This corpus-based study explores how linguistic and sociocognitive factors designate when, why, and how speakers say sorry and I’m sorry in two corpora of American spoken English; namely, Michigan Corpus of Academic Spoken English and Corpus of Contemporary American English. It is well-documented that a given word and its surrounding words or phrases bear a mutual impact on one another’s function. This study tries to cast some light on the shades of differences between sorry and I’m sorry, which seem to be taken for granted by non-native speakers of English, through analyzing one academic and one non-academic corpora. The results of the functional analysis suggest that most instances of the word sorry were used to indicate interruption, self-repair, and expressing regret, whereas I’m sorry was chiefly exploited to express regret and apology. The use of either sorry or I’m sorry was guided by the function intended by the interlocutor. These findings are justified with reference to form–function mappings, sociocognitive alignments, and other discourse considerations.