In fairness, I can't complain about the letters. As I've already disclosed above, it was Arabic script that drew me in. But they are a beast to keep track of for the beginning student: These 28 letters are derived from 86 different forms reflecting embellishments (dots and tails, essentially) on (let's say) 21 basic shapes.

And each letter can have as many as four different forms, depending on whether it is initial, medial, final, or isolated.

And even more galling than that is that handwritten Arabic contains variations independent of printed Arabic:

• In print the dots are separate, but in handwriting two dots are often written as a line, and three dots are often written as a caret. This makes writing easier, and is inoffensive. But more vexingly: In print the characters flow on the line from right-to-left, while in proper handwriting many of the letters are stacked vertically. Compare these printed (unboxed) and handwritten (boxed) versions:
  

  These are the same
  		تجاب

• The medial form of the character ـهـ looks much like a figure-eight. But in handwriting it is often made with a simple downward point. Compare:
  

  These are the same
  		مها

• The character ـسـ (shown here in its medial form) is printed with three unevenly spaced ridges. But in handwriting, it is just a straight line. But ... but ... but ... if you see a long straight line in printed text, it's not the ـسـ; rather, it's just irrelevant spacing for the aesthetics of the printing. Compare the printed (unboxed) and handwritten (boxed) versions below (everything in black is the same word; everything in blue is the same word):
  

  These are the same

  حساب

  ↑

  These are different

  ↓

  حــــاب

  حاب

  These are the same

And it gets challenging – especially in the first weeks – remembering all the details: the four different shapes of هـ; when to write ة, instead of ت; remembering that the final loops for ن, ل, ق, ض, ص, ش, س, and ي swing low, while those for ث, ت, ب, and ف do not. Some letters cut you a break: ط and ظ are the same in all four forms, but what they relinquish in ease of writing they take back in difficulty of pronunciation (see below).

But still, the writing is a sheer joy to me: It's beautiful to look at. And, to be completely honest, it makes me feel smart when I write it. With Cyrillic (which, to be fair, I don't know at all) or Greek (which I only know a little), you're just using a different code. But with Arabic, you're using an entirely different thought process. Moreover, Arabic can't be written without it. I'll confess to a naïve belief earlier in my studies that perhaps it would be better simply to write Arabic in Latin characters, but this would not open Arabic up; rather it would make the linguistic relationship between the words impenetrable. Nonetheless, several systems of transliteration have been developed:

Arabic

SATTS  (US Military)

Library of  Congress

ISO

Qalam

SAS

SM

IPA

Arabic Chat (IM)

ء

E

ʼ, –

ˈ, ˌ

'

ʾ

'

ʔ

2

ا

A

ā

ʾ

aa

a, i, u, ā

aa

aː

a

ب

B

b

b

b

b

b

b

b

ت

T

t

t

t

t

t

t

t

ث

C

th

ṯ

th

ṯ

ç

θ

s, th

ج

J

j

ǧ

j

ŷ

j

ʒ

g, j

ح

H

ḥ

ḥ

H

ḥ

ḥ

ħ

7

خ

O

kh

ẖ

kh

j

x

x

5, 7', kh

د

D

d

d

d

d

d

d

d

ذ

Z

dh

ḏ

dh

ḏ

đ

ð

z, th

ر

R

r

r

r

r

r

r

r

ز

;

z

z

z

z

z

z

z

س

S

s

s

s

s

s

s

s

ش

sh

š

sh

š

š

ʃ

sh

ص

:

ṣ

ṣ

S

ṣ

ṣ

sˁ

S, 9

ض

X

ḍ

ḍ

D

ḍ

ḍ

dˁ

D, 9'

ط

U

ṭ

ṭ

T

ṭ

ṭ

tˁ

TH, T, 6

ظ

Y

ẓ

ẓ

Z

ẓ

đ̣

ðˁ

Z, TH, 6'

ع

`

ʻ

ʿ

`

ʿ

ř

ʕ

3

غ

G

gh

ġ

gh

g

ğ

ɣ

gh, 3'

ف

F

f

f

f

f

f

f

f, ph

ق

Q

q

q

q

q

q

q

q, 8

ك

K

k

k

k

k

k

k

k

ل

L

l

l

l

l

l

l , lˁ

l

م

M

m

m

m

m

m

m

m

ن

N

n

n

n

n

n

n

n

هـ

~

h

h

h

h

h

h

h

و

W

w

w

w

w, ū

w, o

w , uː

w

ي

I

y

y

y

y, ī

y, e

j , iː

i, y

آ

AEA

ā

ʾâ

~aa

ā

'aa

ʔaː

a

ة

@

h, t

ẗ

h, t

t, –

ŧ

a, at

t, –

ى

/

y

ỳ

ae

à

à

aː

a

The problem is: These are all a failure. SATTS makes punctuation impossible because punctuation marks are included. Many of the others include obscure characters that can only be viewed on computers with a full Unicode font set (if you don't see any boxes above, you have a very well-trained computer). Others make you guess as to the original text and thus as to the pronunciation (e.g., Is "sh" in Qalam transcription one letter or two?). And if they allow vowel additions (all except SATTS and Arabic Chat), how are the vowels and other diacritical marks encoded? Just to understand the scope of the problem, look at the familiar name, Omar Khayyam, which speakers of Arabic know to be written as عمر خيام:

		unvocalized		vocalized
Arabic		عمر خيام		عُمَر خَيام
SATTS (US Military)		`MR OIAM
Library of Congress		ʻmr khyām		ʻumar khayyām
ISO		ʿmr ẖyʾm		ʿumar ẖayyʾm
Qalam		`mr khyaam		`umar khayyaam
SAS		ʿmr jyām		ʿumar jayyām
SM		řmr xyaam		řumar xayyaam
IPA				ʕumar xajjaːm
Arabic Chat (IM)		3mr 7'yam

In my opinion, the best is the least scientific: Arabic Chat (those clever kids!). But we're still left with the fact that none of these systems really works. The more useful they are in helping a non-speaker of Arabic, the less reliability they have with a one-to-one (and, thus, easily reversible) correlation to the original text. If English didn't have so many sounds that don't have letters (th as in "think"; th as in "that"; sh), or if Arabic didn't have so many sounds that English represents with a single letter (notably, the emphatic consonants; see below), then a precise correlation (at least between English and Arabic) might be possible. But it was not to be. So we are left with "Omar Khayyam" (in English, that is; the Germans are left with "Omar Chajjam" to refer to the same guy). (His full name, by the way, is غیاث الدین ابو الفتح عمر بن ابراهیم خیام نیشابوری; transliterate that!) (N.B.: In this document I have not been consistent in my use of transliteration; this is in part because I got my base information from different sources that used different systems; but I did not go back through and bring order to it because in some areas a consistent system would actually decrease readability for the casual [non-Arabic speaking] reader.)

John Wilkins didn't care for the writing though. More on him anon – but for now, he wrote in his seminal work, An Essay towards a Real Character and a Philosophical Language: "The Arabic character, though it showed beautiful, yet is it too elaborate, and takes up too much room, and cannot dwell the written small." He shows his ignorance: Arabic is one of the most compact systems of writing ever devised, and takes up significantly less space, for the same conceptual material, than English.

So very early on, I diverged from John Wilkins and made my peace with the Arabic letters. So much of a peace, in fact, that I learned to type them, and then learned them in the Arabic manual alphabet:

Arabic Keyboard:

Arabic Fingerspelling:

ض ص ش س ز ر ذ د خ ح ج ث ت ب ا ة ى

ء ي و هـ ن م ل ك ق ف غ ع ظ ط

A regular sequence of letters is used for ordinary alphabetization, but in lists that are preceded by letters, the same twenty-eight characters are arranged in an entirely different sequence. Headings in this list of Arabic attributes, you may notice, are prefixed in the abjad sequence.

Both sequences are shown below. On the standard sequence, the shading from white to yellow increases incrementally. You can compare the unchanged shading of the letters in the abjad sequence to see how thoroughly shuffled they are. Indeed, it is exactly as if English kept its regular sequence for the dictionary, but prefixed lists in the following order: A, B, E, H, Z, AA, K, F, P, AB, V, W, X, Y, L, R, T, N, U, J, M, C, D, G, I, O, Q, S ("AA" and "AB" to compensate for the fact that Arabic has 28 letters).

First, let us clear up the nomenclature: In English, we do not write with Arabic numerals, no matter what your fourth-grade teacher told you. The confusion arises because we use "Arabic numeration" – that is to say, we organize our numbers in columns of units, tens, hundreds, thousands, etc. But the characters we write with are best called "European numerals," though "Hindu-Arabic numerals" will suffice, as will "Western Arabic numerals" (but only if you are distinguishing from "Eastern Arabic numerals"). In most of the Arabic speaking world (i.e., all but the western-most reaches), the preferred numerical characters are properly called by us "Arabic numerals" or "Eastern Arabic numerals." Now, having dispensed with that ...

I've got a chicken-and-egg uncertainty as to whether the early Arabic world raised mathematical geniuses because of the complexity of their numbers, or whether their numbers are complex because of their mathematical genius. But anyway, they're a smart people with unending oddities in their relationship with the ten digits.

To start: numbers in Arabic are written left-to-right even as the rest of the script is written right-to-left. The following sentence reads, "I have 19 books" and starts at the right:

ç	è	ç
كتاب	١٩	عندي
books	19	I have

Also they are challenging because of their often-false similarity to European numbers. And note that the issue of the handwritten style is present with the numbers as well: The handwritten 3 looks like the printed 2:

Arabic          Numerals	Handwritten
	Printed		٠	٩	٨	٧	٦	٥	٤	٣	٢	١
European Numerals			0	9	8	7	6	5	4	3	2	1

The upshot is that I (like most beginners) make a lot of mistakes, misinterpreting "false cognates" (as it were) such as the following:

١٩٥٥	is not	1900	but rather	1955
٤٠٦	"	3.7	"	406

But the oddest aspect of the numbers is the the grammar related to them. I struggle to bring order to this, especially since I haven't learned (in class, with the appropriate exercises and practice) all of this information. But here goes:

Position

• 1-2, the number is usually omitted, but goes after the noun if added for emphasis.
• 3 and above , the number goes before the noun.

