Page 3 of 21

Re: Soft Sciences Vs. Hard Sciences

Posted: Thu Jan 24, 2019 6:19 am
by teo123
There's just something I think I need to reply to:
However, there are testable components to government and economics (these are much harder sciences than linguistics, actually).
What makes you think this is the case? The truth is, the economists have always failed to predict a crisis, yet alone agree on what the government needs to do to solve it. The most famous case of it was the Great Depression, the economic theories back in the day didn't have much to say about how something like that could happen. Then came Keynes and claimed it was, basically, due to insufficient regulation. Of course, his theory failed to correctly predict future recessions. Then, after every major recession, there came an economist with some "revolutionary" post-hoc explanation of how it could have happened and what is the solution. Of course, the predictive value of that explanation is always very low. But don't take my word for it, here is a paper from IMF saying that:
https://www.imf.org/external/pubs/ft/wp/2000/wp0077.pdf
The failure to predict recessions has been a notable feature of forecasts of the US economy.

Re: Soft Sciences Vs. Hard Sciences

Posted: Thu Jan 24, 2019 7:04 am
by Jebus
teo123 wrote: Thu Jan 24, 2019 4:17 amYou first need to establish using the statistical methods that the languages are related or came into contact before you claim particular words are related.
Which (or both) of the following two statements are you doubtful towards?

Croatian is a language that evolved from Proto-Slavic.

The people who spoke Proto-Slavic were present within the Croatian borders.
teo123 wrote: Thu Jan 24, 2019 4:17 amSo, by that logic, you should have to establish that a significant proportion of the Croatian toponyms is explicable using Croatian, before you claim a particular toponym comes from an obsolete (or even unattested but reconstructed from a Proto-Slavic root) Croatian word.
Perhaps I misunderstand you, but are you claiming that all toponyms originated from around the same time? It seems to me likely that some of the toponyms originated while Croatian was the prevalent language in the region, that some originated while Proto-Slavic was the prevalent language, and perhaps some that originated even before Slavs entered Croatia. It also seems possible to me that those with a different origin have been influenced by the Croatian language to such an extent that it could be easy to mistake them from having Croatian roots.
teo123 wrote: Thu Jan 24, 2019 4:17 amMany modern Croatian toponyms have a well-known Latin etymology. However, I argue that most of those etymologies can't be true because of the known sound laws that would have affected a Latin toponym borrowed into Old Croatian.
There were many toponyms attested in ancient sources that aren't explicable using Latin or Greek. They are usually attributed by mainstream linguistics to the "Illyrian languages", implying there were many and possibly unrelated "Illyrian languages". However, I think that Illyrian was more-or-less a single language (closely related languages at worst), because those toponyms appear to have elements that repeat themselves with certain meanings.
You seem well educated about Croatian toponyms. I'm not but it seems to me like a field that is often limited to educated guesses where it would be really difficult to determine all toponyms with certainty.
teo123 wrote: Thu Jan 24, 2019 4:17 amNow, Brimstone claims that's not a real science. His basic arguments are:
1. My claims are practically unfalsifiable without a time machine.
I don't understand all your claims about toponyms. However, if there is disagreement among the experts on Croatian toponyms, I doubt we will ever know the truth for sure. I would be surprised if there is still much that can be learned about language origins in the area.
teo123 wrote: Thu Jan 24, 2019 4:17 amBrimstone's response to that was that you shouldn't really expect anything because "words easily move".
Seems to me unlikely, but possible (if that's what Brimstone means), that words could completely change meaning over time. However, I would not rule out the possibility that once a word has become extinct that it could be used for something that is different from its original meaning. For example, Ivan wants to name a mountain and he chooses to include the word XXX. The origin of XXX means river, but no one in Ivan's time is aware of that. Ivan just likes the sound of XXX and decides to include it when naming the mountain.
teo123 wrote: Thu Jan 24, 2019 4:17 amMy response was that it's not correct because toponyms don't change the places they refer to randomly, but according to certain laws.


