Search

Abstract Understanding the nature–culture entanglement by combining the methods of natural sciences and humanities is little approached in neither of the fields. With a specific combination of methods from both digital humanities and ecology, we aimed at identifying several of people's life circumstances that relate to their individual sensitivity towards biodiversity. The circumstances with a strong correlation could be considered and targeted by decision‐makers, for example by developing specific education programmes for making people more eco‐conscious or adjusting relevant regulations. We applied machine learning techniques onto a database including information about the frequency of biodiversity mentioned in creative literature (BiL) from 1705 to 1969 as response variable related to metadata about the corresponding works and their authors as predictors, including localisation, age, gender and literature genre. The algorithm determined the response's dependency on each predictor, which can be interpreted as the intensity of this particular sensitivity parameter for biodiversity, and which we also related to time. We recognised that gender, age, region and settlement size are predictors significantly correlated to BiL. Statistically, these predictors can be viewed as starting points of the eventual individual level of awareness for biodiversity. For example, authors from villages exhibit a higher BiL than those from cities, which we interpret as a signal for the dependence of awareness for biodiversity on spatial distance from nature, which in turn can be addressed in urban development. Our conclusion is that applying a machine learning technique on literary data yields meaningful results, thereby showing potential for further similar investigations and the combination of methods from natural sciences and humanities to achieve so far unattainable insights. With our study, these insights could contribute to ecologically based decision‐making processes. Read the free Plain Language Summary for this article on the Journal blog.

Understanding the nature–culture entanglement by combining the methods of natural sciences and humanities is little approached in neither of the fields. With a specific combination of methods from both digital humanities and ecology, we aimed at identifying several of people's life circumstances that relate to their individual sensitivity towards biodiversity. The circumstances with a strong correlation could be considered and targeted by decision‐makers, for example by developing specific education programmes for making people more eco‐conscious or adjusting relevant regulations.We applied machine learning techniques onto a database including information about the frequency of biodiversity mentioned in creative literature (BiL) from 1705 to 1969 as response variable related to metadata about the corresponding works and their authors as predictors, including localisation, age, gender and literature genre. The algorithm determined the response's dependency on each predictor, which can be interpreted as the intensity of this particular sensitivity parameter for biodiversity, and which we also related to time.We recognised that gender, age, region and settlement size are predictors significantly correlated to BiL. Statistically, these predictors can be viewed as starting points of the eventual individual level of awareness for biodiversity. For example, authors from villages exhibit a higher BiL than those from cities, which we interpret as a signal for the dependence of awareness for biodiversity on spatial distance from nature, which in turn can be addressed in urban development.Our conclusion is that applying a machine learning technique on literary data yields meaningful results, thereby showing potential for further similar investigations and the combination of methods from natural sciences and humanities to achieve so far unattainable insights. With our study, these insights could contribute to ecologically based decision‐making processes. Read the free Plain Language Summary for this article on the Journal blog. Read the free Plain Language Summary for this article on the Journal blog. [ABSTRACT FROM AUTHOR]

The article sets up a critique of Sentiment Analysis (SA) tools in literary studies, both from a theoretical and a computational point of view. In the first section, a possible use of SA in narratology and reader response studies is discussed, highlighting the gaps between literary theories and computational models, and suggesting possible solutions to fill them. In the second section, a stratified taxonomy of SA tools is proposed, which distinguishes: (1) the emotion theory adopted by the tool; (2) the method used to build the emotion resources; (3) the technique adopted to accomplish the analysis. A critical survey of six representative SA tools for literary studies (Syuzhet, Vader, SentiArt, SEANCE, Stanford SA, and Transformers Pipelines) closes the article. The article sets up a critique of Sentiment Analysis (SA) tools in literary studies, both from a theoretical and a computational point of view. [ABSTRACT FROM AUTHOR]

What do characters in theater plays know about character relations, and how does the distribution of knowledge evolve over the course of a play? We present a dataset of 30 German plays annotated with information about the distribution of knowledge about character relations (such as "A learns from B that C is the parent of D"). All plays were manually annotated by two independent annotators as part of the Q:TRACK research project, which aims to systematically model character knowledge. The dataset is available on GitHub and Zenodo and can be reused, for example, for systematic studies of knowledge in plays or for analyzing disagreements between annotators.

In recent years, Romanian literary studies took one of its major methodological turns toward distant reading, using either or both quantitative and computational analysis. While quantitative analysis employed lexicographical instruments such as dictionaries and literary chronologies, computational analysis tried to approach the issue from a “data rich” historical perspective (Katherine Bode), while also attempting to build a digital corpus adapted to computational methods. The following paper attempts to survey the main research projects that deal with the computational analysis of Romanian literature in general and the Romanian novel in particular. The first part of the study undertakes a succinct state-of-the-art on past and ongoing endeavours concerned with digital approaches to the study of Romanian literature, their initial findings and potential. The second part will take a more theoretic approach to some of the key concepts related to data supported literary history. Finally, the last part of the study tackles the main challenges of developing a digital corpus of a local literature and the shortcomings related to this literature’s “locality” in terms of computational approaches and the compatibility of the tools developed by Western research projects.

