Arabic for Onomasts: Some Notes

Basil Dragonstrike

        While I make small claim to being a scholar of Arabic, over the years I have picked up some facts about Arabic language and names useful to amateur onomasts. Therefore, I decided to publish some of what I know, so that other amateur onomasts will not have to gain this information by picking up bits and pieces across years. Given the scattered nature of my learning, this article can't be logically arranged. I will focus on what is most useful for those using transliterated names found in translations of in-period works. However, some idea of Arabic letters, orthography and language are needed to make sense of transliteration; I hope I haven't overloaded this article with extraneous information.

        We usually refer to "transliteration". But that's misusing the term. Sensu stricto, a transliteration changes each single letter of one alphabet into one letter of another. For Arabic into Latin letters, a common transliteration method is that of Hans Wehr, which uses such orthographic oddities as t, š, ḡ, etc. There are other transliteration methods that use unusual things like d, ḫ, ǧ, and so on.
        What's more widely used is what's properly called "romanization". This allows one letter of an alphabet to be changed into one or more letters of the "Roman" alphabet, which results in fewer odd letters and easier reading. The premier methods of romanization of Arabic are the Library of Congress method (hereafter LOC), and the Encyclopaedia of Islam method (hereafter EI). These are very similar. The most noticable difference is the EI underlines pairs of letters which come from a single Arabic letter, and LOC doesn't. You'll also notice EI's use of "dj" where LOC uses "j", and "ḳ" in EI instead of "q" in LOC.

        Other romanization methods can be divided into two broad categories: modern methods, which usually resemble LOC or EI, and "Orientalist" methods from the early 18th century through to the early 20th.
        Modern methods usually look like they are based on LOC or EI (but may or may not actually be), with idiosyncratic changes.
        Note that some writers drop the dots that occur under some letters (ḥ, ẓ, etc.) and/or drop the macrons that occur over some lettes (ā, ū, ī), or turn them into circumflexes (â, û, î). This can either be because they don't want to bother, or they lack access to a "printing" method that can show the dots and/or macrons. Also: ʻ and ʼ are romanizations of distinct letters in Arabic. Some people use ` and ', respectively, in the place of ʻ and ʼ; the reason is the same as for not using dots and macrons.
        For example, Philip Hitti, in his An Arab-Syrian Gentleman and Warrior in the Period of the Crusades: Memoirs of Usāmah Ibn-Munqidh seems to use LoC, but does not have a single name ending in a long vowel (-ā, -ū, -ī); he changed them into short vowels (-a, -u, -i), as well as a couple of other peculiarities. Note that there are lots of Arabic names that end with a long vowel; I don't know why Hitti made this change. In my collected list of names from his work -- {Names from Ibn Munqidh} -- I had to research the people named to find out which final vowels should be lengthened. BTW, some modern romanization methods aren't based on, nor do they much resemble, EI or LOC; those, however, are rarely encountered.

        The "Orientalist" methods can typically be recognized by the use of o, oo, ee, or aa, and el- instead of al-. These are, in the main, based on the 18th/19th century Egyptian dialect of Arabic as heard and transcribed by Europeans.
        Of course, I'm sure most of my readers don't intend to go to documents written in Arabic and do their own romanizing. But having some idea of how it's done will be helpful to any amateur onomast going through romanized texts looking for names.
        By the way, if you use Google Translate, the "romanization" it uses (small letters in gray, to one side) is very odd; I know of no system like it. Do not use it if you are forming Arabic names.

        You may come across a name with ö or ü in it, or g. This is very likely a Turkish name. The letter combination gl (or ghl) is another clue. If a name ends in -i (hyphen then i) or -e (hyphen then e), it is very likely a Persian name, and sometimes Persian is romanized with a system that uses K, š, o, v, etc. and perhaps carats instead of macrons.

        Given the wide extent of Islamdom, and the Islamic habit of using Arabic names in areas where Arabic was not the native language, mixed-language names were wide spread. Indeed, one generally accept "rule" is that a person must bear the name of her/his father, even if that is a distinctly non-Islamic and non-Arabic name. As well, sometimes a person would keep her/his "native" personal name, while adding Arabic elements to it.
        For example, a person might keep his Turkish name, but give his son an Arabic name, and thus take an Arabic kunya (see below for what a kunya is). Especially in the cases of Persian and Turkish, mixtures with Arabic are well known. This extended to Arabicized Persian and Turkish name elements, and Persianized/Turkishized Arabic name elements. And, what are usually considered single name phrases can combine languages; Nāṣir-i Khusraw attaches the "enclitic particle" -i, found in Persian, to the Arabic name Nāṣir. The Persian patronymic indicator -zāda and the Turkish patronymic indicator -oglu/-oghlu are found attached to Arabic names, and the Arabic indicators ibn and bint are found before Persian and Turkish names. For example, there is the term "Shaikh-zāda" (meaning "son of the Shaikh") found in Ibn Baṭṭūṭa, used by or referring to descendants of celebrated physicians. In other words, these three languages are mixed together freely. But the SCA largely ignores or denies this---just a warning.

