2020 Workshop

The Right to the Model

Intelligence is the ability to make decisions about decidable alternatives. Consciousness is the ability to decide about undecidable alternatives. Intelligence implies computation and combination, while consciousness implies sensibility (aesthetic and erotic) and ethical judgment.

— Franco “Bifo” Berardi, “(Sensitive) Consciousness and Time: Against the Transhumanist Utopia”

On March 17, 2020, the Province of Ontario invoked the “Emergency Management and Civil Protection Act,” sending all nonessential workers into quarantine. At the press conference, a journalist — they were still physically present — asked “why today?” The Chief Medical Officer’s response noted coordination with his provincial and federal counterparts, but was otherwise opaque.¹CPAC, “Ontario Premier Doug Ford declares state of emergency due to COVID-19 – March 17, 2020.” Dr. Williams response to the journalist’s question runs from 15:50 to 16:30. The decision, and the collaborative process that led up to it, would have involved reviewing multiple epidemiological models. Models often seem like black boxes, in which decision-making processes are concealed within the workings of computation. But models are also authored works; what they reveal about disease transmission depends upon the information that modelers have and the assumptions that they make. Here my aim is to draw out some of this influential practice, and to make it more publicly accessible.

The right to the city, theorized by Henri Lefebvre in 1968, was a radical proposition about how to make decisions within cities: about how to enfranchise the broadest range of residents within decision-making processes, irrespective of citizenship, race, ethnicity, gender, sexual orientation, or age.²Henri Lefebvre, “The Right to the City,” in Writings on the City, trans. Eleonore Kofman and Elizabeth Lebas (Cambridge: Blackwell Publishers, 1996). In revisiting Lefebvre’s argument in 2008, geographer David Harvey rearticulated the importance of social organizing in the 21st century. See David Harvey, “The Right to the City,” in The New Left Review 53 (September–October 2008). The COVID-19 pandemic has produced public policy decisions that have transformed cities around the world through spatial and place-based interventions like social distancing. These spatial recalibrations are byproducts of a decidedly non-spatial process — that of epidemiological modeling. Epidemiological models test how different interventions might change the transmission of a disease by replacing spatial or contextual details with numerical abstractions. Thus I would argue that, much like the right to the city, we need the right to the model in order to understand how models count or account for different lives and the ethical judgements that are in play.³Beth Coleman, “Right to the Smart City: How to Represent, Resist, or Disappear,” in Ways of Knowing Cities, ed. Laura Kurgan and Dare Brawley (New York: Columbia Books on Architecture and the City, 2019), 147–148. Coleman criticizes the framing of representation, access, and inclusion within the concept of rights, as it relies on a coherent state to recognize and grant those rights. Within this short article, given its focus on public health systems, I still find it a useful concept.

I recently talked to epidemiological researcher Ashleigh Tuite to learn more about the modeling process. Based at the University of Toronto, Tuite has been working with Amy Greer and David Fisman to rebuild a model they first created in 2009, in response to the H1-N1 virus, to study the spread of COVID-19 in Ontario today. The original model tested what would happen if different population groups were given the H1-N1 vaccine, the availability of which was limited.⁴Amy Greer, Ashleigh Tuite, and David N. Fisman, “For Debate: Age, Influenza Pandemics and Disease Dynamics,” in Epidemiology and Infection 138, no. 11 (2010): 1542–1549. Even this simple description — of counting and calculating lives infected, recovered, and deceased — underscores that epidemiologists are taking into consideration matters that are not only scientific and statistical but also social and political.

