I am not sure an evolutionary argument gives a 'reason' in the sense you are asking. If it does, the answer to your questions is "Yes, both." Our strong leanings toward capturing simple symbols is just a hardwired mechanism, but it is shaped that way for a reason.
Most of modern experimental psychology would say that we reach out for symbols, and we capture them and then fit them into a context. (Though it requires a very broad sense of 'symbol'.) Our perception does not receive data passively and collate it, but instead projects forward assumptions to meet the incoming data half-way.
For instance, we begin to have emotional reactions before we identify their causes, and then shape the emotional experience to fit the cause. We screen out highly intrusive, obvious and outright scary data when we are acutely focused on something difficult.
Most strikingly to me, we invent object permanence. If we see identical things in two different places in our vision too quickly in succession, we see them as the same thing, we invent a continuous path between them in our vision, and we remember it as real. If the objects are seen in different orientations or in different colors, we think we can remember the exact point in our visual field where 'it' changed color or flipped over, even though 'it' was two different objects, and neither of them was ever in that position.
One way to make sense of this is via evolution. It makes sense that a general reasoning method evolved from a system originally tuned to a narrow range of targets or threats, sought or avoided situations, and not from some more general mechanism that just 'tried to make sense' of the world. As we developed subtler, less pointed goals, we began to see symbols instead of threats or targets. We started to piece our stimuli together into more and more complete packages, and to be less intensely motivated by specific details most of the time. But the basic methodology of scanning for known objects still dominates its structure.
So the theory is that we are not born to learn how to notice two "x" marks as the same thing, but we are born to learn things like how to know two ripe apples as the same thing, or two prowling cats as the same thing, and we have leveraged that for more general purposes.
Once you become attached to this, it is hard to see how other ways learning might evolve would be anywhere near as efficient. It is clearly not the only way of thinking. Many of our more artificial, explicit ways of thinking work directly opposite to this, trying to get the overall sense and not attach to detail or goal first. Resisting the impulse to pre-judge is a primary path to fairness, and whole religions have grow up with that focus. But it seems likely that most of the ways evolution would lead to a brain, would contain some aspect of this bias, since focus on danger is a basic mechanism for survival.
At the risk of continually over-promoting the same book, "Consciousness, Explained" contains this data and the argument much more detail. Dennett lays it out as motivation for his favorite theory of learning, and to make counterarguments to older theories.