The Hebrew Alphabet is a Mess

For the past week or so, I've been learning modern Hebrew on Duolingo, just for fun. Hebrew is a fascinating language, and an amazing example of a nearly-dead language taking on new life to become a flourishing part of everyday life for millions of people. But as my title already says, the Hebrew alphabet is a mess.

The Hebrew alphabet—or alef‑bet—officially has 22 letters, but five of these letters have separate forms at the end of a word. If that were all, it might be OK. But that's not all—not by a long shot. First of all, vowels aren't written in Hebrew, so when you see a word for the first time, even if you know the consonants, there's no way to know what the vowels are, which can completely change the meaning of a word. And knowing the consonants isn't a trivial matter either!

Letters with more than one sound

There are six consonant letters that can each make more than one sound. So bet sometimes makes a b sound and sometimes a v sound; he is sometimes pronounced like an English h and is sometimes not pronounced at all; vav is sometimes pronounced as a v, sometimes as an o, and sometimes as u; yod can either be the y sound or the i sound; kaf can either make a k sound or a kh sound (like Scottish ch in ‘loch’); pe makes either a p or f sound; and shin/sin is either pronounced sh or s. Of course, it's possible to use special dots on the letters to show the exact pronunciation, but this is only common in dictionaries and very basic children's books. This means that as an adult learning Hebrew, when you see a new word, you really have to guess how to pronounce it. For example, seeing the word שכב (shin/sin-kaf-bet) for the first time, there's absolutely no way to know whether it's pronounced sakab, sakav, sakhav, sakhab, shakab, shakav, shakhav, or shakhab. (It's shakhav.) And once again, that's assuming you know the vowels as well. There are some patterns—like that pe, bet, and kaf are more likely to be pronounced as p, b, and k at the start of a word and as f, v, and kh everywhere else—but there are no guarantees!

Letter	Sound(s)
א	(silent)
ב	b or v
ג	g
ד	d
ה	h
ו	v, o, or u
ז	z
ח	kh
ט	t
י	y or i
כ, ך	k or kh
ל	l
מ, ם	m
נ, ן	n
ס	s
ע	(silent)
פ, ף	p or f
צ, ץ	ts
ק	k
ר	r
ש	sh or s
ת	t

Sounds with more than one letter

On top of all this, when you hear a new word for the first time, there's often no way to tell how it should be spelled. It seems like with twenty-two letters, six of which have multiple pronunciations, there should be around twenty-eight different consonant sounds in Hebrew, right? Well, actually there are only twenty consonant sounds. This is because seven of the consonant sounds of Hebrew can be written in multiple ways, so even with multiple pronunciations for some letters, there are still fewer sounds than letters. The v sound can either be written with bet or with vav; the kh sound is either kaf or khet; the t sound is either tet or tav; the k sound is either kaf or kof; the s sound is either samekh or sin; and to top it all off, in spoken Israeli Hebrew, there are two letters (alef and ayn) that are always silent and one (he) that is usually silent at the end of a word, and sometimes silent at the start of the a word. This means that if someone tells you the word for cake in Hebrew is ooga, there's no way to know whether it's אוגה, אוגא, אוגע, עוגה, עוגא, or עוגע. (It's עוגה.) Even for a seemingly straightforward word like katav (“he wrote”), there are eight possible spellings (כתב, כתו, כטב, כטו, קתב, קתו, קטב,  or קטו—the right one is כתב)!

Sound	Letter(s)
(silent)	א, ע, (ה)
b	ב
v	ב, ו
g	ג
d	ד
h	ה
z	ז
kh	ח, כך
t	ט, ת
y	י
k	כך, ק
l	ל
m	מם
n	נן
s	ס, ש
p	פף
f	פף
ts	צץ
r	ר
sh	ש

The vowels are even worse

Of course, these lists don't even include the vowels. They're rarely written, but when they are, they have the same problem as the consonants. By most accounts, there are only five vowel sounds in modern Israeli Hebrew: a, e, i, o, and u, which are pronounced the same as in most European languages. That seems fairly straightforward, except that for those five vowels, there are fourteen different sets of niqqud (the little dots used in children's books and dictionaries). So when you hear the a sound, it could be אַ or אָ or אֲ or אֳ (all of these examples are given with alef as the base letter); when you hear an e sound, it could be אֵ or אֶ or אֱ or אְ; and the i, o, and u sounds can all be written either with or without a helping consonant (yod for i and vav for o and u). In older dialects of Hebrew, all these different symbols made different sounds—some were long and others were short, one a was pronounced further back than the other, etc.—but in modern Israeli Hebrew, there's really no way to tell which of the four a sounds someone is actually pronouncing, or whether or not to write a vav for an o sound. Because of this, hearing a word like zot for the first time, there's no way to know whether the o is just written as a dot (i.e., usually not written at all) or written with a vav. (In this case, it's actually not written at all, and there's a silent alef where the 'vav' would be: זאת [zayin-alef-tav].) A word with multiple vowel sounds ends up having multiple possible spellings, like lehitraot, which has sixty-four (!) different potential spellings, depending on the different versions of e, i, a, and o used. (And that's assuming you already know the right consonants to use!)