Models simplify the complexity of the world by omitting forms of contingency. The Toronto researchers’ model segments Ontario’s 14.6 million residents into age brackets; in doing so the model assumes that constituents within each group are homogeneous — but also avoids addressing the broader social determinants of health.⁵At the start of the pandemic, no Canadian province collected data about the race or ethnicity of people infected with COVID-19. On June 15, Ontario proposed a regulatory change initiating the collection of such data, and on June 29, Public Health Ontario provided case support managers with instructions for how to start collection. See Shanifa Nasser, “Early signs suggest race matters when it comes to COVID-19. So why isn’t Canada collecting race-based data?,” CBC News, April 17, 2020; Kate Allen, “Ontario is starting to collect race-based COVID-19 data. Some worry it could do more harm than good,” Toronto Star, July 2, 2020. While simplification creates speed and legibility — qualities which are arguably the main value of large-scale, generalized models — the right to the model that I am arguing for demands transparency about representational practices, such as the scope and method of data collection, that extends beyond the modelers themselves.

The rebuilt model in Toronto tests the efficacy of different interventions in reducing the rates of contact that might lead to infection. The remaking process began with Fisman reviewing early biomedical literature, emerging from Wuhan, for possible inputs, such as transmission rates. The team then reviewed other available inputs and strategized how to structure the model in order to simulate infection rates and track whether these cases led to hospitalization, recovery, or death. Dialogue and reflection led to decisions that selected specific values over others at each step in the process.

Tuite has also participated in multi-stakeholder modeling exercises in which a wider range of participants work to set the assumptions and values encoded in a model; for instance, she recently collaborated with journalists Nicholas Kristof and Stuart A. Thompson for a New York Times article that showed how the model was being used to answer different questions about the impacts of COVID-19, and the potential narratives that can be constructed.⁶Nicholas Kristof and Stuart A. Thompson, “How Much Worse the Coronavirus Could Get, in Charts,” New York Times, March 13, 2020. This open and public-facing approach, while presenting its own difficulties, encourages a broader stewardship of the data that are being used.⁷See Kevin Donnelly’s history of bias relating to the statistical analysis of biometric data, “We Have Always Been Biased: Measuring the Human Body from Anthropometry to the Computational Social Sciences,” in Public no.60 (March 1, 2020), 20-33. https://doi-org.myaccess.library.utoronto.ca/10.1386/public_00003_7.

Epidemiological models are tools of governance. As such, it is important that we ask basic questions about the assumptions upon which the models are made, so that we can understand both what they reveal and what they conceal. Tuite’s collaboration with journalists demonstrates one way in which models become subject to broader questioning and matters of public concern. As the media scholar Wendy Chun has argued, we need to account for the potential violence inherent in different ways of counting human lives.⁸Wendy Hui Kyong Chun, “Introduction: Race and/as Technology; or How to Do things to Race,” in Camera Obscura: Feminism, Culture, and Media Studies 24, no. 1 (70)(May 1, 2009), 7–35. https://doi.org/10.1215/02705346-2008-013. A brief example, drawn from economics models, makes plain how disturbing the use of calculation is when divorced from questions of representation and ethics. In a memo from 1991, World Bank economist Lawrence H. Summers argued for directing pollution from developed to less developed countries — the consequences of which would prolong the lives of one group, while shortening the lives of another. He based his argument on economic models that made assumptions about the monetary value of the lives of different groups of people. See James A. Swaney, “So What’s Wrong with Dumping on Africa?,” Journal of Economic Issues Vol. 28, No. 2 (1994), 367–377. It is vital to temper intellect (computation), as defined by Berardi in this article’s epigraph, with consciousness and sensibility (ethical judgement). Doing so shows us one pathway towards enacting the right to the model.

Places Journal is supported by an international network of academic partners and generous readers.

Please Subscribe or Donate.

About the Author

Simon Rabyniuk

Simon Rabyniuk received his Master of Architecture degree from the University of Toronto in 2019; the previous year, he received the Howarth-Wright Travel Fellowship supporting research on the integration of civilian drones into city skies. His thesis received an ARC-King Medal for architectural research. Prior to these studies, he was a principal at the research, art, and design studio Department of Unusual Certainties. He currently teaches courses relating to the technology of urbanism at the Daniels Faculty of Architecture, Landscape, and Design at the University of Toronto, OCAD University, and George Brown College in Toronto.