Soft Sciences Vs. Hard Sciences

[teo123]

One of the moderators of AtheistForums also thinks the distinction between hard and soft science is invalid:

I think the distinction between ‘hard’ and ‘soft’ sciences is a false one. It’s a traditional nomenclature, nothing more.

[teo123]

What do you think, can linguists analyzing my alternative interpretation of the names of places in Croatia be said to be doing the same mistake that Sunflowers was doing? Sunflowers had blind faith that physics had nothing to say about what were once exclusively matters of metaphysics (whether something can become nothing). Linguists analyzing my interpretation of the names of places in Croatia seem to have blind faith that informatics and statistics have nothing to say about the names of places in Croatia.
As far as I can tell, the mainstream interpretation of the names of places in Croatia makes one falsifiable claim: that the k-r pattern in Croatian river names is accidental. In other words, that the probability of that k-r pattern occurring by chance is high. That can be falsified, but not in terms of historical linguistics. It can be falsified using informatics and statistics. And I think I have falsified that claim, with my measurements and calculations showing that the probability of that occurring by chance is between 1/300 and 1/17.
And informaticians and statisticians I have asked all say my arguments look good to them. And I guess that the reason my arguments fail to convince linguists is the same as the reason my (and Red's and BrimstoneSalad's) arguments failed to convince Sunflowers that he was wrong.
Linguists studying Croatian names of places in 1930s, such as Antun Mayer, had an excuse for thinking the way they did: informatics basically did not exist back then. Linguists today do not have that excuse.

[teo123]

My communicology professor, Jerko Glavaš, just told me that arguments presented in my paper seem compelling to him, although he would not consider himself to be an expert in the field. He also told me "Samo tako nastavite" ("Just continue doing that."), and that it's important to publish papers in peer-reviewed journals, that it counts for much more than good grades do.

[teo123]

What do you think, can proponents of the mainstream interpretation of the Croatian toponyms be said to be doing the same mistake Allan Savory is doing? Allan Savory is trying to use magical cows to circumvent thermodynamics. Similarly, proponents of the mainstream interpretation of the Croatian toponyms seem to be trying to use magical languages to circumvent informatics.

[teo123]

I've asked a new question slightly related to my paper on a linguistics forum: https://linguistics.stackexchange.com/q/46473/20821

Does a deep orthography (such as English or French) decrease the collision entropy of a written language? An obvious answer seems to be no, because spelling represents how a word was pronounced centuries ago, and it seems absurd to suggest that the collision entropy of a language always increases. But a glance at the data from my paper seems to suggest that's the case. I don't know if anybody has done an actual study on that.

[teo123]

I wonder what people interested in historical linguistics mean when they say that I am being too speculative:

It's one thing to speculate, but have you looked anything up at all?

Historical linguistics is speculation, right?

[teo123]

I have asked the question "Why does my question about why a deep orthography appears to decrease the collision entropy of the written language get downvoted." on Linguistics Meta StackExchange:
https://linguistics.meta.stackexchange.com/q/2054/20821
Draconis there, just like Shain on forum.hr, thinks that my question is off-topic. Well, like I've responded to Draconis:

I asked it on a forum about linguistics because I expected the answer to come from historical linguistics. I expected the answer to be something along the lines of: "Because the mantra 'English is spelt as it was pronounced around the invention of the printing press.' is not true. The silent 'e' is way more common today than final 'e' ever was in spoken English. Look at the numbers 1-10: the silent 'e' in 'five' (Old English 'fif') and 'nine' (Old English 'nigon') was never pronounced.". Or something like that...

[teo123]

I have decided to try to explicitly promote the theories stated in my paper on Linguistics StackExchange. Right now, I have more upvotes than downvotes: https://linguistics.stackexchange.com/q/46552/20821

[teo123]

I have been thinking about this lately...

What if it was an ancient linguist playing a joke on the future?

