‘Grammaticalization’ is Well and Alive

Copyright © 2010 Aleya Rouchdy, All Rights Reserved

This post is by David Wilmsen.
I thought his reply on ‘grammaticalizaion’ ( Arabic-L 11/11/2010) would be an addition to my blog since it can certainly be dissected into more than one topic related to the Arabic language, applied or theoretical.


I find the statement of our colleague about grammaticalization to the effect that "this process is not applicable in the Arabic language" to be very curious (although I recognize the source of its motivation).

First, grammaticalization processes have surely occurred in Arabic over its long history (and prehistory). As such, it is appropriate for us to search for a suitable equivalent for the term 'grammaticaliztion' in Arabic, that we may be able to speak about such historical processes without being obliged always to borrow the European language term (in this case, originally French) when speaking Arabic.

I hasten to add that there is nothing intrinsically wrong with using borrowed technical terms - we are all members of an international community and as such may borrow each others' terms when those are more apt than our own. But it would certainly assist those of us who are teaching students whose native language is Arabic to have a set of agreed upon Arabic terms to use with students who are just entering into our community of scholars. One problem with such terms in Arabic is that there seems to be no consensus amongst scholars and other users over the adoption of technical terminology, and even if there were, such terms often remain opaque. Just consider the discussions that break out occasionally over this forum about the Arabic equivalents of various technical terms! It does little good to teach them to students if students themselves find them opaque and no one else understands them.

Second, grammaticalization processes must surely continue to take place in Arabic as a whole, certainly within the spoken vernaculars, which are, after all, also a part of the entity to which we refer when we speak of "Arabic."

What is more, the process can be seen to have taken place in the development of written Arabic. For example, the word نفس has in some contexts lost (or as those concerned with grammaticalization say, it has undergone the “semantic bleaching” of) its original meaning of 'soul' or 'breath' and is now used to mean 'the same' as in نفس الشيء 'the same thing.' This clear case of grammaticalization is discussed in the book Studies in Semitic Grammaticalization by Aaron D. Rubin (Harvard Semitic Series, vol. 57. Winona Lake, Ind.: Eisenbrauns, 2005, p. 19), who says that the “shift of a noun referring to the body or part of the body into a reflexive pronoun ‘self’” is quite common in Semitic, as indeed it is across world languages.

While some purists decry the construction, it is nevertheless used and found to be quite acceptable by many modern authors, such as Ghassan Kanafani, Naguib Mahfouz, Ahlam al-Mustaghenmi and many others, including that paragon of modern Arabic stylists Taha Hussein.

That some modern-day purists decry such usage must indicate that some change has taken place and that the construction would be encountered less often in earlier but still familiar texts.

To satisfy myself about just this issue, which, I admit, has piqued my curiosity since I first heard it mentioned ages ago while sitting at the feet of my then Arabic teacher استاذ فائز of Yarmouk University of Jordan, I searched arabiCorpus (http://arabicorpus.byu.edu) for the phrase نفس ال in the premodern and the modern literature databases (representing texts from all of the writers mentioned above and many more). After eliminating false hits like النفس البشرية and تنفس الصعداء , the results are as follows:

In the premodern database (comprising 912,996 words), the construction appears 103 times.

In the modern literature database (403,901 words), it appears 272 times.

That is, in a modern literature database less than half the size of that constituting the premodern database, the construction appears almost three times as often.

Using the Quran database (for what it's worth, 84,532 words) as a base line (where the construction appears exactly 0 times), we can see that the grammaticalization of نفس proceeded apace from sometime after the 1st century hijri, becoming increasingly more common between the medieval and modern eras of Arabic writing.

That a process of grammaticalization has clearly occurred, even it if has not gone to completion, is evident in that the function of the word نفس has shifted and its inflectional categories when used to mean ‘the same’ have been bleached to a certain extent. For one may use the plural نفوس with words like بشر or its derivations, but one may not say something like نفوس الشيء ** or نفوس الأشياء ** to mean ‘the same things.’ DeLancey discusses precisely this sort of bleaching and functional shift with respect to the English words ‘top’ and ‘finish’ in an article that may be found on his webpage: http://pages.uoregon.edu/delancey/papers/glt.html

In this instance at least, a discussion of grammaticaliztion is indeed applicable to the history of and current usage in the Arabic language.

This, then, contradicts our colleague's even more curious assertion that "any new grammatical change in Arabic is not acceptable." For, whether or not there is a continued grammaticalization of نفس , other processes of grammatical change do indeed take place even in written Arabic (which is the variety to which he seems to be limiting his statement). One such change, that I myself find grating on the ear, is the non-canonical formation of, if you will, a double valence iḍāfa, or what Badawi, Carter and Gulley (Modern Written Arabic: A comprehensive grammar. London: Routledge, p. 136) call “multiple annexation,” by collapsing what at one time and in higher style is still written as two iḍāfas. To illustrate, I cite a very tongue-in-cheek example adduced by one of my translation teachers, استاذ سوسف الزاهد (now, alas, deceased الله يرحمه). He said:

قد قلت لكم مئة مرة إننا لا نكتب « جزمة وشراب الواد » بل « جزمة الواد وشرابه

Excuse the crudity of expression; this was a mnemonic invented by استاذ يوسف for inculcating in his students how the iḍāfa is generally formed in more elevated registers; he meant it to be humorous so that it stayed in the mind (at which it proved quite effective with me!). Simply that he had to use the mnemonic at all, and, as he was wont to say, had to repeat it a hundred times, should provide ample indication that the construction is found to be acceptable by college-educated native writers of Arabic. Indeed, the construction جزمة وشراب الواد - if not the actual words - can be found very often in current Arabic writing, especially in newspapers.

About the process, Badawi, Carter, and Gulley say, “Although in CA [i.e., Classical Arabic] only one element normally occupies the first position, MWA [Modern Written Arabic] is extending the possibilities,” adding that (p. 138), “MWA is increasingly making use of binomial (or indeed now polynomial) annexation, in which two or more 1st elements are coordinated (by any of the coordinators) before annexation the 2nd element.”

Badawi, Carter and Gulley (137-8) adduce many examples of this, including the binomial annexation عقل وضمير الأمّة , which purists would maintain should be عقل الأمّة وضميرها and I myself, although not a purist, prefer. When they get to polynomial annexation, however, the process becomes harder to dispute, for while we can easily rewrite a phrase such as مشروع تطوير وتحديث مسرح البلوون to read مشروع تطويرمسرح البلون وتحديثه , it is difficult to change a rather unlovely name like غرفة تجارة وصناعة دبي to غرفة تجارة دبي وصناعتها without introducing ambiguity, sounding foolishly pedantic, and doing violence to the name of the institution in question itself.

Therefore, pace our colleague, by the principle that usage defines acceptability, such constructions are apparently an acceptable grammatical change to have occurred recently, let us say within the last fifty years, in written Arabic

غصبا عن عيون النحويين وعن عيون الحرصاء علي الصفاء

If a language has ceased to undergo grammaticalization and grammatical change, it is probably dying or already dead.

David Wilmsen
Associate Professor of Arabic
Department of Arabic and Near Eastern Languages
American University of Beirut

------------------------------------------------------------------------