Why Latin Died Out?

[Pantagathus]

And the point of reconstructing PW is exactly the same as reconstructing a dinosaur or any other fossil--to find out about things evolve and what underlying mechanisms constrain the changes.

 

But at least with dinosaurs we have skeletal fossils...

 

Trying to determine PW is like reconstructing a prehistoric beast based on skeletal evidence from between now and the ice age. I doubt one would arrive at enormous lizard like creatures...

 

So like you said, it's quixotic

[M. Porcius Cato]

Language is not random. Words look and sound the way they do for reasons. That said, I am probably crazy.

Language may not be random as a code (if it were, it would be unlearnable and non-creative), but the arbitrariness of morphemes is one of the defining features of human language.

[Lucius Calpurnius Capitolinus]

Here is where I feel we may disagree. I have real troubles believing that anything is arbitrary-especially things involving human interaction.

[M. Porcius Cato]

Here is where I feel we may disagree. I have real troubles believing that anything is arbitrary-especially things involving human interaction.

 

You may have trouble with the concept, but lots of things are arbitrary. My iPod, for example, has a shuffle function which plays songs randomly--does it sound sometimes like there's an algorithm guessing what would be best to play next? Sure. But there's not one. Sometimes, things are random even when they don't seem to be. Believing in order where there is none is a very human reaction--but it's still wrong.

[Q Valerius Scerio]

Actually, all "random" selecters/generators are driven by algorithms, unless some new technology based on quantum mechanics just came out recently.

 

Also, sarcasm aside, it's not merely that figuring out the first language is hard, but it's impossible, given our current evidence. We have absolutely no idea when or where our language evolved, when we began to speak even. How do we know the first wave of human emigration to southeastern Asian learned to speak while those in Africa remained grunting. How do we know that the grunting was arbitrary? Morphemes? We have no idea when they could have evolved. Furthermore, if they evolved earlier, we have no idea where we might have learned them from. Surely it's absurd to posit humans as the only "speaking" creatures. We had very close relatives a while back. Going beyond 10,000 years is daring, beyond 25,000 years is insane, to the dawn of human awareness plain stupid.

[M. Porcius Cato]

We have absolutely no idea when or where our language evolved, when we began to speak even.

 

Given that you earlier claimed that PW didn't even exist, are you now claiming it did exist but that we simply can't know what it would be? I think this is an absurdly pessimistic view. But give linguistics and allied sciences 50 years, and we'll find out if your pessimism is warranted. The only certainty is that your pessimism won't contribute an iota to the project.

Edited February 16, 2006 by M. Porcius Cato

[Q Valerius Scerio]

No, I never claimed that PW existed either. If you read my post, you'll see a perfectly reasonable alternative to PW. I still call *****. And as for my pessimism, I thought I made it very clear that it is impossible with our current evidence. Instead, you give some fanciful fantasy that one day we'll be able to look at a language that is impossible to figure out with earthly evidence. Save using a time machine, it ain't happening. But in case of a time machine, who knows the possibilities?

[M. Porcius Cato]

Language is a combinatorial, conservative system. Therefore, it is possible to reconstruct elements of previous languages based on current ones. There is no need for a time machine--just massive computational power and loads of data.

[docoflove1974]

Wow...you guys are touching on some deep stuff.

 

I'll just interject on the 'arbitrariness' of language:

 

Ferdinand de Saussure, in the early 20th century, was a noted mathmatician, and turned his attention to language. He surmised (right so) that language is really composed of words and meaning--in his French he used parole and langue, with langue really meaning, well, meaning behind the language. His basic thesis was that there is no reason why a language has certain terms and certain morphology...it is created within the language. There is no reason why English has red while Spanish has rojo and French with rouge, Finnish punainen, Japanese niiro. There's no reason why Latin has nominative/accusative constructions, while Australian languages have ergative-absolutive constructions. Why languages initially form themselves they way they do is arbitrary--we can look at the patterns and theorize about the patterns (which is what my colleagues and I do with language change theories), but ultimately the choices are done with an element of randomness.

 

That being said...with respect to PW...with the level of arbitrariness evident in human language, we do not have the tools to go back much further in time than 7-10K years. Perhaps this will change...but for now, it cannot.

[Q Valerius Scerio]

Language is a combinatorial, conservative system. Therefore, it is possible to reconstruct elements of previous languages based on current ones. There is no need for a time machine--just massive computational power and loads of data.

Sorry, that comment was uncalled for. But I'd like to point out the flaws in your thinking. First, languages are not conservative. They are indeed combinatorial, but not conservative. Actually, they're quite fluent. Language changes every day with each person. Languages are constantly being bombarded with so many different elements daily that it's impossible to keep track of it all the time. What you imagine is on the scale of impossibility.

 

Actually, I've been trying to keep track of "loads" of data for one of my projects - the reconstruction of Latin from Romance languages. Perhaps you don't understand the scope of this project...let me just say it's quite massive, even for such a simple task. Perhaps when we're through, we'll have a better understanding of just how far from the original reconstructed languages are. Anyone who speaks "PIE" and thinks that he is actually speaking something even remotely intelligible to is too far in the clouds. All we can do is get to approximations, nothing exact. And we'll never be exact. Actually, there's too much information to be exact.

