We propose to integrate "within and around" query contexts in information retrieval to determine additional related terms to expand the query. A "within query" context is useful to select relevant term relations, while an "around query" context specifies the domain of interest and the characteristics of a document collection. We implemented this approach using a language modeling framework and we have shown that these factors improve the retrieval effectiveness on several TREC collections. This study is the first integrating multiple contextual factors and showing how they can be beneficial in information retrieval.