Arabic and Hebrew Compared

OBS: first part is a longer, more complete version in English. Below it, you can find the shorter Portuguese version

Arabic and Hebrew Compared

And a Brief Introduction to Semitic Linguistics

By Maurício Carvalho - Last updated August 19th 2000 -

Some updates March 25th, 2001 (William Chomski recognized as Noam’s father, thanks to e-mail)

Back to Main Page

Translated into English from the original in Portuguese on August 2000, and then enlarged.

I wrote this essay basically to learn, so every comment, criticism and correction is welcome, and although much of this essay is a summarized reproduction of what I read, some of the conclusions and explanations are mine. E-mail me.

This Israeli guy was kind enough to e-mail me some corrections of this essay.

Download Hebrew fonts here

Click here for an introduction to the Arabic and Hebrew alphabets (not very good yet)


Why and How I wrote this Essay

Phonetics and Transliteration

An Overview of the Semitic Languages


Nouns and Articles

Adjectives, Pronouns and Adverbs

The Construct State



Appendix 1: Etymology of Biblical Names from Hebrew and Aramaic

Appendix 2: Etymologically Related Words in Arabic and Hebrew

Appendix 3: Portuguese Words of Arabic Origin

Appendix 4: Gentilics and Toponyms

Appendix 5: Semitic and Indo-European

Why and How I wrote this Essay

The Author

My name is Maurício Carvalho, I live in Brazil and am 24 years old. I study languages for the sake of it and attend journalism at the Catholic University of my hometown, Santos, in the State of São Paulo. I work as a language teacher and a tourist receptionist at Santos’ Tourism Bureau.

The first version of the present work was written by me in february 1999; it contained but a highlight on some of the features of Arabic and Hebrew and was the result of my first attempt at unravelling the mysteries of these important languages. I just wanted to understand what it meant a language to be “Semitic” and to discover by myself how close both languages were from each other. I started off with these two languages not only because they are the most important of their group, but mainly because they were the only ones whose grammars and dictionary I had access to, having purchased several books since as early as 1995. My first was the very didactic “Every Day Hebrew” (NTC publ.) at a bookstore called Hugendubel in downtown Munich, Germany. But then I was too unexperienced at linguistics, I had just learned German and was beginning to explore the Slavic languages, an effort whose reasonable result can be appreciated in my 1998 “Polish Grammar”, which was the reason I decided to open my own site and discuss linguistics on the Net.

Until early 1999, I had just learned the Hebrew alphabet (and not too well) and a few words. In January that year I bought at Livraria Cultura in São Paulo (my favorite!) the excellent Teach Yourself Arabic, but as I didn’t believe in my own capacity of learning the Arabic script (even though I had managed to master the more intrincate Devanagari of India) I had to learn for a long time from the very few transliterated examples in that book. As destiny works in mysterious ways, on May 2000 I suddenly got the feeling that I had to learn Persian. A few weeks after buying Routledge’s Colloquial Persian for that purpose, I was already able to comprehend Persian at least in the basic level. Enthusiastic with my newly discovered capacity of learning Asian languages (yes, I know Persian is a not very intrincate Indo-European language, but leaving Europe is exciting even then...) I asked my Brazilian-born Lebanese classmate to return me my Teach Yourself I had lent her some eight months earlier for her to put an end to her illiteracy in her native tongue, so that I could suppress my Colloquial Persian’s greatest flaw: it doesn’t teach the Arabic alphabet, which that language uses since the Islamic conversion over 1,300 years ago. As I said, it was destiny, because she actually returned the book! - and without learning how to write... This doesn’t happen to often in the world, people returning borrowed books... Well, back to the point, as soon as I got that book back I started to learn the script.

But that time I did it - in some five hours’study I was able to remember most letters, combinations and special signs. I could READ ARABIC! If you know the language you can guess that I don’t even know where my Persian book is right now... No offense to Iranians, their language is fabulous and I intend to return to it, but for an amateur Brazilian linguist like me, the possibility of learning Arabic seems far more exciting than that of learning Persian, which does not present many differences from the other Indo-European languages I already knew. Every day then, almost all of my free time was spent sucking all I could from my Teach Yourself and some other books I had bought - notably the “Arabisch-Deutsch Taschenwörterbuch” by Langenscheidt, perhaps the only good beginner’s Arabic dictionary. In about two months, I was already familiar with most of the basic structures of the language and had a 300 + word vocabulary.

I could then turn my attention to that old idea of mine of finding out what being Semitic meant. To that I added new interests such as Biblical and Koranic studies, as well as the origins of the alphabet. The stage was set, and with some help from the scarce Internet resources I could find and the on the other hand vast resources of the São Paulo University library (unfortunately more than two hours away from where I live), I am now able to strech my muscles and give the world what is perhaps the only scientific comparative insight on Semitic languages on the Internet (hey, language-lovers, wake up! There is a lot on Indo-European, but the world is much, much bigger!).


My main linguistic resources to write this grammar were the Précis de Linguistique Sémitique, translated into French from Carl Brockelmann’s book, providing a fresh insight into comparative Semitic linguistics; Gesenius’ Hebrew Grammar, a very complete compendium; Syrische Grammatik, also by Brockelmann (an edition from 1899!), about the literary Eastern Syriac-Aramaic; a Grammar of the Phoenician Language, by Zellig S. Harris (1936), all of whom I could consult at the São Paulo University (USP) Library. I am slowly digesting them into my essay.

If you know of any better resources, let me know.

