Constructing a speech out of language from two or three sources, without crediting the sources or including much original language, is called