Implicit Learning in Spelling: The Case of English Names

Purpose In many languages, spelling is subject to multiple influences. We tested whether adults adhere to two nonphonological hypotheses when spelling an understudied word type: names, specifically, surnames. Word type sensitivity holds that different rules may apply to different word types (e.g., noun ⟨callus⟩ vs. adjective ⟨callous⟩). The distinctiveness hypothesis holds that people prefer spelling different words differently, even when they sound the same (e.g., ⟨bite⟩ vs. ⟨byte⟩).Method We conducted a corpus study and two experimental studies of spelling production (ns = 54 and 46) and two of spelling choice (ns = 67 and 58) with English-speaking Australian undergraduates (overall mean age 26.5 years; 67% female).Results The corpus analysis showed final consonant letters are doubled much more often in surnames (⟨Kidd⟩) than non-names (⟨kid⟩). Mixed-model analyses also revealed a robust word type effect in behavioral spelling production and choice tasks: Participants doubled final consonant letters significantly more often for nonwords presented as surnames (⟨Mr Lodd⟩) than as common nouns (⟨the lod⟩). The corpus analysis also showed that the final letter is doubled more often in surnames that sound like an existing word (⟨Ms Codd⟩) than in surnames that do not (⟨Ms Grodd⟩). However, participants showed the opposite tendency, doubling letters significantly less often in surnames that did (⟨Ms Nodd⟩) than did not sound like words (⟨Ms Lodd⟩).Conclusion Word type influences adults’ spelling. However, the results raise questions about people’s use of distinctiveness and about the extent to which experienced spellers mirror written language statistics in their behavior.