We introduce childPoeDE: the first corpus of German poetry for children comprising poems which are still read today and cover a wide range of topics and authors. ChildPoeDE contains poem texts and both poem-level and token-level metadata. Poemlevel metadata includes information about the anthologies and authors, quantitative text features, rhyme and lexical richness. Token-level metadata covers word length, position and frequency, parts-of-speech, onomatopoeia and sonority. This corpus can be used for computational text analysis, but also as a source for stimulus material in experimental studies. The corpus metadata is freely accessible via Zenodo. The poem texts are protected by copyright. [ABSTRACT FROM AUTHOR]

The TRANSCOMP Dataset of Literary Translations is a collection of document-level word frequencies sampled from 10,631 translations into English of global literary fiction published since 1950, together with a historically matched parallel corpus of 10,682 fictional works originally published in English. We provide CSV files with word frequency counts for 10,000-word samples taken from each text. The associated metadata is available in a separate CSV. These data will be useful to literary scholars and linguists working in translation studies, and those interested in the linguistic, stylistic, and thematic specificity of translations from particular regions.

Abstract Nature's non‐material contributions to people are difficult to quantify and one aspect in particular, nature's contributions to communication (NCC), has so far been neglected. Recent advances in automated language processing tools enable us to quantify diversity patterns underlying the distribution of plant and animal taxon labels in creative literature, which we term BiL (biodiversity in literature). We assume BiL to provide a proxy for people's openness to nature's non‐material contributions enhancing our understanding of NCC. We assembled a comprehensive list of 240,000 English biological taxon labels. We pre‐processed and searched a subcorpus of digitised literature on Project Gutenberg for these labels. We quantified changes in biodiversity indices commonly used in ecological studies for 16,000 books, encompassing 4,000 authors, as proxies for BiL between 1705 and 1969. We observed hump‐shape patterns for taxon label richness, abundance and Shannon diversity indicating a peak of BiL in the middle of the 19th century. This is also true for the ratio of biological to general lexical richness. The variation in label use between different sections within books, quantified as β‐diversity, declined until the 1830s and recovered little, indicating a less specialised use of taxon labels over time. This pattern corroborates our hypothesis that before the onset of industrialisation BiL may have increased, reflecting several concomitant influences such as the general broadening of literary content, improved education and possibly an intensified awareness of the starting loss of biodiversity during the period of romanticism. Given that these positive trends continued and that we do not find support for alternative processes reducing BiL, such as language streamlining, we suggest that this pronounced trend reversal and subsequent decline of BiL over more than 100 years may be the consequence of humans’ increasing alienation from nature owing to major societal changes in the wake of industrialisation. We conclude that our computational approach of analysing literary communication using biodiversity indices has a high potential for understanding aspects of non‐material contributions of biodiversity to people. Our approach can be applied to other corpora and would benefit from additional metadata on taxa, works and authors. A free Plain Language Summary can be found within the Supporting Information of this article.

We introduce childPoeDE: the first corpus of German poetry for children comprising poems which are still read today and cover a wide range of topics and authors. ChildPoeDE contains poem texts and both poem-level and token-level metadata. Poem-level metadata includes information about the anthologies and authors, quantitative text features, rhyme and lexical richness. Token-level metadata covers word length, position and frequency, parts-of-speech, onomatopoeia and sonority. This corpus can be used for computational text analysis, but also as a source for stimulus material in experimental studies. The corpus metadata is freely accessible via Zenodo. The poem texts are protected by copyright.

In recent years, Romanian literary studies took one of its major methodological turns toward distant reading, using either or both quantitative and computational analysis. While quantitative analysis employed lexicographical instruments such as dictionaries and literary chronologies, computational analysis tried to approach the issue from a "data rich" historical perspective (Katherine Bode), while also attempting to build a digital corpus adapted to computational methods. The following paper attempts to survey the main research projects that deal with the computational analysis of Romanian literature in general and the Romanian novel in particular. The first part of the study undertakes a succinct state-of-the-art on past and ongoing endeavours concerned with digital approaches to the study of Romanian literature, their initial findings and potential. The second part will take a more theoretic approach to some of the key concepts related to data supported literary history. Finally, the last part of the study tackles the main challenges of developing a digital corpus of a local literature and the shortcomings related to this literature's "locality" in terms of computational approaches and the compatibility of the tools developed by Western research projects. [ABSTRACT FROM AUTHOR]

This dataset covers 41,341 manual event annotations of six German prose texts from the 19th and early 20th century comprising 290,997 tokens. For each text, the dataset includes annotations by two annotators and gold standard annotations. These annotations were used for the automation of narratological event annotations (Vauth, Hatzel, Gius, & Biemann, 2021), a reflection of inter annotator agreements in literary studies (Gius & Vauth, 2022) and the development of an event based plot model (Gius & Vauth, accepted).

