Does the mental lexicon exist?

One of the central and most intriguing components of language processing to researchers is the mental lexicon. The term was used for the first time by Ann Triesman in 1961 and we still do not have clear answers on how it is structured and how much information it contains, or even if there is something to be called a mental lexicon. For some time, the mental lexicon has been compared to a mental dictionary both storing and organizing word knowledge; however, they are surely different in structure and quantity / quality of information. Neuroimaging studies have also tried to bring contributions to these questions. Some researchers believe that there are many lexicons, one for each level of stored information (ULLMAN, 2007): orthographic, phonological, semantic and syntactic lexicons. Another group of researchers (MCCLELLAND; ROGERS, 2003; SEIDENBERG, 1997, etc.) postulates the existence of only one lexicon where all information levels are integrated. Recently, a new audacious proposal has been done by Elman (2009), the inexistence of a mental lexicon. In this paper, we discuss the different views of the mental lexicon structure and content, in order to question the architecture of the lexical knowledge in the brain as opposed to what can be consciously thought as the speaker's lexical knowledge. We try to proceed on the discussion of Elman's new proposal and confront it to data obtained by behavioral, neuroimaging and computational studies. This theoretical review briefly explains the evolution of the mental lexicon conceptions from the dictionary-like to the no-lexicon proposal.