Tuesday, April 15, 2008

Edenics (Pseudolinguistics and Ideal Languages, Part 3)

The story of the tower of Babel is a rich and enchanting myth full of the conflicts between order and chaos and between the gods and humankind that characterize much Mesopotamian and Levant mythology. When one considers the amount of pre-scientific philological work that examined the myth, it is no surprise that even today many modern popular books on linguistics reference the tower (quite often on their covers). It's also no surprise that some within the contemporary creationist movement has sought to establish the historical accuracy and linguistic plausibility of the myth. Biblical literalism - in part or in whole - requires a Biblically-supportive linguistics as much as it requires Biblically compliant biologies and geologies.

While I cannot speak for the degree to which his work is accepted by the larger creationist community, Isaac E. Mozeson has stepped forward to fill the linguistics gap in modern creationist narratives. He has authored at least two books on the subject: The Word: The Dictionary That Reveals the Hebrew Source of English (1989) and The Origin of Speeches: Intelligent Design in Language (2006). At present, I only have access to the first of these. I may acquire and review the second at a latter date. Mozeson also publishes articles at the website Edenics. While without the pages of data contained in The Word, the website offers his hypotheses in full detail for free if you wish to sample his writing before or after reading this post.

As its full title suggests, The Word sets out to demonstrate that English ultimately descends from Biblical Hebrew. Mozeson has later revised this hypothesis to say all languages descend from Edenic (the language spoken by Adam and Eve in the garden of Eden), his apparent intention being to distinguish Edenic from Biblical Hebrew, which, as a result of several centuries of change, may not be identical to its ancestor. In deference to the revised theory, I will use the word Edenic throughout.

In The Word, Mozeson conflates two related but not identical goals. The first of these is that of the etymologist: Mozeson attempts to rewrite the conventional history of many English words to show that their original source is Edenic. Some of the etymologies suggest a direct relationship (or at least show no medial stage), and others suggest Edenic left its imprint on a language from which English has borrowed, which English is related to, or from which English has descended. Unlike some of the best etymological dictionaries, Mozeson provides only partial histories and no usage citations. While not a feature in most desktop dictionaries, the carefully gathered and dated citations in dictionaries like the Oxford English are part of what make them great references for historical linguists. These citations provide data on when a word was first recorded and on how a word's pronunciation or meaning has changed over time.

To the lay dictionary user, The Word's lack of detailed etymologies for each word may make economical sense, but this allows Mozeson to suggest etymologies that make little historical sense when the full history of each word is explored. For example, the word woman first appears in print as the wifman (c. 1000 CE). Mozeson gives the possible Hebrew cognates WREHEM 'damsel' or 'womb'. Obvious questions arise: where did the /f/ come from and where did it go? Where did the /n/ come from? To convince anyone of his alternative etymology, Mozeson needs to supply well documented answers. Likewise, while Mozeson admits the immediate New World language etymon for skunk, in suggesting that these descend from Edenic (cf. Hebrew tsakhun 'to stink'), he neglects the evidence that the Manhattan and Algonquian words are composed of independent morphological pieces. The final /k/ in skunk originates in an Proto-Algonquian word for fox that combined with a word for urination (ลกek). If the account is to be believed, the Edenic word was broken into unrelated pieces in Proto-Algonquian and then recombined just in time to be borrowed - in whole no less - by English.

Mozeson's second goal - which does not seem to be stated directly in The Word - is to demonstrate that Edenic is the mother of all human languages. I'm hesitant to be heavily critical on this issue with only The Word at hand because he makes no effort whatsoever to support this hypothesis in a convincing manner. However, considering the text of the foreward, I'm left believing Mozeson did feel he had sufficiently supported this claim with accepted methodology.
With all my idol breaking, I have remained too true to conservative linguistic rules to be be iconoclastic. I am grateful for my brief training in linguistics, and for the century of research into Indo-European roots that often made my discoveries possible. (p. 1)
Although inevitably related to the search for accurate etymologies, it is altogether a separate task to reconstruct proto-languages and earlier forms in a single language. Individual words may have widely divergent etymologies. That crag originated as a Welsh word for rock has no bearing on any other English word, but to claim that English (or any other language) descended from the same language as Hebrew requires an analysis that is concerned with details above volume. Primarily, it requires the identification of phonologically plausible rules that relate the oldest English words available to their Edenic sources. Words borrowed from other languages (whether they originated in Edenic or not) are simply distractions. Mozeson excuses himself from this level of detail in incredibly problematic ways.

In the introduction, he dismisses vowels from his analysis entirely.
Vowels are certainly interchangeable, and ought to be largely ignored when comparing words from different languages. In effect, Biblical Hebrew has no vowels; the vowel leters in English (A, E, I, O, and U) are chaotic contrivances that help to make English a nightmare to spell. (p. 7, see note at bottom*)
Being no phonetician, I could forgive Mozeson for expressing reluctance to examine vowels but to ignore them entirely is absolutely unacceptable, particularly when one is making such a radical claim. Vowel changes account for the majority of phonological change in language, and that is what Mozeson should be documenting: regular change in the sound system of Edenics that led to English. By leaving out an account of vowel change, he sidesteps a large chunk of the available data and leaves out a great deal of necessary explanation. True, he could tentatively convince others without closely examining vowels (as other historical theories have), but that is not what he says in the quote above: he thinks vowels are "interchangeable", arbitrary. I won't dwell on this point, but its importance cannot be overstated.

