The aim of this project is to test, as far as is possible in this form of study, the theory of entoptic phenomena as a tool of explanation. A study of this nature has certain limitations: its length precludes the reproduction of all relevant data and original research is not possible. Lewis-Williams and Dowson (1988) made some major assumptions which will be challenged in later chapters. However, it is the function here to summarise the theory laid down by Lewis-Williams and Dowson. The subsequent work by these authors will also be discussed in order to assess how their theory has progressed. Comment will be made on the nature and suitability of its usage by other authors. No attempt will be made to criticise or support their theory at this stage: the sole purpose of this chapter is to lay down the basic framework for the assessment of the chosen research topic.
THE FOUNDATION STONE
In 1988, J.D. Lewis-Williams and T.A. Dowson published a paper in Current Anthropology entitled 'The Signs of All Times'. It was an attempt to use information familiar to physiologists and psychologists for the first time in the arena of archaeology. Their aim was to explain the more 'ritual' side to our forefathers' existence. So much study in the field of archaeology concentrates on the physical evidence of the mundane practicalities of day-to-day lives. Here was a real chance to explore our antecedents thought processes and the ways in which cognition could be altered by surroundings, the possibility of ritual drug use, or altered states of consciousness due perhaps to flickering campsite firelight or stress. Whatever the outcome of the study itself, its possible criticisms and inadequacies, it was, in my view, a major move towards the maturation and separation of cognitive archaeology into a field of its own.
Lewis-Williams and Dowson (1988:201) point out in their opening statement that the signs of European Upper Palaeolithic art have been a persistently intractable challenge to archaeologists. To elucidate their meaning, two main approaches have been employed: the first, using ethnographic analogies to argue that the signs represented traps, huts or shrines inhabited by spirits (Breuil 1952:24); the second, attempting to discover an inherent order and therefore deduce meaning (Laming 1962; Leroi-Gourhan 1968, b; Marshack 1972; Sauvet, Sauvet and Wlodarczyk 1977; Sauvet and Sauvet 1979; p and Faris 1983).
Various patterns have been suggested, some being supported more adequately with data than others. The induction of meaning has been a more hazardous route (cf. the aims of cognitive archaeology: see below p.22). For, as the authors state, it is logically impossible to induce meaning from numerical rock art data, as it is from any data (Lewis-Williams 1983:101; Lewis-Williams and Loubser 1986, as cited by Lewis-Williams and Dowson 1988:201). No explanation as to the meaning of these signs has won general acceptance. The emphasis at the time of this article was not on meaning but on a consolidation of the data which had already been accumulated.
Sieveking (1979:209) and Halverson (1987:70) among others believe it is very probable that the meaning of Upper Palaeolithic art will never be known. From a cognitive archaeological standpoint, the suggestion that the meaning of such art will ever be known is inconceivable. Without the means to be transported back into the past, the thought processes of our antecedents are impossible to determine, but it would be true to say that these very thought processes are an intrinsic part of the study of Archaeology.
Lewis-Williams and Dowson (1988) wished to develop a "model for classifying and addressing Upper Palaeolithic signs that avoids simplistic ethnographic analogy". Neuropsychological research explains the forms of certain depictions; the meanings of some can be established from directly relevant ethnography (Lewis-Williams and Dowson 1988:201). The fact is that Upper Palaeolithic art has no real current counterpart, but Lewis-Williams and Dowson are attempting to prove that a neurological bridge affords some access to the Upper Palaeolithic (Lewis-Williams and Dowson 1988:202).
However, Lewis-Williams and Dowson state that, at the present stage in their research, it is premature to distinguish between phosphenes and form constants. Thus, they have been grouped together and assigned the generic term of 'entoptic phenomena' or entoptics, by way of classifying these largely geometric visual percept. The term 'entoptic' comes from the Greek to mean 'within vision', and the term 'entoptic phenomena' means visual sensations whose characteristics derive from the structure of the visual system (Tyler 1978:1633). Lewis-Williams and Dowson use the term 'hallucinations' to describe more complex iconic visions (Siegal 1977:134; Reichel-Dolmatoff 1978: 12-13). Tyler also makes the point that:
"the nature of entoptic phenomena makes it hard to design highly-controlled, stimulus-bound experiments to specify them. It is therefore appropriate to report them on an observational basis before more indirect outcomes are explored" (Tyler, 1978:1633).
