When talking about Hebrew, in English, it is often handy to be able to show what Hebrew words you're talking about without using Hebrew letters. Sometimes, this is because of the limitations of computer systems in handling Hebrew -- I don't know what fonts you have installed on your computer, which means I don't really know what this website looks like to you. But it can also be useful to transliterate if you're speaking to an audience that is not fully comfortable with the Hebrew alphabet.
My intended audience is mixed: this stuff is written for people who can read Hebrew, and for people who can't. I suspect that many of the people who are interested in the topics I write about know a smattering of Hebrew, and are trying to get some more information, but wouldn't be able to just comfortably read in Hebrew.
We immediately run into a variety of problems. Hebrew has some letters that don't really have a good English equivalent, it has some letters that are sometimes pronounced similarly to each other, and there are even different traditions of pronunciation. Over time, I've developed some opinions about what are better and worse ways to transliterate Hebrew into English, and I figure I might as well write them out, in case any finds it useful.
I think we can summarize roughly three basic ways one might want to transliterate Hebrew, which might perhaps best be explained with an example. Consider the first verse of Genesis:
בראשית ברא אלהים את השמים ואת הארץ
We might also show it with the vowel-points and accents. Your browser may or may not do a good job with these:
בְּרֵאשִׁ֖ית בָּרָ֣א אֱלֹהִ֑ים אֵ֥ת הַשָּׁמַ֖יִם וְאֵ֥ת הָאָֽרֶץ
A lot of that second one looks like a bit of a mess as I'm looking at it now in Obsidian.md, my preferred program for editing web pages.
Anyhow, let's go through our three use cases. First, I might be focused on the consonants, and just want to show you that. For that, I would simply replace each letter with an Latin letter, or in a few cases a symbol. That might look like this:
BRʾŠYT BRʾ ʾLHYM ʾT HŠMYM WʾT HʾRṢ
Now, sometimes we want to talk about the vowels. For example, perhaps we're interested in how the first vowel point in Genesis is a shva. If we want to show the vowels, but we're still interested in providing a pretty precise sort of technical-ish transliteration, we might produce something like this:
bᵉrešit bara ʾelohim ʾet haššamayim wᵉʾet haʾareṣ
That has some advantages -- for one, a lot of people will be able to take a stab at pronouncing it. But the diacritics are still a problem for some purposes. On some sorts of computer systems, they might not display -- although this kind of thing is becoming less of a problem all the time as Unicode has become pretty much universal. But diacritics continue to be a problem for non-specialist readers. How does someone pronounce an š, for example? It may not be intuitively obvious that this is like the 'sh' in 'shin'. So the third sort of transliteration might be one that more readable.
You could then wind up with something like this:
bereshit bara elohim et ha-shamayim we-et ha-arets
The letters of the Hebrew alphabet, each of which represents a consonant, are as follows:
אבגדהוזחטיכלמנסעפצקרשת
If we say their names, they are alef, bet, gimel, dalet, he, vav, zayin, het, tet, yod, kaf, lamed, men, nun, samek, ayin, pe, tsaddi, qof, resh, shin, tav.
I'd like to comment on every one of the letters, but in order to make the discussion easier to follow, let's not do this in alphabetical order, but according to groups of letters that present similar issues, starting with the four letters that present no issues at all.
Here we have four fairly easy letters. Lamed is L, full stop. True, in English we have at least two pronunciations of L -- there's the L's in lulu, and the L's in all. In Hebrew, all the L's are like lulu. Nevertheless, no transliteration scheme is going to teach readers all the differences in pronunciation between their native language and the language transliterated from, so we don't need to worry about that. L works just fine for lamed, and there are no other Hebrew letters that have any sort of L sound, so this one is easy.
In the same way, mem and nun are to be transliterated by M and N, and in transliteration M and N will always represent mem and nun. Resh is universally represented by R, and though the Hebrew letter may be hard to pronounce, it's easy to transliterate, as there are no other letters in Hebrew that make any R sound.
Next, let's consider alef and ayin. English has no obvious equivalent for either. They are conventionally represented by the characters ʾ for alef and ʿ for ayin. These might look like apostrophes, but they're not -- these are Unicode characters represented by the codes 02BE and 02BF. The one big downside of these two symbols is that in many fonts they are very small and easy to miss. If you have Gentium Plus on your computer, that is what this website will display in, and it shows them well.
