A quarter of a million shekels to settle how to spell Petah Tikva on a road sign. That's what the Prime Minister's Office just put out — a notice of intent to contract the Academy of the Hebrew Language to build one unified transliteration system for every government ministry and local authority. Work starts October fourth, runs a full year, and objections to the whole arrangement close this Thursday at eleven thirty at night.
Which is the most bureaucratic deadline I've ever heard. Eleven thirty at night. Who files an objection at eleven thirty?
Somebody who spent the evening driving from Petakh Tikva to Petah Tiqwa and isn't sure they arrived.
Daniel's prompt is about exactly this. The Academy stepping in with the road authorities to standardize how Hebrew place names get rendered in English. His example is Petah Tikva — which he's seen as Petakh Tikva with a K-H and as Petah Tiqwa with a Q — and Ashkelon versus Ashqelon. And he admits something I think a lot of people feel quietly: he's delighted when he sees Tiqwa with a Q because it looks like the older, fuller form of the word. Even though nobody pronounces it that way and hasn't for a long time.
Daniel's got three questions folded in here. First, how has Israel actually grappled with this over the years, because it's not new. Second, why did different practices emerge in the first place — and he suspects competing visions of what transliteration is even for. Third, what's the Academy's role in all this, given how much weight it carries in the life of the language. I want to start with the second question, because I think it explains the other two.
It does. The split at the heart of all of this is pronunciation versus spelling. Do you write the name the way it sounds when a Hebrew speaker says it today, or do you write it as a map of the Hebrew letters themselves? Those are different jobs. The Academy's own stated position is that its rules aim to reflect pronunciation, not necessarily the spelling in the source language. But that's a choice, and for decades plenty of people made the other choice.
The Q is the fossil of the other choice. Kuf is a letter that, in modern Israeli Hebrew, sounds identical to kaf — both are just a K sound. But historically kuf was a deeper, more emphatic consonant, something closer to the Arabic qaf. So a system built on letter correspondence renders kuf as Q. Kuf, Q. One letter, one Latin character. That's where Ashqelon and Tiqwa come from.
And it's not just kuf. Vav got rendered as W for a long time because the letter historically did double duty — it was W in some positions and V in others. Tsadi got rendered as Z in some systems because it's one letter and Z was available, even though it sounds like TS. Chet is the messiest — ch, kh, h, or h with a dot under it, depending on whether you're coming from German, English, or academic Semitics.
So the signs with Q on them aren't wrong. They're artifacts of a different philosophy. They're the letter-by-letter school, and they preserve distinctions that modern Hebrew pronunciation has collapsed. Daniel says he's quietly delighted by Tiqwa because it shows how the word might traditionally have been rendered. The linguists who built that system would say it also shows how the word was traditionally pronounced — just not by anybody alive today in a secular Israeli accent.
And that's the second axis of disagreement the research turned up. Even pronunciation isn't one thing. The same Hebrew letter chet sounds different in mainstream secular pronunciation than it does in Mizrahi pronunciation, or in Arab communities, or in Ashkenazi Orthodox circles. So a system that tries to track pronunciation has to pick which pronunciation. That's not a linguistic question anymore. It's a demographic one.
Then there's the third axis, which is the exonym problem. Jerusalem versus Yerushalayim. Jaffa versus Yafo. Acre versus Akko. Safed versus Tzfat. Those English names don't track Hebrew pronunciation or Hebrew spelling — they're historical English words with their own histories. Caesarea is the wild one: it's appeared on signs as Caesarea, Qesarya, Qesariyya, and Ceysaria. Four renderings of one place, and only one of them tells an English speaker how to say it.
Ceysaria looks like somebody was trying to invent a new letter.
Ceysaria is what happens when a committee compromises with itself.
So Daniel's instinct is right. The differing practices come from different theories of what transliteration should do. Letter fidelity, sound fidelity, historical English convention, and then the political question of whose language counts as the original. All four of those were in play, and different institutions picked different ones.
Let's trace the history, because the timeline is actually the best way to see how the problem resists being solved. It's been declared almost solved many times, and then it comes back in a new form. The British Mandate published the first official trilingual list of geographical names in nineteen thirty. Hebrew, Arabic, Latin.
Then in nineteen forty-nine, Ben-Gurion set up a commission. Its original job was to Hebraize place names in the Negev — give Hebrew names to places that had Arabic ones or none at all. Within about two years it became the nationwide Government Names Committee. So the very first Israeli institution for place names was about renaming, not standardizing. That's a clue about where the energy was.
Nineteen fifty-seven is the key year. The Academy of the Hebrew Language approves its first rules for transliterating modern Hebrew into Latin characters. It includes a simplified system specifically for everyday use — ads, street signs. The Survey of Israel adopts it. Nineteen seventy-seven, the UN recommends it to the whole world.
