Copyright © 1998 by Kevin Sharpe. All rights reserved.
In The Artefact 21 (1998), pp. 95-104.

KEYWORDS: Gallus - Communication – Finger flutings – Mnemonic – Notation

Kevin Sharpe, Mary Lacombe, and Helen Fawbert

Abstract. Sandor Gallus highlighted the significance of the line markings in Koonalda cave in 1956. Further examples, in caves throughout southern Australia, have led to a reassessment of Gallus's suggestion that the lines, dated to the late Pleistocene, were a means of communication – a hypothesis that contravened established opinion. Using the hypothesis of Gallus as a starting point, we describe our hypothesis, experimental method, and results. We suggest that the lines are a mnemonic notation system and, by their artificial reproduction, we will pose questions regarding form, technique, and structural consistencies that will help to elucidate their meaning.

The line markings first identified in Koonalda Cave below the Nullarbor Plain, South Australia, are the subject of a contentious debate in rock art research (Bednarik 1986a). Similar markings occur at sites throughout southern Australia and Europe (Bednarik 1986a; Sharpe and Sharpe 1976, 1978; Wright 1971a, Bahn1991: 97f). The Koonalda markings cannot be dated directly. However, charcoal samples taken from the areas in which the markings occur, have yielded late Pleistocene radiocarbon dates of around 20,000 BP (Gallus, in Wright 1971). Many researchers consider finger flutings one of the earliest forms of intentional marks on rock surfaces (Bednarik 1990).

Koonalda Cave has been the destination of explorers since the latter part of the C19th (Flood 1997: 25), but the first archaeologist to see the cave and highlight its significance was Sandor Gallus in 1956. Though he faced an established opposition (Flood 1997: 30, Aslin et al. 1985: 71), Gallus (1968a, 1968b) proposed that the extensive markings found in the cave were prehistoric finger flutings. For the interpretation of the line markings (a concise description of which can be found in Gallus 1968a), he favored the hypothesis of communication and urged their further analysis. His excavations (Gallus 1971) showed that flint mining occurred in the cave; an activity which was thought to be contemporaneous with the finger flutings. However, if there was a connection it was unlikely to have been a utilitarian one. There are ample supplies of flint in the first main chamber of the cave

The general rejection of not only his theories but also the evidence for Pleistocene occupation was most likely to have been connected to the prevailing opinions in the 1970s (Flood 1997: 32) rather than to the evidence which he presented. In his initial work, he asserted that the intention of the line markings was display, as opposed to the superimposition of animal upon animal observed in the caves of Western Europe. This style in Western Europe might suggest that the physical act of drawing was more important to those involved, than the resultant remaining image (Gallus 1977, Marshack 1997: 64f). Gallus (1968b: 6) observed that in Koonalda, the lines are 'clear and visible . . .an externalism in order to communicate'. He hypothesized that the abstract lines on the walls were expressions of verbs - the part of a sentence difficult to express pictorially (Gallus 1971: 128). In his later work, he combined his theories with those of Webb (pers. comm. 1973-1974, in Gallus 1977) to explain the origin of line markings in general, although concentrating on those at Koonalda. He concluded that, 'the physical act that created the linear externalizations' was 'not only of primary but … of total importance' (Groenfeldt 1985: 30). Gallus opened areas of investigation, which encompassed psychological factors in interpretation and pushed against the established Eurocentric analysis of archaeological evidence. In recognition of his discoveries and contribution, the 'Gallus Site' became the name given to further archaeological investigations in Koonalda (Edwards and Maynard 1969).

This paper describes a systematic analysis to facilitate constructive approaches to the markings. We propose a method to carry out research on the markings themselves and offer a hypothesis in an attempt to elucidate their meaning. Drawing chiefly from our familiarity with the finger markings in Koonalda Cave we propose they are an early notational system, and, as such, are cultural and stylized forms of interpersonal communication. Many of these caves including Koonalda also contain groups of markings scratched and engraved onto walls, the ceiling, and floor boulders. Many of these marks are natural or made by animals. Bednarik (1986a, 1994b) has comprehensively discussed their origins. The flutings generally occur on (formerly) soft calcite deposits known as moonmilk or Montmilch. Moonmilk is comprised of a microscopic, fiber-like lattice of calcite crystals, which can absorb large amounts of water (Bednarik 1990). Sometimes the flutings are termed meanders (Marshack 1977), or macaroni, which pertains particularly to finger lines. They appear to be non-iconic, largely unstructured, and offer little scope for conventional stylistic analysis (Marshack 1977, Aslin et al. 1985, Bednarik 1986a).

An experimental technique for analyzing prehistoric finger markings
Our proposal, that the finger markings found in caves are a mnemonic form of writing, is one that has rarely been endorsed (Bednarik 1986a, Flood 1997). However, work proceeds on methods that differentiate between and falsify certain interpretations. Therefore, before this paper launches into the discussion of our hypothesis and those proposed by others, we would like to present and discuss the methodology of an experimental approach. This extends d'Errico's (1991, 1992, 1995) pioneering techniques of examination of markings on portable objects, to those found on the caves themselves.

