Search

Indonesian has two noun-forming prefixes, PE- and PEN-, that often stand in a paradigmatic relation to verbal base words with the prefixes BER- and MEN-. The central question addressed in the present study is whether the form similarities between PEN- and MEN- make PEN- easier to learn compared to PE-. To address this question, we made use of a computational model, the ‘discriminative lexicon’ (DL) model. We trained this model on 2,517 word forms that were inflected or derived variants of 99 different base words. Of these word forms, 109 were nouns with PE- and 221 words were nouns with PEN-. Both the production and the comprehension networks of the model performed with high accuracy for both prefixes. However, the model was able to provide more precise predictions for PE- as compared to PEN-, implying that PE- should have a processing advantage compared to PEN-. There are two reasons for why PE- is learned more robustly than PEN-. First, PE- words tend to be longer and hence have more discriminative triphones. Second, due to cue competition with MEN-, the prefixal triphones of PEN- are less effective cues than those of PE-. A measure of functional load is proposed that helps clarify the relative importance of the triphones in the prefixes and those straddling the boundary between prefix and stem. Our results shed further light on the productivity paradox, role of junctural phonotactics, and (dis)functionality of affix substitution.

This study addresses whether there is anything special about learning a third language, as compared to learning a second language, that results solely from the order of acquisition. We use a computational model based on the mathematical framework of Linear Discriminative Learning to explore this question for the acquisition of a small trilingual vocabulary, with English as L1, German or Mandarin as L2, and Mandarin or Dutch as L3. Our simulations reveal that when qualitative differences emerge between the learning of a first, second, and third language, these differences emerge from distributional properties of the particular languages involved rather than the order of acquisition per se, or any difference in learning mechanism. One such property is the number of homophones in each language, since within-language homophones give rise to errors in production. Our simulations also show the importance of suprasegmental information in determining the kinds of production errors made.

Using computational simulations, this work demonstrates that it is possible to learn a systematic relation between words' sound and their meanings. The sound-meaning relation was learned from a corpus of phonologically transcribed child-directed speech by using the linear discriminative learning (LDL) framework (Baayen, Chuang, Shafaei-Bajestan, & Blevins, 2019), which implements linear mappings between words' form vectors and semantic vectors. Presented with the form vectors of 16 nonwords, taken from a study on word learning (Fitneva, Christiansen, & Monaghan, 2009), the network generated the estimated semantic vectors of the nonwords. As half of these nonwords were created to phonologically resemble English nouns and the other half were phonologically similar to English verbs, we assessed whether the estimated semantic vectors for these nonwords reflect this word category difference. In 7 different simulations, linear discriminant analysis (LDA) successfully discriminated between noun-like nonwords and verb-like nonwords, based on their semantic relation to the words in the lexicon. Furthermore, how well LDA categorized a nonword correlated well with a phonological typicality measure (i.e., the degree of its form being noun-like or verb-like) and with children's performance in an entity/action discrimination task. On the one hand, the results suggest that children can infer the implicit meaning of a word directly from its sound. On the other hand, this study shows that nonwords do land in semantic space, such that children can capitalize on their semantic relations with other elements in the lexicon to decide whether a nonword is more likely to denote an entity or an action.

The goal of the present study is to understand the role orthographic and semantic information play in the behavior of skilled readers. Reading latencies from a self-paced sentence reading experiment in which Russian near-synonymous verbs were manipulated appear well-predicted by a combination of bottom-up sublexical letter triplets (trigraphs) and top-down semantic generalizations, modeled using the Naive Discrimination Learner. The results reveal a complex interplay of bottom-up and top-down support from orthography and semantics to the target verbs, whereby activations from orthography only are modulated by individual differences. Using performance on a serial reaction time (SRT) task for a novel operationalization of the mental speed hypothesis, we explain the observed individual differences in reading behavior in terms of the exploration/exploitation hypothesis from reinforcement learning, where initially slower and more variable behavior leads to better performance overall.

