the niche, signs of where to be, what to eat, and how to trigger the right internal processes of production of organismic components the right time. The active responsitivity of the agent organism (based upon observable molecular signs) has, as an inner dimension, a quality of feeling, implied by what is in Table 1 called irritability at the level of a single cell. Irritability is a real phenomenon, well-known in biology, logically in accordance with a basic evolutionary matter-mind continuity, rationally conceivable, though impossible for humans to sense or perceive from within or empathetically know what it feels like, say, for an amoeba or an E.coli. It is highly conceivable that synthetic systems analogous to this level of embodiment may be produced some day. Life as biobodies coordinate your cells! Characteristics like multiple code-plurality (involving the genetic code, signal transduction codes, and other organic codes, see Barbieri 2003) and forms of semiotic coordination between cell lines cooperating and competing for resources within a multicellular plant or fungus are characteristics of the evolution of individuality (Buss 1987). The social life of cells within a lineage of organisms with alternating life cycles constitutes a special level of embodied biosemiosis, and special a coupling of evolution and development. It is the emergence of the first biobodies in which the whole body constrain the growth and differentiation of its individual cells (a form of downward causation, cf. El-Hani & Emmeche 2000). This multicellular level of embodiment corresponds to what was called a vegetative principle of life in Aristotelian biology, like that of a plant. Life as animate - moving your self! Here the body gets animated, we see a form of nervous code (still in the process of being decoded by neuroscience), and we see the emergence of animal needs and drives. When we consider animal mind and cognition, the intentionality of an animal presupposes the simpler forms of feelings and irritability we stipulate

in single cells (including the primitive free-living animals, such as protozoa, lacking a nervous system), yet transcends these forms by the phenomenal qualities of the perceptual spaces that emerge in functional perception-action cycles as the animals Umwelt. Proprioceptive semiosis is crucial for phenomenal as well as functional properties of animation (Sheets-Johnstone 1999). More generally, the animal body is a highly complex and specific kind of a multicellular organism (a biobody) that builds upon the simpler systems of embodiment such as physiological and embryogenetic regulation of the growth of specific organ systems, including the nervous system. These regulatory systems are semiotic in nature, and rely on several levels of coded communications within the body and their dynamic interpretations (Hoffmeyer 1996, Barbieri 2003). The expression the body of zoology in Table 1 is used to emphasize both its distinctness as a level of embodiment, and that zoology instead of being simply part of an oldfashioned division of the sciences should be the study of animated movement, including its phenomenal qualities. Life as anthropic - talk about life! With the emergence of humans comes language, culture, division work, desires (not simply needs, but culturally informed needs), power relations etc. The political animal not only lives and makes tools, but talks about it. Within this anthroposemiosis (von Uexkll 1986), the body is marked by differences of gender (not simply sex), age, social groups, and cultures. Life as societal get a life! After humans invented agriculture and states, more elaborated institutions could emerge; and social groups became informed and enslaved by organizational principles of all the sub-systems of a civilized real society (work, privacy, politics, consumption, economy, law, politics, art, science, technology, etc.). Humans discover the culture-specificity of human life, them and us. Reflexivity creeps in as civilization makes more and more ways to get a life. The body becomes societal (marked by civil life) and cyborgian (crucially dependent upon technology, machines). The political animal becomes cosmopolitical. The body is marked not only in the anthropic sense (see above), but also by institutions. The cyborg body is a civilized one, dependent upon technoscience (to keep us healthy and young) and, because of the dominant forms of civilization, ultimately co-determined by the globally unequal distribution of wealth. One can foresee Artificial Life research to play an increasing role in the contest over bodies and biopower as we approach the posthuman condition (cf. Kember 2003). Hybridization and downward causation This tour-de-force through some levels of embodiment makes a note on entanglement and hybridization relevant. The neat linearity implied by the concepts of inclusion and increasing specificity and by the (admittedly idea-of-

13

progress seeming) chain of levels does not hold true unrestricted. For instance, the very possibility of human creation bottom-up of new forms of life seems to suggest some complication (as human purposes may radically inform the natural teleology of what looks like biobodies). Already the culture-determined breeding of new races of cattle, crops, etc. suggests that even though biology should be enough to account for the body of a non-human animal, the human forms of signification interferes with pure biosemiosis, and create partly artificial forms of life like the industrialized pig or weird looking pet dog races. In some deep sense, cows and pigs within industrialized agriculture are already cyborgs, partly machines, partly animals (cf. Haraway 1991). Culture mixes with nature in a downward causation manner, and thus, the hierarchy of levels is tangled (Hofstadter 1979) and natural and cultural bodies hybridize (Latour 1991). We might expect something similarly to apply if we access the status of wet artificial life, as reported by Rasmussen et al 2004 and Szostak et al. 2001. Here, however, we need also to consider not only the biosemiotics of life, but also the special anthroposemiosis of experimental science, and especially the use of models and organisms to study life processes. Models of life Pattee (1989) was emphatic about the distinction between a model of life and a realization of some life process. In the early phase of Artificial Life research, focus was put on the possibility of life in computers, and thus the question of computational simulations vs. realizations was crucial. Considering the possibility of a wet bottom-up synthesis of other forms of life, we need to expand the kind of analysis given by Pattee to include not only the role of computational models in science in general and Artificial Life in particular, but also the very notion of a model in all its variety, and especially the notion of model organisms in biology. It is beyond the scope of this note to make any detailed analysis here, so in this final section only some hints will be given. Let us make a preliminary, almost Borgesian classification of models in biology like the following. Formal models and simulations. Highly relevant for software Artificial Life. Such models are, for their theoretically relevant features, computational and mediumindependent, and thus disembodied, and would hardly qualify as candidates for true or genuine life, from the point of view of organicism or biosemiotics. Semiotically, the map is not the territory; a model is not the real beast. Mechanical and ICT-models. The paradigmatic example here is robots. Robots may provide good clues to study different aspects of animate embodiment, but again, if taken not as models (which they obviously may serve as) but as proclaimed real machine bodies or animats, their ontology is a delicate one. They are build by (often ready-made) pieces of information and communication

