Capturing institutional logics and the construction of digital technologies: A macro-discursive perspective using formal text analysis methods
本文提出一种结合命名实体识别、主题建模和词嵌入的形式化文本分析方法，从宏观话语视角研究数字技术意义的社会建构，并以德国人工智能公共话语为例进行实证演示。
Research suggests that the meanings of new technologies in the digital age are constructed through discourses shaped by diverse actors and a variety of social, material, and institutional aspects. This paper introduces a novel methodological approach that enables large-scale investigations into the construction of the meanings of these technologies from a macro-discursive perspective. This macro-discursive perspective allows researchers to capture and analyze overarching patterns of the aspects that shape a technology's meaning in discourses at the macro level. Specifically, the methodological approach involves analyzing texts to capture the co-occurrence of social actors, problems, and solutions associated with a particular technology, as well as the prevalence of institutional logics' symbolic representations. The approach is grounded in theoretical insights from the social construction of technology (SCOT) and institutional logics. On a technical level, this paper proposes that the methodological approach presented can be implemented by combining formal methods of text analysis. Three methods will be discussed in detail: named-entity recognition, topic modeling, and word embeddings. The practical application of this methodological approach will be demonstrated in an empirical illustration based on the public discourse surrounding artificial intelligence in Germany from 2015 to 2022.