This paper investigates the application of translation alignment algorithms in the creation of a Multilingual Digital Edition (MDE) of Alessandro Manzoni's Italian noIvperlomessi sposi (“The Betrothed”), with translations in eight languages (English, Spanish, French, German, Dutch, Polish, Russian and Chinese) from the 19th and 20th centuries. We identify key requirements for the MDE to improve both the reader experience and support for translation studies. Our research highlights the limitations of current state-of-the-art algorithms when applied to the translation of literary texts and outlines an automated pipeline for MDE creation. This pipeline transforms raw texts into web-based, side-by-side representations of original and translated texts with diferent rendering options. In addition, we propose new metrics for evaluating the alignment of literary translations and suggest visualization techniques for future analysis.
From the very beginning of digital edition creation, there has been a tendency, supported by the power of web technologies, to represent not only the original text but also its translation(s), following the tradition of bilingual printed editions. In this paper, we propose to define multilingual digital editions (MDE) as editions in which translations are not supplementary but essential, intended to enrich both computational analysis and reader experience.
Beyond annotated file accessibility, the MDE should meet additional criteria to be efective. Primarily, the platform must display the original text alongside translations. It is anticipated that there will be a visual correlation between aligned pairs, which will facilitate straightforward comparison and analysis. The accuracy of alignment is, by default, ensuring that the corresponding parts of the texts are properly aligned. Furthermore, the platform should support the visual highlighting of omitted or inserted parts in the translati6o]n,sw[hich will enable users to discern diferences and interpret the nuances of each translation.
These requirements are generally feasible for short, structured texts like poetry or historical documents (examples of MDE publishing strategies are described in Appendix A). The challenge is to develop a flexible, automated system that accurately aligns complex literary texts across multiple languages for computational analysis and user-friendly exploration. The technology should be able to handle the complexities of literary texts, including the splitting, merging, and reordering of sentences, and align text fragments of manageable length, ensuring that they are easy for users to read and understand at a glance to obtain insight into the linguistic and cultural nuances of each version. The automated alignment process should save researchers both time and resources.
For the MDE of Alessandro Manzoni’s nove”Ilpromessi sposi” (The Betrothed), we propose an automatic translation alignment pipeline that adapts state-of-the-art alignment techniques to the objectives of the multilingual digital edition of literary works for educational and research purposes.
A comparative analysis of translations of the same literary work over time can provide valuable insights into the evolution of interpretation and understandingI.p“romessi sposi” is particularly compelling in this context. Not only does it reflect the author’s exploration of the Italian language during a period of significant linguistic evolution, but it has also been translated into many European languages over the past two centuries. This makes it an ideal case study for investigating the influence of temporal factors, linguistic shifts, and the reception of the original novel in diferent cultural contexts.
Two main original editions (1827 and 1840) were translated into European languages and published in parallel in the XIX century. For the development of an automated translation alignment pipeline, we selected and prepared the texts of a wide range of translations of the classic edition of the novel, also known as Quarantana (1840), including English translations from 1845, 1876, 1983, and 2022; Russian translations from 1854 and 1999; a Dutch (1849); a German (1884), a French (1874), a Spanish (1858), a Polish (1882) and a Chinese (1998) (see Appendix 1 for a list of translations).
The core of the MDE creation process is translation alignment, which involves mapping corresponding units (typically words or sentences) between a source and target text. State-of-the-art alignment algorithms have evolved significantly in recent years and now perform optimally in many applications, including machine translation, bilingual dictionary creation, and parallel corpus development.
Modern methods have moved from statistical approache5s,[
One-to-One One-to-Many Many-to-One
Many-to-Many English2022
English1972
Alignment Types Across Different Translations
While the alignment of text and translation at the line level is sufÏcient for poetic, historical, and even verse dramatic texts (see 1[]), where we cannot expect significant variation in the splitting, merging, or reordering of lines, this alignment approach is inadequate for prose due to the extent of restructuring that inevitably occurs in literary prose translation. In such cases, the standard approach is sentence-level alignment. However, it can be challenging, particularly in the case of literary translations, due to the irregularity of the syntactic structure of the original text in another language. Literary translators are not limited to translating a single sentence into another single sentence (this can be described as a one-to-one type of alignment) but are free to manage sentence boundaries and reconfigure sentence structures to better convey the meaning and style of the original text. In this case, in working to achieve the highest similarity score for the aligned pairs, the alignment algorithms are forced to combine several sentences into one, using one-to-many, many-to-one, and many-to-many alignment types.
