DR ELISA VECCHIONE
The Poetry of Science
Restoring the Role of Imagination in Mathematical Modelling
INDEPENDENT SCHOLAR FELLOW: APRIL 2019 – SEPTEMBER 2019
Elisa’s main research interest concerns the relation between science and democracy. She is particularly fond of looking into scientific methodologies – she did it with statistical analysis and risk assessment in the field of genetically modified products, she is doing it again with Integrated Assessment Modelling and impact assessment in the context of climate change – as she believes scientific methodologies provide an important lens for reading society, including its approach to knowledge, uncertainty, order and control. The pursuit of this interest was not an easy journey as it deeply touched upon her identity as a researcher and human being. Be aware, therefore, that the following lines may look like her story was linear, but in reality, it was full of discontinuities, bifurcation, turning back, jumps, accelerations, breaks, and of course, happy moments. This in which I am writing, is indeed a happy moment.
Elisa started her research path in the field of environmental economics and innovation policy at University of Turin. She later pursued her researches in the field of science governance. In 2009, she obtained a PhD in Law, Economics and Institutions at Collegio Carlo Alberto and University of Turin (Italy) with a dissertation on the governance of genetically modified organisms (GMOs) in the US and the EU. Her analysis centred on scientific, political and legal discourses over GMOs risks in the two countries as a way to determine their respective framing of technical evidence and regulatory legitimacy. During this time, Elisa was visiting scholar at the Law School at Cornell University (Ithaca, NY) and the Centre of Management Research at Ecole Polytechnique in Paris (France).
She later joined the Centre for Sustainable development at Sciences Po Paris, where she worked on the role of contested expertise in sustainable development and transnational environmental regulation. This time she focused on climate change and looked at contested expertise ‘from within’: she analysed scientific discourses of climate change as written in the language of Integrated Assessment Modelling. She used that language to extract their embedded values and derive policy-relevant conclusions. During this time, she also became Associate Researcher to the Groupe of Pragmatic and Reflexive Sociology at EHESS Paris.
She then moved to London at the end of 2013 and yet joined another project on the use of scientific knowledge in policy making. The GRIP-Health project at the London School of Hygiene and Tropical Medicine looked at the political and institutional factors shaping the use of evidence in health policies in different country settings. Elisa spent most of her time working on Ghana and Ethiopia and conducted fieldworks within. She is very much grateful for these two experiences, from both a human and professional perspective.
After her maternity leave, Elisa moved to UCL School of Public Policy as Teaching Fellow in 2017. In the meantime, she also became Distant Learning Tutor for two modules at LSHTM, MSc Public Health. In August 2018, she decided that the only way forward with her research career and professional happiness was to do what she wanted to do, with no compromise or surrogated projects. She worked non-stop for around ten days, putting together all the myriads of notes and unfinished papers she has been writing since the end of 2012, and applied for the ISRF Independent Scholar Fellowship.
This project aims at moving forward the discussion about how to improve the use of mathematical models in policymaking. It takes on from the critiques and challenges associated with integrated assessment models of global change and raised in different disciplinary communities, to then gather as many diverse disciplinary positions and rethink the construction and use of models. This initial phase of the study aims to construct a platform of interdisciplinary collaboration. Given the immense potentialities but also risks involved in the use of this mathematical modelling and given the absence of a scientific community on the matter, it is urgent to establish a coherent discussion about how models can be at the same time scientifically sound and policy-relevant. A technical fix alone would not solve the problem: mathematical modelling cannot be fixed in the sense of getting the model right, nor getting the model right is the way to affect the preferences of its users. Therefore, this project proposes to debunk the valueleaden character of models by undertaking a more ‘poetic’ approach to science: the goal is to install a kind of ‘poietic’ process into modelling, so as to extract the hopes, values and meaning that the modeller engages when facing the inescapable limitations to her knowledge in modelling the future. To make poetry out of science, mathematical models will be interpreted as ‘stories’ of the future, and accordingly analysed as narratives. The reason behind this methodological choice lies in its contribution to knowledge. Building on Hayden White’s theory on history and narrativity, narrative analysis allows revealing the consciousness of the storyteller as well as the moral of her story. This project suggests that making the moral character of mathematical models explicit can greatly improve public discussion about the meaning and the urgency of policy action.
