Discriminative learning predicts human recognition of English blend sources

Strict compositionality in morphological theory is problematic for explaining how language-users comprehend phenomena like the partial yet non-decomposable forms in phonaesthemes and in blends like edutainment. An alternative account, based on discriminative learning, proposes that language-users associate linguistic cues (e.g., short segment or letter strings) with multiple simultaneous possible lexical and grammatical meanings. We evaluate this account on off-line human identifications of partial word-forms, using English blend words as our test case. We hypothesize that readers’ ability to parse out source meanings from written blend forms should be correlated with how strongly a naı̈ve discriminative reading model associates the cues in each form with the correct source meanings. We provide evidence for this claim in two experiments, in which the discriminative learning model reliably predicted participants’ success rate in guessing the sources of both attested and novel blends. This finding supports discriminative learning as a realistic model of how readers parse wordforms and map them to meanings. Further, the result points towards a novel, precise account of blend processing.