The ideal alignment type for the MDE is a one-to-one alignment type to maintain the granularity and consistency of the alignment. In our analysis of the sentence-level alignmenIt of promessi sposi (see Figure1 for the diferent translations), while one-to-one alignments are the most common, a significant proportion are more complex types. While this does not inherently complicate the alignment process, as advanced tools such as Bertalign and Vecalign can handle this complexity, the results may be less optimal in terms of meaningfulness. The length of the aligned pairs becomes longer, including several sentences from both the source text and the translation (examples can be seen in the Appendix B). The edge case for this expanded alignment result would be the pairing of the paragraph or even the chapter of the original text with the same of the translation.
That’s why traditional metrics may not be sufÏcient for evaluating the alignment results. The performance of alignment algorithms is typically evaluated using established metrics such as precision, recall, F1 score, and Alignment Error Rate (AE2R3)].[ The first limitation of this approach is that it is based on a ”gold dataset,” which does not provide insight into the performance of the algorithm with respect to other types of tex9t][. A second consequence is that the scores may be high, but the results are not suitable for MDE because the aligned pairs are too large to be analyzed or identified at a glance by a human observer. We, therefore, suggest that, in addition to the increasing importance otfhe distribution of alignment types (one-to-one, one-to-many, many-to-one, many-to-many) as a metric of the acceptability of the resultsh,e number and length of alignment pairs derived from the original sentences should also be considered. A number of aligned pairs close to the number of original sentences would indicate an efective alignment process. Conversely, a significant reduction in the number of aligned pairs would indicate limitations of the sentence-level alignment approach, as it implies that the alignment algorithm is forced to combine more sentences to obtain the appropriate similarity score.
The length of alignment pairs will indicate if they are suitable for human readers. In the
context of creating a digital edition for educational or research purposes with multiple languages,
it is not advisable to present long-aligned texts, given the limited attention span and working
memory of the readers (for further insight, see studies of working memory and comprehension
with multiple text reading1[
To illustrate, if a sentence in the source language (Italian) is 130 tokens long and its corresponding sentence in the target language (English) is 140 tokens long, readers may encounter difÏculties in comparing and understanding such lengthy segments. Even the use of color diferentiation to highlight aligned pairs does not overcome this challenge (see Figu2r).e
In summary, in the context of MDE of literary works, sentence-level alignment still faces a significant challenge. 1) Sentence boundaries are not stable in diferent languages, which leads to a variety of alignment types and doesn’t allow a consistent alignment across the MDE. 2) Strict sentence-level alignment does not fully reflect the variability of the translated texts, such as inserted or omitted parts, and 3) strains readers’ attention spans and working memory and fails to achieve the alignment granularity that is comfortable for the overall reading experience. The alignment process needs to be modified to address these challenges to automated processing and readability.
The alternative methods can provide more accurate and meaningful segmentation of literary texts. In an attempt to move from sentence-level alignment to phrase- or segment-level alignment, here are two promising approaches: • Punctuation splitting: Applying punctuation marks (such as commas, periods, and semicolons) to create initial segments. This method provides natural breaks in the text, preserving the contextual meaning. However, by using this approach and aligning the resulting segments with Bertalign, we achieved a more granular alignment but increased the number of reordering problems that didn’t occur with sentence-level alignment. • Zero/Few-Shot Prompting with LLM Models: The sentences of the original text are segmented using zero-shot prompting OpenAI CPT-4o model19[]. The approved segments are then used as patterns for few-shot prompting to segment the sentences of the translations. This approach provides a robust foundation for universal alignment. The similarity score can be visualized to evaluate segment-level alignment and compare its results with traditional sentence-level alignment. In addition, the visual representation of the similarity score of the aligned segments or sentences allows us to find the semantic outliers.
After extracting high-dimensional embeddings for each aligned line from the original and translated text using the multilingual model (LaBSE), we applied t-Distributed Stochastic Neighbour Embedding (t-SNE) to reduce them to two dimensions. By visually examining the cosine similarity, we can detect anomalies and curious translated fragments (see Fig3u)r,even if the alignment algorithm establishes the correlation between sentences.
