The Proto-Human language (also Proto-Sapiens, Proto-World) is the hypothetical most recent common ancestor of all the world's languages.
The concept of "Proto-Human" presupposes monogenesis of all recorded spoken human languages. It does not presuppose monogenesis of these languages with unrecorded languages, such as those of the Paleolithic or hypothetical Neanderthal languages. Advocates of linguistic polygenesis do not accept the notion of a fully developed Proto-Human language and consider the world's language families independent developments of a proto-linguistic form of communication used by archaic Homo sapiens.
If the assumption of a "Proto-Human" language is accepted, its date may be set anywhere between 200,000 years ago (the age of Homo sapiens) and 50,000 years ago (the age of behavioral modernity).
Contents |
There is no generally accepted term for this concept. Most treatments of the subject do not include a name for the language under consideration (e.g. Bengtson and Ruhlen 1994). The terms Proto-World and Proto-Human[1] are in occasional use. Merritt Ruhlen has been using the term Proto-Sapiens.
The first serious scientific attempt to establish the reality of monogenesis was that of Alfredo Trombetti, in his book L'unità d'origine del linguaggio, published in 1905 (cf. Ruhlen 1994:263). Trombetti estimated that the common ancestor of existing languages had been spoken between 100,000 and 200,000 years ago (1922:315).
Monogenesis was dismissed by many linguists in the late 19th and early 20th centuries, when the doctrine of the polygenesis of the human races and their languages held the ascendancy (e.g. Saussure 1986/1916:190).
The best-known supporter of monogenesis in America in the mid-20th century was Morris Swadesh (cf. Ruhlen 1994:215). He pioneered two important methods for investigating deep relationships between languages, lexicostatistics and glottochronology.
In the second half of the 20th century, Joseph Greenberg produced a series of large-scale classifications of the world's languages. These were and are controversial but widely discussed. Although Greenberg did not produce an explicit argument for monogenesis, all of his classification work was geared toward this end. As he stated (1987:337): "The ultimate goal is a comprehensive classification of what is very likely a single language family."
Notable living American advocates of linguistic monogenesis are Merritt Ruhlen, John Bengtson, and Harold Fleming.
The first concrete attempt to estimate the date of the hypothetical ancestor language was that of Alfredo Trombetti (1922:315), who concluded it was spoken between 100,000 and 200,000 years ago. This estimate happens to agree with current estimates on the age of Homo sapiens.
While earliest known fossils of anatomically modern humans date from around 195,000 years ago, the matrilinear most recent common ancestor shared by all living humans (dubbed Mitochondrial Eve), is dated to about 120-150 thousand years ago. The divergence of the three main descendant lines within Africa, L1/A in Southern Africa (Khoisan/Capoid peoples), L2/B in Central and West Africa (Niger–Congo- and Nilo-Saharan-speaking peoples, Mbuti pygmies), and L3 (East Africa, Out-of-Africa migration), dates to about 100 to 80 thousand years ago.
It is uncertain or disputed whether the earliest members of Homo sapiens had fully developed language. Some scholars link the emergence of language proper (out of a proto-linguistic stage that may have lasted considerably longer) to the development of behavioral modernity towards the end of the Middle Paleolithic or at the beginning of the Upper Paleolithic, roughly 50,000 years ago. Thus, in the opinion of Richard Klein, the ability to produce complex speech only developed some 50,000 years ago (with the appearance of modern man or Cro-Magnon man).
The difficulty in making any statement on particulars of Proto-Human lies in the time depth involved, which is far beyond what linguists can trace back today (between five and ten millennia in the cases of Indo-European and Afroasiatic). Some linguists (e.g. Ruhlen 1994) claim that this difficulty can be overcome by means of mass comparison and internal reconstruction (cf. Babaev 2008).
The relatively few linguists who have discussed the subject disagree on how much can be known of the ancestor language. A conservative position, taken by Lyle Campbell, is that it would have shared the "design features" of known human languages, such as grammar, defined as "fixed or preferred sequences of linguistic elements", and recursion, defined as "clauses embedded in other clauses", but that beyond this nothing can be known of it (Campbell and Poser 2008:391). Less conservative linguists have advanced proposals on the vocabulary and syntax of the ancestor language. There are no serious current proposals on its grammar and phonology.
