It is a conceit of hindsight to see evolution as aimed towards some particular end point, such as ourselves. [Extract from The Guardian] [see also Brothers Judd Blog]Dawkins asks the classic software engineering question: if we were to run the program again, would we get the same results. He points out that some biological features (including flight and sight) have evolved independently many times, and might therefore be expected to crop up whenever the program is run; meanwhile other features (such as the extraordinary ability of the horsefly in some parts of Africa to create a corkscrew-shaped burrow in the mud) seem to be unique, and therefore unlikely to recur in different circumstances.
If we regard evolution as a system, then it is the repeatable and predictable outcomes, such as flight and sight, that might possibly be regarded as the purpose and direction of this system, not evolutionary freaks such as humans and horseflies.
General lesson for POSIWID: we must interpret what a system routinely does, not what it uniquely did on a single occasion, we must interpret what a system does.