AUTHOR: MacNeilage, Peter F.
TITLE: The Origin of Speech
SERIES: Studies in the Evolution of Language
PUBLISHER: Oxford University Press
James J. Jenkins, Department of Psychology, University of South Florida at Tampa
In this much-celebrated year of the bicentennial of Darwin's birth, and the
sesquicentennial of his masterwork, _The Origin of Species_, it is a privilege
to review such an important work in the Darwinian tradition. The title of the
book, _The Origin of Speech_, is deliberately and self-consciously chosen to
evoke its inspiration. In this work Dr. Peter MacNeilage has integrated almost
50 years of his own research in psychology and linguistics and his extensive and
critical reading of the research and theories of others to construct an account
of the evolution of human speaking. In addition, although it is not his chief
aim, he hints at the importance of this evolution as a first step in the
development of the account of the language faculty itself.
This is not a timid undertaking. MacNeilage gives fair warning in his opening
chapter that he means to hew to the Darwinian line. ''It is my intention in this
book, to give an account of the evolution of speech that unflinchingly adheres
to a Neodarwinian perspective -- that contends, in short, that speech didn't
just 'happen' by means of a secular miracle but, instead, evolved by descent
with modification in accordance with the principle of natural selection'' (p 17).
MacNeilage's foil throughout the book is what he calls the ''Classical position''
of Plato, Descartes, Saussure and Chomsky. He sees this position as asserting
that speech and language are special forms, unique to humans. Although such
forms are said to be genetically determined or innate in some unspecified
manner, they are held to be without evolutionary predecessors. Thus, MacNeilage
sets up two possible roots for the origin of speaking, one, Darwinian,
(functionalist) and the other Classical (formalist). He takes as his aim the
careful, detailed explication of the former and the rejection of the latter as
MacNeilage attributes the framework of his analysis to Tinbergen, the famous
ethologist. With regard to any function Tinbergen (1952) asked:
1. How does it work? What are the mechanisms?
2. What does it do for the organism? How does it affect the organism's
capabilities to survive and reproduce?
3. How does it get that way in development? What genetic and epigenetic factors
guide its growth?
4. How did it get that way in evolution? How does the history of the species
help us understand the structure of the trait?
MacNeilage believes that a basic biological orientation must be committed to
finding serious answers to these questions. To shrug these questions off by
retreating into the competence/performance distinction and ignoring them as
''mere performance'' problems, is simply unscientific and a regression into the
age-old mind-body distinction.
He carefully outlines his argument and the structure of the work. The book is
divided into seven parts, each consisting of two or three chapters. Part 1, as
suggested above, sets out the philosophical issues between the two orientations
and outlines the author's position. Chomsky is chosen as the antagonist to
characterize the position of many modern linguists in refusing to come to grips
with the evolutionary questions. MacNeilage concedes that Chomsky's position has
softened a little (see Hauser, Chomsky and Fitch, 2002) but sees him and his
followers as still rejecting serious consideration of evolutionary issues and
finding no place for these issues in their view of syntax and phonology.
In Part 2, MacNeilage's characterization of speaking is laid out in detail. The
basic unit is the syllable. This unit constitutes the frame into which the
consonants and vowels are inserted as content. Thus, he calls his formulation
the frame/content approach. The syllable carries the readily observed peaks of
sonority due to the vowel which characterizes the most open position of the
mouth. The consonants in turn result from the constrictions and changes of the
glottis, lips, tongue and jaw preceding and following the vowel opening. The
most prominent motor characteristic of the syllable is the oscillation of the
mandible as the jaw opens and closes during speaking.
Whence comes this behavior in the ''deep time'' of evolution? Unfortunately, until
the invention of sound recording, speaking left no historic traces. Consequently
we must study current members of related species of primates that presumably
departed from our family tree at earlier times. Here we may search for clues in
their nervous systems and their behavioral characteristics.
There is a growing consensus that speech did not arise from primate vocal calls,
many of which tend to be emotional and perhaps reflexive. This view, which
MacNeilage accepts, looks instead to the motor behaviors involved in chewing,
licking and sucking, all of which are biphasic cyclic activities which result in
the subsequent communicative acts of lip smacks, tongue smacks, lip protrusion,
tongue protrusion, and teeth chatters. This approach is congruent with the
current orientation in psychobiology towards the importance of embodiment in the
understanding of perceptual and cognitive abilities. (Embodiment holds that the
systems of the brain out of which the mind arises are in large part generated by
the experiences and capabilities of the body systems operating in the world.)
