1 Introduction
Every academic field of inquiry is defined by the character and format of its data. As data-types and formats change, new methodologies become relevant. Within the last two decades we have witnessed an explosive growth in born digital data as well as analogue material transformed into digital formats. This development calls for a computational reorientation within the Humanities in general and within the Studies of Religion specifically. In this chapter, I will offer a sketch of trends in the Digital or (as I prefer) the Computational Humanities, zooming in on those most likely to find their way into the research of mainstream scholars of religion dealing primarily with literary, textual objects: Distant reading strategies such as text mining. This contribution, however, holds no detailed presentations of individual methods and methodologies. Instead, my ambition is to offer a guide to what lies beyond the digital horizon. That is: I wish to offer a compass for colleagues who are unacquainted with the field so as to make easier the first steps in navigating the new methods and new types of material in research question-calibrated ways. In order to do so, I follow the show-it-donât-tell-it principle and point to some of my own work as illustrations. Because my work deals with nineteenth-century Danish religiosity, mainly through explorations of the digitized writings of âcultural saintâ, founding father of Danish democracy and modern church father of the Evangelical Lutheran Church in Denmark N.F.S. Grundtvig (1783â1872), I find it constructive to give a brief introduction to Grundtvigâs time, life and work, as well as to the (ascribed) cultural imprints hereof. In addition, I find it instructive to outline the political and scholarly aspects of the process of digitizing Grundtvigâs works. The logic behind delving into this case is that a programmatic contribution dealing with the rise of computational studies of religion must rely on what is tangible and discernable in order to be persuasive â at least more so than would be the case for programmatic papers on procedures already well-known and integrated in this field.
Finally, I will (briefly) venture two prognoses as to which subdisciplines and subfields will be the first to systematically add computational methodologies to the know-how portfolio of the Study of Religion. My predictions are perhaps surprisingly concrete and delineated: 1) Text mining will invite a growth in the study of nineteenth-century religiosity. This prediction does not (only) spring from the fact that this happens to be my own research field; it builds on hard fact â on accessibility and quality evaluations of the available material. 2) Text mining will bring about the return of the philologist â an academic profile essential to the emergence of the historical Study of Religion as a modern discipline in the nineteenth century but also one that, with regional differences, nevertheless has been waning for decades. However, soon Vedic, Avestan, Sanskrit, Sumerian, Hittite, Arabic, Greek, Latin, Norse and other scholars will be in high demand; the growing landscape of digitized corpora will call for exploration rooted in solid linguistic skills.
Thus, this chapter differs from others I have ever written in containing scholarly prognoses as well as pieces of academic advice. I have even included my personal programmatic conceptualizations of the emergent field of Computational Humanities. These conceptualizations constitute my point of departure.
2 Computational Humanities
Let me be clear from the outset. In this chapter, my interest lies with the mainstream (text-based, historical) Study of Religion and the consequences for these that arise from the developments in the growth of computational methods and in the expanding quantity of digital text material. It does not lie with measuring sedimentations in the religious landscapes of what may be regarded as a âDigital Ageâ. Thus, I will not focus on digitization as a cultural process â nor for that matter on related labels that try to capture aspects of the postmodern condition: âmediatizationâ, âacceleration cultureâ, âepidemics of lonelinessâ or âGegenwartsschrumfungâ. Within recent years Quran, Bibles, hymnals etc. have acquired digital alternatives (mainly software applications) to their traditional book-based versions and this could be said to be changing devotional-ritual practices for certain religious groups.1 Similarly, within the last couple of decades, social media have gathered ever-growing virtual communities around fiction-based religions.2 These are just a couple of interesting religious phenomena connected to digitization as a cultural trajectory. These are indeed valid research subjects; but they are not the interest of this contribution. In order to signal a focus on methods and on âdata qualityâ â above subject matter â I therefor prefer the somewhat delimited âComputational Humanitiesâ (CH) over the more common but also more diffuse and open-ended âDigital Humanitiesâ (DH).3 By CH I mean research on human culture and society â that is: subjects deemed relevant for scholars of Humanities â that in a significant way relies on digital data and on computational power in the analysis hereof. In other words, I mean to refer to explorations into âthe human conditionâ (in the broadest possible sense) that import insights and routines from fields such as Computer Science and Statistics.4
Moreover, I see neither DH nor CH as disciplines in their own right. They are better construed as hyper-disciplinary attitudes, trends, experience, and interest aggregations etc. Accordingly, I do not foresee DH or CH coming to life as bachelor or master programs at universities around the world; at least I do not see them attaining any longevity in such formats. Often the best CH results come from the iterative workflow of a close-knit, cooperative unit consisting of, on the one hand, scholars versed in relevant computer science theories and procedures and, on the other, scholars trained in the subject domain in question. Most likely this is the reason that the tendency these years in terms of infrastructure seems to be to establish hub- or center-like constructions based on a group of technically skilled scholars supporting and collaborating with scholars from a variety of humanities disciplines. Obviously, this situation might change. When the bulk of cultural material is born digital, when the bodies of analogue cultural heritage material have all been digitized, and when digitally native generations enroll at university, then CH will most likely become seamlessly integrated into the Humanities. The need for the âcomputationalâ prefix will then dissolve. But for the time being and for the immediate future, we have to accept the need for âleapsâ of interdisciplinary-collaborative trust, for inspiration from first-movers, and for concrete bridging in co-authored publications, if we as traditional scholars of religion want to join the team of early computational adaptors. At least, the acceptance hereof has been a vital step for me â one that has allowed me to solidify, broaden and deepen my studies of the religious landscape of nineteenth-century Denmark in ways I could never have imagined. I will return to this work. Not as a chance for me to showcase it, but springing from the wish to demonstrate the potential of CH methods â and what better way to do this than by presenting what I have done myself?
Since most of these studies gravitate towards the digitized writings of N.F.S. Grundtvig, a brief sketch of this central figure in Danish national and religious history and of the ongoing process of digitizing his published writings is required. This sketch also tells the tale of an emergent digital cultural heritage infrastructure.
3 The Spotless Saints and the Dirty Masses
In Danish public discourse poet, pastor, politician, and romanticist N.F.S. Grundtvig (1783â1872) is regarded as one of the most (if not the) central figure in the nineteenth-century Danish nation building process, as well as in the construction of a modern Danish Christianity: In short, he is regarded as a cultural saint.5 In scholarly literature it is widely acknowledged that Grundtvig sought to stimulate the process of assembling a collective Danish emotional consciousness based on 1) a horizontal-contemporary axis incorporating the different strata within the socially heterogeneous âFolkâ and on 2) a vertical-historical axis connecting present-day Danes with forefathers and legendary characters. In social historian Benedict Andersonâs words, the emotional fabric intended by this attempted interlacing was an âimagined communityâ.6 Nowadays, Grundtvigâs cultural imprints are acknowledged by most Danes: âN.F.S. Grundtvig founded Danish democracyâ; âN.F.S. Grundtvig established the Church of Denmark (folkekirken)â; âN.F.S. Grundtvig is the founder of the Danish school systemâ; âN.F.S. Grundtvig revived the pre-Christian Nordic traditionâ; âN.F.S. Grundtvig is the most important writer of Christian hymns in Denmark.â These are surprisingly recurrent statements in Danish public media, deeming his intellectual activity more culturally important than the work of his world-famous contemporaries Søren Kierkegaard (1813â1855) and Hans Christian Andersen (1805â1875).