And here's the thing — the UN adopting it didn't fix Israel's signs. The UN materials themselves later acknowledged that alternative systems kept being used, including by public bodies. So the standard existed, it was internationally recognized, and the state that created it wasn't fully following it.
Nineteen ninety-five. Education Minister Amnon Rubinstein asks for a review. His argument is that forms like Petah Tiqwa and Qesariyya work terribly as road names. A tourist sees Tiqwa with a Q and has no idea how to ask for directions. The spelling is phonologically faithful to a pronunciation that no longer exists, and it's actively misleading.
So the Academy goes back to work. In two thousand six, the plenum tasks its grammar committee with deciding the simplified transliteration. The committee rules in June. The changes are exactly what you'd predict: vav becomes V instead of W, kuf becomes K instead of Q, tsadi becomes TS, and the automatic E for a non-pronounced schwa gets dropped. So Bene-Beraq becomes Bne-Brak.
That's the moment Tiqwa died, officially.
Officially. In two thousand seven, Israel presents the revised system to the UN, and it's adopted as resolution nine of the ninth conference. And they actually tested it empirically — two hundred twenty tourists who didn't read Hebrew pronounced Latinized place names, and researchers analyzed about fifty-five hundred recordings. The goal was that someone could read a name off a map or a sign and ask for directions well enough to be understood.
I love that they tested it on tourists. That's the right instinct. The whole point of a road sign is that a person who doesn't speak the language can use it.
Right. And then two thousand nine is where it gets political. Transport Minister Yisrael Katz orders road signs changed so English and Arabic names are direct transliterations of Hebrew. Jerusalem becomes Yerushalayim. Nazareth becomes Natsrat. Caesarea becomes Kesariya. His quote was, I will not allow anyone to turn Jewish Jerusalem into Palestinian al-Quds.
And that blew up exactly the way you'd expect. Arab MKs accused him of erasing Arabic. Ahmad Tibi said al-Quds will remain al-Quds and Shfaram will remain Shefa-Amr. The deeper issue is that for Arab towns, the Arabic name isn't a transliteration of Hebrew — it's the original name. So forcing everything through Hebrew spelling was an erasure, not a standardization.
In twenty twelve, the government tried to settle it with a rule. Jewish settlements are transliterated from the Hebrew form. Arab settlements from the Arabic form. And Eilat is an explicit exception, written E-I-L-A-T. I'm not sure why Eilat needed its own carve-out, but it got one.
Eilat is the Switzerland of Israeli toponyms. Neutral, non-aligned, spelled the same in every language.
Twenty twenty-one, MK Eitan Ginzburg proposes a law. The Academy determines spelling of public place names, the state maintains a public online database. Preliminary reading passes thirty-three to nothing. Then it dies somewhere in committee, because that's what happens to things that everyone agrees about in principle and nobody wants to fund in practice.
And the Knesset hearings in twenty twenty-two are where the real state of play comes out. The Government Names Committee approves some names, municipalities handle streets, the Interior Ministry and Transport Ministry each keep their own arrays, the Cartographic Center handles maps, and the railway is responsible for its own objects. That's five or six separate systems.
The Academy's president, Moshe Bar-Asher, said state institutions have ignored the Academy's decisions for years — concerning tens of thousands of names. And the Academy's scientific secretary, Ronit Gadish, said the problem isn't a lack of rules. It's implementation. The Transport Ministry has known about this and failed to fully fix it for fifteen years.
That's the thing Daniel's prompt gets at without quite saying it. The Academy has had a standard since nineteen fifty-seven. It revised the standard in two thousand six. The UN adopted the revision. And the signs are still inconsistent. So the new quarter-million-shekel contract isn't about linguistics. It's about whether anyone can make the state actually follow its own rules.
The scope of the new project is telling. It's not about coining new names. It's about mapping existing transliterations at national and local levels, collecting examples from official documents and public information systems, identifying inconsistencies, and producing a final updated list of official place names. This is an audit with a cleanup crew attached.
And the price — two hundred fifty thousand shekels, about eighty-four thousand dollars. For a year of work. That's... Not a lot of money for what's being asked. Mapping every place name in every government database and reconciling them is a large data project.
There was a smaller precursor in February — a twenty-six-thousand-shekel transliteration consultant contract. So this has been building. Somebody in the Prime Minister's Office decided the piecemeal approach wasn't working and the Academy needed a real mandate.
Let's talk about why the algorithm angle matters, because Daniel raised it and I think it's the part most people haven't thought about. A human driver can see Petah Tiqwa and Petakh Tikva on two different signs and figure out they're the same place. A mapping system can't. It's matching strings, not meanings. If one database says Ashqelon and another says Ashkelon, the system treats them as two different entities.
