176
Robert G. Bednarik
Anthropos 89.1994
at least as old as it is in Europe, and in Asia the sit
uation remains unclear, although the oldest known
traditions are non-iconic. On this very preliminary
basis it would appear that the cognitive processes
leading to iconic graphic art were independently
duplicated in various parts of the world (Fig. 2).
Years BP I Americas
Eurasia
Australia I I Africa
10 000
30 000-
I Non-iconic '-g ¡
I graphic art ! g i
1 t ? r?
50 000"
100 000-
150 000"
200 000"
•&
Non-iconic
graphic
art I
I
Fig. 2: Very preliminary model of early art evolution, based
on the current evidence.
The intercontinental migration patterns during
the Middle Palaeolithic as they are currently ap
parent are of interest here. With first landfall in
Australia before 50,000 BP (Roberts etal. 1990),
and perhaps much earlier, and a similar order of
magnitude for the human settlement duration in the
Americas (Guidon and Delibrias 1986; Bednarik
1989a), it is to be assumed that these first seafarers
possessed traditions of nonfigurative mark making.
That they possessed advanced language is accepted
even by Davidson and Noble. This Middle Palaeo
lithic dispersal model might help to explain the
similarity of the early petroglyphs in Africa, Asia,
Australia, and the Americas. However, there is a
second possibility to account for the uniformity,
perhaps in conjunction with the first.
The prefigurative art of the world, I have
claimed for more than ten years, is derived from
phosphene motifs, and it therefore consists of ar
rangements and combinations of a known series
of form constants (Bednarik 1984, 1986, 1987).
These may occur singly, or as parts of elaborate
“mazes” or geometrical arrangements. Chase and
Dibble (1992: 50) have observed that the phos
phene theory has the distinction of being testable,
of being accessible to refutation. It has remained
the only hypothesis of art origins that is scientifi
cally testable, yet since it has been proposed it has
not been refuted, nor am I aware of any refuting
evidence.
The basic elements of all archaic rock art are
dominated by circles and curvilinear motifs, such
as multiple arcs, concentric circles, meandering
lines; by convergent lines motifs, radial designs,
dot arrangements, sets of parallel lines or grids,
and zigzags or wave lines. Variations and interme
diate forms are common: for instance, the con
vergent lines motif (Rosenfeld’s [1981] trident;
Conkey’s [1978] arrow tip, elsewhere described
as vulva, bird track, etc.) can be transformed into
a radial one by three different mutation processes,
which are evident from motifs of Australian archa
ic linear petroglyphs. Different motif types may
be combined, e.g., sets of parallel lines or grids
enclosed by circles (in Australia), radial design
enclosed by a circle (North America; Bednarik
1988c), or the cross, a radial design, enveloped in
a surround (e.g., South America; Bednarik 1988d).
The phosphene hypothesis is in no way related
to the recent pseudoscientific attempts to interpret
all rock art as shamanic or as trance-induced (Bed
narik 1990c), but it notes instead that precisely
the same range of about fifteen phosphene motifs
(Kellogg etal. 1965) determines all art produced
by children before they draw figuratively, i.e., be
fore the age of four years or so. The findings of
Knoll, Kugler, and colleagues have been indepen
dently confirmed by others who had been unaware
of their work, such as van Sommers (1984) and
Fein (1976). If we now return to my postulate
that art externalizes human concepts of reality and
communicates awareness of perceived reality be
tween humans, we see that, so more “developed”
an art is, so less it can tell us about the past
human quest for comprehending reality. If it were
our aim to explore objective reality, we would
first have to determine how anthropocentricity (the
human concept of reality) was conceived. The
arts of “pre-iconic humans,” at both the ontogenic
and the phylogenic levels, are apparently simi
lar, and they may be capable of providing some
