A new article (open access, peer-reviewed) resulting from The Riddle of Literary Quality has appeared in the journal Language Resources and Evaluation. Previous work already showed that literary novels can be recognized successfully with textual features. This article shows that literature can even be recognized with short fragments (2-3 pages), and also considers judgments of quality and novels that respondents had not read. The textual features are automatically learned document representations that require no feature engineering and are only based on word frequencies.
In addition, the article tests the hypothesis that literary
novels are more complex than non-literary novels. By measuring the similarity
and variety of topics in the novels, literary novels are shown to stand out
more than non-literary novels.
A keyword analysis uncovers some of the stylistic markers
that explain the success of the predictive models. It also highlights certain
biases related to genre and gender, both in the data and the models.
Nevertheless, we find that the greater part of factors affecting judgments of
literariness are explicable in terms of word frequencies, even in short text
fragments and among novels with higher literary ratings.
van Cranenburgh, A., van Dalen-Oskam, K. & van Zundert,
J. (2019), Language Resources & Evaluation. https://doi.org/10.1007/s10579-018-09442-4
Do novel readers want to be swept away by a novel? Or do they seek an intellectual challenge? Or both at the same time?
In 2013, the research team of The Riddle of Literary Quality, led by Professor Karina van Dalen-Oskam in the Netherlands, completed The National Reader Survey. Almost 14,000 respondents shared their opinions about recently published novels. The participants also responded to six statements that related to the question of what they personally considered important to reading fiction.
Today, in the international (open access, peer-reviewed) journal PLOS One, an initial analysis of these reactions appears. The American researcher Allen Riddell applied methods to analyze these responses that were previously used in research into voting behavior. He discovered an interesting pattern in the responses of the representative group of participants in the National Reader Survey. Readers who indicated they read in order to be intellectually challenged also indicated that they wanted to be carried away by a novel. Conversely, this did not apply: not all readers who wanted to be dragged along, also wanted to be challenged intellectually.
Readers who want more than just being swept away by a story are usually ‘literary’ readers. They usually have more demands and higher expectations when they read, they not only restrict themselves to the story, but also want to enjoy the use of language in a novel. This group is often critical of popular books that are “devoured” by others. But this research shows that they, as readers, do report seeking emotional engagement with works: they report wanting to be carried away by a novel. An interesting outcome, not only because it tells us something about the readers of literature but also about the literature itself.
In the article, Riddell and van Dalen-Oskam state that they suspect that the literary reader has come to appreciate this way of reading in the education they have enjoyed or by developing their own reading experience. Literary readers sometimes are critical about ‘non-literary’ readers and the books they read, but this research shows that they do want the same. Only a little more …
Allen Riddell & Karina van Dalen-Oskam, Readers and their roles: evidence from readers or contemporary fiction in the Netherlands. PLOS One, 26 July 2018, https://doi.org/10.1371/journal.pone.0201157
The Riddle is in its final phase. In the next two years, much of the experiments that have been done will be reported on in one more PhD-thesis, scholarly articles and in a book for all those who enjoy reading and may even have participated in our National Reader Survey in 2013. One of the things still on our to do list was applying stylometric tools with the knowledge we gained concerning the literary conventions of contemporary novels on older fiction. In April 2017, Floor Naber has started a short stint at Huygens ING’s Riddle Team to do an experiment with LIWC. She will use the software Linguistic Inquiry and Word Count on the Riddle corpus and on a corpus of late nineteenth-century Dutch novels to test whether old and new compare or where they differ.
2 November 2016: Andreas van Cranenburgh defends his PhD-thesis Rich statistical parsing and literary language.
This thesis studies parsing and literature with the Data-Oriented Parsing framework, which assumes that chunks of previous experience can be exploited to analyze new sentences. As chunks we consider syntactic tree fragments.
After presenting a method to efficiently extract such fragments from treebanks based on heuristics of re-occurrence, we employ them to develop a multi-lingual statistical parser. We show how a mildly context-sensitive grammar can be employed to produce discontinuous constituents, and compare this to an approximation that stays within the efficiently parsable context-free framework. We show that tree fragments allow the grammar to adequately capture the statistical regularities of non-local relations, without the need for the increased generative capacity of mildly context-sensitive grammar.
