… many societies have a symbolism but not a known key to it. Among those that have a key, many reserve it to a minority while the majority are witnesses of and even actors in the symbolic activity. In those societies which have a key and divulge it freely, many symbols are explained neither by that key nor by any other. If the cryptological view of symbolism were valid, it would have to admit that the mass of humanity obsessively manipulates tools whose usage it does not know, and reiterates messages whose sense it is ignorant of.
… In form, symbolic motivations resemble technical ones. Just as we say that a product is good for a certain purpose because it has certain qualities, so we seem to say that an object is good for symbolising this or that because it has these or those properties.
But what characterises a technical motivation, and all rational motivation come to that, is that it is based on a general principle: if one says that glass is good for making bottles because it is transparent and tasteless, one implies that these qualities are desirable for a bottle. If one made a complete list of the desired qualities, of all products having these qualities, one could say that they are good for making bottles, and of a product having none of these qualities, that it absolutely would not suit. A motivation is only valid if it is generalisable in this way.
Symbolic motivations, on the contrary, are absolutely not generalisable.
… freedom of connection is particularly clear in the symbolic use of figurative language. Given any two terms, we can never exclude the possibility that one may become the symbol of the other.
… This same lack of generalisability that makes motivations fail as discourse on symbolism conversely characterises them as a discourse in symbolism, and not at all as a discourse that interprets but, on the contrary, as a discourse that must be interpreted.
… No conscious and shared knowledge justifies a semiological view of symbolism. I have shown that the unshared conscious knowledge, which constitutes the exegesis of symbols, also fails in this respect; exegesis is not an interpretation but rather an extension of the symbol and must itself be interpreted. A final hypothesis remains possible: the second term of the pairs (symbol, interpretation) necessary to a semiological approach might be part of an unconscious knowledge; symbols might be interpreted according to a code that humans share, without being aware of it.
… Ernest Jones, in his essay ‘The Theory of Symbolism’ in which he defends Freud’s position, brings up the problem posed by the multiplicity of symbolic associations. As he points out, these associations are much more numerous and diverse than those proposed by Freud. The latter may even all be missing without this making interpretation impossible. The systematic pairing of several thousand symbols with a hundred or so ideas — says Jones — seems again brought into question. Nevertheless, Jones justifies it by two considerations. Firstly, only unconscious associations are really symbolic: ‘Only what is repressed is symbolised; only what is repressed needs to be symbolises. This conclusion is the touchstone of the psycho-analytic theory of symbolism.’ Secondly, ‘the primary ideas of life [are] the only ones that can be symbolised — those namely concerning the bodily self, the relations to the family, birth, love, and death.’ In other words, he asserts that two criteria — unconscious character on the one hand, and belonging to a short list of possible symbolisations on the other — are co-extensive, and that therefore the data isolated by Freud do constitute an autonomous system.
… if one examines a concrete symbolic system in its entirety, instead of remaining content to assemble isolated examples found here and there that conform to the thesis one wishes to defend, it seems clear that symbolic associations are multiplex, that they may be culturally explicit or implicit, individually conscious or unconscious, inside or outside the domain of interpretation defined by Freud (and after him by Jones) without these three distinctions overlapping. In these conditions, the problem posed by this multiplicity of symbolic associations remains entire, and Jones’ attempt to resolve it fails. The proposed distinction between the symbolic and the metaphorical is based on two criteria that are not co-extensive, and therefore it must be abandoned. The associations proposed by Freud may enter into symbolic interpretation but they are not necessary to it.
It will now be simple to show that in any case, these associations are not sufficient either, and that even if we supposed that they were always made, and made exclusively (which Jones admitted was not the case) they do not constitute an interpretation of symbols. Considered as interpretations, these associations are not, in fact, less mysterious than the symbols themselves.
… Attempts to construct a symbolic system by placing symbols end to end are generally repaid by failure: nothing could be less symbolically efficacious than the Cult of the Supreme Being founded by Robespierre, or the sexual symbols of intellectually pretentious pornographic films. These manipulations of symbols certainly provoke effects, but never quite those that were anticipated. It is not that these symbols once put into play are difficult to decode; on the contrary. It is just that the decoding of symbols is neither necessary nor sufficient to constitute a symbolic system.
The very notion of the symbol is a secondary or cultural development of the universal phenomenon that is symbolism.
… I suggest, therefore, that the notion of the symbol, at least provisionally, be removed from the vocabulary of the theory of symbolism, and be described only as a native notion.
We now see more clearly why the two semiological views discussed so far — the cryptological and the Freudian — are bound to fail. They both agree, without prior discussion, to answer ‘What do symbols mean?’ Yet this question presupposed, firstly that symbols are defined and secondly, that they do have meaning. As these presuppositions are erroneous, the question as posted is impossible to answer. And that is exactly where its symbolic import lies: the inevitable failure of all attempts assures, at one and the same time, their reduplication. Thus, exegetical and psychoanalytical attempts seem to obey a cultural plan — in appearance, to interpret symbolism; in fact to recreate it. For all keys to symbols are part of symbolism itself.