Unity and Variety in Language
The idea of linguistic universals should not seem surprising if we consider that language is the expression or externalization of thought. Given that thought contains universals, and language expresses thought, language should contain universals. If there is an innately (genetically) determined language of thought, that language will contain universals, so learned public languages will naturally reflect the universals contained in the language of thought. Thought has the kind of biological universality found in perception and memory (as well as emotion and desire)—an innately fixed cognitive system. Thus thought always (in humans) involves a finite base of concepts and rules of combination, unbounded creativity of available thoughts, apparatus for identification and ascription, generality and specificity, logical entailment and reasoning, and basic ontological categories (space, time, object, cause, etc). The idea that investigating thinking subjects in alien cultures would reveal the complete absence of such features is both antecedently implausible and empirically unfounded—as it would be to expect massive variation for perception, memory, and emotion. And if spoken language is just the external expression of thought, reflecting its constitutive features, then we can expect that language will display corresponding features—as human languages clearly do. This is especially obvious if we accept that language is essentially a vehicle of thought—a symbolic system designed to aid thinking. If we think with language, then language will need to possess the shape of thought. Linguistic structure recapitulates cognitive structure. If we allow that thought precedes (spoken) language in human history, then language will inherit the properties of the cognitive system that preceded and triggered it. Hence there will be linguistic universals.
This is the standard picture developed by Chomsky (though I have approached the matter from a somewhat different direction from Chomsky): an innate language faculty, universal to the human species, designed initially as a vehicle of thought. Adding an evolutionary perspective, such a universal innate language system is a useful adaptation, given the utility of language. So we would expect such a useful trait to be installed by natural selection and coded in the genes. It would be inefficient to design an organism that had to learn everything about language from scratch, given that it is useful to everyone; to lack knowledge of language at birth would put one at a disadvantage compared to those born knowing language. It is the same for perception and memory: given their evident universal utility it is better to be born with them than have to acquire them laboriously over time. Or if we consider birdsong and whale language, we would expect innately determined universals common to the species; we wouldn’t expect unlimited local variation and a prolonged learning period ab initio. We would expect instinct not culture, genes not environment, fixity not variety—just like bodily anatomy. We would expect hardwired brain-circuit universality not plastic individual variability, and that is precisely what we find.
But this picture raises an awkward question: Why don’t we all speak the same language? If there are innate linguistic universals, why isn’t all of language universal? Wouldn’t it be much simpler to program a single language into the genes so that there was no need to learn a language after birth? Why do the genes make acquiring language so difficult for us? To be sure, children do it relatively quickly and without apparent effort, but why put them to the trouble at all, when it would be so much more convenient to install a single language at birth? Don’t say that it is because there are so many human languages and we don’t want to limit the child to a single language: the question is why there are so many languages to begin with. Why not streamline the whole process so that there is only one language that needs to be mastered? If we make the assumption that human language began in a single language many thousands of years ago, the question is why it diversified into the variety we see today. We don’t see anything comparable in birdsong and whale language—this wild proliferation of languages each unintelligible to anyone other than its native speakers. Wouldn’t any divergence from the original single language be punished by natural selection, given that divergence entails failure of communication? If someone starts speaking a language of his own invention, no one else will understand it, in which case speaking it is a waste of breath; better to stick with the language you know. It would be like a bird suddenly deciding to sing its mating song with a different melody and rhythm—that would not help in the mating game, and such a bird would soon be selected against. Just as no one today thinks it would be a good idea to start speaking an invented alien language to one’s fellow English speakers, so it wasn’t a good idea millennia ago to start spouting an invented language to one’s monolingual associates. The purpose of language (or one of its purposes) is communication, but that presupposes a shared language–so why did variety in languages emerge? Why isn’t every aspect of language like some aspects of language—innate, universal, and fixed? Why so much linguistic variety?
Here someone might start to waffle on about cultural relativity, the Eskimos, human creativity, chance, historical contingency, complexity theory, and what not. But the question is: What is the use of linguistic variety? What biological purpose does it serve? It seems like a waste of time and energy, and it forces us to abandon the plausible universalism suggested by other aspects of language. Clearly linguistic facts do not need to be learned, local, and variable: so why are some of them this way? Where is the evolutionary payoff? Does it arise from some sort of inevitable propensity to error, like errors in gene copying? Is it that we humans are just incapable or sticking to the one language we have inherited? But birds and whales don’t seem to make such copying errors, botching the language of their ancestors. Is it that in the distant past some dunce did the equivalent of pronouncing “book” as “livre” and the error caught on? We need to find some benefit that attaches to linguistic variety, not just a regrettable tendency to get things wrong. Why can it be good not to speak the language of other people? Why would the genes favor someone with the ability to speak in a way that is unintelligible to others? Why can it be an advantage not to communicate?
