LIFE, ARTIFICIAL INTELLIGENCE AND EVOLUTION A DIALOGUE BETWEEN A PHILOSOPHER AND A BIOLOGIST Mikhail Epstein and Eugene Koonin *** The Montréal Review, March 2024 ***
|
|
Science, philosophy, and culture – are these separate magisteria, or can there be beneficial cross-fertilization? This question may be considered almost as old as civilization itself. In antiquity, the boundaries were blurred to the extent that a single individual, such as Aristotle, could make major contributions to biology, logic, and ethics. As centuries passed, though, the demarcation became sharper, and the lack of, at best, limited constructive dialogue between science and humanities came into focus in the second half of the 20th century when science began to grow exponentially, becoming increasingly less transparent to anyone outside the research community. This widening chasm was poignantly and famously captured in the long 1959 essay by Charles P. Snow, “The Two Cultures and the Scientific Revolution.” In the 65 years elapsed, notwithstanding some efforts at bridge construction, the divide has been further widening. In the last few years, however, we have been witnessing and participating in yet another revolution engendered by the coming of age of a powerful class of computer science approaches, deep learning, and artificial intelligence (AI). In a very short time, AI swept through literally all areas of human endeavor, from mathematics to poetry, ushering in equal measure striking new opportunities, challenges, and concerns, and blowing new wind into the old debate on the two cultures. In a sense, AI, although certainly, not being "the theory of everything," prompts us to think in cross–disciplinary terms, on the path to the methodology of "allness." If it is 'everything' that we ultimately pursue through knowledge, the goal cannot be only a mathematical, physical, or biological theory but as comprehensive a theory as its object is. Indeed, the pursuit of a "theory of everything" is not just an endeavor in one specific discipline. It's a quest that inherently demands an all-encompassing methodology, reflective of the vastness and complexity of its subject. AI, in this context, serves as a fascinating tool and a precursor to this broadest scale of understanding. Its capacity to process, analyze, and synthesize information across various disciplines can indeed prepare us for a more integrated way of thinking. What follows is a dialogue between a philosopher, who for years has been deeply interested in the “biologization” of some areas of philosophy and cultural studies (Mikhail Epstein), and a biologist who applies computers to study various aspects of life, but lately, became particularly interested in the generalization of evolutionary concepts based on profound similarities between biological evolution and machine learning (Eugene Koonin). Brought together by a mutual friend, a professor of bioinformatics, Mark Borodovsky at Georgia Institute of Technology, to whom the interlocutors are grateful, Mikhail Epstein and Eugene Koonin engaged in a series of email exchanges trying to clarify and perhaps reconcile their positions on the relationship between the scientific and cultural views of the world. There are major points of convergence: both discussants believe that “Nothing in the world makes sense except in light of evolution” (generalizing on the famous motto of the great geneticist Theodosius Dobzhansky that had “Nothing in biology…”). Significant differences remain, though, with respect to the degree of biologization of culture that is legitimate and desirable. The discussion focuses, in particular, on the applicability of the concept of "learning" to biological systems and the concept of "living" to non–biological systems, such as culture. Above all, Mikhail Epstein and Eugene Koonin wholeheartedly agree that discussing these subjects is timely and every turn of the conversation was fascinating to both. We humbly hope the readers share in this enthusiasm. The dialogue which evolved in the three months from mid–November 2023 to mid–February of 2024 is published as it went, with very minor editing and only a few sentences unrelated to the subject of the discussion being omitted.
I will be happy to discuss with you some important issues on the intersections of philosophy and biology, or “biosophy”, as I sometimes put it, including the possible convergence of artificial intelligence and artificial life.
I have seen some of your work, and although I cannot claim having read it carefully, I am impressed by the breadth of your interests and the reach of your thought. I will be happy to discuss the issues you mentioned, with the caveat that while my understanding of “artificial life” is more or less professional, the same cannot be really said about my knowledge of AI. But, certainly, I have some expertise and many thoughts on evolution and will be glad to discuss.