Lewis-Williams and Dowson distinguish three components in their model:
The first component, types of entoptic phenomena, has had numerous forms distinguished, of which certain types recur. Of these, Lewis-Williams and Dowson have selected six. These have been chosen from the work of neurologists and psychologists (e.g. Kluver 1942:172-77; Knoll and Kluger 1959; Horowitz 1965). Examples of these six images are shown in Table 1 , and are accompanied by corresponding examples from major artistic movements from the Palaeolithic and beyond. In their discussion of Upper Palaeolithic art, Lewis-Williams and Dowson (1988:208) state that "the six entoptic categories do not account for every Upper Palaeolithic sign, the presence of all six suggests that at least some of the art originated in certain altered states and leads on to the assessment of the other two components of our model" . Further examples from the Upper Palaeolithic and Mesolithic are shown in Table 2 .
Their second component is concerned with the principles of perception, of which there are seven as defined by Lewis-Williams and Dowson (1988). These were defined with the aid of neuropsychological studies, and are as follows:
An example of reduplication is illustrated by Reichel-Dolmatoff (1972:91-92) who noted that when Tukanoans were asked to draw their mental imagery, they tended "to fill the pieces of paper he gave them with rows of formalised and reduplicated geometric motifs comparable with their painting of the same motifs on the walls of their houses. The Tukano identified these reduplicated forms as images derived from what they themselves recognised as the first stage of their trance experiences; there can be little doubt of their entoptic origin" (Reichel-Dolmatoff 1978:12-13).
These principles not only apply to entoptics, but also to hallucinations, and in some cases link the two. Also, once these principles come into play, one might be led to believe that a frieze of images is more complicated than it actually is, or that it contains more/different images than it does, simply because of the alterations made to their visual appearance by these seven elements. There are indeed an infinite number of combinations. This hinders rather than helps the recognition of such signs. It would in fact be the exception rather than the rule to find matching images amongst larger groups, as they will undoubtedly have been randomly altered. When one looks at ethnographic examples of art, it is unlikely that their basic imagery has undergone the exact alteration to which the Upper Palaeolithic examples were subjected.
The third component consists of stages in the progression of altered states of consciousness, of which there are three. The three stages and corresponding examples from San, Coso and Palaeolithic art are shown in Table 3 . In the first stage, the subject will only perceive entoptic phenomena. These can be experienced with the eyes open or closed; they also tend to be located at reading distance, and may appear to recede and advance (Siegal 1977:132).
At this stage, these images cannot be consciously controlled to any extent (Kluver 1926:504; Siegal 1977:132). Knoll et al (1963) and Siegal (1977:32) note that they are also characterised by varied and saturated colours. Sometimes, a bright light in the centre of the field of vision obscures all but peripheral images.
The rate of change of the phenomenon is generally very rapid, although it may vary from one hallucinogen to another (Knoll et al 1963:221). The less experienced subjects found it difficult to keep pace with the rapidly changing images. The more experienced subjects had better powers of observation and description (Siegal 1977:134).
The second stage in Lewis-Williams and Dowson's development of mental imagery is where "subjects try to make sense of entoptics by elaborating them into iconic form" (Horowitz 1964:514, 1975:177, 178, 181). A visual image reaching the brain is decoded by being matched against a store of experience. If a 'fit' can be effected, the image is 'recognised' (Lewis-Williams and Dowson 1988:203). However, during altered states of consciousness, the nervous system becomes like any other sense impression (visual, aural, or tactile), which produces a variety of images including entoptic phenomena. The brain acts accordingly and tries to decode these impressions like any other sensory impression during a normal state of consciousness (see Horowitz 1975:177).
As subjects move from the stage where entoptics are manipulated into iconics into stage three, changes in imagery occur. Many of the subjects used during laboratory experiments into entoptic phenomena reported that they experienced a vortex or rotating tunnel that seems to surround them, and there is a progressive exclusion of perceptual information (Horowitz 1975:178).
Also during this transitional stage into iconic imagery, there is an increase in vividness. Subjects stop using similes to describe their experiences and assert that the images are indeed what they appear to be. Even during this iconic stage, entoptics can persist; iconic imagery is "often projected against a background of geometric forms" (Siegal 1977:134).
These three stages are not necessarily sequential. Some of the subjects appear to move directly into the third stage, while others do not progress beyond the first entoptic stage. These are cumulative rather than sequential. This three-stage progression was established by research using mescaline and L.S.D. It is not known if the trajectory of mental imagery is the same for all drug- and non-drug-induced states, but Lewis-Williams and Dowson believe that a broad similarity can be expected.
In summary, Lewis-Williams and Dowson believe that drugs were being used to induce altered states of consciousness as part of a shamanistic ritual, during which time the drug-taker witnessed internally- generated perceptual images which the drug-taker, on regaining a normal state of consciousness, would replicate on a cave wall or some portable artefact.