Considering Grundtvigâs cultural status, it is perhaps no surprise that a consortium including members of the Danish parliament in the late 2000s decided to pave the way for the creation of a scholarly edited, digital version of Grundtvigâs 1073 published writings, making them available to Danish citizens free of charge.7 The writings, published within a period of 68 years (1804â1872), amount to approximately 37 000 standard pages of 2400 units per page. The data set has a median document size of four pages and contains 3 968 841 word tokens distributed over 115 240 word-types.8 This is not an impossible amount of words to read oneâs way through; it is, nevertheless, an overwhelmingly time-consuming job, especially if one sets out to tweak out semantic and sub-semantic structures, as I have done. For such purposes a digitized version of the corpus is a prerequisite.9
By way of OCR (Optical Character Recognition) technology, first editions of each text have been âtranslatedâ from a TIF-format (Tagged Image File Format) into an XML (Extensible Markup Language) document, manually annotated and following TEI (Text Encoding Initiative) standards.10 A crucial step in producing an accurate digital corpus is thus the labor-intensive cleansing and annotation of the raw and oftentimes somewhat dirty OCR results. This work is undertaken by a group of ten Grundtvig-specialized scholarly editors â philologists trained in fields relevant for the domestication of Grundtvigâs prose, such as (obviously) nineteenth-century Danish but also Old Norse, Greek, Bible Studies, hymnology, romantic philosophy, eighteenth century historiography etc. Furthermore, this équipe furnishes the individual texts with contextualizing introductions and glossaries. Their work, piling up on www.grundtvigsvaerker.dk, was initially estimated to consume 200 man-years. This prognosis seems to hold: 10 years and 70 million DKR (app. 10 million Euro) later, the ten scholarly editors are half way through the project.
Such details are highly relevant when drawing up the contours of the computational future of the Study of Religion, because poor OCR quality is what is keeping the major digital archive projects from significantly changing the game of traditional humanities. In other words: the well-known Google Books, the HathiTrust Digital Library, the Project Gutenberg and the ever-increasing amount of national digital archives based primarily on ORC technology will not be the ones to push the CH into the center of humanistic scholarship.11 They do and will continue to pave the way for easy-access to historical documents. For this they must be applauded and appreciated. But the most likely scenario is that small, curated digital editions, strenuously enriched by philologists, will be the key player in the process of integrating CH into traditional scholarship; first and foremost because this material is clean, reliable, and flexible: Thorough markup leaves it open to comprehensive, fine-grained, hermeneutically complex explorations.
Clean, reliable, and flexible but also exclusive. High quality data is, as I have indicated, burdensome to create in terms of time and funding. Not every type of material will move politicians and research foundations to cover the expenses.12 So, in order to bypass the inherent centripetal force of canonical corpora,13 computationally ambitious scholars of religion who are more interested in marginal phenomena than in âdoxaâ will most likely have to be creative and to compensate for the dirt in the big data corpora. (At least for now. Who knows what the algorithmic future of rampant machine learning will bring?) I will return to the subject of hacking the cultural canon below. But first a note on scope.
4 Nothing New under the Sun
The tension between the available and the wished-for empirical material is as old as the academic vocation itself. Furthermore, a wide range of concerns within text-based CH seems to be echoing considerations ingrained in historical disciplines, such the historical study of religions. In fact, a long line of CH studies are perhaps best construed as traditional manual strategies in a new computational key. Four such strategies are worth emphasizing. Firstly, scholars accustomed to traditional text survey tools such as concordances14 â e.g. classic, philologically based scholars of religion â are obvious forerunners of the emergent CH distant readers. Secondly, the question of stylometric profiling and authorship attribution, for so long recurrent in a variety of historical fields such as Bible studies, are now converted into ponderings over statistic similarities between documents or pools of documents. Thirdly, the focus maintained by Annales School, the structuralists, and the post-structuralists on collective mentalities, on the stability of la longue durée, on the synchronic, on discourse, on nodal points etc. seems relatively easily translated into the research designs of CH based studies.15 Fourthly, the diachronic ambition of conceptual and emotional history is a suitable theoretical underpinning of the ever-increasing volume of studies tracing the rise, fall, or dispersion of given concepts and clusters of concepts in a multitude of digital corpora.
In addition: Awareness of how the scope of a given study is situated on the synchronic-diachronic axis is always relevant; within CH it is, however, an acutely important step that secures the basis for choosing the appropriate frames, tools, and strategies. If one combines the aforementioned question of archive quality and size (as noted: perspectives that seem to be inversely proportionate) â that is: If one combines the question of homogenous vs. heterogenous digital material with the logically opposite research scopes, the synchronic and the diachronic, one ends up with the following matrix:
Methodological matrix
© Frøkjær Baunvig
In all its simplicity, this matrix may be helpful for scholars of religion who embark on computational endeavors for the first time. It may tease out reflections on material accessibility, the condition of a given archive, the scope of the underlying research questions â all necessary reflections when seeking to align subject matter, empirical material, and methodology into a robust research design. At least it has done so for me. In the following, by way of brief examples drawn from my own work with numeric representations of large corpora of texts, I will illustrate the four systematically different âideal typesâ of approaches suggested by the matrix â bearing in mind that reality is always messy and that actual studies will always engage differently with these ideals.
4.1 Spiritual and Earthly Structures in Grundtvigâs Writings [A1]
In the first series of studies I would like to present, a group of colleagues and I have sought to draw the contours of Grundtvigâs world-view, as well as to map focal points in his conceptual landscape. In one of these, I have used the text mining strategy of Word2Vec to plot the semantic embedding of the central terms Aand (Spirit) and aandelig (spiritual) within the Grundtvig corpus. Based on the computer linguistic paradigm NLP (Natural Language Processing), the Word2Vec-procedure, in a manner of speaking, âtrainsâ an algorithm that allows you to plot associative structures within a given body of texts.16 In this particular case, lists of the 20 words most common in structure to Aand and aandelig allowed me to demonstrate that, to Grundtvig, âthe âSpiritâ can be described by two kinds of metaphors; on the one hand it is a solid container, on the other hand it is a fierce and dynamic force.â17 Introducing a diachronic sensitivity within the synchronic mainframe, a division of the main corpus into seven sub-corpora divided by publication decade showed that this fundamental metaphorical split is a constant in Grundtvigâs writings. These insights affirm, pin-point and systematize hunches in the existing Grundtvig research literature.