1. Mikhail Epstein: Thank you for your willingness to engage in a discussion. In preparation, I have begun delving into your book "The Logic of Chance," with a special keen interest in its philosophical epilogues. I wonder if you have since published any articles that encapsulate your updated perspective on these matters. Bridging humanistic and scientific thought is a crucial yet mostly overlooked connection on the path to the “Theory of Everything.” Several questions that I would like to raise: Could AI ever fully replicate or outpace the complexity of biological life, which seems sluggish and outdated compared to the rapid advancements in technology? Do AI and AL (Artificial Life), along with synthetic biology, portend a novel evolutionary trajectory for humanity? Such prospects have been recently explored in M. Suleyman’s book "The Coming Wave" (September 2023). What do the advancements in synthetic biology and AI imply for our comprehension of intelligence and consciousness, and in what ways might they challenge our concepts of identity and selfhood? Is it possible for AL and AI to forge evolutionary paths independent of biological processes, and what could this signify for the future of life on Earth? What would be your take on R. C. Lewontin’s assertion:
If Richard Dawkins's concept of the 'selfish gene' and his theory of memes are accurate, and we view organisms as mere vessels for genes, and human brains as carriers for memes, wouldn't AI represent the pinnacle of evolutionary progress? Information systems could evolve at an unprecedented pace, surpassing that of the gradual nature of biological evolution. In this context, is "Synthetic," or "Artificial" Life simply an intermediary phase from organic life to AI, which offers a more proficient medium for the propagation of bits and memes? Alternatively, as I prefer to consider, the 'noofication' of life,—as manifested in the human mind and the creation of AI—might be mirrored by a 'biofication' of AI systems in their future evolution. Biofication means the endowment of non–biological systems, like technology, with biological capacities; noofication, the endowment of physical or biological systems with intellectual capacities. Through my dialogues with systems like GPT-4, I get the impression that AI is open to the possibility of integrating more deeply with biological processes and with emotional life. Looking forward to a stimulating exchange of ideas. Eugene Koonin: You ask a variety of very interesting questions! On some of these, I am in a good position to comment and perhaps provide some explanations, on others, I am out of my depth. The “Logic of Chance” is, I am afraid, seriously outdated. I do not find major errors there but many new ideas and results have appeared since then. I am trying to work on a new version (I cannot really call it a second edition, it will be a new book) but my progress is very slow because I run a large lab and have to work on numerous papers. Meanwhile, I will recommend the following papers that address some of the pertinent issues: Toward a theory of evolution as multilevel learning and Physical foundations of biological complexity. Let me address your questions as best I can – which might not be really satisfying. Caveat emptor for all questions on AI: I am not really an expert. I have used AI considerably in my research and generally seem to understand how it works but I am by no means a professional in AI and/or machine learning. Could AI ever fully replicate or outpace the complexity of biological life, which seems sluggish and outdated compared to the rapid advancements in technology? I can see no obvious limits to the complexity that is achievable by AI. The depth of learning (number of levels) can be increased indefinitely, so that multilevel architectures could indeed evolve enormous complexity reasonably fast, obviously on a much shorter time scale compared to biological evolution. Certainly, it is important that all evolution of AI occurs within computer networks, and its implementation in physical objects is a different matter. Do AI and AL, along with synthetic biology, synthetic biology portend a novel evolutionary trajectory for humanity? Such prospects have been recently explored in M. Suleyman’s book "The Coming Wave." I have not read the book, but yes, I think a new evolutionary trajectory for humanity is easy to imagine. I probably should note here that ‘synthetic biology’ might not be a particularly precise definition. Progress in synthetic biology sensu stricto, synthesizing organisms “from scratch” is quite limited. However, if instead of ‘synthetic biology’, we consider genome engineering, the achievements are already impressive and the potential is enormous. Obviously, speaking of new evolutionary trajectories for humanity, we must add “if civilization is not obliterated or derailed by climate change and other disasters”. What do the advancements in synthetic biology and AI imply for our comprehension of intelligence and consciousness, and in what ways might they challenge our concepts of identity and selfhood? I am really NOT an expert on this. My not uneducated view is that intelligence, consciousness, identity and selfhood are all emergent properties of neural networks, nothing beyond that. Accordingly, they all can and will emerge in the course of evolution of AI. Surely, that is an interesting prospect. Is it possible for AL and AI to forge evolutionary paths independent of biological processes, and what could this signify for the future of life on Earth? AL is Artificial Life, right? At least from what I know, this direction is not progressing that dramatically and certainly has limited potential to create routes of evolution independent of biological ones; all the genome engineering does remains within the domain of biology. Apart from that, there is a direction in computational evolutionary biology known as Artificial Life that models evolutionary processes, in some cases, revealing interesting, previously overlooked evolutionary phenomena. AI is a different matter. Here, there is potential for quite independent evolution, and it is already being realized. I think it can go quite far, changing the face of human civilization. What are the consequences for life on earth… hard to foretell… likely, devastating. What would be your take on R. C. Lewontin’s assertion:
