scholarly journals Modeling morphological priming in German with naive discriminative learning

2019 ◽  
Author(s):  
R. H. Baayen ◽  
Eva Smolka
Keyword(s):  
Discriminative Learning ◽  
Connectionist Models ◽  
Semantic Similarity Measure ◽  
Computational Implementation ◽  
Morphological Priming ◽  
Semantic Compositionality ◽  
Set Up ◽  
Complex Words ◽  
Human Association ◽  
Prime Target

Both localist and connectionist models, based on experimental results obtained for English and French, assume that the degree of semantic compositionality of a morphologically complex word is reflected in how it is processed. Since priming experiments using English and French morphologically related prime-target pairs reveal stronger priming when complex words are semantically transparent (e.g., refill–fill ) compared to semantically more opaque pairs (e.g., restrain–strain), localist models set up connections between complex words and their stems only for semantically transparent pairs. Connectionist models have argued that the effect of transparency should arise as an epiphenomenon in PDP networks. However, for German, a series of studies has revealed equivalent priming for both transparent and opaque prime-target pairs, which suggests mediation of lexical access by the stem, independent of degrees of semantic compositionality. This study reports a priming experiment that replicates equivalent priming for transparent and opaque pairs. We show that these behavioral results can be straightforwardly modeled by a computational implementation of Word and Paradigm Morphology (WPM), Naive Discriminative Learning (NDL). Just as wpm, ndl eschews the theoretical construct of the morpheme. Ndl succeeds in modeling the German priming data by inspecting the extent to which a discrimination network pre-activates the target lexome from the orthographic properties of the prime. Measures derived from an ndl network, complemented with a semantic similarity measure derived from distributional semantics, predict lexical decision latencies with somewhat improved precision compared to classical measures such as word frequency, prime type, and human association ratings. We discuss both the methodological implications of our results, as well as their implications for models of the mental lexicon.

scholarly journals Modeling Maltese Noun Plural Classes without Morphemes

2020 ◽  
Author(s):  
Jessica Nieder ◽  
Fabian Tomaschek ◽  
Enum Cohrs ◽  
Ruben van de Vijver
Keyword(s):  
Neural Networks ◽  
Mental Lexicon ◽  
Classification Model ◽  
Discriminative Learning ◽  
Plural Noun ◽  
Model Classification ◽  
Computational Implementation ◽  
Semitic Language ◽  
Word Forms ◽  
Complex Words

Word and Paradigm morphology proposes that morphologically complex words are stored as wholes in the mental lexicon, and by doing so it avoids problems that are related to the notion of the morpheme. However, it is not yet clear to what degree it is possible to computationally model classification and production of complex word forms without morphemes.We take up this question by modeling the classification and production of the Maltese noun plural system. Maltese is a Semitic language that has two broad classes of plurals: concatenative ones and non-concatenative ones. We model the classification in two models, and the production in a third one. The first model for classification, the Tilburg Memory Based Learner (TiMBL), is a computational implementation of exemplar models. TiMBL is impartial as to the existence of morphemes in the mental lexicon, and allows us to directly compare the classification with and without morphemes. It turns out that the classification with and without morphemes of Maltese nouns is equally good. The second classification model is the Naive Discriminative Learner (NDL). NDL is a computational implementation of discriminative learning. It can be understood as an implementation of the Word and Paradigm model and thus eschews morphemes. It differs from TiMBL in its assumptions about learning, and they way in which the classifications are obtained. NDL, too, classifies Maltese nouns well. A classification task is very different from a production task, and therefore we used neural networks to model the production of plurals. In these models we address the question whether the production of a plural noun for a given singular can be modeled without recourse to morphemes. We used two neural networks architectures (LSTM and GRU) that have been applied to linguistic phenomena, and find that they are able to correctly produce plurals, without making use of morphemes.We conclude that the Maltese noun plural system can be modeled on the basis of whole words alone without morphemes. These results, therefore, support the Word and Paradigm theory of the mental lexicon.

