Language Origins
collects sixteen papers which started as talks at the Fourth International
Conference on the Evolution of Language (2002). The papers thus are academic
and present state of the art perspectives on language evolution. The reader
should have background knowledge in cognitive linguistics and cognitive science

 The book has four parts, each
with a short introduction that complements the general introduction. Part I
deals with spoken language and speech sounds, Part II with the evolution of
grammar (in the narrow sense of syntax and morphology). Part III asks what can
be learned from a comparative perspective with respect to other species. Part
IV considers questions of language learnability and the origins of the
diversity of the natural languages we know.

Language evolves in several ways and in several dimensions. On the one
hand we have to have theories how language was able to evolve at all (e.g. by
processes that exapted
[neural] structures already present for other purposes), and on the other hand
we have to have theories of the stepwise evolution of complexity (e.g. from
proto-grammatical sign chains to phrasal structures) and diversity, where
additional historical (non-biological) influences occur as well.

To give an impression of the theories developed I summarize one paper of
each section.

Michael Arbib defends the Mirror System Hypothesis. It explains
the occurrence of shared signs and symbolization as result of complex mutual
imitation of gestures and pantomime, where this kind of mutual imitation depends
on the presence of a neuronal mirror system. Mirror neurons are neurons that
are involved both in one’s own production of (signing or gesturing) behavior as
well as in the perception of such behavior in others. "[B]rain mechanisms
supporting language evolved from the mirror system for grasping in the common
ancestor of monkey and human" (34). Supposedly proto-languages of such
signing behavior in early humanoids preceded vocal language, which then could
exploit the already present neural structures. Early vocalizations ("protospeech")
and sign use then "feeding off each other in an expanding spiral"(22).
Given non-compositional "unitary utterances" then grammatical
structure had to evolve. Part II of Language Origins carries on from
here.

Maggie Tallermann discusses and criticizes Andrew Carstairs-McCarthies’
thesis that the clauses evolved from the syllable. (Carstairs-McCarthies
himself is present in the collection with a newer paper in which he tries to
trace the evolution of morphology to allomorphy first, and ultimately to the
side effects of natural speech production.) 
All these evolutionary scenarios share the assumption of an early stage
of non-grammatical proto-language. One might speculate that grammatical
structures evolved stepwise from morphemes to syllables to phrasal structures.
Tallermann, however, doubts that the syllable was the basis for the clause, the
"apparent similarities" being only "superficial" (125). For
example, syllables have a nucleus, but sentences don’t. Not all grammatical
categories developed from morphemes (since they need not be expressed at all).
With respect to q-roles and argument structures it has to be
said "that the structure of the syllable does not predict or in any
other way explain the kinds of properties that we see universally in argument
structure" (149). What is also missing with syllables but essential for
phrases, of course in the presupposed transformational approach, is movement.
Dana McDaniel continues here. She argues that "movement was motivated by
properties of language production rather than language comprehension"
(155). A system without movement would have been to costly an overhaul of a production
system based on Merge (in the sense of transformational grammar). Movement
answered to our communicative needs (like focus) given the initial
syntax the production system had. Thus, like and others claim with respect to
derivations aimed for other cognitive interfaces, Move and Merge may be the two
crucial ingredients in the development of proper syntax. Given some initial
complexity of language, cognitive and linguistic complexity increased in
co-evolutionary patterns. Therefore the gap to non-linguistic animals widened
in both areas. Nevertheless there have to be homological traits in humanoids
and primates.

Klaus Zuberbühler, in Part III, traces the prerequisites of human
language in the primate lineage. The "brain regions most heavily involved
in language processes in humans did not arise de novo, but evolved from older
structures already present in the primate lineage" (263). Studying the
linguistic capacities of today’s non-human primates may provide us with
insights on the cognitive capacities necessary, though not sufficient, for
human language. Monkeys are able to produce acoustically distinct vocalizations
in response to discrete external events. As recipients at least apes can deduce
meaning from combinatorial rules.

In Part IV Henry Brighton, Simon Kirby and Kenny Smith propose three
principles underlying the view that language adapts (itself) to be learnable.
Languages themselves survive by adapting their structure to learnability
constraints. These reflect, for example, conditions of cultural transmission.
Further on, generalizable forms have a greater chance to survive. The first
learnability principle stresses that humans have a biologically determined set
of predispositions, the 2second that situatedness and cultural transmission are
as important to language as our cognitive faculties, the third that some
aspects of language do not depend on communication or other functions of
language (but are driven either by adaptive constraints or internal dynamics).

A theory of adaptive forms will thus contain sub-theories of the
conditions of preferred reproduction and the initial learning/extracting
devices in children. Another paper in this part by Matthew Roberts, Luca Onnis
and Nich Chater provides evidence from computer simulations of learning
processes that the idiosyncratic features of individual languages (i.e. all the
quasi-regularities where some forms to be expected are absent) are learned by a
processes of building and choosing the simplest explanatory hypothesis.

Each paper in this collection is interesting in itself and provides
clues to the vast topic of language origin and evolution. Since each paper also
provides a Further Reading section Language Origins is a good
place to start also for graduate students and (academic) non-specialists.

© 2006 Manuel Bremer

Manuel Bremer, Heinrich-Heine-Universität Düsseldorf, Germany

Categories: Genetics, Psychology