``Life, beginning as self-replicating molecules, did not persist in this initial state. Rather, it subsequently elaborated `vehicles' in which the original heritable units became increasingly distanced from direct interaction with the external environment'' [Buss 87] (p.183).
One succession that has been proposed begins with self-replicating molecules as the first stage of the evolution of life (but see Section 2.2). These molecules later grouped together into larger collections of self-replicators, which became enclosed within a membrane. From these `protocells', prokaryotic cells evolved, and collections of prokaryotes subsequently became associated with each other to form eukaryotes. From simple unicellular eukaryotes, multicellular organisms evolved, and single multicellular organisms then grouped together in various social patterns. Such a progression has been proposed by Maynard Smith and Szathmáry ([Szathmáry & Maynard Smith 95], [Maynard Smith & Szathmáry 95]). There is still some concern over the consistency of the criteria used to define each of the transitions (e.g. [McShea 96]), but the majority involve an increase in hierarchical object complexity at least.
In Buss' treatment of the subject, he says ``At Darwin's strong urging, the notion of evolution by natural selection and the notion of progress have been divorced from the outset ... [However] in the limited sense that life evolves hierarchically ... progress is inherent in the evolutionary process'' [Buss 87] (p.186).
If we restrict ourselves to considering only those transitions which do involve an increase in hierarchical object complexity, we see that each entails the absorption of a number of previously free-living individuals into a new, hierarchically higher, unit of selection. When a transition of this nature occurs, two major factors are involved [Buss 87]. The first is a synergism between the interests of the individuals at the lower level and the new `individual' at the higher level. This synergism ``will act to create novel organizations allowing exploitation of the external environment in ways that the lower unit alone could not accomplish'' (ibid. p.185). However, the individuals at the lower level might evolve in ways that disrupt the higher level organisation. Thus, the second factor is a conflict between the two levels of selection. ``The rate and magnitude of such conflicts must be limited, or the higher unit will perish. If variants arise in the lower unit whose effect is to limit the occurrence or magnitude of subsequent variation, then the higher unit will eventually become resistant to further perturbation'' (ibid. p.185).
Once a transition has occurred, then, the new unit of selection must become stabilised if it is not to break down into its component subparts again. (Maynard Smith and Szathmáry also suggest that the genetic relatedness of individuals in the lower level may also play an important role in conferring immediate selective advantage at the time of transition [Maynard Smith & Szathmáry 95] (p.8).) If the new organisation does become stabilised, then it, in turn, is able to participate in the same sort of process, leading to yet higher organisational levels. Bronowski's term ``stratified stability'' captures this picture of progression well [Bronowski 73].2.23
Buss summarises: ``the organization of any unit will come to reflect those synergisms between selection at the higher and the lower levels which permit the new unit to exploit new environments and those mechanisms which act to limit subsequent conflicts between two units'' [Buss 87] (p.viii).
From this point of view, a new perspective on the pattern of evolution emerges. It is a picture of major hierarchical transitions, with long periods in between during which it is debatable whether any consistent notion of progress applies.2.24 Buss claims that ``the major features of evolution were shaped during periods of transition between units of selection'' [Buss 87] (p.188). In the period after a transition, an adaptive radiation of forms is possible to fill the novel evolutionary niches afforded by the new kind of organisation. In the context of organic evolution, Salthe suggests that this expansion continues until a saturation point of number of kinds is reached, ``when the emphasis of the process ... shifts toward coevolutionary elaborations of pairs, guilds, and even more complex symbioses'' [Salthe 85] (p.253). The saturation point can be released if, for example, a subset of the organisms become geographically isolated and able to participate in a new adaptive radiation, but none of these processes will necessarily lead to any general evolutionary progress.
This view of evolution is also consonant with Schwemmler's notion of a constant oscillation between divergent and convergent phases [Schwemmler 89], and, to some extent, with Gould's picture of diversification and decimation [Gould 89].2.25