The processing of morphologically complex words in a specific speaker group

2016 ◽  
Vol 11 (2) ◽  
pp. 308-328 ◽  
Author(s):  
Gunnar Jacob ◽  
Bilal Kırkıcı
Keyword(s):  
Comparison Group ◽  
Native Speakers ◽  
Masked Priming ◽  
Morphological Processing ◽  
Heritage Speakers ◽  
Priming Effects ◽  
Morphological Priming ◽  
L2 Learners ◽  
Complex Words ◽  
Prime Target

The present study investigates to what extent morphological priming varies across different groups of native speakers of a language. In two masked-priming experiments, we investigate the processing of morphologically complex Turkish words in Turkish heritage speakers raised and living in Germany. Materials and experimental design were based on Kırkıcı and Clahsen’s (2013) study on morphological processing in Turkish native speakers and L2 learners, allowing for direct comparisons between the three groups. Experiment 1 investigated priming effects for morphologically related prime-target pairs. Heritage speakers showed a similar pattern of results as the L1 comparison group, with significant priming effects for prime-target pairs with inflected primes (e.g. ‘sorar-sor’ asks-ask) as well as for prime-target pairs with derived primes (e.g. ‘sağlık-sağ’ health-healthy). In Experiment 2, we measured priming effects for prime-target pairs which were semantically and morphologically unrelated, but only related with regard to orthographic overlap (e.g. ‘devre-dev’ period-giant). Unlike both L1 speakers raised in Turkey and highly proficient L2 learners, heritage speakers also showed significant priming effects in this condition. Our results suggest that heritage speakers differ from both native speakers and L2 learners in that they rely more on (orthographic) surface form properties of the stimulus during early stages of word recognition, at the expense of morphological decomposition.

Morphological decomposition in native and non-native French speakers

2014 ◽  
Vol 18 (3) ◽  
pp. 524-542 ◽  
Author(s):  
CAITLIN E. COUGHLIN ◽  
ANNIE TREMBLAY
Keyword(s):  
Native Speakers ◽  
Naming Task ◽  
Native English Speakers ◽  
Masked Priming ◽  
English Speakers ◽  
Morphological Priming ◽  
L2 Learners ◽  
Complex Words ◽  
Orthographic Priming ◽  
Prime Target

This study investigates whether late second/foreign language (L2) learners can rely on mechanisms similar to those of native speakers for processing morphologically complex words. Specifically, it examines whether native English speakers who have begun learning French around the onset of puberty can decompose -er (Class I) French verbs. Mid-to-high-proficiency L2 learners and native French speakers completed a masked-priming word-naming task. Latencies for morphologically related, orthographically related, and semantically related prime–target combinations were compared to latencies for identical and unrelated prime–target combinations. The results reveal the following effects: full morphological priming for both native and non-native speakers, with this effect increasing with French proficiency for L2 learners; partial orthographic priming for both groups; greater priming in the morphological condition than in the orthographic condition for both groups; and no semantic priming for either group. We conclude that L2 learners have access to similar mechanisms to those of native speakers for processing morphologically complex words.

Inflectional and Derivational Morphology in the Mental Lexicon: Symmetries and Asymmetries in Repetition Priming

1992 ◽  
Vol 44 (2) ◽  
pp. 373-390 ◽  
Author(s):  
H. Schriefers ◽  
A. Friederici ◽  
P. Graetz
Keyword(s):  
Repetition Priming ◽  
Mental Lexicon ◽  
Derivational Morphology ◽  
Lexical Organization ◽  
Priming Effects ◽  
Priming Paradigm ◽  
Complex Forms ◽  
Complex Words ◽  
Prime Target

Using a repetition priming paradigm, the interrelations between morphologically related words in the mental lexicon were examined in two experiments. In contrast to most previous studies, in which morphologically complex words occur as primes and stems as targets, derivationally and inflectionally complex forms were fully crossed in prime–target pairs. Experiment 1 showed asymmetries in the pattern of priming effects between different inflectional forms of German adjectives. Such asymmetries are problematic for any theory that assumes that all members of an inflectional paradigm share one entry in the mental lexicon. Experiment 2 contrasted derivational and inflectional variants of the same stems used in Experiment 1. Once again, there were same clear asymmetries in the pattern of priming effects. The implications of these results for models of lexical organization of inflectional and derivational morphology are discussed.