There are several quantitative metrics that can be used to assess the quality of alignment: • Number of resulting aligned pairs with respect to the number of input segments; • Consistency: Ensure that segments are consistently aligned across languages, preserving the meaning and context of the original text. • Number of clusters: The number of clusters indicates how the sentences are grouped.
An ideal alignment would result in clusters where each cluster contains two points corresponding to the embedding of the aligned pair, one from the original text and one from the translation, placed close to each other; • Average similarity: Calculating the average similarity within clusters gives an insight into the coherence of the alignments, with higher values indicating more semantically consistent groupings. • The length of the aligned pair of lines indicates whether the detected lines are within the reader’s attention span and suitable for the user experience.
Similarity visualisation of sentence embeddings in the Chapter 23 language german italian
Comparing sentence-level and segment-level alignment, we can assume that even sentencelevel alignment provides valuable insights into the diferences between the translations and the original text (see Appendix D for examples in Spanish translationIoPfromessi sposi); segmentlevel alignment allows us to go deeper and capture more nuanced variations between the original and the translated text. For example, we can identify the omission of the end of chapter 8 in the German translation (see Tabl1e) and two omissions in the Russian translation of chapter 1 (see Tables3-2), which can be interpreted through the lens of cultural diferences and/or censorship and could not be captured by sentence-level alignment.
Presto, io spero, potrete ritornar sicuri aIch hofe, ihr werdet bald ohne Gefahr in casa vostra; euer Haus zurückkehren können; a ogni modo, Dio vi provvederà, per ilin jedem Falle wird Gott Alles zu euerm vostro meglio; Besten lenken. e io certo mi studierò di non mancare alla grazia che mi fa, scegliendomi per suo ministro, nel servizio di voi suoi poveri cari tribolati.
Voi,» continuò volgendosi alle due donne,Und ihr», fuhr er, zu den beiden Frauen «potrete fermarvi a ***. gewandt, fort, «ihr könnt euch in *** so lange aufhalten. Ai tempi in cui accaddero i fatti che pren- Во время тех событий, которые мы diamo a raccontare, quel borgo, già consid-намерены описать, Лекко было уже erabile, era anche un castello, значительным местечком и маленькой крепостцей; e aveva perciò l’onore d’alloggiare un coв-следствие чего в нем жили комендант mandante, e il vantaggio di possedere una и постоянный гарнизон испанских stabile guarnigione di soldati spagnoli, солдат, che insegnavan la modestia alle fanciulle e alle donne del paese, accarezzavan di tempo in tempo le spalle a qualche marito, a qualche padre; e, sul finir dell’estate, non mancavan mai di spandersi nelle vi- которые занимались собиранием gne, per diradar l’uve, e alleggerire a’ conв-инограда. tadini le fatiche della vendemmia.
Con tutto ciò, anzi in gran parte a cagion di ciò, quelle gride, ripubblicate e rinforzate di governo in governo, non servivano ad altro che ad attestare ampollosamente l’impotenza de’ loro autori, o, se producevan qualche efetto immedi- если декреты ato... действительность...
Несмотря на все, и может быть, потому именно, имели минутную
By providing a more granular and accurate alignment, the segment-level approach also allows the length of aligned pairs to be reduced (increasing their number) and makes the MDE more suitable for reader reception compared to sentence-level alignment (see Fi4g)u. re
Average Lengths for Original and Translation Aligned Pairs in Spanish_1858 spanish_1858 - Original Segment Length spanish_1858 - Translated Segment Length spanish_1858 - Original Sentence Length spanish_1858 - Translated Sentence Length Itron ap1C ap2C ap3C ap4C ap5C ap6C ap7C ap8C ap9C ap10 p11 p12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38
ap ap ap ap ap ap ap ap ap ap ap ap ap ap ap ap ap ap ap ap ap ap ap ap p p
C aC aC C C C C C C C C C C C C C C C C C C C C C C C C aC aC
The automated pipeline for the MDE is proposed as a means of enabling creators to prepare annotated TEI files that are accessible, adaptable, correct, easily parsed by computational tools, and rendered for readers (see Figu5r)e.
We start with the raw texts of the translations in TXT format, obtained after OCR and error checking. For Manzoni’s text, we used TEI files with identifiers assigned to each token. This preparation allows us to take into account the irregular segmentation to be expected due to inconsistencies across the languages.
specifics of the texts to be published, the MDE developers can select the segmentation methodology in accordance with the projected audience and the project’s objectives, enabling alignment at the sentence, phrase, or word level, or a combination, giving readers the flexibility to choose their preferred option.
decision made in the first step, the text can be split into sentences, segments, or even tokens.
Step 4. Applying alignment algorithms: By default, we applied Bertalign with the LaBSE model, trained on 109 languages7[] to the segments obtained at the previous steps. Other multilingual sentence-transformer models, such as BGE M3-Embedding, can also be us4e]d. [
Step 5. Choosing the encoding approach. The encoding approach determines the flexibility of the alignment description for future rendering and for establishing a link between the original and translated texts. For structured texts, where each segment in the original closely corresponds to an equivalent segment in the translations, it may be appropriate to mark each segment with the same identifier. Given the complexity of multilingual alignment, we have taken a diferent approach. The TEI encoding of the original text includes identifiers for each token, providing granular reference points. The TEI-encoded translation text is divided into segments, each referencing the start and end identifiers from the original text, allowing for lfexible and accurate alignment.
Step 6. Encoding. By iterating over the alignment results, we assign the referencing start and end identifiers from the original text to each aligned segment from the translation and generate a new TEI file for the translation, ensuring that all segments are accurately linked to the corresponding elements.
Step 7. Rendering Aligned Texts on the Web. Render the original and translated texts from the TEI files as two columns on a web page with separate XSLT templates for the original and translated text. This interactive interface allows users to click on the original text and see the corresponding translation fragment highlighted, enhancing the user experience by providing an intuitive way to explore and compare the translations side by side.
Step 8. Visualization and evaluation. While the highly unstable text version1s3[] or linelevel aligned translation1s6[] can be efectively visualized with the Sankey diagram or bipartite graph, the alignment results for the modern translations can be visualized with the approach described above, based on the embedding vectors with t-SNE and clustering with DBSCAN. As for the presentation in the user interface, ideally, all multilingual translations should be comparable and aligned with each other, allowing the user to see and interpret the diferences. 7. Future Development and Challenges • Current alignment algorithms face challenges in accurately aligning segments with reordered content. Future work will focus on improving the alignment performance in such cases, ensuring more precise matches even when the original and translated texts difer significantly in structure. • Previous studies on user behavior in digital editions have analyzed log files to understand interaction patterns 3[]. To gain deeper insights, we are using advanced tools such as ReactFlow to study more comprehensively how users interact with diferent elements of MDEs. For example, when readers view two lines in diferent languages side by side, the optimal reading span may difer from traditional reading practices. By analysing user interactions, we aim to determine the most efective segment length for the MDEs.
The proposed pipeline aims to improve the development of Multilingual Digital Editions (MDE) by ensuring that MDE is both methodologically robust and user-centered. By prioritizing user experience and usability, the pipeline adapts existing computational methods and algorithms to the specific needs of educational and research applications.
We have also proposed new metrics for MDEs that focus on the consistency, meaningfulness and granularity of the alignment. These metrics assess the suitability of an alignment for educational and research purposes. By ensuring that the alignment is accessible to human readers while supporting translation studies, the pipeline balances conciseness and reader engagement.
M. Simard, G. F. Foster, and P. Isabelle. “Using cognates to align sentences in bilingual corpora”. In:Proceedings of the Fourth Conference on Theoretical and Methodological Issues in Machine Translation of Natural Languages. Montréal, Canada, 1992, pp. 67–81. url:h ttps://aclanthology.org/1992.tmi-1.7 [22]
This research was supported by the Dipartimento di Filologia Classica e Italianistica, University of Bologna, as part of the project “Manzoni online2: manoscritti e documenti inediti, tradizione e traduzioni” (CUP J34I19003370001, project code 2017CFZFAY_003). For more information on the Leggo Manzoni project, visithttps://projects.dharc.unibo.it/leggomanzo.ni
Decameron web La entretenida Miguel de Cervantes by Furnace and Fugue Project Name Cantar de mio Cid Vincent van Gogh. The Letters Ancrene Wisse Preface The Codex Sinaiticus Project Dafydd ap Gwilym.net Princeton Dante Project Ein Sermon von Ablass und Gnade Secrets of Craft and Na- ture Codex Suprasliensis lines Electronic Beowulf lines Kassák Lajos: The lines Horse Dies the Birds
Fly Away Alignment Type
Comparison Notes
Editions include diferent modes of the text on a single page with a side-by-side view, including the translation.