Do you think there could be some exceptions to these laws?
teo123 wrote: Thu Jan 24, 2019 4:17 am2. That, because the sound laws are supposedly not 100% accurate here, I need to calculate P-values before making any statements relying on them.
Are you making statements with 100% certainty that other experts are not in agreement with? Do other experts you have communicated with make their statements with certainty or do they admit that they are only educated guesses?
teo123 wrote: Thu Jan 24, 2019 4:17 ama) P-values are here almost impossible to calculate because the identification of the ancient Croatian toponyms partly relies on sound laws being exceptionless (that's why the ancient hydronym "Murios", for example, isn't connected to the river Mura).
b) The main reason for the apparent exceptions to the sound-laws is dialectal diffusion, and we can assume it doesn't exist in the toponyms.
c) In the second peer-reviewed paper I published, I did attempt to calculate the P-values of some patterns I saw in toponyms, and they were extremely low.


I agree that it would be very difficult to estimate P-values. This is the same reason I suspect it is difficult to determine toponyms with absolute certainty.
teo123 wrote: Thu Jan 24, 2019 4:17 amd) P-values don't drastically increase the chance of being right here.


How so?

Re: Soft Sciences Vs. Hard Sciences

Posted: Thu Jan 24, 2019 8:02 am
by teo123
The people who spoke Proto-Slavic were present within the Croatian borders.
It's basically given that's false. Proto-Slavic was probably spoken somewhere on the Dnieper river, by a small community before the expansion of the Slavic languages begun. Old Croatian was a dialect of Common Slavic (which was probably not mutually intelligible with Proto-Slavic, given the vast number of sound changes that happened between them), and Common Slavic itself was dialectically very diverse. What the mainstream historical linguistics claims and I doubt is that the toponyms from that time on mostly come from Croatian. I think they mostly come from another language that, despite never being attested, co-existed with Croatian for a very long time. And I think it's the same language the ancient Croatian toponyms come from, the Illyrian language.
Perhaps I misunderstand you, but are you claiming that all toponyms originated from around the same time?
No? Why?
It also seems possible to me that those with a different origin have been influenced by the Croatian language to such an extent that it could be easy to mistake them from having Croatian roots.
Yes, such are common. For instance, the name of the river Vuka (and the toponyms near it, such as Vukovar and Vucedol) appear to be derived from the Croatian word "vuk" (wolf), when it actually comes from the ancient name "Ulca". Similarly, the name of the island Lastovo appears to be derived from the Croatian word for swallow bird, "lasta", when in fact it comes from the ancient name "Ladesta". In fact, I suspect those cases to be even more common than mainstream linguistics claims.
You seem well educated about Croatian toponyms.
Well, thanks! I've put a lot of effort into it.
It would be really difficult to determine all toponyms with certainty.
I agree. Most of the toponyms will remain a mystery for a long time, perhaps forever. However, I think we can determine the meanings of some elements that appear multiple times in toponyms with reasonable certainty.
For example, Ivan wants to name a mountain and he chooses to include the word XXX. The origin of XXX means river, but no one in Ivan's time is aware of that. Ivan just likes the sound of XXX and decides to include it when naming the mountain.
I don't think that's likely. Through what intermediary steps can a word meaning "river" change into a word meaning "mountain"? A change that often happens is from "to foam" to "spring" or from "to grow" to "mountain".
What's documented to have happened, and is somewhat similar to that, is that:
1. A stream is being named after a word meaning "stream", in the documented case "Papuk".
2. The word disappears from the language, but the name of the stream remains.
3. The name "Papuk", since it doesn't mean anything any more in the language, gets applied to a wider area than just the stream it originally named.
4. The area the name "Papuk" refers to gets larger and starts to name an entire mountain, while the stream gets another name, in the documented case "Skakavac".
Do you think there could be some exceptions to these laws?
Well, more exceptions than to sound laws, but still very few. If you find the element *issa~iasa naming the mouth of a river rather than its spring, or an island that is completely waterless, that would strongly suggest my theory is wrong.
How so?
Because we see that, for instance, the element *kar~kur appears in hydronyms way more often than the elements "aqua", "visz" and "voda", and those are the words that demonstrably mean "water" in some attested language widely spoken at some time in Croatia, while the element *kar~kur isn't explicable using any of those attested languages.