The neural network of âEarthâ, âHeavenâ and âHellâ in N.F.S. Grundtvigâs published writings
© Frøkjær Baunvig
In another of these studies expanding the relatively simple Word2Vec embedding procedures, a colleague and I trained an algorithm â or a so-called âneural networkâ â that enabled us to plot the density between three crucial semantic centers of gravitation within Grundtvigâs writings: Jorden ((the) Earth), Himmel (Heaven), and Helvede (Hell). Simply put, this procedure allows for the combination of several word embeddings. In this particular case â and to some surprise in terms of the strength of the findings â it teased out Grundtvigâs fundamentally orientalist world-view: The dense center of his semantic universe is Sol/Soel (Sun), skinner (shines/shining), stiger (rises/rising): The radiant, rising sun.
In these two studies, the main focus is on synchronic relations â semantic structures within the Grundtvig corpus as such. I will now turn to studies calling for a slightly more diachronic approach and for the subtraction of two different types of sub-corpora from the full Grundtvig corpus.
4.2 Emotional Imprints in Grundtvigâs Writings [A2]
Less is more. That was the lesson I drew from the next of my illustrations: A distant reading of letter-s p a c i n g s and exclamation marks in the Grundtvig data. Letter-spacings, also known as âtrackingsâ, was a common tool of emphasis in nineteenth-century Fraktur typesetting used in the vast majority of the 1073 publications by Grundtvig. But where mainstream usage would confine letter-spacings to introductions of names and titles, Grundtvigâs usage was idiosyncratically expansive. Also, his usage of exclamation marks was unusually extensive. Seeking to target what we deemed two traits of a general romanticist stylistic trend introducing oral features into printed matter and saturating typography with emotion, a group of colleagues and I sketched the (synchronic) semantic context of the spacings and the exclamation marks. In order to eradicate possible âdistortionâ from editors and printers we also made a (diachronic) outline of the temporal distribution plotting of spacings and exclamation marks in Grundtvigâs six main publishing houses. We found no signs of distortion. Instead, what we found in both studies was two concurrent, robust semantic clusters confirming the mainstream Grundtvig reception: Grundtvig soaked Christian terms and Danish-Nordic-national terms in emotion. At least he was considerably more prone to making use of spacings and exclamations when writing on these issues.
Simple fetch elements by tag operation, made possible by the high-quality markup of the Grundtvig files, helped us extract a letter-spacing sub-corpus;18 simple parser features allowed us to split the corpus into sentences and thus to extract sentences that ended in exclamation marks to establish a sub-corpus on which we did a so-called topic modelling procedure that helped us to determine the semantic content of the sub-corpus.19
Relatively simple procedures, but by no means trivial findings. This twofold distant reading rooted a duality dominant in the mainstream Grundtvig reception in Grundtvigâs own writings. A highly interesting finding, which makes it possible to suggest a limit to the implications of the notion, held by current Grundtvig scholars such as J.F. Møller, that legacy and reception have distorted the core-content of Grundtvigâs work. At least, this is not evident from our explorations at scale. In fact, they indicate quite the contrary: that Grundtvigâs emotional imprints, his signals of core-content, are very much consistent with the âGrundtvig mythâ and with public reception.
The relationship between the myth and the man, reception and writings, I have further explored in a study even more diachronically sensitive â that is: sensitive to developments within Grundtvigâs writings.
4.3 Sub-Semantic Validation of a Grundtvig Fault-line [B]
As a consequence of Grundtvigâs status as a cultural saint, a large proportion of Grundtvig research has been preoccupied with biographical issues. Sometimes to the point of hagiography. A recurrent theme is that of dating one or several significant reorientations in his religious mindset â moments of repentance, existential crisis and dogmatic re-calibration: 1810, 1824â1826, 1832, 1838. Time and again, dates and noteworthy events have been put forth. In a âneo-biographicalâ study my colleague and I scrutinized the sub-semantic tectonics of Grundtvigâs writings, searching for faultline(s).20 Using American mathematician Claude Shannonâs (1916â2001) relatively simple information theoretical metric, the so-called âShannon entropyâ (H)21 (in our case as a measure for âinformational recurrenceâ), we sought to measure the uncertainty or âpredictabilityâ of a given Grundtvig-piece. Confirming to consensus in the computational linguistic field, which holds that H is a valid measure for lexical diversity or ârichnessâ,22 we tested whether significant shifts in the informational density of Grundtvigâs texts could be used as an indication of a possible Kehre. Adding to our entropic map we introduced the chaos theoretical feature of a so-called ârecurrence plotâ (in this case an âauto recurrence plotâ) in order to identify the possible faultlines in Grundtvigâs oeuvre.
Level of information and auto-recurrence plot in N.F.S. Grundtvigâs published writings
© Frøkjær Baunvig
What we found might be seen as the contours of Grundtvigâs consolidation as a writer: A) Early writings are markedly more unpredictable than later writings and B) in terms of lexical richness Grundtvig underwent a volatile and unstable period from 1824 and onwards, with a climax in 1832.
This development makes it relevant to regard especially 1832 as a faultline. If this finding has not definitively settled a long line of biographical considerations, it has at least qualified them significantly.
Turning from a single oeuvre to cultural mentality â from the relatively homogenous Grundtvig material to a vaster and more diverse corpus of nineteenth-century Danish newspapers â the next illustration poses ideas on how to circumvent the centripetal forces of cultural heritage management.
