Piotr Sobol-Kołodziejczyk

NATURALIZED EPISTEMOLOGY AND ARTIFICIAL LIFE.
A METHODOLOGICAL PERSPECTIVE

Streszczenie

W artykule został podjęty problem związków zachodzących pomiędzy znaturalizowaną epistemologią a dynamicznie rozwijającymi się badaniami nad sztucznym życiem. Zdaniem autora, związek ten może być rozpatrywany zarówno w metodologicznej i ontologicznej perspektywie. Pierwsza z nich jest równoważna akceptacji mocnej wersji natrualizmu Quine’a i wyraża się w stwierdzeniu, że wszystkie doniosłe problemy filozoficzne są de facto problemami naukowymi, a zatem – są one rozwiązywalne za pomocą metod dostarczanych przez nauki szczegółowe (w głównej mierze nauki przyrodnicze). Natomiast perspektywa ontologiczna przejawia się w założeniu, że realnie istniejące przedmioty i własności to tylko te, które są postulowane w ramach teorii naukowych.

W związku z przyjętymi powyżej założeniami w dalszej części artykułu rozważony zostaje problem ontologicznej różnicy pomiędzy naturalnymi i sztucznymi systemami, którymi przypisuje się własność życia. Analiza definicji sformułowanych w ramach biologii teoretycznej prowadzi do wniosku, że różnica ta w zasadzie nie istnieje, ponieważ obydwa rodzaje systemów są układami ewoluującymi. Mechanizm ewolucji jest natomiast opisywalny za pomocą Neumannowskiej teorii automatów komórkowych. W ostatniej części artykułu podejmuje się zaś problem warunków, które muszą być spełnione, aby możliwa była konstrukcja sztucznych świadomych systemów poznawczych.
Keywords: artificial life, naturalism, epistemology, cognition

1.Introduction

Naturalized epistemology is included in a class of the theories of cognition. There are at least three ways of understanding the notion of naturalized epistemology. First, naturalized epistemology can be formulated as a point of view in the field of analytical philosophy. Naturalized epistemology can be also identified with evolutionary epistemology or the theory of cognition based on the theory of hierarchically regulated systems . More generally, the program of naturalism in epistemology is reduced to the belief that natural science constrains philosophy. According to Huw Price: “The concerns of the two disciplines are not simply disjoint, and science takes the lead where the two overlap. At the very least, then, to be a philosophical naturalist is to believe that philosophy is not simply a different enterprise from science, and that philosophy properly defers to science, where the concerns of the two disciplines coincide” [8].

On the basis of the quoted comment, the idea that natural science can be understood as a confirmation of philosophical thoughts or, vice versa all philosophical theories must be based on scientific findings. This idea should be analyzed from two perspectives: (1) the methodological and (2) ontological approach.

The methodological perspective is convergent with Quine’s strong version of naturalism. This approach leads to a conclusion that all legitimate epistemological questions are, in fact, scientific questions. Therefore, epistemology can be reduced to or replaced by science [11]. For example, the Cartesian problem of structure and content of mind is, to a significant extent, a scientific, not a philosophical issue. It is obvious that this question is in the center of interest of cognitive science, evolutionary biology and psychology or artificial intelligence research. The successes in these areas of research imply that some people believe that all legitimate epistemological problems can be solved by means of scientific methods. For example, Włodzisław Duch writes that if we would like to construct a general philosophical theory of mind, we should first see, what can the neurobiologists and cognitive science researchers tell us about the structure and content of mind[3]. Otherwise, contemporary philosophical theories of mind would be in the same relation to the scientific findings as Ptolemy’s astronomy to the heliocentric system. Then, the methodological approach to naturalized epistemology shows that without taking the scientific results and methods into consideration while explaining the structure of cognition, the results acquired by means of epistemological analysis should be rejected.

The ontological approach to naturalized epistemology is convergent with the belief that a class of real existing beings includes only these types of entities, which are postulated by scientific theories. It means that, if any scientific theory about mind and cognition does not postulate the existence of some type of an object, property or situation, we can assume that this object, property or situation does not exist. For example, if we suppose that the structure of mind possesses a combination of “psychons” [7], and any scientific theory of mind does not postulate the existence of “psychons”, we must be able to reject the idea that “psychons” really exist. If we want to be epistemological naturalists, we should stick to the belief that reality constitutes of and can be described only with the aid of scientific models of reality. Therefore, in naturalized theories of mind and cognition there is no place for such philosophical standpoints like, for example, dualism, interactionism or panpsychism.
Taking the above remarks into consideration, the following speculations aim at demonstrating how can the notion of real existing beings be naturalized with the use of some methods given by artificial life research. It seems that naturalizing is important from a philosophical point of view, because it might contribute to demystifying the opinions that the notion of life is an unsolvable problem in the field of philosophy, represented by, for example, Colin McGinn’s “misterianism” [6]. Naturalizing can also show the usefulness of formal methods of epistemological investigation.

