The present study examined whether directional asymmetry in translation priming was found for different-script languages and investigated whether translation ambiguity in the L1 prime or target affected the magnitude of priming. Unbalanced Korean-English bilinguals participated in two masked translation priming tasks with translation-ambiguous and translation-unambiguous word pairs. Results revealed a directional asymmetry in translation priming consistent with previous research. Significant translation priming was obtained in both L1-L2 and L2-L1 directions, with stronger priming effects in the L1-L2 direction. Word pairs with translation-ambiguous primes resulted in weaker translation priming effects compared to translation-unambiguous prime-target pairs in the L1-L2 direction. However, no effects of translation ambiguity on the magnitude of priming were observed when the locus of the ambiguity was in the target for the same word pairs in the L2-L1 direction. These results confirm and extend previous findings on bilingual word recognition suggesting that L1 typically has more impact on L2 lexical processing than vice versa and show that divergence in meaning across a bilingual's two languages causes processing difficulty. Implications for models of bilingual lexical retrieval are also discussed.
Cognitive scientists Spelke and Kintzler (2007) and Carey (2009) identify objects, actions, space and numbers as "core domains of knowledge" that are essential for conceptualizing the world. Gardenfors (2019, 2020) argues that objects, actions and space are characterized by invariances in sensory signals. In this paper, we extend the analysis of invariances to the domain of numbers (understood as positive integers). As a theoretical background, we assume that numbers, as studied in cognitive science, are properties of collections. We claim that the domain of numbers is determined by two types of invariances: (i) the invariance under the location of its objects; (ii) the unconstrained fungibility of objects, which is the determinant invariance of the number concept: If an object in a collection is exchanged for another object, the collection will still contain the same number of objects. We show that invariance under fungibility closely maps onto one-to-one correspondences between collections. Our theoretical analysis is supported by empirical material.
This study investigates vowel-based sound symbolism in Korean ideophones, focusing on how contrasts between yang and yin vowels-associated with perceived size, valence, and intensity- influence lexical processing. Experiment 1, a questionnaire study, confirmed that ideophones containing yin vowels were interpreted as conveying greater negativity, intensity, and size than those with yang vowels. Experiment 2 employed a lexical decision task to examine whether such sound-symbolic associations influence realtime language processing. The results were mixed: accuracy data indicated a mismatch effect, with lower accuracy for yang-vowel ideophones following large-object primes, but no clear facilitation under congruent conditions. Additionally, reaction times for yang-vowel ideophones were slower overall, regardless of prime type. These findings suggest that while explicit sound-symbolic associations are robust, their effects on online lexical processing are context-dependent. Notably, faster responses to yin-vowel ideophones align with prior evidence that conceptually larger objects are processed more efficiently, highlighting the interaction between phonological iconicity and semantic salience in language comprehension.
This survey article reviews the global reception of the Cognitive Revolution and traces the development of cognitive science in East Asia and the Pacific from the 1980s to the present. It focuses on domestic initiatives and international conferences-most notably the ICCS series (from 1997), the founding of the Journal of Cognitive Science (from 2000), and the establishment of the International Association for Cognitive Science (IACS) in 2008- culminating in ICCS Tokyo 2025, after which the conference becomes annual. We survey how cognitive science theories developed. We pay attention to the institutional conditions that enabled sustained intellectual exchange. Focusing on Japan-Korea cooperation in formal linguistics and semantics, the article highlights two successful conference traditions-the Pacific Asia Conference on Language, Information and Computation (PACLIC) and the Japanese/Korean (J/K) Linguistics Conference, as well as LENLS (Logic and Engineering in Natural Language Semantics)-as positive models of archival continuity. The discussion contrasts these cases with more fragmented developments elsewhere in East Asia and argues that long-term success in cognitive science depends not only on theoretical innovation but also on durable archival infrastructures and quality international journals, often maintained through individual scholarly responsibility.