It is most probable that early writing systems were peculiar to a specific geographical area (Salzmann 1993). Therefore, markings within one cave or a group of caves may display patterns or similarities. Finger markings with a discernible similarity to those found in caves from a diverse geographical area may not necessarily be representing the same language. Smith (1992: 28) theorizes that messages encoded in style may 'serve to either integrate groups by visually reinforcing bonds through stylistic homogeneity or to differentiate groups by marking, maintaining and furthering differences through stylistic heterogeneity'. People living in coastal environments - such as probably was the environment of those who used Koonalda Cave - are thought to be primarily concerned with maintaining rights to their own resources (Smith 1992). This, therefore, emphasizes the probability of a small population who had cognitive access to the finger markings and engravings at Koonalda.

Lastly, a subtle division exists between mnemonic markings and communication systems. This lies in the individual nature of the mnemonic marking, designed for remembering an event or story, but unconnected with any standardized symbol. Did writing start as a mnemonic form, shared by a small group of people? In all probability, it was limited both in content and in users, and this limit differentiates it from a writing system. The line markings are analogous, unlike the average writing systems whereby one can differentiate one symbol from another. In the methodology we propose, we hope that the duplication of the line markings will allow us to consider our methods in view of the results. This will enable us to devise further experiments to answer the questions that arise.

Finding a medium that would allow duplication of the line markings was a first concern. We chose plaster of Paris due to its accessibility and its similarity in appearance and consistency to montmilch. The durability of plaster of Paris, once hardened, gave the opportunity to analyze the markings over a period. Using a combination of one part water and two parts plaster of Paris (throughout the experiment), the solution was poured into four 30 x 46 centimeter aluminum trays. This recipe resulted in the plaster being too wet to hold a marking, so a time lapse of 15 minutes was necessary before marking the plaster.

Marker #1
Trays of plaster of Paris were placed on a table. The marking person (male) made the lines by dragging his fingers over the plaster, standing upright. Normally in caves the markings are on a vertical surface in relation to the marker. However, the markings occur overhead also. One adult male made four sets of finger markings to determine whether markings made by one individual were distinctive. If the markings have a common visual thread, defined clusters of markings would suggest the separation of the cave into areas used by different individuals. The finger markings were made in four sets: two sets of vertical markings made with the right hand and two sets of horizontal markings, cutting across the vertical markings. These markings were labeled A, B, C, and D:
A- representing the first marking on the left half of the pan;
B - representing the vertical marking on the right side;
C - representing the top horizontal marking;
D - representing the bottom horizontal marking.

Five fingers on the right hand were used to make marking A. The thumb marking was particularly noticeable due to the distinctive sharp line made by the nail. When all five fingers are used, the lateral position of the thumb ensures that it is dragged sideways over the plaster. The fingernail is therefore, pressed into the plaster producing a sharp line. The thumb mark also starts lower than the middle three fingers. This would help to determine laterality.

The right hand of the same person made mark B, without using the thumb. The spacing between the fingers of both sets was not identical and there were no easily perceptible similarities between the sets of lines made by the same person. Various pressures on the fingers caused the lines to be deeper in some places than in others. The little finger did not make as deep an impression as did the first three fingers. The initial conclusion suggests that it is difficult to determine (using the naked eye) if the same person made a particular set of line markings. Future experiments might look into the possibility of microscopic analysis of line markings. However, there is doubt over the survival of minuscule differences on degraded late Pleistocene surfaces.

Dragging the fingers horizontally across the vertical lines in C and D places the fingers in a different position with different pressure. When moving from right to left with the right hand, the thumb is not used. The twisting of the hand for dragging horizontally causes the thumb to be at too great a distance from the surface to cause a mark.

On the edge of the finger markings, plaster of Paris built up due to the gouging action of the fingers. Moving horizontally over a vertical marking also caused build up of plaster. This pronounced effect, observed in many of the markings, indicated that there was a clear beginning and end. The build-up of plaster at the end of the mark produces this effect by the fingers removed from the surface. The build-up was distinctive at the end, whereas the starting points of the finger markings were smooth.

Marker #2
A female then duplicated the same lines, in an attempt to compare the two trays of finger markings and observe any differences. Many of the same characteristics were apparent as with the male. The thumbnail cut a line into the plaster. The sets of lines made by one person were not noticeably similar and the distance between fingers changed with each set of lines. However, set B made by a female were smaller and cut less deeply into the plaster than those made by the male in set A.

Use of left hand in marking (marker #2)
When looking at the marking produced by the left hand, there were no obvious physical differences than those lines made with the right hand. However, using the thumbnail marking and the lower starting point of the thumb in relation to the other fingers might help determine the laterality of the markers. Laterality might have some importance. Marks made using the left hand might have had a different meaning from those made by the right hand.
The relationship of the marker to the marked surface
We based the final experiment in the series of four on the observation that most cave walls were directly in front of the person making the lines. We placed the tray containing the plaster of Paris in an upright position. Therefore, when sitting on a chair, the wall, (as represented by the tray) was just below shoulder level. Thus, the positioning was more consistent with the relationship of the marker to the cave wall. Initially we observed no difference between the same marks made in a pan lying on the table. However, the thumb made little or no impression when pulling the right hand from right to left. Also, there was some general discomfort felt when moving the right hand in that direction. This led us to speculate that the comfort or maneuverability of the body might impose constrictions on the size and type of line markings made.
In exploring the relationship of the body to the cave wall, it was not possible to use plaster due to the restrictions imposed by the size of the pans. A blank wall was used to determine possible lengths of line markings given certain conditions. A person standing could reach over their head to begin a marking and could pull down to a short distance past the shoulder. The fingernail, rather than the soft flesh of the finger, produced the line if one continued down whilst still standing. Continuing to stand in one spot, the marker could make downward strokes, if right handed. These were approximately 30 centimeters beyond the left shoulder and 60 centimeters beyond the right shoulder. A left-handed person would reverse these lengths. In general, the cluster of marks made by one individual standing still would encompass a maximum area, 60-90 centimeters long, and 90 centimeters wide. If the person continued to squat whilst making the lines, this would lengthen the marks; or, if the person moved sideways, widen the area of the marks. Sideways movement may have caused a disruption of some sort in horizontal markings.