What Lewontin says here, is rather obviously valid. Indeed, the geology and geochemistry we are familiar with, including atmosphere, soil, oceans, have been shaped by life during the 4 billion years of its evolution. If Richard Dawkins's concept of the 'selfish gene' and his theory of memes are accurate, and we view organisms as mere vessels for genes, and human brains as carriers for memes, wouldn't AI represent the pinnacle of evolutionary progress? Information systems could evolve at an unprecedented pace, surpassing the gradual nature of biological evolution. In this context, is AL simply an intermediary phase from organic life to AI, which offers a more proficient medium for the propagation of bits and memes? Alternatively, as I prefer to consider, the 'noofication' of life—as manifested in the human mind and the creation of AI—might be mirrored by a 'biofication' of AI systems in their future evolution. Through my dialogues with systems like GPT-4, I get the impression that AI is open to the possibility of integrating more deeply with biological processes and with emotional life. I think that, understood in this simplistic way (I do not think this is what Dawkins actually meant), considering organisms as mere vessels for genes is wrong and counterproductive. Not only organisms but populations and communities are important as coherent units of selection. Evolution is inherently multilevel and cannot be reduced to the level of genes. And then, again, I would not at all mix or try to merge AI and AL. ‘Biofication’ of AI, though, sounds like an interesting idea. 2. Mikhail Epstein Thank you so much for your clarification of issues that are very important for me. It took me some time to read the two papers: on evolution as multilevel learning and on physical foundations of biological complexity. What I appreciate most about your approach is the examination of various levels of materiality in both biological and non-biological systems, which are governed by digital (or semiotic, as "universalists" of the previous era might assert) laws. Your assertion that SOC (self-organized criticality) inherently implies self-similarity across different organizational levels resonates with me. I find myself aligned with your perspective that there seem to be no discernible limits to the complexity achievable by AI and that the depth of learning and the number of levels can be expanded indefinitely, enabling multilevel architectures to evolve substantial complexity at a rapid pace—undeniably much faster than biological evolution. You write: "Separation of the relevant degrees of freedom into multiple temporal classes is ubiquitous in the universe from composite subatomic particles, such as protons, to atoms, molecules, life-forms, planetary systems, and galaxy clusters". Obviously, culture, intelligence, and creativity (including science) present the highest "degrees of freedom" which prompts us to add them into this list: life-forms, planetary systems, galaxy clusters… and languages, sciences, technologies, cultures, civilizations, etc. Furthermore, you state, "The framework of learning theory employed here seamlessly unifies the descriptions of physical and biological phenomena in terms of optimization through trial and error and loss (fitness) functions." Indeed, a pivotal aspect of this approach is that our general principles are applicable to both living and nonliving systems. Might we venture to extend this to "both physical and cultural" realms? My interest lies in introducing into this digital or semiotic interplay those transphysical and transbiological systems, such as culture, intelligence, consciousness, and creativity. These domains possess their own levels of complexity and autonomy, which are not reducible to the parameters of physical or biological systems. In my view, AI outperforms human intelligence even in the traditionally humanistic domains, like philosophical imagination. I have undertaken several experiments, not strictly scientific, comparing natural and artificial intelligence in executing creative tasks. My preliminary observations suggest AI's superiority over human capability, even at this nascent stage of AI development. Here are references to some of my interviews and articles on AI, describing these experiments and my views on the promises and mysteries of AI: AI as a New Muse: Exploring Artificial Intelligence’s Integrative Creativity and Логика Искусственного Интеллекта в Сравнении с Естественным. As for our common interests, I would like to share some of my earlier musings on the broader, transbiological essence of life. Biology essentially considers life as an organism's active self-development, encompassing energy and information exchange with the environment. However, life exists beyond nature as such, encompassing thoughts, works, theories, arts, human characters, and relationships. I believe this broader understanding necessitates biosophy, a discipline studying life in all manifestations, both material and mental (or