 

Now imagine trying to reconstruct Latin with only Spanish and French. Yeah, that's how possible it is. You're forgetting all the thousands if not ten thousands of languages which are entirely extinct and have left without a trace. You're ignoring the possibility of two groups forming languages separately at different parts of the world. It is quite possible, if not probable, that what you're going to get may be an approximation to something, but to the original language, you're far from it.

Edited February 16, 2006 by Q Valerius Scerio

[M. Porcius Cato]

Look, no one is claiming that we can reconstruct and revive the original PW language--only that it existed and that it in principle it should be possible to identify some elements of the language with at least a small measure of certainty. Given that machine translation of modern langauges is still a ways off (i.e., computer translators suck), it's unlikely that anyone could successfully complete a full reconstruction of languages that are far more modern than Indo-European.

[Q Valerius Scerio]

Look, no one is claiming that we can reconstruct and revive the original PW language--only that it existed and that it in principle it should be possible to identify some elements of the language with at least a small measure of certainty. Given that machine translation of modern langauges is still a ways off (i.e., computer translators suck), it's unlikely that anyone could successfully complete a full reconstruction of languages that are far more modern than Indo-European.

I still disagree with the premise. The only thing we know about PW is that it may have existed. You'll have to prove to me first that it definitely did. You're still avoiding my scenarios that I've given earlier.

[M. Porcius Cato]

I still disagree with the premise. The only thing we know about PW is that it may have existed. You'll have to prove to me first that it definitely did. You're still avoiding my scenarios that I've given earlier.

 

What, in principle, would prove to you that it did? Further, would proto-Indo-European meet your standard of proof or do you also deny the existence of proto-Indo-European?

 

How do we know the first wave of human emigration to southeastern Asian learned to speak while those in Africa remained grunting.

 

There is no requirement of the PW hypothesis that this be true. All that is required of the PW hypothesis is that all modern languages descended from a single ancestor language. The alternative hypothesis is that modern languages descended from more than one independently-emerging languages. I don't think this is particularly likely if the language-learning phenotype evolved only once, and the likelihood of random mutations leading to an identical phenotype multiple times is highly unlikely.

[Q Valerius Scerio]

What, in principle, would prove to you that it did? Further, would proto-Indo-European meet your standard of proof or do you also deny the existence of proto-Indo-European?

Sorry, I should have been more careful with my word choice. What evidence do you give that all languages evolved from a single language? That's all I'm looking for. And as for PIE, well we still aren't sure that it was ever a single language. Plenty of new hypotheses emerge that look at PIE differently, most notably that PIE was actually different languages that never fully emerged, but interacted with each other enough to produce several original PIEs. I won't comment yet on it's validity but it still remains an option.

 

I don't think this is particularly likely if the language-learning phenotype evolved only once

But we don't know exactly when it evolved, how it evolved, or when it was first utilized to form a language.

 

and the likelihood of random mutations leading to an identical phenotype multiple times is highly unlikely.

Why be mutations? Merely having the capacity for language does not equate to actually having a language itself.

[M. Porcius Cato]

Sorry, I should have been more careful with my word choice. What evidence do you give that all languages evolved from a single language?

 

Obviously this won't be definitive, but the basic argument originates in Joseph Greenberg's work.

 

Greenberg was a professor of anthropology at Stanford and a founder of the linguistics department there; first for his seminal work in the 1950s classifying the 1200 languages of Africa and later for his magisterial work on language universals (including his 1962 paper, which is perhaps the most-cited article in linguistics ever), and the reduction of several hundred language families in the New World to only three, Greenberg has been compared to a modern-day Darwin.

 

To my mind, the most compelling evidence for Greenberg's line of reasoning--which ultimately leads to the construction of PW which Greenberg was attempting--is that the mass comparisons that Greenberg conducted for New World languages accurately predicted that the gene frequencies of New World populations would fall into the same three families identified by Greenberg.

 

Much, much more work from biology is needed to determine how well the Greenberg linguistic families map onto data about human migration over time. But the major insight was to bother to do this in the first place, and this insight directly follows from the hypothesis that all human languages have a common ancestor.

 

I don't think this is particularly likely if the language-learning phenotype evolved only once

But we don't know exactly when it evolved, how it evolved, or when it was first utilized to form a language.

Demanding this level of detail is simply absurd. Ulitmately, whether the first mutant was born on a Tuesday or a Wednesday is completely irrelevant to whether all languages descended from a common ancestor. We don't know the birth day of many common ancestors, but it doesn't matter "exactly when it evolved". Any of these complaints might be lodged against any cladistic reconstruction, but if these lead you to doubt evolution, so much the worse for you not for Darwin.

 

and the likelihood of random mutations leading to an identical phenotype multiple times is highly unlikely.

Why be mutations? Merely having the capacity for language does not equate to actually having a language itself.

 

Having the capacity for language-learning plus exposure to a non-syntactic group of symbols (e.g., a pidgin) is a necessary and sufficient condition for language to emerge. On this see Bickerton's discoveries regarding creolization and the the birth of Nicaraguan sign-language; his work on language evolution is also highly relevant.