2. Natural and artificial life

Making an attempt at defining the notion of life, one can encounter serious difficulties. In many philosophical papers, it is often stated that the category of life is undefined. For example, Henri Bergson alleged that life is the élan vital and for this reason it can not be described in scientific terms. If this statement was correct, our attempt at naturalizing the notion of life would suffer a defeat. But if we would pay attention to the conclusions formulated in the previous section, a description of the fundamental properties of life should not present any difficulties, because the only valid approach to the phenomena of life is the scientific approach.

Referring to the definition of life formulated by reductionism, it may be said that life is a characteristic feature of systems (or set of elements), which are able to evolve in a biological sense. It is clear that this definition embraces both organic and non-organic systems. In other words, the above definition of life is too wide because it does not accentuate a structural (ontological) difference between the organic and non-organic systems. If we connect the notion of life with the notion of evolution, we should get a more accurate definition. As Douglas J. Futuyama writes: “Biological (or organic) evolution is change in the properties of populations of organisms or groups of such populations, over the course of generations. The development, or ontogeny, of an individual organism is not considered evolution: individual organisms do not evolve. The changes in populations that are considered evolutionary are those that are ‘heritable’ via the genetic material from one generation to the next. Biological evolution may be slight or substantial; it embraces everything from slight changes in the proportions of different forms of a gene within a population, such as the alleles that determine the different human blood types, to the alterations that led from the earliest organisms to dinosaurs, bees, snapdragons, and humans” [4].

According to this definition, the property of life and the ability to evolve concerns solely the organic systems. From the logical point of view this definition can be misleading. Its acceptation may lead to a statement that the low tier organisms (like viruses) do not evolve. Moreover, this definition does not present the difference between the micro- and macro-level of evolution. With reference to this problem, it is rational to show that micro and macro-evolution are quite different phenomena. Elliot Sober argues that the notion of micro-evolution applies only to the changes, that occur on the species level [10]. Whereas, “macro-evolution” means evolution manifesting itself beyond the species level [5]. It is connected with the formation of new structures, which did not exist before.

In this context, it appears that organic systems evolve both in the macro-and micro- sense. The consequences of micro-evolution are usually the changes in the physical structure of some species. For example, people living 3000 years ago had a different body structure than people living today. The consequences of macro-evolution is the formation of new types of organisms through the processes of mutation, inversion or crossing – over.

3.The problem of animated non-organic systems

Now, the possibility of evolution of non-organic systems should be considered. In other words, we should ask what conditions must be met in order to construct a living artificial system. According to Mark Bedau’s definition, “Artificial life can be situated within an interdisciplinary innovation devoted to understanding the behavior of complex systems. Examples of this new venture include the science of chaos and the work spawned by Wolfram’s studies of cellular automata. By abstracting away from the details of chaotic systems (such as ecologies, turbulent fluid flow, and economic markets), chaos science seeks fundamental properties that unify and explain a diverse range of chaotic systems. Similarly, by abstracting away from the details of life-like systems (such as ecologies, immune systems, and autonomously evolving social groups) and synthesizing these processed in artificial media, typically computers, the field of artificial life seeks to understand the essential processes shared by broad classes of life-like systems. Whereas biology’s focus is to understand the central mechanisms of the life-as-we-know-it, artificial life’s interest embraces all of life-as-it-could-be (artificial life shares philosophy’s characteristic concern with broad essences rather than narrow contingencies)” [2].

There are two approaches to artificial life: strong and weak. Referring to the strong version, it may be said that a correctly programmed computer (if it bears the properties characteristic for animated consciousness systems) should be recognized as an animated consciousness system . This way of thinking leads to the acceptance of the statement that life has a logical, not biological character. It is the consequence of acclaiming John von Neumann’s discovery that every instruction realized by cellular automata approximately actualizes the gene’s activity. Then, using formal methods we can construct real living consciousness systems. In the case of the weak version of artificial life, there is no formulated idea about ontological equivalence between organic and non-organic systems. Computer simulations of life are not animated systems, just like the computer simulation of rain is not rain. By making computer simulations of biological processes we can only progress in the cognition of the real nature of these processes.

In the light of the above distinction, the following question may be formulated: what conditions must be met in order to construct conscious animated artificial cognitive systems? First, it must be assumed that he strong version of artificial life is valid, because (under definition) only animated artificial systems can be conscious. Second, we should assume that these systems are embodied. This is necessary, because, as Michael Anderson points out [1]:
1) Cognition, like every other adaptation, has an evolutionary history that can be useful in understanding its function.
2) Cognition evolved because it was adaptive – that is, it enhanced survival and reproductive success primarily by allowing more effective coping with the environment.
3) Cognition evolved in specific environments, and its solutions to survival challenges can be expected to take advantage of the concrete structure or enduring features of those environments.
4) Cognition evolved in organisms with specific physical attributes, bodies of a certain type with given structural features, and can therefore be expected to be shaped by and to take advantage of these features for cognitive ends.
5) Cognition evolved in organisms with pre-existing sets of behavioral possibilities, instincts, habits, needs, purposes, and the like. The evolutionary process would have taken advantage of these possibilities, preserving some and altering others, and incorporating them into its solutions—for instance, taking advantage of certain pre-existing dispositions to manipulate the environment or one’s relation to it, which dispositions may have evolved for reasons unrelated to cognitive enhancement.

