Substack, Jennifer Marohasy, Aug 12, 2026
In 1922, American journalist Walter Lippmann published Public Opinion, a book that remains one of the clearest accounts of how ‘the average man’ forms some understanding of the world. Lippmann’s central insight was simple and unsettling: the public cannot respond directly to reality; it is too complex. Rather we are programmed to pictures that form in our heads — simplified, incomplete, and often stereotyped images that can become potent symbols.
His point is not that these individuals are stupid, but that the complexity of the modern world makes it impossible for the average citizen to have well-informed opinions on most issues from personal knowledge. They must therefore rely on the simplified pictures and stereotypes supplied to them. He treats this as a structural problem of democracy rather than a personal failing.
These pictures, once established, shape what the public believes to be true and, in turn, what government policies become possible.
More than this, I know from bitter experience that choosing to speak out when the picture is fake science, and forcefully against this, comes at a huge personal cost once the moral imperative has been embedded along with the picture. That my two brothers and my sister won’t discuss these issues with me — all the fakery that ends up becoming part of their decision making process though we disagree, of course, on this too — it caused me some distress earlier this week after my mother again asked when I would next visit with her in Cairns. They won’t even have a cup of tea with me, not any more, not my brother and sister who live in Cairns.
Lippmann was not a detached observer of the process either; he had practised it. During the First World War he worked inside the American propaganda apparatus, helping to shape the messages and choose the pictures that would secure public support for that war. When he later wrote his famous book, he made no fundamental distinction between the wartime propaganda and the ‘manufacture of consent’ in peacetime.
These mental pictures, according to Lippmann, need not correspond closely to reality. They can be selectively framed or even largely invented if they serve the strategic interests of government: aligning popular feeling with the policies that have already been decided upon by those on the inside, with access to what Lippmann referred to as the levers of power. This for me conjures up images of a lever pulled down to begin the printing of the afternoon edition of one of Rupert Murdoch’s newspapers. Perhaps when he sided with Margaret Thatcher against the coal miners, or in support of the Falklands war; how this was framed by the newsprint.
I saw the great efficiency of the process in operation between 1991 and 2006 when the establishment decided to fundamentally change how we think about climate change. For sure Margaret Thatcher and Rupert Murdoch were complicit also in the seeding of this idea and the rollout of the program of propaganda.
In the first IPCC Assessment Report published by Cambridge University Press in 1991, a schematic diagram showed a climate history for one thousand years with the warmest period on record about one thousand years ago. It was based on the work of Hubert Lamb and his colleagues at the Climatic Research Unit (CRU) at the University of East Anglia. It is likely quite accurate for the British Isles.
By the third IPCC Assessment of 2001 also published by Cambridge University Press, this original reconstruction had been replaced by the now-famous ‘hockey stick.’ Pre-industrial centuries were flattened into a relatively level shaft; the Medieval Warm Period and Little Ice Age were gone; and the late-twentieth-century rise became a sharp, anomalous blade. The temperature series had been remodelled in accordance with the predetermined manufactured consensus. This consensus certainly took off in a more meaningful way about the time Margaret Thatcher decided to close-down coal mining and set up Kyoto, and there were others who could see advantage in it.
Before this computer-based remodelling, the past record of climate change was assembled from a variety of different types of evidence including chronicles, grain price records, radiocarbon dating, pollen analysis and vegetation history, tree rings, ocean bed deposits as well, of course, as meteorological instrument records.
Going back to just the 1980s, the consensus amongst the geographers, drawing together evidence from these very different disciplines, was that there was a period about 1,000 years when harvests were so good and summers so long that communities got together and build grand cathedrals across western Europe. That was known as the Medieval Warm Period. Then it got cold again. Of course, if we choose to live an honest life that involves a capacity to reflect some understanding of what went before and also to care about how it might evolve for the next generation (perhaps even above all else), then it is worth having some knowledge of the Medieval Warm Period and to consider the bits and pieces of evidence that would indicate it was likely about a degree warmer back then at least in the British Isles.
Yet this knowledge of the cathedral building during the Medieval Warm Period, supposedly understood by anyone with much education in the West, was all collapsed, to accommodate the emerging manufactured consensus of catastrophic human-caused climate change. The hockey-stick image became the dominant stereotype — repeated in the IPCC Summary for Policymakers, amplified in popular media and even now in school textbooks, in short, treated as settled fact — hardened into the accepted reality. Yet it is all fake, like so much of our modern lives that so many in family and past colleagues do not want to talk about.