Discriminative Learning and the Lexicon: NDL and LDL

2021 ◽  
Author(s):  
Yu-Ying Chuang ◽  
R. Harald Baayen
Keyword(s):  
Lexical Processing ◽  
Morphological Processing ◽  
Discriminative Learning ◽  
Computational Algorithms ◽  
Association Strength ◽  
Computational Theory ◽  
Wide Range ◽  
Multivariate Multiple Regression ◽  
Lexical Learning ◽  
Set Up

Naive discriminative learning (NDL) and linear discriminative learning (LDL) are simple computational algorithms for lexical learning and lexical processing. Both NDL and LDL assume that learning is discriminative, driven by prediction error, and that it is this error that calibrates the association strength between input and output representations. Both words’ forms and their meanings are represented by numeric vectors, and mappings between forms and meanings are set up. For comprehension, form vectors predict meaning vectors. For production, meaning vectors map onto form vectors. These mappings can be learned incrementally, approximating how children learn the words of their language. Alternatively, optimal mappings representing the end state of learning can be estimated. The NDL and LDL algorithms are incorporated in a computational theory of the mental lexicon, the ‘discriminative lexicon’. The model shows good performance both with respect to production and comprehension accuracy, and for predicting aspects of lexical processing, including morphological processing, across a wide range of experiments. Since, mathematically, NDL and LDL implement multivariate multiple regression, the ‘discriminative lexicon’ provides a cognitively motivated statistical modeling approach to lexical processing.

Psycholinguistic Approaches to Morphology: Theoretical Issues

2017 ◽  
Author(s):  
Christina L. Gagné
Keyword(s):  
Human Mind ◽  
Phonological Representations ◽  
Word Structure ◽  
Theoretical Approaches ◽  
Morphological Priming ◽  
Types Of Information ◽  
Complex Words ◽  
Whole Word ◽  
Theoretical Issues

Psycholinguistics is the study of how language is acquired, represented, and used by the human mind; it draws on knowledge about both language and cognitive processes. A central topic of debate in psycholinguistics concerns the balance between storage and processing. This debate is especially evident in research concerning morphology, which is the study of word structure, and several theoretical issues have arisen concerning the question of how (or whether) morphology is represented and what function morphology serves in the processing of complex words. Five theoretical approaches have emerged that differ substantially in the emphasis placed on the role of morphemic representations during the processing of morphologically complex words. The first approach minimizes processing by positing that all words, even morphologically complex ones, are stored and recognized as whole units, without the use of morphemic representations. The second approach posits that words are represented and processed in terms of morphemic units. The third approach is a mixture of the first two approaches and posits that a whole-access route and decomposition route operate in parallel. A fourth approach posits that both whole word representations and morphemic representations are used, and that these two types of information interact. A fifth approach proposes that morphology is not explicitly represented, but rather, emerges from the co-activation of orthographic/phonological representations and semantic representations. These competing approaches have been evaluated using a wide variety of empirical methods examining, for example, morphological priming, the role of constituent and word frequency, and the role of morphemic position. For the most part, the evidence points to the involvement of morphological representations during the processing of complex words. However, the specific way in which these representations are used is not yet fully known.

Semantic transparency affects morphological priming . . . eventually

2018 ◽  
Vol 71 (5) ◽  
pp. 1112-1124 ◽  
Author(s):  
Vera Heyer ◽  
Dana Kornishova
Keyword(s):  
Time Course ◽  
Semantic Information ◽  
Native Speakers ◽  
Masked Priming ◽  
Stimulus Onset ◽  
Semantic Transparency ◽  
Psycholinguistic Research ◽  
Priming Effects ◽  
Morphological Priming ◽  
Complex Words

Semantic transparency has been in the focus of psycholinguistic research for decades, with the controversy about the time course of the application of morpho-semantic information during the processing of morphologically complex words not yet resolved. This study reports two masked priming studies with English - ness and Russian - ost’ nominalisations, investigating how semantic transparency modulates native speakers’ morphological priming effects at short and long stimulus onset asynchronies (SOAs). In both languages, we found increased morphological priming for nominalisations at the transparent end of the scale (e.g. paleness – pale) in comparison to items at the opaque end of the scale (e.g. business – busy) but only at longer prime durations. The present findings are in line with models that posit an initial phase of morpho-orthographic (semantically blind) decomposition.

