Monday, December 16, 2013

Breaking down Creationist bad linguistics

I never suspected this blog would deal with Creationism, but it seems I can't permit a certain psuedo-scientific article go unchecked. This post will dismantle the bad linguistics behind "The Tower of Babel account affirmed by linguistics," an essay by a certain K. J. Duursma.

Duursma's piece has recieved some circulation in Christian apologist circles, such as the Theopedia. This article is not meant as an attack on Christianity or religion in general. This is an attack on bad linguistics. The article is characterized by quote mining of reputable linguists (Crowely et al; R. L. Trask; etc...), less reputable ones (Greenberg; Ruhlen), and ones of no repute (Steel). The factual mistakes and historical distortions betray the ignorance of the author and the foundationlessness of his assertions.

Secular linguists are puzzled by the existence of twenty or so language families in the world today.

Wrong. It's the very first sentence and you are wrong. There are roughly 136 language families according to Ethnologue. Saying 20 language families is grossly ignorant of the facts. The only way you could arrive at a number like 20 is if you confused the clustered families like "Papuan" and "Paleo-Siberian" language families for single language families. They are not. Families like Khoisean are just titles linguists give to refer to dozens or more of unrelated language families that exist in geographically local regions. We're off to a bad start.

The languages within each family (and the people that speak them) have been shown to be genetically related, but few genetic links have been observed between families.

Wrong. Sentence two and wrong again. There are two possible interpretations of this terrible sentence and either interpretation is incorrect:. First interpretation is that Duursma is referring to two subfamilies within a larger macro-family (such as Italo-Faliscan and Germanic families being siblings within the larger Indo-European family). That's patently false as macro-families must have incredibly strong evidence in order to be granted the designation of a genetic relationship. Second interpretation is that Duursma is referring to any two macro-families like Uralic and Indo-European. This is probably what Duursma meant and it's even dumber. There aren't genetic links observed between separated families. By definition a genetic relationship between two language families means they are part of the same family.

But still, if speech did evolve somewhere, somehow, we would expect to find that all languages are genetically related.

World languages may be related; world languages may not be. The rate of language change exceeds the amount of time that humans have been speaking. In other words, languages that are related have mutated past the point of proving a genetic relationship. Furthermore, we needn't expect that all languages must be genetically related even if we evolved. Sometimes languages are birthed spontaneously, as in the case of Nicaraguan Sign Language. The mind is capable of spontaneously creating impressively complex modes of communication even without the influence of language, and the mind is capable of making modes more complex in order to fit one's needs.

Some have therefore suggested that man evolved speech simultaneously in more than one place. This suggestion is beyond belief, considering the dangers involved in the supposed evolution of speech.

No, it's not beyond belief. As I mentioned above, people are capable of creating language ex nihilo.

Only Genesis provides a credible explanation.

Of all the theories for how speech originated and whence our language families came, Genesis is the only incredible explanation.

Central Asia and the rest of northern Asia host the Altaic family, which also contains Turkish.

There are few Altaicists today and since the 1990s they no longer represent a majority view of linguists. The simple fact is that Japonic and Koreanic families do not have even close to enough evidence to support a genetic relatedness. Tungusic as been strongly questioned as well and its position is no longer certain. Duursma is forgetting that, even if we accept the putative Altaic family, there are many, many other north Asian language families like Nivkh, Ainu, and Yeniseian.

The Pacific is host to three or four families.

The Pacific is home to as few as 20 language families but probably many, many more. The only way this sentence makes sense is if the author is counting by Greenberg's typology, which no credible linguist believes.

...the Khosian languages are spoken in the south-west of Africa.

Remember when I predicted that Duursma must be counting language families by confusing 'convenience families' for genetically-related language families. Prediction confirmed. The Khoisean language family is a grouping of convenience. Only a few of the languages within the group are actually related. Nearly all are isolates. We just lump them into a big group because they're close to each other and utilize clicks.

The result shows that 19.5% of the core vocabulary changes every 1,000 years.

No. Some lexicostatisticians had found that the Swadesh vocabulary changes at a rate of 19.5% every 1,000 years. Duursma cites Crowley et al. to "prove" his point, but that very same book demonstrates the problem with that figure just a few pages later. The number was arrived at by testing just 13 of the world's 7000 languages (11 of them were Indo-European). There is a serious problem of statistical significance and contamination due to relatedness and proximity. The simple fact is that the only thing we can definitively say about language change is that change is not constant and can accelerate or decelerate rapidly and unexpectedly.

