=Paper=
{{Paper
|id=Vol-2365/07-TwinTalks-DHN2019_paper_7
|storemode=property
|title=Interdisciplinary Collaboration in Studying
Newspaper Materiality
|pdfUrl=https://ceur-ws.org/Vol-2365/07-TwinTalks-DHN2019_paper_7.pdf
|volume=Vol-2365
|authors=Eetu Mäkelä,Mikko Tolonen,Jani Marjanen,Antti Kanner,Ville Vaara,Leo Lahti
|dblpUrl=https://dblp.org/rec/conf/dhn/MakelaTMKVL19
}}
==Interdisciplinary Collaboration in Studying
Newspaper Materiality==
https://ceur-ws.org/Vol-2365/07-TwinTalks-DHN2019_paper_7.pdf
Interdisciplinary Collaboration in Studying
Newspaper Materiality
Eetu Mäkelä1[0000−0002−8366−8414] , Mikko Tolonen1[0000−0003−2892−8911] ,
Jani Marjanen1[0000−0002−3085−4862] , Antti Kanner1[0000−0002−0782−1923] , Ville
Vaara1[0000−0001−7924−4355] , and Leo Lahti2[0000−0001−5537−637X]
1
Department of Digital Humanities
University of Helsinki, Finland
first.last@helsinki.fi
2
Department of Mathematics and Statistics
University of Turku, Finland
first.last@utu.fi
Abstract. This paper presents a collaboration between computer scien-
tists, linguists and historians studying the material aspects of newspa-
pers and developing a tool for that purpose. The paper describes how the
back-and-forth collaboration in terms of research questions and technical
challenges yielded insights both for solving computational problems as
well as refining historical analysis. In the project, existing metadata was
amended by reconstructing new materiality data from the Finnish digi-
tised newspaper corpora. The analysis of such data is crucial for studying
the development of newspapers, but can also inform other computational
studies on the same data. The use of enriched materiality data allows
for better understanding subdivisions in large corpora such as digitised
newspapers, but also highlight that content and form interact. Content
analysis of newspapers should therefore always take into account ma-
terial properties of the studied material to properly grasp the cultural,
social and political meanings embedded in the sources.
Keywords: Materiality of newspapers · Collaboration · Digital hu-
manities.
1 Introduction
This paper offers a view to the collaboration undertaken at the Helsinki Com-
putational History Group (COMHIS)3 between computer scientists, historians
and linguists on a project that studies the material dimensions of newspapers
and their development [3].
The present day transformation from print to digital is not the first time
newspapers have evolved drastically. Instead, this change of format reminds of
similar transformations when the newspaper first appeared as a distinct material
genre. One influential definition separating a newspaper from a newsbook or
3
http://helsinki.fi/computational-history
56 E. Mäkelä et al.
pamphlet in its early days was that a newspaper was a ”sheet of two or four
pages, made up in two or more columns” [10]. The Dutch had two-column news
at the time, while civil war in Britain saw both the rebels and the crown printing
their propaganda. It took, nevertheless, centuries before journalism became a
profession of its own and newspapers took their particular shape in the mid-
nineteenth century [20,1,2,11,13,23].
In the context of digital humanities, newspapers have become an iconic ex-
ample of “big data” research (cf. [5,15,7], https://numapresse.org/). While in
localised research [8,28] the material can be thought uniform, in the big data
approaches it is striking how little attention is paid to what the data consists of.
A telling example of waking up to this is the Oceanic Exchanges project (https:
//osf.io/wa94s/) where M.H. Beals and Ryan Cordell quickly concluded that
mapping metadata across its many datasets is to be one of its most important
contributions (https://twitter.com/ryancordell/status/1001845719341285377).
Framed against this background, the idea of this paper is to outline how we
developed a tool to uncover and explore the varied materiality of newspapers.
As part of the large-scale digitisation, the accessibility of historical newspapers
has improved drastically, but at the same time much of the information about
the size, shape and feel of the newspapers, that was so central to past readers in
understanding what kind of documents they were perusing, has to a large extent
been hidden from view. Interestingly, the digitised versions of the newspapers
also allow for large-scale study of their material dimensions – an opportunity
that has so far been paid very little attention to. In our case, our focus on
materiality is also just one aspect of the group’s larger interest in studying the
nature of early modern public discourse through the analysis of structured and
unstructured data relating to newspapers and other printed materials.
In what follows, we will first briefly explain the background for this study and
how it fits the group’s publication history. Then, we’ll shortly discuss the type
of data we started our work from, before going into detail on how the research
process that led to the materiality explorer tool actually happened. Finally, we
will describe the tool itself and the tentative results we’ve obtained using it,
before concluding by outlining directions for future work.
2 Studying the Materiality of Newspapers
The first time that data on the materiality of newspapers was extracted and
studied by us at the COMHIS group was as part of the Helsinki Digital Human-
ities Hackathon of 20154 . After that, intermittent analyses on both the content
as well as metadata such as language, location and form of the newspapers was
done as part of the internal dialogue of the research group, in part in the context
of the Academy of Finland funded project on ”Computational History and the
Transformation of Public Discourse in Finland, 1640-1910”5 .
4
http://heldig.fi/dhh15
5
http://www.aka.fi/globalassets/32akatemiaohjelmat/digihum/hanke-esitteet/
salmi-digihum.pdf
Interdisciplinary Collaboration and Newspaper Materiality 57
Slowly, these explorations coalesced into multiple conference presentations on
the subject. Mostly, the actual work happened in sporadic bursts, often with one
of the more computationally oriented researchers in the group being inspired to
run a particular analysis, which then led to back-and-forth exchange between the
historians and the experts in quantitative methods to better interpret and fine-
tune the analysis. In this process analyses were also designed to be more aligned
with research questions pertinent to newspaper history, and new analyses were
requested by the historians.
In time, these explorations led to more focused research questions, dealing
with the modernisation of newspapers in Finland in two main languages. As
newspapers became more frequent, more topical and gained a larger format, they
started resembling the modern newspapers that we encounter today (or perhaps
those of our childhood). In particular, we wanted to trace the asynchronicity
that was present between Finnish-language and Swedish-language papers. Edi-
tors and other intellectuals in Finland operated mostly in both languages, and
thus the newspapers were developed in constant cross-fertilisation across the lan-
guage border, but still the different language spheres developed at different paces.
While Swedish-language papers were generally more advanced up to the 1860s
and 1870s, Finnish-language papers became leading by the turn of the century
1900 due to growth both in terms of readership and places of publication.
A problem with our early explorations was that they had been done in a
haphazard, off-the-cuff manner by different people using different versions of
the data, so they were not mutually consistent and reliable. An impetus to
change this came when one of the conference presentations led to an invitation
to write up the work more formally for the Journal of European Periodical
Studies (JEPS). At this point, it was decided to take one single version of the
data as the source, and calculate all material and linguistic indicators from that.
A more thorough analysis of the trustworthiness of the pipeline and the dataset
itself was also undertaken.
For the JEPS article, the figures and analyses used to inform the content
started as those that had arisen organically as part of the internal dialogue
within the group. However, when polishing the art, a dialogue was held between
the historians and the statistical visualisation experts on what the core message
was. This led to replacing earlier more explorative versions of the visualisations
with ones designed specifically to convey particular arguments. At the same time,
the visual outlook of all graphs was unified.
After working on the JEPS article, the group had a relatively good notion
on what the important aspects of materiality in the data were, and how they
could best be visualised and explored in a unified manner. This led way to
the development of an interactive materiality explorer. Through this, there was
more freedom for the content experts to explore the phenomenon, with much less
frequent need for the computer scientists to run customised analyses or change
the parameters of the exploration.
58 E. Mäkelä et al.
2.1 Extracting and Deriving Material Aspects from ALTO XML
In order to understand what the group was working with, it is relevant to un-
derstand the usefulness of ALTO (Analyzed Layout and Text Object, https:
//www.loc.gov/standards/alto/) files that were luckily available for the project.
