Wednesday, April 27, 2011

An atom's weight of philology

One of the oldest motivations for studying the history of language is to better study the fixed texts of holy books or classics. We try to learn from such texts, but without an understanding of philology we misread them - because, while the words have remained the same, their content has changed. Ibn Quraysh is one case in point; Ruskin offers another:
"[I]n languages so mongrel of breed as the English, there is a fatal power of equivocation put into men's hands, almost whether they will or no, in being able to use Greek or Latin words for an idea when they want it to be awful [ie impressive]; and Saxon or otherwise common words when they want it to be vulgar… [C]onsider what effect has been produced on the English vulgar mind by the use of the sonorous Latin form "damn", in translating the Greek katakrínō, when people charitably wish to make it forcible; and the substitution of the temperate "condemn" for it, when they choose to keep it gentle; and what notable sermons have been preached by illiterate clergymen on - "He that believeth not shall be damned"; though they would shrink in horror from translating Heb. xi. 7, "The saving of his house, by which he damned the world"… "
Standard Arabic has no layer of prestige loanwords corresponding to Greek and Latin words in English - all the classics of the Arab world are themselves in Arabic, and great efforts have been expended to keep the grammar of Standard Arabic roughly constant since the pre-Islamic era. But, thanks to the many new meanings conferred upon old terms during episodes of massive translation - both in the modern era and the Abbasid era - it is fairly susceptible to another of Ruskin's complaints: misinterpreting the words of old texts thanks to their modern meanings.

Once a medical student at Cambridge told me in all seriousness that the Qur'ān anticipated modern science by centuries in mentioning the "atom" (فمن يعمل مثقال ذرة خيرا يره, "for he who does an atom's weight of good shall see it")! Of course, every modern educated Arab knows that a ذرة dharrah is an atom. But looking at a pre-modern dictionary, such as Lisān al-`Arab, gives a rather different picture: a dharrah then was a type of small red ant, a weight equivalent to 1/100 of a barley grain, or a mote of dust (as seen in sunbeams), not an elementary particle of which all matter is composed. In parts of Sudan the first of those meanings is still in regular use: dirr there means a type of ant. But elsewhere they all seem to have faded from away from popular speech.

If I were interested in an English word, I could easily look it up in the OED and find a complete history of its different meanings and the dates at which they were attested. But for Arabic no such dictionary exists; to figure out when and how dharrah came to mean "atom" in the modern sense, I would have to look through a bunch of pre-modern works, or find an article on the subject. It's a gap that would be well worth filling.


John Cowan said...

Lameen Souag et al, eds. A New Dictionary of Standard Arabic On Historical Principles (2081).

David Marjanović said...

Is there any language other than English with such a dictionary?

In German, there's Grimm's, which is now being updated for the first time since Grimm himself. Standard German has changed immensely since then; Standard French looks completely frozen in comparison, and not just because it already had a standardized orthography that practically hasn't changed for several centuries.

David Marjanović said...

...I'm trying to say the published version of Grimm's is pretty much useless and probably hasn't even been reprinted since ever.

Anonymous said...

Learning a language is a defying activity.

We write in a blog about topics related with different foreign languages.

We would very much like to get your comments!


Qifa Nabki said...

Hi Lameen,

I've been enjoying your blog for a while. A historical dictionary of Arabic is one of the great desiderata for Arabists. In this age of electronic databases (cf., such a thing may actually be conceivable.

The continuation of Lane by Manfred Ullmann ( Wörterbuch der klassischen arabischen Sprache) is probably the closest thing we have, for now.

parviziyi said...

If I want to know the history of an Arabic word, I can look up its medieval meaning online in a few keystrokes at That site has searchable copies of the large Al-Sihah Arabic dictionary of Al-Jauhari dated about year 1000, the giant Lisan al-Arab dictionary, the Qamus al-Muhit dictionary, and several other medieval Arabic dictionaries. When I want to know the meaning of an Arabic word in the 19th century, I can get it very quickly by looking it up in the 1852 edition of Richardson's Arabic-English dictionary, which is 1500 pages long, and is free to download at ( also has downloadable copies of the 8-volume work by Lane, but Lane is mostly redundant given the others I just mentioned).

Therefore I think a hypothetical New Arabic Dictionary on Historical Principles would be a lot of work for little reward. The additional info it would contain would not be of great historical value.

By the way, the creators of the OED read few medieval English sources and consequently the OED has a great many errors about attestation dates. To appreciate that that's true, pick a dozen words at random and look them up for attestations in the Middle English Dictionary (which is free online).

semi-expert said...

While agreeing with your assertion of the sore need for an etymological dictionary, I cannot agree with your statement that "Standard Arabic has no...prestige loanwords....great efforts have been expended to keep the grammar of Standard Arabic roughly constant since the pre-Islamic era."