The same psychological susceptibility, combined with the very nature of modern liberal democracies and the way they function, has allowed a complete misrepresentation of a thousand years of temperature history to instil a civilisational-wide fear of climate change.
This is just one of the perhaps lesser-known examples of the general working of how propaganda affects all our lives. It can be easily extended to our understanding of the government’s response to COVID-19 and even the war on Iran as I have been explaining recently.
In fact, what I am describing here about climate change remains mostly hidden. You, dear reader, have some experience of it because you choose to be informed and you have a mostly curious approach to the world.
According to the chatbots, the hockey stick is still the consensus, it is still the official mostly correct historical reconstruction of the last thousand years of climate change. It is me that happens to have an alternative view, in fact one that is anchored in personal knowledge as I have spent so many years putting the pieces together for an historically accurate reconstruction of Australia’s temperature history. Most recently I was revising my charts showing the temperature history of Sandy Cape lighthouse, based on the original temperature measurements as recorded by the lighthouse keeper.
My interest in historical temperature data came from the need to feed historically accurate arrays of data into the artificial neural networks that my husband John Abbot began working-up back in 2011; it is no coincidence that this was the same year that Brisbane flooded. My argument to him at the time was that the flooding was entirely avoidable because the setup across the Pacific considering the key climate indices suggested there would be heavy rains over the coming summer. Wivenhoe dam should never have been kept full of water in the lead up to the 2010-11 summer; it was specifically build for flood mitigation.
Yet all of this was ignored by those managing Brisbane’s water infrastructure, and the meteorologist, who already believed their own propaganda which back then also came in the way of definitive pronouncements by BOM’s David Jones and the everywhere man Tim Flannery, who both predicted endless drought with our reservoirs destined to never fill again.
Meanwhile, away from the very public fake narrative, from the period 2012 to 2017, John Abbot and I produced a series of peer-reviewed papers using his artificial neural networks configurations and my data arrays, to more skilfully predict monthly rainfall up to eighteen months in advance for specific locations. Then there was our last joint paper that considered the last thousand years of temperature change with a forecast.
Specifically, for our 2017 GeoResJ paper, we took several available long temperature proxy records often cited in the technical literature including from tree rings, lake sediments, and some multi-proxy series, to see what would happen if we broke each record into a set of natural oscillatory cycles that best described the temperature variations up to 1830, and then fed the same components into an artificial neural network, a form of machine learning combined with AI, to see what the forecast might be for the twentieth century.
Based on the natural cycles embed in the series, the forecast in every instance, across the six series, was for some warming through the twentieth century. When these projections were compared with the actual proxy and instrumental temperatures after 1880, the largest average difference across the six regions was only about 0.2 °C. From that residual we estimated an Equilibrium Climate Sensitivity — the warming expected from a doubling of atmospheric carbon dioxide — of roughly 0.6 °C. That figure is far lower than the values produced by the models relied upon by the IPCC (typically around 3 °C) and much closer to estimates derived from laboratory spectroscopy.
We concluded that a large part of the warming observed in the twentieth century can be accounted for by the continuation of natural cycles, leaving if any, only a modest role for human-caused greenhouse gases.
For sure, John Abbot and I had taken an altogether different route from the usual climate modelling. Instead of starting with specific assumptions about carbon dioxide and running complex general circulation models, we began with actual long temperature proxy records — tree rings, lake sediments and also multi-proxy series from both the Northern and Southern Hemispheres.
We used spectral analysis to break each record into a set of natural oscillatory cycles (sine waves) that best described the temperature variations up to 1830, before industrial emissions purportedly became significant. Those component cycles were then fed into the artificial neural network, which first learned the patterns of natural variability before making a forecast for the twentieth century.
Our 2017 GeoResJ paper is a very neat study and innovative to boot; it is perhaps the first such study essentially using spectral analysis combined with AI to forecast twentieth century temperatures based on arguably the best publicly available thousand yearlong temperature series.
I was not prepared for the intense social-media backlash that followed, this was after the late Christian Kerr republished a short note about it at the Australian Spectator Online that was subsequently picked up by other online media. High profile mainstream climate scientists (including Gavin Schmit then director of NASA’s Goddard Institute for Space Studies) chimed in, mostly full of derision. This all happened in August 2017, and the lunch planned for my birthday that year was cancelled, that was after Graham Readfearn phoned about for his defamatory article that followed in The Guardian newspaper.
