Professor of Linguistics
University of New Hampshire
Durham NH 03824 USA
karl.diller@unh.edu

– Internal Reconstruction, the Comparative Method, and Hard Evidence.

– The Aquatic Ape syndrome and speech.

– The Preadaptation for Speech

– Implications for the origin of language.

If "human language is an embarrassment for evolutionary theory", as stated in the epigraph at the head of the Web Page for this conference, it is because linguists and anthropologists have not taken seriously the method of internal reconstruction and the evidence it gives for the Aquatic Ape Theory.

Linguists are brought up on the story of how the young Ferdinand de Saussure, in his doctoral dissertation, reconstructed a series of pharyngeal consonants for the proto-Indo-European language, even though these consonants were not present in Sanskrit, Greek, Latin, or any known Indo-European language. The pharyngeal consonants were reconstructed not by comparative evidence but by the method of internal reconstruction. Various alternations in the vowel systems of Indo-European languages could be explained if there had been an original series of pharyngeal consonants.

How reliable is internal reconstruction? Don’t we need harder evidence? Classicists and philologists wedded to the text were highly skeptical. Then a few years later the Hittite language was deciphered – it was an Indo-European language with a series of consonants right where de Saussure had posited them.

Morphophonemic rules which are formulated by internal reconstruction to take account of sound alternations in present day speech end up recapitulating the history of sound changes, as seen in Bloomfield’s classic paper on Menomini morphophonemics and Chomsky and Halle’s synchronic reconstruction of the English vowel shift in their Sound Patterns of English. Bloomfield’s rules mirrored those derived from comparative Algonquin data; Chomsky and Halle’s rules were like those derived from historic texts.

The distinction between the method of internal reconstruction and the comparative method is also seen in molecular biology and genetics. Comparing mitochondrial DNA in various populations to find the source of the human population in an African Eve is a use of the comparative method. The recent reconstruction by Lahn and Page (1999) of the history of how the human Y chromosome came to diverge from its originally homologous X chromosome is a very interesting example of internal reconstruction, involving four separate inversions of DNA on the Y chromosome, ordered in a clear chronological order. The time scale was then made more precise by comparative genetic data and by using the fossil dating of species divergence. The first event, which marked the beginning of X-Y divergence occurred about 240 to 320 million years ago, shortly after the separation of mammals and birds from their common ancestor – pretty far back in the pre-history of humans! The most recent inversion occurred between 30 and 50 million years ago, during the period of primate evolution but before the split between Old and New World monkeys. This is a model study for the relative importance of internal reconstruction and hard evidence. Fossils are incapable of yielding any evidence about the evolution of the Y chromosome. Internal reconstruction defined the problem and provided the explanation. Then comparative evidence and hard fossil evidence provided refinements and certain kinds of precision.

In an inspired example of internal reconstruction, Oxford marine biologist Alister Hardy (1960) argued that several unusual aspects of human anatomy and physiology can be explained if a human ancestor went through an aquatic or semi-aquatic evolutionary phase.

The biological preadaptation of speech – the voluntary control of one’s breathing and vocalization – is a feature of aquatic mammals, as Elaine Morgan has pointed out in three books since 1972, and it fits well with the cluster of other human aquatic features initially pointed out by Hardy. Hardy’s list includes reduced body hair, the orientation of the hair tracts, the layer of subcutaneous fat, sweat glands, upright posture, flexibility of the spine, swimming and diving capabilities (including the slowed metabolism of the diving reflex), and the form of the hands and feet. Morgan adds the features of salt tears, the ability to frown, the shape of the nose, the form of the buttocks and female breasts, face to face copulation, and the foundations of speech.

These are features we share with aquatic mammals and which we do not share with chimpanzees. In particular, the ability to pronounce words of human language is almost completely beyond the capability of the chimpanzee, as is obvious to anyone who has seen films of the Hayes’s chimpanzee Vicky trying with great effort to produce the repertoire of four words she learned to mimic after six years of behaviorist training. The words were mama, papa, cup, and up, pronounced without voicing, with great contortion of the face and upper body, requiring even the help of the hand to keep the lips closed for the whispered /m/ in mama. A chimpanzee would not be a good candidate to initiate the evolution of speech.

Hardy had just returned in 1927 from a two year excursion to study the biology of the Antarctic Seas, when he found this statement in Wood-Jones’s Man’s Place among the Mammals: "The peculiar relation of the skin to the underlying superficial fascia is a very real distinction, familiar enough to everyone who has repeatedly skinned both human subjects and any other members of the Primates. The bed of subcutaneous fat adherent to the skin, so conspicuous in Man, is possibly related to his apparent hair reduction; though it is difficult to see why, if no other factor is invoked, there should be such a basal difference between Man and the Chimpanzee." Hardy immediately recognized the "other factor" that was involved: an aquatic environment. Nearly all aquatic warm-blooded vertebrates have a layer of subcutaneous fat, from ducks and penguins to dolphins and whales. And mammals who have permanently returned to the aquatic environment from land have generally lost their hair, as whales and dolphins have.

