-3

I read a blog post that attempts to show that the size of the territory of a country is purely random and therefore bears no major historical significance.

The argumentation (simplified a bit) is as follows:

  • If we plot sizes of countries on a logarithmic plot the distribution is linear:

enter image description here

(Each bar is a country, the Y axis is the size of countries in km2; blue bars are "large" countries, orange bars are "medium" countries and grey bars are "small" countries. Countries are labelled "large", "medium" or "small" because, according to the blogger, the exponential distribution has different arguments depending on whether the country is below or above 10000 km2, so in fact there are two distributions, as shown by the straight lines in the plot)

  • Therefore, if plot on a linear plot, the distribution will be exponential
  • Exponential distribution occurs if the events are purely random
  • If it was not random whether a country gained or lost a territory we would expect a Pareto distribution instead

This seems to make sense, but... at the same time it is very counteintuitive. Does it follow that a country competently governed is equally likely to lose a territory than a failed country? The blogger himself says this is not the case, but he claims that a country that is governed competently is likely to be governed far worse after a few generations and will start losing territory. One very profound conclusion after another...

Plus, the blogger himself claims that his findings are contrary to established knowledge.

Is it purely random if a country gains or loses a territory?

gaazkam
  • 2,635
  • 3
  • 15
  • 25
  • The blog post isn't in English, nor are the graph axis legends and labels, so it isn't clear what is proposed. – Weather Vane Oct 09 '21 at 17:49
  • @WeatherVane Sadly. Which is why I tried to summarize it in English. Not sure if translating it would be appropriate, but - perhaps - I should quote and translate key parts? – gaazkam Oct 09 '21 at 17:50
  • 3
    "Exponential distribution occurs if the events are purely random" whilst true, not all white birds are swans. There are many claims here. – Jiminy Cricket. Oct 09 '21 at 18:17
  • 17
    Until such time as they define what they mean by “purely random”, this sounds like complete nonsense (to me, a mathematician with expertise in probability theory). It’s common to describe such claims as “[not even wrong](https://en.wikipedia.org/wiki/Not_even_wrong)”. But this is based on your description, which I cannot verify since I can’t read the original post. – Dan Romik Oct 09 '21 at 18:22
  • What is the horizontal axis? – DJohnM Oct 09 '21 at 22:17
  • 1
    I second @DanRomik's point: "purely random" is not a thing. From your other context, what the blog is probably trying say is that land area is the product of many small independent pieces of noise? However, I'd then expect a lognormal distribution rather than exponential? Since, there are an almost unlimited number of explanations that would result in very different distributions, this whole exercise is rather pseudo-scientific. You really can't reach any general conclusions with any certainty based on this limited evidence. At best, you can say an extremely specific hypothesis is wrong. – Sherwin Lott Oct 10 '21 at 00:08
  • 2
    (The hypothesis that it's not "purely random" is woefully underspecific not least because "purely random" is not a defined term.) Also, Enrico Fermi's quote applies here: "with four parameters I can fit an elephant, and with five I can make him wiggle his trunk." – Sherwin Lott Oct 10 '21 at 00:11
  • @gaazkam One last critical point: the log of an exponential random variable does not produce linearity. (More generally, no continuous random variable with support over the real line has a linear cdf). That's simply not how random variables work. The whole premise of the underlying logic is false. – Sherwin Lott Oct 10 '21 at 02:14
  • 1
    I added my vote to close for two reasons: (1) it isn't clear that anyone (apart from the claimant) believes this. This appears to be an obscure blog, and the comments don't appear to be supportive. (2) The limited amount of translated argument is close to gibberish... and yet the OP says it seems to make sense. There is a very great risk that we are tackling unintentional strawmen until we can get a clearer explanation of the claimants argument. – Oddthinking Oct 10 '21 at 10:26
  • 1
    I tried reading a Google Translation of the whole article. There is a lot of context missing (including the exclusion or giant territories), but when it gets to the key argument, it does seem to be as equally confused as the quoted part in the question, so I am feeling more comfortable that the OP has made a valiant effort here. – Oddthinking Oct 10 '21 at 10:31
  • @Oddthinking Regarding notability: I guess you may be right... All I can say is that this particular blogger is kind of infamous in some circles... like I said, for example he once debated professors on the internet on the topic of the global warming. He might not have many supporters, but he does relentlessly argue for his ideas on other forums. – gaazkam Oct 10 '21 at 10:53
  • (Not in this case however - I admit I found this particular article on his blog, and not elsewhere) – gaazkam Oct 10 '21 at 10:54
  • I understand that for many people his ideas may be *obviously* BS, but for many others... there are hard to debunk. I tried debating him a few times on some of his outlandish claims and... i lost every time :( – gaazkam Oct 10 '21 at 10:55
  • 3
    @gaazkam There is no way of winning an argument with those who have fully inoculated themselves against logic and reason. – David Hammen Oct 10 '21 at 14:12
  • Problems I identify: (1) What is the counterclaim? That country size is quantized? (2) What should the influence be? As in, is the ordering in the graph purely from largest to smallest? What's the point of ordering them in this way? – Jan Oct 11 '21 at 12:35

1 Answers1

2

The conclusion of the linked blog post says (via Google Translate):

Contrary to popular belief, no "natural boundaries" exist. All borders between states, with the exception of really great countries, which occupy a large part or even the whole (Australia) of continents, are purely random.