        There is one problem-causing letter in Arabic, called tāʼ marbūtah, which occurs only at the end of a word, and almost always after a short "a". It is usually pronounced like "h" (and is audible at the end of a word), but is sometimes pronounced like "t". This happens (mostly) in what is called iḍāfa, which is when two uninflected nouns are placed one after the other, to show the first is "owned by" the second. NB: before suffixes and in compound words, tāʼ marbūtah becomes tāʼ, pronounced like "t".
        tāʼ marbūtah's romanization has long been a problem. The variability in its sound has led various writers to romanize it differently. Some writers turn it into "t" always, others into "h" always, some ignore/drop it always, some write it "t" or "h" depending on the Arabic pronunciation, and it has even been turned into "ẗ". It is the difference between always "h" and always ignoring (the two commonest ways of dealing with tāʼ marbūtah) that lead to the SCA declaring one name could not mix "end in -a" and "end in -ah", on the basis such a name must be mixing "transliteration" methods. This isn't entirely true, but...well, go fight City Hall. BTW, EI and LoC vary how they treat tāʼ marbutah; EI drops it except in iḍāfa, when it's given as "t"; LoC makes it "h", except in iḍāfa when it is "t".

        You'll see lots of name elements starting with "al-". The hyphen in an artifact of romanization; there is nothing corresponding to it in Arabic. In Arabic, ال (al) is a prefix attached directly to the word. In Arabic, the "l" of "al-" sometimes changes pronunciation to match the following consonant; it only does this before certain consonants, but it always does before those consonants. The consonants that cause the change are called sun letters, the others are called moon letters. The "l" doesn't change before a vowel. Both EI and LoC always romanize the "l" of "al-" as "l", but some other systems romanize it to match the following letter; for instance, the Encyclopaedia Britannica does this. To complicate things even more, the "a" of "al-" is not pronounced if the preceeding word ends with a vowel, and some writters change the "a" to ʼ (a hamza), or (rarely) drop it all together. So, a name that in one method would be ʻAlī al-Thānī, in another would be ʻAlī aʼth-Thānī, or ʻAlī 'th-Thānī, etc. To add to the confusion, the word-ending vowel before "al-", if long, often has its pronunciation changed to the equivalent short vowel; some writers reflect this change, which would give ʻAli ʼl-Thānī or ʻAli ʻth-Thānī or ʻAli ath-Thānī, etc.

        There is one type of ism (the personal name) which is called a "theophoric" (or "theophorous") name. This is a name combining ʻAbd al- with one of "The Beautiful Names of God". The most common is ʻAbd Allāh, with ʻAbd al-Raḥmān running second. ʻAbd means "servant, slave" so these names say the bearer is the servant of God, whether Allāh or one of the other names is used. Also in this category is Hibat Allāh, Khalīl Allāh, and others.

        Let's review the various "pieces" Arabic names can be made of. The first piece has no agreed-upon name -- scholars call 'em by a variety of terms -- but I call them "titulars", because, though they aren't titles per se, they are kind of like titles. These are usually of the form _____ al-Dīn, though _____ al-Dawlā, ­_____ al-Mulk, etc. are found. Translated, they describe the person as "(something admirable) of the faith/government/dynasty/etc."
        The second piece, called the kunya, describes the person as the father or mother of someone, usually the eldest son. The masculine form (father of) is Abū (followed by the name of the son/daugher) and the feminine is Umm. There are a few examples of using the name of a daughter rather than a son; more common (though still rare) is a "metaphorical kunya" such as "Father of a Kitten" or "Father of Craziness". A few writers turn metaphorical kunyas into "The One with (a kitten, insanity, etc)"; this is a poor translation, because it loses the poetic element of these names. Note that it is not unusual for someone to be known only by her or his kunya.
        The third piece is the ism, the personal name, or "given name" as the SCA calls it. A few of these are of the form al-_____. Note that these are not bynames being used as isms; I can't recall seeing any such situation. Note that names without isms are not uncommonly found in Arabic writings; the person had an ism, but for some reason s/he is not refered to by it. In other words, it's not unusual to find someone named by his/her kunya (and perhaps a nasab) plus a byname or two, without the ism being used; this situation is where some people get the wrong idea that a byname is being used as an ism.
        The fourth piece is the nasab, the list of forebears. A woman will use bint followed by the name of her father, or very rarely mother, and a man will use ibn followed by the name of his father, or rarely his mother (ibn is often abbreviated "b." and bint "bt."). More generations can be given, by following the name of the father (or mother) with ibn and another name. This can go on for quite a number of generations. Note that, while the person being named might be known as the son/daughter of his/her mother, that mother's mother will not be found in the nasab; the mother's father will be used. Note also that some people are usually called by the first generation of their nasabs. Sometime a generation, or a few generations, may be skipped, to go directly to an illustrious forebear. Usually the father's or forebear's ism is used in a nasab, but sometimes his/her byname (see below) or kunya or even titular will be used. When a nasab uses a kunya, the word Abū becomes Abī. Note that XXX ibn/bint Abī XXX is extremely rare but not unknown (that is, a person is called the son/daughter of the father of him/herself). As well, occasionally a word other than ibn/bint may be used, to show another sort of relation. Words for "brother of," "sister of," "son of the daughter of," etc. are known. However, you are unlikely to come across any of them.
        The fifth piece is the byname per se; at least, that is the term I prefer. Juliana de Luna has called them "Nickbynames". The problem is, these are supposedly two distinct kinds of names, laqabs and nisbas. And they are, supposedly, entirely different from each other. But, I can assure you that every explanation of the difference I have read disagrees with every other, sometimes in very fundimental ways. Frankly, I've given up trying to figure out who's righter than who; call 'em all "bynames", and I assure you that not a single amateur Arabic onomast will suffer thereby. Note that Arabic grammarians say that titular names are a form of laqab; while this may be right, it is a pain in the neck to amateur onomasts, and I recommend most highly that the distiction between bynames and titular names be kept.

