is impossible in principle, and given the stated reluctance of scientists to discuss findings that are highly uncertain, how could such reports even be attempted? And if global heating constitutes an emergency situation, how should the IPCC respond to it? Evidence suggests that that the IPCC in its assessments has on balance not tried to address those problems but has maintained a conservative attitude when dealing with less understood but plausible high-impact outcomes (Brysse et al. 2013). It may have split into three working groups, of which the first tackles the more tractable physical basis of climate change, but even here there are many processes we do not understand, such as the collapse of ice sheets that can lead to sea-level rise of several metres (Grégoire, Payne and Valdes 2012). In fact, the IPCC’s reported range of 0.3–1.1 metres of sea-level rise by 2100 (IPCC 2019b) is much lower than the 2.4-metre rise that has been identified as the worst-case scenario (Bamber et al. 2019). All of the IPCC’s scenarios of future man-made climate change tend to have one thing in common: the assumption that change happens smoothly and gradually, without any major disruptions. The possibility of such large-scale disruptions – such as ice-sheet collapse, dieback of the Amazon rainforest, or large-scale carbon release from permafrost – is acknowledged by the IPCC. But when the IPCC calculates a possible safe amount of CO2 that could still be emitted, the possibility of such events actually happening is either not taken into account, relegated to footnotes, or labelled ‘low confidence’ in such a way as to make it seem as if those events need not be greatly worried about (Rogelj et al. 2018). But as we have stressed, and as those relatively few (but growing in number) climate scientists attest who speak out openly about the way their fears for the future exceed what they can conservatively prove, that is getting things the wrong way around. To be an effective early warning system, the IPCC or something like it would need to focus pretty strongly on making stark, and as rapidly as possible (rather than after years of slow deliberation), the risks inherent in allegedly low-probability but undoubtedly high-impact events or cascades.
Consider the IPCC report on climate change and land (IPCC 2019b). This assesses the question of how future climate change may impact yields. However, the studies rely on past observed changes or on models that are validated only by past change, and as such cannot take into account that the future climate in many crop-growing regions may be entirely novel, with combinations of extreme heat at high humidity not seen on earth for millions of years (Burke et al. 2018). There is also no assessment of what might happen if the climate becomes fundamentally unstable, as was the case before the adoption of agriculture. Instead, the assessment is limited to linking past climate warming to changes in crop yields, and the use of models to extrapolate those changes into the future. The possibility of major pest outbreaks of a scale not seen in the past, or large-scale drought hitting several major agricultural production regions at the same time (multi-breadbasket failure: Kornhuber et al. 2019), is not taken into account.
Such simple extrapolation of recent observations to make inferences about the future is also used in economic assessments. The extraordinary fact that the 2018 economics Nobel prize was given to William Nordhaus for his work with the DICE model (Nordhaus and Sztorc 2013) illustrates the scale of acceptance and dominance of this approach. Nordhaus’s model includes a function that takes global mean-temperature change as input, and with it predicts the damage caused by climate change as a percentage of global gross domestic product (GDP). The data points used to motivate this damage function are derived either from observations of the GDP of different US states, or from quantification of damages of isolated effects, such as the building costs for dykes, or the costs of health care when there is an increase in the incidence of malaria (Tol 2009). If states with a hotter climate have a lower GDP, then the first method would infer that a warmer climate leads to ‘damage’ (Mendelsohn et al. 2000). As Richard Tol (Tol 2009) states, the authors of all included studies come from no more than three closely related groups of scholars and were published more than 20 years ago. The results of this and related analyses led William Nordhaus to conclude that 4°C warming is optimal for human welfare (Nordhaus 2018), and the IPCC to conclude that the impact of human-caused climate change will be small for most economic sectors (Arent et al. 2014).
It is important to recall that this exercise in ‘foresight’ is based on a number of unstated and unproven assumptions, including: first, that the welfare of a given place in today’s interconnected world is independent of the climate of the rest of the world; second, that damages from climate change can be added up sector by sector, and that interconnectedness and its impact on vulnerability of the modern world can be ignored; third, that changes across time can be inferred from changes across space, for example, if the climate of Massachusetts changes to that of Florida, its GDP per capita would change accordingly; fourth, that it does not matter how quickly dangerous climate change occurs; and fifth, that there are no known thresholds or tipping points that could amplify the impact of climate change/chaos (Lenton et al. 2019). A recent analysis of this body of work also points out that large parts of the economy were excluded from the beginning as supposedly not dependent on climate, as well as other biases (Keene 2020).
A case for the virtue of scientific ignorance
How is it possible that a world of 4°C climate heating is judged to be uninhabitable or leading to societal collapse by some scientists (Steffen et al. 2018), but talked about in a professional matter-of-fact way by others, even seen as ‘optimal’ (Nordhaus 2018)? While the flaws and limitations of the economic approach are obvious, how can we understand the past reluctance of the IPCC to include worst-case scenarios and disruptive events or to accept the limits of its approach largely based on computer models?
Part of the answer may lie in the inherent tendency of professionals, such as climate scientists, to support the status quo (Schmidt 2000), something that clashes with the fact that averting a climate catastrophe requires society to change in radical ways, as advocated by some climate scientists (Rockström 2015). It is important to note that the change advocated is usually not radical enough to include a fundamental shift of the role of the scientists themselves. The possibility that climate science itself may have contributed to the current crisis, by overstating its knowledge and sketching out an often-theoretical road to solving the climate crisis, is seldom entertained. Evidence suggests, however, that climate policy makers and the IPCC have often co-created scenarios in which future, unproven technological solutions have provided justification for delaying action (McLaren and Markusson 2020).
It is obvious that climate scientists have done a generally pretty magnifficent job in trying to detail the way the world has changed and is now under threat. It is obvious that some climate scientists have had to put up with appalling attacks from ignorant or well-funded citizens, media or ‘think tanks’ (sic). It is obvious that a limited but increasing number of climate scientists have become near-heroes for speaking out about the extreme gravity of the threat hanging over our planetary ecosystem. And yet . . . Could climate science, insofar as it has tended not to be properly precautionary, have even, on balance, contributed to a widespread denial of the severity of the threat from global overheating? It is an awkward but necessary question to ask. One experienced climatologist explains that one way ‘we, as scientists, have contributed to the crisis concerns the excessive rationalization of a threat. In other words – we switch off common sense and produce scientific results borne of idealized models – be they mathematical or intellectual.’ He continues:
As citizens we all know the difference between a politician’s words and deeds and we are all painfully aware of . . . a very real risk of a 4 or higher degree of warming. And yet, the IPCC’s various assessment reports have repeatedly relied on highly idealized so-called integrated models that know and admit nothing of these things, and therefore [are] easily bent to produce results that fly in the face of common logic. (Knorr 2019)
Despite the existence of such criticism from within the climate science profession, at the time of writing it has proved to be largely unable to acknowledge the limits of its own approach to knowing about the climate crisis. It appears that the scientific community, while it continues to own the issue of global heating, does not own or abide by the fundamental importance of the precautionary principle for the functioning of human society. There is, however, a radically different approach that can be traced back almost two and half millennia to the ancient Greek philosopher Socrates,