Nevertheless, if you aren't satisfied with ʾ and ʿ , you could use ʔ and ʕ instead, but this is not done very commonly, and it's a bit ugly, so in the interests of maintaining widely used standards, I now use ʾ and ʿ , though in the past I sometimes used ʔ and ʕ.
Though ayin, ʿ , is silent in Modern Israeli Hebrew, it was very much a full consonant in classical and Tiberian forms of Hebrew, and it was a full consonant in all its uses. And so I would just about always recommend showing it in transliteration. Thus, if you wanted to use a Hebrew word for 'evening', in all-consonant mode I'd suggest ʿRB, while in vowels-included mode I would go with ʿereb. If you're wanting to be really beginner-friendly, and aren't concerned about preserving the Hebrew spelling at all, then you could potentially drop it and write ereb or maybe even erev. But we'll get to the pesky issue of b/v distinctions later.
Alef, ʾ, is a little tricker, because it is sometimes quiescent, that is, it is not independently pronounced, but serves as a sort of placeholder for a vowel. This is the case in a particular Hebrew verb meaning 'he created'. In all-consonants form, all consonants should all be retained, so there I'd recommend BRʾ. But if we're adding in vowels and then getting sensitive to details of pronunciation, then I could see a case being made either for baraʾ or bara. I suppose it will come down to the relative importance you're placing on preserving the consonants vs. showing the Tiberian pronunciation. And if you're trying to translate in a less technical way, the case for more often dropping the alef becomes stronger. When writing to a totally non-technical audience for non-technical reasons, I cannot imagine what might be gained by writing ʾelohim instead of elohim.
Next, we come to the letters bet, gimel, dalet, kaf, pe, and tav. In their non-fricative 'normal' form, these letters are conventionally transliterated as b, g, d, k, p, and t. However, each of these letters, at some times and places in the history of Hebrew, have had an alternate fricative or 'soft' pronunciation. The soft b is pronounced like an English v, the soft p like the English f, the soft t like the 'th' in 'through', and the soft d like the 'th' in 'that'. Today's English does not have good equivalents for the soft g or soft k.
Let us start with gimel. In Modern Israeli Hebrew, the 'soft' gimel variant does not appear, and consequently in most academic contexts nobody is going to be trying to pronounce it either. I don't see anyone gaining much by trying to represent it in most cases, so I recommend simply transliterating gimel as G. Likewise, the soft d is no longer pronounced, and I would recommend just transliterating dalet as D across the board.
The case of tav is a little bit harder. The soft tav variant does not appear in Modern Hebrew, but it has appeared recently enough in some parts of the world that it has left a lasting imprint on biblical names as they are rendered in English. Thus, for example, when one reads about a biblical character named 'Seth' in an English Bible, it goes back to an old Hebrew name that ended in a soft t. In any case, in all-consonants mode, where we're not concerned about transliteration, we should just use T. So the first word of Genesis would be BRʾŠYT. Even in other cases, though, I'd recommend just sticking with T. Sticking with T adheres to the pronunciation found in Israel and academia, and it keeps the letter tav and equivalent strictly to one English letter.
The case of b is a little harder. The b/v distinction exists both in the Israeli context and in Hebrew as read aloud in academic settings. In consonant-only mode, of course, B should be strictly retained as B, because when displaying consonants only one is not really concerned about pronunciation. But it could otherwise we very tempting to start transliterating 'soft' b as v. Still, I think one should resist the temptation for a simple reason. The Hebrew vav, conventionally transliterated in biblical contexts as w, is sometimes also transliterated as v. And thus if a reader sees a v, the reader may not know whether they are looking at a bet or a vav. And so it seems to me that it is best to keep rendering bet as b and vav as w, avoiding v completely except in cases where one is explicitly discussing the issue of hard vs. soft b.
In the case of k, the 'soft' k is sometimes written as kh, and I'm sympathetic to this. In consonants-only mode, of course, we will just stick with K. However, kh has the drawback of being potentially confusing to readers, and to introducing one possible source of confusion with het, which people sometimes represent as kh. I might myself use kh in a situation where pronunciation is being especially discussed, but in all other instances I'd recommend sticking with k. People experienced in Hebrew will knew when it is pronounced as a fricative, and people not experienced won't -- but people not experienced with Hebrew are likely to have a similar problem with kh anyhow. It's a close call, but I generally recommend k. I very much recommend against reading a soft k as ch, as this causes even more confusion, both with the letter het, and confusion about pronunciation, as ch in English tends to be an entirely different sound.
There is one hard/soft pair where I do recommend showing the distinction. This is the case with the soft p, which is pronounced f. I do this not because the distinction is particularly important in terms of meaning, but simply because it doesn't possess the drawbacks some of the other pairs have. The p/f distinction is maintained in Hebrew and English, so one doesn't risk teaching an obsolete pronunciation to readers. f is perfectly plain to readers, unlike kh, so it doesn't risk creating unnecessary confusion. And there is no other letter in Hebrew that is an f, so it doesn't create confusion about the Hebrew alphabet. In particular, I prefer f to ph for the soft p, as it continues to render one Hebrew letter by one Latin letter. This is a little idiosyncratic on my part, and I may one day regret this, because it is inconsistent that I should only display the begadkefat phenomenon on one letter. Perhaps for consistency I should just stick with p across the board.
Both he and het might of thought of as 'H' letters, though there are some issues with both. Let's start with he. The he is more of a 'classic' H -- when it is pronounced it sounds like an English H, more or less. So in consonants-only mode, we should spell it as a simple H. Thus, a famous Hebrew term for 'law' would be TWRH, and 'the mountain' would be HHR.
Things get a little more complicated as we move to transliteration styles that include vowels. In Hebrew, the letter H often stands for an -a vowel, at the end of a word, with the H not being pronounced as a consonant at all. Thus, the word ishah, without the final 'h' pronounced, means 'woman'. And because in English -ah at the end of a word typically does not have the 'h' pronounced, that transliteration seems to work out neatly. But this is complicated by the fact that sometimes the h is deliberately pronounced, and this distinction can be meaningful. So, to return to ishah, if you say ishah with the final h pronounced distinctly, that's a different thing in Hebrew, and means 'her husband'. So one could argue that we should omit silent H and show the pronounced H, thus distinguishing isha 'woman' from ishah 'her husband'. On the other hand, by doing that you do a better job of showing pronunciation, but you're dropping a letter from the Hebrew text. In any case, a Hebrew h should be represented by h, if it is represented at all.
We come next to het. This is pronounced in a fashion a bit scrapier than an English h. Just how much scraping you hear will depend on the reading tradition a bit. In some traditions, a het is a bit scrapier than an h, but not as scrapy as a soft k. In Modern Israeli Hebrew, it is indistinguishable in pronunciation from a soft k, but it is still distinguished in spelling. The most common transliteration in academic circles in an H with a dot under it: Ḥ or ḥ. This equivalent has a double advantage: for specialists, it immediately tells them which Hebrew letter they are dealing with, while for non-specialists, anyone who ignored the diacritic will still come up with a pronunciation roughly in line with the correct one. This is not the case with the common non-academic practice of writing het as 'ch', which both confuses het and kaf and has the potential to be seriously misleading for pronunciation.
If you are in a position where you absolutely cannot use ḥ, then I would recommend using a simple h. Though it erases the distinction between he and het, at least it is not too misleading, and it continues to render one Hebrew letter with one English letter.
The letter that in Modern Israeli Hebrew is called vav is sometimes in academic contexts referred to as waw, and in consonantal transliterations it is typically read as 'w', which sticks to a more ancient pronunciation. I recommend sticking with w as opposed to v. This has the advantage of both showing the likely biblical pronunciation of the consonant and avoiding any ambiguity with bet.
Sometimes, a vav is written not to represent the consonant w, but to represent a vowel: o or u. In those cases, only the vowel should be represented in translituation. Thus TWH WBHW should be written tohu wa-bohu. There is another way of doing things, as shown in Strong's concordance, which would suggest rendering both the vowel and the associated consonant: towhu wa-bowhu. But this is a bad idea, as it could mislead the reader into thinking the consonant is somehow still conserved where it stands only for a vowel.
There is a whole menagerie of s/z sounds in Hebrew, and perhaps it is best to start with the simplest of them: samek, which is represented by a simple s. I no of no good reason not to follow this transliteration, which is seems everyone agrees on. Samek is a nice simple s.
Next we come to the letter shin. Though one letter in Hebrew, it has two pronunciations, and this meaningful distinction between pronunciations goes all the way back to before Hebrew was Hebrew. So it is useful to show the distinction between shin and sin. And sin was not always pronounced identically to samek: it used to have a pronunciation which was identical to neither shin nor samek. So it is best to distinguish all three. In technical contexts, this is done by writing sin as Ś and shin as Š. In all-consonants mode, I see no reason to drop this practice.