spiritual). Biosophy explores the dynamics of life in human experiences and actions, questioning what distinguishes the living works or ideas from the dead ones. Life, in its essence, is self-creating (autopoiesis), capable of internal growth, organization, and environmental interaction. It resists entropy, embodying energy and striving for growth and expansion. Life extends beyond biopolymers to semiotic, textual, and symbolic forms, sharing properties like self-organization, growth and reproduction with organisms. Thus, alongside prokaryotes and eukaryotes, we might consider 'noocaryotes'—entities of psychic energies and intelligent life, ideas and texts that multiply and develop in human history. Eugene Koonin: Thank you for this detailed and certainly very interesting response. I fully agree that principles of evolution (and learning) are applicable and must be applied in the study of the history of civilization and culture. I would go further and suggest that attempts to understand and advance those areas without turning to evolutionary thinking are useless and hopeless. I would generalize Dobzhansky’s famous motto in this simple (and outrageous to some) manner:
Indeed, the same principles of evolution govern ‘everything’, from plasma to culture even if their concrete manifestations are enormously diverse. Just in case, I suppose you are familiar with the publications of our compatriot Peter Turchin on Cliodynamics. I find this approach refreshing and promising. Regarding your general definition of life. Such definitions hardly can be true in the ultimate sense, only useful and constructive… or not. In that vein, I think that designating text, works of culture, memes, etc., as ‘living’ is not really useful. I believe it is best to reserve the terms "life," living"," etc. for systems that transform chemical energy. However, those entities belong to a much broader class of evolving systems. Actually, any system with a less than maximum entropy – that is, any system that contains information – or any system away from equilibrium (all synonymous) – evolves, and evolution is arguably the one truly universal process in the universe. I am actually unsure if biosemiotics and even the concept of autopoiesis have ever yielded anything constructive for biology. I am not saying they haven’t, just that I cannot pinpoint any. Possibly, I just need to read and think more. Perhaps, I should note that I am not at all keen on any type of ‘spirituality’. In my experience, whenever this is brought up, the conversation devolves into mysticism, and to me, loses any interest. This is quite different from the problem of consciousness which is certainly important and interesting. However, any ‘non-physical’ interpretations are, again, of no interest to me. 3. Mikhail Epstein: Your response raises an intriguing question about the extension of concepts from one domain to another. You suggest that the terms “living” or “alive” cannot be usefully applied outside of biology: “Designating texts, works of culture, memes, etc., as ‘living’ is not really useful.” Why, then, should we believe that applying the term “learning” to biological entities and even to the Universe as a whole can be seen as a key to explaining all physical and biological phenomena? You write: “The framework of learning theory used here allows us to naturally unify the descriptions of physical and biological phenomena in terms of optimization through trial and error and loss (fitness) functions.” “Key features of the processes of learning and biological evolution imply that this is not a simple analogy but rather a reflection of the deep unity of evolutionary processes occurring in the Universe.” Properly speaking, what is learning? “The acquisition of knowledge or skills through experience, study, or being taught” (Oxford Dictionary). Thus, the article by you and your coauthors “Toward a Theory of Evolution as Multilevel Learning” assumes that the Universe, and organic nature in particular, are based on “the acquisition of knowledge or skills…” Doesn’t this sound somewhat metaphorical and anthropomorphic? This is presented not as a metaphor or even an analogy but as a rigorous scientific concept and the principal thesis. (In your words, “This is not a simple analogy but rather a reflection of the deep unity of evolutionary processes occurring in the Universe.”) Even in Osip Mandelstam's «Грифельная ода» (Slate Ode), the metaphor of learning applied to natural processes (“ученичество миров”) appears to be an overly bold metaphor for poetry. Yet here, it is used as the cornerstone of the unity of physical and biological knowledge, a “theory of everything.” Why then are “living” and “alive,” words that are by no means perceived as unusual or excessively daring in descriptions of certain societies, personalities, cultural phenomena, and creative accomplishments, “not really useful” when applied outside biological phenomena? My point is not to discredit the application of “learning” to cells, planets, galaxies, etc. I am ready to agree that one galaxy learns from another, but why is it a one-way street? Why is it possible to anthropomorphize nature and the natural sciences but not to “biologize” the humanities? I believe that the unity of knowledge can be achieved, or rather sought, in both directions. Here are, in my view, some points of convergence of “living” in the bio– and nonbio– domains:
Thus, although at first glance the use of the word "alive" in these contexts may seem merely a stylistic device, it also reflects the deep connection between different forms of life and living processes, whether in biology, culture, or social life. This connection can be a source of rich analogies that help us understand and interpret various phenomena more deeply. I hope our conversation will prove useful for those interested in exploring the unity of knowledge. Eugene Koonin: Thank you for sharing these interesting and pertinent thoughts. Again, definitions, extension of terms, etc., cannot be objectively right or wrong, only coherent, illuminating and useful – or otherwise. The case of ‘evolution as learning’, though, as I see it, is a different story. Here, what we do, is extending a common theoretical framework to the entire gamut of temporal processes occurring in the universe – in that sense, this is indeed an outline of “a theory of everything”. The essence of this theory is interaction of evolving systems with their environment resulting in extraction of information on the latter and the consequent adaptation leading to enhanced persistence and/or growth. This is learning in the most general, high-level sense. There is nothing at all anthropomorphic here, no implication of “universal consciousness”, of any intentionality, or anything of that sort, whatsoever. If memory serves, in the initial draft of that paper, we had comments to that effect, but had to drop them due to space constraints. As your comments emphasize, I think in some further writing, we ought to make this abundantly clear. As for extending the term "life" to various kinds of social and cultural phenomena, I think this imputes on such phenomena certain meanings that are irrelevant to them. Again, physical compartmentalization, energy transformation, and metabolism are features that are intrinsic to life no less than information processing. The social and cultural phenomena lack that physical side, hence I consider the transfer of terms to be counterproductive. These processes do possess certain life-like features as you point out. From my viewpoint, such features come from the aforementioned general framework of dynamic, evolving, non-equilibrium systems in which category those phenomena definitely belong. 4. Mikhail Epstein: Thank you for clarifying your perspective on learning as a mechanism of evolution. I understand that you're not suggesting the concept of a “universal consciousness” or anything similar. However, the concept of learning inherently presupposes two subjects: someone who learns, and someone from whom knowledge is acquired. If the universe is undergoing self-learning as part of its evolutionary process, that idea resonates with me. You write: “The essence of this theory is the interaction of evolving systems with their environment, leading to the extraction of information about the latter and subsequent adaptation, which results in enhanced persistence and/or growth. This is learning in the most general, high-level sense.” This sounds reasonable. But why are terms such as “adaptation, interaction, persistence, growth, evolution, organization, and transformation” considered not sufficiently adequate for describing this process, while your theory emphasizes learning, implying information exchange between subjects? As a philologist, I believe metaphors are never “innocent,” and the usage of this one suggests a teleological process of information transfer. According to the dictionary, to learn means “to gain or acquire knowledge of or skill in (something) by study, experience, or being taught”. Does this imply that the universe or its various components acquire knowledge or skills from itself or each other during evolution? If no anthropomorphic meaning is intended, why use the term “learning,” which is inherently anthropomorphic? It's important to recognize and critically examine these metaphorical frameworks, as they shape our understanding and interpretation of scientific phenomena. Acknowledging this doesn't diminish the scientific validity of the concept but rather enriches our comprehension of how human perspectives influence scientific discourse. Thank you for your time and effort in elucidating some biological concepts that are crucial for my understanding of the world within the "theory of everything." If this theory aims to encompass “everything,” it should also integrate humanistic knowledge. By no means am I opposed to anthropomorphic metaphors. In fact, I see them as a manifestation of the complementary relationship between the physical universe and consciousness, as particularly expressed in the anthropic principle of cosmology. I wonder why even the boldest and most independent scientific minds, including proponents of the anthropic principle (in both its strong and weak forms), often hesitate to acknowledge that their descriptions of the physical or biological world also involve an anthropomorphic vision through the prism of consciousness and subjectivity (like the "learning" process). I will refer to John Wheeler, one of the founders of the strong anthropic principle who coined the "it from bit" doctrine: all things physical are informational in origin. This justifies the "learning" concept of evolution, but also has much broader implications for the role of consciousness and subjectivity in the universe. According to Wheeler, the universe's existence necessitates observation, implying one or more observers with consciousness (mere mechanical tools of registration are not sufficient). Recognizing that the universe can generate consciousness and needs to be perceived by it, shouldn't we also accept that scientific language can and, in a sense, should be anthropomorphic? Although we originate from different “tribes” and domains of discourse, I still hope that by acknowledging the complementarity of (living) matter and consciousness, we can advance towards a more integrated, cross-disciplinary