        Note that in-period Arabic names could consist of almost any combination of these pieces, though only in the order given above. However, the SCA insists that a "given name" must be present in every registered name, and there must be at least one other name piece.

        By now, you want to know how you can use all this information. Well, there are a number of ways. For example:
        If you see something like Aboo Mohammad el-Tameemee in a book, you now know it's an Orientalist work; such often are poorly researched sources (though a few Orientalist works are still mainstays of Arabic studies).
        Or, if one name source you're using has "Khālid" and another source has "Ghabriyāl", you'll know they're using different romanization methods, because one underlines Kh to show it's one letter in Arabic, while the Gh in the other is not underlined even though it's one letter in Arabic. OTOH, these both are probably from reasonably scholarly sources, and you're likely to be able to find a source with both names romanized using the same method. Or, you can find a table of romanization in one or both books, and show that Khālid can also be romanized Khālid (or Ghabriyāl as Ghabriyāl).
        Say you run into a name containing the byname al-Tabrisi, you would be wise to consider this might not be following all the rules of LOC or EI, because the vast majority of name elements ending in a vowel end in that long vowel. And sure enough, more digging will reveal the LOC and EI romanization methods show this name as al-Ṭabrisī---with not only a long final vowel, but "Ṭ" not "T".
        If you were looking through a book and found plenty of romanized words ending "-at" but almost none with either "-ah" nor "-a", you now know enough to guess the writer has chosen her/his method of dealing with tāʼ marbūtah--namely, to always romanize it as "-t". With that information, you could look for another source of that/those name(s) that uses "-ah" or "-a", or, in your documentation, argue for changing the name element from "-t" to "-ah" or "-a", should you choose to.
        And, if you're confronted by the name Shams al-Dīn Abu 'l-Muẓaffar Yūsuf b. Qīzoghlu Sibṭ ibn al-Jawzī (a real person, BTW), you can figure out a number of things: first, Shams al-Dīn is a titular name, so you couldn't use it in the SCA. Second, Abu 'l-Muẓaffar is his kunya, and this source from drops the "a" in "al-" after a vowel, and shortens that vowel if long---potentially useful information when appraising the rest of that source. Yūsuf is his ism. b. Qīzoghlu is the first generation of his nasab, and it looks somewhat Turkish (remember, -oghlu is a Turkish partonymic marker). Now Sibt ibn al-Jawzī looks odd---if this is a list of ancestors, you would expect an ism of a predecessor should have "b." or "ibn" before it, so why isn't it ibn Sibṭ ibn al-Jawzī? If you do some research, though, you'll find out that "Sibṭ" is one of those "other relationship" words I referred to above. Specifically, it means "son of the daughter of". Combined with "ibn" we have "son of the daughter of the son of". So, this person was the son of the daughter of the son of someone called al-Jawzī; this part of the nasab is based on a byname rather than an ism. BTW, this person is usually refered to by historians simply as Sibṭ ibn al-Jawzi.
        Anyway, this name will document the Arabic names al-Muẓaffar, Yūsuf and al-Jawzī; it will also show Qīzoghlu as a (very likely) Turkish name used in an Arabic context.

        Of course, one thing that would be of enormous help to an amateur onomast would be lists of historical names. Especially in the SCA, the works of Da'ud ibn Auda and Juliana de Luna are often referred to in this regard. While I admire both of them enormously (indeed, I would never have been able to start my own research without the guidance of their works), it must be admitted that the increase in available scholarly books since their time, and the tightening of the standards of the SCA, mean that their work is a bit "behind the times". Especially note that Da'ud ibn Auda gives neither sources nor dates for any of his names, and Juliana de Luna took her names from a very obscure Spanish work that seems not to have given any dates---the only date connected with the work is "700-1200".
        I have gathered my lists of names from specific works, with known dates of composition, or from a highly respected source and that gives dates for the people concerned. Therefore I can, I believe, without undue pride, point to my lists elsewhere on this site as being at least as useful as the efforts of Da'ud ibn Auda and Juliana de Luna. Also, I intend to keep looking for other translated Arabic books, to extract names from, and to keep writing articles about matters useful, interesting, or both.

        Well, I could go on (and on and on and...), but I have only so much room, and I may as well stop here. I hope this will be of use to you.

Return to "General Introduction"