The second part investigates what separates literary from other novels. We work with a corpus of novels and a reader survey with ratings of how literary they are perceived to be. The main goal is to find out the extent to which the literary ratings can be predicted from the texts. We first evaluate simple measures such as vocabulary richness, text compressibility, and the number of cliché expressions. In addition we apply more sophisticated, predictive models: a topic model, bag-of-words model, and a model based on syntactic tree fragments. We find that literary ratings are predictable from textual features to a large extent. While it is not possible to infer a causal relation, this result clearly rules out the notion that these value-judgments of literary merit were arbitrary, or predominantly determined by factors beyond the text.
Link to the thesis: http://dare.uva.nl/record/1/543163
In 2014, the editors of the Utrecht journal for literary studies Vooys:asked project leader Karina van Dalen-Oskam for a short article about the background of the project The Riddle of Literary Quality.. The editors of Vooys gave permission to post an English translation of this article on the Riddle blog. The original publication is titled ‘The Riddle of Literary Quality. Op zoek naar conventies van literariteit’ and was published in: Vooys: tijdschrift voor letteren 32 (2014), 3, p. 25-33.
The Riddle of Literary Quality: The search for conventions of literariness
Recently, the first results of the large reading survey that is part of the project The Riddle of Literary Quality were published on the website of Het Nationale Lezersonderzoek. The survey could be filled in from March 4 until September 27 2013. We are very happy that in total 13,782 readers did this.Some of the first results (in Dutch) can be found on the survey site. More will follow. And ofcourse, we are now starting work on the analysis of as many of the 400 novels in the list as is possible, to find out whether we can find any correlations between features of the texts, readers’ opinions, and also readers’ predominant reading role. Results in this area will take some more time, but we are sure that they will yield interesting new insights into what a book needs to be evaluated as literary or not, good or bad, by different kinds of readers.
On March 4th 2013, the survey of the project The Riddle of Literary Quality was launched on http://www.hetnationalelezersonderzoek.nl/ . This “National Readers’ Enquiry” hopes to reach many thousands of respondents. As can be expected based on the nature of our project, the language of the survey is Dutch. All readers of Dutch among you are welcome to give your ratings on a set of novels (originals and translations) published during the last five years in The Netherlands. The list of novels contains those that were borrowed most from public libraries and that ranked highests on the bestseller lists of the last three years. Enjoy!
The Riddle team has recently become a lot larger. PhD-student Corina Koolen was liaised to the project starting in September 2012, project assistant Fernie Maas joined the 1st of November, and Kim Jautze started work on her PhD on 15 November. More news is sure to follow, the team now being complete!
Andreas van Cranenburgh represented The Riddle at The North-American Chapter of the Association for Computational Linguistics (NAACL, June 3–8 2012, Montreal). He writes:
In the main conference, the first paper that caught my eye was one about a
task called “multiple narrative disentanglement.” This is simply the problem
of recognizing in running text the different narratives with their own sets of
characters and storyline. A method is introduced which is applied to the
famously complex novel ‘Infinite Jest’ by David Foster-Wallace.
Last week, Andreas van Cranenburgh attended the 13th Conference of the European Chapter of the Association for Computational Linguistics (EACL). EACL is one of the foremost conferences in the field of computational linguistics; this year’s acceptance rate was well below 30%. There were only 5 other works about parsing, and all of these were about dependency parsing instead of constituency parsing (different representations for expressing the syntactic structure of sentences). This meant that Andreas’s poster on discontinuous parsing was the only one to focus on constituency structures, which are commonly used in Data-Oriented Parsing and thus relevant to our project.
One paper in particular stood out due to its relevance to our project: Character-based kernels for novelistic plot structure. The paper presented a method to analyze and compare plot structure of novels. For example the relations of characters in a social network can be extracted, as well as their `emotional development’ based on a list of emotion-related words. The resulting information is used to produce a similarity metric for texts. One graph, for example, plotted the emotions of the protagonist of a Jane Austen novel, showing strong peaks corresponding to a proposal, elopement, and marriage of the protagonist. It is encouraging to see that even with a relatively superficial linguistic analysis, interesting details can be revealed of literary texts.