Re: Soft Sciences Vs. Hard Sciences

Posted: Fri Jan 25, 2019 1:49 am
by teo123
Linguistics, as I explained, is a soft science.
Computer science is more of a field of engineering; it doesn't typically apply scientific methodology.
I don't know about that, all I see is that studying linguistics and studying computer science is way more similar to each other than either is to studying Flat-Earthism.
What's distinct for something like "we need a higher energy particle accelerator" is that it's something we can build.
How do you know there isn't some unknown law of physics preventing the higher energy particle accelerators to exist?
That's cool, but it's meaningless without a P value, and P value is meaningless post hoc.
That's right, nothing can be learned from history... Other than that anarchy is to be afraid of, more than the dangers of there being a government that are fully explicable by reason.
You don't get to assume you're right until somebody else shows another theory is just as likely to be right.
But that's exactly what "falsifiable" means, right? That's what makes scientific method better than the philosophically very flawed inductive reasoning.
Linguistics is a soft science. And it's among the softest of the soft.
How did you figure that out? Linguistics, unlike, for example, macroeconomics or the political science, isn't connected to politics for there to be a bias in it. As the IMF paper I linked to says, the predictions the economists (those of them who claim it's possible to make macroeconomic predictions) make are not only incorrect, they are demonstrably biased toward the idea that the current economic policies are good.
Besides, which statement describes human behavior more accurately? The statement "People generally behave rationally and biases are never systematic." (the core premise of economics) or "When people speak, they obey the laws of grammar, and if they don't, other people quickly notice that." (the core premise of linguistics)?
Honest linguists will admit linguistics is a soft science.
So too will honest economists admit they can't predict the economic recessions, explain what caused them, and propose what the government should do about them.
What material value do these languages hold?
Possible evidence to falsify current linguistic theories, among other things. If the Piraha language truly lacks linguistic recursion (and it hasn't been documented to a degree we can actually claim that), that would prove the generative grammar wrong.
Besides, "preserving the languages" usually means simply getting bureaucracy outside of matters of free speech, with their authoritarian notions that some languages are better than others.
You basically have to appeal to a very exaggerated notion of how linguistics affect human cognition
As far as I know, no serious linguist has held to that notion since the advent of the comparative linguistics in the Renaissance (up until then, grammar was considered a part of logic). And don't bring up that "Hopi time"-story, that's just the media vastly misrepresenting linguistics.
Sure, there are linguists who claim to have done the studies about how, for example, people who speak languages with genders or noun-classes have different associations than those who speak English, and there are also studies that fail to replicate that effect. Very few linguists take such notions seriously.
You just can't. Thus soft science.
OK, let's move that into some other context.
A: How far will a ball with a mass m and a radius r go until it touches the ground if you throw it horizontally at the speed v?
B: OK, here is a calculation assuming the ball is in the air at the atmospheric pressure. It will go s meters.
A: With what probability?
B: How do I know? That's simply what the laws of physics tell us. Obviously, they will not calculate the effects of a strong wind, since it could theoretically be there, but we have no reason to believe it's actually there.
A: So, you can't tell me the probability? Well, then that's not a real science.
Does that make sense? Of course not.
How certain? What probability?
Why would the numbers be important here? I just don't see it. You mean like trying to apply the Bayesian Formula to determine how certain the sound changes are? Sorry, the languages just don't work that way. The vocabulary of the languages is divided into layers. The bottom-most layer is called basic vocabulary, it's approximately the words on the Swadesh list. That's where sound changes are derived from, because those are the words that are very rarely replaced by loan-words or neologisms, in other words, they are the oldest words in a language. If it appears that different sound laws operate in basic vocabulary and in some other part of the vocabulary, the right inference from that is that a layer of a vocabulary has been borrowed from a distantly related language.
Consider the Armenian language, for example. From the basic vocabulary, we see that Proto-Indo-European *d turns into 't' in Armenian. If we were to apply the Bayesian Inference to the entire vocabulary, we would get the result that Proto-Indo-European *d remains 'd' in Armenian with low probability of error (since all but around 400 words in Armenian are loan-words, the vast majority of them being from Iranian languages, which didn't undergo the 'd'->'t' sound change).
Do you think that dumping a dictionary into a statistical program would be more scientific than what I am doing now?
Do you also think those phonosemantic studies with low P-values should be taken seriously, despite the fact that their conclusions being true would refute the core premise of etymology?
Why would you assume some scroll is a trustworthy example of a language?
Because I have no good reason to think it isn't. People generally write down grammatical and somewhat sensible (though often theology-related) samples of a language. If some alien claims to be a human being that lived 2000 years ago, that's a good reason to assume he is not telling the truth.
So we need to TEST the idea of changing the minimum wage.
Yeah, everything in science needs to be empirical. Except, that doesn't appear to be the way science works. Should the, for example, the magnetic permeability of vacuum be tested empirically, even though the theory clearly says that it's a constant equal to mu=1/(epsilon*speed_of_light^2)? What would be the point of it? If the theory didn't make the right predictions assuming that value of magnetic permeability of vacuum is true, there is no reason to assume it will suddenly start making the right predictions with an empirically-determined value of "mu".
And that's kind of the point of the legitimate part of the climate change controversy. The theory predicts there will be a negative feedback loop of vapor in the atmosphere decreasing the warming effect of CO2 by a factor of around 0.5. Some empirical data shows there is actually a positive feedback loop of vapor in the atmosphere increasing the warming effects of CO2 by a factor of around 3, assuming all the recent warming was due to the increase of CO2 in the atmosphere. Thereby, the mainstream climate science rejects the "negative feedback loop" prediction the theory makes, and replaces it with the positive feedback loop factor that's been "determined empirically". Is that the right thing to do? The time will show.
I hurt your feelings by denying that you're qualified to evaluate hard sciences because you published a few peer reviewed papers in linguistics.
And insisting that people from "hard sciences" are somehow more competent in evaluating politics (and philosophy) than I am.