The things I learned with the present effort were:

- One of my main surprises was finding out that Arabic is more complex and conservative than Hebrew. I really expected that Hebrew, being much older as a literary language and having being kept artificially and zealously for over 2,000 years after its extinction without suffering the transformations of living languages, would be closer to a common Semitic speech and probably morphologically richer than Arabic, which was first written down only after Christ and was subject to a vast expansions to non-Arab countries, therefore being theoretically more subject to changes. Its true, as I soon found out, that the spoken forms of Arabic are simpler, sometimes getting closer to Hebrew, but even then Arabic is far more intrincate than the language of the Jews.

- Arabic script is does not match exactly the Hebrew script, neither do its sounds. One mistake that prevented me at first to comprehend the nature of Arabic script was the idea that each Arabic letter had a reasonably precise Hebrew correspondence with similar original phonetic value. This is not so because Northwest Semitic languages had already in ancient times lost many of the sounds of Proto-Semitic which were more accurately mantained in Arabic, such as the several emphatic and laryngeal letters. The Arabic letter alif, which I coudn’t think of as anything other than the exact match to Hebrew alef, is actually a totally mute letter, having a mere orthographical role, athough both developed out of the same Phoenician alep. Alef is the soft glottal stop, and later I discovered that that role is played by the Arabic letter hamza.

The fact is that the Arabic script is not a sibling of the 22 perfectly matching Hebrew and Aramaic scripts, but its a cousin, being descended from a script known as Nabatean, this one being a son of the Phoencian alphabet like Hebrew and Aramaic. The precise evolution of the Arabic (and South-Arabian) scripts are not well known, but it indeed have gone much farther from the parent alphabet.

- Modern Hebrew is not much different from the Biblical language. Modern writers of Hebrew may as well use exactly the same constructions found in the Bible, though this is regarded stilted. The modern colloquial or informal usages of the language were shaped in the bosom of the Mishnaic schools of the later ages through which Hebrew adopted a more European character, mainly in Syntax. For example, the old perfective and imperfective conjugations, originally not connected directly to tense, became synonym with past and future in Mishnaic tradition and thus is adopted in the modern usual language. Pronunciation changed a lot, since after the dispersion of the Jews throughout the world the separate communities developed each their own rules. Modern Hebrew pronunciation is based on the tradition of the Sephardic Jews (correct me if I am wrong!). Another insteresting remark on Hebrew is that the language actually never died, it was used by many communities as a means of oral communication, though never as a first language. The struggle of some hard-minded 19th century Israel settlers, especially Ben Yehudah, to establish Hebrew as the sole language of the land was a bitter one, however. Then a part of the Otoman Empire, the Holy Land almost became a land of German, French and English as it was settled by European Jews... The history of the revival of the language is too rich for me to even summarize it here, but I can recomend a booklet called Hebrew: The Eternal Language by William Chomsky (Noam’s father), which taught me a lot.

As for Arabic, it was to my surprise to discover that the language is divided, in its spoken forms, in tens of dialects, although education and written communication is conducted in a common language not too far from that of the Koran. The dialectal variations of the language and its several usage levels are also a matter for an expanded explanation, which I will try and provide as I move on in my research.

1. Phonetics and Transiteration

As a preliminary remark I have to say that, since I learned all of this from books, my Phonetics knowledge is rather weak and imprecise. This is only a rough guide - well, hm hm, like most of my essay!

1.1 Pronunciation and Transliteration

Due to the complications involved in writing Arabic and Hebrew on a computer, and in order to provide those who are not (and don’t want to be) familiar with their scripts with a simpler way of reading this outline, I have devised a straightforward but precise system of transliteration.

1.1.1 Arabic

Arabic only recognizes three vowels: [a], [i] and [u], either long or short, and two diphthongs [ay] and [aw]. Long vowels are nearly always shown in script, whereas only initial and post-initial hamza short vowels are.

The letter alif serves usually in the middle of a word as an indicator of long [a], as well as a carrier of the letter hamza. The hamza goes on top of the alif if followed by [a] or [u] and beneath it if followed by [i]. Followed by a long [a], the alif loses the hamza and gain a broad tilde over it.

The letter waw indicates long [u] and diphthong [aw] and the letter ya indicates long [i] or diphthong [ay]. Then words written mwt may be things like [mawt], [mút], [mawwat], and the word Tayyib, written Tyb, could be Tayib, Tíb,Tayb... you can only know if you know.

Long [a] is missing in a few words, which had originally an alif written above them (alif superscript). The main ones are hádha (this) and lákin (but).

Words ending in my transliteration in [-ay] actually have this ending pronounced long [a], and are spelled with a ya.

Except after emphatic letters (S, T, D, DH, H and q), [a] and [ay] regularly become [e] and [ey]. Sometimes [u] is heard like [o] and [aw] like long [o].

In the examples, long vowels will be represented by á, í and ú. Except, of course, in the case of final [á] from a ya spelling. Doubled consonants are indicated in didactic script by the sign shadda, but left normally unmarked in the regular press. Here they will be transliterated doubled.

And when doubt may arise, the n of nunation (see Declensions of Classical Arabic) will be marked as ñ, and the a of the feminine ending will be seen as à. Dialectal Pronunciation

1.1.2 Hebrew

Transliterating Hebrew is apparently easier than Arabic, but appearances are indeed deceptive. Although the language has fewer sounds and letters than its southern sibling, its orthography has never been efficiently standardized and there are mute alefs and vavs flying in all directions at everyone (well, almost). I will try to represent through my Hebrew transliteration the full spelling of the original words.

In addition, there is a problem I only started thinking about after having written much of the present work, and is therefore a source of imprecision. It’s the fact that some Hebrew letters become altered when preceded by a vowel. These letters are b, t, k, p, d and g. Altered, they originally may have sounded like aspirated consonants (i.e., followed by a h-sound). The Biblical tradition (at least that of my Teach Yourself Biblical Hebrew) teaches that they become respectively v, th (like thick), kh (like Bach), f, dh (like this) and gh (like Arabic gayn). The thing is that Modern Hebrew phonetics only softens b, k and p (to v, kh and f), and many examples I took from Modern Hebrew grammars have been left with no softening of the other letters shown. But this is not very important since softening is mechanically predictable once you know the vowelling.

The final h is only pronouced when doubled (in Hebrew pointed script, it gains a dot inside it, called daghesh forte).

Like in Arabic, Hebrew vowels remain mostly unwritten. However, unlike in its sibling, the rules for vowel carriers are quite crazy. Long [a] is seldom shown by a carrier at all, where as long [u]/[o] and [e]/[i] are more regularly carried respectively by a waw and a yodh. But that is Ancient Hebrew talk. In Modern Hebrew, there is no distinction of vowel quantity (= no long and short) and students go insane with spelling. One rule is that yodh (and waw?) are to be doubled when in consonantal value...

In pointed script (invented in the early 1st millenium), there are gazillions of vowel indicators, each with their name. The Gesenius’ Hebrew Grammar goes on and on for over a dozen pages explaining each of them. That is due to the fact that pointed script was developed by several separate traditions - the one accepted today is of the Masoretic tradition. For practical purposes, one may say that Hebrew in its modern pronunciation possesses the five basic vowels (a, e, i, o, u) plus the so-called shwa’s, extremely short versions of the vowels a, e and o. And there is an ultra-shwa, the one that’s written with two perpendicular dots below the consonant, which is sometimes silent. Here I will not indicate the ultra-shwa, it will be inferred whenever there are two consonants together, as there can be no consonant followed by another one in Hebrew pointed script. The other shwas will be indicated by à, è and ò, but sometimes I could not check which one was right and I wrote only è.

The vowels o and u, when carried by a waw, will be indicated ó and ú. The same for i, transliterated í when followed by a yodh. The combination e + yodh yields the sound [e] (and not [ey]) and will be indicated é.

The ultra-shwa never softens the following consonant.

The presence of an alef, even when it’s a mere orthographical tool, will be indicated always with a semi-colon.

Doubled consonants are only indicated in pointed script, although this distiction is vital in Hebrew grammar, like in Arabic. Here they will be transliterated doubled.! The Hebrew Alphabet

Requires Hebrew and Phoenician fonts. (Hebrew not working yet, wrong fonts!)

1.2 Phonetic Correspondences between Semitic Languages

The evolution of Proto-Semitic sibilants and dentals according to German Semiticist Carl Brockelmann.

2. An Overview of Semitic Languages

This is one of the accepted classifications of the Semitic languages:

Northwest: Hebrew, Phoenician and Ugaritic are very closely related and are called Canaanite languages, whereas Aramaic and its dialects such as Syriac are a more separate branch.

East: Akkadian and its later versions, Babylonian and Assyrian. The Akkadian shown here will be based upon the Assyrian dialect.

South: Arabic in its several forms, Maltese, South-Arabian, Ge’ez, Tigrinya and Amharic - the latter very distant from the rest due to Kushitic influence.

Of the above, only Arabic, its offspring Maltese, Hebrew and Amharic are living literary languages, whereas Tigrinya, South-Arabian and modern Syriac Aramaic are practically only spoken.

The Semitic languages form a branch of the Afro-Asiatic (sometimes called Chamito-Semitic) Language Family. Other branches are:

-Berber (languages of Inner North Africa)

-Egyptian (already extinct, to which belong the Ancient Egyptian language and its later descendant, Coptic)

-Kushitic (several languages of Somalia and neighboring regions)

-Chadic (languages of Inner West Africa, chiefly Hausa)

-Omotic (languages of Southern Ethiopia)

3. Morphology

Arabic and Hebrew morphology, as it is typical of all Semitic languages, is based on the consonant root system. This consists of three (sometimes two or four) skeletal consonants that carry a general concept and that are filled by certain patterns of vowels and may gain affixes (=prefixes, suffixes and even infixes) in order to assume a particular meaning usually within that general concept. For instance, in Arabic the root k-t-b conveys the general concept of writing and things related. Filled with the appropriate set of vowels and sometimes affixes it may become a verb (kataba= he wrote, ;aktub= I write), some nouns (kitáb= book, maktúb= letter, kátib = writer) and other things. The vowel-fillers are usually the same for all roots, so you saw that k-t-b gained prefix ma-+1st consonant+2nd consonant+u+3rd consonant (consonants always remain in the same order), and became maktúb = written, letter, etc. The same can be applied then to the root k-s-r, (idea of breaking) - kasara = he broke, ;aksur = I break, maksúr = broken. Of course not all patterns can be applied to all roots and there are irregularities, oddities and peculiarities that have to be learned as you go along. Besides, not all roots have a clear meaning and not all patterns cause a predictable change of meaning. Hey, these are living languages, not computer languages!

The practical notation for word roots used here will have C standing for consonant, v for a vowel and vv for long vowel - but when it becomes necessary to indicate which of the two, three or four consonants of a root is meant, numbers will be employed: C1, C2, C3, etc. Thus the word maktúb can be noted as maCCúC, and kataba as CaCaCa, whereas kitáb as CiCáC, but if I want to indicate that the T of qaTal is to be doubled, I say “C2 is to be doubled”.