However, when rendering Hebrew in non-technical contexts, it may be helpful to show the correct pronunciation of š by writing it as 'sh'. Thus, in non-technical contexts a reader may be better served by seeing shalom than by being treated to the confusing šalom. Each has its benefits and drawbacks.
The letter tsaddi in Modern Israeli Hebrew is pronounced a bit like the 'zz' in pizza, and is sometimes written as 'tz' in informal contexts, like in the term 'matzah'. I would recommend, for these sorts of informal contexts, 'ts' instead, as it's a bit more precisely accurate in terms of pronunciation. But the normal academic equivalent is ṣ, which should be used whenever diacritics are available.
Finally, there is zayin, a simple z.
The letter tet, in Modern Israeli Hebrew, is pronounced identically to tav. For the sake of distinguishing them, tet is conventionally ṭ. I would recommend following this convention when you can.
As a consonant, yod is y, and as a vowel i or e, as the case may be.
The letter qof is conventionally represented by q, and in Modern Israeli Hebrew qof is pronounced identically to a hard k. However, to maintain the distinction found in the Hebrew alphabet, using q seems like a good idea to me. It's only possible drawback is that occasionally a reader might be confused into adding a u after it, following the English principle that q's precede u's. But I'm not too worried about this.
Modern Israeli Hebrew has basically five vowels, roughly the same as a Spanish vowels a, e, i, o, u. This is somewhat at odds with the strict Masoretic system, but I recommend basically accomodating Modern Israeli Hebrew, and not bothering with diacritics to distinguish vowels.
Thus, qubuts and shuruq should be read as u, holam and qamats katan as o, hiriq as i, segol and tsere as e, and patah and qamats gadol as a. This is enough, generally, to show the meaningful morphological distinctions that a reader might need, without wading too deep into marking up every subtle distinction. It works in Modern Israeli Hebrew speech, and I think it'll work just fine in transliteration.
Shva is a special case. In formal contexts, I'd recommend using Unicode 1D49, the superscript e: ᵉ. Thus, the WYHI of Genesis 1:5 would be written wayᵉhi. I'd recommend using the special Unicode character and not simply using a superscripting tag or function so that if your formatting is one day lost you won't lose the superscript. In cases where the little ᵉ character is unavailable or inappropriate, I'd recommend just using your best judgment on a case-by-case basis to either omit the shva altogether or to use an e for it.
Modern Israeli Hebrew does not pronounce gemination, while older forms of Hebrew did. When in consonants-only mode, I would recommend making no effort to show gemination. Thus 'the sky' in Genesis 2:4 should be HŠMYM, not HŠŠMYM. However, if you're showing vowels, and therefore getting into how the Hebrew is read, then I think the presumption should be in favor of gemination. Thus, in a technical but vowelled format, we'd read ʾišša for 'woman' rather than ʾiša. But if you move down the formality scale and are no longer using diacritics, I'd recommend against writing something as barbaric as ishsha. Better than that would be to simply drop the gemination and write ishah.
Another reason not to use gemination might be if you are separating prefixed articles. So 'sky', šamayim, strictly speaking becomes haššamayim if you want to say 'the sky'. But if you want to make the ha- distinct, you could separate it with a dash, and read ha-šamayim, rather than something weird like haš-šamayim, which might misleadingly suggest the article contains a shin, or the even worse ha-ššamayim, which makes it seem like the word 'sky', by itself, starts with a doubled letter.
It also may make sense to separate the word and from any word to which it is prefixed: thus, wᵉ-et might be a little clearer than wᵉet. A time this should not be done is with a wayyiqtol verb form like wayyomer, where the gemination is an important grammatical feature and the 'and' an integral part of the verb rather than 'just' a conjunction.
One oddity of Hebrew is the occasional silent yod, like in SWSYW, 'his horses'. Here, I'd say just drop the yod if you're providing a vowelled transliteration: susaw rather than susayw, which could mess up the pronunciation for a reader.
Occasionally you might see someone giving consonants in capital letters and vowels lowercase, like perhaps saying MeDaBeReT instead of medaberet. This is ugly, even if there are good intentions behind it. I recommend against this on sheer grounds of ugliness. I did it earlier in this article with BeGaDKeFat, and I already regret doing so.
In any case, transliterating Hebrew is a bit more of an art than a science, unless you want to do something boring like strictly following the SBL's rules.