If this is the same for all languages, it means that statistically all words in a language should be replaced within a period of about 10,000 years. 

This is actually close to the truth, but - funnily enough - it dismantles Duursma's whole point. If genetic links cannot be found past 10,000 years because of natural language evolution, then the fact that there are unrelated language families is not troublesome at all. It certainly is no evidence for the Tower of Babel. There are any number of more parsimonious explanations for the existence of 136 language families in the world. Sorry, I meant twenty.

Trask remains unsure as to how and when this change occurred.

Well, yeah. Trask wasn't an evolutionary biologist and never pretended to be. He was a very, very fine Vasconist - a linguist of the Basque language - but by no means a biologist. We need better sources than Trask to learn about the evolution of the vocal tract and speech.

Again, there is no evidence to back their view that speech evolved.

Demonstrating that speech evolved was never Trask's intention. Trask's book is an introduction to the way language works. It's to introduce the public to how societies and minds works with language. Duursma should not pick up books on Topic A (general linguistics) and blame it for not proving Topic B (the evolution of the vocal tract).

...scholars supporting monogenesis or the relatability of all languages run the risk of being branded Creationists and of therefore having their work disregarded by colleagues.

I completely agree. There is no evidence for a single ancestral tongue of all world languages and those that support such an idea are groundless.

It seems that there is little evidence to support the view that all languages evolved from one or more proto-languages.

A lack of evidence for one theory is not positive evidence for another.

The Babel account suggests that several languages came into existence on that day.

Well, unless only a few people were building the Tower in the Plain of Shinar, there should be several thousand languages created that day, at least. Right? One language for each builder? I guess Duursma could just say that God split them up into three or four language families and then they started fighting. There doesn't seem to be much Biblical evidence for that either. Oh, well, my point here isn't to debate Biblical interpretations, just the linguistic facts, so I'll move on. I just thought it was funny.

The unitary state of Indo-European languages ... [is dated at 3000 BCE].

NO. That is far too young a date. Proto-Indo-European, the ancestral tongue of all IE languages, was last spoken just north of the Caucasus somewhere around 5000 BCE at least. 3000 BCE is an extremely fringe opinion and not even close to representative of the majority viewpoint. I would say that a date of 3000 BCE is about as fringe among linguists as those who believe PIE dates to 9,000 BCE.

Wieland points out that ‘to have such close correlation’s still existing makes little sense if the migrations were as much as 11,000 years ago, as is commonly believed. From the biblical record, they would have been less than some 4,000 years ago’

Umm... okay? Well, you haven't demonstrated that a Tower of Babel event occurred 4,000 years ago, so all of this is evidence-less speculation. And don't tell me that evidence for the Tower of Babel is not the point of the article. You blamed Trask for not proving the evolution of the vocal tract when his intention was to show how people work with language. Regardless, I'm still waiting for this "evidence" I keep hearing about.

Crowley carries on to share how languages can change from sophisticated to simpler versions, and from simpler to more complex systems. He distinguishes between, ‘isolating’, ‘agglutinating’ and ‘inflecting’ languages and shows how languages change in circular patterns.

If languages evolve in a circular pattern then this implies they either don't lose complexity or they regain complexity when they come full circle. Nevermind that none of this is true. Languages do not change in overall complexity; languages become "simpler" for the native speaker and listener as a method of reducing the effort to convey messages, but the languages gain complexity with the production of novel shortcuts that are invisible to the communicants. The point to take home here is that complexity is a constant over time.

Classical Greek was a highly inflected language; it used five cases, as well as Active, Middle and Passive voice. Koine Greek was almost reduced to four cases, and the Middle voice was used rather inconsistently. Modern Greek distinguishes only three cases, but many endings have disappeared. It is a good example of van der Tuuk’s Ruin, as it is slowly becoming an isolated language.

Wow. Duursma confuses the pathway from agglutination to isolation for simplification. This says nothing about complexity. Isolating languages are incredibly complex, it just so happens that they are less complex in terms of inflection. English, for example, reduced the number of cases over time but replaced this with a complex system of word order, idiomatic phrasing, light verbs, and novel case marking (like the fearsome triple genitive that most world languages cannot easily translate in a concise manner). Duursma has no idea what he's talking about.

However, this model cannot be used to explain the origin of highly sophisticated language systems like Sanskrit and Greek.

Yes, it can. The origin of both is Proto-Indo-European and we've known about it for 200 years. In fact, it was first hypothesized by the same Sir William Jones mentioned at the head of this article. Golly.

