A peek into the discursive construction of the Google Search Algorithm: A critical discourse analysis

Abstract

Algorithms are increasingly interwoven with our daily lives. They are everywhere; yet, they are often invisible. Many people are not aware (of the extent) that algorithms decide what to show and what not to show on their Facebook feeds; or that prices for an overnight stay may depend on which browser is used to look for accommodations. While daily life is saturated with algorithms, just few people have access to its actual codes, enabling them to glimpse into or influence what it does and why. Yet, there seems a consensus that algorithmic opacity might lead to unequal power distribution and abuse of power.

This thesis outlines current discussions on search algorithms that address possible implications of algorithmic opacity, such as: bias, discrimination, transparency, (other) concepts that are related with power, and the position of the (cultural) sciences in this field. This framework renders a context for the research of this thesis, which attempts to identify how the Google Search algorithm is discursively constructed by Google and how this discursive construction may reflect power relations between this powerful software giant and its users. Both quantitative and qualitative methods of critical discourse analysis were used to study Google’s statements about Google as a company and its search products.

It was found that Google’s statements were often contradictory: explicit statements were undermined by linguistic structures and specific uses of words that contradicted explicit utterances. Moreover, Google uses words such as ‘natural’ to make their products and processes more agreeable, while mitigating language is used when negative aspects of Google’s search algorithm were mentioned. Lastly, it was found that some topics were completely absent from the Google’s discourse, such as bias, discrimination, algorithmic transparency and ethics, which are topics that are deemed important by academics. This thesis closes with remarks of the difficulty of researching algorithms due to its complex networked system and suggestions for further research.