Given these two constraints on the size of the sets of markings, it allows for the possibility of defined clusters of line markings. These could be used to determine if there are any similarities or patterns within, or between them. This would require further fieldwork, in which photographs of sets of lines falling within these dimensions could be examined for similarities.

Further experiments
A second set of experiments was performed using plaster in trays with a dimension of 60 x 60 centimeters. This allowed for more space in which to make and study the lines.

The initial sets had resulted in a recognizable end mark to each line marking. Material built up on the fingertips when they gouged in to the plaster. A noticeable mark remained when the fingers were taken from the plaster, indicating the end of the line. On the second set of lines, it was found that by pressing a line instead of dragging a line, little plaster was picked up and the beginning and end of each was undetectable. The technique used to make marks on the plaster can control the amount of build-up. One of the participants removed the build-up of plaster at the end of the fingers presuming that it would interfere with the production of the second set of lines. However, this made no difference to the result. Lines made without removing the build-up from the fingers are not discernible from those made with clean fingers. The resulting build-up might have had a function. The use of natural substances (clay, ash, and mud) for ceremonial body marking is well documented in anthropological literature (Layton 1991). Finger lines may be only the remnants of this activity. However, our experiments have shown that very short marks could pick up a substantial amount of material. Though ceremonial activities might have occurred in the cave, it is unlikely that such a volume of montmilch would have been required, or obtained so elaborately.

One experimental pan was used to determine which fingers had the most control and were easiest to use when making finger markings. Our results showed that:

  • The index and middle finger were used with equal ease.
  • To use the third finger, the thumb and little finger caused discomfort.
  • The participants found that to use three or more fingers when making vertical motions was the easiest movement.
  • Horizontal motions with a vertical inclination were easier to make than horizontal motions straight across.
  • When making horizontal marks, both markers felt more comfortable moving from left to right with right hand, and right to left with the left hand.
  • The use of only one or two fingers facilitated the drawing of circles and curves.
  • The form of the hand lent itself to curves and intricate pictures when only one or two fingers were used. This might be an important point to consider in the production of line markings. When sand pictures are made, the figures are drawn with elegant movements of the hand to use the natural shape of the hand in as many ways as possible (Layton 1992b: 138).

Making horizontal and vertical lines
Another experiment had the horizontal markings made first, with the vertical markings made on top of them. A tray was marked going from bottom to top and left to right. The premise behind this is that writing tends to be standardized, often in terms of ease of bodily movement. Both markers felt it more comfortable to drag their fingers from top to bottom and make the horizontal lines second. Dragging the fingers upwards changed the length of the markings in terms of the relationship to the surface. Below the shoulder, the fingertips pressed into the surface; above the shoulder, the fingernails made the markings.
Experimental conclusions and possible field tests
The plaster at the end of fingertips made it uncomfortable to continue making lines for one of the markers. The composition of montmilch may cause this same condition though it is possible (certainly using plaster of Paris) to control the amount that accumulates on the fingertips by using a different finger position or pressure. Neither marker was naturally left-handed. Therefore, the left-handed marks might be different from one who normally uses the left hand.

We suggest that a number of field tests might shed light on these markings. Studies of the markings in caves such as Snowflake, South Australia, or Koonalda would be invaluable. Snowflake contains line markings on montmilch down to the bottom edges where the wall touches the floor. By observing different parts of the cave, it could be determined in which direction the finger markings progressed and from where. From this premise, it would be possible to define whether moving the hand in a downward or upward motion was favored. If the marks appear constructed in a similar way, our hypothesis of a mnemonic writing form gains support. A standardization of marking may have been required, despite the difficulty of certain movements involved. Further conclusions, discussed in the next section, relate to other potential hypotheses proposed for line markings.

Line markings, internal analysis, and systems of notation
We have proposed a hypothesis for line markings, as mnemonic notation systems. Our experiments have been inclined towards the markings at Koonalda, the meaning of which, we hope to elucidate. However, we suggest a series of propositions that break this hypothesis down into a set of testable consequences. These may potentially apply, to other line marking sites, and particularly those in south Australia. This, however, represents only one possible approach to look at line markings as systems of notation.

Many hypotheses have been proposed for line markings (e.g., Clottes 1986, Waller 1993). These have consequently been catalogued and analyzed (Bahn 1988, Bednarik 1986a, 1994). Often, these universal explanations appear to force the empirical data into the hypotheses. In 1988, Bahn suggested efforts should be concentrated towards data collection, with interpretation delayed until the information pool has expanded considerably. Ucko, however, has highlighted the subjective influences even on data collection (1992). Though subjectivity in terms of data collection will frequently hinder the archaeologist, we hope that our approach is relatively free of subjective influence.

