A constructivist neural network model is presented that learns the past tense of English verbs. The model builds its architecture in response to the learning task in a way consistent with neurobiological and psychological evidence. The model outperforms existing connectionist and symbolic past tense models in terms of learning and generalization behavior, and it displays a U-shaped learning curve for many irregular verbs. The trained model develops a modular architecture with dissociations between regular and irregular verbs, and lesioning the different pathways leads to results comparable with neurological disorders. It is argued that the successof the model is due to its constructivist nature, and that the distinction between fixed-architecture and constructivist models is fundamental. Given this distinction, constructivist systems provide more realistic models of cognitive development.

modular connectionist network is described that learns the German verb paradigm. The architecture of the network is in accordance with the rule-associative memory hypothesis proposed by Pinker (1991): it is composed of a connectionist short-term memory enabling it to process symbolic rules and an associative memory acting as a lexicon. The network successfully learns the German verb paradigm and generalizes to novel verbs in ways that correspond to empirical data. Lesioning the model gives further evidence for the rule-associative memory hypothesis: Whe n the lexicon is cut off, the network strongly overgeneralizes the regular participle, indicating that regular forms are produced with the shortterm memory but irregular forms rely on the lexicon. However, in contrast to the rule-association theory, the two paths are not strongly dissociated, but both the short-term memory and the lexicon work together in producing many participles. The success of the network model is seen as evidence that emergent linguistic rules need not be implemented as rules in the brain.

The emerging function of verb inflections in German lenguage acquisition is modeled with a connectionist network. A network that is initially presented only with a semantic representation of sentences uses the inflectional verb ending -t to mark those sentences that are low in transitivity, whereas all other verb endings occur randomly. This behavior matches an early stage in German language acquisition where verb endings encode a similar semantic rather than a grammatical function. When information about the siuface structure of the sentence is added to the input data, the network learns to use the correct verb inflections in a process very similar to children's learning. This second phase is facilitated by the semeintic phase, suggesting that there is no shift from semantic to grammatical encoding, but rather an extension of the initial semantic encoding to include grammatical information. This can be seen as evidence for the strong version of the functionalist hypothesis of language acquisition.

Despite substantial evidence for a bidirectional relationshipbetween language and representation, the roots of this relationshipin infancy are not known. The current study explores thepossibility that labels may affect object representations at theearliest stages of language acquisition. We asked parents to playwith their 10-month-old infants with two novel toys for threeminutes, every day for a week, teaching infants a novel word forone toy but not the other. After a week infants participated in afamiliarization task in which they saw each object for 8 trials insilence, followed by a test trial consisting of both objectsaccompanied by the trained word. Infants exhibited a faster declinein looking times to the previously unlabeled object. These dataspeak to the current debate over the status of labels in humancognition, supporting accounts in which labels are an integral partof representation.

Research on lexical development in Down syndrome (DS) has emphasized a dissociation between language comprehension and production abilities, with production of words being relatively more impaired than comprehension. Current theories stress the role of associative learning on lexical development. However, there have been no attempts to explain the atypical lexical development in DS based on atypical associative learning. The long-term potentiation (LTP) and long-term depression (LTD) of synapses, underlying associative learning, are altered in DS. Here we present a neural network model that instantiates notions from neurophysiological studies to account for the disparities between lexical comprehension and production in DS. Our simulations show that an atypical LTP/LTD balance affects comprehension and production differently in an associative model of lexical development.

The effect of labels on nonlinguistic representations is the focusof substantial debate in the developmental literature. A recentempirical study (Twomey & Westermann, 2016) suggested thatlabels are incorporated into object representations, such thatinfants respond differently to objects for which they know alabel relative to unlabeled objects. However, these empiricaldata cannot differentiate between two recent theories ofintegrated label-object representations, one of which assumeslabels are features of object representations, and one whichassumes labels are represented separately, but become closelyassociated with learning. We address this issue using aneurocomputational (auto-encoder) model to instantiate boththeoretical approaches. Simulation data support an account inwhich labels are features of objects, with the samerepresentational status as the objects’ visual and hapticcharacteristics.