Crichton's 'Aliens Cause Global Warming'

In 2003, Michael Crichton delivered a lecture at Caltech titled "Aliens Cause Global Warming". This lecture has since circulated widely and its arguments often come up in critiques of climate science. An excerpt was even published posthumously in The Wall Street Journal. I wish to comment briefly on two particular sections of the lecture: consensus and climate models.

On Consensus...
Although he never defines exactly what he means by "consensus", Crichton's use of the term seems to rely on a limited and largely incorrect view of the scientific process. He attempts to portray the idea of consensus as irrelevant in science, and assumes that all of science boils down to clear, proven facts: "Science, on the contrary, requires only one investigator who happens to be right, which means that he or she has results that are verifiable by reference to the real world. In science consensus is irrelevant." Unfortunately, complex scientific topics do not fit this "neat and simple" model at all. Climate science is an amalgamation of many different investigators, examining many different lines of evidence that lend strong support to the conclusion that humans are modifying the climate system. This extensive field of research cannot, and will not, ever come down to one investigator being right or wrong. In this case, the vast majority of climate scientists agree on several main conclusions that only emerge with the synthesis of many independent lines of evidence. The reality of climate science is that no single piece of evidence can solely support or refute the conclusions that have emerged. This is exactly why the concept of consensus is important - consensus emerges to indicate what is currently understood and what is not. Rather than being irrelevant as Crichton suggests, consensus is a real and fundamental part of science.
Crichton then argues that scientific consensus has a poor track record and points to examples where the "consensus" was eventually overturned. The implication is that the consensus isn't always correct. But none of his examples are good analogues for the current consensus on climate change. Previously held views on fever, diet, and geology were not based on a robust, mechanistic understanding of empirical evidence, and often rested upon long-held assumptions that had not, or could not, be formally tested. For example, the dominant paradigm in geology going into the 19th century was largely drawn from Biblical accounts of Noah's Flood. In the mid 19th century, Charles Lyell challenged this view with a theoretical framework of more gradual change based upon extensive observations. In the 20th century, Alfred Wegener's theory of continental drift was indeed controversial at first, not because the consensus failed to "acknowledge what any schoolchild sees", but because he did not have a robust mechanisms for why the continents should move, nor was there any evidence that the continents were indeed moving. With the accumulation of more independent lines of evidence and the development of a more mechanistic theory, the consensus around plate tectonics was eventually formed (even before the movement of continents was empirically observed). In his simplistic portrayal of consensus, Crichton fails to see that the processes that formed the current paradigms in medicine and geology (and overturned the old ones) are same processes that have formed the consensus on climate change. The consensus used to be that human activity is too insignificant to alter the climate. Science advances and things change.
Finally, Crichton states that "consensus is only invoked in situations where the science is not solid enough." He revisits his idea that science boils down to clear "yes or no" answers by suggesting that it's obviously unnecessary to invoke consensus around something like E = mc2. According to Crichton, this is apparently real science with a right answer. Again, this is not very analogous to something like climate science, but it's also not a valid example in its own right. Einstein originally proposed the idea of mass-energy equivalence largely based on theoretical ground and without strong direct empirical evidence. Support grew around the theory, but if you wanted rock-solid "proof", you probably would have had to wait almost three decades until the discovery of the positron (the observed conversion of energy into matter and antimatter). While it currently may seem silly to think of there being a consensus around E = mc2, there was a time when this "fact" needed to build up its own evidence and consensus. And in contrast to climate science, this particular case does largely boil down to a relatively straight-forward test that can convincingly support or refute the theory. Crichton's broader implication that all of climate science needs to be 100 percent "solid" is really a strawman argument: an impossible version of reality that is therefore easy to pull apart. Instead of indicating science that's not solid, the consensus on climate change informs us which aspects are solid and which are less so.
On Climate Models...
Crichton's only direct critique of climate science in the lecture is the "overt reliance that is being placed on models". This is a version of the common claim that the evidence for global warming is based on computer models. Unfortunately, this claim is not true. Climate models are one among many tools and methodologies used in climate science. They are extremely useful for certain applications (especially in predicting future climate), but the conclusion that humans are significantly contributing to climate change through greenhouse gas emissions would remain robust in the absence of computer models. To hold them up as the keystone of climate science is another strawman argument.
Even still, Crichton's critique of climate models is off base, starting with his confusion of weather (which is chaotic) and climate (which is not). Most importantly though, climate models do not attempt to "predict the world of 2100," but instead project different climate scenarios based on a range of assumptions of what the world might look like in terms of population, greenhouse gas emissions, land use patterns, etc. Their assumptions, caveats, and uncertainties are openly discussed. In the absence of computer models, we can still make valid probabilistic projections of mean global temperatures based on future greenhouse gas concentrations (it will get warmer). However, without climate models we cannot have any real sense of regional differences, particularly with regard to precipitation. Climate models thus provide a very important means of examining potential local impacts of climate change, but they are far from the the keystone of consensus that Crichton implies.
Aside from his discussion of consensus and climate models, Crichton builds an indirect case against climate change science based on guilt by association, although he never convincingly demonstrates association, nor guilt. It is a tremendous logical leap of faith to conclude that the search for extraterrestrial radio signals and one group of scientists' research on the potential impacts of nuclear war somehow invalidate decades of climate research by thousands of individuals. Overall, Crichton's lecture is primarily supported by his rhetorical skills, not his arguments.
Another take on Crichton's speech with more examination of some of the facts he presents can be found here.

Other presentations by Alden Griffith can be found at Fool Me Once.

Posted by Alden Griffith on Tuesday, 8 February, 2011

Creative Commons License The Skeptical Science website by Skeptical Science is licensed under a Creative Commons Attribution 3.0 Unported License.