Typically, there is no alignment, but the text length is usually just one page, making comparison straightforward. These viewers enable comparison of two or more versions of the text.
The original text in Old Church Slavonic is directly followed by its corresponding parallel Greek text.
When the special view type and option are selected, and the user hovers the mouse over a line, the translation appears in a special area.
The page displays side-byside views of the original text and its translations into two other languages, highlighting the corresponding translated line when the mouse hovers over the original line. 4. Interlinear text/translation alignment 5. Dynamic alignment display + + + + + + + + + + + Project Name The Community of the sentences Realm in Scotland lines
+ + The side-by-side viewer of the Latin text with its English translation aligns the sentences, allowing users to click on the sentence number in the original text, which automatically scrolls the other side of the page to the corresponding sentence. The TEI Publisher tool allows the user to highlight corresponding parts and automatically scroll when hovering over the lines.
1Sì; ma com’è dozzinale! com’è sguaiato! 1Oui; mais comme il est commun! comme com’è scorretto! il est inégal! comme il est incorrect! id2Idiotismi lombardi a iosa, frasi della linio-tismes lombards à foison, phrases de gua adoperate a sproposito, grammatica ar-la langue employées à rebours, construcbitraria, periodi sgangherati. tions arbitraires, périodes boiteuses; et 3E poi, qualche eleganza spagnola semi- puis quelques petites élégances espagnoles nata qua e là; e poi, ch’è peggio, ne’ lu- semées ça et là; et puis, ce qui est bien pis, oghi più terribili o più pietosi della storiad,ans les endroits les plus terribles ou les a ogni occasione d’eccitar maraviglia, o diplus touchants de son histoire, à chaque far pensare, a tutti que’ passi insomma che occasion d’exciter la surprise ou de faire richiedono bensì un po’ di rettorica, ma ret-penser, à tous les passages enfin qui detorica discreta, fine, di buon gusto, costui mandent, il est vrai, quelques fleurs de rhénon manca mai di metterci di quella sua torique, mais d’une rhétorique sobre, fine, così fatta del proemio. de bon goût, ce digne homme ne manque jamais d’y mettre quelque chose dans le genre de son début.
1 Né alcuno dirà questa sij imperfettione del 1 Und es wird gewiß niemand sagen, dies Racconto, e deformità di questo mio rozzo sei ein Geschichtsfälscher und eine EntstelParto, a meno questo tale Critico non sij per- lung dieser meiner einfältigen Erzählung, sona afatto diggiuna della Filosofia : che es sei denn der Tadel ein Mann, der aller quanto agl’huomini in essa versati, ben Weltweisheit vollständig bar wäre. vederanno nulla mancare alla sostanza di2 Denn man wird bald sehen, daß in detta Narratione. Beziehung der darin vorkommenden Per2 Imperciocché, essendo cosa evidente, sonen am Wesentlichsten der besagten e da verun negata non essere i nomi se Erzählung nichts fehle;zumal es eine non puri purissimi accidenti... augenfällige, von niemand gelenkte
1 Però alla mia debolezza non è lecito 1 Doch mijn’ geringeren krachten is het solleuarsi a tal’argomenti, e sublimità peri- niet gegeven zich tot zoo hooge vlugt, tot colose, con aggirarsi tra Labirinti de’ Politici zulk eene gevaarvolle verhevenheid te vermaneggj, et il rimbombo de’ bellici Ori- hefen, en zich te wagen in den doolhof der calchi: solo che hauendo hauuto notitia di staatkundige spitsvondigheden of te midden fatti memorabili, se ben capitorno a gente van het geschal der schorre krijgsklaroenen. meccaniche, e di piccol afare, mi accingo 2 Naardemaal ’er dus eenige merkdi lasciarne memoria a Posteri, con far diwaardige gebeurtenissen ter mijner tutto schietta e genuinamente il Racconto, kennis gekomen zijn, welke, wel is waar, ouuero sia