The government's own rationale is about shared identifiers. Multiple databases exist, and one state dataset on the open data portal explicitly warns that the National Digital Directorate isn't responsible for its accuracy. So you have the same place name stored differently in different systems, and no single authoritative list to reconcile them against.
I should be careful here — I haven't seen a study that quantifies how many mapping errors are caused specifically by transliteration variants. The failure is real in principle, and the government clearly believes it's happening, but I don't have a number for how many wrong turns it's caused. The human cost is better documented. The Transport Ministry's own planning chief said in two thousand nine that the lack of uniform spelling impairs drivers' ability to find their way. And there was a Knesset testimony about an American tourist driving to the Kinneret who saw Sea of Galilee on one sign and Sea of Kinnereth on another and called to ask if he was still going the right way.
Kinnereth with a T-H at the end. That's the letter tav rendered as TH, which is an old convention from when tav was thought to have a softer pronunciation in some positions. Nobody says Kinnereth. But the sign was following a letter-fidelity rule.
The signs are a museum of dead phonological distinctions. Q for kuf, W for vav, TH for tav. You can date a sign by which extinct consonants it still believes in.
That's actually a beautiful way to put it. The road network is a stratigraphic record of transliteration philosophy. Drive from Jerusalem to Haifa and you pass through the nineteen fifties, the nineteen seventies, and the two thousand six reform, all still standing.
Now, the Academy itself. Daniel calls its role fascinating and outsized, and he's right. It was established by Knesset law in nineteen fifty-three, succeeding the Language Committee that Ben-Yehuda started in the eighteen eighties. The Language Committee coined something like thirty thousand terms across sixty subject areas. Bialik personally coined the words for airplane, camera, and reaction.
Bialik coined matos, matzlema, and tguva. The national poet was also the national neologist. And the Academy inherited that role — it's the official authority on Hebrew, including transliteration. Its nineteen fifty-seven system became the basis for the UN standard and the American and British government standards. So this isn't a small academic body. It's the closest thing Hebrew has to a legislature.
But here's the tension. The Academy can rule, but it can't enforce. Bar-Asher's complaint about tens of thousands of ignored names is the whole story. The Academy is a standards body with no police force. It can publish the correct spelling of Ashkelon, but it can't make the municipality repaint the sign.
And the road authorities have their own legitimate concerns. The Transport Ministry has argued that overly long spellings distract drivers — you're reading a sign at highway speed, and a linguistically precise rendering might be too long to parse quickly. So there's a genuine conflict between what's linguistically accurate and what's safe on a road sign. A language academy and a roads authority have openly competing mandates.
The road sign is a safety device before it's a linguistic document. That's not a compromise people like to admit. The Academy wants the name to be correct. The Transport Ministry wants the driver to not die.
There's also a human cost buried in the history that I think about. The head of the Cartographic Center who participated in the reform work, Aviel Ron, was killed in a suicide bombing in two thousand two, along with two children. The serious negotiations on the reform didn't resume until two thousand five. The standardization effort literally stopped for three years because one of the key people was murdered.
That's the part of the story that doesn't make it into the news coverage about the letter Q.
Right. And it explains why the timeline is so long. This isn't just bureaucratic inertia. There's genuine tragedy in the middle of it.
Let's get back to the Q, because I think it's the emotional heart of Daniel's prompt. He says he's quietly delighted by Tiqwa. I understand that. The Q looks like the word is wearing its history on its sleeve. It's the difference between a name that's been sanded smooth for convenience and a name that still shows its grain.
But here's the counterpoint. The Q doesn't preserve how the word was spoken. It preserves how the word was spelled. And for a road sign, the spoken form is what matters. A tourist who sees Tiqwa and asks for directions to Tiqwa will be met with confusion. The person they're asking says Tikva. The Q isn't helping anyone communicate.
Unless the person they're asking is a linguist, in which case they'll get a lecture on emphatic consonants.
And that's the thing. The old system was built by and for people who cared about the letter-level structure of Hebrew. The new system is built for people who need to drive somewhere. Both are legitimate. They're just for different audiences.
There's a two thousand two letter to the editor that Haaretz resurfaced — someone complaining that the Public Works Department was desperately in love with the letter Q, and that someone else likes to repeat the same letter twice for no reason. The examples were Yarqon versus Yarkon, Elyaqim versus Elyakim, Mevaseret Ziyyon versus Zion.
Mevaseret Ziyyon with a double Y and a double Z. That's the letter-fidelity system at its most extreme. Every Hebrew letter gets a Latin equivalent, even when the result looks like a typo to an English speaker.
Ziyyon looks like a sneeze.
