So in order to start an open-ended description-based replicator (one that is physically capable of what we generally refer to as Darwinian or biological evolution) you have to be able to specify multiple objects (among alternatives) using a common transcribable medium. This requires an irreducible organization made up of rate-independent memory tokens (symbols) and a set of non-integrable constraints, operating together in a semantically-closed system. The products of this system must successfully specify and produce a very particular dissipative process. The objects in this dissipative process must use the laws of nature to cause the medium to be processed, the products to be produced, and the memory to be copied and be placed inside a separate replicant along with a complete set of constraints. And for that pathway to be successful (i.e. semantically closed) requires a simultaneous coordination between the individual segments of the medium that describe the constraints and the individual segments of the medium that describe the various constituents of the dissipative process (i.e. changing the arrangement of one segment, changes the product of all the other segments). These requirements aren’t merely a mouthful, they are an accurate (and heavily abbreviated) summary of what physics and biology have taught us through logic, prediction, and confirmation via experimental result.