The answer I want to suggest is secrecy. The reason distinct human languages exist—the reason they evolved—is that it was advantageous to keep secrets from other people. If two people are plotting against a third person, it is helpful to speak in a language that third person cannot understand. Or if you are a member of closely-knit tribe and are planning with others to engineer a power grab, it pays to speak a language impenetrable to others; it solves the problem of eavesdroppers. You can whisper or meet in secret too, but this is risky if people suspect that something is afoot–better to speak in a new language. Similarly, if you know where there is a good supply of food but you don’t want other members of your tribe to know about it, you might devise a language limited to your family to talk about such matters. Thus linguistic diversity arises from group secretiveness. It arises from facts of social psychology—facts with a clear rationale and evolutionary payoff. It isn’t just a matter of serendipity or human fallibility or novelty for its own sake. Each new language is a kind of secret code in the war of man against man. To put it differently, languages vary because of the fear of being found out. And if the secretiveness is mutual—they are plotting against you, as you are plotting against them—your languages will grow apart even more. There is even selective pressure to make the new language as impenetrable as possible to outsiders so as to prevent code breaking. If someone were incapable genetically of such linguistic novelty, they would be at a disadvantage compared to more linguistically versatile group members; they would be unable to take part so effectively in plots that require secrecy. It is useful to be able to speak a “private language”. So you need linguistic flexibility not fixity. If whales were always out to deceive each other and engage in inter-group warfare (from mild to life-threatening), we might expect some variety in whale language comparable to what we find in human languages; but the social psychology of whales is not like that. We humans started speaking in diverse tongues when we began living in large enough groups to encourage keeping secrets from each other (the family is not large enough because of overlap of genetic interests). To put it simply (but misleadingly): English differs from French because the English and the French were always at war.  More accurately, our ancestors began to diverge linguistically because they needed to keep secrets from each other in conditions of social conflict. No doubt this mechanism began very early in the history of language, long before anything like nations formed. It is evident enough that humans are a sly and deceptive species, prone to faction and feud, and linguistic divergence is one aspect of our propensity to dissemble.
A consequence of not having a single fixed language common to all human societies is that it becomes necessary to learn each language. In order to be flexible the language faculty cannot be completely fixed innately. So there is a price to be paid for being able potentially to speak multiple languages—the labor of learning. Presumably the price is worth paying: it would be convenient to be born knowing a shared human language (as it might be, everyone is born speaking English), but the problem is that everyone would be stuck with that and so linguistic privacy would be impossible. Someone born knowing only one language and not capable of learning any other would be at a disadvantage compared to more adaptable speakers. We may imagine an original position in which people can speak only one language, and this language might be innately known (like birds and whales), but speakers who could branch out into inventing other languages had the advantage over less flexible speakers in the area of secrecy. They can speak in a secret code that baffles others. There is no comparable advantage to being able to perceive or think differently from other people, since these activities cannot be monitored by observers. But when it comes to speaking, observers can tell what you are saying if you share a language, so secrecy will require something that disguises the message (not so for hidden inner thoughts). Espionage requires camouflage. And don’t we see vestiges of this in children’s behavior today—as they contrive to keep their communications hidden from adults by using a “private language”? Sub-groups will often invent a dialect that is closed to outsiders, so that they can keep things to themselves. Slang achieves this purpose, as does jargon. We want to be intelligible to some but not to all—we want to be unintelligible to certain individuals, particularly our enemies. A specially constructed language can serve to make our message invisible to all but the select few (hence secret codes in war time). True, this language has to be acquired, which has its costs, but it is worth it considering that a universal language would be far too transparent for comfort. If human beings had no desire to plot and deceive—if they had no political secrets—they could happily speak a single universal language; but plotting and deceiving is the human way, like it or not. Hence we speak many languages and have to learn each of them (the part that distinguishes languages, not their universal features). By contrast, we do not think with distinct sets of concepts, analogous to different languages, and the language of thought is universal. Linguistic variety is really an anomaly biologically, but it has its biological rationale. It operates, in effect, as a form of camouflage.
It is often said that the purpose of language is communication. That is not incorrect, except if it neglects the function of language as a vehicle of thought. But we should add that the function of human languages, in all their variety, is also to enable lack of communication—to exclude others from understanding everything one says, to camouflage one’s thoughts. What we seek is selective communication, including some and excluding others. Language makes thought public, but we often want our thoughts to be unknown to others—because we desire privacy. A language faculty that revealed our thoughts to any attentive listener would not serve all of our purposes. It would make us too transparent. We need to conceal as well as to reveal. Linguistic universals are useful, but linguistic particularity is also useful. Hence humans speak languages that are both unified and diversified—reflective of thought but also capable of concealing thought.
If these suggestions are on the right lines, we can now say not just that human languages are both unitary and diverse but why they have both these characteristics. They are unitary because of their intimate connection to thought, which itself is unified; and they are diverse because of the social requirement of secrecy, which requires selective communication. If language were used for nothing except as an instrument of thought, we might expect it to be invariant across the human species, as well as innate and unalterable. But its social role in selective communication ensures that languages vary. 
 Of course, I intend this as pithy exposition not literal historical fact: the French and English were speaking different languages well before they went to war with each other. Let me also add that many other factors no doubt contribute to the diversity of human languages; I am simply trying to get at the biological roots of the phenomenon.
 It isn’t easy to think of other aspects of human culture that display this kind of deception-based diversity, language being our primary mode of communication, but we can cite the prevalence of culture-bound gestures such as varying handshakes and hand signals. Such gestures can serve the social purpose of inclusion and exclusion, and they often trade upon indecipherability to outsiders.