Indeed, what has been briefly outlined here in our thought-experiment is valid not only for tree-forms - which, incidentally, appear to be extremely rare in rock art according to the current interpretation of the data. It may apply to other, seemingly natural forms alluding more or less clearly to certain animals and human figures. All the drawings in Figure 4, which are reproduced from illustrations and photographs from Japanese folklore literature, show territorial markers of local communities in Japan. Primarily of abstract geometrical form, these technogenetic (bundling stalks) markers were transformed by secondary processes -that is to say, by accumulation of new cultural elements - into plant-, animal-, or human-like form, or even into secondary topo-, cosmo-, or technomorphous types. This paper tries to use this ethnographic material for the interpretation of palaeolithic and neolithic petroglyphs and pictographs.
There is another set of qualities which aesthetically characterises rock art representations. It extends from naturalistic, realistically proportioned art works to more or less stylized types, and finally to designs which are highly abstract, showing geometric forms which can no longer be identified as representing objects (Fig. 5). In the current archaeological literature, it is generally assumed that naturalistic representations are primary and that the prototypes of the painter were natural or instrumental forms. The conditions of 'style' are usually taken for granted and barely discussed. Purely geometric forms, if not otherwise identified, are interpreted in the frame of abstractly conceived Euclidian geometry.
Anyone intending to discuss rock art in a world-wide context must necessarily refer to a theoretical framework of world-wide validity. Conventionally, this framework consists of three elements. First there is the formal expression - that is to say, the rather naive opinion that formal units indicate a direct relation with some natural or cultural object. This relation, it is believed, can still be identified today. The second element deals with the question of the meaning of such formal units in themselves or in the context of other forms. The third element concerns the purpose, the function, the social or human conditions that produced such forms. Interpretation in general is based on the concepts established by various other disciplines, such as history of religion, history of art, economic anthropology, etc.
One of the main problems of interpretation stems from the fact that we know very little about the material conditions of prehistoric societies. Archaeology presents us merely with what was durable, the 'remains'. Consequently, analogies with known ethnological conditions have been increasingly used for the interpretation of prehistorical data (e.g., Narr 1973); but, particularly with regard to rock paintings, the analogies do not seem to be sufficient. The ethnographic study of semantic architecture might provide some clues for the reinterpretation of rock art morphology.