A fairly large number of words have been tentatively traced back to the ancestor language, based on the occurrence of similar sound-and-meaning forms in languages across the globe. The best-known such vocabulary list is that of John Bengtson and Merritt Ruhlen (1994), who identify 27 "global etymologies". The following table, adapted from Ruhlen (1994b), lists a selection of these forms.
| Language | Who? | What? | Two | Water | One/Finger | Arm-1 | Arm-2 | Bend/Knee | Hair | Vagina/Vulva | Smell/Nose |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Khoisan | !kū | ma | /kam | k´´ā | //kɔnu | //kū | ≠hā | //gom | /ʼū | !kwai | č’ū |
| Nilo-Saharan | na | de | ball | nki | tok | kani | boko | kutu | sum | buti | čona |
| Niger–Congo | nani | ni | bala | engi | dike | kono | boko | boŋgo | butu | ||
| Afroasiatic | k(w) | ma | bwVr | ak’wa | tak | ganA | bunqe | somm | put | suna | |
| Kartvelian | min | ma | yor | rts’q’a | ert | t’ot’ | qe | muql | toma | putʼ | sun |
| Dravidian | yāv | yā | iraṇṭu | nīru | birelu | kaŋ | kay | meṇḍa | pūṭa | počču | čuṇṭu |
| Eurasiatic | kwi | mi | pālā | akwā | tik | konV | bhāghu(s) | bük(ä) | punče | p’ut’V | snā |
| Dené–Caucasian | kwi | ma | gnyis | ʔoχwa | tok | kan | boq | pjut | tshām | putʼi | suŋ |
| Austric | o-ko-e | m-anu | ʔ(m)bar | namaw | ntoʔ | xeen | baγa | buku | śyām | betik | iǰuŋ |
| Indo-Pacific | mina | boula | okho | dik | akan | ben | buku | utu | sɨnna | ||
| Australian | ŋaani | minha | bula | gugu | kuman | mala | pajing | buŋku | puda | mura | |
| Amerind | kune | mana | p’āl | akwā | dɨk’i | kano | boko | buka | summe | butie | čuna |
Based on these correspondences, Merritt Ruhlen (1994b:105) lists these roots for the ancestor language:
In a 2003 paper, Murray Gell-Mann and Merritt Ruhlen argued that the ancestral language had subject–object–verb (SOV) word order. The reason for thinking so is that the world's major language families nearly all reconstruct back to SOV word order in their earliest stages. Their proposal develops an earlier one made by Talmy Givón (1979:271-309).
If this thesis is correct, it would have wide-ranging implications. Since a key article by Joseph Greenberg in 1963, it has been known that SOV word order is commonly associated with a series of other phenomena (Gell-Mann and Ruhlen 2003:3-4). Among these, some of the most important are:
For example, instead of saying The man goes to the wide river, as in English, Proto-Human speakers would have said Man wide river to goes.
Many linguists reject the methods used to determine these forms. Several areas of criticism are raised with the methods Ruhlen and Gell-Mann employ. The essential basis of these criticisms is that the words being compared do not show common ancestry; the reasons for this vary. One is onomatopoeia: for example, the suggested root for 'smell' listed above, *čuna may simply be a result of many languages employing an onomatopoeic word that sounds like sniffing, snuffling, or smelling. Another is the taboo quality of certain words. Lyle Campbell points out that many established proto-languages do not contain an equivalent word for *putV 'vulva' because of how often such taboo words are replaced in the lexicon, and notes that it "strains credibility to imagine" that a proto-World form of such a word would survive in many languages.
Using the criteria that Bengston and Ruhlen employ to find cognates to their proposed roots, Lyle Campbell finds seven possible matches to their root for woman *kuna in Spanish, including cónyuge 'wife, spouse', china 'girl', and cana 'old woman (adjective)'. He then goes on to show how what Bengston and Ruhlen would identify as reflexes of *kuna cannot possibly be related to a proto-World word for woman. Cónyuge, for example, comes from the Latin roots for 'with' and 'yoke', so its origin had nothing to do with the word 'woman'. China is a loanword from Quechua and thus cannot be an example of a Spanish inheritance from Proto-World. Cana comes from the Latin word for 'white', and again shows a history unrelated to the word 'woman' (Campbell and Poser 2008:370-372). Campbell's assertion is that these types of problems are endemic to the methods used by Ruhlen and others.
There are some linguists who question the very possibility of tracing language elements so far back into the past. Campbell notes that given the time elapsed since the origin of human language, every word from that time would have been replaced or changed beyond recognition in all languages today. Campbell harshly criticizes efforts to reconstruct a Proto-human language, saying "the search for global etymologies is at best a hopeless waste of time, at worst an embarrassment to linguistics as a discipline, unfortunately confusing and misleading to those who might look to linguistics for understanding in this area" (Campbell and Poser 2008:393).