Experimental work in psycholinguistics has been dedicated to issues such as whether semantic context facilitates recognition of words. To provide answer to this question, we have to observe semantic priming, with the help of which researchers can study the structure of semantic memory. First of all, however, we need to understand where other information about words (pronunciation, spelling, part of speech) is stored and what the term ‘mental lexicon’ actually refers to. Psycholinguists have come to certain conclusions by considering what they have observed in the studies of the processes of lexical retrieval. From these findings, they have constructed several models of lexical access, out of which I will be presenting three major ones: two direct access models and a search model. To see how these models work, we have to examine how they explain the basic experimental findings: the effects of frequency, non-word, context, and degradation. I will also be listing the problems the models have with the explanation of the combined effects of frequency, context and degradation.