Masked morphological priming of compound constituents

2009 ◽  
Vol 4 (2) ◽  
pp. 159-193 ◽  
Author(s):  
Robert Fiorentino ◽  
Ella Fund-Reznicek
Keyword(s):  
Word Recognition ◽  
Visual Word Recognition ◽  
Masked Priming ◽  
Word Formation ◽  
Visual Word ◽  
Regular Form ◽  
Morphological Priming ◽  
Complex Words ◽  
Open Class ◽  
New Evidence

Recent masked priming studies suggest that complex words are rapidly segmented into potential morphological constituents during initial visual word recognition. Much of this evidence involves affixation or other formally regular operations, leaving open the question of whether these effects rely heavily on the identification of a closed-class affix or other formal regularity. In two masked priming experiments with English transparent and opaque bimorphemic compound primes consisting solely of open-class morphemes, we find significant constituent priming, but no significant priming for purely orthographic overlap. We conclude that masked morphological priming generalizes across word-formation types to include compounds with no affix or other regular form. These results provide new evidence for across-the-board morphological-level segmentation during visual word recognition and for morpheme-based compound processing.

scholarly journals Evidence for Early Morphological Decomposition: Combining Masked Priming with Magnetoencephalography

2011 ◽  
Vol 23 (11) ◽  
pp. 3366-3379 ◽  
Author(s):  
Minna Lehtonen ◽  
Philip J. Monahan ◽  
David Poeppel
Keyword(s):  
Time Course ◽  
Lexical Processing ◽  
Morphological Structure ◽  
Masked Priming ◽  
Brain Regions ◽  
Decision Task ◽  
Morphological Decomposition ◽  
Lexical Representations ◽  
Complex Words ◽  
Prime Target

Are words stored as morphologically structured representations? If so, when during word recognition are morphological pieces accessed? Recent masked priming studies support models that assume early decomposition of (potentially) morphologically complex words. The electrophysiological evidence, however, is inconsistent. We combined masked morphological priming with magneto-encephalography (MEG), a technique particularly adept at indexing processes involved in lexical access. The latency of an MEG component peaking, on average, 220 msec post-onset of the target in left occipito-temporal brain regions was found to be sensitive to the morphological prime–target relationship under masked priming conditions in a visual lexical decision task. Shorter latencies for related than unrelated conditions were observed both for semantically transparent (cleaner–CLEAN) and opaque (corner–CORN) prime–target pairs, but not for prime–target pairs with only an orthographic relationship (brothel–BROTH). These effects are likely to reflect a prelexical level of processing where form-based representations of stems and affixes are represented and are in contrast to models positing no morphological structure in lexical representations. Moreover, we present data regarding the transitional probability from stem to affix in a post hoc comparison, which suggests that this factor may modulate early morphological decomposition, particularly for opaque words. The timing of a robust MEG component sensitive to the morphological relatedness of prime–target pairs can be used to further understand the neural substrates and the time course of lexical processing.

scholarly journals Morphological priming of inflectional suffixes

2020 ◽  
Author(s):  
Katarina Marjanovic ◽  
Davide Crepaldi
Keyword(s):  
Priming Effect ◽  
Word Identification ◽  
Masked Priming ◽  
Visual Word ◽  
Priming Paradigm ◽  
Morphological Priming ◽  
Visual Word Identification ◽  
The Rich ◽  
Complex Words

Morphologically complex words are processed through their constituent morphemes during visual word recognition. While this has been primarily established through the stem priming paradigm, the role of shared affixes is more controversial. Also, most evidence on affix priming comes from derivation, while inflectional priming remains largely unaddressed. Here we present two lexical decision, masked priming experiments filling this gap. Taking advantage of the rich inflectional pattern of Slovene, we assessed inflectional suffix priming (mestam–HALJAM), and compared it to the well-established stem priming effect (haljov–HALJAM): while the latter is solid as expected, the former seems to be weak to non–existing. Results further indicate that there is no interaction between sharing a stem and sharing an inflectional suffix—neither stem nor suffix priming is boosted when primes and targets also share the other morpheme. These data indicate an important difference between stems, derivational affixes and inflectional affixes, which we consider in the context of models of visual word identification and information theory.

Sign in / Sign up

Export Citation Format

Share Document