Etymologies and Then Some
(This is a response to the comment on the previous posting.)
Paul, thanks for your confidence! There are still many problems with etymology, though, even within eclectic and poetic nonlinear approaches...
First, the total number of roots is probably lower than 8000--waw-waw-waw, for example, wouldn't work as a root, and there are tons of other nonsensical combinations of three consonants. On the other hand, there are some quadrilateral roots, and also some new roots that started out as nouns borrowed from Turkish, Persian, French, etc., that were then "Arabized" into a morphological standard, even though many of these haven't been accepted into formal Arabic yet. Finally, there are tons of loan words that keep getting rejected, but many have slipped through into orthodox texts after many years. Many of these are nouns, and are not Arabized and elaborated into verbs and adjectives, and thus their etymologies are a lot easier to spot.
One interesting related process is the creation of "false etymologies." Because no colloquial system that I know of reproduces all the consonants of Arabic "faithfully," many differences in written form are masked in colloquial pronunciation. For example, the dhal and tha sounds (voiced and voiceless linguadental fricatives?) turn into very strange and distinctive z and s sounds in Lebanese Arabic. In Moroccan Arabic, they much more regularly turn into simple d and t sounds, even though these already exist in formal Arabic as dal and ta. Lebanese like to argue that even though they say tha wrong (being too influenced by Aramaic), they're still less wrong than Moroccans, who would not be able to distinguish tha from ta in everyday speech, apart from context. In other words, in all spoken Arabic you find some homophones, and in everyday etymologies these distant roots are joined because they sound alike. There's a long and celebrated history of some notorious French and American scholars who worked in North Africa without bothering to learn how to read and write Arabic. They got snookered into buying whole-heartedly into some of these everyday etymologies and they went on to elaborate some really charming little theories. Over the centuries, these scholars have provided the Arabic literati with no end of amusement.
Further, for hundreds of years, Arabic linguists did not really pay much attention to the Arabian peninsula (where today pronunciation of spoken varieties continues to diverge sharply from formal Arabic), and were instead based in Baghdad (Persian influence), Damascus (Aramaic influence), and Cairo (Coptic influence). As a result, sometimes the semantic boundary lines between roots are quite obscured, and implications and textures cross back and forth. Even for Arabists, important questions remain unanswered.
For example, in Impasse of the Angels (Chicago, 1997, p. 269), linguistic anthropologist and Arabic language scholar Stefania Pandolfo comes across the Moroccan word "frag" and explains its origin. Since g does not exist in formal Arabic, we must search for the third radical in the root: F-R-?. Egyptian Arabic gets its g from jim (J), but Gulf Arabic gets its g from qaf (Q). Since in Moroccan Arabic "frag" implies "a gap, a separation," Pandolfo proposes F-R-Q, to separate, part, or divide. Since Morocco is closer to Egypt than to the Gulf, though, when I read the word I was thinking about F-R-J, which means to open, part, separate, or cleave. While FRQ and FRJ seem to be related, I thought FRJ was a more likely root, since associated derivations show the emphasis in FRQ to be on the sides that have been divided, while the associated derivations of FRJ seems to focus more on the empty space itself. But then I thought of ghayn. What about F-R-Gh?
I thought of ghayn because I have found that many French speakers, including Moroccans, tend to mispronounce this velar fricative as a uvular fricative ("The French R"), pushing it farther and farther back from g. This tendency has had direct implications on my little field of Gnawa studies, since the "uvular ghayn" obscures the relationship between "Ghana"--pronounced in Arabic the way it's written, "ghana"--and Gnawa. For that matter, such arguments also separate the spelling of Ghana from the way it is currently pronounced by its citizens: "gana." In other words, French speakers tend to discount a relationship between gh and g, because they see a closer relationship between gh and r. Which is pretty far-fetched from a strictly Arabic perspective, where the ra is a rolled r, as in Spanish.
But F-R-Gh is also related in meaning to FRQ and FRJ: to be empty, void, to be vacant, used up, exhausted. Here, the focus is exclusively on the empty space, and its condition upon being emptied.
So there appears to be some relationship between the three roots, and moreover it appears to progress from fullness to emptiness or from wholeness to fragmentation, FRQ to FRJ to FRGh. Now I'm wondering whether F-R-G wasn't some original root in another language, perhaps Coptic, which was mispronounced by subsequent Arabic speakers who created FRQ, FRJ, and FRGh as ways to compensate for their inability to reproduce the g sound. Or maybe, horror of horrors, the origin is the Latin fragmentere... (And this is not out of the question, since of course most Medieval Arabic linguists were conversant in Greek and Latin, and many of the people who became Mediterranean Arabs had contact in one way or another with speakers of these languages.)