In the third lecture, we looked at studies that have attempted to teach symbols and language, mostly artificial, to non-human primates. We drew on a classification due to Charles Sanders Peirce, and revisited by Terence Deacon and Andreas Nieder, which distinguishes several types of signs, i.e. signifier/signified relations :
- iconic " signs, for which there is a non-arbitrary relationship between signifier and signified ;
- indexical " signs, for which there is a correspondence between each isolated arbitrary sign and its referent, but no Combinatorics system into which these signs fit ;
- finally, genuine " symbols ", i.e. signs linked together by a vast Combinatorics system of relations. Insofar as these relations between signifiers are isomorphic with those linking the signifieds themselves (mental contents), it becomes possible, by simple symbolic manipulation, to carry out mental operations that retain a systematic correspondence with reality. A typical example is the correspondence between the symbols of Arabic numerals and the corresponding quantities : purely formal arithmetic calculations can be used to anticipate changes in quantities.
In this context, experimental studies show that many animals (primates, parrots, dogs, etc.) can acquire several hundred arbitrary signs, reaching Pierce's level 2. However, these acquired signs function more like cues than symbols. In fact, they have two major differences from the symbols of human languages or semiotic systems. Firstly, they lack reversibility: animals do not spontaneously understand that the signifier/signified relationship is reversible, and that it is therefore possible to move from meaning to symbol and vice versa, in either direction. On the other hand, no animal species has proven capable of learning a genuine system of complex rules linking signs to each other.
