Text Encoding

The intersection between archive theory and digital technologies has generated more than technical innovations—it has challenged our assumptions about what archives are and do. Projects such as The Walt Whitman Archive, The Shelley-Godwin Archive, the Samuel Beckett Digital Manuscript Project, and the Digitale Faustedition demonstrate a new understanding of the archive's role in literary studies. These initiatives raise a critical question: Should we understand digital archives as reproductions of physical collections, or as entities with their own distinct logic? This paper attempts to answer this question, based on the experience gained from the project Solomos’ Digital Archive. Solomos' holographs are the most authoritative source for his incomplete works. Responding to one of the most important challenges of Modern Greek Philology, Linos Politis proposed ―and in 1964 carried out― the photographic reproduction and diplomatic transcription of the manuscripts. Today, some sixty years later, the digital edition being implemented at the Aristotle University of Thessaloniki is part of that longstanding editing tradition and at the same time follows contemporary developments in the field of digital scholarly editions. We discuss the challenges posed by the very concept of the literary archive in the digital condition, which often approximates that of an edition, as it consists not only of digital surrogates and their metadata but also of encoded transcriptions based on editorial decisions (according to the TEI standard); the work of encoding and annotating the transcribed text is what allows for multiple forms of textual representation (diplomatic, critical, genetic, reading) and for textual analysis via methods of distant reading.

This article examines the transregional life and typographic legacy of a 24-point Ottoman naskh printing type (MI-24), developed in 1867 by the Ottoman-Armenian punch-cutter Oḥannes Mühendisyan in collaboration with the court calligrapher Ḳāżīʿasker Muṣṭafá ʿİzzet Efendi. Celebrated for its exceptional calligraphic fidelity and mechanical refinement, MI-24 emerged as the pinnacle of Arabic-script typography in the Ottoman empire. Drawing on previously unexamined archival materials, this study reconstructs MI-24's production, dissemination, and adaptation across a wide geographyfrom Istanbul to Beirut, Cairo, Tehran, Kabul, and even Chicago, London and Oxford. It explores how regional actors including Jesuit missionaries, Armenian type-founders, Arab Nahḍ a intellectuals, and European Orientalists engaged with and modified this typographic model to suit diverse cultural, religious, and technological needs. In tracing these typographic networks, the article situates MI-24 at the intersection of Islamic visual culture, print modernity, and global knowledge circulation. Far from being a product confined to Istanbul's imperial presses, MI-24 became a mobile and malleable artefact-reshaped by local aesthetics and political agendas. Ultimately, this article reframes Arabic-script typography as a dynamic site of visual negotiation and transimperial collaboration, contributing to broader discourses on print culture, design history, and the materialities of Islamicate text production.

La Training School internazionale ‘Interdisciplinary Methods for Textual Analysis. From Ancient to Medieval Texts’ nasce con l’obiettivo di offrire a studenti e giovani ricercatori, professionisti e interessati, una preparazione avanzata sui nuovi metodi di indagine applicati al testo, considerato nella sua accezione più ampia: dalle iscrizioni antiche fino ai manoscritti tardo-medievali e umanistici. Il programma è concepito per mettere al centro una riflessione sul metodo come strumento di mediazione tra innovazione tecnologica e materialità del testo. Le lezioni frontali presenteranno una panoramica critica delle più recenti tecnologie digitali applicate allo studio dei testi, illustrandone potenzialità, limiti e implicazioni interpretative in ambiti quali l’epigrafia, la filologia, la paleografia e la linguistica storica.

Our cities today represent a complex intersection of material and immaterial culture, which can be described and promoted through possible heritage itineraries. Among the various topics of these itineraries, a topic of great interest for us is the writings traced throughout history on the wall surfaces of the city's buildings to convey messages and opinions for the most diverse purposes. Their value is not only in the quality of the single artefact but also because they can serve as unique keys to understanding the society that produced them in depth, i.e., to comprehend the culture, customs, and ways of communication of a specific historical period. From this point of view, the inscriptions produced by the ruling classes, the secular and religious power are less exciting than those traced by single or common people. Therefore, these traces of the past are a fundamental historical resource since the people's voice resides in them; they are an intangible heritage transmitted through the materiality of the supports and architectures on which they were traced or placed. Unfortunately, many of these inscriptions are hidden or barely visible within cities. Besides, their decoding is related to the type of handwriting, the engraving technique, and the content conveyed. Finally, in today's society, there is a prevalent dominance of the image with clearly conveyed content, relegating these kinds of cultural products to an omitting background. The study presented in this paper, at an early stage of development, is developed within Rome, a city characterised by a complex and deep cultural stratification. The research aims to map some graffiti and 'minor' epigraphs (i.e., non-monumental or particularly solemn) and build narrative itineraries throughout the city, starting from a study of these particular traces within the architecture. By enhancing these historical sources through these routes, the architecture will become a narrative system capable of telling stories, giving voice to real-life stories and the immaterial culture condensed in those places, and rediscovering a stratified and hidden intangible heritage in the city.