The present study proposes an experimental exploration of the Romanian novel written between 1920 and 1940 through the use of stylometry, a method of distant reading employed for the statistical analysis of style. Drawing from the most recent advances in the field of computational stylistics, we select a formal standpoint from which we seek to investigate the relation between the Romanian novelistic canon and minor, tertiary novels published in the same period. In our test cases, we will attempt to establish some of the more promising aspects of stylometric analysis, as well as single out the experiments that yield no relevant result. Because of the relative novelty of the method, the purpose of our investigations is to offer a kind of pilot experiment that can illustrate the benefits of using computational methods on Romanian literary corpora.

The TRANSCOMP Dataset of Literary Translations is a collection of document-level word frequencies sampled from 10,631 translations into English of global literary fiction published since 1950, together with a historically matched parallel corpus of 10,682 fictional works originally published in English. We provide CSV files with word frequency counts for 10,000-word samples taken from each text. The associated metadata is available in a separate CSV. These data will be useful to literary scholars and linguists working in translation studies, and those interested in the linguistic, stylistic, and thematic specificity of translations from particular regions. [ABSTRACT FROM AUTHOR]

This dataset covers 41,341 manual event annotations of six German prose texts from the 19th and early 20th century comprising 290,997 tokens. For each text, the dataset includes annotations by two annotators and gold standard annotations. These annotations were used for the automation of narratological event annotations (Vauth, Hatzel, Gius, & Biemann, 2021), a reflection of inter annotator agreements in literary studies (Gius & Vauth, 2022) and the development of an event based plot model (Gius & Vauth, accepted). [ABSTRACT FROM AUTHOR]

Since Franco Moretti coined the successful term distant reading, quantitative/computational text analysis methods have gained wide circulation in literary studies. The diffusion of distant reading approaches has raised a lively debate and has attracted various criticisms, both from “traditional literary scholars” and from self-critical adopters. One important reason underlying these critical positions is the fact that it lacks sound and coherent rationales from the point of view of the theory: distant reading is the first methodology in literary studies that does not come with a theory of literature embedded in it. Consequently, all distant reading studies derive their theoretical frameworks and terms from literary theories that mostly rely on the notion that literary texts can be explained only by the way of interpretation. On what grounds, then, can we construct a theory of literature amenable to distant reading methods? I think that the better theoretical frameworks are the cognitive and bio-evolutionistic approaches to literature and cultural evolution studies. These theoretical approaches require a change in the level of description of the literary domain and justify the move from "interpretation" to "explanation" as the real aim of the scholarly inquiry.

The taxon label richness analysed so far (Figure 7, top left) represents the richness of taxon labels, ignoring the correlation to their actual taxon and thus including and distinguishing synonyms. Only one other frame within the corpus, contained in a book of animal anecdotes (Goodrich, 1845), exceeds 100 taxon labels gl Similar to a rank abundance curve in ecology, Figure 4 shows the one hundred most abundant taxon labels in the corpus, representing about two thirds of all occurrences of taxon labels. If we compare the richness trend of taxon labels to the development of the general lexical richness (Figure 7, bottom left) by calculating the ratio of taxon labels' richness and lexical richness (Figure 7, bottom right), we observe a relative decrease in taxon labels. Keywords: biodiversity in literature; computational literary studies; cultural ecosystem services; environmental humanities; historical biodiversity; Nature's Contributions to People; non-material contribution; text mining EN biodiversity in literature computational literary studies cultural ecosystem services environmental humanities historical biodiversity Nature's Contributions to People non-material contribution text mining 1093 1109 17 10/11/21 20211001 NES 211001 INTRODUCTION Our planet is losing biodiversity at unprecedented rates due to land-use change, direct exploitation, climate change, pollution and the invasion of exotic species (Cardinale et al., 2012; IPBES, 2019; Millennium Ecosystem Assessment, 2005; Tilman, 1999). 31.6% of all frames contained no taxon label and 6.5% of all frames contained 10 or more taxon labels. [Extracted from the article]

The present study proposes an experimental exploration of the Romanian novel written between 1920 and 1940 through the use of stylometry, a method of distant reading employed for the statistical analysis of style. Drawing from the most recent advances in the field of computational stylistics, we select a formal standpoint from which we seek to investigate the relation between the Romanian novelistic canon and minor, tertiary novels published in the same. In our test cases, we will attempt to establish some of the more promising aspects of stylometric analysis, as well as single out the experiments that yield no relevant result. Because of the relative novelty of the method, the purpose of our investigations is to offer a kind of pilot experiment that can illustrate the benefits of using computational methods on Romanian literary corpora. [ABSTRACT FROM AUTHOR]