ABSTRACT
This article discusses the lexical and semantic representation through the collocation that appeared in the Coronavirus Corpus. This research investigates the frequent collocates that appeared together with the node word Corona and find out how those collocates construct the meaning through the linguistic system and conceptual system as they are involved in lexical representation. This research offers a new insight into teaching language using Lexical Concepts and Cognitive Models (Evans, 2009). The data collected are from the Coronavirus corpus by using the corpus-based method. The strength of the keywords and collocate is measured by using Mutual Information (MI). The MI was set in 5, therefore, three lexemes resulted, i.e., cases, patients, and outbreak. The data were analyzed using the lexical concept and cognitive model proposed by Evans (2009). The research results, in general, reveal that the information is coming from the "authorized institution" and "government's representative", it needs "the official approval or agreement" before publishing to the media, and in the passive form, it describes the foregrounded information and agentless informational assertion. Therefore, the information should be accurate and firm. However, another lexical concept reveals that the information is terrorizing and terrifying, the educated and trained person should also conduct the treatment.