Language change, as Crowley’s model shows, would be unlikely to produce consistent endings for the whole of the Inflecting Language.

This does not make sense.

The fact remains that the Greek/Sanskrit parent was utterly consistent...

No, it wasn't. What are you talking about, Duursma? Ever hear of Early and Late PIE? Narrow PIE? The S-mobile? The isogloss mysteries and wave theory? The shift from animate/inanimate-neuter to masculine/feminine/neuter? This article is a bad joke that keeps getting worse.

If chance, then, did not make this Proto Language, where did it get its consistency from?

First of all, PIE is a reconstructed language. That means we can only reconstruct what is shared among its daughter languages. Part of what that means is that we invariably fail to capture the full nuance of PIE, including its inconsistencies. Regardless of the fact that we already know PIE was not consistent and far from it. Why wasn't it consistent? Cause it was a normal language just like Etruscan or Tagalog: constantly in flux. If it were consistent (which is a stupid idea to begin with), it would be a conlang.

It suggests a Designer.

It doesn't suggest anything other than it was a ~7000 year old language with a number of differing dialects.

In Babel one of the groups was given the sophisticated, and utterly consistent, Proto Indo-European language.

This implies that PIE was a first language at the Tower of Babel event. Does Duursma realize that there are fragments of Pre-Proto-Indo-European that can be reconstructed which point to an even earlier stage in the language's history?

Sadly, as people in a fallen world began to use this language, it slowly began to lose some its consistency, as grammatical mistakes became fashionable.

I already pointed out that PIE speakers frequently made mistakes, such as the S-mobile, which arose from a confusion of the inflectional ending *-(V/C)s with the start of the next word. Much like a napron was confused for an apron in English, PIE confused words like -os teuros "the bull" for -os steuros. Thus we get tauros in Greek but steer in English.

The facts we observe today are consistent with the Tower of Babel account in Genesis 11, but this does not prove the correctness of the account.

Habla mucho pero dijo nada.

Since the history of languages cannot be reconstructed beyond 10,000 years, evidence for (and against) alternative views is limited.

I agree. But this undermines the point of the entire article because nothing affirming of the Tower of Babel was presented. That is what you were trying to do when you titled your piece "The Tower of Babel account affirmed by linguistics," right Duursma?

...if we take an objective look at the facts at our disposal we cannot but draw the conclusion that the Bible account has far more going for it than the alternatives, for which there is little, if any, evidence.

You didn't discuss linguistic dating and the reason why is obvious, that would be evidence against a putative Tower of Babel. Some language families pre-date a possible Tower of Babel event. Afrasiatic is roughly 10,000 years old; PIE is 7000; Algic is roughly 8-9000.

We therefore wholeheartedly believe that the findings of historical and comparative linguistics have served indeed to affirm the Tower of Babel account recorded in Genesis 11, beyond reasonable doubt

Wow, that's hilarious. Duursma failed to provide any evidence in favor of Genesis 11. All he said was that there is no evidence against his theory. Beyond reasonable doubt? Let's hear some reasonable evidence first.

Believing this account, however, requires believing in God, and the denial of the evolution theory, which suggests that all animals, humans, and even human language, arose by chance. For many, this might prove too big a price to pay, despite the evidence.

What a capstone to a paper laden with errors and inconsistencies, hilarious mistakes and ridiculous assertions.


  1. Thank you for taking the time to write this.

  2. I would agree with your article if it wasnt for the assumed datings of the languages.

    Paleontologists and histirians tend to overestimate anything that is older than 3000 years.

    We have no rock solid proof that ANYTHING existed before 4000 bc which is mind blowing.

    The bigbang is a mathematical theory with no physical evidence that cannot used for proving a date.

    Geological evidence would be compromised by a worldwide flood which many cultures all over the world describe

    And there is no evidence of human history written or spoken older than 5000 bc.

    Gee i wonder why.

    P.S. Donosaur bones are recently found to have large amount of soft tissue and blood which discarts the theory of "fossils" not to mention the fact that paleontologists refuse to carbon date them

    So we are left with one question.
    What do we really know?

    1. I have a question about your worldwide flood. According to the Bible, your invisible ghost murdered nearly every human on Earth and nearly all the animals who apparently must have sinned as well. But my question isn't about Him since I don't believe your imaginary fwend exists outside your head. My question is about you ...

      Why would any moral human being worship such a hideous brutal creature?

      And a follow-up question: Since you do, what does that say about you?