3.1 Nouns and articles

Arabic and Hebrew nouns can be preceded by a definite article, al- in Arabic and ha- in Hebrew, which are prefixed to the word they define. Furthermore, the article is invariable in gender and number, i.e., it serves for all words. The addition of the article to a word causes the following peculiarities:

-In Arabic, we find the assimilation of the l of the article bofore dentals, sibilants and n (namely n, DH, T, D, S, sh, z, r, dh, d, th, t, called the sun letters, for shams, sun, begins with one of such letters). The result is that the initial consonant of such words becomes doubled in pronunciation, although in writing the article is still written al-. See: al + rajul (man) = alrajul pronounced [arrajul] - al + shams (sun) = alshams pronounced [ashshams].

-In Hebrew the initial consonant of the word becomes doubled. When the first vowel is “e”, the article becomes he-.

Nouns can either belong to the masculine or the feminine gender, as it is thought to have been in Proto-Semitic. The final phoneme that characterizes feminine nouns (and most adjectives) in Arabic and Hebrew is [a], which being short is obviously not written, but normally indicated in writing by (a) in Arabic a special letter called tá; marbúTa, actually an h with the two dots of tá; above it and theoretically following the [a] in pronunciation but usually silent, and (b) in Hebrew by an h which would also follow the [a] theoretically but is also silent. These special feminine markers revert to a letter t in orthography whenever a pronominal suffix is attached to the word, and in the construct state (see below) they are pronounced [t] but mantain their special spelling.

The historical reasons for that is that the feminine ending was originally [at] in Proto-Semitic, but the final [t] wound up being left out of pronunciation whenever it was the last phoneme in the absolute state (see construct state below).

This leads us to the question of Semitic declension, which today survives in Koranic Arabic and partially in Written Standard Arabic. Words, even feminine words with their original final [at] used to be followed by an ending that indicated one of three cases: nominative, accusative and genitive. So the Arabic word today pronounced [madína] (city) was originally [madínatun] in the nominative case. As the endings disapeared in the spoken language, the final [t] was exposed and eventually was lost. The same phenomenon might have happened in Northest Semitic languages before historical times, with only a few fossils remaining (see below table). This is one of the many indications we have that Arabic is far more conservative than Hebrew, Aramaic and perhaps other Semitic languages, even though it was the last of them to enter History, in the 4th century A.D.

The Declensions in Classical Arabic

This declension is thought to have existed similarly in Proto-Semitic. The case endings were, according to the Précis, the same as in definite Arabic, but long (-ú, -á, -í), with an extra adverbial case in -u, of which there are still remains in Arabic. Actually, in the family terms ;ab, ;akh and the word dhú(1) , in construct state, still mantain the long case endings. The -n of the indefinite forms (-m in Akkadian) is a posterior addition and could be regarded as an indefinite article.

(1) This particle means possessor of and is used to form adjectives such as dhú dhakí (= possessor of intelligence = intelligent)

The origin of these ending are:

- the pronoun (he, this) for the nominative ending;

- the demonstrative for the accusative;

- and the genitive ending, although it has an obscure origin, may as well be related to the adjectivating suffix -í.

Hebrew still possesses a suffix -ah which can be added to certain nouns to indicate direction, being possibly a remain of the old accusative. Examples: ;ereçah (lit. “to the land”, i.e., to Israel). But the special preposition “;eth” plays the role of an accusative marker for definite nouns: “qaH ;eth habasar” (=take the meat < “basar”= meat; “qaH”= imperative of verb “laqaH”, to take). Another example is the first sentence of the Book of Genesis: bereshíth bara; ;èlohím ;eth hashamayim we;eth ha;ereç (In the beginning, God created the heavens and the earth).

For the dual, there were the endings -áni [alif + nún] for nominative and -ayni [yá; + nún] for both accusative and genitive. The final n is lost in the construct form (see below). These same declension endings are used in Classical Arabic in plural nouns, except for the external plurals (see plural below). In the dual and external plurals declension still survives in the written language for the oblique cases (accusative and genitive) are spelled differently from the nominative, Colloquially, however, only the oblique formas are used

Arabic nouns ending in a consonant are masculine except for a few like ;umm, ;ukht, shams, yad, ríH (mother, daughter, sun, hand, wind). As for Hebrew, most nouns denoting abstract concepts are feminine, including some ending in a consonant other than the h feminine marker, especially the abstract suffix -uth. The feminine of adjectives ending in -eh is usually -eth, and of those ending in -í is íth.

The plural

Plurals in Hebrew are quite straightforward. One adds -ím to masculine nouns and -óth to feminine, with the transformation of some vowels into shwas. Tov > Tovím - Tovah > Tovóth (good).

Arabic on the other hand possesses an intrincated system of plural shapes. Apart from complex words such as those consisting of more then four consonants, which take an external plural, most Arabic words form their plurals by changing their vowel patterns and sometimes gaining an affix. Amongst these words, the only ones entirely predictable are the four-letter words. Possessing a short vowel between C3 and C4, their plural shape is then CaCáCiC. Otherwise they goCaCáCíC. Note as well that the feminine ending is dropped in such plurals:

Obs: CvCCvvC nouns refering to people optionally have a CaCáCiCa plural shape: ;ustádh > ;asátidha (professor).

As for three-consonant words, there are no steadfast rules. One can only ascertain a word’s plural safely from a dictionary. I will try however to summarize the main tendencies governing such maverick plurals:

- CiCCa words tend to build plural CiCaC: qitca > qitac (piece) - nukta > nukat (joke).