14

technology; they may realize a certain kind of machine semiosis (Nth 2003), but their form of embodiment is radically different from real animals (see also Ziemke 2003, Ziemke & Sharkey 2001). Evolutionary models. This label collects a large class of dynamic models not only across the previous two categories (because they may be either computational or mechanical, cf. also the field of evolutionary robotics) but also combining evolutionary methods with real chemistries. Many sessions of previous Artificial Life conferences have been devoted to these models. Model organisms. The standard notion of a model here is to study a phenomenon, say regulation of cancer growth in humans, by investigating the same phenomenon in another but in some senses similar organism like the house mouse. In experimental biology, it has proved highly important to a fruitful research programme to choose the right organism for the right job. Drosophila genetics is a well-known case in point. The lineage or population of model organisms is often deeply changed during the process of adapting it to do its job properly, and it is apt to talk about a peculiar co-evolution of this population and the laboratories using it in research. E.g., Kohler (1994) describes how Drosophilae was introduced and physically redesigned for the use in genetic mapping and sees the lab as a special kind of ecological niche for a new artificial animal with a distinctive natural history. Stripping-down models. A method of investigating the minimal degree of complexity of a living cell by removing more and more genetic material to see how few genes is really needed to keep autopoiesis going (cf. Rasmussen et al. 2004). The problem is, of course, as is well-known from parasitology, that the more simple the organism becomes, the more complex an environment is needed, so by adding more compounds to the environment, you can get along with fewer genes. The organism is always part of an organism-environment relation, which makes any single measure for complexity such as genome size problematic. Bottom-up models. The term bottom-up may be used for all three major areas of Artificial Life, in relation to software, hardware and wet models. Considering only the later here (e.g., Szostac et al. 2001), the crucial question is to distinguish between, on the one hand, a process aimed at by the researches which is truly bottom-up emergent, creating a new autonomous level of processes such as growth and self-reproduction pertaining to biofunctionality and biobodies, and on the other hand, something more similar to engineering a robot from pre-fabricated parts, that is, designing a functioning protocell but under such special conditions that one might question its exemplifying a genuine agent or organism. Just as exciting as they are as examples of advances in wet Artificial Life research, just as perplexing are they as possible candidates for synthetic

15

true organisms, because their process of construction is highly designed by the research team. In this way, they are similar to the model organisms in classical experimental biology, but with the crucial difference that no one doubts the later to be organisms, while it is question begging to proclaim the former to be. The Life-Model entanglement problem A special kind of hybridization is of interest here; the co-evolution of human researchers and a population of model organisms. As hinted at above, also in the case of wet Artificial Life systems, the real life and the model of life gets entangled. This raises questions not only about sorting out, or purifying as Latour (1991) would say, biosemiosis from anthroposemiosis to the extent that this is possible at all, but also considering more in detail the nature of the very entanglement. The hybridicity of human design top-down and natures openended, evolutionary design bottom-up creates a set of complex phenomena that needs further critical study. Conclusion From an organicist perspective, real biological life involves complex part-whole relationships, not only regarding the structured network of organism-organscells relationships, but also regarding the environment-(Umwelt)-organism relations. The biosemiotic trend in organicism is needed to understand natural life (the plants and animals we already know) from a scientific perspective, but is not enough to evaluate the complex question of what is life? as recently raised by synthetic chemistry approaches to wet artificial life. Here, also more ontological, metaphysical, and philosophy of science (and scientific models) inspired considerations are needed. Some of these have been presented, other just hinted at. AcknowledgementsI thank Frederik Stjernfelt, Simo Kppe, Charbel Nio El-Hani, Joo Queiroz, Jesper Hoffmeyer, Mia Trolle Borup and Tom Ziemke for stimulating discussions. The work was supported by the Faculty of Science, University of Copenhagen, and the Danish Research Foundation for the Humanities.

16

References

Barbieri, M. 2003. The Organic Codes. An introduction to semantic biology. Cambridge: Cambridge University Press. Bedau, M., McCaskill, J., Packard, N., Rasmussen, S., Adami, C., Green, D., Ikegami, T., Kaneko, K. and Ray, T. 2000. Open problems in artificial life. Artificial Life 6: 363-376. Buss, L.W. 1987. The Evolution of Individuality. Princeton: Princeton University Press. El-Hani, C. N. and Emmeche, C. 2000. On some theoretical grounds for an organismcentered biology: Property emergence, supervenience, and downward causation. Theory in Biosciences 119 (3/4): 234-275. El-Hani, C. N., Queiroz, J., and Emmeche, C. 2004. A Semiotic Analysis of the Genetic Information System. Semiotica 160 (1/4): 1-68.