4.4 The Rise of the Demi-Goddess âDanaâ in Danish Nineteenth-Century Newspapers [C1]
In a study of the rise of an allegorical-poetic personification of Denmark â the demi-goddess Dana â in Danish newspapers 1800â1870, I accepted OCR dirt and, with considerable help from my astute student assistant Stine Kylsø Pedersen, compensated with hours of manual cleaning, annotation and coding.23 This study was based on data made available through the Danish Royal Libraryâs digital archive Mediestream.24 The Danish Royal Library have scanned 35 464 209 Danish newspaper and periodical pages published between 1666 and 2013; by way of OCR software this material is made searchable in an interface allowing for date delineations. This enabled a straightforward list-generation conveying the total number of occurrences of âDanaâ in Danish newspapers within the period from 1 January 1800 through to 31 December 1869. A distant reading of the distribution showed a significant and steady increase in occurrences of the name up until the 1850s and a boom in the 1860s:
âDanaâ on the rise in Danish nineteenth-century newspaper material
© Frøkjær Baunvig
Close reading controls of the Dana-hits allowed us to drop false positives (not included in the table above). Mediestream provides a facsimile display permitting such control. Using this feature, we have rejected false hits. But more importantly, reading our way through the âtrue Danasâ we systematically coded each of them to gain an overview of their genre and semantic context.25 This was a time-consuming but rewarding procedure. It made it possible to outline the âcoming of ageâ of a national spirit: Dana began her life in Danish conscience collective â that is, in the nineteenth-century Ãffentlichkeit-material par excellence: newspapers â as a young nymph-like creature in the first decades of the nineteenth century, when the reading public was still thoroughly familiar with the fantastic creatures of antiquity. Gradually, however, she became ever more associated with the landscape and climate of Denmark. In times of the national crisis that culminated in the First Schleswig War in the late 1840s and early 1850s, she changed into a stern motherly-martial character and grew in popularity. This was her heyday in national poems and songs printed in the papers. From hereon she broadened her activities and morphed into steamboats, tonics and even potato-grading machines. In other words, she was thrown into the murky commercial waters. The main observation here is that Danaâs figurative contours began to blur at exactly the point in time when the imagined social conception of âDenmarkâ had become reasonably robust and widespread.
An important appendix: This procedure has an inbuilt, self-evident blind spot: False positives are relatively easily detected. But there is no way of telling how many Danas the OCR dirt made us miss. Here we must simply take comfort in the assurance that quantity can give â in the assumption that the size of the corpus secures an acceptable statistical significance â as the CH analogy to a Social Science and Humanities survey, perhaps? When it comes to hacking the structures of cultural heritage formations that make it difficult to find digital material suited for the exploration of mainstream (as opposed to elitist-canonical) mentalities and discourses, this procedure is simply the one that seems to be the best and most feasible at the moment (and in the near future).
4.5 Measuring Fluctuations in Grundtvigâs Posthumous âRelevance Levelâ [C2]
In a series of investigations, I have relied on the Danish Royal Libraryâs digital archive Mediestream for yet another diachronic task: I have sought to locate a proxy for what one could refer to as Grundtvigâs âposthumous relevance levelâ, his âcultural imprintâ or his âreception historyâ in the years between his death in 1872 and today. All Danish newspapers taken together allow us to indicate fluctuations in the interest payed to Grundtvig in Danish public opinion. Due to copyright regulations, it is not possible to access facsimile versions of newspapers published after January 1 1921.26 One can, however, search the corpus as a whole. This affords evaluations of the frequency and distribution of âGrundtvigâ compared to other prominent nineteenth-century personalities, such as e.g. (Søren) âKierkegaardâ (1813â1855). Luckily both surnames are relatively infrequent, so one can assume that the main bulk of occurrences will in fact refer to the Grundtvig and the Kierkegaard.
Distribution of âGrundtvigâ and âKierkegaardâ in Danish newspaper material
Accessed at: http://labs.statsbiblioteket.dk/smurf/Note: Mediestream covers material from 1666 to 2013; the so-called âSmurfâ search tool at the Danish Royal Libraryâs KB Labs covers material published from 1750 to 2011
Following the spike marking the bicentennial anniversary of Grundtvigâs birth in 1983, from the mid-1980s through to 2011 Grundtvig and Kierkegaard take turns to receive the most attention in Danish newspapers. Historically, however, Grundtvig have dominated the material. In his life and day, Grundtvig was (among a variety of other things) a politically engaged writer of newspaper columns and a public character, whereas Kierkegaard was decidedly not politically interested. To a certain extent, this explains the significant differences in the representations of Grundtvig and Kierkegaard in newspapers in the mid-1800s â that is: before Kierkegaardâs death in 1855. Peaks of interest and relevance are evident in years marking jubilees of birth or death: For Kierkegaard 1913 (marking his 100th birthday) and 1955 (marking the 100th anniversary of his death) represent modest crests; an increase in public attention in the years of Grundtvigâs birthday anniversaries (1883, 1933, 1983) is evident, as it is detectable in the years of anniversaries of his death (1922, 1972). A Grundtvig revival during the Second World War is furthermore apparent, whereas Kierkegaardâs relevance seems relatively stable from the 1920s onwards.
Distribution of âGrundtvigâ and âKierkegaardâ in Google Books âFrench (2019)â Corpus
Accessed at: http://books.google.com/ngrams
Distribution of âGrundtvigâ and âKierkegaardâ in Google Books âGerman (2019)â Corpus
Accessed at: http://books.google.com/ngrams
It is no great surprise that the tables turn when searching through the 40 million books scanned by Google Books (so far).27 Kierkegaard, on his part, experiences a breakthrough in the first third of the 20th century in the English, German and French 2019-corpora (Tables 6, 7 and 8). This turns into a general, accelerating interest in the 1940s and 1950s. Overall, Kierkegaard seems to be least relevant in the German corpus; his earliest peak occurs in the French corpus; in the French as well as the German corpus he suffers a significant decrease from the 1950s onwards. Kierkegaardâs relevance level has, however, increased remarkably in the English corpus since the late 1980s. In these respective heterogenous digital text corpora, Grundtvig, however, is virtually absent (Tables 6, 7 and 8).
Distribution of âGrundtvigâ and âKierkegaardâ in Google Books âEnglish (2019)â Corpus
Accessed at: http://books.google.com/ngrams
Such corpus material combined with the ready-made display tools they offer are seductively easy to use. With no access to the back-end â to the underlying text-material of the datasets â in effect, they are black boxes suited only for heuristic purposes and must be handled with a great deal of hermeneutic caution. But when one is disciplined in terms of limiting the stakes of the search (when one does not rely on them for answers to central research questions), these big, dirty and obscure datasets can be the foundation of quite indicative sketches, as I will argue in the next section.
4.6 The Acceptable Reference Corpus [D]
âSearches with Google Books Ngramâs German/English/French sub-corpus will testify to this claimâ; âsearches with the Danish Royal Libraryâs Mediestream newspaper corpus will testify to this claim.â Reports from my different explorations of trends in nineteenth-century religiosity are full of such statements. Oftentimes the trend in focus will be operationalized linguistically, and often it is found in the Grundtvig corpus. My reason for consulting large and somewhat dirty corpora such as Google Books or any other given OCR-based archive of a considerable volume, such as Mediestream, is that they by virtue of their âbig dataâ nature offer acceptable material for backdrop analyses â for straightforward checks of overall trends in the distribution of given terms or clusters of terms.28
Allow for a brief interlude concerning the concept of âbig dataâ: âData used to mean documents and papers, with maybe a few photos, but it now means much more than that.â29 Now, it means the accumulation of unfathomable points of information on commercial, healthcare and social issues, changing how we conceptualize societal infrastructure and the conditions for analyzing it. Dawn E. Holmesâ âBig Data. A Very Short Introductionâ (2017) invites reflections on what data is as a concept and as a phenomenon, its history and its future applications.30 As a computationally dedicated scholar of religion it is, however, important to emphasize that such underlying triumphalism is difficult to maintain when confronted with ever thickening barriers of data access regulations: Commercial data (including social media data), on the one hand, is hard for non-corporate, university researchers to access due to copyright regulations; healthcare and social security data, on the other, is extremely difficult (for humanist scholars in particular) to access due to GDPR directives.