This study investigated the neural correlates of morphological processing in Sino-Korean prefixed derivations in the visual word recognition. Using event-related potentials (ERPs), brain responses to morphological, semantic and orthographic priming effects were examined on the three time windows: N250 (200-270 ms), N400 (340-380 ms), and a late positivity component (LPC) (380-450 ms). Forty-three native Korean speakers participated in a masked priming lexical decision task while their EEG was recorded. The overall results revealed robust and sustained neural effects of N250, N400 and a late positivity for morphological priming, but absence or very weak effects for semantic and orthographic priming, indicating early morphological decomposition and subsequent reanalysis. The following topographic analysis further confirmed that the morphological priming effects were elicited at language related brain regions. These findings were taken as supporting evidence for the early sub-lexical and continuous engagement of morphological analysis of the lexicon in visual word recognition, extending our understanding of morphological decomposition beyond alphabetic languages.
This study examines how Vietnamese learners of English as a foreign language acquire comparative ellipsis, focusing on two subtypes: Gapping and Stripping. It investigates learners' sensitivity to structural constraints as well as the roles of L1 influence and structural priming. Since Vietnamese lacks Gapping-like constructions, such structures were expected to present acquisition challenges. Previous research has not clearly distinguished how second-language learners process Gapping versus Stripping. To address this gap, both comprehension and production tasks were employed. Results showed that learners consistently performed well with Stripping, suggesting its greater accessibility, likely due to parallels with discourse-level omission in Vietnamese. By contrast, Gapping posed initial difficulties. Nevertheless, exposure through priming improved learners' production and, to a lesser extent, their grammaticality judgments, indicating short-term priming effects. These findings support the Full Transfer/Full Access model: learners begin with L1-based representations but can restructure their grammar with sufficient input. However, performance with Gapping remained less stable, reflecting the impact of structural complexity and typological distance. The study underscores the need for targeted pedagogical support for constructions absent in learners' L1. Instructional implications include explicitly contrasting Gapping and Stripping and providing guided, contextualized practice. Future research should investigate the durability of priming effects and the effectiveness of instructional interventions in fostering mastery of L2 ellipsis.
This study examines how readers process evidential dependencies in Korean, a language where marking the source of information is morphologically encoded but often optional. Using a combination of offline and online measures, we investigated the extent to which readers adhere to evidential consistency during incremental processing. Our findings indicate that while readers generate probabilistic expectations based on early cues, these expectations are distinctly modulated by semantic compatibility. The observation that evidential mismatches are frequently accommodated suggests that evidentiality in Korean functions as a soft constraint to prevent communication breakdown. We conclude that evidential processing is not driven by rigid grammatical algorithms but is a flexible mechanism that balances probabilistic expectation with semantic congruence.
This study investigates the morphosyntactic and semantic conditions underlying imperative negation in Korean, focusing on the auxiliary verb mal-, which expresses prohibitive force ('Don't do'). Traditionally, malhas been analyzed as a morphological fusion of the negative morpheme an(i) 'not' and the light verb ha-'do', realized under the presence of deontic modality (Han & Lee 2007). While this Distributed Morphology approach captures the co-occurrence of negation and obligation, it fails to explain why malalternates with an.h-in various deontic contexts and why it is excluded from others that also encode deontic meanings, such as evaluative or permissive expressions. Building on cross-linguistic research in the syntax of imperatives, this paper proposes that the occurrence of malis determined not solely by modality but by person features and speech act structure. Specifically, malappears when negation operates within a directive configuration, where the Person (Pers) head agreeing with the subject interacts with the Addressee node in the Speech Act Phrase (SAP) (i.e., the Mood Phrase in Korean, in conventional terminology) to encode the participant roles of speaker and addressee. This interaction yields the morphological realization of mal-in imperatives, propositives, and related directive environments, but blocks its use in factual or evaluative clauses. The analysis demonstrates that the distribution of mal-reflects a syntactic encoding of speaker-addressee relations rather than deontic modality alone. By linking negation, person agreement, and speech act structure, this study provides new evidence that Korean morphosyntax directly encodes communicative roles, offering broader implications for the grammar of directives and negation across languages.