Emmeche, C. 1994. The garden in the machine. The emerging science of artificial life. Princeton: Princeton University Press. __1997. Defining life, explaining emergence. Online paper at: http://www.nbi.dk/ ~emmeche/ __2000. Closure, Function, Emergence, Semiosis and Life: The Same Idea? Reflections on the Concrete and the Abstract in Theoretical Biology. In: J. L. R. Chandler & G. Van de Vijver, eds.: Closure: Emergent Organizations and Their Dynamics. Annals of the New York Academy of Sciences 901: 187-197. __2001. Does a robot have an Umwelt? Reflections on the qualitative biosemiotics of Jakob von Uexkll. Semiotica 134 (1/4): 653-693. __2004. Causal processes, semiosis, and consciousness. In: J. Seibt (ed.), Process Theories: Crossdisciplinary Studies in Dynamic Categories. p. 313-336. Dordrecht: Kluwer. Gilbert, S. F. & Sarkar, S. 2000. Embracing complexity: Organicism for the 21st Century. Developmental Dynamics 219: 1-9. Haraway, D. 1991. Simians, Cyborgs and Women: The Reinvention of Nature. New York: Routledge. Hoffmeyer, J. 1996. Signs of Meaning in the Universe. Bloomington: Indiana University Press. Hofstadter, D.R. 1979. Gdel, Escher, Bach: an Eternal Golden Braid. London: The Harvester Press. Hulswit, M. 2002. From Cause to Causation. A Peircean Perspective. Dordrecht: Kluwer. Kauffman, S. 2000. Investigations. Oxford: Oxford University Press. Kember, S. 2003. Cyberfeminism and Artificial Life. London & New York: Routledge. Kohler, R.E. 1994. Lords of the Fly. Drosophilae Genetics and the Experimental Life. Chicago: The university of Chicago Press. Kull, K. (ed.) 2001. Jakob von Uexkll: A paradigm for biology and semiotics. Berlin & New York: Mouton de Gruyter (= Semiotica 127(1/4): 1828). __2000. An introduction to phytosemiotics: Semiotic botany and vegetative sign systems. Sign Systems Studies 28: 326-350. __1999. Biosemiotics in the twentieth century: A view from biology. Semiotica 127(1/4): 385414. Latour, B. 1991. We have never been Modern. New York: Harvester Wheatsheaf. Liszka, J. J. 1996. A general introduction to the semeiotic of Charles Sanders Peirce. Bloomington: Indiana University Press. Mayr, E. 1997. This is Biology. The science of the Living World. Cambridge: Harvard University Press. Nth, W. 2003. Semiotic machines. S.E.E.D. Journal (Semiotics, Evolution, Energy, and Development) 3 (3): 81-99. Pattee, H.H., 1989. Simulations, realizations, and theories of life. In: Artificial Life (Santa Fe Institute Studies in the Sciences of Complexity, Vol. VI), C.G. Langton (ed.). pp. 63-77. Redwood City: Addison-Wesley. Rasmussen, S., Chen, L., Deamer, D., Krakauer, D., Packard, N., Stadler, P., Bedau M. 2004. Transitions from nonliving and living matter. Science 303: 963-965. Reynolds, A., 2002. Peirces Scientific Metaphysics. The Philosophy of Chance, Law, and Evolution. Nashville: Vanderbilt University Press. Rosen, R. 1991. Life Itself. A Comprehensive Inquiry Into the Nature, Origin, and Fabrication of Life. New York: Columbia University Press. Searle, J. 1992. The Rediscovery of the Mind. Cambridge, Mass.: MIT Press. Sheets-Johnstone, M. 1999. The Primacy of Movement. Amsterdam & Philadelphia: John Benjamins.

17

J. Szostak, J., Bartel, D., Luisi, P. 2001. Synthesizing life. Nature 409: 383-390. Star, S.L. & Griesemer J.R. 1989. Institutional ecology, translations, and Boundary Objects: amateurs and professionals in Berkeleys Museum of Vertebrate Zoology, 1907-39. Social Studies of Science 19: 387-420. Stjernfelt, F. 2002. Tractatus Hoffmeyerensis: Biosemiotics expressed in 22 basic hypothesis. Sign Systems Studies 30(1): 337-345. Van de Vijver, G., S. Salthe, S. and Delpos, M. (eds.) 1998. Evolutionary Systems: Biological and Epistemological Perspectives on Selection and Self-Organization. Dordrecht: Kluwer. von Uexkll, T. 1986. Medicine and semiotics. Semiotica 61 (3/4): 201-217. Wicken, J.S. 1987. Evolution, Thermodynamics, and Information. Extending the Darwinian Program. Oxford: Oxford University Press. Ziemke, T. 2003. Whats that thing called embodiment?. In: R. Alterman and D. Kirsh (eds.), Proceedings of the 25th Annual Meeting of the Cognitive Science Society. p. 13051310. Mahwah, NJ: Lawrence Erlbaum. Ziemke, T. and Sharkey, N. E. 2001. A stroll through the worlds of robots and animals: Applying Jakob von Uexklls theory of meaning to adaptive robots and artificial life. Semiotica 134(1-4): 701-746.

18

Semiosis and living membranesJesper Hoffmeyer

External or internal explanation? The idea that the world is such a thing that it can be objectively described so to say in the view from nowhere, as Thomas Nagel has expressed it (Nagel 1986), has nourished the scientific spirit for centuries, and there can be no doubt that this strategy has been enormously successful. Undeniably however this century has witnessed a decline in the general confidence in this idea. The deeper science digs into the material basis of nature the more sophisticated and complex are the problems it must confront. Thus at the most fundamental level we have witnessed a gradual disintegration of the concept of elementary particles into an ever-growing series of very abstract entities threatening our idea of materiality. But even more disquieting perhaps are the subtle problems posed by the observer himself as a human being: how can it be the case that one of the people in the world is me? (Nagel 1986:13). This strange feeling of being a first person singular seems to encompass qualities (sometimes called qualia), which are not describable in a language that cannot go beyond the categories of the 3rd person singular or plural (Searle 1992). The error of confusing 1st person experience and 3rd person experience runs easy in our thinking as was illustrated by a recent newspaper story brought under the heading: Scientists will soon be able to see consciousness. The facts behind this alarming title showed up to be a quote from an expert in mathematical modeling working with brain scanning: I am rather sure that one day we will have a picture on our scanner of the activity patterns constituting consciousness he told the newspaper. But will he? Let us imagine him scanning my brain while I living in the dark winter of Denmark have an experience of longing for summer. Here I personally have no difficulty in believing that this experience might somehow have been evoked by a brain activity that can be visualized on the scanner. For the sake of the argument let us now further assume (though I take it to be not very likely) that one day our expert would be able to tell his colleagues while scanning me, that now Jesper Hoffmeyer has an experience of longing for summer, and also that I did in fact have this experience at exactly this time. Even then of course the expert did not see my longing, for it cannot be seen it can only be felt and only by me. Biology has had a long and painful tradition of resistance towards the mechanistic implications of the objectivistic ontology. The obligation to analyze and explain living systems as if they were purely physical, i.e. non-living, systems has seemed contra-intuitive to many biologists, and the recurrent emergence of vitalistic ideas in ever-new disguises may testify to this. The vestigia terrent