A companion note to the Keshif Database outlining the TEI P5 encoding decisions developed for digital editions of Ottoman-Turkish micro-editions — covering metadata structure, entity markup (persons, roles, places, dates), genre and format encoding, and editorial apparatus. The note also addresses the status of these decisions as a working framework open to revision.

Il progetto Atti Chiari, volto a raccogliere il primo grande corpus italiano di atti di parte, presenta stringenti requisiti di ordine legale e numerose peculiarità sul piano della lingua e dei contenuti, che hanno reso necessario progettare e implementare una serie di processi e di strumenti ad hoc.
In particolare, al fine di eliminare ogni dato personale dai documenti, senza tuttavia distruggerne il tessuto linguistico e comprometterne la leggibilità, si è creata una procedura di pseudonimizzazione basata su un preliminare stadio di annotazione, che aggiunge informazione per poterla poi rimuovere in modo diverso a seconda dei fini (analisi linguistica, analisi giuridica, ecc.). Nel contempo, questa leggera annotazione contribuisce a fornire dati funzionali non solo alla pseudonimizzazione del testo, ma anche alla sua conversione dal loro originale formato di presentazione a uno semantico basato su TEI. I documenti così preparati vengono poi centralizzati in un corpus, destinato a indicizzazione a scopo di ricerca linguistica. Dati i molteplici criteri di ricerca di cui si richiede la combinazione, quale che sia la loro provenienza e modello, è stato utilizzato un nuovo tipo di motore di ricerca, disegnato anzitutto in ambito filologico, per ottenere la necessaria apertura e granularità dei metadati.

This paper presents a comprehensive examination of SAEL (Structured Ancient Egyptian Language), a novel encoding system designed for the representation and computational treatment of Egyptian hieroglyphic texts. SAEL emerges as a response to long-standing challenges in digital Egyptology, particularly the conflation of semantic content with visual presentation that characterizes traditional encoding systems. This document provides a detailed technical specification of SAEL's notation, operators, and semantic principles, while offering a systematic comparison with established systems such as the Manuel de Codage (MdC) and the Revised Encoding Scheme (RES). Furthermore, this study addresses the fundamental challenges inherent in digital representation of ancient languages, with particular emphasis on the complex ligature systems of Egyptian hieroglyphs and the broader implications for Unicode-based typography. The analysis reveals SAEL's distinctive approach to separating linguistic content from artistic representation, its partial compatibility with existing standards, and its potential contribution to the advancement of computational Egyptology and digital humanities scholarship.

Scribes using the Arabic alphabet for writing Arabic, Persian or early varieties of Turkic used non-obligatory 'distinctive' diacritics beside the standard diacritics of the writing systems of their languages. After quoting Onur (2024) on the presence of such diacritics in Arabic, Persian, Khāqānī Turkic and Early Middle Turkic, the author documents their appearance in 13 th-14 th century epitaphs in Arabic and Atil Turkic, a language closely related to Chuvash, spoken in the Volga-Kama area of Russia, and in 11 th-12 th century legal documents in Uyghur script discovered in Yarkand, East Turkestan. The distinctive diacritics consist either of dots or of small Arabic letters placed either below or above the line of writing.