Thus, the majority of Mozeson's comparisons rest of similarities between consonants. Many of the most famous and successful hypotheses of 19th century historical linguistics concern consonants, but there is one key difference between these hypotheses and Mozeson's: they propose sets of changes that affect phonemes across all words in a language, Mozeson proposes changes that affect words. The confusion is clear when Mozeson cites one of the most famous 19th century linguists, Jacob Grimm, in support of his methods.
Grimm's Laws, established by the same Jakob Grimm (d. 1863) who brought us those grim and bloodthirsty fairy tales, allow linguists to compare and historically link all letters formed by the same part of the mouth. D and T are called dentals because the tongue must touch the teeth in order to make the sound. Just as M and N are interchangeable nasals or sounds made in the nose, so D, T and TH may be considered the "same" letter. (p. 6)
Again, that word "interchangeable." Neither Grimm's Law nor any other description of sound changes in Indo-European languages provide room to link just any sound with others produced in a similar place or manner. Grimm proposed a cycle of sound changes that affected the entire inventory of Germanic sounds. D and T didn't randomly interchange: Proto-Indo-European /d/ became /t/, just as all voiced stops became voiced stops, and /t/ became /th/ just as all voiceless stops became voiceless fricatives**.

Mozeson's lack of establishing regularity in sound change allows him to link various Edenic sounds to more than one English sound, and vice-versa. It broadens the possibility that he will find correlations: based on its place of articulation, the sound /d/ could be correlated with /t/, both /th/ sounds, and /n/. The same is true for nearly every sound in Hebrew (except only ayin and aleph, so far as I can see). Mozeson does not make all of these leaps, but he has not ruled them out either. This makes his theory unfalsifiable: there's no way for anyone to propose exceptions or identify inconsistencies.

Mozeson also allows suspicious amounts of metathesis of consonants in support of his hypothesis. Metathasis is a linguistic phenomenon in which two or more sounds switch places. As mentioned on the Wikipedia page, in English this frequently happens around r's (e.g., when one says comfterble for comfortable). Some languages make use of metathesis for regular phonological or morphological reasons (e.,g., for inflection or to derive new words from old roots). In all cases, metathesis occurs for a reason, even in the English examples. Mozeson, on the other hand, offers no hypothesis to predict when one can expect to find it. Not only does this make his larger hypothesis less falsifiable, it also makes it overproductive to the point of vacuousness.

Without any specific mechanisms limiting the possibilities in sound change, it's no surprise that Mozeson is able to correlate many English words with Biblical Hebrew words: where there is no look-alike readily available, a small reordering or deletion dramatically increases the chance of finding one.

The foreword and introduction to The Word are full of Abrahamic ideology and jabs at linguists for failing to come to the same conclusions he has.
Because the majesty of Hebrew is only faintly visible in its offspring, it is no wonder that intelligent men can still maintain that most words are arbitrary and meaningless, or that language is the result of cavemen grunting. Of course, some of these same brilliant academicians will insist that a chimp at a typewriter will come up with a Shakespearian sonnet if given enough time (about eight billion years). (p. 4)
He even goes so far as to attribute the accepted account to racism because it separates 'white' Indo-European languages from Afro-Asiatic languages. He simply overlooks that this separation is based on both clear systems of sound changes within families and no clear connections across families despite years of trying. Instead, as with creationist argumentation, there are simply allegations of academics sweeping contrary evidence under the rug.
for the past several decades, Western historical linguists have been the proud Dr. Frankenstein creators of a proto "Indo-European" language that curiously favors the Germanic element. Who would research Hebrew as the root language when even the Ph.D's in Semitics hung Hebrew out on a limb called West Semitic? Nobody uncovered a clay tablet of Proto-Semitic, but surely, the argument went, Hebrew evolved from older more cumbersome languages. The de-evolution of words, and the ongoing corruption of humankind, was simply not considered. (p. 2)
There are no Proto-Semitic tablets because that language (if it existed) was spoken before the invention of cuneiform. Linguists do not seriously consider the "de-evolution of words" in most cases because living languages are constantly growing due to innovation and borrowing. The only languages we regularly see "de-evolve" or become measurably simpler are dying languages spoken in places like Central America where children no longer use the local language in all social contexts.

Like creationists insisting that a global flood is plausible without doing all the necessary math, Mozeson insists that Edenic is a possible ancestor for all human languages without doing the documentation and modeling necessary to demonstrate his case. I can only second what was suggested by Sabzi Aash in this thread:
This is my suggestion for Mozeson. It's the most logical thing to do and the fact that he hasn't done it can only be labeled "suspicious". Forget about English. Just prove that Aramaic or Phoenician or Ugaritic, or even Arabic, comes from Hebrew. These languages are so similar to each other that if one of them were the source of the others, it should be a cinch to demonstrate. If he could show that every linguist had gotten such a basic relationship wrong, that alone would turn the field on its head, and then he'd have a basis to go on to greater claims.
That this has not been done yet is as telling as the lack of paradigm-shifting research coming from the Discovery Institute.

* Biblical Hebrew, the spoken language, certainly has vowels. What Mozeson means is that the Biblical Hebrew written language is based on consonants (although it does have some vowel markings). Except in academia, its rare for new words to be coined based on writing conventions in another language. That Mozeson confuses spoken and written languages is rather telling of the poverty of his linguistics training.
** Other rules were at play. So my saying 'all' here overlooks some mostly-understood complications created by context and other concurrent changes.

No comments: