Dropsafe

by Alec Muffett

Science is the Art of the Demonstrable

Sometime back in 1996 I was in a lovely cafe in Palo Alto, having breakfast with Diffie; Whit slid a stapled printout of an academic paper across the table, and asked me what I thought of it. The paper was: “Transgressing the Boundaries: Towards a Transformative Hermeneutics of Quantum Gravity”.

I remember the scene quite distinctly, because it was unusual for Whit to ask my opinion of a paper, and I so wondered if it might be something interesting like crypto, network security, or time travel – Whit being an avid classic SciFi nut. I skimmed the first 4 or 5 pages or so, dredging up old physics and the horrors of classic QM from (then) a decade previous — when I hit something, stopped reading, and dropped the paper on the table.

“It smells like bullshit”, said I.

“Why do you say that?”

“Well he’s saying that the laws of physics are relative not merely to your frame of reference, but also to your perception; relativity doesn’t work like that.”

“Well, there may be something in it…” said Whit — but with his standard-issue twinkle in his eyes, making me wonder if I was having my leg pulled.

We changed topic; back then I was not connected enough to know about Sokal Affair, and I had no access to NPR so I didn’t know that on the day of its publication in some Post-Learnist journal the paper had been splashed by its author as a complete fabrication designed to wind-up the postmodernist establishment.

The Wikipedia article on the Sokal Affair led me however to a further article on the Science Wars of which I was equally ignorant — I was busy fighting a different war at the time.[1]

Quote:

Until the mid-20th century, the philosophy of science had concentrated on the viability of scientific method and knowledge, proposing justifications for the truth of scientific theories and observations and attempting to discover on a philosophical level why science worked. Already Karl Popper had begun to attack this view. He denied outright that justification existed for such concepts as truth, probability or even belief in scientific theories, thereby laying fertile foundations for the growth of postmodernist attitudes.

I know people who fête Popper, thus this surprised me because the only thing I know about Popper is the Doctrine of Falsifiability, which prima facie appears to be a useful tool for science:

Falsifiability or refutability is the logical possibility that an assertion can be shown false by an observation or a physical experiment. That something is “falsifiable” does not mean it is false; rather, that if it is false, then this can be shown by observation or experiment. The term “testability” is related but more specific; it means that an assertion can be falsified through experimentation alone.

For example, “all men are mortal” is unfalsifiable, since no finite amount of observation could ever demonstrate its falsehood: that one or more men can live forever. “All men are immortal,” by contrast, is falsifiable, by the presentation of just one dead man.

This (to me) echoes the right approach to science:

  • We have a theory of Gravity: it works, it can be demonstrated to work, really well, and if somebody eventually discovers Cavorite then several hundred years of work will require “substantial modification”.

  • We have a theory of Evolution: it works, it can be demonstrated to work, really well, and it has been extrapolated plausibly that all life on Earth has sprung from this mechanism; nobody has yet found a plausible counterexample to evolution, so until someone finds a bunch of architectural diagrams from God in a coal seam, I’m happy to go with the Evolutionists.

  • We have a theory of God: however the existence of a God is not falsifiable; you cannot prove that there is no God since he may be hiding under a pebble, or behind a leaf, or in a sunbeam or something – and because you cannot prove a negative, not to mention the highly debatable question about what being a “God” entails — all Atheists are stuck with being 99.999999…% agnostics, what Dawkins calls a “Tooth Fairy Agnostic”.

Returning to that Popper assertion: a little digging into the citations and I found myself seething about people playing word games. It’s the sort of thing you have to understand in order to cope with spin-doctors, but when it’s applied to well-understood scientific theories it can upset me. One minute they appear to be explaining “proof by induction” and the next minute they appear to be deriding, er, “proof by induction” — and I’m not sure that they know the difference between the two; although I side with Hume that that inductive can be risky, hence why I am a “Tooth Fairy Agnostic” Atheist.

I have no axe to grind pro-Popper or anti, and no particular reason to defend him, but I have the impression that the above “The Science Wars are Popper’s fault” citation is yet more wilfully misapprehended bullshit. I presently don’t wish to research the matter further because if you follow the train of clicks too far you end up on pages covered in pictures of Australian Black Swans, somehow declaring them to be both cosmically meaningful and unobvious, when I instead think of them in their reality as an agricultural pest.

However there’s another aspect left out of my “theory of Gods” bulletpoint above: demonstrability, which I’d like to return to for a moment.

Gravity is easily and repeatedly demonstrable. Evolution is easily and repeatedly demonstrable. In general, scientific theories which are very strong are easily and repeatedly demonstrable.

So now to the present day, and we have a bunch of people at the Climate Research Unit, who have some data that has been severely hacked-about, who work in an academic echo chamber, and whose data got released, I suspect by dint of Murphy’s Law rather than “hacking”.

That we are experiencing a period of pretty wild climate variation is demonstrable fact, and anthropogenic (ie: fossil-fuel-burning) climate change may or may not be the cause; but for reasons above I am not going to inductively try to link the two, because the Earth’s climate is a huge and chaotic system that almost defies measurement, and certainly defies long-term prediction in much the same way that evolution defies prediction.

And that a bunch of scientists make money by trying to predict long-term trends in the behaviour of a chaotic system?

Mmmm…

—
[1] Incidentally, the Crypto Wars really deserve a Wikipedia page

⊞

Comments

6 responses to “Science is the Art of the Demonstrable”

  1. Robin Wilton

    If you haven’t read it, you might find “Wittgenstein’s Poker” an entertaining read. From the starting point of a notorious, brief clash between Wittgenstein and Popper in Cambridge, it goes on to explore the various social, academic and personal differences between the two.