NASA talks about this in: When Rivers are Borders

Rivers also serve as borders in several African countries: Benin, Nigeria, Senegal, Mauritania, and Niger in West Africa; the Central African Republic, the Democratic Republic of the Congo, Tanzania, and Mozambique in Central Africa; and Botswana, Zambia, Zimbabwe, and Namibia in Southern Africa. In Europe, Spain, Portugal, France, Germany, Poland, Norway, Bosnia, Romania, and Ukraine are just some of the many countries with international border rivers.

The patterns are not random. Smith explained that the continents where European powers established colonies and exerted great influence tend to have more borders defined by rivers because European explorers, cartographers, politicians, and diplomats found rivers to be a convenient way to divide territories. Asia, in contrast, has fewer river borders than other continents (16 percent) because European influence and colonialism was more limited. The exception is South Asia, where Great Britain and France remained colonial powers into the 20th century.

Which shows that the blog's assertion is nonsense.

Weather Vane
  • 7,653
  • 1
  • 32
  • 42
  • 2
    The blog's assertion is wrong because NASA says something else? Umm... how do I put it... this particular blogger is pretty infamous for challenging the claims of established science. In his mind he has shown he is right, if scientists say something else *scientists are wrong*. I bet that this case is no different: he says in his post "Te doniosłe wydarzenia historyczne, szczegółowo opisywane i analizowane w podręcznikach szkolnych, akademickich i solennych opracowaniach naukowych, okazują się nie mieć, per saldo, absolutnie żadnego znaczenia." – gaazkam Oct 09 '21 at 19:13
  • Which I'd translate into English in this way: "*These profound historical events, thoroughly described in school and academic handbooks and in meticulous scientific papers turn out to have no meaning whatsoever*" – gaazkam Oct 09 '21 at 19:15
  • 7
    The author has cherry-picked a section of the graph that is roughly linear, and found arbitrary reasons to exclude the parts that are not, and dressed it up in a long ramble of supposition, backed up by a quote from the Bible. – Weather Vane Oct 09 '21 at 19:16
  • That this blogger is in conflict with mainstream science is not news... on other occasions he was even debating profesors in the internet on global warming, which he claimed was not anthropogenic. My problem with this blogger is that... however outlandish some of his claims might be... he does support his claims with lots of plots, charts and formulas... and reasoning that at least has semblances of being rational. Often I'm stuck with questions like... "*so he challenges mainstream science and he must be wrong... but *how* is he wrong?!*" – gaazkam Oct 09 '21 at 19:19
  • You say: "*The author has cherry-picked a section of the graph that is roughly linear*" Umm... the correlation coefficient of 99,7% is "*roughly*" linear? "* found arbitrary reasons to exclude the parts that are not*" - only 6 largest countries; he himself notes that his findings do not apply to countries that are this large – gaazkam Oct 09 '21 at 19:22
  • 2
    I disagree: he supports his case with quasi-science like *This rule, however, has its limits of applicability. When the area of ​​a country exceeds approximately 5 million square kilometers, the braking mechanism is activated, as already mentioned, and the probability of losing a territory becomes greater than the probability of joining another. The second border is at the area of ​​10 thousand. square kilometers, where a specific type of phase transition takes place and the exponent of the distribution changes abruptly.* His limits are chosen from the data, not established from his "reasons". – Weather Vane Oct 09 '21 at 19:40
  • "*His limits are chosen from the data, not established from his "reasons".*" True. He's split the chart into two parts, so that each part has an exponential distribution. Is doing so pseudo-scientific? – gaazkam Oct 09 '21 at 19:49
  • I think you are trying to draw a red herring across this answer, which states that the blog's conclusion about natural boundaries is **wrong**. This isn't a matter of opinion, or proof, or anything like that, but fact. There **are** many rivers which form political boundaries, besides the Rio Grande. – Weather Vane Oct 09 '21 at 20:06
  • 4
    I will point out that both NASA and the author can be correct: if the areas between rivers are randomly distributed, then the area of countries occupying those areas can also be randomly distributed. Note that I am saying "can" not "are" because the conditions are necessary, but not sufficient, to show correlation. – bishop Oct 09 '21 at 20:51
  • @gaazkam Re *He's split the chart into two parts, so that each part has an exponential distribution. Is doing so pseudo-scientific?* An all-caps and emboldened **YES** is the only appropriate answer to that question. – David Hammen Oct 10 '21 at 08:49
  • 3
    @DavidHammen (and gaazkam): "Well actually... No" He splits the data into THREE parts, then removes the inconvenient outliers of Australia and larger before plotting it. – Oddthinking Oct 10 '21 at 10:20
  • @WeatherVane "*I think you are trying to draw a red herring across this answer*" - I'm just playing the devil's advocate here. I sincerily think that arguing the wrong case can be a good way to learn how and why exactly it is wrong. – gaazkam Oct 10 '21 at 10:41
  • 3
    @bishop you don't need to go as far as the bolded parts. The first howler is *Contrary to popular belief, no "natural boundaries" exist.* This is untrue, even if you disqualify the undefined 'really great countries' in the blog. Actually not undefined: the blog defines them *from the data*. – Weather Vane Oct 10 '21 at 15:22