The problem of subjectivity in the recording of Australian rock art has long been recognized (Layton 1992a, Marshack 1989). Even an attempt to record art accurately must be influenced by the art tradition of one's own society (Layton 1992a: 145). Art in this context is used as a convenient label for the deliberate marks and images found on rock surfaces. We do not wish to imply any western views on the definition of art although subjectivity is often difficult to avoid. Any theories archaeologists entertain about the line markers' meanings and intentions affect how they collect data and what data they collect. Ideal conditions for objective data collection seldom do exist. Data are laden with theory. We want to emphasize the empirical nature of the theories; they must lead to questions researchers can ask of the markings and they must attempt to synthesize the more-or-less objective data already known. Gallus (in Groenfeldt 1985: 29) warned against anachronistic approaches to subjects under observation. 'In the center of an interpretative activity stands the modern observer who applies his own cognitive universe, the categories of which, he feels, uncover reality¼ . The emphasis today should be on understanding, not on a reorganization of traditional ways of thinking'. Since art is a cultural expression, the interpretation is difficult for those outside the culture (Layton 1991: 171f). Therefore, there are problems inherent in our accurately attempting to translate any forms of non-western art (Layton 1992b). An attempt can only be made, through anthropological study, to build up a picture of the way people in another society view the world. This is essential in understanding the art of that society (Coote and Sheldon 1992: 10). However, there is a problem. To what extent can we rely on living indigenous people to elucidate the meanings of ancient forms, whatever our definitions of art?

To begin with a very basic premise, we suggest that the line markings had a purpose and significance for their makers. Aesthetic considerations may have played some part due to the situation of the lines around naturally present holes, cracks, shells, and other features on the rock surface, apparently emphasizing the sculptural form of the rocks. However, a parsimonious explanation, such as aesthetic appreciation or a desire to embellish (Hamilton-Smith 1986) (based on the application of Occam's razor) is too simplistic. The desire to make a simple phenomenon more complex and stimulating does not fully explain the extent to which people of the late Pleistocene period used finger markings. A large-brained mammal, in need a constant mental stimulation, would soon tire of this repetitive behavior had it no other meaning or intention.

A further observation in support of a form ritual behavior is suggested by the difficulty in reaching the interior of the cave. The upper chamber lies at nearly an hour away from the surface. It is accessible using steel ladders, cut paths, lights (the upper chamber is pitch black), and other modern paraphernalia. Access and maneuverability once inside is difficult. It is unlikely that the line markers would have risked their safety to reach a surface to doodle on, if they had no other more serious intentions for being there. Further, the prehistoric visitors cleared floors in the surface rubble of the upper chamber, placing stones neatly around the outside of the floors. This suggests intentioned, serious, and continued use of the cave.

The sea level during the late Pleistocene was about 90 meters lower than today along the coastal area directly to the south of Koonalda cave, exposing a coastal plain lower than the one in which Koonalda is situated (Wright 1971a). Instead of Koonalda sitting on the edge of the 20 kilometer-wide coastal belt, it sat 180 kilometers or so beyond it (Wright 1971a). Study of sediments washed into a lower portion of the cave suggests a similar climate and flora 20, 000 years ago as today (Frank 1971). We could therefore expect the plain south of today's coast to be as arid and inhospitable as the plain north of today's coastal belt. If the Koonalda people of 20, 000 years ago were anything like the Mirning, who in recent times inhabited the land adjacent to Koonalda, the only part of the region on which they could live was the coastal belt. The land beyond the coastal belt was seldom visited and in fact aroused fear (Wright 1971b). Thus, the nearest habitable land for the Koonalda people possibly lay around 180 kilometers south of the cave. The line markers would have traveled a long way over inhospitable land to reach the cave. These factors, combined with the difficulty of access, suggest a considerable motivation to reach the cave.

Clottes (1986:161), discussing the lines found in Réseau Clastres (Pyrenees), suggests that the lines might have been formed (deliberately or instinctively) to 'leave a trace of one's passage inside a remote cavern'. This may be applicable for Réseau Clastres, but does not warrant consideration for the extensive coverage of finger markings found at sites such as Koonalda. If, as our results indicate, there is a formal structure to the markings, this would be a highly elaborate way to leave one's trace.

Psychological models for motivation would suggest that the marks were made to make the markers feel safe or to assert themselves in a potentially threatening environment. Bednarik (1986d: 168) suggests that the assertion was that of the 'exploring, stimulus-seeking human mind', rather than that of humans in an alien-environment. Waller (1993) suggests that sound reflection could explain the content and context of rock art. Bednarik (1994c) has refuted this to some extent. However, Waller's theory might be more useful when considering the appearance of finger tracings in specific areas of caves. Acoustics would have been important in the oral aspect of myth-stories.

Bednarik (1986c, 1994b) dicusses the emergence of perceived symbolism in the human cognitive system, as relating to the survival value of visual ambiguity. He suggests that the recognition of resemblance between unrelated objects may have provided the necessary cognitive stimulus for hominids to form analogies between a mark and a taxonomically unrelated object. These marks were made with materials that were easily obtainable, and the information contained in them was limited to the persons of a particular geographic and cultural region (Goody 1968). From this premise, one can suggest that the line markings created by people of the late Pleistocene were used as a form of communication. The conveyance of information might have been achieved using simple signs, pictures, or markings which linguists describe with such terms as communication inscription, cueing system, or notation system. Writing as we know today in the modern world is probably the result of a gradual evolution from these forms. We suggest that looking for structural consistencies in the line markings might be more helpful than attempting to find direct analogies with modern written script.

