At each time in our development, we have a limited view of the world, a limited horizon, both as individuals and as groups or cultures. Similarly, each organism has, in a sense, a limited world view, a horizon of perception and action defined by its structure. The world view, in turn, is the embodiment of a certain body of knowledge.
In the history of philosophy, we can discern two basic types of views or schools of thought about how knowledge arises. On one side are empiricists who think of the human mind as a blank slate on which experience writes something. On the other side we have rationalists, who emphasize the role of pure reason, independent of experience, and typically assume that there must be some pre-existing structure of the mind to make it possible to make any sense of experience. In classical forms of rationalism, this innate structure defines a limited world view as well, and a horizon beyond which we cannot look.
There are obviously processes of perception and thought, and there should be some active part of the mind that does the processing. We might think of this as something like programs. However, in classical rationalist theories, this pre-existing structure of the mind is normally seen as something static.
My own position in these matters is that while there is an active structure at any time, this structure is changing. It is self-modifying and developing. There should be something like an initial version of it that is innate, but this is not a fixed structure, something like a fixed set of laws of thinking or a fixed program. Instead it is just a nucleus of development, a starting point, an initial structure from which we depart. Some of the innate structures might be “switched on” or “ripening” at a time when others have already be changed by learning processes (a process that is going on at least until we finish puberty), so the innate structures might never be there as a complete initial state, but we might still think of it as an initial structure (with some initially dormant parts activated later). The innate developmental nucleus might look different in different people, so that they show different personality structures right from the beginning. However, I think that there are no unchangeable structures of cognition.
So, while in most animals, let alone simpler organisms, the cognitive horizon is fixed or can only change by small amounts, humans are able to move and change, changing their point of view and widening and shifting their horizon. The metaphor of a horizon might be a bit misleading here or must be taken as a simplification, however: we are not only widening or shifting the part of the world we are perceiving but we can completely change the language and concepts used to describe the world and thus the ontology of the world as we perceive it, i.e. its structure and composition in terms of entities that constitute it.
Some parts of the brain seem to have a higher degree of hard wiring, especially parts involved in primary processing of sense data and parts involved in direct control of our muscles. However, even these structures must be trained and can be modified. For example, people who suffer brain damage as children can learn to compensate the loss to some extent, repurposing different parts of the brain.
If the learning processes have such a potential of compensation, evolution cannot even produce the perfectly working logical mind that classical rationalism assumes. As soon as it evolves a developmental nucleus that can be rounded up by creative learning processes, evolutionary pressure would stop, at least if children are protected inside their community and are therefore not forced to function perfectly immediately after birth. If plasticity opens up new adaptive possibilities, evolution might even remove specialized structures. So the main thing that might have happened in human evolution might not be the development of new modules of cognition for novel specific types of thinking but instead the development of unspecific, freely “programmable” neuronal networks and the reduction of specialized structures. New types and ways of thinking would then have arisen not by genetic change but by cultural innovation. What we might find is that there where mutations to increase the size of the brain by just producing more unspecialized brain tissue, by increasing the density of neurons or of connectivity between neurons, or other unspecific, general changes increasing freely programmable brain power, e.g. general mechanisms to ease the automation of processes into habits or to ease reflexive analysis of thought processes by some feedback loop. We might find that during evolution towards humans, very little specific, sophisticated structures have arisen and that such specific, specialized structures instead where reduced.
I think that even the developmental nucleus that probably exists in the human child is not a necessary precondition of learning. While Kant, for example, thought that certain structures where necessary preconditions for any acquisition of knowledge, I think that for each more sophisticated developmental core you can start with, a simpler one could be found that could develop into the more sophisticated one in a learning process. However, the simpler such a core is and the less information such systems contain, the less economic or efficient the will tend to be (Kurt Ammon is speaking of a “principle of cognitive economy”). The structures Kant thought to be necessary could be implemented in terms of programs and these programs could be developed in a learning process. However, too simple a starting structure would be too uneconomic in terms of biological evolution. Learning would simply take too long. We can therefore expect some developmental core to be there as a result of evolution, but it will not be a very sophisticated structure. And any cognitive structure it represents could be replaced by a more sophisticated, different one as a result of learning, giving it just a role comparable to scaffolding. But theoretically, we could replace each initial structure with a simpler one. The Core is, in a sense, empty, formless or undifferentiated, as Ammon has noted.
This view is certainly neither classical empiricism nor classical rationalism. One could call it empiricist rationalism or, rather rationalistic empiricism. It is the view that there are some (simple) innate structures but that these structures are temporal and are modifiable, perhaps beyond recognition, during learning. In this view, creativity is at the core of cognition. This is the reason I am speaking of “creativistic philosophy” or “creativism”.
In this view, the mind, at any stage, is active, not just a passively recording blank slate, but the active component is changing and the possible changes are not predetermined since the process changing them is itself changing. A complete and exact formal description of these processes is not possible, any such description is incomplete. There are no fixed structures, laws or categories of thought and perception, no fixed, unchangeable ontology, no unreachable thing-in-itself, no horizon that cannot be shifted. There is no fixed knowledge representation language or fixed “mentalese” language of thought (except for neural networks in general or universal programming languages). There are no linguistic universals. There is no Chomskyan fixed structure underlying all language and no fixed “language acquisition device” (instead, I think, there is, if anything, at most a simple developmental nucleus of such a system that is itself modified and extended during the process of language acquisition, but that is a topic for another article). The logical formalisms developed in analytical philosophy and artificial intelligence may be used as partial descriptions of specific aspects, but they are not universal (and consequently, analytic philosophers cannot agree on them and they never stabilize).
This re-programmability means that cognitive psychology cannot be a science in the sense that it is dealing with fixed laws. There are laws of neurology, but just like the electronic parts of a computer don’t tell you much about what the computer is doing, the neurology does not tell you much about what is going on at the psychological level. What the computer is doing depends on the software and what the brain is doing depends on learned structures that incorporate information from the outside world that cannot be derived from the laws of neurology. These learned structures are a result of our biographies and of the histories of the societies and cultures we are embedded into.
One could say that rationalism is appropriate for most animals, if you look at what these animals know and how they arrive at that knowledge. There are, however, some animals whose behavior is partially creative and cultural, including some wales, some birds and some apes. For those animals, rationalism does not fit so well again, and empiricism enters the picture. For human beings, these creative and cultural components of cognition become dominant. Here, there might be a remnant of rationalism in the form of a developmental core of cognition.
However, rationalism in its classical form does not fit the “animal rationale”.
(The picture is from http://commons.wikimedia.org/wiki/File:Geocentric.jpg. It shows a representation of a geocentric world view, as an example of a historic state in the history of ideas, a bunch of knowledge defining a certain world view that was later replaced by a different one.)
 See https://creativisticphilosophy.wordpress.com/2014/10/08/world-views-and-evolution/. The views discussed there are connected to the theories of the biologist Jacob von Uexküll, whose theories about the “Umwelt” of animals defined by their structure may be seen as a generalization of Kant’s views on epistemology.
 A note for those readers not familiar with philosophical nomenclature: the word does not derive from “rational” (as opposed to “irrational”) but from the Latin “ratio”. Of course, the term “rational” is in turn derived from “ratio”. So the opposite of “rationalism” in the context of epistemology is not “irrationalism” but “empiricism”.
 Ammon, Kurt: “The Automatic Development of Concepts and Methods“, Doctoral Dissertation, University of Hamburg, 1987, p. 77.
 Ibid., p. 174-175.