ALTO files contain a description of the visual organisation of content on a page,
at the core of which are the individual words and their page coordinates. At the
same time, the words are also grouped into blocks, often corresponding to para-
graphs or columns. The format also contains general layout information, such
as the sizes of margins and main printed area.
The usefulness of ALTO for analysing materiality crucially depends on the
choice of the measurement unit in which all coordinate and size information is
given. Here, the format gives a choice from three options: mm10 (tenth of a
millimeter, the default value), inch1200 (1200th of an inch) or pixel. Of these,
the first two directly relate all measurements to actual physical dimensions,
while the pixel coordinates do not. However, even then, the information on
original physical dimensions can be recovered if the DPI value of the image
is known, information given in the METS metadata files originally often ac-
companying the ALTOs. Unfortunately, many collections such as the Dutch
Delpher (https://delpher.nl/) and French Gallica (https://gallica.bnf.fr/) pro-
vide their ALTO data specifically using pixel coordinates, while not giving out
the METS files (which would also contain logical segmentation information, sep-
arating the text into articles and adverts). Similarly, the National Library of
Finland (http://digi.kansalliskirjasto.fi/), while providing the METS files, ex-
plicitly removed scanning information from them until requested otherwise.
These examples highlight how little thought is given to the material dimen-
sion of the newspapers in most digital processing pipelines even before the user
interface layer. Luckily, the ALTO files of the National Library of Finland had a
MeasurementUnit of mm10. Given this, we could easily extract page size, printed
area and character and words counts for each page. Besides these, the ALTO
file also contains some style information that can be extracted. Currently, we
disregard the information on left/center/right alignment, but do extract font
information. Directly given are the size, face, style (bold/italic/underline) of
each font used, to which we add the calculated number of characters and words
written using that font, as well as the overall page area covered.
For each page, we also extract all text box coordinates (visualised in Figure
1). While these are primarily meant to locate text visually on the page in reader
interfaces, they can be processed to yield layout information. First, we extract
column counts using a lighter-weight process than the computer vision approach
used in [6]. We scan the page from top to bottom, for each Y coordinate counting
the number of text boxes present there. This yields a distribution associating
all column counts with the area they control on the page. Mapping shifts in the
amount of columns seems to be one of the clearer indicators of changes in layout.
This is useful both for assessing the general development of newspaper layout,
but also for identifying particular instances in which editors felt they needed to
introduce changes to the layout. Columns obviously roughly correspond to page
Interdisciplinary Collaboration and Newspaper Materiality 59
Fig. 1. ALTO text blocks overlaid on a newspaper page.
60 E. Mäkelä et al.
size, but changes in the width of columns are also indicative of how newspapers
explored issues of readability.
2.2 Developing the Materiality Explorer
The Helsinki Computational History Group sits along the same corridor at the
University of Helsinki. This physical presence is an important part of the group’s
work, but so is Slack. As a tool, Slack is an effective way of communicating while
sharing research ideas and findings, but it also has the benefit of functioning
as a means of documenting much of the group’s efforts. To provide an example
of this, we will present shortly below an analysis of our Slack communication
relating to developing the materiality explorer.
On this particular project, the intensive work started – according to the
comments on Slack – on 30 October 2018. It began when Eetu Mäkelä posted
first images of a general visualisation unifying multiple aspects of materiality
data. From the beginning, it was clear that the point of the materiality explorer
was to experiment with different ways to define gross materiality categories in
newspapers. It took however few days before the work on the development got
going seriously.
Nevertheless, by 12 December 2018, there were altogether 355 different mes-
sages (8-9 on average / day) on the group’s slack channel dedicated to newspapers
about this work. Altogether 9 people participated in this online discussion with
different kinds of input. While some people just posted one or a few notes, two
group members had more than 100 messages each devoted to this project. There
was also, of course, actual human interaction in real life, which is unfortunately
not recorded. What drove the work was a looming deadline for the DH2019
conference at the end of November.
