TY - JOUR A1 - Rabovsky, Milena T1 - Change in a probabilistic representation of meaning can account for N400 effects on articles BT - a neural network model JF - Neuropsychologia : an international journal in behavioural and cognitive neuroscience N2 - Increased N400 amplitudes on indefinite articles (a/an) incompatible with expected nouns have been initially taken as strong evidence for probabilistic pre-activation of phonological word forms, and recently been intensely debated because they have been difficult to replicate. Here, these effects are simulated using a neural network model of sentence comprehension that we previously used to simulate a broad range of empirical N400 effects. The model produces the effects when the cue validity of the articles concerning upcoming noun meaning in the learning environment is high, but fails to produce the effects when the cue validity of the articles is low due to adjectives presented between articles and nouns during training. These simulations provide insight into one of the factors potentially contributing to the small size of the effects in empirical studies and generate predictions for cross-linguistic differences in article induced N400 effects based on articles’ cue validity. The model accounts for article induced N400 effects without assuming pre-activation of word forms, and instead simulates these effects as the stimulus-induced change in a probabilistic representation of meaning corresponding to an implicit semantic prediction error. KW - N400 KW - ERPs KW - prediction KW - neural networks KW - cue validity KW - meaning Y1 - 2020 U6 - https://doi.org/10.1016/j.neuropsychologia.2020.107466 SN - 0028-3932 SN - 1873-3514 VL - 143 PB - Elsevier CY - Oxford ER - TY - JOUR A1 - Rabovsky, Milena T1 - Change in a probabilistic representation of meaning can account for N400 effects on articles: a neural network model JF - Neuropsychologia N2 - Increased N400 amplitudes on indefinite articles (a/an) incompatible with expected nouns have been initially taken as strong evidence for probabilistic pre-activation of phonological word forms, and recently been intensely debated because they have been difficult to replicate. Here, these effects are simulated using a neural network model of sentence comprehension that we previously used to simulate a broad range of empirical N400 effects. The model produces the effects when the cue validity of the articles concerning upcoming noun meaning in the learning environment is high, but fails to produce the effects when the cue validity of the articles is low due to adjectives presented between articles and nouns during training. These simulations provide insight into one of the factors potentially contributing to the small size of the effects in empirical studies and generate predictions for cross-linguistic differences in article induced N400 effects based on articles’ cue validity. The model accounts for article induced N400 effects without assuming pre-activation of word forms, and instead simulates these effects as the stimulus-induced change in a probabilistic representation of meaning corresponding to an implicit semantic prediction error. KW - N400 KW - ERPs KW - prediction KW - neural networks KW - cue validity KW - meaning Y1 - 2019 VL - 143 PB - Elsevier CY - Amsterdam ER - TY - GEN A1 - Rabovsky, Milena T1 - Change in a probabilistic representation of meaning can account for N400 effects on articles: a neural network model T2 - Postprints der Universität Potsdam : Humanwissenschaftliche Reihe N2 - Increased N400 amplitudes on indefinite articles (a/an) incompatible with expected nouns have been initially taken as strong evidence for probabilistic pre-activation of phonological word forms, and recently been intensely debated because they have been difficult to replicate. Here, these effects are simulated using a neural network model of sentence comprehension that we previously used to simulate a broad range of empirical N400 effects. The model produces the effects when the cue validity of the articles concerning upcoming noun meaning in the learning environment is high, but fails to produce the effects when the cue validity of the articles is low due to adjectives presented between articles and nouns during training. These simulations provide insight into one of the factors potentially contributing to the small size of the effects in empirical studies and generate predictions for cross-linguistic differences in article induced N400 effects based on articles’ cue validity. The model accounts for article induced N400 effects without assuming pre-activation of word forms, and instead simulates these effects as the stimulus-induced change in a probabilistic representation of meaning corresponding to an implicit semantic prediction error. T3 - Zweitveröffentlichungen der Universität Potsdam : Humanwissenschaftliche Reihe - 731 KW - N400 KW - ERPs KW - prediction KW - neural networks KW - cue validity KW - meaning Y1 - 2019 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:517-opus4-526988 SN - 1866-8364 ER - TY - JOUR A1 - Haßler, Gerda T1 - Meaning and Function BT - the Place of Coseriu’s Linguistic Theory in Functional Linguistics JF - Concordia discors vs. discordia concors : researches into comparative literature, contrastive linguistics, translation and cross-cultural strategies N2 - The use of the word functional in the most diverse theories and approaches has contributed in no small measure to the confusion in linguistics today. This article does not claim to give an overview of the different directions of functionalism in linguistics. Rather, the aim is