Questions and approaches
Work based on classical methods of historical linguistics has brought about a wealth of insights into the substance and structure of earlier forms of human languages. But this work gives us access to only a fairly small phase in the history of human languages: Linguistic reconstruction becomes notoriously fuzzy and conjectural once we are dealing with a time depth exceeding 8,000 years. While there exists a wide array of opinions on how far back reconstruction can be pushed and on how the genetic and areal relationship patterns among earlier languages may have been, there appears to be general agreement on the following point: The languages that were spoken 8,000 ago were typologically not dramatically different from what they are today. Thus, compared to the biologist, the linguist is in a deplorable situation:
In other words, the central aspects of language—syntax and phonology—have no evident homologs. In that sense, language is an emergent trait (or ‘‘key innovation’’) and poses, along with all such traits, particularly difficult problems for evolutionary biology. Likewise, there are no archaeological digs turning up specimens of the language of 100,000 years ago. While the fossil record has given us a reasonably clear picture of the evolution of the vocal tract, grammatical structure, needless to say, is not preserved in geological strata. (New meyer 2003: 61)
What we factually know about the neurology, cognitive capacity, and sociocultural environment shaping early language development is severely limited. Fossil bone findings—for example on the enlargement of the thoracic vertebral canal, the descent of the larynx, and an increase in encephalization—have been used in linguistic reconstruction work. But the exact significance of such findings remains largely unclear, and some of these findings, such as the descent of the larynx, have recently been shown to be entirely irrelevant (Fitch 2002). Davidson (2003: 144) therefore concludes that anatomical evidence has proved a poor guide to speech abilities, and has contributed little to the understanding of the emergence of language.
That we know essentially nothing about the evolution of language in pre-modern times is certainly not because not enough research has been done on it, and also not because there is a lack of plausible hypotheses on this issue: on the contrary, scholars have been working on it for at least three centuries, if not longer, and have come up with many fascinating hypotheses. Still, in spite of all this work, no empirically really satisfactory information has surfaced from this research on any of the following questions that we consider to be important to understand language genesis and evolution:
(1) Open questions
a. Why did human language evolve, and what purpose did it serve?
b. When and where did it evolve?
c. Who were the creators of early language?
d. Wasitsorigin mono-genetic or poly-genetic, that is, do the modern languages derive from one ancestral language or from more than one?
e. Were the forms and structures characterizing early language motivated or arbitrary?
f. Did language originate as a vocal or a gestural system?
g. Can language genesis be related to the behavior of non-human animals?
h. Was language evolution abrupt or gradual?
i. Which is older—the lexicon or grammar?
j. What was the structure of language like when it first evolved?
k. How did language change from its genesis to now?
l. How long did it take to develop a structure that corresponds to what we find in modern languages?
m. How did phonology evolve?
n. How did the properties believed to be restricted to modern human languages arise, in particular syntax and the recursive use of language structures?
Not surprisingly therefore, many linguists believe that these questions are beyond the scope of their discipline, or of any discipline, and are therefore—in the wording of Chomsky (1988: 183)—‘‘a complete waste of time’’. This view has a long tradition: and in much the same spirit, the Linguistic Society of Paris and the British Academy urged their members in the 1860s to refrain from discussing the origin of language, as a reaction to the speculation that had dominated discussions on this issue in linguistics and philosophy.
But the situation has changed in the course of the last decades: The question of how human language evolved has become a Weld of intense research. First, observations on present-day languages and cultures have come to be used by contemporary linguists as indirect evidence in favor of different hypotheses of how human language could have originated. For example, in a number of highly non-trivial works, Haiman (1994, 1998) argues that linguistic signs could have gradually—via repetition and ritualization—originated from motivated symptoms. Accordingly, all linguistic structures can be placed on a functional spectrum, whereby the more iconic a structure, the more likely it is to be at the expressive/ directive end; the more arbitrary a structure, the more likely it is to be at the phatic end of that spectrum (Haiman 1998: 166; see also Givo̒n 1979b, 1979c). Second, new technologies offered by computer sciences, findings in biology, psychology, neuroscience, brain imaging, genetics, palaeo-anthropology and archaeology, together with more detailed information on restricted linguistic systems such as pidgin languages, home sign systems invented by deaf children of non-signing parents, systems arising in specific conditions of first and second language acquisition, etc., have encouraged scholars to make a fresh attempt to look for answers to questions such as the ones listed in (1), and a plethora of books and articles have appeared, offering a wealth of stimulating hypotheses.
The problem now facing research on language evolution is not that there are no answers to the questions listed in (1); on the contrary, there are perhaps too many of them, with the effect that the novice in this research Weld may find it hard to decide which of the answers—many of them mutually contradictory—is the most credible. And there is also a problem with the question that has worried linguists over centuries, namely whether it will ever be possible not only to provide answers but to provide answers based on empirically sound knowledge rather than on conjecture. Even for a question like (1a), which appears to be readily accessible to empirical research, there is no uncontroversial answer, leading Fitch, Hauser and Chomsky (in press: 5) to the conclusion that ‘‘from an empirical perspective, there are not and probably never will be data capable of discriminating among the many plausible speculations that have been offered about the original function(s) of language.’’