The Classic view is, of course, that there are innate distinctive features that
are the basic units of speech. MacNeilage regards features as convenient units
of taxonomic analysis but rejects them as innate units of mental structure or
behavioral atoms. He points out that language surveys show no evidence of
converging on a fixed number of distinctive features. Instead, such surveys
reveal an astonishing variety of speech sounds distributed continuously along
various parameters. Fifty percent of speech sounds occur in only one language
and no single sound appears in all languages (see Ladefoged, 2006, and Ladefoged
and Maddieson, 1996). This is an unlikely outcome if distinctive features are
supposed to be innate. Other evidence in studies of speech errors strongly
suggests that speech sounds move as phonetic units to corresponding places in
adjacent syllables, thus supporting both the notion of the syllable frame and
the functional unity of the phoneme.
MacNeilage further notes that the generative approach has no time dimension in
either evolutionary time or in developmental time. The mental systems are said
to appear full blown at some one point in the species and to manifest themselves
in infant development as soon as performance capabilities permit them. Current
genetics knows no parallel to such phenomena in complex behaviors or traits and
current biological thinking is quite incompatible with such a concept. Recent
views on development focus on the organism acting in its environment and
emphasize dynamical systems and principles of self-organization. The genetic
heritage is seen not as a blueprint that specifies everything in advance, but as
a recipe in which the relevant ingredients intermix in sequences that interact
with each other and with environmental events throughout the course of
development. Gene with gene, gene with organism, and gene with environment
interactions are everywhere present.
In Part 3, MacNeilage spells out his view of the developmental sequence in the
child. The basic issue is seen to be Lashley's (1951) old problem of serial
order in behavior. MacNeilage answers this with the frame which he sees emerging
in babbling. The preferred form is consonant-vowel (CV) here just as it is most
common syllable in most languages of the world. He argues that babbling is
already somewhat mimetic i.e., imitative. Research shows that infants match
speaking faces with heard vowels, and spontaneously imitate tongue and mouth
gestures. Further, they show a lower rate of nasals in babbling than would be
expected by chance, echoing the low frequency of nasals in speech inventories
Data analysis shows that in babbling and first words the pairing of consonants
and vowels is not independent: labial consonants tend to go with central vowels
(''ba ba''); coronal consonants go with front vowels (''dee dee'') and velars are
paired with back vowels (''go go''). This correlation is also true of VC pairings.
MacNeilage regards this as evidence that this is a stage of pure frames, not yet
the assembly of independent units.
When frames are assembled in sequences as the next stage begins, it is argued
that such constructions are easier when they start with a pure frame such as the
labial consonant-central vowel syllable. Data from first words support this
claim. Finally, there is a shift away from the high frequency of reduplicative
syllables to variegated syllables characteristic of adult speech. This shift
develops as the learner begins to acquire more words and is forced to achieve
the more varied syllabic production. For this MacNeilage appeals to a general
purpose mimetic capacity in humans (see Donald, 2001). The general argument is
that word growth begins with the traditional baby talk (''mama'' and ''papa'' as the
canonical forms) and then proceeds to elaborate via mimesis under the pressure
to develop more word possibilities.
The origin of words themselves is believed to arise from important social
behaviors such as vocal grooming (see Dunbar, 1996). This in turn leads to the
pairing of sounds with already existing concepts. There is ample evidence of the
existence and use of concepts in chimpanzees and gorillas so it is a question of
environmental pressures, the general purpose mimetic abilities and the ability
to produce varied utterances that made the first words possible. MacNeilage
suggests that the big breakthrough took place in the family unit as a result of
these abilities and the increased time of nurturance required by the neonate,
compared to related species. Following Jakobson (1960) and Murdock (1959) he
regards the terms for mother and father as candidates for first words and first
contrasts; nasal stops for mother and oral stops for father. There is abundant
evidence that these terms are omni-present in today's languages and, indeed, it
appears that they are frequently reinvented as languages change over time.
