Moving Fingers Write History and Having Writ Become Digital: Towards a Big Data Framework for the Analysis of Parliamentary Proceedings

Abstract

AbstractParliamentary proceedings provide political scientists and historians with traces of evidence about the attempts of politicians to shape the destiny of their electorate. Researchers in history and politics refer to this evidence. However, the recorded proceedings constitute large volumes of data making it difficult to comprehend and analyse in a systematic way. Leveraging big data technologies and computational linguistics to harness the digital record of parliamentary proceedings enables accomplished and novice political scientists and historians alike to gather shreds of evidence about historic events or politicians by multi-dimensional analysis of the traces. We report our preliminary work on looking at the proceedings of the UK (& Ireland) Parliament between 1801–1921 - the contribution of the parliamentarians have been made available by the UK Parliament - some 2 million individual contributions by 27,000 parliamentarians: We have chosen a subset of 4 key members of the parliament - 2 Irish and 2 English politicians with a total output of 7.4 million tokens in 16822 contributions. There is a statistically significant difference at three levels of linguistic description - lexical, terminological/semantic, and affect levels. This work is a prelude to the analysis of the whole 120-year corpus. Our paper helps people to analyse the distribution of words used in parliamentary speeches together with information and statistical significance of that word or a class of words. Our paper does not attempt to comment on historical inferences, but rather just observes them, leaving the commentary and inference to historians. The parliamentary exchanges can be characterised as having the 5 key features of Big Data - value, variety, veracity, volume, and velocity.KeywordsParliamentary proceedingsBig data frameworkMulti-dimensional analysisHansard