view of “everything.” Eugene Koonin: I am no philologist, not by any account, but I tend to agree with you that metaphors and more broadly, connotations of terms matter. In that regard, it was not without some hesitation that my colleagues and I used the term ‘learning’ in the articles we discuss. However, at the end of the day, I still believe it was a good decision. Metaphors can be helpful, even impactful in science, drawing non-trivial but relevant connections between different fields, or else, can be misleading, leading in false directions. In the case of ‘learning’, the incentive to use this metaphor to describe evolution is the deep analogy, down to the mathematical isomorphism, between the theory of machine learning and the theory of adaptive evolution (and both these theories show striking parallels to thermodynamics). In the theory of machine learning, the term and concept are already cleared of the anthropomorphic, teleological connotations, the AI theorists have done that work for us. So I believe that scientific concepts and the language used to present them should not be anthropomorphic, this is a fallacy, from my perspective. Furthermore, I do not think that consciousness is a fundamental phenomenon that is, in any sense, complementary to the physical universe. Neither are conscious observers or consciousness in any form or guise necessary for the existence and evolution of the universe. I strongly doubt Wheeler meant this literally. This whole observer problem in physics is quite subtle, though. I certainly do not claim to understand it in its entirety, but one aspect that we highlight in that very article on evolution as learning is that, to be observable (explorable, amenable to study, whatever), the universe needs to meet certain criteria that are also conducive to evolution of complexity including life. This is in a sense a manifestation of the weak anthropic principle (certainly not the strong one!) that, however, has nothing to do with a “requirement” for a conscious observer. Speaking of the anthropic principle, I cannot call myself a proponent (some years ago, I would – in that respect, my views have evolved), in the sense, that as scientists, we must always seek non-anthropic explanations of any phenomenon, turning to weak anthropic principle only as the explanation of last resort. But I do think it has its place in the scientific worldview. As for the strong version, to emphasize once again, to me, this is beyond reasonable discussion. Consciousness… I am certainly no expert, and I do recognize that this is a major issue, particularly difficult due to its inherently self-referential nature. Yet, I have no doubts that consciousness, or perhaps more precisely, self-awareness is an emergent property of complex brains. I think using this concept and term in a broader context is misleading. Finally, about ‘theories of everything’. Scientists normally do not use this phrase in earnest, more in a tongue-in-cheek manner. The best we can strive for is a maximally general theoretical framework within which diverse phenomena can be analyzed. Usually, these general theories are of no immediate use for specific research, even though it is good to be aware of the foundations of your field. As I think rather elegantly put in the following paper by very prominent physicists (which I recommend). I think I have presented my position coherently and, I hope, clearly. I am happy to continue the dialogue if there are additional points to discuss, or halt it if not, or shift to other, perhaps, more specific subjects. 5. Mikhail Epstein: It seems that our correspondence indeed is now reaching a conclusion, at least for this first round. Our positions have been articulated quite clearly, both in substantive and methodological terms. I must acknowledge, albeit with some disappointment, that my ideas regarding life beyond organic nature – concepts such as transbiology, or biosophy, which I detailed in the Russian file (comprising articles from my 2017 book 'A Projective Dictionary of Humanities') and included in a previous correspondence dated December 4, 2023 – did not elicit a positive response from you. In conclusion, I would like to remind that my view is based on a well–established philosophical tradition. For example, Henri Bergson in his Creative Evolution extends his concept of the "élan vital" beyond biological organisms to encompass human culture and consciousness. He sees culture, artistic creation, and thinking as manifestations of this vital impulse, as parts of the evolutionary process driven by the "life force" that you prefer to view as evolution through learning. Incidentally, these concepts resonate, albeit only partially, with "artificial life," which I mentioned in my first letter. "Artificial life," according to its founder Christopher Langton, is the functionality of the living outside of biological substance, for example, in computer simulations. Here is his definition:
What I proposed, in my view, was more realistic than Langton's projections, as it is not about computer simulations of life, but the manifestation of life in culture: the property of vitality in such cultural subsystems as language, literature, art, philosophy, society, etc.. Just to summarize my argument: I am talking about living thoughts, living communication, living speech, living characters, and so on. This is not "artificial life" in Langton's sense, nor "synthetic biology" or genome engineering dealing with living nature and its possible transformations in bioengineering – possibilities you find valid (as do I). Yet what I mean is "trans–organic life," the living beyond organisms. From such a perspective, nothing particularly new needs to be invented, since the category of life and living is routinely applied to many phenomena of cultural life, including the very concept of "life" that I just used in conjunction with culture. "Literary life," "life of an artistic work," "life of an image," "living legacy," etc. The problem is to find common grounds between this humanistic, so far rather metaphorical understanding of life – and the approach established by modern evolutionary biology. Can we imagine that the living in culture and the living in nature possess objectively common properties/functions? When Boris Pasternak pronounces as a motto: "to be alive, alive only, alive and only until the end," he expresses not only a good wish to people, but also a certain vector of the evolution of the universe: from the non-living to the living in nature, and further to the living in culture (where, as in nature, there is also a lot of dead, inorganic stuff). If you believe that the basis of the living is learning, then why not consider that living, as it manifests itself in culture, in human learning, can be the focus of biosophic, humanistic interest, as one of the most important forms and constants of the evolution. Not only life as evolution is learning, but also learning itself is one of the forms of the evolution of life. If a transfer of meaning in one direction, from culture to nature, from pedagogy to biology, is relevant, then why not in the other? Interestingly, in your latest rejoinder you explain the use of the concept of learning in biology, in your theory of evolution, by its transfer from the theory of machines. You write: "In the theory of machine learning, the term and concept are already cleared of the anthropomorphic, teleological connotations, the AI theorists have done that work for us." The theory of machines was defined by Norbert Wiener as "cybernetics," concerned with "control and communication in the animal and the machine." In cybernetics, the machine is conceptualized as analogous to an animal; thus, why indeed not make the reversal, to apply what has been achieved in the theory of machines to the science of nature, of animals, as you suggest? Cybernetics learns from biology so that biology later could learn from cybernetics. This is quite a feasible transfer of knowledge from one discipline to another, even as far distanced as cybernetics and biology are. But there is a field that in fact mediates between nature and technology, and is closer to both of them. This is culture, which bridges the natural and artificial. Why then would applying both biological and technological concepts to culture be deemed illegitimate? If you apply technology to biology, as Wiener applied biology to technology (in his cybernetics), why to discredit the transfer from both of them to the intermediary field of cultural studies, of the humanities? Again, this field, also as a resource for methodological bridges and mediations, is much closer to both biology and technology than the latter two to each other? Humans are natural beings who through cultural evolution create technology; these spheres are mediated exactly by culture. Culture is "natural artificiality" or "artificial naturalness," at the transition from the natural, pre-cultural to the supernatural, technically artificial. Culture, by its ontological status, is commensurate to the human as such, as a naturally artificial being, generated by nature and generating culture. I will be glad to receive your response to these musings, but you may consider it optional. In any case, we can assume that all positions have been stated fully enough, and at this point we can put a period so that the volume of correspondence does not exceed the attention span of our potential readers. Eugene Koonin: Thank you for the response. I also thought that our exchange (yes, the first round) was coming to conclusion, but now I think your comments call for a reply because I want to make it clear and explicit what I accept in your thoughts and what I don’t. I actually wholeheartedly agree that applying ideas, concepts and approaches of evolutionary biology to explore the evolution of society and culture (fields in which I am, obviously, no expert) is not only possible but could be essential to achieve progress in understanding these phenomena (in that regard, are you familiar with Cliodynamics-- in my opinion, a serious and promising, even if not fully mature, attempt to apply quantitative methods in historical study?) I also highly appreciate the study of artificial life and digital organisms, an interesting field that helped uncover some previously overlooked but important features of biological evolution itself, for example, in this classic paper: Evolution of Digital organisms at High Mutation Rates Leads to Survival of the Flattest. Speaking more generally, the same or closely similar laws govern evolution in natural and artificial systems, not limited to biology. What I can never accept, though, is any flavor of “"élan vital”, any kind of (quasi)mystical “life force”. Any invocation of these inescapably leads into the domain of pseudoscience. Actually, to be frank, I have a bit of an allergy to the very name of Bergson (strangely enough, I have read “Creative Evolution” – or rather, looked through because I found it unreadable). By the way, if you have not read Jacques Monod’s great book “Le Hasard et la Necessite” where he presents a brief but scathing critique of Bergson, I strongly recommend it. I hope these comments clarify my position, and should that be the case, perhaps, they indeed conclude (this stage of) our discussion. ***
*** |
|
||
The Montréal Review © All rights reserved. ISSN 1920-2911