Part 4 is devoted to brain organization and the evolution of speech. This
section begins with a tutorial on the brain that will help the uninitiated
follow the discussion. Current literature suggests, contrary to earlier beliefs,
that related primates have predominantly left hemisphere dominance for ingestive
behaviors and controlled routine motor behaviors. The great apes appear to be
both right handed and even more strongly right footed. Examination of the
literature on our primate relatives suggests that the area governing vocal calls
in other primates is not located in a homologous area to that which governs
human speech. However, communicative non-emotional behavior in primates is in
the left hemisphere both for production and perception of the various
oral-facial gestures (lip smacks, tongue smacks etc.). Further, recent
neurological evidence has revealed that there are mirror neurons in this region
involved in both ingestive behaviors and visuofacial communicative behaviors
(see Rizzolatti and Craighero, 2004).
In further discussion of the neurological capacity of humans, MacNeilage finds a
candidate area for the generation of frames, the supplementary motor area,
located above the sensory-motor strip in the left hemisphere. When this area is
galvanically stimulated in brain explorations in humans, it yields repetitive,
cyclical, motor and speech behaviors which last beyond the duration of the
stimulation. No other area of the brain is known to have such response to
stimulation. These findings and the steadily increasing capacity for general
purpose mimetic ability furnish the neurological foundation for the emergence of
the articulatory skills found in speech.
Part 5 is devoted to a critique of generative phonology and its inability to
deal with either the development of speech in the human child or the origins of
speech. The shortest form of the argument is that phonology is basically
descriptive but unscientific, looking for regularities and then using the
regularities as ''rules'' to ''explain'' the same data. He notes again the lack of
real cross-language solutions to the nature and number of distinctive features
and comments on the poverty of the notions of markedness. He accuses linguists
of accepting phonetic data when it confirms their beliefs and of rejecting such
data as ''mere performance data'' when it disagrees with their ''rules.''
Part 6 tackles questions concerning the nature of sign language. Can the
existence of sign language be taken as evidence that the externalization of
language is modality independent as Chomsky asserts? MacNeilage examines what he
takes to be the non-parallel characteristics of vocal-auditory language and the
manual-visual form and concludes that they are fundamentally different. The
speaking code is linear and sequential. The manual system is simultaneous.
Although both are babbled under appropriate circumstances, they are not
synchronous in onset or in progressive development, nor is their recognition
based on the same rhythmic structures. Further, comprehension of sign language
seems to depend to a greater extent on right hemisphere properties than does speech.
Part 7 assays a review of the argument in terms of Tinbergen's fourth question:
How did speech get there phylogenetically? First, MacNeilage points to the
unlikelihood of direct genetic control of any aspects of universal grammar or
any other specific gene-to-particular-phenomenon of language. Although the gene
FOXP2 was briefly considered to be an instance of such a connection, it now
appears to have a much more general sphere of influence, general motor control.
No other candidates are in sight.
Bird song gives evidence for innateness in the selection and vocalization of
specific songs, but there is no evidence that humans have anything like a
parallel language-specific innateness. Birdsong does, however, show a
frame/content organization which suggests some form of convergent evolutionary
device to solve the problem of sustained, repeated vocalization. Humans do, in
addition, show innate imitation of facial expressions and, particularly,
movements of the tongue and mouth.
Over evolutionary time the oral-facial gestures and phonation are believed to
serve in vocal grooming, facilitate infant-parent vocal interaction and
labeling, and eventually lead to the coupling of sounds and concepts. This was
the monumental social discovery that ultimately was transmitted as part of
culture (a meme) and replicated itself through the general mimetic capacity of
the species, perhaps even giving impetus to the enhancement of working memory
capacity through the phonological loop (Baddeley, 1986). The overall picture is
one of bodily functions that permit certain kinds of actions being recruited in
the service of social needs and consequent selection advantages in adaptation.
All of this leads to further interactions of genes and memes and to the eventual
result of language.
Finally, MacNeilage concludes: ''I hope the Darwinian approach to the evolution
of speech I have presented here will become part of the framework enabling the
phonological component of speech to enter the mainstream of modern science where
it deserves to be, considering its importance in getting us to be who we are''
Why should linguists be interested in this book? First, it is a serious
scholarly work; a study integrated across many areas in linguistics, psychology,
ethology, neurology, genetics, and epigenetics. It includes a valuable set of
references in these fields (30 pages, approximately 500 citations) to which the
reader is directed for further information and evidence concerning the author's
claims. In this reviewer's opinion the book is much more soundly based in data
than most works in ''evolutionary psychology'' that are being offered today.
