I extend thanks to Inna Kouper for her response to “(Google-) Knowing Economics”, which I co-authored with Fabien Medvekcy. The points she raises are fascinating and important, as is this whole area of study. How we come to know with Google is only becoming more important as we rely ever more heavily on the search engine and its associated applications. I greatly appreciate the chance to think and write further on these issues … [please read below the rest of the article].

Image credit: Bytemarks via Flickr / Creative Commons

I extend thanks to Inna Kouper for her response to “(Google-) Knowing Economics”, which I co-authored with Fabien Medvecky. [1] The points she raises are fascinating and important, as is this whole area of study. How we come to know with Google is only becoming more important as we rely ever more heavily on the search engine and its associated applications. I greatly appreciate the chance to think and write further on these issues.

Kouper raises a number of important points around google-knowers and google-knowing. In particular, she would like to see google-knowers and google-knowing more clearly defined.[2] She would also like to see more acknowledgement of the role that the complex lines power (and algorithms) play in making certain knowledges visible. I believe our goals in this are aligned, and these projects must remain a work in progress.

Google-knowers and google-knowing are always changing, because Google is changing too. As a business it seeks to keep in front of technological development, and quietly tweaks its algorithms as it sees fit.[3] That said, there are ways of approaching google-knowing I think will remain relevant, and I discuss them below. Kouper’s experiences are mostly with science communication and science knowledge, and these are the useful examples she draws upon, but as I will suggest in the conclusion, google-knowing economics hints at an even deeper problem.

Given the dominance of the search engine—with Google search holding 87% of worldwide market share for desktop computing in January 2020 and a whopping 95% market share on mobile devices—developing a sense of how to approach google-knowledge is vital. [4],[5] What is needed, we think, is a way to embrace the situated individual knower AND understand their google-knowledge as part of the political economy of Google. (Fabien Medvecky and I are working on a book-length exploration of google-knowing attempting to do just that.)

Google-Knowers

We have called for a more agent-centred methodology to better understand how knowledges (in this case about economics) are made public, especially now that Google search plays such a major role in how people know. We think an agent-centred approach to studying google-knowing is needed for several reasons: the personalisation of search results, the associated dis-jointing of knowledge across populations, the sense of choice agents have over how they navigate search results, and the changing result profiles over time. If we look with the eyes of a single searcher then we can gain a snapshot of Google results at one time. It means we can know with knowers, rather than about them. And as knowers entwined physically and mentally with computers, phones, search engines and their algorithms, we can know with Google too. We can, like all google-knowers, embrace our embodied joining with our technology as the cyborgs we are.[6]

We have said that it is important to recognise that each knower is situated in a body, and in time and space. But to accept our situated-ness shouldn’t mean to maximally limit our possibilities for seeing further. To this end, we take the step of using the (somewhat) anonymising incognito search. This means we are reducing the effect of personalisation on our search results, while still retaining some located-ness. We want to suggest that ‘situated’ and ‘objective’ knowledges are not binary, rather they exist at the two ends of a continuum. Our google-knower is situated for Google (in a physical place and time by an incognito search) though they are not as situated as they would be if all of Google’s personalisation tools had been activated. The Google search results we study are not objective; they never could be. But they are more generalisable than if we had embraced personalisation. Our google-knower is in time and place, but without heavy baggage from past searches. Embracing contamination to the maximum would come at the cost of this generalisable information. So instead we acknowledge a more limited contamination that lets us see just that bit more broadly.

Google-Knowledge

Using the google-knower as a methodological figure lets us into a world of particularity and nuance—as compared to the picture we would get if we looked at big data. It also allows us a glimpse of what other knowers might be seeing without relying on what they say they saw.

One thing we are able to see as google-knowers is the ways that boundaries between forms of communication are blurred—from Wikipedia, to YouTube videos, to entertainment and shopping. Google doesn’t just trade in information as a public good, but as a commodity to attract users—eyes for advertisers.[7] I would argue (but why do I believe this?) that Google is sincere in its mission statement—‘to organise the world’s information and make it universally accessible and useful’.[8] But it is also the case that for Google information is what attracts users, and the accuracy of that information is not the only consideration. For Google-the-advertising-business ‘good’ information is enticing and distracting, it confirms what we think and feels satisfyingly ‘truth-y’.[9]

We don’t know what the algorithms look like that govern the information that shows up for us about economics (or about multitudes of other important types of knowledge). These are Google’s trade secrets. But we do know the results of the algorithms, and to a large extent this is enough. Algorithms have often been metaphorically understood as ‘black boxes’, as something we need to ‘see under the hood’ of.[10] The metaphor implies that only when we understand the structure of what is underneath can we know how to resist it. But this can serve to make us passive in the face of Google and other technologies. If we don’t understand, we can’t fight. But this is not the case: what we need to understand is what we are shown, it should be enough that we understand the results of the algorithms that guide our knowing.[11] Simply, we need to know about what comes up on Google’s first results page—that page that few people bother to go past.[12] That is what we have attempted to describe, and in describing, we hope to challenge the contours of the picture of economics we are being shown. This is how a google-knower, knowingly blinkered by personalisation and located-ness, can respond to the political economy of Google search.

Testimonial Knowledge

Kouper has picked up on our claim that all google-knowledge is testimonial in form. It is knowledge from someone else, or from several someone elses in the case of Wikipedia. It is not empirically gained: not gained from our own experiences or perception of the world. But I don’t think this is obvious to most Google users, partly because this form of knowledge has become increasingly dominant in a world known by reading and by digital materials. It means that when we are assessing the truth of a claim, a big part of what we are judging (often unconsciously) is the trustworthiness of the speaker or website author.