Re: Soft Sciences Vs. Hard Sciences

Posted: Sat Jan 26, 2019 6:22 pm
by brimstoneSalad
teo123 wrote: Thu Jan 24, 2019 6:19 am What makes you think this is the case? The truth is, the economists have always failed to predict a crisis,
A bubble can sustain itself on false consumer/investor confidence for years or even decades. It's an issue of chaotic human psychology nobody can predict.
An economist's evaluation can tell you whether something is stable, or if it's over or under valued. But there's no way to really predict a recession because it's all about the confidence flawed and often irrational human beings have in the market and their behaviors are beyond the scope of the hardest parts of economics.

There is HISTORICAL analysis in economics, where you try to plug in a bunch of variables and find correlations in history to attempt a prediction of what the human beings in the system will do... but that's only weakly predictive, because there are so many differences today (like the internet) and we don't have that much sample data.

That's why much of economics in terms of what it can predict is still pretty soft. However, there are aspects that are on the harder side, like basic supply and demand economics (without any weird consumer confidence bubbles), larger trends, and broadly the ideal behavior of rational agents.
teo123 wrote: Thu Jan 24, 2019 6:19 amyet alone agree on what the government needs to do to solve it.
There is general agreement that the government needs to spend money to stimulate the economy in a recession. WHAT you want to spend it on can vary. Infrastructure is usually a good bet, because it has value to the economy after spending it. Some prefer spending in the form of tax breaks which in theory might incentivize companies to grow and that would in turn benefit the employees... that is, "trickle down economics".

