The third chapter of my book "Sounds from the Garden of Eden" discusses the theme of language relations. The common, widely accepted theory tells us that almost every language is a part of a family of languages. Like, for instance, English belongs to the Germanic family, which by itself is a part of the Indo-European family.
Every student of linguistics or anything that has to do with languages studies this theory as if it were an absolute truth.
The theory of language family was invented two hundred years ago to be applied to the language families then known, the Indo European and Semitic. later all languages of the world were assembled in families, and also in macro families, like the Afro Asiatic which connects 6 different families of languages, 5 from Africa, and the Semitic.
The theory principle is very simple. In the deep far past there was a language spoken by a quite small group of people. This language was split to two or more daughter languages, and those in their turn split to more granddaughters, and more splits were carried on until the contemporary state of language map has gained its form. The connections between the languages of a family are usually drawn to show a "family tree".

Here is the academic "family tree" of Semitic languages:      

  Though widely accepted, still this theory suffers weaknesses. For instance, this scheme shows at the same level the Ugaritic, the Canaanite, the Aramaic, the Arabic and the Ethiopian. But Ugaritic was spoken through the second millennium B.C. and passed away at about 1200 B.C. Canaanite, which comprises Biblical Hebrew and Phoenician was spoken starting two hundred years later, and Arabic appeared at the 5th century A.D. So why showing all this languages at the same level? And when and how did each language have connections with each other, if ever? the "family tree" is wrong in most  important aspects of the history of the languages.

The third chapter of "Sounds from the garden of Eden" suggests a very different principle to describe and explain that history.
The great similarity between the various Semitic languages misled scholars to conclude that all these languages branched out from one ancestor. Yet here, we must be guided by
the Fundamental Rule of Languages History:
every language, without exception, is the result of merging features of earlier or neighboring languages.

This scheme shows the connections over time between the languages, while the thickness of lines suggests the assumed connection strength.

0.  Laguage system in the area where Semitic language were spoken, preceding the earliest evidence.
1 . Earliest Semitic dialects.             undocumented

2.  Old Akkadian, Eblaite,  2700 B.C.E. 2350, B.C.E, Contributing languages: Mainly extinct Semitic dialects Sumerian, other languages.
4.   Akkadian (Babylonian, Assyrian).   2000 B.C.E. Contributing languages: Old Akkadian, Sumerian, other Semitic dialects.                                                                                                                                  
5.  Ugaritic.  1500 B.C.E. Mostly Semitic dialects, Asia Minor languages, Akkadian.

7.  Hebrew and Phoenician 1000 B.C.E.. Biblical Hebrew. 700  B.C.E.   Cotributing languages:  Semitic dialects, Ugaritic, Egyptian, Archaic Greek, Hittite.
8.  Old Aramaic. 900 B.C.E. Contributing languages: Mainly extinct Semitic dialects, Akkadian, Hittite.

9.   Trans-Jordanian and Arabian desert languages. 900 B.C.E. Contributing languages:   Hebrew, Aramaic, other Semitic dialects.
10.  South Arabian languages. 500  B.C.E. Contributing languages: Mainly Semitic dialects, Akkadian, probably African languages.
11. Classical Arabic. 500 C.E. Contributing languages:  Most of contemporary Semitic  dialects, Persian,      other contemporary languages.
13.  Ge'ez  400 C.E. Contributing languages: South Arabian and other Semitic dialects, African languages.
The main insight that should be read in this scheme is that the languages did not branch from a single common stem, but instead, each language is always, the consequence and result of of mergers of traits of, from, other languages. And when we look at the history of any family, we must bear in mind that the similarity of the languages is the result of mergers of close dialects.
When Semitic dialects merged with a non Semitic language, namely Sumerian, (and later with African languages)  the resulting language, the Akkadian, showed  many differences compared to all other Semitic languages: Akkadian lost all guttural consonants, unlike any other Semitic language, and the verb in Akkadian, just like in Sumerian is at the end of the sentence, opposing all other ancient Semitic which start the sentence with the verb.
What does the "tree" explanation do with such facts? Nothing at all. It puts Akkadian at east and all other languages at west, pretending to have resolved this problem.
By no means is it true that Classical Arabic of the 5th century C.E. has the same roots in the ancient extinct Semitic dialects as Old Akkadian or Ugaritic. The traits that seem common to the languages of different generations were transformed through mergers which always inflict changes, and the seemingly sameness of  characteristics is a kind of deception that scholars are fed with due to use of written characters that camouflage the great differences of real speech.
Here I have dealt with Semitic languages, but the same insights must be taken to all language families. There in no exception:
every  language, without exception, is the result of merging features of earlier or neighboring languages.
The "family tree" must be substituted with a scheme that takes into account all traits of a language, not excluding traits introduced to the language by ex-family languages, because in many cases, like that of Akkadian, such influence is fundamental.


  1. Classical Arabic should be defined as the Arabic of the age of classical Arabic literature, roughly the 8th through 13th - or 14th or 15th -centuries AD. Its precursor - pre-classical Arabic - may only have appeared around 500 CE, but there must have been a people (or peoples) who either called themselves Arab or were named so by others. The first mention of Arabs comes from 853 BC in the Kurkh monolith of Shalameser III, which enumerates the members of alliance of peoples against Assyrian expansionism, including a certain Gindubu of the land of the Arabs (gin-du-bu māta arba-a-a). One hundred years later, Assyrian records place these arba-a on their southern borders in the Haurān plateau (incorporating the southern regions of modern state of Syria - including the Golan - and northern Jordan). These people may or may not have spoken a dialect related to or the parent of classical Arabic, but they were probably speaking a language that would be called the language of the arb-a-a-a or Arabic. So your contention that Arabic and the other Semitic languages arose as dialects of each other and that they absorbed influences from the other Semitic varieties in which they came into contact (often intense contact) and from surrounding languages is no-doubt true - but certainly not a new way of thinking about such things. But its implying that Arabic is no more than 1,100 years old cannot be true. So too is it mistaken in assuming that the modern Arabic dialects descended from that particular variety of Arabic, probably already an inter-dialectal variety of elocution, oracle, and oratory by the 5th century AD and not the native tongue of anyone.

  2. Thank you for your comment. That an Assyrian monolith has the name Arab does not make a language. Surely there were Semitic dialects close to known Arabic of 6th century C.e. but we can be quite sure they differed from Arabic no less than Biblical Hebrew differed from the pre-Hebrew language known to us by the Amarna letters. My use of "classical" Arabic was meant to difer the written Arabic from any ancient or modern dialect. Have you read cautiously you coud not have concluded that I suggested that modern Arabic dialects descended from that particular variety of Arabic. My main point is that any language is always, with no exception a result of mergers of traits from other, older and contemporary languages, that meet at some area. And by that my attitude differ and new compared to the "tree". There has never been any "family tree", and therefore Arabic cannot be put at same level with much more ancient languages.


Post a Comment

Popular posts from this blog

meaning, not syntax, is the essence of language