Špela Arhar Holdt
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Špela Arhar Holdt has a total of 147 co-authors. Here are the top co-authors by number of shared publications:
- Iztok Kosem
University of Ljubljana (Faculty of Arts & Faculty of Computer and Information Science), Ljubljana, Slovenia
4 shared publications - Damjan Popič
University of Ljubljana, Faculty of Arts, Slovenia
3 shared publications Iztok Kosem
University of Ljubljana
3 shared publications- Mojca Stritar3 shared publications
- Nataša Logar
University of Ljubljana
3 shared publications - Aleksandra Bizjak Končar
Research Centre of the Slovenian Academy of Sciences and Arts (ZRC SAZU)
2 shared publications
Latest Publications
- Open Access
- 22 December 2025
- Published by Institut za novejso zgodovino in Contributions to Contemporary History
- Vol. 65 (3)
Potencial ChatGPT pri razvoju Slovarja sopomenk sodobne slovenščine
In this study, we examine how well ChatGPT-4 performs in two lexicographic tasks: (a) cleaning the list of automatically retrieved synonym candidates and assigning synonymic material to lexical senses, and (b) generating dictionary entries, including sense division, definitions, and examples, based on different input data. As a gold standard, we consider the lexicographic decisions recorded in the Digital Dictionary Database for Slovene. In the first experiment, we analyse the results for 246 dictionary entries and find that ChatGPT processed the data identically to lexicographers in 41.9 % of cases, while in 58.1 % of cases, it made different decisions. When assessing the relevance of synonym candidates, ChatGPT was more permissive than the gold standard. Differences in synonym placement (assignment to a different sense in 14.6 % of entries, missing placement in 19.9 %) can be partly attributed to input data characteristics, such as task complexity and the brevity of semantic indicators. In the second experiment, we test ChatGPT’s ability to autonomously generate dictionary entries for 116 headwords. The analysis of generated sense divisions and definitions reveals that the system performs moderately well: in 57 % of cases, it identified all senses, almost 80 % of generated entries received an average score of 3.5 or higher, and 19 % received the highest score from both evaluators. The main challenges include excessive splitting of senses, failure to recognise figurative meanings, and reduced predictability of results. We conclude that ChatGPT has potential for speeding up manual lexicographic work if its results are properly monitored and refined.
- 22 December 2025
- Published by Institut za novejso zgodovino in Contributions to Contemporary History
- Vol. 65 (3)
Open AccessPotencial ChatGPT pri razvoju Slovarja sopomenk sodobne slovenščine
In this study, we examine how well ChatGPT-4 performs in two lexicographic tasks: (a) cleaning the list of automatically retrieved synonym candidates and assigning synonymic material to lexical senses, and (b) generating dictionary entries, including sense division, definitions, and examples, based on different input data. As a gold standard, we consider the lexicographic decisions recorded in the Digital Dictionary Database for Slovene. In the first experiment, we analyse the results for 246 dictionary entries and find that ChatGPT processed the data identically to lexicographers in 41.9 % of cases, while in 58.1 % of cases, it made different decisions. When assessing the relevance of synonym candidates, ChatGPT was more permissive than the gold standard. Differences in synonym placement (assignment to a different sense in 14.6 % of entries, missing placement in 19.9 %) can be partly attributed to input data characteristics, such as task complexity and the brevity of semantic indicators. In the second experiment, we test ChatGPT’s ability to autonomously generate dictionary entries for 116 headwords. The analysis of generated sense divisions and definitions reveals that the system performs moderately well: in 57 % of cases, it identified all senses, almost 80 % of generated entries received an average score of 3.5 or higher, and 19 % received the highest score from both evaluators. The main challenges include excessive splitting of senses, failure to recognise figurative meanings, and reduced predictability of results. We conclude that ChatGPT has potential for speeding up manual lexicographic work if its results are properly monitored and refined.