to present what Coseriu‘s view characterised as functional in his time and to what extent his theory outlined a path that still makes sense in functional-cognitive linguistics today. This will involve an examination of Coseriu‘s difficult-to-identify concept of function. Furthermore, the article will also show that functional thinking is relevant for current grammatography. KW - function KW - meaning KW - grammatography KW - oppositions KW - functional unit Y1 - 2022 SN - 2065-4057 SN - 2457-8835 VL - 16 IS - 2 SP - 17 EP - 50 PB - Ştefan cel Mare University Press CY - Suceava ER - TY - JOUR A1 - Hassler, Gerda T1 - Meaning and Function BT - the Place of Coseriu’s Linguistic Theory in Functional Linguistics JF - Concordia discors vs. discordia concors : international journal for researches into comparative literature, contrastive linguistics, cross-cultural and translation strategies N2 - The use of the word functional in the most diverse theories and approaches has contributed in no small measure to the confusion in linguistics today. This article does not claim to give an overview of the different directions of functionalism in linguistics. Rather, the aim is to present what Coseriu‘s view characterised as functional in his time and to what extent his theory outlined a path that still makes sense in functional-cognitive linguistics today. This will involve an examination of Coseriu‘s difficult-to-identify concept of function. Furthermore, the article will also show that functional thinking is relevant for current grammatography. KW - function KW - meaning KW - grammatography KW - oppositions KW - functional unit Y1 - 2021 SN - 2065-4057 SN - 2457-8835 IS - 16 SP - 17 EP - 50 PB - Ştefan cel Mare University Press CY - Suceava ER - TY - JOUR A1 - Rabovsky, Milena A1 - McClelland, James L. T1 - Quasi-compositional mapping from form to meaning BT - a neural network-based approach to capturing neural responses during human language comprehension JF - Philosophical transactions of the Royal Society of London : B, Biological sciences N2 - We argue that natural language can be usefully described as quasi-compositional and we suggest that deep learning-based neural language models bear long-term promise to capture how language conveys meaning. We also note that a successful account of human language processing should explain both the outcome of the comprehension process and the continuous internal processes underlying this performance. These points motivate our discussion of a neural network model of sentence comprehension, the Sentence Gestalt model, which we have used to account for the N400 component of the event-related brain potential (ERP), which tracks meaning processing as it happens in real time. The model, which shares features with recent deep learning-based language models, simulates N400 amplitude as the automatic update of a probabilistic representation of the situation or event described by the sentence, corresponding to a temporal difference learning signal at the level of meaning. We suggest that this process happens relatively automatically, and that sometimes a more-controlled attention-dependent process is necessary for successful comprehension, which may be reflected in the subsequent P600 ERP component. We relate this account to current deep learning models as well as classic linguistic theory, and use it to illustrate a domain general perspective on some specific linguistic operations postulated based on compositional analyses of natural language. This article is part of the theme issue 'Towards mechanistic models of meaning composition'. KW - language KW - meaning KW - event-related brain potentials KW - neural networks KW - N400 KW - P600 Y1 - 2019 U6 - https://doi.org/10.1098/rstb.2019.0313 SN - 0962-8436 SN - 1471-2970 SN - 0080-4622 VL - 375 IS - 1791 PB - Royal Society CY - London ER - TY - THES A1 - Hümmer, Christiane T1 - Synonymie bei phraseologischen Einheiten : eine korpusbasierte Untersuchung T1 - Synonymy of Multi-Word Expressions : a corpus-based study N2 - Die vorliegende Dissertation widmet sich dem Thema Synonymie im besonderen Fall der phraseologischen Einheiten. Es handelt sich um eine korpusbasierte, empirische Untersuchung, die sich insbesondere mit der Frage beschäftigt, ob und inwiefern es möglich ist, sich typisch semantischen Kategorien wie Bedeutung, Idiomatizität, Bildlichkeit etc. über die Untersuchung typischer Verwendungsmuster zu nähern. Diese Themenstellung motiviert sich aus bisher in der linguistischen Literatur strittigen grundsätzlichen Aspekten der Diskussion um Bedeutung und Synonymie: Sie ist zum einen als Beitrag zum besseren Verständnis des Verhältnisses zwischen Verwendungsdaten und Bedeutung sowie zum Status traditioneller wörterbuchähnlicher Bedeutungsangaben innerhalb einer gebrauchsbasierten Semantik gedacht: Zum anderen geht es darum, detaillierte Erkenntnisse über die Übertragbarkeit des Konzepts Synonymie von Einzellexemen auf phraseologische Einheiten zu gewinnen. Unter der Annahme, dass menschliches Lernen bzw. Erschließen von Bedeutung primär empirisch funktioniert, ist die Analyse der Breite der Varianz des tatsächlichen kontextuellen Verhaltens phraseologischer Einheiten bei gleicher oder ähnlicher Bedeutung dazu geeignet, detaillierte Erkenntnisse über die Korrelation zwischen Bedeutungs- und Verwendungsaspekten sowie über den Einfluss phraseologiespezifischer Eigenschaften zu gewinnen. Ausgangspunkt der Untersuchung ist eine Gruppe phraseologischer Einheiten, die in Wörterbüchern als bedeutungsähnlich bzw. synonym klassifiziert werden. Unter diesen phraseologischen Einheiten