Let me try something dangerous and talk about intelligence without really defining it; there are many different kinds of intelligence and the arguments here will hold for most definitions I can think of. The necessary requirement is that intelligence is an emergent property of individuals (not necessarily humans; not even necessarily biological life formsbut for now constrained to life forms in general–in the sense of mutating auto-replication and pursuit of survival) that allows them to adapt to changing conditions on an intra-generational scale (evolution, for example, is a mechanism for adaptation on an inter-generational scale). I believe (though, quite frankly, haven’t thought hard about it) this is sufficient to go on.
Is intelligence a necessary artifact of evolution? To expand on this, what set of circumstances make intelligence a much more desirable trait than other traits, and how likely is intelligence to emerge? Evolution deals with randomness — it’s a greedy random walk, favoring changes that increase the species’ chance of survival. What makes intelligence better than, say, a stronger set of legs? I have two theories. First, as life forms evolve and strengthen their physical characteristics, it becomes inefficient to continue the physical growth; either it leads to massive energy needs which begin to outweigh the individual’s abilities to gather food, or it leads to side effects inherent in the mechanics of a body (stronger legs may lead to worse injuries). Evolution, essentially, runs out of avenues to pursue and non-physical development becomes the most energy-efficient. Secondly (now I realize the two theories are related), evolution’s greatest limitation is its speed — it must act over generations; and with complex enough organisms the generation cannot be very short. If the natural circumstances favor quick adaptability (for example, a series of ice ages come and go too quickly for any single species to evolve around them), evolution must replace itself with intelligence.
Of course, I may be wrong and intelligence could just be a fluke.
Regardless, if I wanted to have a particular characteristic evolve, I could manufacture a world which favors that characteristic and watch nature come up with it through a process of evolution. In the extreme, if all I had was plants and wanted the species to be able to walk, I would provide incentives for the plants to displace themselves (maybe an ever-moving source of food). Early species will probably simply grow fast, or maybe have the ability to detach themselves from the soil and attach themselves back, propelled by wind. Ultimately species would develop self-propulsion (I could help them by providing a negative incentive to simply go where the wind takes them). Nature would “cheat” and use water as an interim medium — it’s easier to be able to walk if you are already swimming — and so we can see how ultimately we would have species able to walk.
Similarly, what would I have to do to favor intelligence?
I did make an assumption that the life form evolves, that is, life replicates itself (a “species” composed of a single individual that doesn’t die cannot evolve) with mutations between successive generations. In order for evolution (that is, a long-term progression) to take place, there must be survival of the fittest, and with it, the favoring of life to non-life by the individuals. That second assumption is interesting because I’m not quite sure how it came about and why it holds true for species. With intelligent species such as humans you could make an argument that the will to live is an outcome of consciousness — a constantly running narrative of our life, created thanks to the development of memory and the ability to make connections (non-intelligent species have memory but they can’t connect it into a narrative) — but for all other species, it’s not so black-and-white.
It’s, obviously, just as fascinating to talk about why intelligence exists as how it exists — I think that we tend to focus too much on the latter and not enough on the former (and the theories above are just a small step towards that thought). But, on the how, we can learn a lot just by drawing a parallel between it and other non-mental features of evolution. Intelligence requires the environment — and with it sensory inputs and the feedback element with the environment. Intelligence is a wonderful example of a (relatively — all purists calm down) binary characteristic that nevertheless came about gradually from non-intelligence (just as flight came from non-flight; the outcome is clearly distinct but it’s not immediately clear how non-flight evolved into flight).