- CuCCa words tend to build plural CuCaC: lucba > lucab (toy) - súra (written s-w-r-a) > suwar (image).

- CvCC words may to build plural CuCúC: qalb > qulúb (heart) - cilm > culúm (science) - jund > junúd (troops) or ;aCCáC: film > ;aflám (film) - shakl > ;ashkál (shape) - lawn > ;alwán (colour).

- CáCiC words may to build plural C1uC2C2áC3: sákin > sukkán (dweller) - sá;iH > suwwáH (tourist). Some of these words have the option of a CaCaCa plural if denoting people: Tálib > Tulláb or Talaba (student).

- CáCiC (or CáCiCa) words tend to build plural CawáCiC: sáHil > sawáHil (coast) - qácida > qawácid (rule).

- Words with the feminine ending and long vowel on C2 tend to build plural CaCá;iC: jarída > jará;id (newspaper) - jazíra > jazá;ir (island, in the definite plural aljazá;ir it also means Algeria).

Tired? Well, the author of the Précis also seems annoyed with such an intricate plural system in Arabic, who comments that in this language (and in Ge’ez), internal plurals have evolved monstruously, limiting the role of the Semitic external plural formation. So, unlike what some may have thought, it wasn’t the Northern languages that simplified the plurals shapes, but the Southern ones that complicated it after the breakup of Proto-Semitic.

The construct

Typical of all Semitic languages, the construct state (called thus traditionally in Western Grammars) is a common way of expressing what in some Indo-European languages is expressed with the preposition of or the genitive case, although it has some other functions. In modern Hebrew it must be noticed that the construct is normally only used to make set phrases such as football game, the merely occasional possessive function being performed by the prepostionshel. This is no suprise to me, since the Hebrew construct is quite intrincated in what concerns voweling and suffixes, being in this respect more complicated than Arabic.

To form a construct one simply puts the possessor after the thing possessed: the book of the man would be “book the man” (Arabic kitáb alrajul, Hebrew Sefer ha;ísh). Although the first word of the construct is regarded as definite, it can never take the definite article.

As stated earlier, the ancient feminine ending [at] is brought back to pronunciation if in the first word of a construct: alkitábat alcarabí (= the writing of arabic; kitába = writing). Other suffixes like those of the dual and plural lose their final nasal in the construct state, with vowelling changes in Hebrew.

External plural and dual endings suffer some changes in the construct state, mainly the loss of the final -n (Arabic) and -m (Hebrew). Hebrew words also undergo a change in their vowels standing before this lost -m: bením (sons) becomes beney.

The normal state of the noun is denominated absolute state.

3.2 Adjectives, pronouns and adverbs

In Arabic, adjectives, pronouns and demonstratives (well, verbs too) agreeing with inanimate plural nouns assume the feminine singular shape. I found no satisfactory explanation of this phenomenon but I believe that it has to do with the traditional inferior status given to women in all grammars. Thus female=thing(s). In the Semitic sphere, this is unique to Arabic.

Adjectives, with few exceptions, take [a] to become feminine in both languages. The thing is that Semitic languages do not distinguish between nouns and adjectives so well (much like Latin languages and Greek), so the following peculiarity has arisen: instead of saying “the big book”, they started saying “the book, the big” like when in English one say “give me the suitcase, the green one”. In Semitic adjectives don’t need to be followed by an equivalent of “one”, so what happened is that today Arabs say alkitáb alkabír (kabír = big), i.e., they say the noun first, the adjective later and the article comes with both. The same goes for Hebrew: haSefer hagadol (Sefer= book; gadol= big). Note that the adjective must come after the noun.

The degrees of comparison

The Arabic adjective possesses a comparative of superiority with the formula ;aCCaC. But when the two last consonants are the same (C2=C3), the formula employed is ;aCaCC. So we have: kabír (big) > ;akbar (bigger) - kathír(very) > ;akthar (more) - jadíd (new) > ;ajadd (newer). Complex adjectives such as maftúH (open), burtugalí (Portuguese), muD-Hik (ridiculous), masíHí (Christian) etc. cannot undergo such rules because they have more then three consonants. In such cases, one uses a noun related in meaning to the adjective plus the word ;akthar (more). The Arabic equivalent to English than is the preposition min: aljarída ;ajadd min alkitáb = the newspaper is newer than the book - huwa ;akbar min alwalad = he is bigger than the kid.

The Hebrew adjective, once again confirming the greater simplicity of this language in comparison to its Southern sibling, possesses no special form for the comparative of superiority. One simply uses the normal adjective and the preposition min: hú; gadol min hayeledh = literally “he is big from the kid”, but meaning “he is bigger than the kid”. But this is the formal rule. Modern Hebrew uses the word yóter (more) and also the prep. min (or shortened mi), or the more elegant adjective + yóter construction: yóter Ham or Ham yóter = warmer.

The demonstratives

I was always puzzled with the enormous differences that seemed to separate the demonstratives in Arabic and Hebrew, as I learned them in my grammar books. But the Précis de Linguistique Sémitique provided me with the following explanation:

The most ancient demonstrative known in Semitic is , still used in Arabic as an interjection. In Hebrew and Aramaic, it became the definite article, although in Aramaic it is suffixed and ended up fossilized at the end of nouns: bayt > * baythá > baitá (the house, house). In Arabic, Hebrew and Aramaic, the demonstrative joins the 3rd person pronoun, which in Arabic serves as an elongated demonstrative (that one over there), but is rarely used.