This study investigates how three GPT-based large language models-ChatGPT-3.5, ChatGPT-4o, and ChatGPT-5.0-resolve globally ambiguous structures in English and Korean, focusing on relative clause (RC) attachment and temporal adverbial adjunct (TAA) attachment ambiguities. Across both constructions, all three models produced near-categorical low-attachment interpretations in English. Although English speakers generally prefer low attachment, human judgments are gradient; thus, the models' uniformly categorical responses reflect an overgeneralized heuristic rather than human-like syntactic parsing. In Korean, where human readers are known to favor high attachment, the models again exhibited strong low-attachment preferences for both RC and TAA structures, with ChatGPT-4o and ChatGPT-5.0 showing especially rigid, near-ceiling low-attachment behavior. These cross-linguistic patterns indicate that GPT models do not acquire language-specific attachment strategies but instead extend an English-derived low-attachment heuristic across typologically distinct languages. The findings underscore key limitations in current autoregressive LLMs' cross-linguistic syntactic generalization and highlight the need for training and modeling approaches that more effectively capture typological variation in structural interpretation.
This paper aims to explore the dynamic aspect of deepening our understanding of concepts and reality that involves experiences and thought at later stages of our lives along personal and social dimensions. We take a linguistic approach, using corpus data focusing on examples of Japanese hont & omacr;no {jibun/ shiawase} 'true {self/ happiness}' and their lexical variants. Insights about the concept and reality of hont & omacr;no {jibun/ shiawase} are identified and organized in terms of classification and contrast. Then qualitative and quantitative analyses of the corpus data that are annotated with the identified insights are provided. The analysis demonstrates that the meaning of jibun and shiawase extends beyond lexical and encyclopedic knowledge and reflects the process of seeking and approaching truth with a critical stance toward the world around us (e.g. a common sense of values). Hont & omacr;no is fundamentally involved in this process and functions as a contrast marker reflecting human critical thinking. The discussion includes a comparison between the empirical insights about shiawase and Aristotle's philosophical thought on eudimonia 'happiness' as presented in Nicomachean Ethics. The findings suggest that the proposed corpus-based analysis serves as an effective means for obtaining and utilizing empirical data for the exploration of philosophical issues as well as lexical semantics and pragmatics.
This paper examines anaphoric bare nouns in Korean, concentrating on their interaction with the topic marker NUN. Unlike in English, Korean bare nouns (without determiners) can function as anaphoric definites by referring back to their antecedents introduced in prior discourse. However, Kim (2023) and Tomioka (2024) recently note that when marked with NUN or WA, the displaced bare nouns in apparent topic position require a preceding definite demonstrative to serve this function, a pattern observed in both Korean and Japanese. By focusing on Korean cases and comparing them with Chinese counterparts, we argue that the definite demonstrative ku 'that' on the displaced NUNmarked bare nouns in apparent topic position is necessary for reasons beyond simply encoding anaphoric definiteness. In a noncontrastive context where only one referent has been introduced in prior discourse, NUN marking on the displaced anaphoric bare noun in apparent topic position in the following sentence results in infelicity, as it inherently induces a contrastive topic reading. However, this effect is neutralized when the noun at hand is preceded by the definite demonstrative.
This study examines the linguistic competence of large language models (LLMs) by evaluating their performance on anaphoric binding in Korean, a language that permits both local and longdistance binding. Using 280 test sentences adapted from prior experiments on native speakers' preferences for local versus long-distance binding, we tested three Generative Pretrained Transformers-GPT-3.5-turbo, GPT-4, and GPT-4o-on their ability to resolve reflexives and pronouns. Across all models, we observed a strong and consistent preference for long-distance binding, even in contexts where native Korean speakers reliably favor local antecedents. This pattern was especially pronounced with morphologically complex anaphors such as caki-casin and ku-casin, for which the models selected long-distance antecedents in 66.7% of cases for GPT-3.5-turbo, 72.5% for GPT-4, and 50% for GPT-4o. These results suggest that the models do not replicate the nuanced patterns of native speaker judgments, particularly in contexts involving morphological or syntactic complexity. However, despite these discrepancies, the models-especially GPT-4o-did exhibit some sensitivity to semantic constraints. In gender-controlled conditions, GPT-4o selected the appropriate local antecedent for ku-casin in 95% of cases in Experiment 1 and 70% in Experiment 2, and the correct long-distance antecedent for kunye in 90% of both experiments. This indicates that while LLMs do not fully internalize binding principles in a native-like way, they are capable of utilizing lexical-semantic cues such as gender to guide antecedent resolution. The findings highlight both the potential and limitations of current LLMs at the syntax-semantics interface and underscore the importance of evaluating these models on linguistically rich and typologically diverse phenomena to gain a deeper understanding of their representational capabilities and theoretical adequacy.