Second, it provides a plausible and persuasive account of the origin, evolution
and development of speaking. Many current linguists (following Chomsky) ignore
the challenge of accounting for the origins of language or discount the problem
as being completely and permanently beyond investigation. MacNeilage argues that
we must take a biological approach and concern ourselves with these questions.
In his view it is a question of whether linguists are going to win a place for
their field as a modern science or languish in the role of describers and
classifiers in the old Linnaeus tradition, leaving others to explain the
regularities that they find. Recent advances in genetics, microbiology and brain
sciences are revolutionizing our understanding of behavioral matters. The fields
are being massively rewritten every decade. Current literature can scarcely keep
up with the discoveries being made in these fields. Linguists must not fall behind.
Third, this book explicitly challenges much of current linguistic thought and
practice at a basic level. It argues that linguistic explanation is often
fundamentally circular, a process that many psychologists and linguists have
objected to for years. The linguist searches for regularities and, having found
them, appeals to his generalization in the form of a rule as the explanation for
the data. The observation and the generalization are of key interest, of course,
but the subsequent ''explanation by rule'' is non-causal. MacNeilage argues
against the casual acceptance of distinctive features and markedness, as if they
were universal realities of speaking. In MacNeilage's opinion, surveys of the
world's languages fail to confirm the universality hypothesis. He finds little
real support for modern phonology and recommends more attention to phonetics and
less to abstract, supposedly innate, categories that are sometimes only
distantly related to observable data.
Parts of the book, though interesting, seem to stray away from the central
argument. Parts Five and Six are directed at counter-arguments that the reader
may or may not be concerned with. For the reader interested in the evolutionary
account, the first four parts are the crucial ones.
All accounts of evolutionary development are to some extent ''Just So Stories.''
Rigorous proof is not possible in most cases of complex traits. Every story must
assemble what evidence it can find into a plausible account. This book does a
masterful job of assembling and interpreting all of the evidence we have
concerning the evolution of speaking. In the long run it may not be the final
word, but until we have a better story, this is the one that must be the prime
Baddeley, A. D. (1986). _Working Memory_. London, Clarendon Press.
Donald, M. (2001). _A Mind So Rare_. New York: Norton.
Dunbar, R. I. M. (1996). _Grooming, Gossip, and the Evolution of Language_.
Cambridge, MA: Harvard University Press.
Greenberg, J. H. and Jenkins J. J. (1964). Studies in the psychological
correlates of the sound system of American English. I and II. _Word_. 20, 157-177.
Greenberg, J. H. and Jenkins, J. J. (1966). Studies in the psychological
correlates of the sound system of American English: III and IV. _Word_, 22, 207-242.
Hauser, M. D., Chomsky, N. and Fitch, W. T. (2002). The faculty of language:
what is it, who has it, and how did it evolve? _Science_, 298, 1569-1579.
Jakobson, R. (1960). Why ''Mama'' and ''Papa''. In B. Caplan and S. Wapner (eds.)
_Essays in Honor of Heinz Werner_. New York: International Universities Press,
Ladefoged, P. (2006). Features and parameters for different purposes.
Ladefoged, P. and Maddieson, I. (1996). _The Sounds of the World's Languages_.
Murdock, G. P. (1959). Cross-language parallels in parental kin terms.
_Anthropological Linguistics_, 1, 1-5.
Osgood C.E. and Sebeok, T. A. (1954). _Psycholinguistics A Survey of Theory and
Research Problems_. Baltimore: Waverly Press.
Rizzolatti, G. and Craighero, L. (2004). The Mirror-Neuron System. _Annual
Review of Neuroscience_, 27, 169-192.
Tinbergen, N. (1952).Derived activities: Their causation, biological
significance, origin and emancipation during evolution. _Quarterly Review of
Biology_, 27, 1-32.
ABOUT THE REVIEWER
James Jenkins is Emeritus Distinguished Research Professor of Psychology at the
University of South Florida. He was one of the original group, calling
themselves psycholinguists, who met at Indiana University in 1953 and brought
forth the first survey of psycholinguistics (Osgood and Sebeok, 1954). He is
probably best known to linguists for his collaboration with the distinguished
linguist Joseph Greenberg in studies of the psychological correlates of the
sound system of American English (Greenberg and Jenkins, 1964, 1966). In the
last 30 years his work has largely been concerned with speech perception,
especially perception of American English vowels by native speakers of English
and by second language learners.