Varieties of Emergence in Artificial and Natural Systems* Achim Stephan Institut für Philosophie der Universität Karlsruhe (TH), Postfach 6980, D-76128 Karlsruhe at present: Hanse-Wissenschaftskolleg, Fischstr. 31, D-27749 Delmenhorst Z. Naturforsch. 53c, 639-656 (1998); received April 20, 1998 Emergence, Reduction, Connectionism, Qualia, Synergetics In different disciplines such as philosophy of mind, dynamical systems theory, and connec­ tionism the term ‘emergence’ has different jobs to perform. Therefore, various concepts of emergence are developed and examined. While weaker versions are compatible with prop­ erty reductionism, stronger versions are not. Within philosophy of mind, particularly within the qualia debate there is a need for a strong notion of emergence, while in discussions of emergent properties of connectionist nets or of dynamical systems one can do with weaker notions of emergence. Introduction quences of this controversy is a great confusion In addition to having various technical uses, the about what is really meant by an ‘emergent prop­ term ‘emergence’ also has a use in ordinary lan­ erty’, when this term is used in such different disci­ guage. Thus, sometimes people use the expression plines as theories of self-organization, philosophy ‘the emergence of x’ just to mean that x has ap­ of mind, dynamical systems theory, or connec­ peared or that x has come up. The term ‘emer­ tionism. gence’ is used in this way in book titles such as Therefore, the second section of this article is “The Emergence of Symbols” (Bates, 1979) and intended to discuss, in a systematic way, several “The Emergence of Probability” (Hacking, 1975). theories and concepts of emergence of different Of course, one could speak about ‘the emergence strengths. It will be shown that the weaker ver­ of animals with brains’ or about ‘the emergence of sions are compatible with property reductionism. robots’ in this sense of ‘emergence’. However, I In contrast, stronger versions are incompatible will not focus on this ordinary use of ‘emergence’ with property reductionism. Also, the important in what follows. Instead, I will focus on the techni­ distinction between synchronic and diachronic cal uses of ‘emergence’. theories of emergence is developed within this sec­ In most technical uses, ‘emergent’ denotes a sec­ tion. ond order property of certain first order properties In the third section, I examine the different con­ (or structures), namely, the first order properties cepts of emergence, distinguished in section 2 , as that are emergent. However, it is controversial they apply to several natural and artificial systems. what the criteria are by which emergent properties It will become evident that the concepts perform are to be distinguished from non-emergent prop­ very different jobs, and so, one needs to be clear erties. Some criteria are very strong, so that few, if about which concept of emergence one wants to any, properties count as emergent. Other criteria employ. For example, within philosophy of mind, are inflationary in that they count many, if not all, particularly within the qualia debate there is a system properties as emergent. One of the conse- need for a strong notion of emergence, while in discussions of emergent properties of connec- tionist nets or of dynamical systems one can be * This communication is a contribution to the workshop content with weaker notions of emergence. on “Natural Organisms, Artificial Organisms, and ‘Emergent’ is not only attributed to properties Their Brains” at the Zentrum für interdisziplinäre in a strict sense, but also to dispositions, behavior, Forschung (ZiF) in Bielefeld (Germany) on March 8 -1 2 , 1998. and structures. To simplify my presentation, I will use the concept of a property in a wide sense to Reprint requests to Dr. A. Stephan, Nauheimer Str. 39, D-65428 Rüsselsheim. apply to dispositions (e.g., being breakable) and E-mail: [email protected] behavior: Of a system which behaves such and 0939-5075/98/0700-0639 $ 06.00 © 1998 Verlag der Zeitschrift für Naturforschung, Tübingen • www.znaturforsch.com. D Dieses Werk wurde im Jahr 2013 vom Verlag Zeitschrift für Naturforschung This work has been digitalized and published in 2013 by Verlag Zeitschrift in Zusammenarbeit mit der Max-Planck-Gesellschaft zur Förderung der für Naturforschung in cooperation with the Max Planck Society for the Wissenschaften e.V. digitalisiert und unter folgender Lizenz veröffentlicht: Advancement of Science under a Creative Commons Attribution Creative Commons Namensnennung 4.0 Lizenz. 4.0 International License. 