2026

Big Data and Cognitive Computing 10(5)

Salazar Alva, R.G., Núñez, M., López Del Alamo,C. & Martín Arista, J.

Abstract: Preserving ancient languages is essential for understanding the cultural and linguistic heritage of humanity. Old English, however, remains critically under-resourced, which limits its accessibility to modern natural language processing (NLP) techniques. We present a scalable framework that uses advanced large language models (LLMs) to generate high-quality Old English texts to address this gap. In this study, we specifically employ state-of-the-art models, including Llama-3.1-8B and Mistral-7B, as our foundation models, which are then adapted to the unique characteristics of Old English. Our approach combines parameter-efficient fine-tuning (Low-Rank Adaptation (LoRA)), data augmentation via back-translation, and a dual-agent pipeline that separates content generation (in English) and translation (into Old English). Evaluation with automated metrics (BLEU, METEOR, and CHRF) shows improvements over baseline models, with BLEU scores increasing from 26 to over 65 for English-to-Old English translation. Expert human assessment confirms high grammatical accuracy and stylistic fidelity in the generated texts, with average scores of 9.0/10 for inflection and word order, 9.1/10 for lexical authenticity, and 7.8 for semantic coherence. These results demonstrate that the framework can reliably expand limited historical corpora while maintaining linguistic integrity, with immediate practical applications in digital humanities research, computational philology, and the development of educational resources for Old English study. Beyond expanding the Old English corpus, our method offers a practical blueprint for revitalizing other endangered languages, thus linking AI innovation with the goals of cultural preservation.

2026

AI 7(9)

Martín Arista, J. & Núñez, M.

Abstract: This paper raises a methodological question: How can we assess machine-made Old English when there is no parallel reference text and the standard metrics do not fit the task? We propose a pipeline with five measuring layers plus two compliance components, including lexical attestation with form linking, frequency-profile diagnostics, character-level comparison, word-embedding geometry under a verified mapping and dependency parsing, with bootstrap confidence intervals around the main contrasts. We apply the pipeline to a machine-made version of Gregory’s Dialogues: 4217 sentences, one for each sentence of the Old English original, generated under hard constraints. The unattested residue is two word types and 0.003% of tokens. The frequency profile diverges from the original by 0.008, less than the original diverges from the background corpus. At character level, in embedding space and in parsed syntax, the generated text stands at the same distance from the Dictionary of Old English Corpus as the original itself does. We propose an overall metric G, the geometric mean of seven bounded components, which scores the text at 0.991 with a confidence interval of [0.991, 0.992]. Two blind detection experiments with expert judges place the index externally: roughly two-thirds of generated sentences pass as authentic to specialists, so the divergence the pipeline measures is real at corpus scale but not available to sentence-by-sentence reading. The main contribution is a reusable evaluation method for historical language generation, together with a single interpretable score that subsumes the partial metrics without hiding them.

2026

International Journal of Humanities and Arts Computing 20(1)

Martín Arista, J. & Metola Rodríguez, D.

Abstract: We evaluate a Stanza-based pipeline for Old English that combines character-level processing with language-specific word embeddings derived from the Dictionary of Old English Corpus. On a 25,000-word dataset annotated with Universal Dependencies, character-level models yield consistent gains over token-based variants for lemmatization and dependency parsing (test UAS 88.92 per cent; LAS 79.65 per cent). Relative to a multilingual baseline for Old English, our best parser improves LAS by approximately 20 per cent; compared with a recent monolingual spaCy baseline on similar data sizes, gains are approximately 5 per cent on our split. We complement scores with qualitative analyses to illustrate strengths and limitations, and we provide implementation details to support replication. The main conclusion of this work is that, under low-resource conditions, language-specific embeddings and character-level modelling are more effective for Old English processing than cross-linguistic transfer learning.

2026

Procesamiento del Lenguaje Natural 77

Martín Arista, J.

Abstract: Lemmatisation is an essential step in the computational analysis of historical corpora, such as Old English, which presents problems of spelling variation, widespread inflectional morphology, and limited annotated resources. This paper conducts a detailed comparison of neural architectures for Old English lemmatisation: encoder-decoder (BiLSTM with attention) and Transformer under three experimental conditions (corpus-only, dictionary augmentation without POS tags, with POS tags) and advanced techniques (character-level augmentation, learning cycles with restarts). The Transformer achieves 66.65% development accuracy and our improved model 67.58% test accuracy. Dictionary augmentation without POS degrades performance (?53% unknown POS); with POS it recovers but does not exceed corpus-only training. Pre-trained embeddings (Word2Vec) and BPE tokenization reduce accuracy (up to ?12.67% and ?12.92%), showing that lemmatisation is a morphological task where semantic signals introduce noise.

