CHAPTER THREE

METHOD

Understanding how a company both promotes and maintains its image, values, or stances on issues can be accomplished through discourse analysis. For this thesis, I accessed Google’s documents that any person with Internet access and interest would also be able to find related to data centers, sustainability, and renewable energy. These materials are presented on separate company websites and blogs, available through web links and attachments on these sites. In this thesis, the ‘public’ in public-facing materials exists because, like Warner (2002) asserts, people are addressed through discourse. For the purposes of this thesis, it is worth noting that Google presents its sustainability measures in a manner that does not require advanced experience with the topics, though those with insights into topics may better understand what is described.

This thesis articulates Google’s data centers to energy and purported sustainability policies as the result of analysis of the discourse the company has issued through their dedicated public-facing materials, the policy documents the EPA has written, and the environmental reports that Google has published independently and through collaborative efforts with other agencies. To accomplish this, I analyzed 26 different Google webpages, two EPA reports, and five joint publications of external agencies including Oxford Economics, Lawrence Berkeley National Laboratory, The World Resources Institute, The Renewable Energy Buyers Association, and the Ellen McArthur Foundation that analyze Google and their data centers. The analysis in this thesis draws from the frameworks about critical discourse analysis (CDA) developed by scholars like Raymond Williams (1959; 1977), André Brock (2018), Michel Foucault (1992b), and Norman Fairclough (1992a; 1992b; 2001). Critical discourse analysis builds articulations, or relationships, between items because it allows for contextualization beyond face-value of these documents. Understanding that Google is a company that relies on the physical environment to power and cool their machines that enable search engine functionality goes beyond the claims that are explicitly made in these documents. Relationships between Google, regional power grids, and renewable energy are revealed once one works to review what is or is not publicly said about those things. By evaluating the discourse they provide, matching it further with documents they provide through partnerships with outside sources, and by analyzing government materials related to data centers, I am able to contextualize and evaluate the claims, key terms, and to understand more holistically how Google situates and promotes itself through their disclosure.

Discourse, as Fairclough (1992a) argues, “is socially constructive, constituting social subjects, social relations, and systems of knowledge and belief, and the study of discourse focuses upon its constructive ideological effects (p. 36).” Discourse is “socially constructive” because it consists of language, social values, and social ideologies. Language comprises discourse, and language appears through “signification;” Williams argues that “signification” is “the social creation of meanings through the use of formal signs, is then a practical, material activity; it is indeed, literally, a means of production” (Williams, 1977, p. 38). Signs are created and arranged through human interpretation into meaning, which is a means of production because it requires labor in presentation, maintenance, and through reproduction. It is through the interpretation of signs, or the words, letters, grammar, etc., that we use that we construct our thoughts, ideas, and understandings of the world around us. In this way, interpretation is both subjective and socially constructed and maintained, portrayed materially through mediation, whether via paper, computers, devices, etc. In this thesis, discourse through the publicly available language used on Google’s websites, blogs, and reports reveals overall stances, beliefs, and values of the company related to the topics like energy, infrastructure, resources, and revenue. Fairclough (2001) asserted that CDA builds from Williams’ view, where it utilizes the view of “semiosis as an irreducible element of all material social processes” (p. 1). Here, “semiosis” refers to the idea of meaning making, where words, letters, and their formations constitute understanding of the world. Fairclough further argued that “analysis of discourse attends to its functioning in the creative transformation of ideologies and practices as well as its functioning in securing their reproduction” (Fairclough, 1992a, p. 36). When studying discourse, researchers can make informed discoveries about cultural values, understandings, positionality, power dynamics, and more because of the semiotic nature of language. Critical discourse analyses “offer interpretations of how a text can become polysemous and effective when placed in the public domain of cyberspace” (Mitra & Cohen, 1999). Brock (2018) further asserts that CDA appreciates “hermeneutics,” reminding readers that CDA relies on interpretation as an analytical tool (p. 1019). For my thesis, I use CDA to interpret the materials, looking specifically for keywords, repeated phrases, or explicit mentions of values to draw commonalities between the documents, to track the ways they changed over time if possible, to interpret how Google positions themselves to the resources they consume.

One key aspect of conducting CDA also involves understanding what discourse is or what constitutes discourse. Michel Foucault is lauded as one of the key scholars in discourse and the methodology of discourse analysis. Michael Arribas-Ayllon and Valerie Walkerdine (2011) summarize Foucault’s approach as one concerned with “genealogy” or “mechanisms of historic inquiry; “the archeology of “the mechanisms of power and offers a description of their functioning” (p. 92). Fairclough (1992b) summarizes Foucault’s understanding of “discourse” as “concerned with analysing ‘statements’” and these are concerned with “specifying sociohistorically variable ‘discursive formations’ (sometimes referred to as ‘discourses’), systems of rule which make it possible for certain statements but not others to occur at particular times, places, and institutional locations” (p. 40). Essentially, Fairclough (1992b) summarizes how Foucault’s understanding of discourse is closely tied to power, systems of power, and power in society. Discourse can be interpreted to reveal positions and systems of power, to contextualize some of the social values and ideologies.

