Why am I no longer an anarchist

[brimstoneSalad]

a simple lack of control in a potential confounding variable is a valid criticism

So, do you think that, to the argument I presented in my paper Etimologija Karašica, the neuralbeans'es response "You are ignoring the possibility that the nouns (river names being nouns) in the Croatian language have a significantly lower collision entropy than the rest of the words in the Aspell word-list for the Croatian language." is a proper response?

Teo, I'm not an expert in linguistics, but as far as I can tell, yes, it seems an appropriate response. That is probably the most common scientific criticism any publication gets. It is not Neuralbeans' job to prove it, it's your job to show that this confounding variable is not an issue because you're making the claim by submitting the paper.
Now if Neuralbeans DID prove it, that would be a much stronger rebuke, but just pointing out a plausible confounding variable that is not accounted for is enough to undermine the conclusions and require more work on your part to support them more fully.

To me it seems like a blatant ad-hoc hypothesis.

It's a null hypothesis. A critic must assume the conclusion is false, and look at plausible explanations. Technically speaking the criticism is "ad-hoc", as in made to purpose, that's the point of criticism. But from my uninformed reading it was not inane and did not seem to make any wild assumptions characteristic of what would be criticized as ad-hoc arguments in support of a conspiracy theory or other quackery.

Maybe if the Croatian language had a Swahili-like grammar (where nouns can only start with 18 prefixes, while verbs can start with whatever sounds the phonotactics allows), that would be a proper response, but the way things are it isn't.

Well then there's your response. Can you edit the paper with an explanation of that, and come to some kind of number on the probability of that alternative being correct? Could you choose some number of other noun categories or whatever at random and look at the variation in collision entropy to find a P-value?
That would strengthen your case.

Do you think that, to the Eratosthenes'es experiment to measure the curvature of the Earth, the typical Flat-Earther response ("You are ignoring the possibility that the Sun is tiny and close.") is a proper response?

No, that's one of the axioms. In a paper, it would start by stating assumptions like that. The experiment does not purport to discover the location of the sun, it takes it as a given due to vast amounts of pre-existing external information.

[teo123]

It is not Neuralbeans' job to prove it, it's your job to show that this confounding variable is not an issue because you're making the claim by submitting the paper.

Then we have vastly different ideas about how burden of proof works.
When The Mad Revisionists claim the Moon doesn't exist, we give them a proof that the Moon exists that you see it. By giving them a proof, we shift the burden of proof onto them. Now when they claim the Moon is a hologram, the burden of proof is on them.
When Flat-Earthers claim the Earth is flat, and we give them a proof that ships disappear bottom first, we shift the burden of proof onto them. Some Flat-Earthers claim the ships disappearing bottom first is an illusion caused by waves, but then the burden of proof is on them to prove that doesn't happen when there are no waves.

Furthermore, who is being more rigorous here, mainstream onomastics or me?
Mainstream onomastics claims that this k-r-pattern in the Croatian river names is a coincidence, but provides no statistical model of a language that suggests there is a non-negligible probability of such a pattern occurring by chance.
I show that it follows from basic information theory (collision entropy and birthday paradox) that the probability of that k-r pattern occurring by chance is between 1/300 and 1/17, and I provide an explanation for that pattern (that *karr~kurr was the Illyrian word for "to flow", for example, that Karašica comes from Illyrian *Kurr-urr-issia, "flow-water-suffix").
Complaining that I am not being rigorous enough is, well, besides the point.

[teo123]

No, that's one of the axioms. In a paper, it would start by stating assumptions like that. The experiment does not purport to discover the location of the sun, it takes it as a given due to vast amounts of pre-existing external information.

As far as I understand the history of astronomy, if you asked Eratosthenes how he knew the Sun was far away, he would have responded with "Well, assuming that makes it easy to calculate the size of the Earth.". If you asked him how he knew the Earth was round, he could give you a relatively-rigorous response such as "Ships disappear bottom first as they recede over the horizon.", but for the Sun being far away... He lived centuries before Aristarchus, and even the Aristarchus'es ideas weren't really mainstream science all until the Renessance. Eratosthenes couldn't give you a rigorous response to the objection that perhaps the Sun is close.

[brimstoneSalad]

Then we have vastly different ideas about how burden of proof works.

You have some work to do on your understanding, then.

When Flat-Earthers claim the Earth is flat, and we give them a proof that ships disappear bottom first, we shift the burden of proof onto them. Some Flat-Earthers claim the ships disappearing bottom first is an illusion caused by waves, but then the burden of proof is on them to prove that doesn't happen when there are no waves.

The counterclaim that it's an illusion needs to be plausible, which it is not. That's not how waves work. In the very least I would expect a diagram to show how it actually works with waves when intuitively that makes no sense when visualizing line of sight.

IF ships disappearing from the bottom were literally the only reason we suspected the Earth to be round, then challenges to that would surely be taken more seriously as they probably were in the era of that discovery.