But back to the main point, which is that I have often found it helpful (enlightening, informative) to consult big data archives as reference corpora in order to tease out a reasonably reliable insight in whether a given trend I am investigating conforms to main developments or whether it differs from them. For instance, when seeking to evaluate Grundtvigâs usage of the term âSpiritâ (Aand), I have found it valuable to consult Danish nineteenth-century newspapers in order to answer questions such as: Does the frequency of the term in the Grundtvig corpus map onto or diverge from a general linguistic trend? Does the semantic distribution in the two corpora converge? For tentative use, such backdrop sketches with OCR-based, big data corpora represent acceptable sources of information.
Through the above examples of work with numeric representations of large corpora of texts, I believe that I have indicated future contours of research designs â future approaches that will find their ways into the mainstream historical study of religions. From these demonstrations of procedures oriented towards digital material and to varying degrees driven by computational methods, I will move on to the concluding paragraphs of this chapter â to my predictions of what CH will encourage in the historical study of religions.
5 Prediction 1: Nineteenth-Century Religiosity
Data quality and copyright regulations: These are the Scylla and Charybdis of any Computational Humanist. As a scholar of nineteenth-century religiosity, I cannot help but take comfort in the fact that material from the long nineteenth century represents a sweet spot between the two. Too old to be covered by (most) copyright regulations, the material already digitized is unburdened by access restrictions. Most twenty-first-century material, born digital data and a large proportion of twentieth-century digitized material is covered by a variety of regulations, making the accessibility process tiring and often unsuccessful.31 This is not the case with material of the long nineteenth century. Moreover, typographically similar to contemporary texts, nineteenth-century print is easier to manage than e.g. early-modern print or medieval and antique manuscript material.32 Furthermore, produced in a historical epoch considered critical in the formation of modern nation-states, public authorities and private foundations around the world seem to take pride in funding digitization projects dedicated to nineteenth-century material: There is simply more of it available in digital formats than material from other historical periods.
As mentioned, this trend is strongest among canonical works and oeuvres of persons deemed culturally significant, such as e.g. N.F.S. Grundtvig (1783â1872) in Denmark,33 Henrik Ibsen in Norway (1828â1906),34 and Rabindranath Tagore in India (1861â1941).35 Gradually, however, pulp fiction and pamphlets but first and foremost newspapers are emerging digital archives as well.36 I predict that the growth in richly annotated, small corpora combined with the growth of acceptable, big data archives eligible as reference corpora will change the conditions for text-based studies of cultural and religious developments in the nineteenth century before it changes the conditions for text-based studies of drifts within other periods.37
As I hope will have become clear by know, I consider the careful quality evaluation of a given digital text corpus essential in order to assess which computational tasks would be safe to demand of it and in order to assess the appropriate emphases one can ascribe to results. But the main point here is that, in terms of available material, the long nineteenth century is privileged. Not least when it comes to European and North American material. Based on these simple facts, I predict that the next decade will witness a variety of studies of nineteenth-century representations of lay and elite Christianity, qualifying, nuancing, or contesting theoretical claims such as, for instance, Linda Woodheadâs neo-classic description of the nineteenth century as the onset of two social processes that conditions present-day Christianity: The âsubjective turnâ and the âfeminization of pietyâ.38 One way of approaching such investigations could be to assess whether there are significant âmacro driftsâ in subject and pronominal terms dominating the religious-devotional discourses embedded in diverse forms of text material. Another field inviting computational exploration could be the one outlined by Max Weberâs notion of modernity. It would be very interesting to put his classic notion of âEntzauberungâ (enchantment) and rationalization â of the gradual, historical process culminating in the late nineteenth century by dividing the world into independent value-spheres (religion being one of them)39 â to the test. A way to operationalize such a task could be to trace dispersions, contractions, and reconfigurations of religious discourses (i.e. neural networks between relevant terms)40 in a vast and genre-heterogenic material.
In any case, on the basis of the digital material at hand, I predict that investigations at scale of the transformation of religiosity in the long nineteenth century will be a growing field within the historical Study of Religion in the decades to come.
6 Prediction 2: Return of the Philologist
There was a time, intellectual historian James Turner argues, when philology âreigned as king of the sciencesâ;41 when philological competences were âthe pride of the first great modern universities.â42 This time, however, is no longer. In the course of the last few decades, comparative-philological activity seems to have been undermined from within academia â a process among many other things reflected in university recruitment strategies. Yet, the so-called âmassificationâ of higher education43 may also be central to the erosion of (the need for and relevance of) philological erudition, a line of post-structural, critical theorists have with growing strength challenged the acceptability of scholarly pursuits focusing on religious world-views and languages in post-colonial areas.44 In other words, the problem is that philology to some extent âowes its existence to European imperialism.â45
So, since the middle of the twentieth century the status of philology has changed significantly. This is James Turnerâs assessment: In his monograph âPhilology: The forgotten Origins of the Modern Humanities,â he outlines two basic types of philological endeavor: comparative philology and textual philology46 â respectively diachronic and synchronic in scope.47 The oscillation between these two philological modes, he argues, drew forth the comparative study of religions, when it came into being as an academic discipline proper in the latter half of the nineteenth century. I propose that it might soon do so again. I agree with Turner that philology did not succumb to the post-colonial pressure; that it âdid not vanish,â but only âwent undergroundâ48 for a while. And I think that now might be the time for its return? At least, I see signs that philology is once again on the rise. These signs are mainly based on funding drifts and material accumulation. In addition to biblical material and the aforementioned nineteenth-century material, we see a growing number of digital archives, repositories, and editions of material relevant to the study of religions accumulating online. The quality â that is: the extent of the preprocessing procedures going into establishing this material â vary greatly, but the main trend is evident. A few examples could be projects such as these:
Digital archives relevant for studies of religions
© Frøkjær Baunvig
Every academic field of inquiry is defined by the character and the format of its data. As data-types and formats change, new methodologies and competences becomes relevant. But when old material comes to life in this way, old skills become relevant once again. In the years to come, Vedic, Avestan, Sanskrit, Sumerian, Hittite, Arabic, Greek, Latin, Norse scholars will, thus, once again, be in high demand; the growing landscape of digitized corpora will call for exploration rooted in solid linguistic proficiencies.