19

20

of vitalism probably also explains the strangely heated atmosphere generated by any scientific criticism of the Darwinian orthodoxy. Darwinian theory, when formulated in a hypothetic-deductive scheme, is in fact a true representation of the view from nowhere in as much as it is not concerned about the material and temporal causative details of the evolutionary process but contents itself to the level of statistical description. The scheme of natural selection is so precious to biologists exactly because it seems to implement the birds eye view solidly in biology or, in other words, to safeguard the position of biology as one of the true sciences. But what about qualia? If one does not believe in miraculous creation then human beings must be the products of evolution and the neo-Darwinian theory is therefore challenged to account for the evolution of qualia. Can a purely mechanical process create qualia? One may of course dismiss such questions by simply claiming qualia to be an illusion, as did philosophers such as Patricia and Paul Churchland and Daniel Dennet (Churchland 1986, 1991, Dennet 1991). Rather than following this eliminitavist strategy (see Searle (1992) for a criticism of eliminitavism) one might suggest that biology, and science in general, reconsiders its reasons for adopting the ontology of the view from nowhere in the first place. As Nagel himself has put it: the fundamental idea behind the objective impulse is that the world is not our world. This idea can be betrayed if we turn objective comprehensibility into a new standard of reality. That is an error because the fact that reality extends beyond what is available to our original perspective does not mean that all of it is available to some transcendent perspective that we can reach from here. But so long as we avoid this error, it is proper to be motivated by the hope of extending our objective understanding to as much of life and the world as we can. (Nagel 1986: 18). Nagel was lead to accept a modern kind of dual aspect theory. Dual aspect theory (also called the dual-attribute theory) was originally forwarded by Spinoza as the view that the unitary substance God is expressed in the distinct modes of the mental and the physical. To talk about a dual aspect theory is largely hand waving writes Nagel It is only to say roughly where the truth might be located, not what it is. If points of view are irreducible features of reality, there is no evident reason why they shouldnt belong to things that also have weight, take up space, and are composed of cells and ultimately of atoms. But: the main question, how anything in the world can have a subjective point of view, remains unanswered. (Nagel 1986: 30). In a deep sense we will perhaps always feel that this question remains unanswered but I do think that the question could be approached in an evolutionary frame if science was persuaded to give up its unfounded belief that reality is in a narrow sense identical to objective reality. A scientifically consistent dual aspect theory might be raised on the basis of what I have called semiotic materialism (Hoffmeyer 1997a). Semiotic materialism claims that our universe

has a built-in tendency (consonant with modern interpretations of the 2. law of thermodynamics) to produce organized systems possessing increasingly more semiotic freedom in the sense that the semiotic aspect of the systems activity becomes more and more autonomous relative to its material basis (Hoffmeyer 1992, 1996, Swenson 1989, Swenson and Turvey 1991). The semiotic dimension of a system is always grounded in the organization of its constituent material components, and cannot exist without this grounding, but evolution has tended to create more and more sophisticated semiotic interactions which were less and less constrained by the laws of the material world from which they were ultimately derived. And this same process finally lead to the creation of selfconscious and intelligent beings, and the religions and cultures they made (or which made them). Thus, rather than seeing human subjectivity or qualia as a unique and utterly unexplainable feature of human existence (modus Searle) or, reversely, as a seducing linguistic illusion (modus Dennet), semiotic materialism sees qualia as a highly evolved instantiation of a semiotic freedom which was latently present in our universe from the beginning and which has been gradually unfolding in the course of organic evolution (Hoffmeyer 1995, 1996, 1998a). Fundamentally the view from nowhere is an externalist way of understanding things. Mechanical dynamics are seen as guided by global (simultaneous) relations, which are held to be prior to their sequential realization. Koichiro Matsuno has shown that this global or externalist way of describing nature does not capture the inner workings of what it describes (Matsuno 1989, 1996; Matsuno and Salthe 1995). The problem is the need for global simultaneity, i.e. the availability of global initial conditions at any arbitrary moment. Global simultaneity cannot be assured a priori for the simple reason that nothing no possible kind of communication or co-ordination propagates beyond the velocity of light. Because of this, global description in a mechanistic framework, however consistent logically, will turn out to be self-contradictory if the objects described are supposed to be located in the material world (Matsuno and Salthe 1995:317). If we want to explain and not only describe material processes we therefore have to apply an internalist view, i.e. giving priority to concrete particulars, seeing things not from without but from within: Local dynamics is the more inclusive perspective because it incorporates something completely lacking in the global dynamics; it is concerned with just how, in the rough and tumble of actual situations, global integration will materialize in detail based on local behaviors, while global dynamics idealistically takes attainment of consistent global integration for granted from the outset (Matsuno and Salthe 1995: 313). Continuous time, duration, and not discrete time is the medium required by local dynamics, which thereby exhibits agency, i.e. the capacity for making

21

contingent choices internally. But this does not preclude the cumulation of the local dynamics into a global record of seemingly simultaneous operations. According to Matsuno and Salthe the necessity for securing the law of the excluded middle is a form of final cause, while locality itself locates an observer. Observation interacts with local dynamics to bring about history, which is absent from global mechanics; any global history must be constituted out of prior local records (Matsuno and Salthe 1995: 333). Surprisingly then the view from nowhere and the mechanical model of the world which nourishes it turns out to be anti-materialistic, while a true materialism gives priority to the fundamentally semiosic nature of material processes. Life: the invention of externalism In the Umwelt theory of Jakob von Uexkll, animals posses internal phenomenal worlds, Umwelts, which they project out into their surroundings as experienced external guiding marks for activity (Uexkll 1982 [1940]). More recently the term Umwelt has acquired a slightly broader interpretation allowing for all kinds of organisms to posses some sort of species specific Umwelt (Anderson et al. 1984). Even bacteria may be said to posses Umwelts in the sense that tens of thousands of receptor protein molecules at their surfaces bind to selected molecules in their environment thus mediating measurements of the outside chemistry to patterns of activity at the inside (figure 1). If the bacterium enters a nutrient gradient it will start moving upstream, and if it enters a gradient of bacterial waste products it will know to move downstream. The bacterium in other words has evolved a capacity to make distinctions based on historically appropriated cyto-molecular habits built into the dynamic macromolecular architecture of the cell and its DNA (Hoffmeyer 1997a). Seen from the human observers point of view the Umwelt of an organism is a kind of world model (Meystel 1998), but seen from the organisms own point of view all there is situated activity, eventually accompanied by a sense of awareness or even anticipation in the case of the most sophisticated animals. Thus, to describe living systems in terms of possessing Umwelts is still part of an externalist discourse even though it is an attempt to deal with the world as seen from the animals point of view. This is because it is only in the historical perspective to which the animal has not itself access that the Umwelt can be described as a model. Retrospectively, and thus externalistically, we can say that evolution has molded the Umwelts of organisms and created species possessing still more sophisticated Umwelts matching still deeper levels of environmental dynamics. There is no way to escape externalism in science, but one should be aware of the limitations this fact poses on science, namely that we can only approach an understanding of historical processes after the fact, i.e. retrospectively.