finden sich Ausdrücke unterschiedlicher Bildlichkeit, Idiomatizität und morphosyntaktischer Struktur, von denen einige aus mehreren Inhaltswörtern bestehen, wie etwa jmd. schüttelt etw. aus dem Ärmel, jmd. hat etw. mit der Muttermilch eingesogen und jmd. weiß Bescheid, während andere lediglich Verbindungen eines Inhaltswortes mit einem oder mehreren Funktionswörtern und dem Verb haben oder sein darstellen. Zur letzteren Gruppe gehören unter anderem Ausdrücke wie jmd. hat das Zeug zu etw., jmd. ist auf Zack und jmd. ist vom Fach. Diese Heterogenität der zu untersuchenden Ausdrücke bezüglich ihres phraseologischen Status macht sie zu einem geeigneten Gegenstand der differenzierten Betrachtung der Rolle phraseologiespezifischer Eigenschaften für die Beschreibung von Bedeutung und Synonymie. Die Untersuchung besteht im ersten Schritt in einer detaillierten Annotation aller Vorkommenskontexte der phraseologischen Einheiten im Korpus des DWDS (Berlin-Brandeburgische Akademie der Wissenschaften, www.dwds.de) auf Basis einer Reihe vordefinierter Analysekriterien. Aus dieser Annotation wird in einem zweiten Schritt eine Beschreibung der typischen Verwendungsmerkmale jeder einzelnen lexikalischen Einheit gewonnen. Gleichzeitig enstehen nach einer festgelegten Methode Bedeutungsbeschreibungen in Form von Paraphrasen, die auf elementare Bedeutungsbestandteile reduziert werden. Diese Bedeutungs- und Verwendungsbeschreibungen der untersuchten phraseologschen Einheiten bilden die Basis für den dritten Teil der Arbeit, in der die beiden Beschreibungsebenen aufeinander bezogen werden. Im Ergebnis zeigt die Arbeit, dass a) Zwischen Verwendung und Bedeutung lexikalischer (phraseologischer) Einheiten identifizierbare systematische Korrelationen bestehen, die einen datenzentrierten Zugang zur Untersuchung und Beschreibung lexikalischer Semantik ermöglichen. b) Innerhalb einer Gruppe von Synonymen so wie auch innerhalb einer erweiterten Menge quasisynonymer Ausdrücke jedem einzelnen Element ein eigener Platz zukommt, der dieses Element von allen anderen Elementen unterscheidet. c) Die Verwendungsdaten einer phraseologischen Einheit positive Evidenz für den individuellen Grad der Relevanz der Merkmale Festigkeit, Idiomatizität und Motiviertheit einer phraseologischen Einheit liefern. N2 - The present Ph.D.-dissertation addresses the subject of synonymy in the special case on Multi-Word Expressions. As a corpus-based, empirical study, it focusses around the question if and how it is possible to approach semantic categories such as meaning, idiomaticity and metaphoricity by analysing the characteristic usage patterns of a lexical unit. The study is driven by basic unresolved issues in the linguistic discussion on meaning and synonymy: On the one hand, it contributes to a better understanding of the relationship between usage data and meaning as well as on the status of traditional dictionary-like definitions in empirical semantics. On the other hand, it provides detailed insights on the applicability of the linguistc concept of synonymy - as it has been discussed for single lexemes – to lexicalized Multi-Word Expressions. On the basis of the assumption that human learning and grasping of meaning is a basically empirical process, the analysis of the variability of contextual behaviour of Multi-Word Expressions with identical or similar meaning constitutes a suitable way of gaining insights on the correlation between meaning and use as well as on the influence of specific properties of Multi-Word Expressions. The study starts out from a group of phraseological units that have been classified as synonymous in German phraseological dictionaries. These expressions differ their degree of idiomaticity, metaphoricity and morphosyntactic structure. Some of them – e.g. 'jmd. schüttelt etw. aus dem Ärmel' or 'jmd. hat etw. mit der Muttermilch eingesogen' (sb. does sth. with great ease) - comprise several content words whereas others are composed of only one content word together with one or more function words and the verb 'haven' (have) or 'sein' (be) – as it is the case for 'jmd. ist auf Zack', jmd. hat das Zeug zu etw.' and 'jmd. ist vom Fach' (sb. is able/competent in sth.). Their heterogeneity makes them a very good starting point for the research on the role played by phraseology-specific properties for meaning and synonymy. As a first step, the author realizes a detailed annotation of all occurrences of the these Multi-Word Expressions in the corpus of the DWDS (Berlin-Brandenburg Academy of Sciences and the Humanities, www.dwds.de), using a series of predefined analysis criteria. From this analysis, she gains a description of their typical usage traits. In parallel, a neatly defined methodology leads to a set of meaning descriptions on the basis of dictionary paraphrases that are reduced to their elementary meaning components. In the last part of the study., these meaning and usage descriptions are then correlated yielding the following conclusions: a) Between meaning descriptions and usage patterns, systematic correlations can be devised that allow a data-driven approach to the semantics of lexical units. b)Within a group of synonyms, each element has its unique place, determined by its characteristic usage patterns. c) The usage data of a multi-word unit provide positive evidence forrr the degree of idiomaticity, morphosyntactic stability and metaphoricity of the expression. KW - Synonymie KW - Bedeutung KW - Empirie KW - Phraseologie KW - synonymy KW - meaning KW - usage data KW - phraseology Y1 - 2007 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:517-opus-51052 ER -