Il capitolo illustra le fasi della redazione di un'edizione critica digitale evidenziando il ruolo cruciale degli strumenti digitali. Si parte da un nuovo approccio alla collazione dei manoscritti, che integra un database per archiviare le lezioni e analizzarle preliminarmente. Grazie alle formule e ai filtri del foglio di calcolo, è possibile elaborare grandi quantità di dati per ottenere rapporti di filiazione tra i manoscritti per mezzo di calcoli statistici e orientare l'analisi filologica, finalizzata alla definizione dello stemma codicum. Questo metodo offre uno strumento accessibile, integrabile e potente per gestire e analizzare dati complessi. Successivamente, si approfondisce l'organizzazione dei dati per l'edizione critica digitale, illustrando il progetto PRIN RETI (Rendering TExts and Images). Si descrivono le modalità di marcatura in XML/TEI e l'uso di EVT, un software dedicato per la visualizzazione delle edizioni. Si illustrano le scelte fatte per classificare le varianti testuali, evidenziando le potenzialità e che gli strumenti digitali pongono alla filologia, con esempi pratici, al fine di mostrare come questi approcci innovativi possano contribuire alla ricerca filologica, fornendo soluzioni avanzate e flessibili.

By combining traditional historical enquiry with TEI XML encoding and decoding in a corpus analysis phase, the project aims at addressing research questions mainly related to the French and British positions on the topics of armament design and production and of armament control within the Western European Union (WEU) from 1954 to 1982. The paper focuses on the annotation of speakers (different countries and institutional representatives) and their discourse in a selection of institutional documents (minutes, notes, studies, memoranda) (encoding phase) and the identification of linguistic patterns on armament issues in their discourse, as well as the interpretation of results (decoding phase). From a larger perspective, the study considers the TEI encoding as adding to the original text a “material” layer that further supports both machine and human interpretation (decoding). In this sense, this study may move closer to the concept of “material hermeneutics,” by understanding code, ...

S ocial engagement refers to participation in collective activities and today represents a key strategy through which museums and other cultural heritage organisations can both empower their connections with visitors, and improve their own eicacy as cultural and scientiic institutions. As explained by Waterton and Watson, understanding the role that heritage plays in a particular society allows us an insight on how the communities engage with it. That makes it possible for museums and collections to connect with their social context, improving their ability to serve as dynamic social spaces for community engagement and action. Social engagement shows therefore a strong geographical base because 'where' , 'who' and 'how' afect any process oriented to encourage local participation in heritage valorisation. Such connections appear even more evident when applied to a geographical collection, like the one present in the University of Padua. This paper originates from the authors' experience on structuring the scientiic project of the future Museum of Geography of the University of Padua in Italy. Even though the museum has not been formally established, two years' work on historical research, collection reconnaissance and space organisation has led to an idea of a 'collections and connections' museum, where the collections become strategic to highlight social and spatial connections to both historical and present research practices and visitors' personal experiences.

Molti probabilmente non ci crederanno, ma la sicurezza dei mainframe è incredibilmente semplice da comprendere e il suo modello potrebbe essere facilmente adottato da altre piattaforme e applicazioni. L'unica cosa di cui abbiamo bisogno è una spiegazione rapida ed efficace per dissipare l'idea errata che la sicurezza dei mainframe sia estremamente complessa. Questa convinzione errata è probabilmente legata alla percezione che la maggior parte delle persone ha della complessità del mainframe stesso. Chiarire questi preconcetti è esattamente ciò che vorrei fare con questo breve documento.
PS. Se lo preferite, potete trovare qui anche la versione originale in inglese.

The Digital Fauvel is an interactive facsimile-edition of the satirical “Roman de Fauvel” (ca. 1314) as preserved in the manuscript Paris, Bibliotheque Nationale fr. 146. In this source, the story of Fauvel (an allegorical horse that stands for the worst aspects of humanity) is augmented by hundreds of images and musical items in a variety of genres. To “read” Fauvel in a way that approximates the rich experience its original audiences could have derived from the object requires a working knowledge of Middle French, Latin, medieval music notation, and 14th-century iconographic conventions. As a result, modern scholars have dissociated the object into more recognizable volumes—editions that group like with like and necessarily leave context behind. We have used new media to re-synthesize the source, making it more navigable, intelligible, and meaningful to the untrained reader and expert researcher alike. In its current form, The Digital Fauvel software runs on a large multitouch tab...

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

The contents of this site is subject to the French law on intellectual property and is the exclusive property of the publisher. The works on this site can be accessed and reproduced on paper or digital media, provided that they are strictly used for personal, scientific or educational purposes excluding any commercial exploitation. Reproduction must necessarily mention the editor, the journal name, the author and the document reference. Any other reproduction is strictly forbidden without permission of the publisher, except in cases provided by legislation

We report on experience gained from our ongoing multi-year project to produce an Electronic Variorum Edition of Cervantes' Don Quixote de la Mancha. Initially designed around a custom database representation, the project's evolution has lead to the adoption of a TEI-based format for information interchange among the project's major components. We discuss the mechanics of this approach and its benefits.

