59. Society as a CompIex Adaptive System WALTER DUCKLEY WE HAVE ARGUED at some length in another place1 istic is its functioning to maintain the given struc- that the mechanical equilibrium model and the ture of the system within pre-established limits. It organismic homeostasis models of society that involves feedback loops with its environment, and have underlain most modern sociological theory possibly information as well as pure energy inter- have outlived their usefulness. A more viable changes, but these are geared principally to self- model, one much more faithful to the kind of regubtion (structure maintenance) rather than adap- system that society is more and more recognized tation (change of system structure). The complex to k,is in process of developing out of, or is in adaptive systems {species, psychological and socio- keeping with, the modern systems perspective cultural systems) are also open and negentropic. (which we use loosely here to refer to general But they are oppn "internally" us well as externally systems research, cybernetics, information and in that the interchanges among their components communication theory, and related fields). Society, may result in significant changr.s in the nature of or the sociocultural system, is not, then, principally the components theinselves with important con- an equilibrium system or a homeostatic system, sequences for the system as a whole. And the but what we shalt simply refer to as a complex energy level that may be mobilized by the system adaptive system. is subject to relatively wide fluctuation. Internal To summarize the argument in ove~lysimplified as well as external interchanges are mediated form: Equilibria1 systems are relatively closed and characteristically by infirmation flows (via chemi- entropic. In going to equilibrium they typically cal, cortical, or cultural encoding and decoding), lose structure and have a minimum of free energy; although pure energy interchange occurs also. they are affected only by external "disturbances" True feedback control loops make possible not and have no infernal or endogeno~q sources of only self-regulation, but self-direction or at least change; their component elemenm are relatively adaptation to a changing environment, such that simple and linked directIy via energy exchange the system may chanre or elaborate its structwc (rather than information interchange); and since as a condition of survival or viability. they are relatively closed they have no feedback We argue, then, that the sociocultural system or other systematic self-regulating or adaptive is fundamentally of the latter type, and requi~s capabilities. The homeostatic system (for example, for analysis a theoretical model or pers-iw the organism, apart from higher cortical function- built on the kinds of characteristics mentioned. In ing) is open and negentropic, maintaining a what follows we draw on many of the con- moderate enerm IeveI within controlled limits. But and principles presented throughout this sou- for our purposes here, the system's main character- book to sketch out aspects of a complex adap~w system model or analytical framework for tbt sociocultural system. It is further argued thar r Many of the ideas expressed here appear in more number of recent socioiogical and social ps- extended form in the author's Sociology and Modern Systems Theory (Englewood Cliffs, N.1.: Prentice- logical theories and theoretical orientations arti- Hall, 1467). culate well with this modern systems perse SOCIETY AS A COMPLEX ADAPTIVE SYSTEM 49 1 and we outline some of these to suegest in addition In these terms, then, the paradigm under- that modcrn systems research is not as remote lying the evolution of more and more complex from the social scientists' interests and endeavors adaptive systems begins with the fact of a poten- as many appear to believe. tially changingenvironment characlerized by variety with constraints, and an existing adaptive system or organization whose persistence and elaboration to higher levels depends upon a successful mapping Complex Adaptive Systems: A Parndi~:m of some of the environmental variety and con- straints into its own organization on at least a A feature of currenr general systems research semi-permanent basis. This means that our is the gradual development of a general paradigm adaptive system-whether on the biological, of the basic mechanisms underlying the evolution psychalogical, or sociocultural level-must mani- of complex adaptive systems. The terminorogy of fest (I) some degee of "plasticity" and "irrita- this paradigm derives particularly from informa- biiify" vis-a-vis i2s environment such that it carries tion theory and cybernetics. We shall review these on a constant interchange with environmental concepts briefly. The environment, however else it events, acting on and reacting to it; (2) some may bc characterized, can be seen at bottom as a source or mechanism for variejy, to act as a set or ensemble of more or less distinguishable potential pool of adaptive variability to meet elements, states, or events, whether the discrimina- the problem of mapping new or more detailed tions are made in terms of spatial or temporal variety apd constraints in a changeable environ- relations, or properties. Such distinguishable ment; (3) a set of selective criteria or mechanisms differences in an ensemble may be most generally against which the "variety pool" may be sifted referred to as "variety." The relatively stable into those variations in the organization or system "causal," spatial and/or temporal relations between that more closely map the environment and those these distinguishable elements or events may be that do not; and (4) an arrangement for preserving generally referred to as "constvaint." If the ele- and/or propagating these "successful" mappings.* ments are so "IooseIy" relatcd that there is equal It should be noted, as suggested above, that probability of any element or state being associated this is a relational perspective, and the question of with any other, we speak of "chaos" or complete "substance" is quite secondary here. (We might randomness, and hence, lack of "constraint." But also note that it is this kind of thinking that gives our more typical natural environment is charac- such great significance to the rapidly developing terized by a relatively high degree of constraint, relational logic that is b~omingmore and more without which the development and elaboration important as' a technical tool of analysis.) Also, of adaptive systems (as well as "science") would as suggested, this formulation corresponds closely not have been possible. When the internal organi- with the current conception of "information" zation of an adaptive system acquires features vieu~ed as the process of selection-from an that permit it to discriminate, act upon, and ensemble of variety--of a subset which, to have respond to aspects of the environmental variety "meaning," must match another subset taken from and its constraints, we might generally say that a similar ensemble. Communication is the process the system has "~napged"parts of the environ- by which this constrained variety is transmitted mental variety and constraints into its organization in one farm or another between such ensembles, as structure and/or "information." Thus, a subset and involves coding and decoding such that the of the ensemble of constrained variety in the original. variety and its constraints remains environment is coded and transmitted in some relatively invariant at the receiving end. If the way via various channels to result in a change in source of the "communicatjon" is the causalIy the structure of the receiving system which is constrained variety of the natural environment, isomorphic in certain respects to the original and the destination is the biological adaptive variety. The system thus becomes selectively system, we refer to the Darwinian process of matched to ~tsenvironment both physiologically natural selection whereby the information en- and psychologically. It should be added that two coded in the chromosomal material (for example or more adaptive system;, as well as an adaptive the DNA) reflects or is a mapping of the environ- system and its natural environment, may be said mental variety, arid makes possible a continuous to tx selectively interrelated by'a mapping process and more or less successful adaptation of the in the same terms. This becomes especially former system to the latter. If the adaptive system important for the evolution of social systems. in question is a (relatively high-level) pqcboIogical 492 SELF-REGULATION AND SELF-DIRECTION IN SOCIOCULTURAL SYSTEMS or cortical system, we refer to "learning," whereby biological processes as genetic information is the significant environmental variety is transmitted gradually aawented by cortically imprinted via sensory and perceptual channels and decodings information, and finally by entirely extrasomatic to the cortical centers where, by selective criteria cultural depositories. The implications of these (for example, "reward" and "punishment") related shifts, and others that could be included, are to physiological andlor other "needs" or "drives," obviously far-reaching. relevant parts of it are encoded and preserved as One point that will require more discussion "experience" for varying periods of time and may may be briefly mentioned here. ahis is the refafive promote adaptation. Or, on the level of the discontinuity we note in the transition from the symbol-based