Word frequency is a strong predictor in most lexical processing tasks. Thus, any model of word recognition needs to account for howword frequency effects arise. The Discriminative Lexicon Model (DLM)models lexical processing withmappings between words' forms and their meanings. Comprehension and production are modeled via linear mappings between the two domains. So far, the mappings within the model can either be obtained incrementally via error-driven learning, a computationally expensive process able to capture frequency effects, or in an efficient, but frequency-agnostic solution modeling the theoretical endstate of learning (EL) where all words are learned optimally. In the present study we show how an efficient, yet frequency-informed mapping between form and meaning can be obtained (Frequency-informed learning; FIL). We find that FIL well approximates an incremental solution while being computationallymuch cheaper. FIL shows a relatively low type-and high token-accuracy, demonstrating that the model is able to process most word tokens encountered by speakers in daily life correctly. We use FIL to model reaction times in the Dutch Lexicon Project by means of a Gaussian Location Scale Model and find that FIL predicts well the S-shaped relationship between frequency and the mean of reaction times but underestimates the variance of reaction times for low frequency words. FIL is also better able to account for priming effects in an auditory lexical decision task in Mandarin Chinese, compared to EL. Finally, we used ordered data from CHILDES to compare mappings obtained with FIL and incremental learning. We show that themappings are highly correlated, but that with FIL some nuances based on word ordering effects are lost. Our results show how frequency effects in a learning model can be simulated efficiently, and raise questions about how to best account for low-frequency words in cognitive models. [ABSTRACT FROM AUTHOR]

Günther et al. (2022) investigated the relationship between words and images in which they concluded the possibility of a direct link between words and embodied experience. In their study, participants were presented with a target noun and a pair of images, one chosen by their model and another chosen randomly. Participants were asked to select the image that best matched the target noun. Building upon their work, we addressed the following questions. 1. Apart from utilizing visually embodied simulation, what other strategies subjects might have used? How much does this setup rely on visual information? Can it be solved using textual representations? 2. Do current visually-grounded embeddings explain subjects' selection behavior better than textual embeddings? 3. Does visual grounding improve the representations of both concrete and abstract words? For this aim, we designed novel experiments based on pre-trained word embeddings. Our experiments reveal that subjects' selection behavior is explained to a large extend on text-based embeddings and word-based similarities. Visually grounded embeddings offered modest advantages over textual embeddings in certain cases. These findings indicate that the experiment by Günther et al. (2022) may not be well suited for tapping into the perceptual experience of participants, and the extent to which it measures visually grounded knowledge is unclear. [ABSTRACT FROM AUTHOR]

Arnon and Snider ((2010). More than words: Frequency effects for multi-word phrases. "Journal of Memory and Language," 62, 67-82) documented frequency effects for compositional four-grams independently of the frequencies of lower-order "n"-grams. They argue that comprehenders apparently store frequency information about multi-word units. We show that "n"-gram frequency effects can emerge in a parameter-free computational model driven by naive discriminative learning, trained on a sample of 300,000 four-word phrases from the British National Corpus. The discriminative learning model is a full decomposition model, associating orthographic input features straightforwardly with meanings. The model does not make use of separate representations for derived or inflected words, nor for compounds, nor for phrases. Nevertheless, frequency effects are correctly predicted for all these linguistic units. Naive discriminative learning provides the simplest and most economical explanation for frequency effects in language processing, obviating the need to posit counters in the head for, and the existence of, hundreds of millions of "n"-gram representations.

Two auditory lexical decision experiments document for morphologically complex words two points at which the probability of a target word given the evidence shifts dramatically. The first point is reached when morphologically unrelated competitors are no longer compatible with the evidence. Adapting terminology from Marslen-Wilson (1984), we refer to this as the word's initial uniqueness point (UP1). The second point is the complex uniqueness point (CUP) introduced by Balling and Baayen (2008), at which morphologically related competitors become incompatible with the input. Later initial as well as complex uniqueness points predict longer response latencies. We argue that the effects of these uniqueness points arise due to the large surprisal (Levy, 2008) carried by the phonemes at these uniqueness points, and provide independent evidence that how cumulative surprisal builds up in the course of the word co-determines response latencies. The presence of effects of surprisal, both at the initial uniqueness point of complex words, and cumulatively throughout the word, challenges the Shortlist B model of Norris and McQueen (2008), and suggests that a Bayesian approach to auditory comprehension requires complementation from information theory in order to do justice to the cognitive cost of updating probability distributions over lexical candidates. (Contains 16 tables and 9 figures.)