Analyses undertaken on development versions of the materiality explorer soon
led us to realise that some of our data was problematic. Here, an important point
to notice is that computational processing of the data did not start with us, but
included also the scanning and OCR of the pages, as well as the metadata work
done on the collection at the National Library of Finland. What we found out
was that the National Library of Finland had used altogether 22(!) versions
of scanning software. A key problem for us was that some of these did not
differentiate between Fraktur and Antiqua fonts. By using metadata to analyse
which newspapers were scanned with which version, we determined that reliable
font identification could only be had up to the year 1910. We also employed some
spot checking to compare algorithmic results to the manually keyed metadata,
and for example decided to use the raw data directly for page size and date range
estimation instead of the same information as keyed.
After a few days of pondering about the effects of these technical problems
for analysis, we started focusing more on the question of cramming information
on one sheet of newspaper – thinking also about the readability of the text on the
page. At the same time, a more extensive reading of relevant secondary sources
begun to figure out the technological development (especially with the DH2019
conference submission in mind). The reason for doing this was to find possible
Interdisciplinary Collaboration and Newspaper Materiality 61
identifiable markers to flag differences and effects caused by changes in printing
techniques. For example, the emergence of lithography offset printing was one
such technique whose effects we could clearly identify also in the data.
We also soon advanced to thinking about layout and the relevance of the
front-page. The idea was to figure out ways of detecting typographic changes
on the front page within the context of a single newspaper to understand its
development. At this time, it came as an idea to try to identify an instance of a
(statistically) typical front page for each decade over time for both Finnish and
Swedish language newspapers. Once we knew that this is possible based on the
tools at hand, several different kinds of experiments to find “typical” newspaper
proportions using the materiality explorer were made. Our deliberations partic-
ularly echoed those by Myllyntaus [21], who has done a huge amount of work
on these issues without the statistical apparatus that we have on hand today.
What was visible in our data was that importing the rotary press and offsetting
technology to Finland changed the newspaper layout in the papers that could
afford this technology in a very short period of time. We were able also to see
that the linguistic and geographic diversity in Finland led to a situation where
print runs were smaller and there was more type-setting ongoing than in some
larger European countries.
We realised also that we could group different language newspaper published
by the same publisher in the same year at the same location together in order to
study their layout and content. This would help us to understand how news possi-
bly circulated from one language to another and how different advertisements for
example are presented in different languages in Finland. Many previous scholars
have been interested about different language profiles of newspapers in different
Finnish towns. What these scholars haven’t realised is that the question of type,
layout etc. can also have intellectual relevance. So, to ask if parallel newspapers
are coming from the same publishing house (as they at times do) is a relevant
question to ask.
On Sunday 25th of November, Eetu Mäkelä posted an image of the mean
front page of Helsingin Sanomat in 1907. This also marked the saturation point
of the development phase of this part of the work. There were still new ideas
coming in, for example, about terseness of language in newspapers in order to
allow cramming, but the main thing for us at this point was to prepare for the
DH2019 deadline that was on 27th of November. Perhaps we need to wait for
the next deadline to get back seriously to this project.
2.3 The Materiality Explorer Interface
As it currently stands, the materiality explorer has three main functionalities,
each aimed at a different use cases. Common to all views are a set of selectors,
allowing to limit the set of newspapers under study. Currently, these hold facili-
ties for limiting study by 1) time, 2) newspaper language, 3) newspaper lifetime,
4) printing location and 5) individually by title.
In the overview view shown in Figure 2, first presented is the absolute amount
of data. This is important, as all the other graphs display their information as
62 E. Mäkelä et al.
Fig. 2. The materiality explorer.
