In the case of the mnemonic image, we are first tempted to think of it as a "perceptive" trace (stored in and retrieved by our head, we don't really know how) that functions by pure "correspondence" with what it refers to ( model of memory as a repository). Then we are tempted to think of it, on the contrary, as a mere constructed sign, whose link with what it represents is merely conventional and which functions by mere "coherence" with what we "know". The first temptation is clearly that of the mnemonic image as a photograph of the past, if we underline the indexical relationship (as a trace) of said iconographic device that links it to the reality it represents. The second temptation could also be metaphorized by photography if we underline what is built in it (or symbol: its artificial chromaticism, the incorporated Renaissance perspective, the framing convention, etc.). But if we believe Wittgenstein when he says that the image mnemonic is not like a photograph, we are accepting that both temptations lead to an erroneous (or at least incomplete) idea of ​​the nature and functioning of it.