The article presents a prototype of a digital edition of diplomatic material, exemplified through the scarce yet typologically and in terms of textual tradition diverse documentary heritage of King Stefan Uroš III (Dečanski). The methods of encoding metadata and texts of charters and letters in XML format are discussed, and the application of the widely adopted standard Text Encoding Initiative (TEI), supplemented by appropriate concepts from the Vocabulaire international de la diplomatique (VID), is demonstrated. The functions of the XPath and XQuery expression languages as well as the application of the Keyword in Context (KWIC) method are examined and shown through selected examples. The preliminary conclusions regarding the contribution of digital methods in diplomatic research are offered, along with proposals for further work on Serbian documentary sources.

O n peut aujourd'hui reconstituer la naissance de la critique des variantes -qui fut plus tard nommée « critique des paperasses » par son fondateur Gianfranco Contini (1912-1990) 1 -et le débat qui suivit son affirmation en Italie à l'aune de documents inédits du fonds Contini et grâce au travail de catalogage et de valorisation de ses papiers et de sa correspondance soutenu par la Fondation Franceschini et par l'École normale supérieure de Pise, sous l'impulsion de Lino Leonardi et de Claudio Ciociola 2 . Il s'agit d'un moment critique fondamental de l'histoire culturelle italienne. En effet, cette confrontation dialectique avec la vieille génération (celle qui avait résisté au fascisme et dont Benedetto Croce, avec son Manifeste des intellectuels antifascistes de 1925, est le plus célèbre représentant) a dominé la culture italienne jusqu'aux années 1960 -une confrontation parfois âpre, menée par la nouvelle génération des intellectuels italiens, celle de la Résistance, qui tout en admirant Croce ne pouvait pas se reconnaître dans l'idéalisme qu'il représentait. Gianfranco Contini fut un protagoniste central de ce moment critique. Spécialiste de philologie romane et critique militant, il a uni l'étude philologique du passé à un intérêt pour les poétiques de la modernité, en particulier pour le symbolisme, et a introduit en Italie depuis les années trente une nouvelle idée du texte, entendu comme processus et non plus comme état. C'est justement l'orientation prise par les travaux de l'ITEM -l'institution qui en Europe a le plus oeuvré pour garantir la continuité de cette idée -, que pourrait synthétiser la citation placée en exergue de cet article, tirée d'un auteur français qui fut très cher à Contini. Notre propos est de retracer l'histoire de la naissance de cette idée, en rappelant les liens entre la critique des variantes et la poésie symboliste française, puis les origines communes de la critique des variantes et de la critique génétique, dans le contexte culturel italien de l'après-guerre encore dominé par la personnalité de Benedetto Croce. Ce fut une confrontation très dure entre générations, une véritable révolution culturelle, mais qui, à la différence

Natural Language Processing (NLP) tasks are usually performed word by word on textual inputs. We can use arbitrary symbols to represent the linguistic meaning of a word and use these symbols as inputs. As "alternatives" to a text representation, we introduce Soundex, MetaPhone, NYSIIS, logogram to NLP, and develop fixed-output-length coding and its extension using Huffman coding. Each of those codings combines different character/digital sequences and constructs a new vocabulary based on codewords. We find that the integration of those codewords with text provides more reliable inputs to Neural-Networkbased NLP systems through redundancy than text-alone inputs. Experiments demonstrate that our approach outperforms the state-ofthe-art models on the application of machine translation, language modeling, and part-ofspeech tagging. The source code is available at https://github.com/abdulrafae/coding nmt.

This paper introduces a novel way of enhancing NLP prediction accuracy by incorporating model interpretation insights. Conventional efforts often focus on balancing the trade-offs between accuracy and interpretability, for instance, sacrificing model performance to increase the explainability. Here, we take a unique approach and show that model interpretation can ultimately help improve NLP quality. Specifically, we employ our learned interpretability results using attention mechanisms, LIME, and SHAP to train our model. We demonstrate a significant increase in accuracy of up to +3.4 BLEU points on NMT and up to +4.8 points on GLUE tasks, verifying our hypothesis that it is possible to achieve better model learning by incorporating model interpretation knowledge.