Interdisciplinary Collaboration and Newspaper Materiality 63
proportions of the whole. Depending on a user selected option, this proportion
may be calculated by year, by month or by week. In addition, the user can select
whether they want an observation to be titles, issues or pages. Here, the choice
depends on what one is interested in. Counting by titles treats each newspaper as
a single unit, allowing exploration of the breadth of newspapers without regard
to how often they appeared or how large they were. On the other hand, if one
is more interested in the amount of information consumed by an end reader,
then possibly counting by issue or even by page is appropriate. Another use case
where observing by page or issue may be more interesting is when studying the
development of a single newspaper, where the differing publication frequencies
and page sizes no longer matter, but instead even singular aberrant pages are
interesting.
After this absolute count, a baseline measure of text per month is given,
against which all the materiality information can be contrasted. This baseline
was developed in consultation between the computer scientists, historians and
linguists to provide a language-neutral measure for throughput. By counting the
number of characters each newspaper produces in a month without regard to how
they are divided between issues or pages, this measure shows how much content
needs to be transmitted. As this quantity rises, newspapers must respond with
material innovations, whether by increasing page count, page size or publication
frequency, or by cramming more material into available space by decreasing font
size, line breaks or margins.
A second view allows grouping the data by a combination of material dimen-
sions, thereby allowing exploration of archetypal materiality categories. Finally,
two distinct views allow the user to explore respectively page and issue-level
material anomalies in the data: for example pages which have much more text
than others or pages with abnormal layout, or issues with appendixes or which
appear on the same day as another issue. These both lead the way for interesting
qualitative analyses, but can also be used to remove abnormal data from further
quantitative computational analyses of either form or content. In our project,
the anomaly detection served as a central method for exploring the data as well
as identifying errors in the code or metadata. Here historians had a rich source
for detecting counter-intuitive findings, and often those findings led to feedback
that could further improve coding efforts.
At present, we are using the interface to exploratively develop hypotheses
on common development patterns as well as archetypal materiality categories.
Both of these are interesting in themselves as objects of study, but can also be
used later to partition datasets for other computational processing such as OCR
retraining or content analysis. While this current stage of explorative hypothesis
development is interactive, visual and qualitative, our plan for the next stage is
to explore statistical validation of such hypotheses using for example Granger
causality and archetypal coverage measures. Once developed and tested, these
again will be added to the interface to enable further self-sufficient analysis in a
more trustworthy manner.
64 E. Mäkelä et al.
3 Discussion
At the outset of this project, we asked in particular how the modernisation of
newspapers published in Finland could be better understood by looking at the
form, shape, location and publication frequency in newspapers published in dif-
ferent languages (Finnish and Swedish being the main publication languages).
The project produced one article that pays particular attention to the different
speed in development with regard to Swedish-language and Finnish-language
newspapers in Finland. Further, we produced an interactive materiality explorer
that helps researchers understand the development of material aspects of news-
papers. We also developed preliminary hypotheses that will be shortly discussed
below with regard to different categories of materiality.
For the Finnish newspapers, the data shows a general order in how they
expanded: first, layout was changed to include more words per page; second,
page size was increased; third, publication frequency was increased and only
after that was the amount of pages increased. This last step often coincides with
the introduction of rotary presses, which allowed newspapers to more easily be
composed of more than four pages, and also allowed them to move back from
large page sizes to more easily handled formats. Simultaneously, the data shows
also high variability, where papers not only frequently printed supplements, but
could switch back and forth between formats inside a single week, or cram text
into a special issue through diminished line breaks. Similar shifts took place also
with regard to fonts. Newspapers explored different Fraktur and Antiqua fonts to
try out readability, but also because fonts were oftentimes used to signal that the
contents was aimed for a particular audience. While there are plenty of exceptions
to this, it seems that Fraktur was more often used when dealing with economy
and religion, whereas Antiqua was reserved to politics, philosophy and the high
arts. To test such hypotheses about different uses for fonts and relating that
to the overall development of newspapers, we still need more robust statistical
information. We also aim to compare used fonts and with other factors, such
as language frequency and size of newspapers. (For the history of newspaper
layout and design, see [4,17,22,24,26,27,12].) Compared to earlier studies, our
data driven approach gives us a great opportunity to evaluate the main findings
of earlier historical studies of newspaper materiality [18,30,21,9,16,32].