22

The understanding that biology models the activity of model-building organisms is at the core of biosemiotics of course. Where bacteria are considered the subtleties of the situation stop here because the Umwelt of bacteria is mostly concerned with chemistry. But considering the Umwelts of more sophisticated organisms it becomes clear that these organisms have developed models of their surroundings which are very much aimed at the activity of other organisms and thus of other model-builders producing a semiotic web of infinite complexity. It remains an open question to which extent animals may understand that those other animals, which they model in their Umwelts, do themselves act on the basis of models. As we know even some human beings, e.g. many scientists, have not been too willing to admit that much. We can safely say, however, that the evolutionary road from the most primitive externalist models of the world as possessed by bacteria to the appearance of the first models capable of approaching an internalist perspective

23

Figure 1. Binding of a nutrient molecule to the chemoreceptor blacks CheY phosphorylation allowing for the dissociation of CheY from the switch complex. The switch complex now changes its conformation and induces a counterclockwise movement of the flagellum. Result: the bacterium moves straight forward.

24

has taken billions of years to pass. No wonder then, that internalism is still rejected by many hard scientists. It certainly takes some amount of intellectual and emotional sophistication to enter the unruly inner worlds of otherness. The fact that even prokaryotic organisms like bacteria have Umwelts must influence our understanding of the origin of life. As pointed out by Stanley Salthe neither self-reference nor other-reference can be said to be an exclusive property of life, since even tornadoes may be said to posses a primitive kind of both (Salthe 1998). Yet I would resist the temptation to ascribe Umwelts to tornadoes or to any other pre-biotic systems because none of these systems have yet acquired means for an evolutionarily productive interaction between these two kinds of reference. The self-reference of a tornado is too short-lived and unstable to allow for a true evolutionary process of self-modification in response to other-reference. I shall suggest that it is the stable integration of self-reference and other-reference which establishes the minimum requirement for an Umwelt and thereby sets living systems apart from all their non-living predecessors. It is this double referential or semiotic character of living systems which is the true challenge to theories of the origin of life. And to my knowledge none of the present theories have tried to confront this most central aspect of life: the semiotic core of what it is like to be living. Mathematical modeling of complex systems indicates that the formation of life-like chemical systems may not have been such an unlikely event as was formerly believed (e.g. Monod 1971). In rejecting the magical molecule or RNA approach to the origin of life problem Stuart Kauffman from the Santa Fe Institute has pointed to the remarkable fact that the simplest known free living cells, socalled pleuromona, are already very complex, containing an estimated number of genes of a few hundred to about a thousand. And he suggests that the reason for this might be that a certain minimum complexity is necessary for life to appear (Kauffman 1995). His work with mathematical modeling of combinatorial chemistry has shown that when a large enough number of reactions are catalyzed in a chemical reaction system, a vast web of catalyzed reactions will suddenly crystallize. Such a web, it turns out, is almost certainly autocatalytic almost certainly self-sustaining, alive (Kauffman 1995: 58). Complexity thus is a prerequisite to autocatalytic closure, which again is a prerequisite to life. And Kauffman confidently concludes that The secret of life, the wellspring of reproduction, is not to be found in the beauty of Watson-Crick pairing, but in the achievement of collective catalytic closure. The roots are deeper than the double helix and are based in chemistry itself. So, in another sense, life complex, whole, emergent is simple after all, a natural outgrowth of the world in which we live (Kauffman 1995: 48). While this scheme stands as a convincing alternative to the predominant RNA scenarios for the origin of life, it does not confront the question I have posed as central here: How could a system with the ability to make a productive model

of its external environment appear? What is the origin of the Umwelt? I have dealt with this question elsewhere (Hoffmeyer 1998b) and claimed that what is missing in Kauffmans model is the concept of an asymmetry between inside and outside, and the first precondition for the establishment of such an asymmetry would be the formation of a closed membrane around a complex autocatalytically closed web of interacting molecules (step 2 in figure 2). Also Bruce Weber has emphasized the importance of membrane formation (Weber 1998a and b).ORIGIN OF LIFE Five necessary steps 1. Autocatalytic closure (Kauffman) 2. Inside-outside asymetry (closed surface) 3. Proto-communication (a comunity of surfaces) 4. Digital redescription (code duality) 5. Formation of an interface (inside-outside loops) Figure 2. Five necessary steps on the road to the origin of life. See text for discussion.

The main fabric of the kind of membranes possessed by living systems is a so-called LIPID BILAYER, a continuous sheet, no more than five to six nanometers thick, made of two layers of amphiphilic molecules mostly phospholipids, glycolipids, and sulfolipids joined laterally, as well as tail to tail, by van der Waals interactions between their hydrophobic hydrocarbon chains (figure 3). Christian de Duve characterizes lipid bilayers in the following words: Lipid bilayers are fluid structures of almost limitless flexibility. They behave as twodimensional liquid crystals within which the lipid molecules can move about freely in the plane of the bilayer and reorganize themselves into almost any sort of shape without loss of overall structural coherence. Furthermore, lipid bilayers are self-sealing arrangements that automatically and necessarily organize themselves into closed structures, a property to which they owe their ability to

25

Figure 3. Lypid bilayers. The hydrophilic heads and tails of phospholipids cause them to form bilayers when dispersed in water. Bilayers tend to form closed vesicles.