Line markings as story telling
Gestures, inflections, tone, eye contact, head movements, pitch, loudness, and body language add a rich meaning to words (Allman 1994). The live voice lies worlds away from the printed page and spoken words convey meanings absent in printed words (Bolinger 1980). A writing theory supported by Wurth describes gestures as writing in air, and many written signs as fixed gestures (Vygotsky 1978). Our reliance on a sound-symbol relationship in writing makes it all the more difficult to think of a different type of association. When we look at line markings, therefore, we might consider some other relationship with language than the one we assume for our own. 'Visual designs function as self-contained systems of communication acting directly on the beholder and are not fully translatable into linguistic messages', Groenfeldt (1985: 21).

Writing, or more accurately, notation can serve different purposes for those who do not necessarily follow the concept of modern writing. An experiment by Thorndike asked small children who could not write to remember difficult things (Vygotsky 1978). The children created apparently meaningless and undifferentiated squiggles and lines, but would read them as though they were spelled-out words. The children referred to specific marks and repeatedly indicated, without error, which marks denoted which phrase. The marks became mnemonic or remembering symbols.

Educational experiments conducted with Aboriginal children (Seagrim and Lendon 1980) have shown that there appears to be a cultural conditioning which renders the western form of writing irrelevant and useless. Seagrim and Lendon (1980: 205) suggest a valuable point to consider. The skills required for reading are foreign to the child. 'He is never called upon to use comparable skills which he might transfer to this situation'. The traditional Aboriginal education has shown that considerable demands are made on the memory for detail in respect of 'myths, songs, dances, ceremonies, kinship rules, tracking, hunting, identifying plants, insects, grubs and animals, the local and … the distant geography of their land … and for both sexes the correct performance of religious practices' (Seagrim and Lendon 1980: 30). The skill of tracking - that is, learning to read tracks - might have some correspondence with learning to read cursive handwriting. However, there is a clear difference. Tracks are symbolic in that they partake of the characteristics of the foot or paw that made them. Many written words derive their meanings by an arbitrary attachment to their significance. Therefore, we should appreciate, when looking at line markings, that there are many ways of deriving meaning from a variety of markings. The story telling observed by Seagrim and Lendon (1980) at Hermannsburg was often accompanied in the sand by the graphic representation of events. These are largely symbolic markings and the precise meaning is not given in their form. Munn (1973: 58) suggests that the sand markings made by the Walbiri have to be examined in relation 'to a wider graphic art of narration'. She notes that 'to accompany one's speech with explanatory sand markings is to talk in the Walbiri manner. We can propose that the markings at Koonalda might have a connection, in evolutionary terms, with the later use of graphic forms, which have become the media of social interaction. The graphic forms used are simple in appearance, yet when verbal accompaniment is reduced, they can carry on the basic meaning of the story, combined with the gestural signs, quite comprehensively. Each graphic element covers a range of possible meaning (Munn 1973: 64, Layton 1991: 185). The specific items meant depend upon the narrative context or the associated motifs (Mithen 1996: 157). When looking at the markings in Koonalda Cave, the general lack of apparent diversity can be explained in this way. Other forms of non-verbal communication - Braille, for example - appear to have little diversity when read by the untrained eye or fingers (Salzmann 1993: 213). They are however, fully comprehensive communication systems, with meanings to those familiar with them. This parallels the finger markings. To a small group of people in possession of specific type of knowledge, the variety of markings might have acted as a mnemonic notation system, reminding them of crucial and specific occurrences in a myth story.

Story telling could have been an important form of communication for pre-Historic peoples. Therefore, it is possible that the first notation systems have a connection to story telling. The line markings in the caves may have acted as organizing or memory devices for their story-telling creators, in the way that squiggles and lines do for young children (Bahn 1988). Caves would have provided suitable gathering places for story telling and teaching, particularly if the stories were restricted or for initiation purposes. Certain dreamings are restricted to specific members of Aboriginal groups today (Layton 1992b: 141f). Munn (1973:63), in her study of the Walbiri, noted that the children did not tell stories as a pastime, but at five or six years could reproduce and identify the basic graphic forms used in narration. Sand drawing was frequently performed in front of babies and small children; thus, the observation of sand drawing is part of the early perceptual experience. Layton (1992a: 54) notes that among the Aranda of Central Australia, older men are fond of using sand drawings when instructing initiates in sacred myths. As mnemonic devices, early notation systems might have helped those involved to organize and remember important stories or pieces of information (Coulmas 1989).

It is likely that a specific ritual was peculiar to a certain area - this also appears in ethnographic studies of Australian Aborigines (Gould 1969). Koonalda Cave was probably visited repeatedly for one distinct ritual. One might suggest from this that the same mnemonic lines would appear on the cave walls. In Walbiri sand drawings, graphic elements often overlay each other (Munn 1973: 72). Erasure rarely occurs and sequential actions, such as sitting and then lying down, are drawn one over the other. Munn (1973: 73) notes that a particular story can never be looked at as a unitary whole, and 'no retelling is likely to reproduce the exact arrangements and scene cycles again'. One might suggest that in the case of the Walbiri, the lines were not used as a mnemonic. However, one might suggest that the use of graphic construction, evolved from an earlier tradition of mnemonic use.

The large numbers of finger lines that exist on the walls of the Upper Chamber of Koonalda Cave suggest that the behavior of wall marking was repeated through time. This indicates that a community or at least certain members within that group shared a cultural activity. One can postulate, therefore, that the markings are a socially conveyable system of notation. This assumption would mean that a line marker did not express complete individualism, and so someone else within the same cultural tradition as the marker would have understood the marker's meaning. Thus, since we assume the finger markings are the inscribed form of the same stories, repetition or consistencies may exist somewhere in the finger markings.

