Academic journal article Journal of American Folklore

Automatic Enrichment and Classification of Folktales in the Dutch Folktale Database

Academic journal article Journal of American Folklore

Automatic Enrichment and Classification of Folktales in the Dutch Folktale Database

Article excerpt

There is an increasing interest in creating digital collections of folk narratives because of the potential to make large numbers of narratives accessible to scholars and-if properly annotated-to allow investigators to search on dimensions of interest such as theme and location (la barre and tilley 2012). in addition, by applying computational analysis and visualization methods to such collections, previously unperceived patterns may be revealed, leading to new insights (Abello, broadwell, and tangherlini 2012).

In this paper, we focus on one such digital collection: the Dutch Folktale Database, an online collection of folktales maintained by the meertens instituut in Amsterdam. The Dutch Folktale Database serves both as a digital archive and as a digital research instrument. to enhance its usefulness as a resource for scholars, the folktales in the collection have been annotated with metadata, ranging from the name of the narrator to the international tale type. So far, these metadata have been assigned manually, a process that is both time-consuming and prone to errors. recently, we have started to investigate the possibility of automating this process using various machine learning techniques that enable the computer to automatically learn from data to, for example, generate metadata on its own. in addition, we have experimented with the computational modeling of motifs in our collection. in the following sections, we present the primary results of these efforts, and discuss the implications for future research. First, we give an overview of the Dutch Folktale Database and its development over the years.

The Dutch Folktale Database 1.0-3.0

The Dutch Folktale Database (DFD) was first developed as a stand-alone database in 1994; it was mainly used by researchers at and visitors to the meertens instituut in Amsterdam. The database was designed to offer insight into folktale1 traditions in the netherlands, and to provide access to extant synchronic and diachronic variations. to meet this goal, complete story texts gathered from written historical sources or transcribed from audio recordings were included in the database. relevant metadata were added to each individual folktale: Who was the narrator? Where was the story told? When was the story told? in what language or dialect was the story told? Who collected the story, and how was it recorded? other metadata used to enrich the tales include a list of personal names, place names, keywords, a brief summary of the tale in modern Dutch, the genre and tale type numbers where appropriate, and motifs in the story according to Stith Thompson's Motif-Index.

The most important international catalog used initially to identify stories in the DFD was The Types of the Folktale by Antti Aarne and Stith Thompson (1964). This index works appropriately for classifying widespread, traditional fairy tales and anecdotes, but not for local tales, anecdotes, traditional legends, modern jokes, and urban legends. For identification of folktales within these genres, it was necessary to use additional catalogs such as (in order of importance): the Frisian folktales compiled by Jurjen van der kooi (1984), the traditional legends and saints' legends compiled by Jacques rudolf Willem Sinninghe (1943), the urban legends compiled by Jan harold brunvand (1994), and an additional general index compiled by Theo meder.2 The coverage of these additional catalogs is too limited to identify all stories; in many cases, it is impossible to classify unique stories, in particular family stories or personal narratives with a tale-type designation. in 2012, our group estimated that it was possible to identify 60 percent of the stories in the database with the folktale indices, leaving 40 percent of the tales unclassified.3 As the DFD 2.0 was being developed, the new index Types of International Folktales by hans-Jörg uther (2004), an expanded and updated index based on Aarne-Thompson, appeared. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.