2025

Big Data and Cognitive Computing 9 (8)

Martín Arista, J. Ojanguren López A. E. & Domínguez Barragán, S.

Abstract: This study presents the first systematic empirical comparison of neural architectures for Universal Dependencies (UD) parsing in Old English, thus addressing central questions in computational historical linguistics and low-resource language processing. We evaluate three approaches—a baseline spaCy pipeline, a pipeline with a pretrained tok2vec component, and a MobileBERT transformer-based model—across datasets ranging from 1000 to 20,000 words. Our results demonstrate that the pretrained tok2vec model consistently outperforms alternatives, because it achieves 83.24% UAS and 74.23% LAS with the largest dataset, whereas the transformer-based approach substantially underperforms despite higher computational costs. Performance analysis reveals that basic tagging tasks reach 85–90% accuracy, while dependency parsing achieves approximately 75% accuracy. We identify critical scaling thresholds, with substantial improvements occurring between 1000 and 5000 words and diminishing returns beyond 10,000 words, which provides insights into scaling laws for historical languages. Technical analysis reveals that the poor performance of the transformer stems from parameter-to-data ratio mismatches (1250:1) and the unique orthographic and morphological characteristics of Old English. These findings defy assumptions about transformer superiority in low-resource scenarios and establish evidence-based guidelines for researchers working with historical languages. The broader significance of this study extends to enabling an automated analysis of three million words of extant Old English texts and providing a framework for optimal architecture selection in data-constrained environments. Our results suggest that medium-complexity architectures with monolingual pretraining offer superior cost–benefit trade-offs compared to complex transformer models for historical language processing.

2025

Encyclopedia 5(3)

Martín Arista, J.

Abstract: This entry presents a comprehensive overview of the computational study of Old English that surveys the evolution from early digital corpora to recent artificial intelligence applications. Six interconnected domains are examined: textual resources (including the Helsinki Corpus, the Dictionary of Old English Corpus, and the York-Toronto-Helsinki Parsed Corpus), lexicographical resources (analysing approaches from Bosworth–Toller to the Dictionary of Old English), corpus lemmatisation (covering both prose and poetic texts), treebanks (particularly Universal Dependencies frameworks), and artificial intelligence applications. The entry shows that computational methodologies have transformed Old English studies because they facilitate large-scale analyses of morphology, syntax, and semantics previously impossible through traditional philological methods. Recent innovations are highlighted, including the development of lexical databases like Nerthusv5, dependency parsing methods, and the application of transformer models and NLP libraries to historical language processing. In spite of these remarkable advances, problems persist, including limited corpus size, orthographic inconsistency, and methodological difficulties in applying modern computational techniques to historical languages. The conclusion is reached that the future of computational Old English studies lies in the integration of AI capabilities with traditional philological expertise, an approach that enhances traditional scholarship and opens new avenues for understanding Anglo-Saxon language and culture.

2025

Procesamiento del Lenguaje Natural 74

Martín Arista, J. Ojanguren López A. E. & Domínguez Barragán, S.

Abstract: The aim of this article is to assess three training corpora of Old English and three configurations and training procedures as to the performance of the task of automatic annotation of Universal Dependencies (UD, Nivre et al., 2016). The method is aimed to deciding to what extent the size of the corpus improves results and which configuration turns out the best metrics. The training methods include a pipeline with default configuration, pre-training of tok2vec step and a model of language based on transformers. For all training methods, three training corpora with four different sizes are tested: 1,000, 5,000, 10,000, and 20,000 words. The training and the evaluation corpora are based on ParCorOEv2 (Martín Arista et al., 2021). The results can be summarised as follows. The larger training corpora result in improved performance in all the stages of the pipeline, especially in POS tagging and dependency parsing. Pre-training the tok2vec stage yields better results than the default pipeline. It can be concluded that the performance could improve with more training data or by fine-tuning the models. However, even with the limited training data selected for this study, satisfactory results have been obtained for the task of automatically annotating Old English with UD.