Each such consistency represents the repetition of a myth story ritually recounted in the upper chamber. These may represent different traditions of line marking. Similarly, the finger markings may represent several cultural traditions of notation systems. Certain sites may have been in use over tens of thousands of years (Bednarik 1990). If so, then finger markings may not conform to one set of structural patterns but exhibit several. Further, as millennia went by the stories recounted in the Upper Chamber and the method of depicting them probably changed. The markings may represent the same, yet developing traditions of myth stories and notation system.

In maintaining an approach that is as objective as possible, we must consider other variables. Consistencies could derive from elsewhere than the markings' meaning; for instance, the markings could move in one direction because of where the line marker had to position him or herself to draw. Nevertheless, suppose that we cannot convincingly explain as accidental or functional the structures isolated and associated with the meaning of a notation system. Thus, we need to distinguish between these two types of consistencies: those functionally or accidentally based, and those not. We suggest examining consistent structures in situ to see if the situation readily explains them. Those that evade accidental or functional rationales then become candidates for us to consider as products of the markings' meaning.

Certain animal markings may show a functional consistency. Bednarik describes a variety of animal scratchings and marks (1994b: 36), but we could not say they were intentional in the sense that humans intend meaning in their notation systems. Faulstich (1986) suggests that animal claw marks were the stimulus that originally prompted finger markings, though we would debate his assertion that this represents a 'quantum leap' in Homo sapiens' expanding penchant for abstract thinking. Many animals exhibit mimicry, therefore one might suggest that the capacity to copy was firmly fixed in the early, yet fully sapient cognitive system. As Clottes (1986: 161) has suggested and which we have mentioned previously, the marking of walls to leave a trace of one's passage inside an unknown place, might have been an initial reaction. Alternatively, the deliberate mimicry of animal markings might have been an original intention. Nevertheless, this does not invalidate our hypothesis. The former activity evolved into the ritualized markings, which appear at sites such as Koonalda, demonstrating a deliberate and repeated intention.

Marshack, Bednarik, and d'Errico have contributed an important technique of analysis to the study of pre-Historic line markings (Bednarik 1986a, 1986b, 1994; d'Errico 1989, 1991, 1992a, 1992b, 1993, 1994, 1995; Marshack 1991), d'Errico focusing on engravings on portable artifacts. They suggest a close examination of the line markings themselves, looking at their cross-sections — depth, width, and shape — over their length, and of the points at which they meet or overlie. Different cross-sections of engraved lines would imply the use of a different tool, perhaps by different people, and perhaps at different times. As our experiments have shown, an examination of finger line junctions tells which lines overlie others, and hence the temporal sequence of their compilation. For finger markings, using our techniques, we can also gain an idea of the size of the hand (the distances between the digits) and the direction in which the hand moved over the medium. We may perhaps discern the age and handedness of the line maker (Bednarik 1986a). From techniques such as these, we can recreate the story of the construction and structure of the lines: the beginning and endings of lines, the order of their creation, and those fashioned with different tools or by different individuals.

This analytic technique and internal analysis remains an ideal; in practice, it faces problems. Some sets of lines contain none that overlap or meet, and hence this method loses the temporal dimension of their construction. Further, the rock may now have a quite granular surface, which means we can carry out this analysis only to a certain size dimension. At some point the rock surface swallows up the lines and renders them indistinguishable from background marks. This happens especially with weathered lines. The rock may also contain surface cracks. They can sometimes be short and of the same dimension as engraved lines. Cross-sections and overlays cannot always be ascertained with certainty. Unfortunately, this will involve a degree of subjectivity and error in the structural analysis of the line markings.

Summary and conclusions
In summary, the hypothesis we propound can be broken down into a number of proposals, which we suggested previously (Sharpe and Lacombe 1998).
Proposition 1.
The finger markings in Koonalda Cave are a system of notation, a mnemonic depiction of myth stories.
Proposition 2. As an inscribed form of stories associated with a ritual, the finger markings represent the same set of stories, repeated numerous times.
Proposition 3.
There exist structural consistencies to the finger markings.
Proposition 4(i). The finger markings can be divided into clusters.
Proposition 4(ii). Each cluster of finger markings can be analyzed into cycles.
Proposition 4(iii). There are clusters in which each set of cycles is similar.

These propositions become successively more difficult to refute the further one ascends from 4(iii). That is, number 4(iii) is probably more easy to refute than 4(ii) which in turn is more easy to refute than 4(i), and so on.

We suggest that the line markings are a notation system that act as mnemonics for the telling of stories. We suppose, in other words, that the markings communicate in an interpersonally recognizable manner. Suppose the Koonalda people's culture expected each line marker to add arbitrarily a stream of finger marks to those already there and thus to express his or her place and participation in the ritual tradition. This is not a notation system. Thus, while the line markers may have followed cultural norms, we can imagine circumstances in which the markings do not present the stories in a notation system and the remainder of the propositions do not apply. If the list of propositions holds, this does not prove the notation-system nature of the markings, though it lends support for the idea. It does suggest, if the people communicated in finger lines, how they might have conveyed their meaning.

