The initial stage of language comprehension is a multi-label classification problem. Listeners or readers, presented with an utterance, need to discriminate between the intended words and the tens of thousands of other words they know. We propose to address this problem by pairing a network trained with the learning rule of Rescorla andWagner (1972) with a second network trained independently with the learning rule of Widrow and Hoff (1960). The first network has to recover from sublexical input features the meanings encoded in the language signal, resulting in a vector of activations over the lexicon. The second network takes this vector as input and further reduces uncertainty about the intended message. Classification performance for a lexicon with 52,000 entries is good. The model also correctly predicts several aspects of human language comprehension. By rejecting the traditional linguistic assumption that language is a (de)compositional system, and by instead espousing a discriminative approach (Ramscar, 2013), a more parsimonious yet highly effective functional characterization of the initial stage of language comprehension is obtained.