- Open Access
- 14 December 2025
- Published by Elsevier in Journal of Responsible Technology
- Vol. 25
Corplus: A new concordancer for exploring authentic texts with language corrections
- 14 December 2025
- Published by Elsevier in Journal of Responsible Technology
- Vol. 25
Open AccessCorplus: A new concordancer for exploring authentic texts with language corrections
- Open Access
- 31 October 2025
- Published by Zveza društev pedagoških delavcev Slovenije in Sodobna Pedagogika
- Vol. 76 (3) , 162-176
Grammatical error correction of Slovenian school essays using large language models
Grammatical error correction (GEC) is the task of automatically detecting and correcting grammatical errors in text. Large language models have enabled the development of accurate automated methods for detecting and correcting certain types of errors. In the educational domain, the aim of GEC is to aid teachers in correcting student errors. Excessive paraphrasing is a property of Generative Pre-trained Transformer-based models and is undesirable in the language education context. To avoid this, we develop multiple Slovenian models for correcting errors in spelling, word case (capitalization), word form, and word order. We describe the training data construction, training process, and model evaluation approach using the Šolar-Eval 1.0 corpus of school essays authored by primary and secondary school students. Our quantitative evaluation shows that the developed models have reasonably high accuracy levels, and our qualitative evaluation highlights the strengths and weaknesses of the models and the evaluation process. The analysis reveals multiple challenges and promising future directions for improving both model development and the evaluation process.
- 31 October 2025
- Published by Zveza društev pedagoških delavcev Slovenije in Sodobna Pedagogika
- Vol. 76 (3) , 162-176
Open AccessGrammatical error correction of Slovenian school essays using large language models
Grammatical error correction (GEC) is the task of automatically detecting and correcting grammatical errors in text. Large language models have enabled the development of accurate automated methods for detecting and correcting certain types of errors. In the educational domain, the aim of GEC is to aid teachers in correcting student errors. Excessive paraphrasing is a property of Generative Pre-trained Transformer-based models and is undesirable in the language education context. To avoid this, we develop multiple Slovenian models for correcting errors in spelling, word case (capitalization), word form, and word order. We describe the training data construction, training process, and model evaluation approach using the Šolar-Eval 1.0 corpus of school essays authored by primary and secondary school students. Our quantitative evaluation shows that the developed models have reasonably high accuracy levels, and our qualitative evaluation highlights the strengths and weaknesses of the models and the evaluation process. The analysis reveals multiple challenges and promising future directions for improving both model development and the evaluation process.
- Open Access
- 31 October 2025
- Published by Zveza društev pedagoških delavcev Slovenije in Sodobna Pedagogika
- Vol. 76 (3) , 86-106
Jezikovno popravljanje v digitalnem okolju: kvalitativna študija z učiteljicami in učitelji slovenščine
Prispevek predstavlja kvalitativno raziskavo o izkušnjah, potrebah in preferencah uporabe digitalnih orodij za jezikovno pregledovanje in popravljanje pisnih besedil pri pouku slovenščine. V raziskavi je sodelovalo 18 učiteljev in učiteljic slovenščine z različnih vrst šol (osnovnih šol, strokovnih in poklicnih šol ter gimnazij). Učitelji so se seznanili z orodjem za jezikovno pregledovanje v programu Microsoft Word in novorazvitim orodjem za strojno popravljanje slovenskih besedil, nato pa so v strukturiranih intervjujih poročali o svojih izkušnjah z digitalno podprtim popravljanjem, preferencah glede funkcionalnosti in vizualizacije strojnega popravljanja ter stališčih glede vloge generativne umetne inteligence v procesu razvoja pisne zmožnosti učencev. Rezultati so razkrili ključne pomanjkljivosti strojnih rešitev, kot sta omejena natančnost in prilagodljivost pedagoškim potrebam (vizualizacija jezikovnih popravkov, vrsta in količina popravkov za prikaz, podajanje povratne informacije, možnost formativnega spremljanja itd.). Na drugi strani so rezultati ponudili uvid v tista mesta popravljanja, kjer bi bila digitalna podpora najbolj učinkovita in zaželena. Ugotovitve dajejo vpogled v omejitve obstoječih tehnologij in smernice za razvoj naprednejših digitalnih rešitev, prilagojenih potrebam učiteljev v izobraževalnem procesu.