Multilingual neural machine translation (MNMT) jointly trains a shared model for translation with multiple language pairs. However, traditional subword-based MNMT approaches suffer from out-ofvocabulary (OOV) issues and representation bottleneck, which often degrades translation performance on certain language pairs. While byte tokenization is used to tackle the OOV problems in neural machine translation (NMT), until now its capability has not been validated in MNMT. Additionally, existing work has not studied how byte encoding can benefit endangered language translation to our knowledge. We propose a byte-based multilingual neural machine translation system (BMNMT) to alleviate the representation bottleneck and improve translation performance in endangered languages. Furthermore, we design a random byte mapping method with an ensemble prediction to enhance our model robustness. Experimental results show that our BMNMT consistently and significantly outperforms subword/word-based baselines on twelve language pairs up to +18.5 BLEU points, an 840% relative improvement.

Currently, no ‘canonical’ classical text with a multi-testimonial tradition has a digital scholarly edition based on a complete digital transcription of all primary sources, and on the automated collation of those transcriptions. Most classicists simply do not feel that they need such editions. I argue that this is ultimately due to the ‘canonization’ of the corpus of classical texts. Classicists are more focussed on the ‘Text’ than on the documents (manuscripts) and their texts: they tend not to consider the textual variance in the manuscripts as culturally meaningful in itself, but merely instrumental in view of the constitutio textus. I suspect that we will not have ‘comprehensively digital’ editions of ‘canonical’ classical texts with a multi-testimonial tradition until classical philology broadens its research agenda. 1. We have a problem There is a problem which seems to be mostly going unnoticed. There is no (not one) ‘comprehensively digital’ scholarly edition of a ‘classica...

Recent breakthroughs in natural language processing (NLP) have come with escalating model sizes and computational costs, posing significant challenges for deployment in realtime and resource-constrained environments. We introduce EMBYTE , a novel byte-level tokenization model that achieves substantial embedding compression while preserving NLP accuracy and enhancing privacy. At the core of EMBYTE is a new Decompose-and-Compress (DECOMP) learning strategy that decomposes subwords into fine-grained byte embeddings and then compresses them via neural projection. DECOMP enables EMBYTE to be shrunk down to any vocabulary size (e.g., 128 or 256), drastically reducing embedding parameter count by up to 94% compared to subword-based models without increasing sequence length or degrading performance. Moreover, EMBYTE is resilient to privacy threats such as gradient inversion attacks, due to its byte-level many-to-one mapping structure. Empirical results on GLUE, machine translation, sentiment analysis, and language modeling tasks show that EMBYTE matches or surpasses the performance of significantly larger models, while offering improved efficiency. This makes EMBYTE a lightweight and generalizable NLP solution, well-suited for deployment in privacy-sensitive or low-resource environments.

In the AI era, the integration of artificial intelligence into lexicography and language service technologies has become inevitable. Lexicographers and language technology developers are actively exploring ways to embed AI into their research and products to remain relevant and competitive. This study investigates how frequent users of Chinese-English reference tools perceive current translation software and dictionaries, as well as their expectations for future improvements. User reviews reveal that while current reference tools provide decoding functions, there is still a significant unmet demand for encoding support. Questionnaire results further show that users strongly expect improvements in electronic dictionaries and AI tools, especially regarding cultural adaptability, specialised terminology, and the quality of example sentences. These findings offer new insights into the evolving needs of dictionary users in the AI era. Our study aims to support the development of next-generation language tools that are efficient, user-friendly, and culturally adaptive, particularly for Chinese users learning English as a second language. To this end, we also evaluate the DeepSeek-R1 model for generating advanced Chinese-English translations and lexicographic content. Compared to Google Translate and DeepL, DeepSeek-R1 shows superior performance in handling specialised terms, syntactic accuracy, and textual coherence. Furthermore, in the specific task of providing in-depth explanations of specialised terms, although GPT-4o performs competently, DeepSeek-R1 demonstrates a stronger lexicographic capability than both Gemini-2.5 and GPT-4o.

The contents of this site is subject to the French law on intellectual property and is the exclusive property of the publisher. The works on this site can be accessed and reproduced on paper or digital media, provided that they are strictly used for personal, scientific or educational purposes excluding any commercial exploitation. Reproduction must necessarily mention the editor, the journal name, the author and the document reference. Any other reproduction is strictly forbidden without permission of the publisher, except in cases provided by legislation