When looking at other animals, which appear to lack the same sort of complex language abilities, it opens the question of how human language evolved out of non-language. How do you get from a creature with no complex language to us?

Vocal Origin

Consider our close cousins, the non-human primates. They communicate important information through a varied set of vocalizations. Vervet monkeys, for example, use different calls for eagles versus panthers, and listeners react differently to these calls. These differential vocalizations map to the world in a meaningful way (different calls discriminate between eagles and panthers - in some sense mean that an eagle is near or a panther is near - rather than just being random), and they are arbitrary (the eagle call does not look like an eagle nor sound like an eagle, but comes to be associated with eagles through social learning early in a vervet's lifetime). In this way their vocalizations share similar features with our language.

It is not hard to see how meaningful (that is, symbolic) semantic communication could come about in early humans. Certain great apes pant-hoot when excited or happy; other times they grunt or bark. These vocalizations correspond to internal states (or at least, from the observer-ape's view, they correspond to future behavior that suggests excitement or anger). Perhaps our early hominid ancestors had their own set of grunts and other vocalizations that had differential roles in communicating information. It is not hard to imagine "ungh ungh" turning eventually into "ma ma" if those particular vocalizations came to correspond to a particular object (the speaker's mother).

However, notice that while a vervet's call can distinguish between an eagle or a panther (sharing different information), it apparently cannot communicate that a panther ate an eagle or an eagle ate a panther. Their communication lacks the syntactical structure necessary to communicate such a relationship. Likewise, while we can imagine certain symbolic communication arising out of differential vocalizations gaining different associations through use, it still does not explain how early humans could have developed the syntax that gives modern language so much flexibility and power.

One problem with the idea that language evolved out of grunts and similar vocalizations is that in primates these calls typically arise in the limbic system. Our complex spoken linguistic processing takes place elsewhere in the brain (Broca's and Wernicke's areas, e.g.; many of the same areas as fine motor control). So this makes it seem strange that our language use could have evolved (in a mere six million or so years since we diverged from chimpanzees) from the grunts and howls of our common ancestor with the other primates.

The alarm calls of a vervet monkey, while they discriminate in a meaningful way, seem to be based on affective-physiological reflexes (if you see a panther, you automatically freak and call out). Jane Goodall documented a chimpanzee who discovered a pile of bananas left by researchers. The chimpanzee made a food call as soon as it came upon the treasure, and soon the troupe came and took away all the food, leaving little for the one who had discovered it. When that same chimpanzee later came upon a similar pile, he started to make a food call, but was observed to put his hands over his mouth and fight himself from making the call. Supposedly it was in the more recently evolved 'higher'-functioning systems of the brain that such inhibition was possible to override the more automatic limbic-system call that his body wanted to make simply in response to seeing the food. By holding back his call, he was able to eat to his heart's content.

The key is that if these calls - meaningful and symbolic as they might be (the food call means there's food, not just anything) - arise unbidden in the limbic system, it is not likely that our higher language skills evolved directly from them, else they would share the same brain structures. We would find our own limbic system lighting up (rather than Broca and Wernicke's areas, say) when we used language.

So due to both this neurological evidence and the lack of explanation for syntax, it is hard to imagine how limbic-system calls and other fairly simple vocalizations could become the sort of complex language we use.

Gestural Origin

Consider a different angle then, defended by primatologist Roger Fouts. He thinks our language began gesturally. We know that our ancestral skill set includes gestures. Great apes use them widely. Chimpanzees in the wild use hand gestures to beg for food, to tell another to move out of the way, to order another to stop moving closer by holding up an open palm. We use them ourselves as a fall-back when traveling somewhere we do not speak the native language. How does the use of gestures help explain how syntactic language could have come about?

Let us say you want to communicate the fact that a hawk caught a gopher. With spoken language, you would have to start with the symbolic vocabulary (a symbol for hawk, a symbol for gopher, a symbol for the act of catching; the symbols could be grunts, screams or more subtle spoken sounds) and then somehow add on a grammar/syntax (the subject-verb-object relation implicit in the idea we want to convey when we connect those words together). It is hard to see how to get there in an evolutionary continuity, as explained above. However, with gestures the syntax is actually implicit in the gestures themselves; the relationship is already there.

Try a thought experiment. Hold both of your hands up in front of you. Hold one finger upright on your left hand. Hold your right hand out in front of you, open. Now glide the right hand forward in front of your face and close your fingers around the upright finger on your left hand. A hawk just caught a gopher. If we can get the abstract representational part of things going (that a hand can be a hawk or gopher), that is to say, if we have the ability to use symbols, then grammar comes naturally and for free in a gestural system. The visual representation can communicate such information because it can map many such relationships to the world.

Gestures demonstrate the syntactical information implicitly, not as an alteration of first learning individual vocabulary words and then putting them together with rules out of the blue. And once this basic system is in use (and it seems that being able to communicate such information would serve a useful function and potentially catch on), it is easy to get more complicated information across. Use two fingers, say, to represent two gophers getting caught. Other alterations would soon lead to a very broad repertoire, and the basis for innovation to create further abstraction. This makes it easier to explain how language could come from non-language. Syntax does not need to be invented out of the blue to connect together distinct grunts or calls; it is there all along.

But, one might object, humans use spoken language ubiquitously now, and signed languages are the exception. If language became gesturally, how did we end up here? Perhaps it is just a matter of starting to slowly tie vocalizations in with the main gestural communication. Vocalizations would never have gone away when early hominids began using gestures to communicate extra information. It would be very natural once the latter developed to end up with a bi-modal communication that featured both gestural and vocal information. Spoken sounds might start as a sort of supplement at first, but since it frees up the hands for other things, more and more sounds would come to replace the established functions of particular hand movements (that is, freeing up the hands provides a selective advantage, so it is easy to see why eventually we would end up more with spoken language than signed language).

Interestingly, as Fouts points out, this explains the observation that motor coordination of the tongue mirrors the motor coordination of the hands (as Darwin noted long ago - watch a person sewing and their tongue often follows the movements of their hands). It also explains why the same areas of our brain that are activated for language are the ones involved in certain forms of complex motor coordination (particularly stringing together chains of motor actions).

There is further potential support for the gestural origin of language hypothesis. Consider how human infants are able to learn sign language earlier than spoken language. Our brains seem wired to be ready for gestural communication. Also, a 2006 study by Meguerditchian, et al. connected lateralization (that is, preference for one side of the body, like being right-handed) in baboon gestures with lateralization in their brains. Specifically, when the analogue of the human language area was bigger, it correlated with more preference for use of the hand on the corresponding side of the body for communicatory gestures. The same did not hold for non-communicative motor activities, like simply reaching for food.

Right now this idea merely stands as one hypothesis for how complex language (both syntax and semantics) could have evolved in humans from non-linguistic behavior in our shared ancestry with non-human primates. Much further investigation and testing will need to be done to see if it provides as powerful explanation as it promises, but for the moment it appears to offer one plausible explanation for the evolutionary origin of language.