Large Language Models (LLMs) are deep learning-based text generation tools that have shown remarkable improvement in producing coherent discourse and displaying unexpected abilities. This article explores how LLMs can contribute to our understanding of human language and cognition. The author argues that viewing LLMs merely as word predictors that mimic human language behavior oversimplifies their underlying mechanisms and representational capabilities. LLMs function as information acquisition and processing systems and, from a connectionist perspective, can serve as useful models of human cognition-or at least of certain aspects of it. The article begins by providing a brief account of criticisms concerning LLMs' limitations, followed by an examination of the nature of representations within these models and a discussion on their architectural components. It further presents LLMs as generalpurpose systems, highlighting their emerging non-linguistic capabilities. It is suggested that LLMs may have the potential to capture and effectively apply cultural constructs, which are primarily conveyed through language and encapsulate useful 'programs' for addressing a variety of tasks. Additionally, the article briefly examines the possibility of top-down processing and consciousness in these models. Ultimately, the author proposes that current and future generations of LLMs can contribute to a of human and encouraging cognitive scientists to view them as more than mere engineering tools designed to mimic human language.
While much is documented and theorized about how children acquire the meaning of small numbers expressed in simplex numerals (number words that cannot be decomposed further), relatively little is known about how children come to understand complex numerals (that are composed of simplex numerals). We aimed to investigate how children make sense of complex numerals by assessing how they compare large numbers. In a novel task, children compared the numerical values of two large numbers that differed only in one syntactic position (e.g., four hundred twenty chairs vs. six hundred twenty chairs). Together in two studies, with Korean-and English-speaking children, we demonstrate that children including those with limited counting fluency (e.g., not counting past 49 or 99) are able to compare large numbers in the hundreds and the hundred-thousands and that this ability correlates with their counting fluency. At the same time, children find it more difficult to compare two complex-numeral phrases (as the example above) than two simplex-numeral phrases with the same number of words (e.g., four little yellow chairs vs. six little yellow chairs). These results suggest that syntactic complexity of complex numerals influences children's numerical thinking, which is consistent with the idea that children's implicit understanding of numerical syntax contributes to building generative number concepts.
Facial expressions are primary cues for emotion perception, yet their interpretation is influenced by contextual factors. Color is one such factor, carrying affective meanings that may affect how emotional expressions are recognized. The present study examined how red and yellow facial colors affect the recognition of anger and happiness expressions in Korean participants, and whether these effects are modulated by semantic priming. Participants first completed a Stroop task designed either to strengthen color-emotion associations (congruent priming condition) or to present colors in a non-associative manner (incongruent priming condition). They then judged facial expressions overlaid with red or yellow hues. Results showed that red facilitated recognition of angry faces, whereas yellow facilitated recognition of happy faces, with these effects being robust under the semantic priming condition. Reaction time analyses further revealed a facilitative effect of yellow on happy face recognition, while the facilitation of red on anger was less consistent. These findings suggest that color acts not only as a perceptual cue but also as a conceptually grounded signal that interacts with higher-order semantic processes in shaping emotional face perception.