In Arabic, the historically most primitive demonstrative is dhá (m) and dhí (f), with secondary forms tá/tí, in Ge’ez, on the contrary, it is which is feminine and ze (from ancient dhi) is masculine. In Hebrew it winds up as zeh (m) andzóth (f), this -th being the feminine ending, while in Aramaic is a relative pronoun of common gender and a demonstrative (fem. ). Their plural in Semitic comes from an entirely different root: ;ulá or ;ulá;i in Arabic and ;élle / ;ellin Biblical Aramaic and Biblical Hebrew.

Arabic still joins both and dhá to make the by far most used demonstrative of closeness (this): hádhá (m) and hádhi (or hádhihi - f), plural both genders há;ulá;i.

But the story is more complicated. Arabic, as well as Aramaic and Ge’ez, possess a demonstrative of distance (that), obtained with the addition of a [k]. In Arabic we have dháka (m) and tika (f), but which exist today most often with the extra strengthening element [l]: dhálika (m) and tilka (f). In Hebrew, was simply joined to the 3rd person pronouns.

In Arabic, they precede the noun phrases, whereas in Hebrew they follow them. In both languages they demand the definite article.

For sake of simplicity, the following table shows the demonstratives used in Standard Arabic and Hebrew:

In Arabic: hádha alwalad alSagír [hadha lwalad aSSaghír] = this smal boy.

In Hebrew: hayeledh haqaTan zeh = this small boy.

If you want to say “this is the small boy” in Arabic you have to insert the 3rd person pronoun to serve as a copula: hádha huwa alwalad alSagír.

The pronouns

As possessive personal adjectives, instead of separate words, Arabic and Hebrew, as well as other Semitic languages, use a set of suffixes attached to the word possessed. These suffixes are:

The object personal pronouns (me, thee, him, etc.) also come as suffixes, which happen to be, with one exception, exactly like the possessive suffixes above. The exception is the 1st person singular pronoun (me), which, as in all Semitic languages, assumes the form - when it is attached to verbs. In Hebrew, due to its rich vocalic system (at least in writing), the adition of these suffixes causes the first vowel of some longer (2 syllables or more) words to become one of the four shwa’s. And it’s never too much to remind you that these suffixes bring back the feminine ending [at] to pronunciation and spelling: Arabic madína > madínatí (city > my city).

Not only prepositions in general take these prefixes, but also several interrogative words, i.e. Arabic kayfak = how are you (kayf = how)

Similarly to what happens with the construct state, Modern Hebrew normally uses the preposition shel, roughly equivalent to “of”, plus the above endings to build the possessive adjectives, leaving the noun+suffix construction to lofty-stiled speech. Thus an Israeli says normally Sefer shelí as “my book” but could use Sèferí if he wanted to be sophisticated.

The subject personal pronouns are:


-These are used only for emphasis once the verbal endings sufficiently point out the subject of a clause. In Modern Hebrew, the feminine plural pronouns are not used colloquially, a phenomenon that occurs also in several forms of Spoken Arabic.

-The difference of the two 1st person pronouns in Hebrew is, according to my Teach Yourself Biblical Hebrew, that the longer form is used in “pause” (?). Modern Hebrew uses only the short forms.

-The Syriac dialect of Aramaic possesses a gender distinction on the 2nd person plural pronouns: ;attón (m) and ;attén (f).

3.3 Verbs

Verbs are conjugated also by changing root voweling and by the addition of affixes. The dictionary or base-form of a verb is the 3rd person masculine singular perfective (yes, Semitic verbs vary in gender!), which is the simplest form. There are only two stems, perfective and impefective, to which prefixes and suffixes are added for the agreement of person and number.

The simple verb conjugated:


-for katabtu, katabta and kataba the final short vowel is today only mantained when a pronominal suffix is added. In the imperfect the final short -u is also dropped.

-there is also a dual in Arabic.

-final long ú’s are spelled waw+alif in Arabic.

The second person plural feminine of Hebrew is used only in elevated style today. In several dialects of Arabic the feminine forms of the plural of verbs also disapear. The vowel C2 receives in the Arabic imperfective is, in the case of simple verbs, a matter of memorization, and many Arabs themselves make mistakes, as I read once. Arabic dialects might have a very different voweling, even in the perfective.

The usages of the two conjugations in Arabic and Hebrew languages, for practical purposes, can be said to nearly always coincide with tense. Thus the perfective aspect, which denotes completed actions, is equivalent to the past tense of European languages, and the imperfect, describing actions incomplete, is used for present and future tenses. Aditionally, Arabic adds the prefix sa- or the word sawfa to a verb to put it in the future, whereas Modern Hebrew uses the imperfective only for the future and employs the old present participle as a present tense verb, which declines as an adjective: kotev, kotvah, kotvím, kotvót.

In Arabic the imperfective also has two extra moods (the one given above is the default mood called indicative): the jussive and the subjunctive. The distiction is quite simple: instead of gaining final -u vowel, the sujunctive gains -u and the jussive gains no vowel. Additionally, the verb loses the final -n. Hebrew knows extra moods as well, thogh they are confined to a very literary use (Biblical, I’d say...). They are the jussive and the cohortative moods.

The negative of Hebrew verbs is obtained with the word lo; (no) before the verb. The same also holds for Arabic, which puts before imperfective verbs, but in the perfective negation is done with lam plus the imperfective jussive or less commonly with má plus the perfective verb.

Simple verbs with special roots

Verbs presenting a semilvowel, a glotal stop or some other letters in their original root as well as those whose two last consonants are the same, like H-b-b (“to love” in Arabic), are somewhat peculiar in their conjugation.