What we also aim to do with these patterns is to develop evidence-based
archetypal categories of newspapers across history. We are then able to trace
and compare these through time and place, but also use them to study the
evolution of individual newspapers. These categories will also help us understand
the newspapers as objects of intellectual activity, creating a theory of different
historical maturity levels of newspapers. This in turn will help us chart the
development of public discourse over time.
Besides presenting the research process regarding the material development
of newspapers as a genre in itself, we argue that content and form interact, and
thus big data approaches to newspaper analyses also need to pay attention to
material differences in order to accurately understand the subdivisions in large
corpora. Here, this paper continues on a path previously charted by for example
Interdisciplinary Collaboration and Newspaper Materiality 65
[19,29,14], while providing an orthogonal axis to those expanding study from
text to visual elements [25,31]. For example, using the metadata we can create
meaningful subsets of the data that are balanced by paper type for for example
topic modelling or teaching automated transcription algorithms.
Here, Finland makes an intriguing case for digital history because its public
sphere is bilingual, with newspapers in both Swedish and Finnish. One inter-
esting phenomena that arises from this are publishers publishing newspapers in
both languages. For example, in Kotka there are both Finnish and Swedish news-
papers by the same publisher with identical layouts and advertisements. Such
could be used to create parallel corpora, interesting for the study of common-
alities and differences between the different language public spheres, but also
perhaps as material for machine translation.
References
1. Allen, J.E.: The Modern Newspaper: its typography and methods of news pre-
sentation. [With illustrations. New York & London, Pp. ix. 234. Harper & Bros.
(1940)
2. Baldasty, G.J.: Commercialization of News in the Nineteenth Century. University
of Wisconsin Press, Madison (2014)
3. Bode, K.: The Equivalence of “Close” and “Distant” Reading; or, Toward a New
Object for Data-Rich Literary History. Modern Language Quarterly 78(1), 77–106
(Mar 2017). https://doi.org/10.1215/00267929-3699787, https://read.dukeupress.
edu/modern-language-quarterly/article/78/1/77-106/19924
4. Broersma, M. (ed.): Form and style in journalism: European newspapers and the
representation of news 1880-2005. Peeters, Leuven, Dudley, MA (2007)
5. Buntinx, V., Bornet, C., Kaplan, F.: Studying Linguistic Changes over 200 Years
of Newspapers through Resilient Words Analysis. Frontiers in Digital Humanities
4 (2017). https://doi.org/10.3389/fdigh.2017.00002
6. Buntinx, V., Kaplan, F., Xanthos, A.: Layout analysis on newspaper archives. In:
DH2017 abstracts (2017), https://dh2017.adho.org/abstracts/193/193.pdf
7. Cordell, R., Smith, D.: What News is New?: Ads, Extras, and Viral Texts on the
Nineteenth-Century Newspaper Page. In: DH2017 abstracts (2017)
8. Cristianini, N., Lansdall-Welfare, T., Dato, G.: Large-scale content analysis of
historical newspapers in the town of Gorizia 1873–1914. Historical Methods: A
Journal of Quantitative and Interdisciplinary History 51(3), 139–164 (Jul 2018).
https://doi.org/10.1080/01615440.2018.1443862
9. Gustafsson, K.E., Rydén, P.: A History of the Press in Sweden. NORDICOM,
Göteborg (2010)
10. Hutt, A.: The changing newspaper; typographic trends in Britain and America
1622-1972. Gordon Fraser, London (1973)
11. Høyer, S., Pöttker, H.: Diffusion of the news paradigm 1850-2000 (2014)
12. Kapr, A., Forssman, F., Willberg, H.P.: Fraktur: Form und Geschichte der gebroch-
enen Schriften. Hermann Schmidt, Mainz (1993)
13. Kutsch, A.: Journalismus als Profession: Überlegungen zum Beginn des journalis-
tischen Professionalisierungsprozesses in Deutschland am Anfang des 20. Jahrhun-
derts. In: Blume, A., Böning, H. (eds.) Presse und Geschichte: Leistungen und
Perspektiven der historischen Presseforschung, pp. 289–325 (2008)
