Skip to content

Our statement on climate models…


Our statement on climate models (Joseph Norman, Rupert Read, Yaneer Bar-Yam). We have *only one* planet and need to learn to live with imperfection of models.

Source: Timeline Photos – Nassim Nicholas Taleb


  1. Richard Lilton wrote:

    “Even a risk with a very low probability becomes unacceptable when it affects all of us.”

    That’s a central premise in your argument, but it’s not exactly true. To see why, consider that a global zombie apocalypse is improbable, yet possible. It’s a risk, and it affects us all. But should we therefore invest large sums of money in zombie defenses to defeat the risk? Of course not. We accept this risk, even though it affects all of us, contrary to your principle above.

    That principle is therefore false, and your argument resting on it deflates like an old balloon. We cannot set aside the reliability of models when deciding whether to spend large sums of money to defeat the risks of climate change. Probability matters, and so too do the models. Without them, it’s reckless to sink so much wealth into defeating what may be a phantom risk.

    Saturday, May 23, 2015 at 10:27 am | Permalink
  2. Rupert Read wrote:

    Richard – global zombie apocalypse is an _unreasonable_/_insane_ worry, one is that is absurd to even imagine looking for evidence for or against. Climate apocalypse is not. That is a categorial difference.

    Friday, November 6, 2015 at 2:33 am | Permalink
  3. Carlos wrote:

    Richard, I thought the same at first, but the key in their reasoning is this “The scale of the effect must be demonstrated to be larg e enought to have impact. Once this is shown, AND HAS BEEN the burden of proof of absence of harm is on those who would deny it”

    Monday, January 11, 2016 at 12:47 pm | Permalink
  4. If models were accurate,our response could be measured in appropriate quantities and qualities. Since models are not very accurate, Taleb argues we must not discount the issue to zero (inaction) but consider what to do if the worst might happen.

    But therein lies the challenge. All responses have immediate and continuing costs. And they too might well have unknown consequences. In the short term, they have known consequences for the economies of nations.

    In fact, by Taleb’s reasoning, we must also be prepared for global cooling–unlikely, but certainly within the scope of climate response to increased reflection of energy, snowfall, etc.

    The responses and plans for responses should be divided into two categories: adaptation and management.

    Like the Dutch who have been adapting to rising sea level for centuries, we now have many more ingenious ways to adapt.

    Management of the climate is geo-engineering, terraforming and the such. Whether this means neutralizing ocean acidification or pumping GHGs out of the atmosphere, we should be doing the research and controllable experiments now. Developing geo-engineering will prepare us for any significant direction of climate change.

    Sunday, April 24, 2016 at 6:06 pm | Permalink
  5. John wrote:

    Very well put Wallace! I agree, we should be building a tool set using controllable experiments.

    Sunday, April 24, 2016 at 9:24 pm | Permalink
  6. Rupert Read wrote:

    Yes, we should also consider the possibility of global cooling, which (ironically) can be induced by global warming. But there is a strong precautionary argument against that consideration taking the form of leaping to hubristic scenarios of geo-engineering, which point in the opposite direction to the ‘via negativa’ suggested in our statement, above.
    See e.g.

    Monday, April 25, 2016 at 6:45 am | Permalink
  7. David Glynn wrote:

    John and Wallace, how do you plan to control these experiments? The global climate is a complex system so this is like claiming you can do randomized controlled trials on a single patient in order to decide on his future treatment.

    Monday, August 8, 2016 at 2:55 pm | Permalink
  8. John Mauer wrote:

    By the arguments in the short article, precautionary measures should be taken with any risk of catastrophic effect. (By that argument, we should all see doctors regularly to avoid the risk of debilitating disease.) This argument totally avoids the view that small amounts of CO2, as we are currently producing, is a positive for our environment. I’m disappointed in the inconsistency of your arguments.

    Friday, July 28, 2017 at 3:42 am | Permalink
  9. Deoxy wrote:

    This sounds all nice and sciency, but it is actually a huge case of begging the question:

    “Once this is shown, AND IT HAS BEEN, the burden of proof of absence of harm is on those who would deny it”

    So, all we have to do is accept as proven that which is the whole point of contention, and everything becomes obvious…

    Monday, July 31, 2017 at 12:28 pm | Permalink
  10. Jeremy Batchelder wrote:

    If Wallace Kaufman wants to “use” Taleb’s reasoning, they should probably understand it first. I recommend they start with The Black Swan and then proceed to Antifragile.

    1)In Taleb’s book Antifragile, he outlines his belief that Mother Nature and natural systems/processes are the most timed tested mechanisms we have. We should aim to design our society around these systems that have been around for many millions of years and survived. If they hadn’t succeeded the tests of many years, they wouldn’t still be here today.

    2)In order to do this, we must conduct policies of subtraction. We must “subtract” or remove the things that humans have created in the last 500 years that have caused damage to these natural processes. Building bigger walls and pumping chemicals in the ocean won’t do this.

    3)By geo-engineering natural processes, we are putting ourselves into even more risky situations. It is nearly impossible to tell what will happen if we try to neutralize the whole ocean. We do not know how these chemicals will affect complex systems, and we would have to wait many years to find out.

    Friday, August 11, 2017 at 6:59 pm | Permalink
  11. David L wrote:

    JEREMY BATCHELDER, thank you, best comment so far. Removing stressors from a organism by trial and error is a more robust strategy than adding medicine to a system by trial and error. The problem is not the adding of medicine but the unpredictability of the reaction of complex systems.

    you can only predict when an event is predictable. So the model is only valide for a certain range called normality

    in NORMAL cases we can easily modell the effects of stressors like toxicants to a complex system (we just dont use good modells, because we stick to linearity :D)

    but it is not the normal case which you have to consider, it is the unlikely (not to unlikely) event. And it makes no sence to imagine which event it could exactly be…we just dont know. Black Swans do not exist when there is no subjekt which suffers or benefits. So we should better be on the safe side <– that is the point

    you can say "fack future", there is no problem with this. Its not that fragility is wrong or something. If we will fail because we collectively "decide" to be fragile than something else, "the whole" will benefit. :)

    Monday, November 13, 2017 at 10:55 am | Permalink

Post a Comment

Your email is never published nor shared. Required fields are marked *