THE PROGRESSION OF A THEORY
In 1993, Lewis-Williams and Dowson published an article which expanded on their previous study. In no way did they attempt to alter their previous work in order to recognise some of the criticisms which had been made. They have merely added points in support of their established views. It is in this article that they follow Bradley (1989) and Patton (1990) in applying their theory to a data set composed of the art forms found on Megalithic monuments. Examples of Megalithic motifs as compared to laboratory-generated entoptics are shown in Table 4 . Examples of depictions from San, Coso and Neolithic contexts are shown in Table 5 .
In this later paper, they use the work of Bourguignon (1968, 1973) to support their theory of a universal phenomenon (Lewis-Williams and Dowson 1993:55). Bourguignon found that as many as 437 of the 488 societies she surveyed had some form of institutionalised altered states of consciousness. Her sample ranged from foragers to more complex societies (Lewis-Williams and Dowson 1993:55; Noll 1985:447).
On Megalithic art, Lewis-Williams and Dowson state that, because there are fewer iconic depictions than in Upper Palaeolithic art, research is restricted in general to their stage 1. They also note that when there is little diversity in the occurrences, such as in the case of cup and ring marks, "it is impossible to make a convincing case for entoptic origins" (Lewis-Williams and Dowson 1993:56). Instead of looking at the implications of this problem, they dismiss it by saying "it is better to exclude these very simple motif assemblages until the more persuasive motifs have been studied or until other evidence suggests that the motifs may have been associated with altered states of consciousness" (Lewis-Williams and Dowson 1993:56). They regrouped Shee Twohig's megalithic motifs into entoptic and non-entoptic forms (see Table 6 ). Their fondness for excluding contradictory evidence from their data is also illustrated in their previous work (eg. Lewis-Williams and Dowson 1988:208), where they state that the exclusion of certain errant data strengthens their argument as opposed to flawing it.
To say that entoptic phenomena, when seen reproduced by either modern humans or our ancestors, are not culturally biased would be a mistake. Once these entoptic images, be they phosphenes or form constants, have left the safe harbour of the host's mind, they are subject to any form of cultural bias. As the authors point out, when discussing their comparison of Neolithic art and laboratory-produced images, "the elements established under laboratory conditions are not actually entoptic phenomena; rather, they represent the ways in which Western subjects perceive them and then, with pencil and paper, record their perceptions. The subjects' familiarity with geometry and a wide range of geometric forms together with the 'clinical' rather than emotional ambience of laboratory work make for greater simplicity and regularity than we should expect to find in art produced in more highly charged circumstances" (Lewis-Williams and Dowson 1993:56).
They go on to say that, at the present time, they are not concerned with the cultural predilections which govern the proportions of certain entoptics and choice in their selection, but they are concerned simply with their presence.
During the 1993 study, they note that the placing of a body in a grave, such as a chambered tomb, recalls the laboratory subjects' reports on the mental imagery being related to a tunnel-like perspective. They report a bright light in the centre of their field of vision that becomes the focus of this tunnel (Lewis-Williams and Dowson 1993:60).
They go on to say that, "there is a need to examine in detail the placing of motifs within the architecture of the tombs (e.g., Eogan 1986), the development of 'styles' within specific sites, and the reuse of decorated stones (e.g. O'Sullivan 1986, 1989), in the light of their possible association with specific components of altered states of consciousness" (Lewis-Williams and Dowson 1993:60-61).
There have been two main patrons of Lewis-Williams and Dowson's theory: Bradley (1989) and Patton (1990). They have had enough faith in its plausibility to publish articles where they have used it on their own data sets. However, they do not agree with all of the points which Lewis-Williams and Dowson have made concerning the theory and the conclusions drawn from their research. A summary of their views on the theory, and their use of it, follows.
Bradley (1989) voices two main criticisms; his intention being to establish his chosen data set as a more suitable candidate for the application of the theory. Firstly he points out that not enough is known about the Upper Palaeolithic in general, as well as the chronology of the art itself, and that Lewis-Williams and Dowson's theory rests on global constants and cannot be tested archaeologically. Secondly, since Upper Palaeolithic art contains a striking mixture of the abstract and the naturalistic, placing too much weight on the signs ignores the realistic depictions. He goes on to say that the very nature of the art may mean that it is not the best candidate for this form of study (Bradley 1989:69). Bearing this in mind, Bradley puts forward Megalithic art as a candidate for testing this theory, for the following reasons:
The art inside the passage graves was not always easy to locate, and would have needed deliberate lighting. In the second symbolic phase (see Patton below p.16), the non-representational signs, almost all can be matched to Lewis-Williams and Dowson's account of entoptic phenomena. (Bradley 1989:71). As opposed to the tumuli, (where there is little if any art), the chambered tombs remained 'open' to the outside world for a considerable period, and it is here that the entoptic images are depicted (Bradley 1989:71).
"At different times there have been claims that the Irish art includes solar and anthropomorphic motifs (Brennan 1983; O'Sullivan 1986: 81-82), but for the most part it is made up of abstract designs almost all of which again reflect the entoptic imagery discussed by Lewis-Williams and Dowson (figs. 3 and 4)" (Bradley 1989:73).
In favour of Lewis-Williams and Dowson, Bradley notes that, neither Breton evidence or that from Britain and Ireland can prove their case, but the ideas they discuss have certainly helped to shed light on some of the more intractable problems experienced by those working in an entirely different field. That makes their study a particularly persuasive one. (For British and Irish Megalithic motifs see Table 9 ).
Lewis-Williams and Dowson's second disciple is Patton (1990). He takes his study from the work carried out at the Table des Marchand passage grave at Locmariaquer in Brittany. He says that "this complex includes a series of carved menhirs, of which Le Grand Menhir Bris is one. A second menhir, with a carved anthropomorphic motif and a series of 'shepherds' crooks" , was left in situ when the passage grave was built and incorporated as the back stone of this monument. (For examples from Brittany see Table 8 ).
"Studies of megalithic art (Shee Twohig 1981) suggest that motifs became increasingly stylised, symbols whose meaning had been self-evident became accessible only to the initiated. All of the evidence suggests increasing restriction of participation in ritual and access to ritual knowledge" Patton (1990:555).
He also points to the fact that only a select few were buried in these graves. Patton (1990) believes that it was the people with ritual knowledge o Hf the supernatural and of altered states of consciousness, and those who experienced entoptic phenomena who would have been considered privileged.
It is Gavrinis which has (according to Patton, 1990:556), the greatest concentration of entoptic motifs of any grave in the Armorican area. He goes on to say that one finds culturally-significant motifs against a background of entoptics, as was stated by Lewis-Williams and Dowson in their description of altered states of consciousness.
Patton (1990:557) concludes by stating that, at the end of the 5th millennium BC in southern Brittany, there was an emergence of marked social differentiation; and that the emergence of entoptic phenomena in Breton megalithic art was a part of this development, geographically and chronologically. However, entoptic images disappear from the repertoire C of Armorican megalithic art in the second half of the 4th millennium BC. This coincides with the shift from 'Carnac tumuli', such as Gavrinis, to gallery graves and lateral-entrance graves. In these new graves, the images are representational, and include hafted axes and stylised female forms: no entoptics are present.
The theory of entoptic phenomena and its relationship to Upper Palaeolithic art was devised by Lewis-Williams and Dowson in 1988. Previous to this, any suggestion of a link was both brief and tenuous. The authors were the first to tackle the problem in any depth, and for this they must be admired, even if at times their model appears far-fetched. They put forward the notion that past populations, for example during the Upper Palaeolithic, were engaging in the generation of drug-induced tranc Be states, which altered consciousness and were intrinsically linked to some form of shamanistic ritual.
Their model has three main components. Firstly types of entoptic phenomena involved the selection of six entoptic forms from a variety outlined by neuropsychological studies carried out by Knoll (1963) among others. Secondly they listed the seven principles governing their perception, these being: replication; fragmentation; integration; superpositioning; juxtapositioning; reduplication; and rotation.. Thirdly they defined the three stages in the progression of altered states of consciousness, these being entoptic (non-culturally biased entoptic images, predominantly geometric in form), iconic (representations which may or may not have been culturally biased) and hallucinogenic. All three stages may or may not have been entered into, but they would have been sequential, no matter how many stages the subject encountered.
Lewis-Williams and Dowson's research has been based in two main areas, neuropsychological/physiological and ethnographic studies. The primary assumptions generated by this study, when applied to our predecessors, are that they would have had the physical ability to generate phosphenes and form constants (entoptics); that they had the drugs available to them to induce hallucinations, during which time entoptics would have been perceived; that these trance states were all part of some ancient shamanistic ritual; and that these images would have been depicted in the form we know today from Upper Palaeolithic cave art contexts. It is the aim of this project to attain evidence to refute or support these primary assumptions.
This theory has attracted both support and criticism, and has subsequently been used to explain the symbolic content of Megalithic art. It was first used by Bradley (1989), then by Patton (1990) and finally by the originators themselves (Lewis-Williams and Dowson 1993). However, it appears that since its first publication, in 1988, no further work has been carried out to enhance the basic assumptions generated by the model, nor has there been any attempt made to pacify the critics. Subsequent studies appear to have been based on foundations which have not been established as being firm in the first instance.
In future chapters, I shall examine each assumption in turn and gather
information in order to support or refute them and thus produce a more concrete
foundation on which future studies can be built.