 The search model is a two-stage model with three access options, phonological, orthographical and semantic. Where the stimulus is first matched one by one against the representations in the respective access bin. At a second stage, the identified input activates the target as well as related input in the master file. Let us illustrate this. The spoken input horse is identified in the phonological access file and activates the target and all phonologically related elements, such as horse, force, cause, etc. If the input is semantic, that is, if we identify the image over horse, the master file would also contain elements such as zebra, rite or even cow. In all cases, the final choice, of course, is context dependent.