66 E. Mäkelä et al.
14. Lahti, L., Marjanen, J., Roivainen, H., Tolonen, M.: Bibliographic Data Science
and the History of the Book (c. 1500–1800). Cataloging & Classification Quarterly
0(0), 1–19 (Jan 2019). https://doi.org/10.1080/01639374.2018.1543747
15. Lansdall-Welfare, T., Sudhahar, S., Thompson, J., Lewis, J., FindMyPast News-
paper Team, Cristianini, N.: Content analysis of 150 years of British periodicals.
Proceedings of the National Academy of Sciences 114(4), E457–E465 (Jan 2017).
https://doi.org/10.1073/pnas.1606380114
16. McReynolds, L.: The News under Russia’s Old Regime: The Devel-
opment of a Mass-Circulation Press. Princeton University Press (1991).
https://doi.org/10.2307/j.ctt7zth51
17. Moen, D.R.: Newspaper layout and design. Iowa State University Press, Ames
(1989)
18. Moran, J.: Printing Presses: History and development from the Fifteenth Century
to Modern Times. University of California Press (1973)
19. Moreux, J.P.: Innovative Approaches of Historical Newspapers: Data Mining, Data
Visualization, Semantic Enrichment (Aug 2016), https://hal-bnf.archives-ouvertes.
fr/hal-01389455/document
20. Morison, S.: The English Newspaper, 1622-1932: An Account of the Physical De-
velopment of Journals Printed in London. Cambridge University Press, 1 edition
edn. (Oct 2009)
21. Myllyntaus, T.: Suomen graafisen teollisuuden kasvu 1860-1905. University of
Helsinki, Helsinki (1981)
22. Olson, K.E.: Typography and Mechanics of the Newspaper. D. Appleton and Com-
pany (1940)
23. Pettegree, A.: The Invention of News: How the World Came to Know About Itself.
Yale University Press (2014)
24. Presbrey, F.: The history and development of advertising. Doubleday, Garden City,
N.Y. (1929)
25. Smits, T.: Illustrations to Photographs: Using computer vision to analyse news
pictures in Dutch newspapers, 1860-1940. In: DH2017 abstracts (2017)
26. Sutton, A.A.: Design and makeup of the newspaper. Prentice-Hall (1948)
27. Swanson, G.: Graphic Design & Reading: Explorations of an Uneasy Relationship.
Allworth Press (2000)
28. Tilles, D.: The Use of Quantitative Analysis of Digitised Newspapers to Challenge
Established Historical Narratives. Roczniki Kulturoznawcze 7(1), 83–97 (2016).
https://doi.org/10.18290/rkult.2016.7.1-4
29. Tolonen, M., Lahti, L., Roivainen, H., Marjanen, J.: A Quantitative Approach
to Book-Printing in Sweden and Finland, 1640–1828. Historical Methods: A
Journal of Quantitative and Interdisciplinary History 0(0), 1–22 (Dec 2018).
https://doi.org/10.1080/01615440.2018.1526657
30. Tommila, P., Salokangas, R.: Sanomia kaikille: Suomen lehdistön historia. Kleio ja
nykypäivä, Edita, Helsinki (1998)
31. Wevers, M., Smits, T., Impett, L.: Modeling the Genealogy of Imagetexts:
Studying Images and Texts in Conjunction using Computational Meth-
ods – DH2018. In: DH2018 abstracts (2018), https://dh2018.adho.org/en/
modeling-the-genealogy-of-imagetexts-studying-images-and-texts-in-conjunction-using-computational-methods/
32. Wilke, J.: Belated modernization: form and style in German journalism 1880-1980.
In: Broersma, M. (ed.) Form and style in journalism : European newspapers and
the presentation of news, 1880-2005, Groningen studies in cultural change, vol. 26.
Peeters, Leuven, Dudley, MA (2007)