ABSTRACT

To date, the evolution of archaeological knowledge production and theory has been discussed and analyzed using qualitative methods by reading vast amounts of archaeological texts in search of specific discourses or framings of the past. In this paper, we present text mining methodologies from digital humanities that can be applied to large corpora of archaeological texts to trace and evaluate changing knowledge practices. Such a big data approach is imperative. Due to the rapid increase of archaeological publications, qualitative research into the intellectual history of archaeology has become complicated and highly selective. The big data methods presented in this study were tested on a large corpus (4,811 texts totaling over 51 million words) of different types of archaeological texts from the Dutch-speaking part of Belgium. The different text mining tools were successful in identifying theoretical trends. Our tools were also successful in charting the decrease in quality due to changed organizational circumstances (developer-led archaeology). Furthermore, we could also map changing banal nationalist framings of the past.

摘要

迄今为止，通过阅读大量考古文本以寻找过去的特定话语或框架，已经使用定性方法讨论和分析了考古知识生产和理论的演变。在本文中，我们介绍了数字人文学科的文本挖掘方法，这些方法可以应用于大型考古文本语料库，以追踪和评估不断变化的知识实践。这种大数据方法势在必行。由于考古出版物的迅速增加，对考古学思想史的定性研究变得复杂且具有高度选择性。本研究中提出的大数据方法在来自比利时荷兰语区的不同类型考古文本的大型语料库（4,811 篇文本，总计超过 5,100 万个单词）上进行了测试。不同的文本挖掘工具成功地识别了理论趋势。我们的工具还成功地绘制了由于组织环境变化（开发人员主导的考古学）而导致的质量下降情况。此外，我们还可以绘制过去不断变化的平庸民族主义框架。