Where does metonymy begin? Some comments on Janda (2011)

In a recent paper published in this journal, Laura Janda makes a number of claims about metonymy, specifically about metonymy in word-formation as part of grammar. In a nutshell, what she says is that suffixed nouns such as Russian saxarnica (from saxar ‘sugar’) ‘sugar bowl’, Czech břicháč (from břicho ‘belly’) ‘person with a large belly’, or Norwegian baker ‘baker’, are metonymic extensions from saxar ‘sugar’, břicho ‘belly’, and bake ‘bake’, respectively. It is our contention that this claim about metonymy being involved in word-formation phenomena such as suffixation is misconceived and leads to an overuse of the term ‘metonymy’. We first comment on Janda's views on cognitive linguistic research on metonymy in grammar and word-formation, and then evaluate the evidence that she provides to support her central claim – from some general claims about metonymy and grammar to the way she identifies metonymy in word-formation. Finally, we point out a series of problems ensuing from the concept of word-formation metonymy. The analytical parts of Janda's article are in our view a more or less traditional cross-linguistic inventory of suffixation patterns that do not exhibit metonymy as such. However, some genuine metonymies that crop up among her examples are glossed over. In other words, we claim that her analysis ignores metonymies where they appear and postulates metonymies where they do not exist.