As an instrument for the study of the early stages of evolution, we introduce evolutive systems, defined as systems that have the capacity to evolve given appropriate conditions in their environment. They consist of building blocks (e.g. monomers) that are either stable or in steady supply, and of transient assemblies (e.g. polymers) that are entities of great variety, some of which are capable of function. Evolution leads to the accumulation of structure within the transient assemblies during repeated cycles of disintegration (partial or total) and reassembly, on account of the selective advantages associated with transient assembly functions. Transient assemblies must be either inherently unstable or subject to disintegration by agents in their environment. Evolutive systems must have access to a negentropy input in the form of energy in packets larger than typical thermal energies. Reproduction, although not a prerequisite, greatly affects the capacity of evolutive systems to evolve, and thus can be expected to appear in an evolutive system if at all possible. Similarly, functions that require the expenditure of negentropy (for example mobility, breathing, circulation, sensing, communicating, etc.) are not prerequisites for evolution, but can be expected to become established in evolutive systems during evolution through the selective advantages that they confer. A computer-based evolutive automaton is used to explore possible evolutionary scenarios. In the presence of spatial and temporal inhomogeneities, one can construct a multitude of evolutionary scenarios through which various functions, such as the operation of genetic code, can become established within the evolutive automaton. This variety of possible evolutionary scenarios is all the more remarkable because the automaton does not include many important physical processes that would be present in a real system and would greatly multiply the number of possible evolutionary mechanisms and scenarios. Some evolutionary mechanisms are based on survival related selection, while others are based on generation related selection. Previously explored scenarios for the initiation of life have been based mostly on generation related selection. In this paper, we give particular emphasis to survival related selection which is more general in that it does apply to structures and functions related to reproduction but, unlike generation related selection, it is not limited to them. Some of the most basic features of terrestrial living systems can be seen either as prerequisite features of an evolutive system (such as the mortality of living organisms, instability of biological polymers, imperfect reproduction caused by mutations, and the need for a negentropy input) or as features that one can reasonably expect to become established in an evolutive system (such as reproduction and the multitude of living functions that require expenditure of negentropy). This suggests the possibility that an independent definition of living systems may not be necessary if features of living systems substantially overlap with features that one may expect to find in evolutive systems.