This paper set out to suggest that some prehistoric line markings found in European and Australian caves are a notation system, or a mnemonic system of communication. To amplify this hypothesis, we proposed an empirical means, based on Marshack, Bednarik, and d'Errico's methods of internal analysis, to help elucidate our theory, and suggested a series of testable propositions for Koonalda Cave and its finger markings. If they hold up, they suggest that each set of clusters of markings with similar structures depict a myth story told in the cave. We also described the methodology and results of an experimental approach to the analysis of finger markings. If these propositions are deemed tenable, they may help to build a case for other sites with line marks. By focusing on Koonalda, we do not mean to suggest that this argument would apply to all sites of line markings; separate cases would need constructing for each of them. However, we believe that our approach is relatively objective and will facilitate the progress of future research confirming the importance of line markings.


  1. ALLMAN, W. F. 1994. The stone age present. Simon & Schuster, New York.
  2. ASLIN, G.D., E.K. BEDNARIK & R.G. BEDNARIK 1985. The 'Parietal Markings Project' - a progress report. Rock Art Research 2(1):71-74.
  3. P.G. 1991. Pleistocene Images outside Europe. Proceedings of the Prehistoric Society 57(1): 91-102.
  4. P.G. and VERTUT, J. 1988. Images of the ice age. Facts on File, New York.
  5. BEDNARIK, R. G. 1986a. Parietal finger markings in Europe and Australia. Rock Art Research 3(1): 30-61.
  6. BEDNARIK, R.G. 1986b. Cave use by Australian Pleistocene man. Proceedings of the University of Bristol Spelaeological Society 17(3): 227-45.
  7. BEDNARIK, R.G. 1986c. Comment on W. Davis, 'The Origins of Image Making'. Current Anthropology 27: 202-3.
  8. BEDNARIK, R.G. 1986d. Reply to 'Parietal finger markings in Europe and Australia'. Further Comments. Rock Art research 3(2): 159-170.
  9. BEDNARIK, R. G. 1990. The cave petroglyphs of Australia. Australian Aboriginal Studies 2: 64-68.
  10. BEDNARIK, R. G. 1994a. On the scientific study of palaeoart. Semiotica 100(2): 141-68.
  11. BEDNARIK, R.G. 1994b. The discrimination of rock markings. Rock Art Research 11(1): 23-44.
  12. BEDNARIK, R.G. 1994c. Further comment. Epistemology and palaeolithic rock art. Rock Art Research 11(2):118-121.
  13. BEDNARIK, R. G. 1997. The global evidence of early human symboling behavior. Human Evolution 12(3): 147-168.
  14. BOLINGER, D. 1980. Language: The loaded weapon. Longman, London.
  15. CLOTTES, J. 1986. Parietal finger markings in Europe and Australia. Further comments. Rock Art Research 3(2): 159-170.
  16. COOTE, J. and SHELDON, A. 1992. Introduction. In J. Coote and A. Sheldon (eds.), Anthropology, Art and Aesthetics, pp. 1-11. Clarendon Press, Oxford.
  17. COULMAS, F. 1989. The writing systems of the world. Basil Blackwell, Oxford.
  18. D'ERRICO, F. 1989. A reply to Alexander Marshack. Current Anthropology 39: 494-500.
  19. D'ERRICO, F. 1991. Microscopic and statistical criteria for the identification of prehistoric systems of notation. Rock Art Research 8(2): 83-93.
  20. D'ERRICO, F. 1992a. A reply to Alexander Marshack. Rock Art Research 9: 59-64.
  21. D'ERRICO, F. 1992b. Technology, motion and the meaning of Epipalaeolithic art. Current Anthropology 33: 94-109.
  22. D'ERRICO, F. 1993. La vie sociale de l'art mobilier paléolithique: manipulation, transport, suspension des objets en os, bois de cervidés, ivoire. Oxford Journal of Archaeology 12: 145-74.
  23. D'ERRICO, F. 1994. L'art gravé azilien: De la technique … la signification. XXXIŠme supplément … Gallia Préhistoire. CNRS, Paris.
  24. D'ERRICO, F. 1995. A new model and its implications for the origin of writing: the La Marche antler revisited. Cambridge Archaeological Journal 5: 3-46.
  25. EDWARDS, R. 1971. Art and Aboriginal Prehistory. In D. J. Mulvaney and J. Golson (eds.), Aboriginal Man and Environment in Australia. ANU, Canberra.
  26. EDWARDS, R. and MAYNARD, L. 1969. Prehistoric Art in Koonalda Cave (Australia). Bollettino del Centro Camuno di Studi Preistorici IV: 117-130.
  27. FAULSTICH, P. 1986. Parietal finger markings in Europe and Australia. Further comments. Rock Art 3(2): 159-170.
  28. FLOOD, J. 1996. Culture in Early Aboriginal Australia. Cambridge Archaeological Journal 6(1): 3-36.
  29. FLOOD, J. 1997. Rock Art of the Dreamtime. Images of Ancient Australia. Angus and Robertson, Australia.
  30. FRANK, R. 1971. The sediments. In R. V. S. Wright (ed.), Archaeology of the Gallus Site, Koonalda Cave, pp. 31-44. Australian Aboriginal Studies 26, Australian Institute of Aboriginal Studies, Canberra.
  31. GALLUS, A. 1968a. Parietal Art in Koonalda Cave, Nullarbor Plain, South Australia. Helictite 6(3): 43-49.