640 A. Stephan • Varieties of Emergence in Artificial and Natural Systems such, one can always say it has the property (or the bearers of emergent properties (or structures) disposition) to behave such and such. I will, how­ consist of material parts only. According to the ever, discuss emergent structures separately. As we thesis, all possible candidates for emergent proper­ will see, they are particularly important for the in­ ties, such as, e.g., being alive or being in a mental terpretation of dynamical systems and connec- state, are instantiated only by material systems tionist nets. with a sufficiently complex physical microstruc­ ture. It excludes all vitalistic positions which hold that properties like being alive can be instantiated Weak Emergence, Synchronic Emergence, only by a compound consisting of an organism and and Diachronic Emergence some supernatural entity, e.g. an entelechy or an There are three theories among the different elan vital} Thus, all substance-dualistic positions varieties of emergentism deserving particular in­ are rejected; for they base having cognitive states terest: synchronic emergentism, diachronic em er­ on supernatural bearers such as a res cogitans? gentism, and a weak version of emergentism. For Hence, the thesis of physical monism denies that synchronic emergentism the timeless relationship there are any supernatural components responsi­ between a system’s property and its microstruc­ ble for a system’s having emergent properties. Par­ ture, i.e. the arrangement and the properties of the ticularly, this means that living or cognitive sys­ system’s parts, is in the center of interest. For such tems - whether artificial or natural - consist of a theory, a property of a system is taken to be the same parts as lifeless objects of nature. There emergent, if it is irreducible, i.e., if it is not reduci­ is no reason to suppose that there are some spe­ ble to the arrangement and the properties of the cific components that belong just to those systems system’s parts. In contrast, diachronic emer­ which are alive or able to cognize, but are missing gentism is mainly interested in predictability of in systems which are lifeless or unable to cognize. novel properties. For such a theory, those proper­ Instead, it is nothing but specific constellations of ties are emergent that could not have been pre­ physico-chemical processes that show vital be­ dicted in principle before their first instantiation. havior or have mental qualitites. By the way, these two stronger versions of emer­ gentism are not independent of each other, since (i) Physical monism. Entities existing or coming irreducible properties are eo ipso unpredictable in into being in the universe consist solely of principle before their first appearance. Hence, syn- material parts. Likewise, properties, disposi­ chronically emergent properties are diachronically tions, behaviors, or structures classified as emergent, too, but not vice versa. emergent are instantiated by systems con­ Both stronger versions of emergentism are sisting exclusively of physical parts. based on a common ‘weak’ theory, which at the present pervades emergentist theorizing mainly in Embracing a naturalistic position, emergentists connectionism and theories of self-organization. subscribe to a scientific empiricist position, but in Its three basic features - the thesis of physical mo­ so doing, they do not subscribe to reductionism. nism , the thesis of systemic (or collective) proper­ While the first thesis puts the discussion of ties, and the thesis of synchronic determinism - are emergent properties and structures within the compatible with reductionist approaches without framework of a physicalistic naturalism, the se­ any problems. The stronger versions of emer­ cond thesis delimits the type of properties that are gentism can be developed from weak emergentism by adding further theses. ^Supernatural’ properties are meant to be hyperphysi­ cal, i.e., as independent from (physical) nature and Weak emergentism their laws. 2 In the history of emergentism, however, there were The first feature of contemporary theories of theories of emergence that did not claim the thesis of emergence - the thesis of physical monism - is a physical monism; instead, they took mental or neutral building blocks as fundamental (cf. Broad, 1925, thesis about the nature of systems that have emer­ p. 610-653). Anyway, the thesis of physical monism is gent properties (or structures). The thesis says that not questioned by main stream debate today. A. Stephan • Varieties of Emergence in Artificial and Natural Systems 641 possible candidates for emergents. It is the thesis While the first thesis restricts the type of parts of systemic properties. out of which systems having emergent properties This thesis is based on the assumption that gene­ may be built, and while the second thesis charac­ ral properties of complex systems fall into two dif­ terizes in more detail the type of properties that ferent classes:3 (i) properties which some of the might be emergent, the third thesis specifies the system’s parts also have, and (ii) properties that type of relationship that holds between a system’s none of the system’s parts have. Examples of the micro-structure and its emergent properties as a first class are properties such as being extended relationship of synchronic determination : and having a velocity; sometimes such properties (iii) Synchronic determination. A system’s proper­ are called hereditary properties (however, 'heredi­ ties and dispositions to behave depend nomo- tary' is not used in a biological sense).