… we have now established three species of representations: copies, signs, and symbols; of the last of which only logic treats. A second approximation to a definition of it then will be, the science of symbols in general and as such. But this definition is still too broad; this might, indeed, form the definition of a certain science which would be a branch of Semiotic or the general science of representations which might be called Symbolistic, and of this Logic would be a species.
http://www.commens.org/dictionary/entry/quote-harvard-lectures-logic-science-lecture-i-0