This paper explored the modal verb shall in formal and informal writings in academic and fiction registers. It focused on the frequencies of shall across academic and fiction domains in contemporary American English and the differences in the usage of shall between academic and fiction registers of contemporary American English. The researchers used a corpus linguistic method. Data were collected from the Corpus of Contemporary American English (COCA) and analysed using Hanks’ (2004) Corpus Pattern Analysis technique. All occurrences of shall in academic and fiction writing styles of COCA were retrieved, and 400 concordance lines consisting of 200 texts from each domain were collected. The texts were analysed and described in accordance with their syntactic, stylistic, and semantic characteristics. Results showed that shall was rare in COCA’s academic and fiction registers as the overall frequencies were 59.77 and 68.34 words per million, respectively. From all the 400 tokens being analysed, the researchers found that shall in the observed data could be classified as rules and regulations, direction, prediction, volition, and etc. The uses of shall in both domains in COCA varied syntactically, semantically, and stylistically.