The Research Idea
Mathematical modelling (MM) is a technique for representing how different systems behave through the means of mathematical equations. The use of mathematical equations seemingly confers MM the status of science in the classical sense of predictive science, but its properties betray this conclusion. The complexity of the systems modelled make MM closer to a crafting art, in which the modeller creates stories about the future by interrelating different factors in a coherent way; and at the same time, justifies that a certain history of the future can be made by following a specific sequence of events.
With few exceptions, these properties of MM are largely dismissed among modellers due to the discomfort created in the validation of their work and, by transmission, in its uptake for policy decision. In order to overcome this discomfort, improve the scientific quality and the policy-relevance of MM, the role of creativity and imagination should be reconsidered in the science-policy paradigm. The project will thus take on from Ravetz’s plea for rediscovering the metaphors contained in mathematical models. It will move on by applying narrative analysis to mathematical models as in the use made by Hayden White’s in historiography. By building a parallel between the historiographer and the modeller, the project will thus restate the question raised by White: can narrative, along with its imaginary and poetic character, provide the basis for the scientific method? And what would the implication be for the collaboration between science and policy?
The use of computer modelling for policy decisions is widespread, but the quality of their outcomes does not live up to their potentialities. By interrelating many factors at the same time and coherently, mathematical models do not make predictions; however, they provide lessons about how present decisions prepare us to possible futures. In spite of these properties, MM is largely performed within the modern paradigm of science for policy control, responding to an insatiable quest for objectivity and evidence. This situation affects models’ quality and the chances of policy manipulation and misuse. The post-normal science (PNS) community has embraced the technical problem and the potentialities of MM at the same time. Ravetz proposes to rediscover the metaphorical styles embedded in models so as to release their learning potential. Saltelli and Giampietro propose a new methodology called ‘quantitative storytelling’, which audits specific choices in selected mathematical models in order to identify choices that are technically incorrect or absolutely implausible and unjustified. Outside PNS, Morgan and Wise similarly argue in favour of the explanatory power of narratives in science. The use of storytelling is indeed a promising way to harness the learning potential of MM by conferring a more idealistic character to science and possibly, following Kahan, a more favourable cognitive and public engagement with it. But how would this be possible in practice? Unexplored is indeed the question of how storytelling can translate into a public tool of learning and decision-making.
The use of narrative analysis for MM promises to improve their scientific quality and policy relevance. The use of MM in Integrated Assessment of global changes, for instance, is marred by the uncertainty surrounding too many factors, including data, parameters and the choice of mathematical frameworks. A classical scientific perspective sees these uncertainties as engendering the risk of a ‘cascade’ of errors compromising the outcomes of the model. Hence the need to ‘manage’ uncertainty. Instead, a narrative perspective will ‘embrace’ these uncertainties and treat them as building blocks in the construction of future scenarios. Therefore, a narrative approach would invite a reflexive discussion of these uncertainties among modellers, pushing them to reveal the values and hopes they mobilise when they choose among different parameter values and mathematical frameworks. In sum: how uncertainty makes the difference in each of the story told in the model and what moral is associated with how the story ends. Narrative has a power to translate knowing into telling by fashioning human experience – in this case, the thought experience in the future – into a form assimilable beyond cultural and epistemic differences. Cognitively speaking, narrative is more engaging than any other technical or philosophical discussion. Therefore, narrative can improve interdisciplinary collaboration among modellers, and between modellers and policymakers. The restoration of values within the scientific practice of modelling will provide a new bridge with policymaking and especially, with democratic policymaking. How? By explicating the public meaning of suggested models and opening it to public debate.
