then shaped and modified by the supralaryngeal vocal tract,tongue,nasal
and oral chambers,and lips.These anatomical structures and chambers
have long been analyzed in living humans (Negus 1929),so that we know
exactly where and how various sounds (formants) are produced in the
vocal tract (see Borden,Harris,and Raphael 1994).It becomes quite
another matter to extend similar kinds of analyses to the fossil record
when,at best,essential parts are estimated from contingent bony areas.
In fact,no patent skeletal structure directly predicts laryngeal anatomy
or the shape and size of the supralaryngeal vocal tract,tongue,or lips.
Correlations exist,but these are essentially inferences that suggest
anatomical structure,and not necessarily the specificity and effectiveness
of linguistic functions leading to speech.
Other problems relate to the incompleteness of the fossil record,both
in the uneven chronological representation of various hominid forms and
the fragmentary nature of specimens that constitute the fossil record.
These shortcomings are constantly being overcome by field work.For
example,the hyoid bone that sits atop of the larynx was recently added
to the inventory,but only western Eurasia and later aspects of the fossil
record are represented.It is also important to recognize that anatomical
regions used to infer speech ability (e.g.,the external cranial base) are
commonly missing in fossils or,if present,are often distorted or incom-
plete.The consequence is that in some cases a fragmentary,distorted area
is estimated in the original fossil,and this estimated region is used to
“estimate”further the shape of the supralaryngeal vocal tract.Such
extrapolations are far removed from the bony anatomy of the fossil,let
alone from the soft tissue that made up the vocal tract when the fossil
was alive.They are also susceptible to fundamental errors in anatomical
detail,since it is not always a straightforward matter to repiece or recon-
struct a complicated area such as the external cranial base.Given all this,
it should be apparent that it is not a simple matter to marshal the fossil
record and review various anatomical regions to determine when speech,
let alone language,arose.Further complicating the issue of origins is the
fact that it is not even clear that modern human speech abilities arose
only once,since a few markers thought to be associated with a modern
vocal tract occur in fossils considered (without dispute among paleoan-
thropologists) to be offshoots from the main branch leading to modern
Homo sapiens.
Even with a perfect fossil record and perfect correlation between bony
anatomy and soft tissue sound-production units (neither of which
exists),problems remain in determining when language arose.In our
view,this is because language is primarily neurologically based and not
simply predicted by the ability to produce the necessary sounds.It is
important to remember that humans can conduct language without a218 David W.Frayer and Chris Nicolay