The Utility of Cognitive Plausibility in Language Acquisition Modeling: Evidence From Word Segmentation

The informativity of a computational model of language acquisition is directly related to how closely it approximates the actual acquisition task, sometimes referred to as the model's cognitive plausibility. We suggest that though every computational model necessarily idealizes the modeled task, an informative language acquisition model can aim to be cognitively plausible in multiple ways. We discuss these cognitive plausibility checkpoints generally and then apply them to a case study in word segmentation, investigating a promising Bayesian segmentation strategy. We incorporate cognitive plausibility by using an age‐appropriate unit of perceptual representation, evaluating the model output in terms of its utility, and incorporating cognitive constraints into the inference process. Our more cognitively plausible model shows a beneficial effect of cognitive constraints on segmentation performance. One interpretation of this effect is as a synergy between the naive theories of language structure that infants may have and the cognitive constraints that limit the fidelity of their inference processes, where less accurate inference approximations are better when the underlying assumptions about how words are generated are less accurate. More generally, these results highlight the utility of incorporating cognitive plausibility more fully into computational models of language acquisition.     

The role of corpus size and syntax in deriving lexico-semantic representations for a wide range of concepts

One of the most significant recent advances in the study of semantic processing is the advent of models based on text and other corpora. In this study, we address what impact both the quantitative and qualitative properties of corpora have on mental representations derived from them. More precisely, we evaluate models with different linguistic and mental constraints on their ability to predict semantic relatedness between items from a vast range of domains and categories. We find that a model based on syntactic dependency relations captures significantly less of the variability for all kinds of words, regardless of the semantic relation between them or their abstractness. The largest difference was found for concrete nouns, which are commonly used to assess semantic processing. For both models we find that limited amounts of data suffice in order to obtain reliable predictions. Together, these findings suggest new constraints for the construction of mental models from corpora, both in terms of the corpus size and in terms of the linguistic properties that contribute to mental representations.     

Brain reflections of words and their meaning

The neurobiological organization of meaningful language units, morphemes and words, has been illuminated by recent metabolic and neurophysiological imaging studies. When humans process words from different categories, sets of cortical areas become active differentially. The meaning of a word, more precisely aspects of its reference, may be crucial for determining which set of cortical areas is involved in its processing. Word-related neuron webs with specific cortical distributions might underlie the observed category-specific differences in brain activity. Neuroscientific principles can explain these differential topographies.     

Word frequency and root-morpheme frequency effects on processing of Korean particle-suffixed words

Two experiments investigated the roles of the frequency of the root morpheme and the frequency of the whole word for a particular type of suffixed word in Korean in which the suffixed word can be thought of as a phrase (e.g., grandson-with). In both experiments, sentence frames were constructed so that they could have one of two target words that varied on frequency characteristics in the same location in the sentence. In Experiment 1, the frequency of the root morpheme was varied with the frequency of the word controlled, and in Experiment 2, the frequency of the word was varied with the frequency of the root morpheme controlled. Word frequency had a significant effect on fixation times, whereas root morpheme frequency did not. The results were surprising as native Korean speakers view the root morpheme as the “word” (analogous to how English readers would view a noun followed by a prepositional phrase).     

Cross‐Situational Learning of Phonologically Overlapping Words Across Degrees of Ambiguity

Cross‐situational word learning (XSWL) tasks present multiple words and candidate referents within a learning trial such that word–referent pairings can be inferred only across trials. Adults encode fine phonological detail when two words and candidate referents are presented in each learning trial (2 × 2 scenario; Escudero, Mulak, & Vlach, 2016a ). To test the relationship between XSWL task difficulty and phonological encoding, we examined XSWL of words differing by one vowel or consonant across degrees of within‐learning trial ambiguity (1 × 1 to 4 × 4). Word identification was assessed alongside three distractors. Adults finely encoded words via XSWL: Learning occurred in all conditions, though accuracy decreased across the 1 × 1 to 3 × 3 conditions. Accuracy was highest for the 1 × 1 condition, suggesting fast‐mapping is a stronger learning strategy here. Accuracy was higher for consonant than vowel set targets, and having more distractors from the same set mitigated identification of vowel set targets only, suggesting possible stronger encoding of consonants than vowels.     

Word order and information status in child language

In expressing rich, multi-dimensional thought in language, speakers are influenced by a range of factors that influence the ordering of utterance constituents. A fundamental principle that guides constituent ordering in adults has to do with information status, the accessibility of referents in discourse. Typically, adults order previously mentioned referents ("old" or accessible information) first, before they introduce referents that have not yet been mentioned in the discourse ("new" or inaccessible information) at both sentential and phrasal levels. Here we ask whether a similar principle influences ordering patterns at the phrasal level in children who are in the early stages of combining words productively. Prior research shows that when conveying semantic relations, children reproduce language-specific ordering patterns in the input, suggesting that they do not have a bias for any particular order to describe "who did what to whom". But our findings show that when they label "old" versus "new" referents, 3- to 5-year-old children prefer an ordering pattern opposite to that of adults (Study 1). Children's ordering preference is not derived from input patterns, as "old-before-new" is also the preferred order in caregivers' speech directed to young children (Study 2). Our findings demonstrate that a key principle governing ordering preferences in adults does not originate in early childhood, but develops: from new-to-old to old-to-new.     

Being forward not backward: lexical limits to masked priming

This study investigated whether masked priming is mediated by existing memory representations by determining whether nonwords targets would show repetition priming. To avoid the potential confound that nonword repetition priming would be obscured by a familiarity response bias, the standard lexical decision and naming tasks were modified to make targets unfamiliar. Participants were required to read a target string from right to left (i.e., "ECAF" should be read as "FACE") and then make a response. To examine if priming was based on lexical representations, repetition primes consisted of words when read forwards or backwards (e.g., "face", "ecaf") and nonwords (e.g., "pame", "emap"). Forward and backward primes were used to test if task instruction affected prime encoding. The lexical decision and naming tasks showed the same pattern of results: priming only occurred for forward primes with word targets (e.g., "face-ECAF"). Additional experiments to test if response priming affected the LDT indicated that the lexical status of the prime per se did not affect target responses. These results showed that the encoding of masked primes was unaffected by the novel task instruction and support the view that masked priming is due to the automatic triggering of pre-established computational processes based on stored information.