I think that it’s hard to overstate the importance of this observation. It means that even when we think we are talking about the objective truth of a claim, what we end up talking about are actually subjective truth claims. We end up wondering not whether something is true, but whether it is reasonable to believe what someone says. Then our discussion is really about authority in a world where money and power sneak in often imperceptibly into what we are told. Kouper has suggested that the problem with trusting testimony is needing to know about the ‘rules that govern the speakers’ behaviour and the intentions and motivations of those who are advancing the knowledge claim’.[13] We agree heartily, with the caveat that none of this matters for google-knowers or google-knowledge if those testimonies don’t make it onto Google’s first page. It is the rules that govern Google’s algorithmic behaviours and assessments of authority that matter.

Conclusion

Kouper has said that Google is ‘not really to blame’ for the challenges of making economics public. She is pointing to the complex and concerning world in which some groups can push their websites further up and onto Google’s first page. In this world, she is suggesting, it is the non-neutrality of groups pushing their messages (about vaccination, climate change etc) that we should understand. In this world the idea that information is simply made available for citizens to make good individual and collective decisions is naïve.

She is right: but in fact, the problems that citizens face when trying to gather a picture of economics using Google only makes the challenge of making good decisions greater. The non-transparency of Google’s algorithms, and the ways some information is made easily available (and the way some is presented as more authoritative), mean that people may struggle to grasp not only economics as a set of disciplinary knowledge but also the economy in which Google operates. This means google-knowers have a very knotty problem: in a world dominated by Google, knowledge of economics AND knowledge of how economics impacts what we know about economics is determined by algorithms out of our sight. And for researchers who want to use Google to know what google-knowers know … What we can do for a start is stand in the same place as google-knowers and survey the landscape.

Contact details: Vicki Macknight, University of Otago, vicki.macknight@gmail.com

[1] Macknight, Vicki, and Fabien Medvecky. 2020. “(Google-) Knowing Economics.” Social Epistemology 34 (3): 213-226.

[2] For a philosophical attempt, see Gunn, Hanna Kiri and Michael P. Lynch. 2018. “Googling.” In The Routledge Handbook of Applied Epistemology, edited by David Coady and James Chase, 41-53. Routledge.

[3] And this is even more so now with the machine learning RankBrain becoming one of the three most important factor in producing search results: Graham, Richard Norroy. 2017. “Understanding Google: Search Engines and the Changing Nature of Access, Thought and Knowledge within a Global Context.” PhD thesis, 197, University of Exeter.

[4] Worldwide desktop market share of leading search engines from January 2010 to January 2020
https://www.statista.com/statistics/216573/worldwide-market-share-of-search-engines/. Accessed 18 June 2020. See also https://gs.statcounter.com/search-engine-market-share/desktop/worldwide.

[5] Mobile search engine market share worldwide, https://gs.statcounter.com/search-engine-market-share/mobile/worldwide. Accessed 18 June 2020.

[6] The figure of the cyborg, of course, links us to Haraway’s Cyborg Manifesto (1985). It has been used productively to understand people’s relationships with Facebook, see for example Waite, Catherine and Lisa Bourke. 2015. “Using the Cyborg to Re-Think Young People’s Uses of Facebook.” Journal of Sociology 51 (3): 537-552.

[7] For a Marxist reading of Google’s business model, see Fuchs, Christian. 2019. “A Contribution to the Critique of the Political Economy of Google.” Fast Capitalism 8 (1): http://www.fastcapitalism.com. http://fuchs.uti.at/wp-content/uploads/Google_FastCapitalism.pdf. For a discussion of commodification of information in Google’s blog, see Hodson, Jaigris. 2014. “Google Me: A Discourse Analysis of the World’s Leading Search Engine.” International Journal of Technology, Knowledge & Society: Annual Review 9 (2): 125-138.

[8] ‘Google search: our mission’, https://www.google.com/search/howsearchworks/mission/. Accessed 18 June 2020.

[9] In 2019 Google’s profit from advertising was approximately 71% of total profit, amounting to almost 135 billion U.S. dollars. Advertising revenue of Google from 2001 to 2019, https://www.statista.com/statistics/266249/advertising-revenue-of-google/. Accessed 22 June 2020.

[10] Lee, et al (2019) discuss the limits we set upon how we understand algorithms when we use different metaphors. Lee, Francis and Lotta Björklund Larsen. 2019. “How Should We Theorize Algorithms? Five Ideal Types in Analyzing Algorithmic Normativities.” Big Data & Society 1-6. https://doi.org/10.1177/2053951719867349.

[11] As Taina Bucher has put it “what people experience is not the mathematical recipe as such but, rather, the moods, affects and sensations that the algorithm helps to generate” (32). Bucher, Taina. 2017. “The Algorithmic Imaginary: Exploring the Ordinary Affects of Facebook Algorithms.” Information, Communication & Society 20 (1): 30-44.

[12] According to eye-gaze studies, see Pan, Bing, Helene Hembrooke, Thorsten Joachims, Lori Lorigo, Geri Gay, and Laura Granka. 2007. “In Google We Trust: Users’ Decisions on Rank, Position, and Relevance.” Journal of Computer-Mediated Communication 12 (3): 801-823.

[13] Kouper, Inna. 2020. “Googling as Research: A Response to ‘(Google)-Knowing Economics’”. Social Epistemology Review and Reply Collective 9 (6): 19-24. https://wp.me/p1Bfg0-567, 22.

Categories: Critical Replies

Tags: epistemic status, epistemology, Fabien Medvecky, Google-Knowers, Google-knowing, Google-Learners, Inna Kouper, knowing, knowing in the age of Google, testimonial knowledge, Vicki Macknight