26

join with each other by fusion and to divide by fission, while always maintaining a closed, vesicular shape. (de Duve 1991). Now, while such closed membrane systems may have formed rather easily in the pre-biotic world, they would probably also die out relatively quickly unless they had acquired a capacity to canalize a selective flow of chemicals (nutrients and waste) across the membrane. But in themselves lipid bilayers are very impenetrable and in cell membranes as known today the transport of chemicals in and out of the cell is only possible because of the inclusion into the membrane itself of thousands of rather complex protein molecules (figure 4). For this reason de Duve in his scenario for the origin of life does not favor membrane encapsulation as an early step. Here however we are concerned with the chemistry of the process only in the trivial sense that what is suggested should conform to the chemical evidence we have, and that evidence is certainly very speculative. Whatever did happen at the chemical level and whatever the order of succession actually was what I claim here only is that from a conceptual point of view membrane formation was THE decisive step in the process. Before membrane closing occurred there could be no inside-outside asymmetry and thus no communication between systems. And my guess is that it is only because pre-biotic systems reciprocally dragged each other into a communicative network that they could muster the creativity needed for the gradual construction of a true cell. The continuous interaction between processes of invention and interpretation, which I have termed SEMETIC INTERACTION, would then have been put in play (Hoffmeyer 1997b). If for instance at some locality conditions allowed for the production of swarms of such closed membrane systems one might eventually obtain a higher level autocatalytic closure so that the outputs from one entity served as inputs to other entities and vice versa. In such a swarm one might say that the closed membrane systems had acquired a germ form of other-reference or protocommunication (step 3 in the figure 2). Still, in my terminology this would not qualify for Umwelt possession since at this stage the system still has no self-referential dynamics. For selfreferential dynamics to occur the system needs what Kauffman has called a written record, i.e. the spatially organized components of the system should somehow become re-described in the digital alphabet of DNA or RNA thereby forming what Claus Emmeche and I have called a code-duality (step 4 in figure 2). Code-duality refers to the idea that organisms and their DNA are both carriers of a message sent down through generations: Living systems form a unity of two coded and interacting messages, the analogically coded message of the organism itself and its re-description in the digital code of DNA. As analog codes, the organisms recognize and interact with each other in the ecological space, giving rise to a horizontal semiotic system, while as digital codes they (after eventual

recombination through meiosis and fertilization in sexually reproducing species) are passively carried forward in time between generations (Hoffmeyer and Emmeche 1991). So far a system has been established which seen from the observers point of view has an obvious interest in maintaining the needed flow of chemicals across the surface. But the system still has no way to assist the fulfillment of its own interest, it has no mechanism for goal-oriented modification or action. Thus the system is not an agent in its own interests. It doesnt matter to the system whether it can distinguish features of its environment or, in other words, it has not yet acquired the capacity for making distinctions. What is needed in addition to the DNA-record is the formation of a feedback link between DNA and environment, so that events outside the system become translated into appropriate events inside the system (Weber et al. 1989). The membrane in other words must turn into an interface linking the interior and the exterior (step 5 in figure 1). Only then does the systems understanding of its environment matter to the system: relevant parts of the environment becomes internalized as an inside exterior, the Umwelt, and in the same time the interior becomes externalized as an outside interior in the form of the semiotic niche, i.e. the diffuse segment of the semiosphere which the lineage has learned to master in order to control organism survival in the semiosphere (Hoffmeyer 1996). I see this as the decisive step in the evolutionary process of attaining true semiotic competence, i.e. the competence to make distinctions in space-time where formerly there were only differences. The semiotic looping of organism and environment into each other through the activity of their interface, the closed membrane, also lies at the root of the strange future-directedness or

27

Figure 4. Model of membrane structure showing the integration of specialized proteins into the lipid bilayer. Membrane bound proteins may perform a variety of functions such as recognition of molecular messages from other cells or transport of specific molecules across the membrane. A hypothetical channel for water and certain ions between several protein subunits is shown.

intentionality of life, its striving towards growth and multiplication. The spatial asymmetry between the inside interior and the outside exterior is coupled to the time asymmetry implicit in the self-referential mechanism of DNA re-description followed by cell division. The landscape of membranes As far as is known cyto-membranes never form de novo by self-assembly of their constituents (as they must nevertheless have done at least once in the distant past); they always grow, in an essentially homomorphic fashion, by accretion, that is, by the insertion of additional constituents into pre-existing membranes. The corresponding patterns are transmitted from generation to generation by way of the cytoplasm (e.g. of egg cells), which contains samples of the different kinds of cyto-membranes found in the organism. The ordinary textbook talk of DNA as governing cellular or even organism behavior is therefore rather misleading. In fact if any entity should be thought of as a governor of cellular activity this should certainly be the membrane. DNA contains the recipes for constructing the one-dimensional amino acid chains, which form the backbones of enzymes, and among them the enzymes needed for catalyzing the formation of the constituents of lipid bilayers and assembling them. But whether these recipes are actually read and executed by cellular effectors depends on membrane bound activity. All major activities of cells are topologically connected to membranes. In the prokaryotes (bacteria) the plasma membrane (the active membrane inside the cell wall) is itself in charge of molecular and ionic transport, biosynthetic translocations (of proteins, glycosides etc.), assembly of lipids, communication (via receptors), electron transport and coupled phosphorylation, photoreduction photophosphorylation, and anchoring of the chromosome (replication) (de Duve 1991: 63). In the more modern and much bigger eukaryotic cells these tasks has been taken over by specific sub-cellular membrane structures of mitochondria, chloroplasts, the nuclear envelope, the golgi apparatus, ribosomes, lysosomes etc. Many if not all of these membranes are themselves descendants from once free-living prokaryotic membranes, which perhaps a billion years ago became integrated into that co-operative or symbiotic complex of prokaryotic membranes which is the eukaryotic cell. Membranes also are the primary organizers of multi-cellular life. The topological specifications necessary for growth and development of a multicellular organism cannot be derived from the DNA for the good reason that the DNA cannot know where in the organism it is located. Such knowledge has to be furnished through the communicative surfaces of the cells. Morphogenesis is mostly a result of local cell-cell interactions in which signaling molecules from one cell affect neighboring cells. Animal cells, for instance, are constantly exploring their environments by means of little cytoplasmic feelers called