Bibliography
Anderson, Benedict. Imagined Communities. Reflections on the Origin and Spread of Nationalism. London: Verso, 1983.
Ayyadevara, Kishore. Pro Machine Learning Algorithms: A Hands-On Approach to Implementing Algorithms in Python and R. Apress, 2018.
Barats, Christine, Valérie Schafer, and Andreas Fickers. âFading Away ⦠The Challenge of Sustainability in Digital Studies.â Digital Humanities Quarterly 14, no. 3 (2020): 165â189.
Bartlett, Robert. Why Can the Dead Do Such Great Things? Saints and Worshippers from the Martyrs to the Reformation. Princeton: Princeton University Press, 2013.
Baunvig, Katrine Frøkjær. âFictional Realities of Modernity: The Fantastic Life of the Demi-Goddess Dana in the Emerging Nation State of Denmark.â In Mythology and Nation Building in the Nineteenth Century: N.F.S. Grundtvig and His European Contemporaries, edited by Lone Kølle Martinsen, Sophie Bønding, and Pierre-Briece Stahl, pp. 97â134. Aarhus: Aarhus University Press, 2021.
Baunvig, Katrine Frøkjær. âThe Spiritualist.â Journal for the Cognitive Science of Religion. Sheffield: Equinox, forthcoming.
Baunvig, Katrine Frøkjær, and Kristoffer Laigaard Nielbo. âKan man validere et Selvopgør?â In Textkritik som analysmetod, edited by Paula Henrikson, Mats Malm, and Petra Söderlund, pp. 45â67. Stockholm: Svenske Vitterhetssamfund, 2017.
Baunvig, Katrine Frøkjær, Oliver Jarvis, and Kristoffer Laigaard Nielbo. âEmotional Imprints: Letter-Spacings in N.F.S. Grundtvigâs Writings.â Digital Humanities in the Nordic Countries Proceeding (2020): 192â202. http://ceur-ws.org/Vol-2612/short3.pdf.
Baunvig, Katrine Frøkjær, Oliver Jarvis, and Kristoffer Laigaard Nielbo. âEmotional Imprints: Exclamation Marks in N.F.S. Grundtvigâs Writings.â Digital Humanities in the Nordic Countries Post Proceeding (2021): 156â169. http://ceur-ws.org/Vol-2865/short7.pdf.
Baunvig, Katrine Frøkjær and Stine Kylsø Pedersen. âThe Dana Repository.â In The Grundtvig eStudies Pamphlets 1 (2021). https://grundtvigcenteret.au.dk/fileadmin/ingen_mappe_valgt/The_Grundtvig_eStudies_Pamphlets_1.pdf.
Brügger, Niels. âDigital Humanities in the 21st Century: Digital Material as a Driving Force.â Digital Humanities Quarterly 10, no. 2 (2016) http://www.digitalhumanities.org/dhq/vol/10/3/000256/000256.html.
Christiansen, Johanne Louise, and Katrine Boserup Jensen. âDa Koranen blev en App ⦠En genforhandling af Koranens Status som helligt Objekt.â Religionsvidenskabeligt Tidsskrift 70 (2020): 112â129.
Cohen, Margaret. The Sentimental Education of the Novel. Princeton: Princeton University Press, 1999.
Davidsen, Markus. The Spiritual Tolkien Milieu: A Study of Fiction-based Religion. PhD thesis, University of Leiden, 2014.
Descola, Philippe. Beyond Nature and Culture. Chicago: The University of Chicago Press, 2013. English translation of Par-delà nature et culture, published in 2005.
Franzini, Greta, Melissa Terras, and Simon Mahony. âA Catalogue of Digital Editions.â In Digital Scholarly Editing. Theories and Practices, edited by Matthew James Driscoll and Elena Pierazzo, pp. 161â182. OpenBook Publishers, 2016. https://library.oapen.org/bitstream/id/9668bc0d-eb07-4b4f-a5e5-0cf335188694/633780.pdf.
Foster, Charles. Being a Beast. An intimate and radical look at nature. London: Profile Books, 2016.
Gumbrecht, Hans Ulrich. The Powers of Philology: Dynamics of Textual Scholarship. Washington: Library of Congress Cataloging-in-Publication Data, 2003.
Hadju, Péter. âDigital Editions of Canonic Texts: Do such Projects Affect the Canon?â Texto Digital 6 (2010): 85â95.
Holbraad, Martin, and Morten Axel Pedersen. The Ontological Turn: An Anthropological Exposition. Oxford: Oxford University Press, 2017.
Holmes, Dawn E. Big Data: A Very Short Introduction. Oxford: Oxford University Press, 2017.
Jørgensen, Finn Arne. âThe Internet of Things.â In A New Companion to Digital Humanities, edited by Susan Schreibman, Ray Siemens, and John Unsworth, pp. 42â53. London: Blackwell, 2016.
Kilbride, William. âSaving the Bits: Digital Humanities Forever?â In A New Companion to Digital Humanities, edited by Susan Schreibman, Ray Siemens and John Unsworth, pp. 408â419. London: Blackwell, 2016.
Lee, Haimin. â15 Years of Google Books.â 2019. https://www.blog.google/products/search/15-years-google-books/.
Moretti, Franco. âThe Slaughterhouse of Literature.â Distant Reading. London: Verso, 2013.
Møller, Jes Fabricius. Grundtvigs Død. Aarhus: Aarhus University Press, 2019.
Pierazzo, Elena. Digital Scholarship in the Humanities 31, no. 3 (2016): 513â516.
Rosado, Delia Langa, and Miriam E. David. ââA massive University or a University for the Masses?â Continuity and change in higher Education in Spain and England.â Journal of Education Policy 21, no. 3 (2006): 343â365.
Said, Edward W. Orientalism. New York: Pantheon, 1978.
Schreibman, Susan, Ray Siemens, and John Unsworth. A New Companion to Digital Humanities. London: Blackwell, 2016.
Shannon, C.E. âA Mathematical Theory of Communication.â The Bell System Technical Journal 27 (1948): 379â423, 623â656.
Thoiron, Philippe. âDiversity Index and Entropy as Measures of Lexical Richness.â Computers and the Humanities 20, no. 3 (1986): 197â202.
Turner, James. The Forgotten Origins of the Moderns Humanities. Princeton: Princeton University Press, 2014.
Weber, Max. âWissenschaft als Beruf.â In Schriften: 1894â1922. Stuttgart: Alfred Kröner Verlag, 2002 [1918].
Wilcken, Patrick. Claude Levi-Strauss: The Father of Modern Anthropology. London: Penguin Books, 2010.