Dutch children, from the second and fourth grade of primary school, were each given a visual lexical decision test on 210 Dutch monomorphemic words. After removing words not recognized by a majority of the younger group, (lexical) decisions were analysed by mixed-model regression methods to see whether morphological Family Size influenced decision times over and above several other covariates. The effect of morphological Family Size on decision time was mixed: larger families led to significantly faster decision times for the second graders but not for the fourth graders. Since facilitative effects on decision times had been found for adults, we offer a developmental account to explain the absence of an effect of Family Size on decision times for fourth graders. (Contains 3 tables and 4 footnotes.)

A 2-layer symbolic network model based on the equilibrium equations of the Rescorla-Wagner model (Danks, 2003) is proposed. The study first presents 2 experiments in Serbian, which reveal for sentential reading the inflectional paradigmatic effects previously observed by Milin, Filipovic Durdevic, and Moscoso del Prado Martin (2009) for unprimed lexical decision. The empirical results are successfully modeled without having to assume separate representations for inflections or data structures such as inflectional paradigms. In the next step, the same naive discriminative learning approach is pitted against a wide range of effects documented in the morphological processing literature. Frequency effects for complex words as well as for phrases (Arnon & Snider, 2010) emerge in the model without the presence of whole-word or whole-phrase representations. Family size effects (Moscoso del Prado Martin, Bertram, Haikio, Schreuder, & Baayen, 2004; Schreuder & Baayen, 1997) emerge in the simulations across simple words, derived words, and compounds, without derived words or compounds being represented as such. It is shown that for pseudo-derived words no special morpho-orthographic segmentation mechanism, as posited by Rastle, Davis, and New (2004), is required. The model also replicates the finding of Plag and Baayen (2009) that, on average, words with more productive affixes elicit longer response latencies; at the same time, it predicts that productive affixes afford faster response latencies for new words. English phrasal paradigmatic effects modulating isolated word reading are reported and modeled, showing that the paradigmatic effects characterizing Serbian case inflection have crosslinguistic scope. (Contains 1 footnote, 23 tables, and 13 figures.)

This eye-tracking study explores visual recognition of Dutch suffixed words (e.g., "plaats+ing" "placing") embedded in sentential contexts, and provides new evidence on the interplay between storage and computation in morphological processing. We show that suffix length crucially moderates the use of morphological properties. In words with shorter suffixes, we observe a stronger effect of full-forms (derived word frequency) on reading times than in words with longer suffixes. Also, processing times increase if the base word ("plaats") and the suffix ("-ing") differ in the amount of information carried by their morphological families (sets of words that share the base or the suffix). We model this imbalance of informativeness in the morphological families with the information-theoretical measure of relative entropy and demonstrate its predictivity for the processing times. The observed processing trade-offs are discussed in the context of current models of morphological processing. (Contains 4 tables and 2 figures.)

In this study, we investigate the processing of morphologically complex words in Danish using auditory lexical decision. We document a second critical point in auditory comprehension in addition to the Uniqueness Point (UP), namely the point at which competing morphological continuation forms of the base cease to be compatible with the input, henceforth the Complex Uniqueness Point (CUP). Suffixed words with later CUP elicited longer response latencies. We also observed an interaction between suffix frequency and whole-word frequency. Both suffix and whole-word frequency were facilitatory, except for words for which both frequencies are high. For such words, we observed inhibition, and most clearly so for female compared with male participants. Finally, a comparison of complex with simple words revealed that, other things being equal, complex words have a processing advantage compared to simple words. We discuss the consequences of these findings for models of morphological processing. (Contains 6 figures and 7 tables.)

This paper explores the time-course of morphological processing of trimorphemic Finnish compounds. We find evidence for the parallel access to full-forms and morphological constituents diagnosed by the early effects of compound frequency, as well as early effects of left constituent frequency and family size. We also observe an interaction between compound frequency and both the left and the right constituent family sizes. Furthermore, our data show that suffixes embedded in the derived left constituent of a compound are efficiently used for establishing the boundary between compounds' constituents. The success of segmentation of a compound is demonstrably modulated by the affixal salience of the embedded suffixes. We discuss implications of these findings for current models of morphological processing and propose a new model that views morphemes, combinations of morphemes and morphological paradigms as probabilistic sources of information that are interactively used in recognition of complex words. (Contains 3 figures, 8 footnotes and 9 tables.)