https://en.wikipedia.org/wiki/Trickle-down_economics

There have been more higher quality studies recently showing that trickle-down doesn't really work in practice (as it's been put into practice in the U.S.), but theory is more complicated because very high taxes CAN negatively affect economic growth. The empirical question is really about what the ideal tax rate is, and how, psychologically, to encourage more investment and spending. These are complicated questions that we don't always have good answers to. However, a lack of answers doesn't mean those answers can't be found with good research.
teo123 wrote: Thu Jan 24, 2019 6:19 amThen, after every major recession, there came an economist with some "revolutionary" post-hoc explanation of how it could have happened and what is the solution.
People want answers, and whatever economist will give it to them (even if that's acting in bad faith) will get put on blast by news organizations that don't know any better.
It's kind of how all of these celebrity doctors offering a cause to all illnesses and a miracle cure become famous and end up selling a bunch of books and go on daytime talk shows (like the evil Candida, and "just don't eat carbs and everything will heal!", or dietary fat, and the respective "no oil!" miracle cure). If you were to judge medicine by celebrity TV doctors you'd be making a very big mistake. Don't make the same mistake about economics.

The bottom line, Economics isn't a hard science like physics because it doesn't involve some very hard to predict things like human psychology. But it's a harder science than linguistics because it also involves some pretty hard math and relies less on ancient history we have little surviving evidence from. It's somewhere in between.

Re: Soft Sciences Vs. Hard Sciences

Posted: Sat Jan 26, 2019 8:25 pm
by Red
While we're on the topic of economics (which is probably the hardest of all social sciences), I don't really see much respect for the discipline. For example, RationalWiki refers to economics as "The Dismal Science" (taken from a Thomas Carlyle quote).
https://rationalwiki.org/wiki/Economics

I think economics is a particularly useful discipline, so I don't see the disrespect here. Is it because people get too political when it comes to such issues? Or when people who study politics assume they know about 50% of it?

And since you brought up math, would you say it's a rule of thumb in science that the more math a discipline involves, the harder it is?

Re: Soft Sciences Vs. Hard Sciences

Posted: Sun Jan 27, 2019 3:50 pm
by Lay Vegan
Red wrote: Sat Jan 26, 2019 8:25 pm And since you brought up math, would you say it's a rule of thumb in science that the more math a discipline involves, the harder it is?
Not necessarily, although to some extent this is true. It’s not only the presence of math that makes a science “hard” or “soft” but the reliability of the math. It’s the degree to which variables can be controlled for in an experiment. In a field like physics or astronomy, researchers can control for all conditions so that anybody who sets up the same experiment with those same conditions should expect to get the same quantitative result (replication).

One of the key elements in really any scientific experiment is the statistical evaluation of outcomes between a test group and a control group, something that “soft scientists” are indeed familiar with. Economics is not considered soft because of a lack of math, but because economists can’t properly identify or control for all variables that effect the outcomes of an experiment.

Even if in a carefully controlled experiment, my state (unlike all 49 other states) suddenly raises taxes then suffers massive unemployment, who’s to say there isn’t some unknown variable affecting that outcome? How replicable could that experiment really be? And how generalizable is the math? It’s no different with other “soft sciences.” Psychologists can use mathematical models to predict and explain human behavior extensively, but the question is how replicable and how generalizable their theories are.

I’d also add that this doesn’t make soft sciences any less important than the “hard sciences.” In reality, they’re extremely important because of their practical applications to our society, and the unreliability of their laws and theories relative to harder sciences doesn’t invalidate their utility.

Re: Soft Sciences Vs. Hard Sciences

Posted: Sun Jan 27, 2019 7:29 pm
by Red
Lay Vegan wrote: Sun Jan 27, 2019 3:50 pm Not necessarily, although to some extent this is true. It’s not only the presence of math that makes a science “hard” or “soft” but the reliability of the math. It’s the degree to which variables can be controlled for in an experiment. In a field like physics or astronomy, researchers can control for all conditions so that anybody who sets up the same experiment with those same conditions should expect to get the same quantitative result (replication).