Willkomm, Jens, Christoph Schmidt-Petre, Martin Schäler, and Michael Schefczyk. âUsing Ngrams to Develop a Query Algebra for Conceptual History.â Digital Humanities Conference Proceedings (2019), https://staticweb.hum.uu.nl/dh2019/dh2019.adho.org/index.html.
Woodhead, Linda. Reinventing Christianity. Nineteenth-century Contexts. London: Routledge, 2001.
Zhang, Yang. âEntropic Evolution of Lexical Richness of Homogeneous Texts over Time: A Dynamic Complexity Perspective.â Journal of Language Modelling 3, no. 2 (2015): 569â599.
Ãlender, Maurice. The Language of Paradise. Cambridge. MA: Harvard University Press, 1992.
Cf. e.g. Johanne Louise Christiansen and Katrine Boserup Jensen, âDa Koranen blev en app ⦠En genforhandling af Koranens status som helligt objekt.â Religionsvidenskabeligt Tidsskrift, nr. 70 (2020): 112â129.
Cf. e.g. Markus Davidsen, The Spiritual Tolkien Milieu: A Study of Fiction-based Religion, PhD thesis (University of Leiden, 2014).
Digital Humanities came into general usage in the first decade of the 2000s. Terminologically it surpassed âHumanities Computingâ. In introductions to and readers on Digital Humanities, such as for instance A New Companion to Digital Humanities edited by Susan Schreibman, Ray Siemens and John Unsworth (London: Blackwell, 2016), this change is deemed necessary in order to âshift the emphasis from âcomputingâ to âhumanitiesâ.â This is necessary particularly if one wishes to include into the field studies âof digital technologies, their creative possibilities, and their social impactâ (Schreibman et al., A New Companion, xvii). I do not, however, find this broadening of the field helpful at a practical-didactic level.
The human conditionâ is not an unchallenged scope. In fact, the anthropocentrism of the Humanities seems to be under growing pressure. Philosophers are leading the way in the post- and transhumanistic field and anthropologists foresee an âontological turnâ (cf. Martin Holbraad and Morten A. Pedersen, The Ontological Turn: An Anthropological Exposition (Oxford: Oxford University Press, 2017)) and engage themselves in ânon-humanâ and ârewildingâ perspectives (cf. Philippe Descola, Beyond Nature and Culture (Chicago: The University of Chicago Press, 2013. English translation of Par-delà nature et culture in 2005)). Outside academia one finds popular versions of the latter trend. Veterinarian and natural historian Charles A. Foster who has conducted âfieldworkâ as a badger, an otter, a fox, a deer etc., described in his monography Being a Beast (2016), is a bestselling example.
This âsainthoodâ is not a banal, cosmetic analogy. Reverence and quasi-ritual structures have been built around Grundtvig as a Great Dead (cf. Robert Bartlett, Why Can the Dead Do Such Great Things? Saints and Worshippers from the Martyrs to the Reformation (Princeton: Princeton University Press, 2013)). Grundtvig has a cathedral named after him: the Copenhagen Grundtvigs Kirke; every year his birthday (almost coinciding with the day of his death) is celebrated in Grundtvig-relevant institutions; one such celebration entails the opening of his crypt at the small cemetery Claraâs KirkegÃ¥rd on the outskirts of the Sealandic town of Køge. Moreover, Grundtvigâs Death (Grundtvigs død) is a commodity â at least it is a recent title in a popular book series by Aarhus University Press written by Grundtvig scholar Jes Fabricius Møller (2019).
Cf. Benedict Anderson, Imagined Communities. Reflections on the Origin and Spread of Nationalism (London: Verso, 1983).
As such this is a typical so-called first-generation Digital Humanities project, digitally preserving a canonic element of a given cultural heritage. For an introduction to the theme of digital preservation, see William Kilbride âSaving the Bits: Digital Humanities Forever?â In A New Companion to Digital Humanities, edited by Susan Schreibman, Ray Siemens and John Unsworth (London: Blackwell, 2016), 408â419.
In comparison, Grundtvig beats Charles Dickens. A small win, but a win: 3 859 231 is the total amount of word tokens in Dickensâ writings. A level significantly above other canonic writers, 835 997 is the total amount in William Shakespeareâs plays; Jane Austenâs novels contain 744 318 word tokens.
The data are available at: https://github.com/centre-for-humanities-computing/grundtvig-data. Furthermore, a custom XML parser to facilitate third-party data exploration is available at: https://github.com/centre-for-humanities-computing/GrundtvigParser.
Cf. Elena Pierazzo, Digital Scholarship in the Humanities, Volume 31, Issue 3, September (2016): 513â516.
A main problem for historical texts predating the late nineteenth century is that of typography. For European material âFrakturâ and other types of âgothicâ typesetting pose significant challenges to the OCR âtranslationsâ. For non-European languages and for non-Latin alphabets the situation is even more unsettled and the material reliant on human curation.
The Great Unread vs. The Canon has long been a topic in Literary History (Margaret Cohen, The Sentimental Education of the Novel (Princeton: Princeton University Press, 1999), Franco Moretti, âThe Slaughterhouse of Literature.â Distant Reading (London: Verso, 2013)).
Cf. Péter Hadju, âDigital editions of canonic texts: do such projects affect the canon?â Texto Digital 6 (2010): 85â95.
A concordance is a cross-indexing âtoolâ â that is: an alphabetical list of the keywords of a given book or a body of works, listing every occurrence of each keyword in its immediate semantic context.
A dedication to semiotic patterns, in fact, made cultural anthropologist and father of structuralism Claude Lévi-Strauss long for a computational-like machinery to aid him in his categorizations of Native American myths. In his mindâs eye he saw a machine: âConsisting of a series of two-metre-long, and one-and-a-half-metre-high upright boards, on which cards containing mythic elements could be âpigeon holed and moved at willâ. As the analysis moved into three dimensions, the cards would need to be perforated and fed through IMB equipment. The whole operation would require a substantial atelier, along with a team of dedicated technicians working to divine âthe generic law of the mythââ (Patrick Wilcken, Claude Levi-Strauss: The Father of Modern Anthropology (London, 2010), 265). Now, such a machine sits on my desk.
The procedure is based on so-called âunsupervised machine learningâ. As indicated by the name, Word2Vec-analyses map the multidimensional vector-space by which a given term is nested within a given corpus; this space is annotated as a cosine similarity (cf. Kishore Ayyadevara, Pro Machine Learning Algorithms: A Hands-On Approach to Implementing Algorithms in Python and R (Apress, 2018), 167â178).
Katrine F. Baunvig, âFictional Realities of Modernity: The Fantastic Life of the Demi-Goddess Dana in the Emerging Nation State of Denmark.â In Mythology and Nation Building in the Nineteenth Century: N.F.S. Grundtvig and His European Contemporaries, edited by Lone Kølle Martinsen, Sophie Bønding, and Pierre-Briece Stahl (Aarhus: Aarhus University Press, 2021), 97â134.