CDA has been used in other ways by critical scholars, including Michelle Lazar (2005) in her works on feminist critical discourse analysis. Lazar (2005) asserts that “CDA is known for its overtly political stance and is concerned with all forms of social inequality and injustice” (p. 2). I draw inspiration from Lazar, in that understanding Google’s position and stance with the ongoing climate crisis matters to me, and I believe corporations should be held responsible for the impacts they make on the environment. Further, Cubitt (2017) also reminds readers that environmental regulations are political, that “media are not only passive channels of communication: Parts of no part excluded from their own governance, they seek to speak for themselves” (p. 182). Energy regulations will impact Google’s data centers. André Brock (2018), advances the CDA framework to conduct his research on “technocultural” matters: critical technocultural discourse analysis (CTDA). In his work, Brock (2018) asserts that CDTA uses CDA to isolate and examines “site-based online discourses” (p. 1025). Though this thesis does not utilize CDTA, an understanding of CDA’s inspiration and applicability for online discourses is helpful in understanding the framework I used during my analysis of websites, linked materials, and web-based reports. Brock’s (2018) work, though specifically pertaining to Black Twitter, provides inspiration for conducting critical discourse analysis via these online, mediated sources.

I also use lenses of inquiry from other communication infrastructure and technology scholars like Sean Cubitt (2017), John Durham Peters (2015), Nicole Starosielski (2015), and Harold Innis (Young, 2017). Cubitt (2017) theorizes the way technology in its current model is unsustainable, depleting earth’s finite resources, and how technology is unsustainable with its cyclical obsolescence. Peters (2015) urges for a material understanding of media; “to understand media we need to understand fire, aqueducts, power grids, sewage systems, DNA, mathematics, sex, music, daydreams, and insulation” (p.29) because all of these things coexist to make media possible. Peters (2015) argues, “the digital changes of our times are impossible without mines and minerals, clouds and electrical grids, habits of human want and labor, and global patters of human inequality and abuse” (p. 377). This viewpoint is particularly salient for this thesis, as I am interested in understanding the energy and resources that enables Google, through its data centers, to meet its enormous demands. Starosielski (2015) asks questions about the invisibility of essential infrastructures, specifically undersea cables that are the “backbone of the global Internet” (p. 3). The question of “why have undersea cables, as the backbone of the global Internet, remained largely invisible to the publics that used them?” (p. 3) inspired this research altogether; Starosielski (2015) posed a question I applied to data centers and their energy consumption, since I had not heard about these infrastructures or been exposed to the figures of energy usage ever before, despite relying on servers most of my life growing up in a digital age.

To compile the archive of websites, reports, blog posts, and linked attachments, I conducted a search on Google Chrome using “data centers,” “sustainability,” and “renewable energy” as search terms. I chose these terms because I felt they best captured the questions I am asking, while also being broad enough to generate many results. The results of these searches revealed entirely separate websites dedicated to these search terms; Google does not include this information on their company webpage and instead redirects interested parties to these other websites/blogs on which all of their information is housed. For example, Google has separate sites for Sustainability and data centers, though the two redirect to one another frequently. These accessible materials featured everything from power usage graphics, images of Google data centers around the world, real-time updates about energy usage at specific locations, and easily navigable extensions to other pertinent materials. Blog posts were typically given authorship by high-ranking Google employees. The most common authors include Urs Hölzle, Google’s Senior Vice President of Technical Infrastructure; Michael Terrell, Google’s Head of Energy Market Development; Kate Brandt, Google’s Chief Sustainability Officer; and Ruth Porat, Google’s Senior Vice President & Chief Financial Officer. I decided to analyze these materials because I wanted to critically analyze the language, discourse, as well as the relationships between such discourse, the environment, and the positions of corporate power that Google asserts through these public-facing materials. These materials were assessed using the method of critical discourse analysis to reveal keywords in the following findings section and some resulting arguments about public relations serving as policy for Google, which are explored in the analysis section.

To do such critical analysis, I read through each of the documents. After doing an initial reading of multiple items, I reread the documents and pulled quotes that I found were similar or repetitive, offered any key phrases, or explained Google’s values explicitly or inexplicitly. After doing this with each material in the archive I constructed, I then went back through and analyzed the quotes I pulled and began synthesizing the findings. From there, I did some secondary research to learn more about some of the technicalities of the power grid, renewable energy credits and buying, matching versus converting energy consumption, and referenced the reports through other agencies. From there I completed the Findings and Results arguments that follow.

It is worth noting that my work is not a study in public relations. The method of CDA allows me to focus on and contribute to infrastructure studies, as the documents analyzed would not exist or be necessary without infrastructure. The CDA process allowed me to establish an understanding of Google’s discourse, how they situate themselves to the public and the climate crisis, and to make some arguments about Google’s purported values within the world by way of their infrastructure. That is the critical nature of my work; to use discourse as a means to building the articulations of Google, the environment, and their data centers; to bring in the context of all of their documents, filled with the language of their values, and situate these things against the climate crisis, in context.