Old Arabic contains many borrowed words. A few examples are these:

ṣirāṭ 'path' < Aram isṭrātiyā < Gk stràa < Lt strata
burj < 'tower' < Syr būrgā < Gk púrgos
ṣalāt 'prayer' < Aram ṣlōṯā
tīn 'fig' < Aram tīna
sifr 'book' < Aram sifrā
ṣanam 'idol' < South Arabian ṣnm, Safā'itic ṣnmt
ṣaḥīfa 'page of writing'/ṣuḥuf 'scriptures' (from which muṣḥaf) < South Arabian ṣḥft < Ethiopic s' ḥ f 'to write'
zanjabīl 'a well in paradise' < Syr zangabīl < Pahlavi singaḇēr 'ginger'
(many botanical terms come from Persian, e.g., Ar warda 'rose' < Aram wardā < Avestan varǝḏa

All of these words are found in the oldest attested Arabic, being Jahaliyya poetry and the Quran. Many volumes have been written about borrowed worlds in the Quran, including by native Arab commentators from the classical era of Arabic writing.

I suppose by "grammar" you mean syntax. As this is a posting about etymology, I shan't divert the discussion away from borrowed words; suffice it to say that, despite the great effort (about which you are indeed correct) the syntax of written Arabic has been changing, albeit slowly, from its earliest days until now. Such change can be easily detected. Anyone interested may look here for a demonstration:

semi-expert said...

sorry, that should read Gk stràta <

Lameen Souag الأمين سواق said...

"Standard Arabic has no...prestige loanwords" means something entirely different from what I actually said, "Standard Arabic has no layer of prestige loanwords corresponding to Greek and Latin words in English." The former statement is uncontroversially false; the latter, uncontroversially true. Arabic has no common pattern of near-synonyms in which one synonym is a prestigious loan and the other is a non-prestigious native form, such as Ruskin discusses. And even the individual loanwords that are prestigious come from a variety of very different languages and lack any conspicuous unifying factor, in contrast to English Latinate vocabulary, or Persian Arabic-derived vocabulary.

Some of the specific etymologies you cite, however, are a bit problematic. Safaitic is more closely related to Arabic than to Old South Arabian; if s.anam involves an *l > n shift rather than vice versa, and if this shift only happened once, in OSA, then it must be a borrowing in Safaitic too (and if either of those assumptions isn't true, then on what basis can we conclude that it's a loan?) Likewise, Ethiopic, whose writing system derives from OSA, hardly seems likely to be the _source_ of OSA writing terminology (although Arabic borrowing such terminology from OSA is eminently likely.)

Berkman Center for Internet & Society said...

Dear Blogger,

Hi! This is an invitation from the Berkman Center for Internet & Society at Harvard University and Global Voices Online to participate in a survey we're carrying out jointly. We're asking questions about online safety for bloggers in the Middle East and North Africa, and we need your help!

We are particularly interested in learning about your experiences with social media, the steps you take to protect your privacy online, and your perceptions of online threats. We are sending surveys to approximately 600 influential bloggers throughout the region in the hopes of learning more about how bloggers view and approach the issue of online safety.

The survey will take approximately 25 minutes to complete. We will not share personal information — or the fact that you've participated in this survey — with anyone else, and you’ll be the first to know when we release the results of the survey. Will you help us?

Survey link:

With appreciation,
The Berkman Center & Global Voices Online

semi-expert said...

To be quite honest, I wasn't sure what was meant by "layer of loan words" Now that you explain it, the question is still debatable. The dichotomy you speak of is one between Germanic words and Latinate words, quite true, but by way of French not by way of the coining of Latinate technical terms .

There is of course an analogous situation in Arabic: the dichotomy between fusḥā and 'amiyya, where indeed there do exist vulgar alternatives to the rarefied "standard" forms, some of them native some not. The one difference is that here the borrowed forms are devalued.

semi-expert said...

Along, I should add, with the native alternates.

languagehat said...

There is of course an analogous situation in Arabic

It is not analogous at all; I think you still don't understand what Lameen is saying.

The one difference is that here the borrowed forms are devalued.

Well, that and the fact that the "borrowed" terms are not borrowed, which kind of cuts the ground out from under the attempted analogy.

semi-expert said...

That may be so. In re-reading the original entry and in re-reading the blogger's reply to my earlier post, I still have difficulty in understanding precisely what he is trying to say. Except that it is a preface to calling for an etymological dictionary for Arabic, an enterprise, you may notice, that I heartily endorse.

As to your remark, a great deal of the literature on code-switching in Arabic between oral approximations of the written language and the matrix language (that being a spoken vernacular) treats words, often but not always technical terms of all varieties, coming from the more formal code as borrowings, especially when they become lodged in the regular discourse of matrix language, effectively becoming part of a vernacular. Perhaps this is not exactly analogous to borrowings into English from Latin and Greek, but most analogies are not exact, and it is still legitimate to call such terms borrowings.