Nearly every major group of vertebrates from the time of the dinosaurs has had a member species return to an aquatic or semi-aquatic environment. Paleontologists have no trouble identifying the skeletons of aquatic dinosaurs, the ichthyosaurs and plesiosaurs. Since humans never became as completely aquatic as the whales, we do not have a completely aquatic-looking skeleton. But note the elongated streamlined body, which is well adapted to swimming and which requires upright posture on land.

The theory is that the human ancestor was forced into an aquatic environment – perhaps marooned on an island – and because of this geographical isolation, diverged into a new species of primate. Presumably this animal always slept on land, but spent a major part of its waking hours in the water, much of the time wading upright with its head out of the water, searching for shellfish and other marine creatures for a diet that came to be increasingly derived from animal sources. This diet, rich in Omega-6 and Omega-3 fatty acids, would have permitted the increase in brain growth that was initiated at this time.

The proposals are actually quite specific. Leon P. LaLumiere first located the island to what is now the Danakil Alps in the Afar region of northeastern Ethiopia near the fossil bed of Lucy, the australopithecus afarensis. The Danakil Alps are surrounded by salt plains, and were isolated by the Afar Sea at the time of the divergence of the human line from chimpanzees. Morgan quotes the geologist Paul Mohr, that "What is more or less agreed upon is that by the late Miocene [seven million years ago] a marine basin had become established over northern Afar...and these conditions persisted until the isolation and desiccation of the salt plain arm of the sea some 70,000 years ago (CNR-CNRS team, 1973)" (quoted in Morgan 1990, 51). This area was a forested primate habitat at the time of the inundation, and would have been an obvious setting for the separation of the human line from the other apes, and for the shaping of the human line in a forced aquatic setting.

The implication is that whether or not the Australopithecines were direct descendants of humans, or whether they escaped from the aquatic environment at a divergent time, they were shaped in the same or in a similar aquatic setting. Along with their bipedal gait, the Australopithecines presumably had most of the other aquatic features of human anatomy including hairlessness and the layer of subcutaneous fat and the other anatomical features that separate us so markedly from Chimpanzees. But of course if the Australopithecines were not in the direct ancestral line of humans, they may not have developed all these aquatic features as much as the human line did.

Mammals forced from land into aquatic environments develop voluntary control of breathing and specific adaptations to prevent water from entering the lungs through the nose. Elephants, excellent swimmers and another species with an aquatic interlude before returning to land, developed an extremely long nose which they can hold up out of the water. The human nose is not so spectacular, but it is impressively large compared with Chimpanzees, and is streamlined appropriately for a swimmer. The human nose is similar to that of the Proboscis monkey, a semi-aquatic primate living in the mangrove swamps of Southeast Asian islands. Seals developed valvular nostrils; humans developed the velum which closes off the back of the nose. Human diving capabilities are formidable, as seen for example by traditional pearl divers in Japan and Korea, and Hardy points out that humans have a partial diving reflex in which metabolism is slowed. A swimming and diving creature needs to be able to take in quickly a large amount of air, and for humans this is facilitated by the descent of the larynx which enables and enhances the taking in of large breaths through the mouth. (See also the contribution by Verhaegen and Munro in this conference proceedings on the aquatic preadaptations for speech).

The disreputable theories of language origin, the bow-wow theory, the ding-dong theory and the like, were hung up on the problem of how voluntarily controlled speech could have emerged from instinctive stimulus-bound cries. This problem disappears if the earliest bipedal ancestors of humans were preadapted for voluntary control of vocalization. If these human ancestors also had the ability of present-day chimpanzees to use manual and visual symbols and to understand vocal symbols, we are at a very good starting point for a system of vocal language without a lot of hand-waving.

With preadaptation, we have a very long timescale to work with. Swadesh (1971) argued that there were eoglottic, paleoglottic, and neoglottic eras in the prehistory of language, corresponding to the eolithic, paleolithic, and neolithic eras of toolmaking. The eoglottic/eolithic period, he argued, went back to the australopithecines, and was a very long period in which the elements of language were as limited and primitive as the hand tools. Preadaptation makes it plausible that pre-humans actually used that entire period of time in developing language. There need not be any direct connection between toolmaking and language to suggest that language development might have taken as long in pre-human culture as toolmaking ability did. If so, we presumably have a situation where the evolution of language was longer and more graded than in the model of two relatively quick and almost discontinuous stages of protolanguage and full language. We may have had eo-protolanguage, paleo-protolanguage, and neo-protolanguage. Given the long timescale and what we know about grammaticalization, we also may have had an appreciable amount of proto-grammar in those earlier eras. After all, a preadaptation for voluntary use of the vocal tract implies the existence of a basic functional Broca’s area in the motor association cortex, and just as agrammatism is associated with Broca’s Aphasia, basic grammar ability may well be associated with this elementary but functional Broca’s area.

 Conference site: http://www.infres.enst.fr/confs/evolang/