Well, some ancient linguist making a fake Rosetta Stone convincing enough to fool modern linguists is highly implausible. If you try to produce a long meaningless text which looks like a language without knowing modern linguistics, chances are, your text will have statistical properties vastly different from a real language. That's why, if you claim the Voynich Manuscript is a hoax, you need to show an algorithm which a medieval person could use to generate such a hoax (matching statistical properties of a real language so closely). So far, nobody has done that. (That's not to say Voynich Manuscript is not statistically weird. Its first-order entropy is very similar to Hebrew, but its second-order entropy is very low, even lower than Hawaiian. But the question remains, which algorithm could a medieval person use to produce such a hoax?)
Now, if some ancient linguist was making fake glosses (translations of single words or short phrases), the chances of that getting accepted by modern linguists are a bit better, but it's still far from certainly fooling modern linguists. Constantine Porphyrogenitus made, as far as I know, one such fake gloss: claiming that "Croat" means "one who has a lot of land" in Slavic. He was obviously mistaken, there is no similar word with such a meaning in any modern Slavic language.
Now, whether some ancient linguist could fool modern linguists by making fake glosses from a language we know little about... It's hard to tell. The author of Ravenna Cosmography claimed that there are rivers in Pannonia called Ira and Bustricia. We have no idea which rivers those are. But we are pretty sure Bustricia is a real river name, because it has quite an obvious etymology: it is related to Proto-Slavic *bỳstrъ meaning "quick". Now, whether there was a river named Ira... It's hard to tell. What could the name Ira mean? So, yeah, it is quite possible the author of Ravenna Cosmography made up a river name and fooled modern linguists.
I think that the study of names of places relies too much on the works of ancient linguists. Historical sources can be useful, sure, but it's primarily to dismiss false etymologies. The claim that the river name "Karašica" comes from the Croatian fish name "karaš" (crucian carp, from Latin "carassius") can be proven wrong by the early attestations of that name, such as "Karassou", containing a different suffix. But even without that, I think it should be obvious that etymology is unlikely. Before Gatski Kanal was built in 1875, Karašica was a lot faster, making it unlikely there were ever many crucian carps there. Furthermore, why would the first two consonants of many Croatian river names be 'k' and 'r', respectively (I call that "k-r pattern"), if those river names do not share the same root (or at least the same prefix)? The claim that "Poreč" comes from Croatian for "at the river" can be disproven by pointing that its ancient name, long before there were Croatians there, was "Parentium". But even without that, I think it should be obvious that etymology is unlikely. Which river would motivate it? Also, in the local dialect, Proto-Slavic *ě (found in Proto-Slavic word for river, *rěka) changed to 'i', rather than to 'e'. To claim that "Poreč" comes from Croatian for "by the river", one would need to invent that some different dialect of Croatian was once spoken there, which is an obvious ad-hoc hypothesis.
But, like I've said, I think the study of the names of places should rely less on the works of ancient linguists and more on mathematical methods. As I've written my latest paper (there is an English summary on my blog), I think I have proven by measuring the collision entropy of the Croatian language and doing birthday calculations that the probability of that k-r pattern occurring by chance is between 1/300 and 1/17. I think that proves beyond reasonable doubt that those river names share the same root, or at least the same prefix. Obviously, some ancient linguist confirming that would be excellent, but I think that even absent that, it is reasonable to believe my theory.

[teo123]

I asked the following question on Philosophy StackExchange: "If calculating the p-value post-hoc is meaningless, why is it reasonable to believe the Grimm's Law (and other laws of historical phonology) is true?".
Because I think that there is a contradiction between those two claims:
1. Calculating the p-value post-hoc is meaningless, like @brimstoneSalad claims (I hope I am not straw-manning).
2. It is reasonable to believe the Grimm's Law is true, or at least an approximation (in the same sense that Newton's Laws of Motion are an approximation).
The most obvious answer to that contradiction seems to be that "Calculating the p-value post-hoc is meaningless." is not true.