    Reply
  2. Simon

    “And that a bunch of scientists make money by trying to predict long-term trends in the behaviour of a chaotic system?”

    This “parody” is rather below you Alec.

    Even chaotic systems have parameters that affect the shape of the attractor, and thus make gross properties of the system predictable.

    Is the proportion of greenhouse gases in the atmosphere likely to be such a parameter for climate?

    Have the predictions of climate scientists from the 1980’s been far from the mark of what was subsequently observed?

    What would the odds of them being right with such an odd prediction if the system were chaotic and merely fluctuating around a set average global temperature?

    Reply
  3. alecm

    @Simon – it’s not a parody, it’s a statement of uncertainty; yes chaotic systems swing around a mean, in the same way that measured over a sufficiently long period of time a random walk can appear to have both a source and a destination rather than averaging-out to leave you where you started.

    I doubt that atmospheric temperature will be 1000C one day and 42C the next, but I am pretty confident that any claim for what weather I’ll face this December will be more likely wrong as it is right, unless the error-bars are sufficiently big.

    As for the 80s, I dunno because I wasn’t hacking NCAR at the time; but I remember this from the 70s if it helps: http://en.wikipedia.org/wiki/Global_cooling

    Reminder: I am still not saying that AGM does not exist; I am just saying that I don’t trust the scientific establishment who work with dodgy data in a small, incestuous community; that they are trying to predict the future of the climate is moreso amusing to me, akin to how I can retrospectively view Michael Fish circa 1987.

    Reply
  4. Simon

    Yes, but they aren’t forecasting weather for December, they are forecasting climate. There is a difference. What they are predicting is the error bars, not the value.

    Since you don’t know how accurate climate forecasting is by your own admission it rather weakens your comments about climate science. If science is about the demonstrable, one should know what has been already demonstrated instead of throwing peanuts from the stands.

    Sure there are issues with climate data, but the basic premise of climate change is based on established science and observation. That CO2 is a significant greenhouse gas, established in the late 19th century, and that we’ve increased the concentration of atmospheric CO2 (observed).

    One needs data to validate the climate models, and to establish historic climate behaviour, but the basic principals are physics and not dependent on historic data.

    Did you even read the Global Cooling article at Wiki? I’m thinking you didn’t.

    Reply
  5. alecm

    >Since you don’t know how accurate climate forecasting is by your own admission it rather weakens your comments about climate science. If science is about the demonstrable, one should know what has been already demonstrated instead of throwing peanuts from the stands.

    Fair comment; and I should also enjoy greater confidence and pay more attention if there was more surety that the folk in the middle of the ring were not peddling snake-oil or religion.

    And yes, any astronomer who’s read up on Venus knows about greenhouse gases.

    Re: the Wiki site, I find it very telling that the old graph for Global Cooling says “temperature change” http://en.wikipedia.org/wiki/File:Global_cooling.jpg

    …and the new one says “temperature anomaly” http://en.wikipedia.org/wiki/File:Instrumental_Temperature_Record_(NASA).svg

    Perhaps you’ld like to explain what the anomaly means in terms of an unbiased original hypothesis, and how it is calculated? I presume it’s not just average global temperature in Kelvin with errorbars, else they would have said. I presume it’s differences from some notional mean, but would like to know how the mean is calculated and whether the mean is an ascending, flat or descending line/curve that is quietly “corrected for solar input” or some other fudge-factor.

    You worked at the Met Office. Maybe you can say.

    If for instance they had a descending mean and the absolute temperature remained constant, you’ld have a first-order derivative that would be rising. Is that what the “anomaly” Y-axis represents? Could you please point me at a graph of average absolute temperature over time?

    Above all I reiterate that you’re treating me like an antagonist of global warming; whereas all I am saying is a) from observation I don’t trust the community b) from having downloaded it off bittorrent and taken a look at it myself, I don’t trust the data, c) because of (b) I don’t trust the statistical analyses and anyway d) I don’t believe in the concept of “normalcy” as applied to the entire Earth, and so disbelieve in measuring diversions from the normal.

    Maybe you can be an ambassador for the concept, rather than attack?

    Reply
  6. Simon

    The CRU data is not the only data used in climate change, so items (b) and (c) are largely irrelevant to the AGW debate.

    (a) You don’t trust the community. Well I’m sure 97% of climate scientists will be surprised but it is science it doesn’t solely rely on trust. But you can read most of their research online for free, or summarised in the IPCC reports.

    I could be an ambassador, but I wasn’t a climate scientist at the Met Office and you’d be better of reading it from climate scientists in the IPCC reports. My understanding of climate, and how it works, was pretty poor before the current crop of denialists turned up. With a few points of better understanding where the Hadley center lectures I attended at the Met Office covered an issue.

    As regards the diagram the footnotes explain what the anomaly is compared to (1961-1990 mean temperature) – although the original source has 1951-1980 mean. According the NASA GISS the best estimate for 1951-1980 mean land-ocean surface temperature is 14.0C – so simply add 14.0 to the anomaly for an absolute mean temperature (if that means anything).

    However I did use to process observational data for the Met Office so I know how far removed these averages are from the reality of day to day collection of observational data. Much of the period in question is also covered by satellite data, which whilst they have their own calibration issues, generally have better coverage. The satellites of course measure radiance, and the process of reconciling these records with other observations and the models was not smooth. But I’d expect the satellite record to be less influenced by things like urban heat islands- which have to be corrected out of the land station data.

    http://en.wikipedia.org/wiki/Satellite_temperature_measurements

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *

More posts