Me Tarzan, you Jane. Let's snog

Do we owe our linguistic complexity to silver-tongued forebears? Geoffrey Sampson delves into the origins of language

Both these books are about how language evolved - a subject that was once taboo for linguists (because it seemed so wholly speculative) but has been rehabilitated over the past 20 to 30 years. Robbins Burling is concerned chiefly with what we know (or can reasonably surmise) about how our remote ancestors developed from languageless apes into creatures using a stock of meaningful words. Guy Deutscher discusses the later and more accessible process by which the "me Tarzan, you Jane" stage of language evolved into the systems we know today, with their wonderful structural intricacy and logical expressiveness.

Deutscher stunned many linguists in 2000, at a time when it was widely taken as axiomatic that "the earliest written documents already display the full expressive variety and grammatical complexity of modern languages" (to quote Ray Jackendoff), by demonstrating that this simply is not so: the written records of Akkadian, an ancient language of the Middle East, show complement clauses evolving under the pressure of new communicative needs in a language that previously lacked them. More recently, Deutscher's finding has been somewhat overshadowed by Dan Everett's even more remarkable discovery that a language of our own day, Pirahã, spoken by a monolingual tribe in the southern Amazon basin, is astonishingly simple in logical structure - not only does the language entirely lack clause subordination of any kind (early Akkadian did have relative clauses) but, for instance, it cannot express any quantification concepts (no "all", "most", numerals and so on).

This all helps to make the evolution of grammar, which would once have seemed too mysterious to say anything useful about, a topic whose time has come. Once we are prepared to entertain it as a realistic subject for research, many current languages turn out to contain fossilised clues to how abstract grammatical features evolved out of words for concrete realities long before written records began. The slice of time that we can inspect directly is not long enough to see much detail of how modern grammars evolved, but the fossil clues from different languages point in a sufficiently consistent direction to give us confidence about how in principle it must have happened.

Through my work as a panellist for the Ask-a-Linguist public information service, I know that no aspect of language or languages fascinates the public as much as questions about their ultimate origins, and Deutscher has accordingly aimed his book at a very broad readership. Sometimes his funky style verges on the embarrassing - like when a vicar insists on hauling out a guitar to make divine service "cool". If a topic is intrinsically interesting, to my mind it is best expounded in a non-pompous but sober style and allowed to stand on its own feet. But I suspect Deutscher would argue that his unusual style serves a very specific purpose, in helping the reader to contemplate well-known phenomena through new eyes.

Take, for instance, his extended treatment of Semitic verb inflection, whereby roots consisting solely of consonants are put into various tenses, moods and so on by interdigitating various patterns of vowels among the root consonants. In On the Origin of Species , Charles Darwin used the eye as an example of an organ that one would suppose must have been designed as a finished whole, yet which in reality could be shown to have evolved by many small steps, each of which was independently adaptive. In a similar way, Deutscher feels that if he can convince readers that Semitic verb inflection might have emerged through the same kinds of processes that we can hear happening in our own language today, then he will have won them over to his general point of view.

A young Jew beginning Hebrew school would typically be introduced to verb inflection via some standard Biblical root such as l-m-d, "to learn". Instead, Deutscher invents a hypothetical root s-n-g, "to snog", which yields Hebrew or Akkadian forms such as asnug "I snogged", ushasnag "I shall cause to snog", sunnag "he was snogged intensively" and so on. The structure of a classical language commonly seems to us to possess a monumental quality that is worlds away from the everyday sloppiness by which, say, an English phrase such as "going to" is reduced to a slovenly "gonna" and used to express futurity. By describing Semitic verb forms in terms of snogging, Deutscher makes it easier for the reader to get his mind round the idea that these verb forms, and all other grammatical structures, are products of that same sloppiness. In millennia to come, schoolchildren may find themselves studying the "gonnative" tense of 21st-century Classical English in much the same spirit of resentful awe with which youngsters in recent times have contemplated Greek aorists, Latin perfects and pluperfects, or the Niphals and Hiphils of biblical Hebrew.

The bulk of Deutscher's book shows how the three factors of expressiveness, analogy and economy (or sloppiness) can between them explain the full growth of a structurally complex language over many millennia, from an initial "me Tarzan" stage at which there were only concrete words and no grammar or methods for expressing abstractions. But to speakers of modern European languages, it often seems that in the historical period language evolution has been proceeding in the opposite direction. English in particular, but also modern continental languages, feel structurally simpler than the languages of classical antiquity. Famously, August Schleicher (the scholar who in 1861 first formulated the family-tree concept of language relationships) explained this simplification in terms of Hegelian philosophy. Nations in the prehistoric period evolve their intellects, and with them their languages, to a point where they can achieve free will and rise above the blind laws of nature to create their own histories; after that, they are no longer dependent on the structure of their languages, which accordingly withers away as a plant withers after its time of ripeness has passed. For most of the subsequent century and a half, linguists have scoffed at these ideas as unscientific, romantic nonsense. Remarkably, in an epilogue Deutscher argues tentatively that Schleicher may have been on to something. Deutscher believes that it may indeed be characteristic of advanced modern civilisations to reverse the arrow of language evolution.

Burling's book contains none of Deutscher's verbal fireworks. It is clearly written, but if Burling had occasion to discuss Semitic verb inflections he would surely use the verb "learn" rather than "snog". Burling gives a workmanlike and up-to-date account of what we can infer about our distant ancestors' initial acquisition of language: he assumes no prior knowledge, although his readers will probably be people with an interest in the human sciences. Burling draws attention to controversies, but he consistently suggests that we should prefer reasonable, biologically realistic scenarios over wild if influential hypotheses, such as Derek Bickerton's idea that the full complexity of modern language emerged overnight from a genetic mutation in a single individual.

Although Burling is concerned chiefly to present consensus views of how our species became verbally articulate, there are interesting issues that come more into focus in his book than in other comparable writings. One of these is the relative importance of the ability to comprehend language, compared with the ability to produce it. As Burling rightly says, discussions of language acquisition usually emphasise the speaker's rather than the hearer's side of the equation, for the obvious reason that speaking is so much easier for a third party to observe. But Burling argues that the acquisition process in the individual child is led by comprehension, with speaking abilities following on after a delay; he suggests that the mistaken emphasis on production may have systematically distorted our understanding of the evolution of language in general.

Burling's book also gives full attention to an idea that has emerged in recent years, that the evolutionary advantage of language for the earliest speakers had to do with social relationships. People had long taken it for granted that the reason why language promoted survival was because it helped human groups to co-operate in manipulating the non-human world, for instance co-ordinating their behaviour when hunting or transmitting technological knowledge. But Burling, who has spent many years living in a tribal society in north-eastern India, comments from personal experience that language is not in practice crucial to the execution of such functions in primitive societies.

A number of recent writers (notably Geoffrey Miller) have argued that the true reason why those early members of our species who were skilled in using language tended to leave many descendants (which is the definition of evolutionary fitness) was that speaking well is the key to social popularity and to sweet-talking members of the opposite sex. Understanding and manipulating the external world - which to the modern academic seems such a central use of language - could have developed as a secondary function after language was well established. Thus, it may be that Deutscher was spot on when he chose "snog" as his archetypal ancient verb root.

Geoffrey Sampson is professor of natural language computing, Sussex University.