Besides phonotactic principles, orthographies entail graphotactic rules for which the reader must convert a phonological representation on the basis of spelling adaptation rules. In the present study, the learnability of such rules will be investigated with reference to Dutch. Although Dutch orthography can be considered highly regular, there are graphotactic rules that change letter sequences in plural noun formation. In a lexical decision experiment, the acquisition and use of such rules were examined. Participants were groups of 31 children from Grade 3 and 34 children from Grade 6, and 25 adults. The results showed that both children and adults are significantly less accurate and slower in recognizing plural word forms which undergo vowel change as a consequence of pluralization. It is concluded that graphotactic rules in Dutch orthography complicate Dutch word identification from an early stage of development and continue to play a complicating role in the word identification process of adult readers. In the discussion it is shown that current models fail to fully explain the processing of graphotactic rules in visual word identification.

A computational model for the comprehension of single spoken words is presented that builds on an earlier model using discriminative learning. Real-valued features are extracted from the speech signal instead of discrete features. Vectors representing word meanings using one-hot encoding are replaced by real-valued semantic vectors. Instead of incremental learning with Rescorla-Wagner updating, we use linear discriminative learning, which captures incremental learning at the limit of experience. These new design features substantially improve prediction accuracy for unseen words, and provide enhanced temporal granularity, enabling the modelling of cohort-like effects. Visualisation with t-SNE shows that the acoustic form space captures phone-like properties. Trained on 9 h of audio from a broadcast news corpus, the model achieves recognition performance that approximates the lower bound of human accuracy in isolated word recognition tasks. LDL-AURIS thus provides a mathematically-simple yet powerful characterisation of the comprehension of single words as found in English spontaneous speech. [ABSTRACT FROM AUTHOR]

Presents results of four experiments that show that verbs have family size effect independently of nominal conversion alternants, that this effect is a strict type frequency effect and not a token frequency effect, that the effect is co-determined by the morphological structure of the inflected verb, and that it occurs irrespective of the orthographic shape of the base word. (Author/VWL)

Indonesian has two prefixes, PE- and PEN-, that are similar in form and meaning, but are probably not allomorphs. In this study, we applied a distributional vector space model to clarify whether these prefixes have discriminable semantics. Comparisons of pairs of words within and across morphologically defined sets of words revealed that cosine similarities of pairs consisting of a word with PE- and a word with PEN- were reduced compared to pairs of only PE- words, or of only PEN- words. Furthermore, nouns with PE- were more similar to their base words than was the case for words with PEN-. The specialized use of PE- for words denoting agents, and the specialized use of PEN- for denoting instruments, was also visible in the semantic vector space. These differences in the semantics of PE- and PEN- thus provide further quantitative support for the independent status of PE- as opposed to PEN-. [ABSTRACT FROM AUTHOR]

In structuralist linguistics, compounds are argued not to constitute morphological categories, due to the absence of systematic form-meaning correspondences. This study investigates subsets of compounds for which systematic form-meaning correspondences are present: adjective–noun compounds in Mandarin. We show that there are substantial differences in the productivity of these compounds. One set of productivity measures (the count of types, the count of hapax legomena, and the estimated count of unseen types) reflect compounds' profitability. By contrast, the category-conditioned degree of productivity is found to correlate with the internal semantic transparency of the words belonging to a morphological category. Greater semantic transparency, gauged by distributional semantics, predicts greater category-conditioned productivity. This dovetails well with the hypothesis that semantic transparency is a prerequisite for a word formation process to be productive. [ABSTRACT FROM AUTHOR]

Finnish nouns are characterized by rich inflectional variation, with obligatory marking of case and number, with optional possessive suffixes and with the possibility of further cliticization. We present a model for the conceptualization of Finnish inflected nouns, using pre-compiled fasttext embeddings (300-dimensional semantic vectors that approximate words' meanings). Instead of deriving the semantic vector of an inflected word from another word in its paradigm, we propose that an inflected word is conceptualized by means of summation of latent vectors representing the meanings of its lexeme and its inflectional features. We tested this model on the 2,000 most frequent Finnish nouns and their inflected word forms from a corpus of Finnish (84 million tokens). Visualization of the semantic space of Finnish using t-SNE clarified that a 'main effects' additive model does not do justice to the semantics of inflection. In Finnish, how number is realized turns out to vary substantially with case. Further interactions emerged with the possessive suffixes and the clitics. By taking these interactions into account, the accuracy of our model, evaluated with the fasttext embeddings as gold standard, improved from 76% to 89%. Analyses of the errors made by the model clarified that 7.5% of errors are due to overabundance (and hence not true errors), and that 16.5% of the errors involved exchanges of semantically highly similar stems (lexemes). Our results indicate, first, that the semantics of Finnish noun inflection are more intricate than assumed thus far, and second, that these intricacies can be captured with surprisingly high accuracy by a simple generating model based on imputed semantic vectors for lexemes, inflectional features, and interactions of inflectional features. [ABSTRACT FROM AUTHOR]