Furthermore, who is being more rigorous here, mainstream onomastics or me?
Mainstream onomastics claims that this k-r-pattern in the Croatian river names is a coincidence, but provides no statistical model of a language that suggests there is a non-negligible probability of such a pattern occurring by chance.

Chance is the null hypothesis, so you have the burden of proof to show it's something else.
However, because you've selected river names specifically out of a grab bag of many hundreds or more of comparable noun categories, you need to do better on your P-value to make chance implausible.

https://xkcd.com/882/

I show that it follows from basic information theory (collision entropy and birthday paradox) that the probability of that k-r pattern occurring by chance is between 1/300 and 1/17,

Other noun categories:
Town names
Hill names
Valley names
Stream names
Bridge names
People names
Tool names
Animal names
Business names
Whatever names

Can you find 17 general categories?
If so, your observation that rivers are an outlier (with a 1/17 chance of being coincidence) is not exceptional.
If there are 17 similar categories that you could have assessed instead, but intuitively you chose rivers because you saw the pattern, then it's not chance that you chose the one that was an outlier (being an outlier was due to chance).

You need a P-value vastly disproportionate to the sample you drew from.
Not even 1/300 seems terribly convincing given that issue.

and I provide an explanation for that pattern (that *karr~kurr was the Illyrian word for "to flow", for example, that Karašica comes from Illyrian *Kurr-urr-issia, "flow-water-suffix").
Complaining that I am not being rigorous enough is, well, besides the point.

I will assume your explanations on your P-value in isolation is credible, I don't know anything about the linguistics. But incidentally (due to your pattern finding intuitions being good) cherry picking an outlier when you are familiar with a wide array of samples you could have picked instead skews your actual P-value in real world practice.

If you had no knowledge of Croatian whatsoever and blindly picked "river names" and only ever "river names" out of a hat with a million categories and found this association on your first try, that would be different.

[brimstoneSalad]

As far as I understand the history of astronomy, if you asked Eratosthenes how he knew the Sun was far away, he would have responded with "Well, assuming that makes it easy to calculate the size of the Earth.".

That's unlikely. He probably had good reasons to assume this beforehand, as I understand it this was already common knowledge since at least Aristarchus (some 100 years earlier, although the distance calculation attempts were under-estimations).
It is also not hard to show with a small desktop model (like a large cardboard box with precise measurements) that the sun's rays are effectively parallel, which supports the basic premise required for the experiment.

Again it seems like you're just jumping to wild assumptions about scientists and the belief that they're operating on ignorance.
If that were his only reason, then obviously that would be an easy criticism (and a valid concern).

Eratosthenes couldn't give you a rigorous response to the objection that perhaps the Sun is close.

I think he probably could have, but if he could not have then he probably would have admitted that, and that his findings are contingent on that possibility. I can't see why he would have bothered undertaking it if he did not already know the sun's rays are parallel on Earth.

[teo123]

You have some work to do on your understanding, then.

Let's see if we can reach of a consensus via discussing it.

The counterclaim that it's an illusion needs to be plausible, which it is not. That's not how waves work. In the very least I would expect a diagram to show how it actually works with waves when intuitively that makes no sense when visualizing line of sight.

Right. I think that the best response to the Flat-Earthers claiming that ships disappearing bottom first is caused by waves is something along the lines of: "I assume you know basic trigonometry, try to calculate how high the waves would have to be hide a ship that's 5 km away from you. I'll give you a hint: they would have to be slightly higher... than your eye level. Not to mention the same thing obviously happens when there are no waves.".
But, even so, the burden of proof is on one claiming there is some weird illusion going on, to prove that such an illusion is at least possible. It's difficult to prove that sunrays appearing to converge is an illusion (rather than proving the Sun is close), but it's easy to prove that such an illusion is at least possible (for example, with railroads).

IF ships disappearing from the bottom were literally the only reason we suspected the Earth to be round, then challenges to that would surely be taken more seriously as they probably were in the era of that discovery.

Right. If ships were disappearing bottom first, but the same constellations were visible from all places on Earth, that would be difficult to explain.

Chance is the null hypothesis, so you have the burden of proof to show it's something else.

My friend, to me it seems like that k-r pattern in the river names is probably the only statistically significant pattern in the Croatian names of places.
Mainstream onomastics considers that s-r that repeats in the Croatian toponyms such as Sirmium (ancient name for Srijemska Mitrovica), Serota (ancient name for Virovitica), and Serapia (ancient name for Bednja) to have been the Illyrian word for "to flow", but I think that pattern is not statistically significant. The primary reason for assuming that pattern is not a coincidence appears to be the existence of the Indo-European root *ser meaning "to flow". Now, how they make that belief compatible with the belief that Illyrian is the direct ancestor of Albanian (where there was a very old sound change of Indo-European *s to 'gj', happening even before the satemization) is beyond me, but that's irrelevant here.

However, because you've selected river names specifically out of a grab bag of many hundreds or more of comparable noun categories, you need to do better on your P-value to make chance implausible.