One of the key elements in really any scientific experiment is the statistical evaluation of outcomes between a test group and a control group, something that “soft scientists” are indeed familiar with. Economics is not considered soft because of a lack of math, but because economists can’t properly identify or control for all variables that effect the outcomes of an experiment.

Even if in a carefully controlled experiment, my state (unlike all 49 other states) suddenly raises taxes then suffers massive unemployment, who’s to say there isn’t some unknown variable affecting that outcome? How replicable could that experiment really be? And how generalizable is the math? It’s no different with other “soft sciences.” Psychologists can use mathematical models to predict and explain human behavior extensively, but the question is how replicable and how generalizable their theories are.

I’d also add that this doesn’t make soft sciences any less important than the “hard sciences.” In reality, they’re extremely important because of their practical applications to our society, and the unreliability of their laws and theories relative to harder sciences doesn’t invalidate their utility.
I agree with all of what you said here. Good post, though I'm not quite sure what to add.

I just want to clarify though; You do acknowledge that the social sciences are lacking in rigor, and until then, they can't serve a significant purpose in society.

Re: Soft Sciences Vs. Hard Sciences

Posted: Sun Jan 27, 2019 8:57 pm
by esquizofrenico
I don't think that it is a problem of lacking in rigour, it is simply that the scientific method, which works marvels when studying inanimate and non personal objects, it's much less useful in the kind of objects social sciences study. This is most clearly understood in the case of economics, which has already been commented. In my opinion, the existence of a perfect and completely reliable economic theory it's a contradiction, because the existence of a perfect and completely reliable economic theory would have a dramatic effect in the economic market. This is something that Wittgestein realized about psychology and why he believed it would never be a real science. You can find a formula that perfectly describes the motion of an electron and the electron won't give a damn. But if you find a formula that perfectly described the motion of a person, that person will be extremely interested in reading about it and will act in a certain way, maybe rebelling (making it a self-contradicted prophecy) or even trying to follow the formula (making it a self-fulfilled prophecy). All systems with this level of feedback are known to be extremely chaotic.

Another problem is, of course, that some experiments that would be very easy to do with objects that are considered by the majority as non personal, are impossible to perform with personal objects. To kidnap a group of babies and putting them in an isolated chamber for their whole life to see if they develop a new language would be an extremely useful experiment, that would settle a lot of discussion about the origin of language. But good luck doing that.

So in my opinion, it is not that social sciences lack in rigour, it is simply that the scientific method has very limited use for the kind of objects they study. In fact, I believe that fields like linguistics are the most likely social sciences to be able to give a very rigorous science, since they deal with an object that was settled long ago and therefore their observations will have no effect on it. I have gone to a lot of complexity science conferences in which their main topic was linguistics and what they were doing (studying language through graph theory and entropy analysis) looked extremely rigorous and give very convincing conclusions about certain features that all languages share. That could give us predictive power to know if a certain document contains a language or simply babble, and in my opinion the only thing that matters about a science is whether it has predictive power or not.

Re: Soft Sciences Vs. Hard Sciences

Posted: Sun Jan 27, 2019 9:54 pm
by Lay Vegan
Red wrote: Sun Jan 27, 2019 7:29 pm I just want to clarify though; You do acknowledge that the social sciences are lacking in rigor, and until then, they can't serve a significant purpose in society.
What I will concede is that social sciences make soft (mostly unreliable) predictions whereas the physical sciences make hard (reliable) predictions. Obviously some fields like psychology can’t operate like chemistry, since it involves studying intangible things (what is happiness? And how is that operationally defined?) but with an abundance of replicable studies where terms are clearly defined and controlled for, perhaps such fields can become more reliable as a whole?

Unfortunately, social systems introduce far too many confounding variables that are impervious even to randomized controlled experiments (something of a gold standard in science). And of course, even if social scientists were to increase their use of randomized controlled experiments, this would introduce serious ethical concerns, as @esquizofrenico points out.

However, instead of condescendingly dismissing these sciences as “useless” or “insignificant” thereby stigmatizing them, we should be increasing funding toward those disciplines and push to improve their methodology.