This paper will consider a number of recently offered procedures and models of metaphor identification in the literature. The purpose is to evaluate their possible significance for application to the detection of conceptual metaphors in corpora. It is on the basis of such an evaluation that a new corpus-based model of metaphor identification is proposed. Compared to most previous ones, this model goes beyond the mere act of tracking the surface linguistic forms of metaphors in discourse to identifying the underlying conceptual metaphors that give rise to them. In so doing, a combination of automatic and manual means is employed. The two together have proved to be efficient in accelerating the process of revealing the conceptual bases underpinning patterns of linguistic metaphors in a particular corpus of texts. The proposed conceptual metaphors operate and can be identified in a discourse context. It can therefore be utilized in the development of a simplified and generalizable framework capable of facilitating the researcher's ability to detect and analyze metaphorical language.
Twenty-three adult native speakers of Korean participated in an appropriateness judgment task involving pre-recorded utterances of exact and fuzzy number words. Our results indicate that the use or interpretation of a fuzzy number word is judged to be appropriate whenever the corresponding point value is anywhere inside the range bounded by the natural numbers constituting the fuzzy number word. Results also suggest that the use of a fuzzy number word might be considered more natural as an input in the form of a request or command than as an output in the form of an estimate.
Combinatory Logic (CL) of H. B. Curry is a formalism where any operators can be composed and transformed by abstract operators (called combinators), without using bound variables. The actions of combinators are presented by introduction and elimination rules in the Gentzen's.natural deduction style. By explicit deductions using combinators, it is shown how linguistic expressions are analyzed and linked to their structured semantic and cognitive representations where different types of cognitive primitives (operators and relators) are composed together. CL is used in the theoretical model of GRACE (GRammar of Applicative, Cognitive and Enunciative operations) to explain how activity of language, expressed by natural languages, works with changes of metalinguistic representations between different levels of processing: from a linguistic level towards a semantic-cognitive interpretation, according to a bottom up processing and from cognitive schemes towards their linguistic expressions, according to a top down processing. Some illustrating examples of changes of representations are given; one example is about a semantic representation of the verbal unit 'go go out of ', other examples are about the semantic representations of aspects , tenses.
This study examines the morphological errors made by L1 Korean speakers with L2 English causative verbs-change of state verbs and psych verbs-in the argument structure alternation (i.e., causative/inchoative alternation) to identify possible influence from their L1. Driven by Montrul's (2001) findings, where L1 Turkish-L2 English learners transferred their L1 morphological patterns with psych verbs but did not do so with change of state verbs, we aim to partially replicate the experimental design of Montrul (2001) to examine whether her results of Turkish speakers are repeated with Korean speakers in our study given the resemblance between Turkish and Korean. A picture-based acceptability judgment task was implemented to see whether Korean speakers are biased towards certain morphological verb forms, possibly constrained by their L1 patterns. The results align with those of Montrul (2001) as traces of L1 influence were discovered with psych verbs while only marginal of them were observed with change of state verbs. Indeed, in case of change of state verbs, Korean speakers' strong preferences for morphologically marked verb forms in intransitive contexts (e.g., The door got opened) suggest that L2-developmental factors (viz., overpassivization) may have played a greater role in their morphological errors.
The purpose of the article is to present an overview of how convexity serves as a constraint on the semantics of natural language. I begin by presenting four question that a semantic theory should be able to answer. The semantic theory I propose is based on conceptual spaces, which are geometrical or topological structures provided with a betweenness relation. Previously, I have proposed the criterion that a natural concept is a convex region in a conceptual space. I show how this criterion can be expanded to an analysis of how convexity plays a role in the semantics of different word classes. I will show that the convexity criterion also improves the learnability of word meanings. Voronoi tessellations based on prototypes of categories can function as an efficient mechanism behind such learning processes. For many applications, a Euclidean or a city-block metric provides this relation, but for some word classes, for example color words and prepositions, polar convexity is required. Another topic concerns how the mappings between words and regions in conceptual spaces can be aligned between different individuals - in other words, how we know that we mean the same things when we use a word. If continuity and convexity of the mappings are assumed, Brouwer's fixpoint theorem assures that there exist "meetings of minds" as regards word meanings.
