Predefined Organism Structure

Organisms in Tierra, Cosmos and similar systems have a predefined, and generally fixed, structure. In Tierra, this structure includes the program's code, various registers, an instruction pointer and a stack; in Cosmos, it is somewhat more complicated (see Section 4.3). It has already been suggested that an attempt to engineer over the early history of life would be problematic. In particular, it is difficult to say how appropriate these structures are as models of any given stage of biological or chemical evolution, or indeed how appropriate they are for modelling living organisms at all.

This is of course related to the lack of theory underlying the model. By imposing a fixed, predefined structure upon the organisms, designers of such systems risk introducing their own prejudices, and restricting the evolutionary potential of the system (as mentioned in Section 3.3.3). In particular, the organism structures used in most of these systems would appear to restrict their potential for modelling some of the important evolutionary phenomena discussed in Chapter 2, such as symbiogenesis and hierarchical evolution. The lack of a theory to govern the design of the environment and the kinds of ways in which organisms may interact does not help in this respect either.

 In addition to the predefined organism structure, the decoding of instructions in Tierra is also `hard-wired' into the system's operating system. Now, from an epistemological point of view, Howard Pattee points out that symbolic information (such as that contained in an organism's genes) has ``no intrinsic meaning outside the context of an entire symbol system as well as the material organization that constructs (writes) and interprets (reads) the symbol for a specific function, such a classification, control, construction, communication ...'' [Pattee 95b]. He argues that a necessary condition for an organism to be capable of open-ended evolution is that it encapsulates this entire self-referent organisation (Pattee refers to this condition as semantic closure). From this it follows that organisms should be constructed ``with the parts and the laws of an artificial physical world'' [Pattee 95a] (p.36).<sup>7.2</sup> In other words, the interpretation (phenotype) of the symbolic information (genotype) of an artificial organism should be constructed and act within the artificial physical environment of the system. Additionally, if the system is to model the origin of genetic information, then the genotype itself must also be embedded within the environment; that is, the complete semantically-closed organisation--the entire organism--must be constructed and act within the physical environment. This issue of the `embeddedness' of organisms in their environment will be analysed in more detail in Section 7.2.3.