When one of the roots is n

In Northwestern Semitic verbs with C1=n lose this n when they gain prefixes, either personal or from the derived stems: nathan > yathan (gave > gives). When C3=n this also happens: nathan (he gave) > nathathi (<nathan+ti) I gave.

When C2 is a semivowel (w, y):

The middle radical disapears in the perfective leaving an “a” between C1 and C3, and in the perfective it resurges as a vowel carrier.

Arabic examples: q-w-m (to stand up) = qám, yaqúm; n-w-m (to sleep) = nám, yanúm; m-w-t (to die) = mát, yamút

Hebrew examples: q-w-m (to stand up) = qam, yaqúm.

When C2=C3:

The two last radicals merge, reapearing separately when the personal ending begins with a consonant. In Arabic, the formula for the perfective is CaCC and for the imperfective yaCvCC. In Hebrew, CaCC and yiCvCC.

3.4 The derived stems

The derived stems of Semitic verbs are a rich means of creating new verbs from the simple verb you saw above. They are formed by changing the voweling of the simple verb and/or adding prefixes to the stem. Most kinds of derived stems tend to cause a general change in meaning to the root, but many cannot have their meaning predicted.

Arabic possesses nine derived conjugations, each with both active and passive forms. Hebrew has six, of which three are passive and three active (of course the simple verb is also active).

The Arabic derived stems are as follows, using the base verb lamasa, to touch, the imperfective given without perfixes:

The Hebrew verbal stems are called traditionally binyaním (structures), whose formula is normally given through the verb pacal (to do), but since guttural letters such as c cannot be doubled, the patterns will be given here with the verbqaTal (to kill). Thus we have the following derived binyaním:

niqTal - often functioning as the passive of the binyam qaTal (or simple verb) - imp. yiqqaTel, imperative hiqqaTel

hiqTíl - often causative

húqTal - often the passive of hiqTil

qíTTel - often transitive intensive

qúTTal - is the passive of qiTTel

hithqaTTel - often intransitive of qiTTel

3.4.1 Correspondences between Arabic and Hebrew derived stems

Not all derived stems agree in the two languages as regarded to morphology and semantics. Here are the ones that do:

The intensive stem is formed similarly in both languages, as well as in Phoenician, Ge’ez and Aramaic. One doubles the second consonant of the basic stem: Arabic qatal (to kill) becomes qattal (to massacre), Hebrew qaTal becomesqaTTel. In the imperfect, C1 does not lose its vowel (this loss only occurs in the simple verb): Arabic yuqattil, Hebrew yèqaTTel. Note that in all derived stems formed by the doubling of a radical, Arabic prefixes are vowelled -u- instead of normal -a-.

Remember that, since vowels are left unwritten and doubled consonants are left unmarked, fuguring out the wether a verb is in its intensive form of simple form is a matter for pros.

The causative stem is obtained with a prefix, ;a- in Arabic and hi- in Hebrew, and the loss of C1’s vowel - and in Hebrew, C2 is vowelled -í-: ;aqtal, hiqTíl (to cause to kill).

The n-prefix is characteristic of the reflexive stem: inqatal in Arabic and niqTal in Hebrew = to kill oneself.

The reflexive of the above-mentioned intensive stem is obtained with a ta- prefix in Arabic and a hith- prefix in Hebrew added to the intensive stem: taqattal, hithqaTTal = to massacre oneself.

Note: some of these forms of qatal-qaTal do not actually occur, but can occur with other verbs. The verb qatal-qaTal is given here only as a model.

Other remarks on verbs

As the verb “to be” (kán in Arabic and Hayah in Hebrew) is not used in the imperfective, in order to render “the book is big”, simply remove the article from the adjective: alkitab kabir; haSefer gadol (although Hebrew has some “being particles” that can be used too, like yesh). On the other hand, the negated imperfective of verb to be is obtained with special separate verbs. Arabic has laysa (perfective in form but imperfective in meaning, i.e., “is not”) and Hebrew has ;én - the latter can also be used to negate present tense (the present participle used as present verb) more sophisticatedly.

Another verb “lacking” in Semitic languages is “to have”, which doesn’t exist at all. In order to express this idea one uses a preposition meaning “to” or “next to, by” plus the possessor, then adding the thing possessed. In Arabic this preposition can be either cind(a) or li, in Hebrew it’s only l-. The idea is to say “to the man there is a book” where in English one says “the man has a book”. As the verb “to be” (and consequently “there to be”) is not used in the imperfective, one actually says “to the man a book” (actually, with no indefinite article, the sentence is “to the man book”...).

In Arabic: cind alrajul kitáb (Classical cind alrajuli kitábañ)

In Hebrew: l’ha;ísh Sefer. Or with the being-particle: l’ha;ísh yesh Sefer.

Arabic, Phoenician and Aramaic have developed a way of forming a sort of plus-quam-perfect, (i.e. I had done) by using the perfective of the verb to be. This is kán (root k-w-n) in Arabic, so kán katab = he had written. From this we can go to the verb string structure in Arabic. With no infinitive, Arabic verbs in strings just go agreeing with the subject all the way: huwa lá zál yaktub = “He didn’t cease to write, he is still writing” (lit. “he not ceased he writes”, where zál = he ceased). Arabic however has the so-called verbal noun, always irregular for simple verbs but regular for derived stems, which can fill in some of the functions of an infinitive.

6. The numerals

An odd charcteristic of Arabic and Hebrew numerals from 3 onward is that it is the morphologically feminine numerals that go with masculine nouns and the morphologically masculine nouns that go with feminine nouns. The Encyclopaedia Britanica was the only place where I found a mention to this feature - they say it has never been satisfactorily explained by Semiticists. The Précis is silent on the issue, it only mentions it exists.