Well, I used to think everybody would agree that the p-value of that k-r pattern is probably closer to 1/300 than to 1/17. However, my information theory professor Franjo Jović didn't agree with that. He told me that he thinks that the p-value is closer to 1/17 than to 1/300 because he thinks that not a lot of entropy in the Croatian language goes to morphology (I estimated the collison entropy of morphology to be at most 1.572 bits per pair of consonants, but I couldn't give a lower bound).
But even so, I don't think the exact numbers are that much relevant. Mainstream onomastics uses a methodology of which one of the basic principles is that the etymologies from languages we know a lot about (Croatian, Latin, Turkic, Celtic...) are more probable than the etymologies from languages we know little about (Illyrian...). And it's not obvious what the mathematical justification of that principle would be. Not only that, that methodology appears to be incompatible with the information theory. It gave the result that this k-r pattern in the Croatian river names is a coincidence (Mainstream onomastics claims that Karašica comes from Turkic "kara sub" meaning "black water", that Krapina comes from Germanic word for the carp fish, that Korana comes from the Celtic word *karr meaning "stone", that Kravarščica is named after the village Kravarsko rather than vice versa, that Krbavica is related to the archaic Croatian word "hrbat" meaning "mountain", and that "Krka" is the Illyrian for "oak tree" previously referring to some town rather than to the river.), while basic information theory says that such a pattern is unlikely to be due to chance (that the probability of such a pattern occurring by chance is somewhere between 1/300 and 1/17). The right thing to do is to throw away the mainstream methodology and start using some methodology that doesn't run right in the face of the information theory, right?

I will assume your explanations on your P-value in isolation is credible

Well, it's hard to tell. A forum.hr user called DarkDivider told me that my supposed Proto-Slavic form of the river name Karašica, *Kъrъrьsьja (from Illyrian *Kurrurrissia), is impossible because Proto-Slavic phonotactics didn't allow four syllables with yers to be consecutive. I asked him where he had read that, but he didn't answer me. Furthermore, as far as I understand it, mainstream onomastics is also proposing etymologies which imply Proto-Slavic forms with four yers. The town name Cavtat is widely thought to derive from the Latin phrase "(in) civitate", and the Proto-Slavic form was therefore *Kьvьtъtь. So, overall, I'd say it's credible.

It is also not hard to show with a small desktop model (like a large cardboard box with precise measurements) that the sun's rays are effectively parallel, which supports the basic premise required for the experiment.

Really? You can tell that way that the Sun is further away than around 5'000 kilometers, which would be the explanation of the Eratosthenes'es experiment if the Earth were flat?

[teo123]

It's hard for me to objective now because I'm invested into the belief that that paper "Etimologija Karašica" is right. I wrote that paper, before that I did the measurements and numerical calculations described in the paper, I published that paper in two journals (Valpovački Godišnjak and Regionalne Studije, and Valpovački Godišnjak required me to heavily edit it before it could be published), I made two YouTube videos about it, I shared that paper with many people in real life (my Algorithms and Data Structures professor, my Computer Science professor, my Communications professor, two of my Information Theory professors, my Computer Architecture professor, my friend who dropped out of his PhD in computer science but is still interested in it, my high-school teachers...) and all of them claimed my arguments seem compelling to them... If that paper is critically flawed, then it would only be ethical to contact all of those people I misled and explain that to them. That would definitely destroy my reputation.

[teo123]

That is probably the most common scientific criticism any publication gets.

Well, you know, on another forum, I claimed that one of the basic principles of social sciences is to never doubt the statistics provided to you by the government. Then @Red told me he thinks that's not the case. I responded by citing Katherine Hirschfeld as she said "Social scientists frequently cited favorable health statistics supplied by these regimes (Cuba...), without critically looking at the ways these were created and maintained by state power.". Then @Red told me:

Just because something frequently happens doesn't mean it's supposed to happen.

Yeah, I think the same response applies here.

[teo123]

Chance is the null hypothesis, so you have the burden of proof to show it's something else.

I am not sure that's a good principle. What do you think about the Levy's objection to Semmelweis that perhaps the puerperal fever is seasonal and that the mortality rate dropped coincidentally because of seasonality? If chance is the null hypothesis, we should presume Levy is right. But it seems obvious that Levy was being unreasonable here.

[brimstoneSalad]

Let's see if we can reach of a consensus via discussing it.

I don't really have time to help you with that Teo. You need to do more work on your own.

My friend, to me it seems like that k-r pattern in the river names is probably the only statistically significant pattern in the Croatian names of places.

Seeming is not adequate, and cherry picking "place names" as a category to search in is not enough.

But even so, I don't think the exact numbers are that much relevant.

You don't think they're relevant because you don't want to think that, because that's bad for your conclusion.

Really? You can tell that way that the Sun is further away than around 5'000 kilometers, which would be the explanation of the Eratosthenes'es experiment if the Earth were flat?

You can determine that they are essentially parallel. Measuring the angle with more precision would probably require a larger and more carefully constructed apparatus. You know basic trig, so you could probably figure out what kind of experimental apparatus would be adequate. I'm not doing the math for you.