64 possible ways to write lehitra'ot:
לֵהִיתרָאוֹת, לֵהִיתרָאֹת, לֵהִיתרַואֹת, לֵהִיתרַאֹת, לֵהִיתרֲאוֹת, לֵהִיתרֲאֹת, לֵהִיתרֳאוֹת, לֵהִיתרֳאֹת, לֵהִתרָאוֹת, לֵהִתרָאֹת, לֵהִתרַאוֹת, לֵהִתרַאֹת, לֵהִתרֲאוֹת, לֵהִתרֲאֹת, לֵהִתרֳאוֹת, לֵהִתרֳאֹת, לֶהִיתרָאוֹת, לֶהִיתרָאֹת, לֶהִיתרַאוֹת, לֶהִיתרַאֹת, לֶהִיתרֲאוֹת, לֶהִיתרֲאֹת, לֶהִיתרֳאוֹת, לֶהִיתרֳאֹת, לֶהִתרָאוֹת, לֶהִתרָאֹת, לֶהִתרַאוֹת, לֶהִתרַאֹת, לֶהִתרֲאוֹת, לֶהִתרֲאֹת, לֶהִתרֳאוֹת, לֶהִתרֳאֹת, ל`הִיתרָאוֹת, ל`הִיתרָאֹת, ל`הִיתרַאוֹת, ל`הִיתרַאֹת, ל`הִיתרֲאוֹת, ל`הִיתרֲאֹת, ל`הִיתרֳאוֹת, ל`הִיתרֳאֹת, ל`הִתרָאוֹת, ל`הִתרָאֹת, ל`הִתרַאוֹת, ל`הִתרַאֹת, ל`הִתרֲאוֹת, ל`הִתרֲאֹת, ל`הִתרֳאוֹת, ל`הִתרֳאֹת, לֱהִיתרָאוֹת, לֱהִיתרָאֹת, לֱהִיתרַאוֹת, לֱהִיתרַאֹת, לֱהִיתרֲאוֹת, לֱהִיתרֲאֹת, לֱהִיתרֳאוֹת, לֱהִיתרֳאֹת, לֱהִתרָאוֹת, לֱהִתרָאֹת, לֱהִתרַאוֹת, לֱהִתרַאֹת, לֱהִתרֲאוֹת, לֱהִתרֲאֹת, לֱהִתרֳאוֹת, לֱהִתרֳאֹת

Vowel	niqqud, etc.
a	אַ, אָ, אֲ, אֳ
e	אֵ, אֶ, אְ, אֱ
i	אִ, יִ
o	אֹ, וֹ
u	אֻ, וּ

Time for a reform?

When you add the vowels' ambiguity to the consonants' ambiguity, you end up with even more possible spellings for any new word you hear. The reformist in me wishes we could just pare it down to one sound per letter, and one letter per sound. So, for example, instead of bet sometimes being pronounced as b and sometimes as v, and vav also sometimes being pronounced v, bet would always be pronounced b and vav would always be pronounced v. As long as we're thinking about changing things, there's really no reason for five letters to have different word-final forms while the other letters don't, so we could get rid of the word-final forms, too. This would give us exactly twenty consonant sounds and twenty letters:

Sound	Letter(s)
(silent)	א
b	ב
v	ו
g	ג
d	ד
h	ה
z	ז
kh	ח
t	ט
y	י
k	ק
l	ל
m	מ
n	נ
s	ס
p	פ
f	ף
ts	צ
r	ר
sh	ש

If we did this, we could keep the word-final form of pe for the f sound. This is the only alternate sound that can't be written with a separate letter, so keeping the final form would be a reasonable thing to do. Also worth noting is that kaf, ayn, and tav would be removed completely, since their sounds are covered by other letters (kof, alef and tet).

This all seems well and good. If we did this, a word like shakhav could be unambiguously written as שחו (shin-khet-vav), and a person who's never seen the word before would know exactly how to pronounce it. (Well, the consonants, anyway.) But our spelling reform would have some unfortunate unintended consequences. Just like English, Hebrew's weird spelling system actually has one benefit: it makes it easier to see the relationships between words. A good example of this in English is the word electric and the related word electricity. If we got rid of the letter C and just used k or s instead, we could spell these words elektrik and elektrisity, making the pronunciation more closely tied to the spelling, but also making it ever so slightly less obvious that these words are related. This is the case even more in Hebrew, where words like pa'al and nif'al are less obviously related when the p and the f are completely separate letters.

A compromise

For this reason, along with a host of others, Hebrew speakers haven't implemented any major alphabetic reforms. But they have introduced a few conventions that make it ever so slightly less impossible to figure out the correlations between spelling and pronunciation. One of these conventions—using yod for the I sound and vav for the o and u sounds—actually dates back centuries. Without it, a word like ooga (which means ‘cake’, from a few paragraphs back) would simply be written עגה (ayn-gimel-he), and we would have absolutely no idea what the vowels were. With the vav, we can at least narrow the first vowel down to one of two possibilities. Along with this, when vav does make a v sound instead of a vowel sound, it's often written as a double vav, as in the word barvaz (ברווז) ‘duck’, which, without the double vav could be confused for barooz or baroz. Similarly, when yod makes an ay or ey sound, it's often written as a double yod, as in tsaharaym (צהריים) 'noon'.

While these conventions don't solve the problem—it still feels very much like the Hebrew alphabet was designed for a completely different language—they help a little. And short of convincing every Hebrew speaker everywhere to change the way they do things, I don't think there's much I can do other than just view reading and writing in Hebrew as a special sort of challenge. After all, if I can figure this out, that's saying something, right?