Relatione. slechts menschen van gering bedrijve en 2 Nella quale si vedrà in angusto Teatrolage geboorte betrefen, maar des alnietluttuose Traggedie d’horrori, e Scenetemin eene rijke vertooning opleveren van di malvaggità grandiosa, con intermezi droevige en vreesselijke ongevallen, voord’Imprese virtuose e buontà angeliche, op-beelden van drieste boosheid, doormengd poste alle operationi diaboliche. met vrome ondernemingen en verheerljkt door het zielesterkend schouwspel van hemelsche deugd, in onophoudelijken strijd met de gruwelijke aanslagen der helle, zoo heb ik besloten mij aan te gorden om daarvan der nakomelingschap een getrouw en nauwkeurig Verhaal ofte
Relaas achterlaten. 1 Ma che? quando siamo stati al punto 1 Pero, ¡oh cielos! llegado el momento de redi raccapezzar tutte le dette obiezioni e capitular las objeciones y sus respuestas y el risposte, per disporle con qualche ordine, de ordenarlas, hallamos, que habíamos hemisericordia! venivano a fare un libro. cho un libro: visto lo cual, abandonamos 2 Veduta la qual cosa, abbiam messo da nuestro intento por dos razones, que sin parte il pensiero, per due ragioni che il letd-uda alguna el lector considerará oportutore troverà certamente buone: la prima,nas. — che un libro impiegato a giustificarne un al- 2 La primera, porque temimos que el hacer tro, anzi lo stile d’un altro, potrebbe pareurn libro para justificar otro, ó solo su estilo, cosa ridicola: la seconda, che di libri bastaparecería cosa ridícula. uno per volta, quando non è d’avanzo. 3 La segunda, porque creemos que es suficiente, cuando no excesivo, el publicar un solo libro á la vez.
Gertrude domandò sommessamente e tre- Gertrudis con mucha timidez pidió la explimando, che cosa dovesse fare. cación de aquellas palabras y lo que debía hacer en consecuencia.
Il principe (non ci regge il cuore di dargli in questo momento il titolo di padre) non rispose direttamente, ma cominciò a parlare a lungo del fallo di Gertrude: e quelle parole frizzavano sull’animo della poveretta, come lo scorrere d’una mano ruvida sur una ferita. Continuò dicendo che, quand’anche... caso Él continuó diciendo que... “á pesar de lo mai... che avesse avuto prima qualche ocurrido... en el caso en que... hubiera intenzione di collocarla nel secolo, lesiido con la intención de establecerse en el stessa ci aveva messo ora un ostacolo in- mundo, ella había contraído un lazo indissuperabile; giacché a un cavalier d’onore,oluble y había creado un obstáculo invencom’era lui, non sarebbe mai bastato cible. Hombre de honor como era, jamás l’animo di regalare a un galantuomo unase habría atrevido á presentarla á ningún signorina che aveva dato un tal saggio di caballero después de tales antecedentes”. sé. «Ebbene, non si parli più del passato: tutto En hora buena; no hablemos más de lo è cancellato. pasado: todo está olvidado ya. Avete preso il solo partito onorevole, conveniente, che vi rimanesse; ma perché l’avete preso di buona voglia, e con buona maniera, tocca a me a farvelo riuscir gradito in tutto e per tutto: tocca a me a farne tornare tutto il vantaggio e tutto il merito sopra di voi.
Ne prendo io la cura.»
«Brava! bene!» esclamarono, a una voce, — Muy bien, muy bien, exclamaron á la par la madre e il figlio, madre é hijo. e l’uno dopo l’altra abbracciaron Gertrude; la quale ricevette queste accoglienze con lacrime, che furono interpretate per lacrime di consolazione.
Allora il principe si difuse a spiegar ciò che farebbe per render lieta e splendida la sorte della figlia. Parlò delle distinzioni di cui goderebbe neElntonces el príncipe habló de las distinmonastero e nel paese; ciones que Gertrudis habría de tener en el convento y en el país. che, là sarebbe come una principessa, come la rappresentante della famiglia; che, appena l’età l’avrebbe permesso, sarebbe innalzata alla prima dignità; e, intanto, non sarebbe soggetta che di nome.