  32. GALLUS, A. 1968b. Archaeological excavations at Koonalda, Nullarbor Plain, 1957-1967. Journal of the Anthropological Society of South Australia 6(7):4-8.
  33. GALLUS, A. 1970. Expanding horizons in Australian prehistory. Twentieth Century 25: 66-75.
  34. GALLUS, A. 1971. Results of the exploration of Koonalda Cave, 1956-1968. In R. V. S. Wright (ed.), Archaeology of the Gallus Site, Koonalda Cave, pp. 87-133. Australian Aboriginal Studies 26, Australian Institute of Aboriginal Studies, Canberra.
  35. GALLUS, A. 1977. Schematisation and symboling. In P.J. Ucko (ed.), Form in Indigenous Art, pp. 370-386. Australian Institute of Aboriginal Studies, Canberra.
  36. GOODY, J. R. 1968. Literacy in traditional societies. Cambridge University Press, Cambridge.
  37. GOULD, R. A. 1969. Yiwara: Foragers of the Australian desert. Charles Scribner's Sons, New York.
  38. GROENFELDT, D. 1985. The interpretation of prehistoric rock art. Rock Art Research 2(1): 20-47.
  39. HALLAM, S. 1971. Roof markings in the Orchestra Shell cave, Wanneroo, near Perth, Western Australia. Mankind 8(2): 90-103.
  40. HAMILTON-SMITH, E. 1986. Parietal finger markings in Europe and Australia. Further comments. Rock Art Research 3(2): 159-170.
  41. LAYTON, R. 1991. The anthropology of art. Cambridge University Press, Cambridge.
  42. LAYTON, R. 1992a. Australian Rock Art: a new synthesis. Cambridge University Press, Cambridge.
  43. LAYTON, R. 1992b. Traditional and Contemporary Art of Aboriginal Australia: Two case studies. In J. Coote and A. Sheldon (eds.), Anthropology, Art and Aesthetics, pp. 137-159. Clarendon Press, Oxford.
  44. MARSHACK, A. 1975. Exploring the mind of ice age man. National Geographic 147(1): 64-89.
  45. MARSHACK, A. 1977. The meander as a system: the analysis and recognition of iconographic units in upper palaeolithic compositions. In P. J. Ucko (ed.), Form in Indigenous Art, pp.286-317. Australian Institute of Aboriginal Studies, Canberra.
  46. MARSHACK, A. 1989. Methodology in the Analysis and Interpretation of Upper Palaeolithic Image: theory versus contextual analysis. Rock Art Research 6(1): 17-53.
  47. MARSHACK, A. 1991. The roots of civilization: The cognitive beginnings of man's first art, symbol and notation (2nd ed.). Moyer Bell, New York.
  48. MARSHACK, A. 1997. Palaeolithic Image Making and Symboling in Europe and the Middle East: A Comparative View. In M. W. Conkey, O. Soffer, D. Stratman and N. G. Jablonski (eds.), Beyond Art: Pleistocene Image and Symbol, pp. 53-92. Memoirs of the California Academy of Sciences, no. 23. San Francisco, California.
  49. MAYNARD, L., and EDWARDS, R. 1971. Wall markings. In R. V. S. Wright (ed.), Archaeology of the Gallus Site, Koonalda Cave, pp. 59-80. Australian Aboriginal Studies, no. 26. Australian Institute of Aboriginal Studies, Canberra.
  50. MITHEN, S. 1996. The Prehistory of the Mind. A search for the origins of art, religion and science. Thames and Hudson: London.
  51. N. 1973. Walbiri Iconography. Cornell University Press, Ithaca.
  52. SALZMANN, Z. 1993. Language, culture and society. An introduction to linguistic anthropology. Westview Press, Oxford.
  53. SEAGRIM, G. N., and LENDON, R. J. 1980. Furnishing the mind. A comparative study of cognitive development in Central Australian Aboriginies. Academic Press, Sydney.
  54. SMITH, C. 1992. The articulation of style and social structure in Australian Aboriginal art. Australian Aboriginal Studies 1: 28-34.
  55. SHARPE, C. E. and SHARPE, K. J. 1976. A preliminary survey of engraved boulders in the Art Sanctuary of Koonalda Cave, South Australia. Mankind 10(3): 125-30.
  56. SHARPE, K. J. and SHARPE, C. E. 1978. The upper chamber of Koonalda Cave, South Australia: a second report. Parts one and two. Unpubl. manuscript, Auckland.
  57. SHARPE, K. J. and LACOMBE, M. 1998. Line markings as systems of notation? Unpubl. paper presented at the Rock Art Conference in Turin.
  58. P. J. 1992. Subjectivity and the recording of palaeolithic cave art. In T. Shay and J. Clottes (eds.), The limitations of archaeological knowledge, pp. 141-79. Études et Recherches Archéologiques de l'Université de Liège 49. Université de Liège, Service de Préhistoire, Liège.
  59. VYGOTSKY, L. S. 1978. Mind in society. Harvard University Press, Cambridge, Massachusetts.
  60. WALLER, S.J. 1993. Sound reflection as an explanation for the content and context of rock art. Rock Art Research 10(2): 91-101.
  61. WRIGHT, R. V. S. (ed.) 1971a. Archaeology of the Gallus Site, Koonalda Cave. Australian Aboriginal Studies 26, Australian Institute of Aboriginal Studies, Canberra.
  62. WRIGHT, R. V. S. 1971b. An ethnographic background to Koonalda Cave prehistory. In R. V. S. Wright (ed.), Archaeology of the Gallus Site, Koonalda Cave, pp. 1-16. Australian Aboriginal Studies, Australian Institute of Aboriginal Studies, Canberra.