After a brief introduction of myself and an overview of my expertises, I shall present an
example of \textit{quantitative linguistics} as might be intended by a physicist.
I'll show an overview of some interesting quantities, like dictionary growth and 
correlations, that can be calculated from a stream of words in case of a text, or a 
stream of tags in case of folksonomies. The hope is that those quantities may
help to uncover the semantics embedded in words and tags.
Eventually, I'll sketch some possible ideas for future works.