Plural

• 1, the noun modified is singular.
• 2, the noun modified is dual (a separate type of plural only for two of an item).
• 3-10, the noun modified is plural.
• 11 and above, the noun modified is singular.

Gender

• 1-2, the number shows gender agreement.
• 3-10, the number shows gender polarity (i.e., uses the opposite gender).
• 11-19, the number shows gender agreement.
• 20 and above, ending in 0, the number does not take a gender.
• 20-99, ending in 1-2, the number shows gender agreement.
• 20-99, ending in 3-9 , the number shows gender polarity.

Declension

• 1-2, the number is an adjective, and thus unaffected by declension
• 3-10, the numbers are diptotes (two cases in a language that normally has more than two cases), with one case dropped in oral usage.
• 11-99, the noun is declined in the accusative.
• whole 100s 1,000s, etc., the number is the first term of a genitive construction.

As an example, you can see increments of كتاب, a masculine noun, and specifically, where it is modified by feminine adjectives, where it is plural or singular, and where the dual plural is added.

As with other Semitic languages, the structure of Arabic is built around a root set of three (or rarely two or four) consonants. In English, we think in terms of prefixes or suffixes. But in Arabic, because the length of the root is fixed, the transformation of the root into related words is tremendously more complex, with modifications before, after, and within the root. An example of a root is ك ت ب; which references writing; to the right you can see some of the permutations of this root.

The fascinating, sublime part is that you can take a different trilateral root and apply the same permutations to it, and thus achieve (to some extent) parallel related meanings.

(Here, be forewarned, I stray a bit from Arabic, but I hope you'll find the digression worthwhile.) In the year before the trip, one of my few major projects that was not actually related to the trip was reading the superb novel Cryptonomicon and its even more superb prequel, The Baroque Cycle (by Neal Stephenson; read them!). The latter delves into the work of natural philosophers (proto-scientists) at the time of the British Civil War, and one of the side stories introduced me to the (very real) work of John Wilkins on the "Real Character," in which words were augmented based on their logical connections. For example, "de" referred to an element (in Wilkin's time, fire, air, water, and earth); "deb" referred to fire, the first element; "debα" (with an alpha at the end, not an a) referred to the flame, which Wilkins counted as the first "species" of the "genus" deb; and "ded," "deg," and "dep" were, respectively, air, water, and earth. To use a different example, "zi" referred to beasts; "zit" referred to "RAPACIOUS beasts of the DOG kind" (emphasis copied) ; "zitα" (with an alpha) referred to dogs and wolves; "zita" (with an a) referred to foxes. it all seems very orderly until you learn that "RAPACIOUS beasts of the DOG kind" also included "zite" (seals) and "zito" (armadillos). In fairness to Wilkins, he divided "RAPACIOUS beasts" into only two camps: "DOG kind" and "CAT kind," and sorted between them based on the shape of the head. So his fuzzy logic (I use this as a scientific term and not as a value judgment) is internally consistent: Stipulating that a seal is not a dog, the shape of its head is still more dog than cat. Even so, work on the Real Character failed to advance because it could not withstand the degree of conjecture, arbitrariness, and scientific revision that would be necessary for such a language. But the Arabic language evolved in just such a way. Had Wilkins been more aware of it, we might be speaking, or at least might be familiar with, a variant of the Real Character today.

But back to Arabic: The great challenge to the student with this structure comes in using a dictionary. Words are sorted according to the root, and then by part of speech. Thus, all of the words to the right would be collected on the same page of a dictionary. It is as if an English dictionary showed on the same page, "bespoken," "misspoke," "speaker," "spoke," "well-spoken"; it works just fine if you know the language – but if you don't know which of the letters form the root, and which ones are the add-ons, you've got several possibilities to cross-check.

The writing is a form of art, but not at the sacrifice of being a vehicle for communication, as this graphic shows. The text is the signature of Mahmud II and reads:

محمود حن بن عبد الحميد مظفر دائما
Mahmud Han bin Abdulhamid muzaffer daima
(Mahmud Han son of Abdulhamid is forever victorious)

Source: Wikipedia; permission granted on source page for re-use.

You can watch the graphic fill in on its own, or you can hover your mouse over the colored words to see the coordinately colored sections of the graphic. Hover over "Black and White" to the right to see the unembellished version as Mahmud II would have written it.

As the graphic fills in, you may notice sections that never turn from the background gray color. This is because they are not characters in the signature, but, rather, are artistic embellishments to complement the rest of the design.