Schmidt’s main public criticism was that one of the proxy series we used (the one drawn from Moberg) appeared to have its time axis shifted by about 35 years, so that the period we treated as the twentieth century actually ended around 1965. He argued this meant our projection missed the stronger recent warming. Our response was straightforward: we digitised the series exactly as it had been published and presented in the peer-reviewed literature we cited. We did not alter the dates or the values; we took the published figure and the associated data as they stood. Any apparent discrepancy therefore reflects how that series was already being shown in the mainstream climate science sources available at the time, not a manipulation on our part.
A broader objection, raised by Schmidt and others, concerned the use of artificial neural networks themselves. Critics argued that training a machine-learning model only on pre-industrial temperature cycles and then projecting it forward is invalid once a new factor — rising carbon dioxide — enters the system. In their view, the network cannot properly ‘know’ about a forcing that was absent from its training data, so any residual warming it fails to explain cannot safely be attributed to human influence. This has been characterised as a methodological disagreement about whether statistical pattern-learning from past natural variability can be used to test the size of the anthropogenic contribution going forward. Of course, it can unless you are thinking backwards, unless you fundamentally have a problem with letting the historical pattern influence your thinking about what could occur in the future!
At the same time there was a formal request to the journal’s editor to have our paper retracted with immediate effect. Preventing that happening proved all time consuming and involved writing detailed justifications, always hoping the editor would stand firmly on the side of fairness and rational argument rather being intimidated by the mad men massing against us, including Gavin Schmidt.
Then, shortly afterwards, the journal was closed by well-known publisher, Elsevier. The editor was out of a job; the journal ceased to exist. Retraction Watch covered the closure at the time and noted that the journal had drawn attention because of the paper by Abbot and Marohasy after it was cited in a Breitbart column.
The several high-profile climate scientists that attacked us very publicly online and in private correspondence to the journal editor, never went to the trouble of publishing any formal criticism in the peer-reviewed literature.
The emotional intensity of their reaction presumably reflected our threat to what they knew was a politically useful stereotype that could not easily withstand honest scrutiny.
So, rather than any calm assessment of our data and methods they used attack dogs like Graham Readfearn and Gavin Schmidt in an attempt to shame us both. For sure, at that time several friends and colleagues indicated they now questioned my motivations and a well known defamation lawyer indicated in private correspondence to John Abbot that he would not defend climate deniers, clearly we are nutters.
All of this is consistent with Walter Lippmann’s theory of the pseudo-environment, developed in his 1922 book. Lippmann argued that in complex modern societies people do not respond to reality directly, but to simplified pictures — stereotypes — that are supplied to them and that come to stand in for the world itself. Once a dominant picture is established in this case the hockey stick image of the last thousand years, contradictory evidence is resisted or marginalised, not through detailed technical rebuttal, but because it threatens the shared image that in this case, underpins the case for treating climate change as a civilisational emergency that justifies rapid, large-scale intervention not to mention the careers of so many not only within science but also now within our financial systems.
In policy terms the easily ridiculed hockey stick has helped legitimise national and international net-zero targets (especially the 2050 goal); aggressive renewable-energy mandates, subsidies and feed-in tariffs; accelerated phase-out of coal, oil and gas; carbon pricing, emissions trading schemes and regulatory restrictions on fossil-fuel investment; and large-scale public spending and central-bank/financial-regulatory pressure to redirect capital toward ‘green’ technologies.
On the investment side, the scale of the financial commitment is enormous. Mainstream estimates of the cumulative capital required for a global energy transition consistent with net-zero pathways by mid-century typically fall in the range of $75–150 trillion depending on the organisation, the end date and what is included — power generation, grids, storage, electrification of transport and industry, etcetera.
The hockey-stick image persists in part because it justifies that scale of mobilisation even if it is fake.
The replacement of this fake reconstruction with something more realistic will not be easy. John Abbot and my work from 2017 was a valuable attempt to show a better way. The paper has just 46 citations nearly ten years on.
It is disappointing that our paper, and this work more generally was not supported, or promoted by organisations like the Institute of Public Affairs. The work was funded by the B.Macfie Family Foundation through the IPA, but I relied almost entirely on the enthusiasm of the late Christian Kerr for the relatively small amount of publicity that it got back in 2017.
There was specific funding to continue this type of work at the IPA provided by Bryant Macfie, but the Board chose to discontinue and prematurely terminate both my and John Abbot’s contracts in March 2025. Scott Hargreaves indicated that this type of primary research, particularly combined with my plans to integrate it into a whole new theory of climate resilience had no place at his think tank, at least not going forward.