- 31 October 2025
- Published by Zveza društev pedagoških delavcev Slovenije in Sodobna Pedagogika
- Vol. 76 (3) , 86-106
Open AccessJezikovno popravljanje v digitalnem okolju: kvalitativna študija z učiteljicami in učitelji slovenščine
Prispevek predstavlja kvalitativno raziskavo o izkušnjah, potrebah in preferencah uporabe digitalnih orodij za jezikovno pregledovanje in popravljanje pisnih besedil pri pouku slovenščine. V raziskavi je sodelovalo 18 učiteljev in učiteljic slovenščine z različnih vrst šol (osnovnih šol, strokovnih in poklicnih šol ter gimnazij). Učitelji so se seznanili z orodjem za jezikovno pregledovanje v programu Microsoft Word in novorazvitim orodjem za strojno popravljanje slovenskih besedil, nato pa so v strukturiranih intervjujih poročali o svojih izkušnjah z digitalno podprtim popravljanjem, preferencah glede funkcionalnosti in vizualizacije strojnega popravljanja ter stališčih glede vloge generativne umetne inteligence v procesu razvoja pisne zmožnosti učencev. Rezultati so razkrili ključne pomanjkljivosti strojnih rešitev, kot sta omejena natančnost in prilagodljivost pedagoškim potrebam (vizualizacija jezikovnih popravkov, vrsta in količina popravkov za prikaz, podajanje povratne informacije, možnost formativnega spremljanja itd.). Na drugi strani so rezultati ponudili uvid v tista mesta popravljanja, kjer bi bila digitalna podpora najbolj učinkovita in zaželena. Ugotovitve dajejo vpogled v omejitve obstoječih tehnologij in smernice za razvoj naprednejših digitalnih rešitev, prilagojenih potrebam učiteljev v izobraževalnem procesu.
- Open Access
- 25 August 2025
- Published by Linkoping University Electronic Press in Swedish Aerospace Conference, FT2025
- Vol. 216, 71-85
CLARIN.SI, the Slovenian node of CLARIN
The paper presents the organisation and services offered by the Slovenian research infrastructure for language resources and technologies CLARIN.SI. We introduce the governance, organisational structure, and technical components of the infrastructure, followed by a description of its web applications with a focus on the repository and concordancers. Next, we provide an overview of support activities offered by CLARIN.SI, which includes services of the CLASSLA knowledge centre for processing South Slavic languages, financial support for projects, and facilitation of conferences and workshops. We also present the involvement of CLARIN.SI in national and European projects, with its sister national infrastructure nodes of DARIAH and CESSDA, and in the work of CLARIN ERIC.
- 25 August 2025
- Published by Linkoping University Electronic Press in Swedish Aerospace Conference, FT2025
- Vol. 216, 71-85
Open AccessCLARIN.SI, the Slovenian node of CLARIN
The paper presents the organisation and services offered by the Slovenian research infrastructure for language resources and technologies CLARIN.SI. We introduce the governance, organisational structure, and technical components of the infrastructure, followed by a description of its web applications with a focus on the repository and concordancers. Next, we provide an overview of support activities offered by CLARIN.SI, which includes services of the CLASSLA knowledge centre for processing South Slavic languages, financial support for projects, and facilitation of conferences and workshops. We also present the involvement of CLARIN.SI in national and European projects, with its sister national infrastructure nodes of DARIAH and CESSDA, and in the work of CLARIN ERIC.
- Open Access
- 1 January 2025
- Published by Stellenbosch University in Lexikos
- Vol. 35, 294-317
Lexicography and Neology: Building Bridges through a Multilingual Approach with ENEOLI
Neology, the study of new word formations and new word meanings, remains an underrepresented yet critical dimension of linguistic research. This paper explores the efforts of the European Network on Lexical Innovation (ENEOLI) COST Action, launched in October 2023, in addressing key gaps in the study of lexical innovation in connection with lexicography. With 343 participants from 51 countries as of March 2025, ENEOLI seeks to build a comprehensive, multilingual approach to neology research. This initiative focuses on developing digital resources, comparative studies, and professional training to elevate the understanding of neology within the linguistic community, including the development of a multilingual glossary of the meta-terminology of neology. By integrating lexicographic and computational methodologies, ENEOLI sets a benchmark for multilingual research. This initiative not only contributes to a deeper understanding of neology but also provides resources that support its application across diverse academic and professional contexts, including translation, education, and digital humanities. Through its collaborative framework, ENEOLI highlights the significance of lexical innovation in understanding language evolution and its cultural and social dimensions. Keywords: ENEOLI, neology, lexicography, lexical innovation, language change, multilingual glossary, terminology Die leksikografie en neologie: Die smee van bande deur 'n veeltalige benadering met ENEOLI. Neologie, die studie van nuwe woordvormings en nuwe woordbetekenisse, bly 'n onderverteenwoordigde dog kritieke aspek van taalkundige navorsing. Hierdie artikel ondersoek die pogings van die Europese Netwerk vir Leksikale Innovasie (ENEOLI) COST-aksieplan, wat in Oktober 2023 van stapel gestuur is, om sleutelkwessies in die bestudering van leksikale innovasie rakende die leksikografie aan te pak. ENEOLI streef met 343 deelnemers uit 51 lande (teen Maart 2025) daarna om 'n omvattende, veeltalige benadering tot neologienavorsing te volg. Hierdie inisiatief fokus op die ontwikkeling van digitale hulpbronne, vergelykende studies en professionele opleiding om die begrip van neologie binne die taalkundige gemeenskap te bevorder, insluitend die ontwikkeling van 'n veeltalige glossarium van die metaterminologie van neologie. Deur leksikografiese en rekenaargebaseerde metodologieë te integreer, stel ENEOLI 'n maatstaf vir veeltalige navorsing. Dié inisiatief dra nie net by tot 'n dieper begrip van neologie nie, maar voorsien ook hulpbronne wat die toepassing daarvan oor diverse akademiese en professionele kontekste, insluitend vertaling, opvoedkunde, en digitale geesteswetenskappe, ondersteun. Met sy samewerkingstruktuur beklemtoon ENEOLI die belangrikheid van leksikale innovasie in die verstaan van taalontwikkeling en die kulturele en sosiale dimensies daarvan. Sleutelwoorde: ENEOLI, neologie, leksikografie, leksikale innovasie, taalverandering, veeltalige glossarium, terminologie