28

filopedia (filamentous feet) that extend out from the cell. These cytoplasmic extensions that drive cell movement and exploration are expressions of the dynamic activity of the cytoskeleton with its microfilaments and microtubules that are constantly forming and collapsing (polymerizing and depolymerizing), contracting and expanding under the action of calcium and stress writes Brian Goodwin (Goodwin 1995:136). A developing organism has to generate its own form from a simple initial shape, and this process can be described as highly parallel sequences of bifurcations, i.e. transitions from states of higher symmetry (lower complexity) to states of lower symmetry (higher complexity). A key factor in this process seems to be calcium ion transport across cellular membranes. Calcium ion is bound by special proteins, and it has been shown that the interaction between calcium and the cytoskeleton could result in the spontaneous formation of spatial patterns in the concentration of free calcium and the mechanical state of the cytoplasm, i.e. the kind of bifurcations needed for explaining developmental processes (Goodwin 1995). What emerges from the work of Goodwin and others is an understanding of developmental processes as a formation of relational order arising from a complex pattern of basically semiotic interactions between the constituents of the developing organism. The medium for these interactions is the landscape of membranes at all levels of complexity forming the organism from the mitochondrion to the skin. Dynamic boundaries If the prototype tool is a hammer the prototype organism is probably a dog (in the western world at least). Like ourselves dogs have relatively well defined boundaries, they are mortal individuals and they cannot be at two places at the same time, but have to move in order to get food. Such organisms have been called determinate organisms. Most organisms in the world however are not at all like this. The life of fungi for instance is a constantly changing interplay between dissociation and association generating varied patterns in the interconnected, protoplasm-filled tubes (hyphea) that spread through and absorb sources of nutrients. The hyphea branch away from one another (i.e. dissociate) most prolifically when nutrients are freely available, but re-associate to form such structures as mushrooms when supplies are depleted. Fungi exemplifies what might been called indeterminate organisms, and according to the British biologist Alan Rayner The fungi are an entire kingdom of organisms: their total weight may well exceed the total weight of animals by several times and there are many more species of them than there are of plants! In many natural environments fungi provide the hidden energy-distributing infrastructure like the communication pipelines and cables beneath a city that connects the lives of plants and animals in countless and often surprising ways (Rayner 1997: vii).

29

30

Indeterminate organisms possess expandable or open boundaries that enable them to continue to grow and alter their patterns indefinitely. Such organisms are potentially immortal. Thus, in a Canadian forest one individual rhizomorphic fungus, Armillaria bulbosa, is reported to have produced a network some 1500 ha in area and estimated to weigh 100 tones and to be 1500 years old (Rayner 1997: 130). Only extreme conditions would kill this exemplar. The extreme emphasis on the regenerative aspect of life, proliferation and reproduction, which is the essence of modern gene centered evolutionary thinking does not work well in the world of indeterminate creatures. Here processes of boundary-fusion, boundary-sealing and boundary-redistribution all provides means for reducing dissipation, allowing energy to be maintained within the system rather than lost to the outside. Such processes lead to more persistent organizations in which individuality is blurred. For illustration let us consider a few of the many cases offered in Alan Rayners book (Rayner 1997): Heartrot is intuitively conceived as a disease but in actual fact it is part of a quite normal recycling process, i.e. a process where internal partitioning allows resources to be redistributed from locations that no longer participate in energy gathering or exploration to sites where these processes are being sustained. Heartrot is caused by the degradation by fungi of the predominantly dead wood of mature trees. Heartrot thus results in the hollowing of tree trunks which provides a huge variety of habitats for animals as well as allowing the tree to recycle itself by proliferating roots within its own internal compost of decomposing remains that accumulate within the cavity (Rayner 1997:179). Another illustrative example is the partnership formed by the majority of higher plants with fungi, so-called mychorrhizas. Here the funguses not only provide their plant partners with improved access to mineral nutrients and water in exchange for organic compounds produced by photosynthesis. The mycorrhizal mycelia are also thought to provide communication channels between plants enabling adult plants to nurse seedlings through fungal umbilical cords to reduce competition and to enhance efficient usage and distribution of soil nutrients. There is a risk to this invention,, however, because pirating plants, by tapping into mycorrhizal networks, may indirectly divert resources from the participating plants (Rayner 1997: 63). These are both cases of mutual symbiosis, of course, and symbiosis in general is probably the most underestimated aspect of evolutionary creativity (cf. Sapp 1994). The one-eyed focus on the reproductive aspect of life, the proliferation of successful genes, systematically weeds out the heterogeneous contexts in which organisms always live. These contexts do not only consist in material interactions but also always include subtle semiotic interactions mediated by environmental cues of all kinds. Thus, traditional symbiosis should be seen as just a particular kind of a much more widespread eco-semiotic integration (Hoffmeyer 1997c).