Many biological examples may serve as evidence that natural consciousness cognitive systems satisfy the mentioned conditions. These conditions can also be satisfied by an artificial embodied system (like for example Rodney Brooks’ COG) if it:
(A) exists in the same complex, noisy, and cluttered environment which people inhabit. The robot must have sophisticated sensing systems to deal with the complexities of the natural environment without artificial simplifications such as static backgrounds, highly engineered workspaces, or restrictions on the shape or coloring of objects in the world. Furthermore, the robot must also interact safely with people and objects in the environment. The robot’s control systems must be powerful enough to perform routine tasks (such as lifting small objects), but must incorporate multiple levels of safety protocols [9];
(B) is capable to construct non-programmed and non-projected structured. By other words, the process of robot’s evolution must be stochastic. It means that this process should allow the possibility of occurrence an innovation’s elements, could change robot’s evolution in not planned way;
(C) the robot can adapt to the environmental circumstances;
(D) robot’s software must have the property of self-organization and optimalization. Self-organization permits the modifications of connections between the software elements, adaptation the system’s action to the environment and thereby – emersion the emergent attributes in system’s structure.

4. Conclusions

In this paper the theoretical conditions, which must be satisfied for the construction of conscious non-organic systems have been considered. Neither formal methods using in the process of construction AL- systems, nor the existing AL – systems were discussed in the paper, as it was not its main objective. We wanted to describe a situation, in which the border between philosophy (especially – epistemology) and natural science is very hard to determine. It was determined by the assumption that some problems are inter (or multi) disciplinary by its nature. In the process of the construction of AL-systems, this knowledge must be exploited, if we want to postulate the equivalence between natural and artificial systems due to the possibility of making the same cognitive operations. Then, one might say that AL-research confirms the assumptions of naturalized epistemology. Is this statement correct? This is a topic for further discussion.

Piotr Sobol-Kołodziejczyk
Zakład Logiki i Metodologii Nauk, Międzywydziałowy Instytut Filozofii Uniwersytetu Rzeszowskiego al. Rejtana 16C, 35-959-Rzeszów, e-mail: pi.kolo@wp.pl.

References:
[1] Anderson, Michael (2007), How to study the mind: An introduction to embodied cognition, In: Brain Development in Learning Environments: Embodied and Perceptual Advancements, ed. by Flavia Santoianni and Claudia Sabatano, Cambridge Scholar Press, pp. 65-82.
[2] Bedau, Mark (1998), Philosophical Content and Method of Artificial Life, In: The Digital Phoenix: How Computers are Changing Philosophy, ed. by Terrell Bynum and James H. Moor, Basil Blackwell, pp. 135-152.
[3] Duch, Włodzisław (1998), Platonic model of mind as an approximation to neurodynamics, In: Brain-like computing and intelligent information systems, ed. by Sachiko Amari and Nikola Kasabov, Springer, pp. 491-512.
[4] Futuyama, Douglas (1988), Evolutionary biology, Sinauer Associates Inc.
[5] Mayr, Ernst (1991), One Long Argument: Charles Darwin and the Genesis of Modern Evolutionary Thought, Harvard University Press.
[6] McGinn, Collin (1999), The Mysterious Flame: Conscious Minds in a Material World, Basic Books.
[7] Popper, Karl and Eccless, John (1977), The self and its brain, Springer.
[8] Price, Huw (2004)., Naturalism without Representationalism, In: Naturalism in Question, ed. by David Macarthur and., Mario de Caro M., Harvard University Press, pp. 71-88.
[9] Scasselatti, Brian (2001), Foundations of the Theory of Mind for Humanoid Robots, Ph. D. Dissertation, Department of Electrical Engineering and Computer Science MIT.
[10] Sober, Elliot (2000), Philosophy of Biology, Westview Press.
[11] Stich, Steven (1993), Naturalizing Epistemology: Quine, Simon and the Prospects for Pragmatism, In: Philosophy and Cognitive Science, ed. by Chistopher Hookway and Donald Peterson, Royal Institute of Philosophy, Supplement No. 34, Cambridge University Press.

Tekst pierwotnie opublikowany w “Poznańskie Forum Kognitywistyczne. Teksty Pokonferencyjne.” Tom 3, Poznań 2009, s. 155 – 161. Przedruk za zgodą redakcji ze strony: Poznańskie Forum Kognitywistyczne. http://pfk.wikidot.com/nasze-wydawnictwa

——————————————————————————————–
Materiał udostępniany na zasadach licencji

Creative Commons 2.5
Uznanie autorstwa-Użycie niekomercyjne
-Na tych samych warunkach 2.5 Polska
——————————————————————————————–