- 1 January 2025
- Published by Stellenbosch University in Lexikos
- Vol. 35, 294-317
Open AccessLexicography and Neology: Building Bridges through a Multilingual Approach with ENEOLI
Neology, the study of new word formations and new word meanings, remains an underrepresented yet critical dimension of linguistic research. This paper explores the efforts of the European Network on Lexical Innovation (ENEOLI) COST Action, launched in October 2023, in addressing key gaps in the study of lexical innovation in connection with lexicography. With 343 participants from 51 countries as of March 2025, ENEOLI seeks to build a comprehensive, multilingual approach to neology research. This initiative focuses on developing digital resources, comparative studies, and professional training to elevate the understanding of neology within the linguistic community, including the development of a multilingual glossary of the meta-terminology of neology. By integrating lexicographic and computational methodologies, ENEOLI sets a benchmark for multilingual research. This initiative not only contributes to a deeper understanding of neology but also provides resources that support its application across diverse academic and professional contexts, including translation, education, and digital humanities. Through its collaborative framework, ENEOLI highlights the significance of lexical innovation in understanding language evolution and its cultural and social dimensions. Keywords: ENEOLI, neology, lexicography, lexical innovation, language change, multilingual glossary, terminology Die leksikografie en neologie: Die smee van bande deur 'n veeltalige benadering met ENEOLI. Neologie, die studie van nuwe woordvormings en nuwe woordbetekenisse, bly 'n onderverteenwoordigde dog kritieke aspek van taalkundige navorsing. Hierdie artikel ondersoek die pogings van die Europese Netwerk vir Leksikale Innovasie (ENEOLI) COST-aksieplan, wat in Oktober 2023 van stapel gestuur is, om sleutelkwessies in die bestudering van leksikale innovasie rakende die leksikografie aan te pak. ENEOLI streef met 343 deelnemers uit 51 lande (teen Maart 2025) daarna om 'n omvattende, veeltalige benadering tot neologienavorsing te volg. Hierdie inisiatief fokus op die ontwikkeling van digitale hulpbronne, vergelykende studies en professionele opleiding om die begrip van neologie binne die taalkundige gemeenskap te bevorder, insluitend die ontwikkeling van 'n veeltalige glossarium van die metaterminologie van neologie. Deur leksikografiese en rekenaargebaseerde metodologieë te integreer, stel ENEOLI 'n maatstaf vir veeltalige navorsing. Dié inisiatief dra nie net by tot 'n dieper begrip van neologie nie, maar voorsien ook hulpbronne wat die toepassing daarvan oor diverse akademiese en professionele kontekste, insluitend vertaling, opvoedkunde, en digitale geesteswetenskappe, ondersteun. Met sy samewerkingstruktuur beklemtoon ENEOLI die belangrikheid van leksikale innovasie in die verstaan van taalontwikkeling en die kulturele en sosiale dimensies daarvan. Sleutelwoorde: ENEOLI, neologie, leksikografie, leksikale innovasie, taalverandering, veeltalige glossarium, terminologie
- Open Access
- 19 November 2024
Primerjava sistemov za označevanje jezikovnih popravkov v štirih slovenskih besedilnih korpusih
- 19 November 2024
Open AccessPrimerjava sistemov za označevanje jezikovnih popravkov v štirih slovenskih besedilnih korpusih
- 1 September 2024
- Published by Springer Nature
- p. 85-96
Neural Spell-Checker: Beyond Words with Synthetic Data Generation
- 1 September 2024
- Published by Springer Nature
- p. 85-96
Neural Spell-Checker: Beyond Words with Synthetic Data Generation
- Open Access
- 18 July 2024
- Published by Springer Nature in Language Resources and Evaluation
- Vol. 59 (2) , 1151-1177
Šolar, the developmental corpus of Slovene
The paper presents the Šolar developmental corpus of Slovene, comprising the written language production of students in Slovene elementary and secondary schools, along with teacher feedback. The corpus consists of 5485 texts (1,635,407 words) and includes linguistically categorized teacher corrections, making the corpus unique in reflecting authentic classroom correction practices. The paper addresses the corpus compilation, content and format, annotation, availability, and its applicative value. While learner corpora are abundant, developmental corpora are less common. The paper bridges the gap by introducing the evolution from Šolar 1.0 to 3.0, emphasizing improvements in text collection, error and correction annotation, and categorization methodology. It also underlines the challenges and unresolved issues of compiling developmental corpora, most notably the lack of openly available tools and standards for different steps of the compilation process. Overall, the Šolar corpus offers valuable insights into language learning and teaching, contributing to teacher training, empirical studies in applied linguistics, and natural language processing tasks.
