| Popis: |
Even though proper name derivation is mostly regarded as weak, even null, the scan of TLFi data leads to an important - although heterogeneous - corpus of derivatives. From the description and analysis of these data, we propose a classification, built as follows : first, on the main character of the basic proper name (embodied or disembodied) ; then, for embodied based derivatives, on the semantic relationships established between the basis and the derivative. In numerous cases, the meaning variations can be explained by tropaïc relationships which apply on each element of the derivation. |