Number one is never used, the singular of the indefinite noun is enough. Two is unnecessary as well, since both languages have a dual. Colloquially, however, sometimes speakers may use 2 followed by the plural. From 3 to 10, the plural (in Arabic, plural genitive) is used. From 11 to 100, the singular (another oddity) is used (in Arabic, in the accusative).


-Colloquially, only the feminine numerals are employed in Arabic. Actually, according to some of my sources, even in official announcements the frozen feminine is used throughout.

-In Gulf Arabic, thintayn is colloquial for 2. And, of course, ithnán(i) being a dual, has only its oblique form is used in speech (ithnayn)

-The tens are formed like in German: two and twenty, three and twenty, etc. Besides, they are regarded as constructs.

Bonus: Numerals in other Afro-Asiatic languages:


-Hyphens appear where original voweling is unknown.

-Egyptian voweling is deduced.

-The South Arabian is of Harsusi.

-In Semitic languages, the morphologically feminine numbers are given.

Source: Numbers in over 4000 Languages with transliteration adapted to my system as far as I could figure out the ones on my source, since there was no standard. In Maltese, which is written in Latin script, the gh seems to correspond to the ayn of Arabic. Somali and Hausa are written with our alphabet as well.

4. Syntax

This section will treat only the syntactical issues not already covered in the previous sections.

The syntax of Modern Written Arabic is quite complicated, for it mantains most of the Classical language’s characteristics. The main oddities are:

-The normal word order is verb-subject-object, and verbs in sentences like this do not go to the plural or dual no matter what their subject is.

-Certain verbs of being or becoming, as well as some particles as ;inna, require that their semantic subjects go to the accusative.

-The conjunctions and relative pronouns introducing subordinate clauses cannot be modified by a preposition. That causes the preposition to go to the end of the subordinate clause and then gain an appropriate pronominal suffix: albayt allathí ;askun fihí = lit. “the house which I live in it”.

Hebrew syntax follows the same Arabic rule of subordinate clauses. In addition, Hebrew has a construction, common in the Biblical language, known as waw consecutive. This is the concatenation of several verbs with the help of the conjunction w- (and), in which the first or main verb is in the perfective and the rest goes to the imperfective, or vice versa.


Modern Hebrew, An Essential Grammar (ed. Routledge)

Every Day Hebrew (ed. NTC)

Teach Yourself Biblical Hebrew

Teach Yourself Arabic

Colloquial Arabic of the Gulf and Saudi Arabia (ed. Routledge)

Arabic for Travellers (Berlitz)

Taschenwörterbuch Arabisch (ed. Langenscheidt)

Teach Yourself Arabic Verbs and Essential Grammar

Apendix 1: Etymology of biblical names from Hebrew and Aramaic.

Although only the Old Testament was written in Hebrew (but books of Ezra and Daniel in Aramaic), most names from the New Testament (which was written in Greek) are of either Hebrew or Aramaic origin. These names, however, were hellenized by the authors. As Greek did not have many of the sounds found in the original names, they suffered many transformations in order to be written in the Greek alphabet.

As exceptions, we have a few sentences of the New Testament which were preserved in their original Aramaic and then translated into Greek. The most famous of which is perhaps the one we find in Mark 15:34, when he cites Jesus’ last words: elwi, elwi, lima sabacqani (eloi, eloi, lima sabakhthani?), probably from an original Aramaic ;elí, ;elí, lamá sabaqthaní? (my God, my God, why have you left me?). The Aramaic verb sabaq (he left, in Hebrew alsosabaq) was conjugated by Jesus in the 2nd person masculine singular of the perfective = sabaqtha, to which he added -ní, the object 1st person singular suffix = me. The Aramaic word limá or lamá (= why), in Arabic lima(dha) and in Hebrew lammah, are actually the word ma (what) preceded by the preposition li (for). The word ;elí is ;el (God, in Hebrew also ;el and in Arabic ;iláh) plus the possessive suffix of the 1st person singular -í (= my - just like in Arabic and Hebrew). The Greek spelling eloi is probably due to the fact that the Classical Greek diphthong [oi] was already being pronounced as it is in Modern Greek, namely [i].

main source: On-Line Bible (actually a CD-rom)

Apendix 2 - Etymologically related words in Arabic and Hebrew

As it always happens when we compare two languages from the same family, many words which share a particular root have no longer exactly the same meaning. When this happens in the following table, you will find two translations separated by a hyphen in the third column. The first one refers to the Hebrew meaning and the second one to the Arabic. When separated by a comma, it means that both the Arabic and the Hebrew words have both meanings.

On the Hebrew column, related Aramaic words that I have bumped into are inserted between parentheses with the abbreviation Aram.

Apendix 3 - Portuguese Words of Arabic Origin.

The Arabic words are given without the definite article al-. Remember that they usually got into Portuguese with the article, and also that the l of the article is assimilated if the sound following it is a sun letter (n, l, DH, T, D, S, sh, s, z, r, dh, d, th, t).

When no meaning is given in the Arabic column, the Arabic word means the same as in Potuguese. As you will see, many English words share the Arabic origin of their Portuguese counterparts, although usually without the definite article.

The Brazilian spelling and meaning are given.

Appendix 4 - Toponyms, Gentilics and Cultural Words in Arabic

The following table will include as many words as I discover but will be concentrated in the Middle Eastern and North African Worlds.

Arabic- speaking world

Appendix 5 - Semitic and Indo-European