- 18 July 2024
- Published by Springer Nature in Language Resources and Evaluation
- Vol. 59 (2) , 1151-1177
Open AccessŠolar, the developmental corpus of Slovene
The paper presents the Šolar developmental corpus of Slovene, comprising the written language production of students in Slovene elementary and secondary schools, along with teacher feedback. The corpus consists of 5485 texts (1,635,407 words) and includes linguistically categorized teacher corrections, making the corpus unique in reflecting authentic classroom correction practices. The paper addresses the corpus compilation, content and format, annotation, availability, and its applicative value. While learner corpora are abundant, developmental corpora are less common. The paper bridges the gap by introducing the evolution from Šolar 1.0 to 3.0, emphasizing improvements in text collection, error and correction annotation, and categorization methodology. It also underlines the challenges and unresolved issues of compiling developmental corpora, most notably the lack of openly available tools and standards for different steps of the compilation process. Overall, the Šolar corpus offers valuable insights into language learning and teaching, contributing to teacher training, empirical studies in applied linguistics, and natural language processing tasks.
- Open Access
- 24 August 2023
- Published by Springer Nature in Research Square
Šolar, the developmental corpus of Slovene
The paper presents the Šolar developmental corpus of Slovene, comprising the written language production of students in Slovene elementary and secondary schools, along with teacher feedback. The corpus consists of 5,485 texts (1,635,407 words) and features linguistically categorized teacher corrections. Notably, the annotation of language problems is based on actual teacher corrections, making the corpus unique in reflecting authentic classroom correction practices. The paper addresses the corpus compilation, content and format, annotation, availability, and its applicative value. While learner corpora are abundant, developmental corpora are less common. The paper bridges the gap by introducing the evolution from Šolar 1.0 to 3.0, emphasizing improvements in text collection, error and correction annotation, and categorization methodology. It also underlines the challenges and unresolved issues of compiling developmental corpora, most notably the lack of openly available tools and standards for different steps of the compilation process. Overall, the Šolar corpus offers valuable insights into language learning and teaching, contributing to teacher training, empirical studies in applied linguistics, and natural language processing tasks.
- 24 August 2023
- Published by Springer Nature in Research Square
Open AccessŠolar, the developmental corpus of Slovene
The paper presents the Šolar developmental corpus of Slovene, comprising the written language production of students in Slovene elementary and secondary schools, along with teacher feedback. The corpus consists of 5,485 texts (1,635,407 words) and features linguistically categorized teacher corrections. Notably, the annotation of language problems is based on actual teacher corrections, making the corpus unique in reflecting authentic classroom correction practices. The paper addresses the corpus compilation, content and format, annotation, availability, and its applicative value. While learner corpora are abundant, developmental corpora are less common. The paper bridges the gap by introducing the evolution from Šolar 1.0 to 3.0, emphasizing improvements in text collection, error and correction annotation, and categorization methodology. It also underlines the challenges and unresolved issues of compiling developmental corpora, most notably the lack of openly available tools and standards for different steps of the compilation process. Overall, the Šolar corpus offers valuable insights into language learning and teaching, contributing to teacher training, empirical studies in applied linguistics, and natural language processing tasks.