This study investigates the phenomenon of defectiveness in Russian case and number noun paradigms from the perspective of distributional semantics. We made use of word embeddings, high-dimensional vectors trained from large text corpora, and compared the observed paradigms of nouns that are defective in the genitive plural, as suggested by Zaliznjak (1977), with the observed paradigms for non-defective nouns. When the embeddings of about 20,000 inflected forms were projected onto a two-dimensional space, clusters of case and number within case were found, suggesting global semantic similarity for words with the same inflectional features. Moreover, defective lexemes were characterized by lower semantic transparency, in that inflected forms of the same lexeme are semantically less similar to each other, and their meanings are also more idiosyncratic. Furthermore, compared to non-defective lexemes, inflected forms from defective lexemes are further away from the idealized average case-number meanings, obtained by averaging over the vectors of all inflected forms of the same case-number combination. As a consequence, the semantics of defective forms are predicted less precisely by a simple model of conceptualization that assumes that the meaning of a given Russian inflected form is approximated well by the sum of pertinent embeddings of the lexeme, case, and number within case. We conclude that the relationship between defectiveness and semantics, at least the kind captured by word embeddings, is stronger than has been anticipated previously. [ABSTRACT FROM AUTHOR]

This study addresses the relation between morphological productivity and semantic transparency. Using distributional semantics, we compare German word formation using particles with derivational word formation. We observed that derivational suffixes, but not particles, tend to make strong independent semantic contributions to their carrier words. In two-dimensional t-SNE maps, complex words show clustering by affix, but not by particle. Furthermore, the semantic vectors of suffixed words are predictable from their base words with higher accuracy than is possible for particle verbs. For particle verbs, but not affixed verbs, semantic similarity within the set of complex words correlated negatively with the number of types. Furthermore, only for particle verbs, a greater number of observed types predicted a reduced probability of observing unseen types. We propose that particle verbs primarily serve the onomasiological function of labeling, resulting in relatively idiosyncratic semantic vectors. By contrast, words sharing derivational affixes form distinct clusters in semantic space while maintaining strong and consistent semantic relations with their base words. This enables these words to serve not only as labels, but also allows them to be used with an anaphoric function in discourse. [ABSTRACT FROM AUTHOR]

We used word embeddings to study the relation between productivity and semantic transparency. We compiled a dataset with around 2700 two-syllable compounds that shared position-specific constituents (henceforth pivots) and some 1100 suffixed words. For each pivot and suffix, we calculated measures of productivity as well as measures of semantic transparency. For compounds, productivity (P) was negatively correlated with the number of types (V) and with the semantic similarity between non-pivot constituents and their compounds. Conversely, the greater semantic similarity of the pivot with either the compound or the non-pivot constituent predicted higher degrees of productivity. Visualization with t-SNE revealed clustering of suffixed words' embeddings, but no by-pivot clustering for compounds, except for a minority of pivots whose regions in semantic space did not contain intruding unrelated compounds. A subset of these pivots was found to realize a fixed shift in semantic space from the base word to the corresponding compound, a property that also emerged for several suffixes. For these pivots, no correlation between P and V was present. Thus, Mandarin compounds appear to realize, at one extreme, motivated but unsystematic concept formation (where other pivots could just as well have been used), and at the other extreme, systematic suffix-like semantics. [ABSTRACT FROM AUTHOR]

As a second step, the word-document frequency table is subjected to a dimension reduction technique (singular value decomposition), resulting in a matrix of words by I n i latent dimensions. As a consequence, the vector with a word's document frequencies provides a semantic fingerprint of that word. Clearly, representing words' "meanings" as high-dimensional numeric vectors informed by how words are used in text has significant advantages. Distributional semantics, a research area at the intersection of artificial intelligence, psychology, and computational semantics, represents words' meanings by means of high-dimensional vectors of real numbers calculated from large corpora. [Extracted from the article]