Katrine F. Baunvig, Oliver Jarvis, and Kristoffer L. Nielbo, âEmotional Imprints: Letter-Spacings in N.F.S. Grundtvigâs Writings.â Digital Humanities in the Nordic Countries Proceeding (2020): 192â202.
Katrine F. Baunvig, Oliver Jarvis, and Kristoffer L. Nielbo, âEmotional Imprints: Exclamation Marks in N.F.S. Grundtvigâs Writings.â Digital Humanities in the Nordic Countries Post Proceeding (2021): 156â169.
Katrine F. Baunvig and Kristoffer L. Nielbo, âKan man validere et selvopgør?â In Textkritik som analysmetod, edited by Paula Henrikson, Mats Malm, and Petra Söderlund (Stockholm: Svenske Vitterhetssamfund, 2017): 45â67.
C.E. Shannon, âA mathematical theory of communication.â The Bell System Technical Journal 27 (1948): 379â423, 623â656.
Philippe Thoiron, âDiversity Index and Entropy as Measures of Lexical Richness.â Computers and the Humanities 20:3 (1986): 197â202, Yang Zhang, âEntropic Evolution of Lexical Richness of Homogeneous Texts over Time: A Dynamic Complexity Perspective.â Journal of Language Modelling 3:2 (2015): 569â599.
Baunvig, âFictional Realities of Modernityâ (2021).
http://www2.statsbiblioteket.dk/mediestream/avis [accessed 29 January 2021].
A list of the 4517 individual Danas furnished with weblinks as well as the table containing our coding is available as a data repository online (Katrine Frøkjær Baunvig and Stine Kylsø Pedersen, âThe Dana Repository.â In The Grundtvig eStudies Pamphlets #1 (2021)).
Though EU copyright regulations protect intellectual property until 70 years after the death of a given author, in the Danish Mediestream case a uniform solution for the material has been reached and the line is (for the time being) drawn at 1921: cf. https://www2.statsbiblioteket.dk/mediestream/info/2.
Cf. Haimin Lee, â15 Years of Google Booksâ (2019). https://www.blog.google/products/search/15-years-google-books/.
cf. Jens Willkomm, Christoph Schmidt-Petre, Martin Schäler and Michael Schefczyk, âUsing Ngrams to Develop a Query Algebra for Conceptual History.â Digital Humanities ConferenceâProceedingsâ(2019),âhttps://staticweb.hum.uu.nl/dh2019/dh2019.adho.org/index.html.
Dawn E. Holmes, Big Data: A Very Short Introduction (Oxford: Oxford University Press, 2017), xv.
The explosive data growth and the development within the area of machine learning has turned futuristic, technological enthusiasm into a tangible, cultural phenomenon in itself. The public interest in the (controversial) business incubator and consultant service âSingularity Universityâ is one among many examples hereof. In their own words, they have âa massive transformative purposeâ and their mission is to prepare corporations and the general public for âexponential opportunitiesâ (cf. https://su.org/about/). Others speak with joy of the so-called âinternet of thingsâ (cf. e.g. Finn Arne Jørgensen, âThe Internet of Things,â in A New Companion to Digital Humanities, edited by Susan Schreibman, Ray Siemens, and John Unsworth (London: Blackwell, 2016), 42â53). But overall dystopic reactions to the data and technological developments are visible as well. The Home Box Office series Westworld (2016â) is just one popular instantiation of a general caution and sceptical attitude towards developments within artificial intelligence.
Cf. Niels Brügger, âDigital Humanities in the 21st Century: Digital Material as a Driving Force.â Digital Humanities Quarterly Vol. 10, No. 2 (2016) http://www.digitalhumanities.org/dhq/vol/10/3/000256/000256.html.
Cf. Greta Franzini, Melissa Terras, and Simon Mahony. âA Catalogue of Digital Editions.â In Digital Scholarly Editing. Theories and Practices, edited by Matthew James Driscoll and Elena Pierazzo (OpenBook Publishers, 2016), 178.
A long line of projects similar to the Danish Mediestream have sprung to life. A brief list includes: ZEFYS Zeitungsinformationssystem (zefys.taatsbibliothek-berlin.dk); TROVE, National Library of Austria (trove.nla.gov.au); The British Newspaper Archive (britishnewspaperarchive.co.uk); Chronicling America. Historic American Newspapers (chroniclingamerica.loc.gov/newspapers/); Delpher Dutch digital library (delpher.nl); Gallica archive of French newspapers (allica.bnf.fr/); The digital library of Danish American newspapers and journals (box2.nmtvault.com/DanishIM/jsp/RcWebSearchResults.jsp). See also: https://en.wikipedia.org/wiki/Wikipedia: List_of_online_newspaper_archives covering 90 countries.
At the same time, growth of data does, however, seem to call for systematic reflection and sustainable, infrastructural solutions to maintenance. Data âpreservation has recently attracted new attention within research: questions of sharing data and reproducibility of science, open access and maintenance have become more and more pertinent as the websites and digital productions come of age. â404 not foundâ messages now replace a growing number of hyperlinks, and the Web as scientific platform is full of digital wastelands, caused by the end of research projectsâ (Christine Barats, Valérie Schafer, and Andreas Fickers, âFading Away ⦠The challenge of sustainability in digital studies.â Digital Humanities Quarterly Vol. 14 No. 3 (2020): 165â189).
Cf. Linda Woodhead, Reinventing Christianity. Nineteenth-century contexts (London: Routledge, 2001).
Cf. e.g. Max Weber, âWissenschaft als Beruf.â Schriften: 1894â1922 (Stuttgart: Alfred Kröner Verlag, 2002 [1918]), 510.
Pinpointing the relevant keywords for such a procedure would be the obvious first task. God, angel, spirit, providence, blessing, damnation, prayer, contemplation, heaven, hell, afterlife etc. could be a starting point.
James Turner, The Forgotten Origins of the Moderns Humanities (Princeton University Press, 2014): x.
Turner, The Forgotten Origins, x.
Cf. Delia Langa Rosado and Miriam E. David, ââA massive university or a university for the masses?â Continuity and change in higher education in Spain and England.â Journal of Education Policy Vol. 21 No. 3 (2006): 343â365.
E.g. Maurice Ãlender, The Language of Paradise (Cambridge. MA: Harvard University Press, 1992), Edward W. Said, Orientalism (New York: Pantheon, 1978).
Turner, The Forgotten Origins, 370.
Cf. Hans Ulrich Gumbrecht, The Powers of Philology: Dynamics of Textual Scholarship (Washington: Library of Congress Cataloging-in-Publication Data, 2003), 2.
Turner, The Forgotten Origins, 380.
Turner, The Forgotten Origins, 380.