The P600 ERP component as an index of rational error correction within a noisy-channel framework of human communication

Abstract Models that aim to explain electrophysiological responses to linguistic processing (e.g., the N400 and P600 components) typically assume that the representation of linguistic input is error-free. Recent evidence suggests that language processing mechanisms are well-adapted to input corrupted by noise (e.g., speech errors, mishearing) and readily correct for it via rational inference over possible intended sentences and probable noise corruptions. We hypothesize that a P600 ensues whenever the probability that the message was corrupted by noise exceeds the probability that it was produced intentionally and perceived accurately. We show that semantic violations that are attributable to noise—for example, in “The storyteller could turn any incident into an amusing antidote ”, the implausible word “antidote” is orthographically and phonologically close to the intended “anecdote”—elicit a P600 (rather than N400). Further, the magnitude of this P600 is shown to relate to the probability that the comprehender will retrieve a plausible alternative.