The absurdity of conflating all of this into one simple measure of genetic fitness becomes especially obvious when considering the world of indeterminate organisms where individuality and mortality are only loosely connected, and where the dynamic boundaries in space and time are not defined by their genetic set-up. The evolution of boundaries and the evolution of the contexts in which they put themselves are assisted by, not caused by, genetic inventions. Membrains The distinction between determinate and indeterminate organisms is itself indeterminate of course, no organism is completely determinate or completely indeterminate. Complicated functions such as photosynthesis, capture of prey, ingestion of food and reproduction requires a high degree of adaptive refinement and in order to handle these tasks even indeterminate organisms regularly produce highly prescribed determinate offshoots, e.g. flowers, fruits, leaves, and fungal fruit bodies. While these offshoots often attract the interest of the human observer it is nevertheless the indeterminate part of the system that generates the offshoot and regulates the interrelationships between offshoots by providing interconnected pathways whose variably deformable and penetrable boundaries outline the channels of communication within the system. Determinate superstructure is integrated by indeterminate infrastructure (Rayner 1997: 80, my italics). This pattern can be generalized since even clearly determinate organisms such as dogs or human beings are dependent on indeterminate infrastructure in the form of the vascular system and the nervous system. The pattern of development of nervous and vascular infrastructures resembles that of a foraging mycelium says Rayner: Like the hyphal tubes of a mycelium they may be variably partitioned and cross-linked, and their lateral boundaries are variably insulated so as to receive and distribute input with minimal dissipation (Rayner 1997: 141). There seems to be a deep logic in this arrangement. Nervous systems and brains never developed in plants or fungi. From the beginning these structures were connected to the need of animals to move for purposes of flight or foraging. Nerve cells became specialized for the long distance communication needed in order to co-ordinate the activities of body parts too far apart for quick interaction through traditional cell to cell communication. Thus, determinacy of the body structure, was compensated through the indeterminacy of their movements. As Rayner observes: Animals...follow and create trajectories in their surroundings as they change their position and behavior over time and interact with one another. If these trajectories are mapped, they often exhibit a clear but irregular structure, similar to the body boundaries of indeterminate organisms like plants and fungi (Rayner 1997: 70). The plasticity of the sensorimotor system of determinate organisms, in

31

32

other words, assures the capacity for capitalizing on opportunities which in indeterminate organisms are assured simply through the plasticity of their boundaries. It follows that the brains, which gradually evolved to guide the activity of animals, could not themselves be determinate but would have to retain a plasticity, which could match the unpredictable semiotic heterogeneity of the environment. While morphogenesis in general is guided by local cell to cell interactions neurons can be in direct contact with many cells that are located quite far apart from one another, by virtue of their long output (axons) and input (dendrite) branches. This allows populations of cells distant apart from one another in the brain directly to interact and influence one another so that a non-local developmental logic is superimposed on top of the local regional differentiation that preceded it. And this is the key to the adaptive plasticity of the developing brain since it makes it possible for the nervous system as a whole to participate actively in its own construction (Deacon 1997: 195). Loosely sketched what happens is a kind of neuronal selection process. The developing brain produces a surplus of nerve cells and each of these nerve cells produces far more branches of their growing axons than will finally become functional synaptic connections. Only a fraction of the newly produced connections will happen to get involved in persistent coordinated activity while the remainders are eliminated in a competition between axons from different neurons over the same synaptic targets. Nerve cells, which do not succeed in making any synaptic contributions, are persuaded to commit suicide, so that only some 60% of the totality of nerve cells originally produced will survive into adulthood. That experience from the outer world influences the pattern of neuronal cell death and synaptic strengthening was illustrated by some rather harsh experiments done to newborn kittens. If the right eye of kittens were sewed shut during a critical period of early life they would develop functional blindness on that eye for the rest of their lives. It could be shown that this was not due to lack of experience as such. What happened was that synaptic competition eliminated all those connections to the visual cortex that were derived from the passive right eye leaving all the available synaptic space for left eye connections to capture. The patterns of impulses emitted through the optical nerve from the right eye never reached the visual cortex (Gilbert 1991: 644). The indeterminacy of the brain is its strength. In Terrence Deacons words: Cells in different areas of the brain are not their own masters, and have not been given their connection orders beforehand. They have some crude directional information about the general class of structures that make appropriate targets, but apparently little information about exactly where they should end up in a target structure or group of target structures. In a very literal sense, then, each

developing brain region adapts to the body it finds itself in....There need be no pre-established harmony of brain mutations to match body mutations, because the developing brain can develop a corresponding organization on line, during development (Deacon 1997: 205). For decades neuroscientists have tried to model the brain as a computational organ. Brain cells were seen as logical gates, adding and subtracting input spikes until some threshold level of charge is breached, at which point they convulse to produce a spike of their own. This all-or-nothing nature of a nerve cells firing was thought to overcome the usual soupy sloppiness of cellular processes, i.e. to bring it into an area of digital calculation. But 30 years of research along these simplistic ideas has not been able to give any real answers to how the brain works. And it has now become clear that the output of any individual neuron depends on what the brain happens to be thinking at the time. Not the single firing nerve cell but the self-organizing and ever-changing global pattern of activity all over the brain seem to hold the key to an understanding of its capacity to produce the mental phenomena we all experience. Thus on top of the indeterminacy of the growing brain comes an indeterminacy at the level of synaptic connections, which furthermore makes up for an indeterminacy of global activity pattern formation. And finally on the top of all this plasticity comes yet another level of indeterminacy: Language. Language ran its hyphae far into the nervous system allowing, today, no hope of excision not even in theory. Language does not think through us but it has become a part of us. And yet language is common property and, hence, extraneous to us (Hoffmeyer 1996: 112). Terrence Deacon in his recent book The Symbolic Species suggests that languages have adapted to childrens brains much more than the brains have evolved to become linguistic. This would explain the mystery of how children can learn to talk in spite of the often postulated unlearnability of language: Human children appear preadapted to guess the rules of syntax correctly, precisely because languages evolves so as to embody in their syntax the most frequently guessed patterns. The brain has co-evolved with respect to language, but languages have done most of the adapting. (Deacon 1997: 122). Thus, there is no need for postulating innate linguistic knowledge. With the invention of speech the Umwelts of individual organisms gradually gave way to the idea of the one and only world. The searching membranes of life had at length found the means for a productive non-local association with other membranes thereby creating what the American neurophysiologist Walter Freeman has called societies of brains (Freeman 1995). And this is where the idea of objectivity is rooted, i.e. in the social nature of human knowledge. Our ingenious membrains communicate directly with other membrains attempting to construct the world in the image of the collective, i.e. i