This project applies the concept of ‘meta-history’ by White to MM in order to put forward the historical and narrative characters of MM and analyse them accordingly. The project brings together two ideas: that of mathematical models as being governed by a kind of ‘historical’ science, in which events can never be predicted, however always retro-dicted; and that of mathematical models as building ‘stories’ of the future. The application of the concept of meta-history to MM also responds to the same concern by White of demonstrating that historical narratives depend on the consciousness of the storyteller about the values, meaning and norms of her time. An extensive literature has discussed the valueleadenness of MM, but in no case a methodology has been formulated that extracts such values, if not for suppressing them and rendering science its objective character. Instead, this project aims at restoring the imaginative character in science. Like White’s historiographer revealing more or less implicitly her consciousness about the present ‘contest’ in which history unfolds, the modeller brings her own consciousness in the modelling exercise by shaping her story of the future. In both cases, the imagination of the storyteller about, respectively, the past and the future plays out in building events into a specific story. In modelling, such imagination is encapsulated in the modeller’s choices over deeply uncertain parameters and variables, making uncertainty the scientific and moral site for discussing the contest in which we are preparing for the future.
This project will take climate change as case study, as this still represents one of the most contested issues in science policy collaboration. The project will proceed along three main phases:
1. A theoretical framework will be constructed for applying narrative analysis to mathematical modelling in the context of science policy debates. This will bring together the disciplines that have employed narrative analysis in the study of science policy relations, including economic history, sociology and philosophy of science, and public policy, and use Hayden White’s approach to narrative analysis for bringing coherence to the whole framework.
2. A literature review will be made of the most debated and uncertain parameters in IA models. The review will draw from the Fifth Assessment Report (AR5) of the Intergovernmental Panel on Climate Change, which provides the most comprehensive assessment of climate change knowledge. This review will be supported by survey questions sent to selected modellers across the three working groups involved in the AR5 process. One respondent will be selected out of each group according to her availability for interview. During individual interviews, the respondent will be asked to indicate the authors s/he is in major disagreement with across the working groups, with the aim to eventually interview one of them. The final target is to conduct six interviews in total.
3. Narrative analysis will be performed on the models built by the six selected authors based on the new theoretical framework and supported by the interviews.
The three phases of the project as described in the methodology, are conceived to retrieve and bring to order, a set of personal ideas and studies about the use of modelling in policymaking. These ideas are partly contained in a working paper of 2012 (“Deliberating Beyond Evidence: Lessons from Integrated Assessment Modelling”), and partly in a series of conference papers produced for different disciplinary audiences.
Phases 1 and 2 need not be consequential. Phase 3, instead, will occur after the previous phases have been completed.
The project is expected to be implemented within a period of seven months, with phase 1 and 2 completed by month 5, and phase 3 by month 7.
Each phase will have one output corresponding to the preparation of an article on:
1. the relevance of Hayden White’s theory for improving the making and use of mathematical modelling in policymaking;
2. the relevance of uncertain parameters in IA models to reveal the role of values within scientific discussions and debates;
3. the translation of selected debates into narratives about the future.
Depending on the collaboration of the selected interviewees, these articles will be submitted either independently or as a special edited issue in one peer review journal – e.g. Futures.
4. A fourth output is envisaged: the elaboration of a mini-compound on article 1 and 2 to be circulated at the IPCC, in order to share information on how to anticipate possible debates within and across the three working groups, and how to overcome them.
This projects aims at setting up a network of collaboration between researchers working on mathematical modelling from different disciplinary perspectives, including economic history, sociology and philosophy of science, and public policy. The project, indeed, responds to an interdisciplinary vocation of improving the dialogue between different scientific communities and this, for the benefit of public debate. Ideally, the study of the use of mathematical models for policy would tackle more directly the problem of improving democratic decision-making, namely by designing new forms of stakeholders engagement based on the use of narratives. Therefore, the proposed study goes in the direction of two hopes, which need not be exclusive but are actually synergetic: one, preparing and leading an application for new collaborative grants, such as European COFUND or ERC; two, consolidate my knowledge and build around it a new professional identity with which I can perform the function of mediator of scientific debates, and/or facilitator of public debates over scientific research and innovation.