The two thousand six reform killed all of that. Vav is V, kuf is K, tsadi is TS. The goal was that a person who doesn't read Hebrew can look at the sign and produce a sound that a Hebrew speaker will recognize. That's the whole test. Not faithfulness to the source. Faithfulness to the listener.
And that's why I think Daniel's quiet delight is worth examining. He's a person who cares about the language, who laments phonological contraction. The Q feels like resistance. But the Q is also a dead end. It's preserving a distinction that no longer exists in the speech of most Hebrew speakers. The contraction already happened. The Q is just a monument to it.
There's a parallel with English spelling reform debates. The people who want to keep the silent letters in knight and through are preserving etymology. The people who want to simplify are preserving communication. Both are real values. They just can't both win.
And Hebrew has the added complication that the etymology is sacred to a lot of people. The letters aren't just letters. They're the letters. Changing how a word is written feels different in Hebrew than it does in English.
That's why the Academy has outsized importance, as Daniel put it. In English, no single body decides how words are spelled. The dictionaries describe usage; they don't legislate it. In Hebrew, the Academy actually does legislate. It's a prescriptive authority in a descriptive world. And when it rules on transliteration, it's ruling on how Hebrew presents itself to people who don't read it.
The Academy has also drawn some boundaries. It's on record that it doesn't consider Hebrew-to-Arabic transliteration its own competence. That's a political boundary as much as a linguistic one. Arabic place names in Israel have their own authority structure, and the Academy has decided not to touch it.
Which connects back to the two thousand nine blowup. When Katz tried to force everything through Hebrew, he was asking the state to override Arabic names. The Academy, to its credit, has stayed out of that particular fight. It rules on Hebrew names. Arabic names are someone else's business.
So where does the new contract leave us? The Academy gets a year to audit every government database, find the inconsistencies, and produce one authoritative list. That's the deliverable. But the enforcement question remains. The Academy can publish the list. Can it make the Transport Ministry repaint the signs?
The signs are ten to fifteen years old, many of them. The Transport Ministry said during the Ginzburg bill hearings that a five-year replacement timeline was unrealistic and proposed at least ten. So even if the Academy produces a perfect list tomorrow, the physical signs won't match it for a decade or more.
And the digital databases could match it much faster. That's the asymmetry. A database entry can be updated in an afternoon. A road sign has to be manufactured, shipped, installed. The physical world moves slower than the digital one.
There's also the exonym problem, which no standard can solve. Jerusalem versus Yerushalayim. The official directory deliberately preserves some historical English names. So the one standard is structurally impossible — you can't have a single rule that gives you both Jerusalem and Yerushalayim. You have to make exceptions.
And every exception is a place where someone will disagree. Why is Jerusalem an exception but not Jaffa? Why is Acre an exception but not Ashkelon? The rule is clear; the exceptions are politics.
I think the honest summary is that Israel has been trying to solve this for almost a century, and each attempt solves part of it and creates a new problem. The nineteen fifty-seven rules solved the letter-fidelity question. The two thousand six reform solved the pronunciation question. The twenty twelve decision solved the Hebrew-Arabic question. And the signs are still inconsistent. So now the state is paying the Academy to do the unglamorous work of auditing what's actually out there and reconciling it.
It's less like a linguistics project and more like a data governance project. The language part was settled years ago. The hard part is the databases.
And there's something almost moving about the fact that the Academy is the one being asked to do it. This is the institution that revived the language, coined its modern vocabulary, and set its standards for seventy years. Now it's being asked to reconcile spreadsheets. That's not a demotion. It's the logical endpoint of being the authority.
The authority ends up doing the filing.
The filing is the authority. That's what enforcement looks like in practice. Not a decree from on high. A list that everyone agrees to use.
Hilbert, you've been quiet back there.
Hilbert: The word is transliteration. You keep saying transliteration. The Academy's own documents say romanization. Different thing.
Tell me the difference.
Hilbert: Transliteration is letter for letter. You take the Hebrew letters and you map each one to a Latin character. That's where the Q comes from. Romanization is broader — it's any way of writing a language in the Roman alphabet, including by sound. The Academy's nineteen fifty-seven document says romanization. The UN group that handles this is the Working Group on Romanization Systems. The Americans and British call their standard the BGN-PCGN romanization system for Hebrew. Nobody who does this for a living says transliteration unless they mean the letter-for-letter kind specifically.
So the whole debate we've been describing — letter fidelity versus sound fidelity — is baked into the terminology. Transliteration is the letter camp. Romanization is the umbrella.
Hilbert: That's right. And the two thousand six reform was a move from transliteration to romanization. They stopped mapping letters and started mapping sounds. The Q disappeared because kuf doesn't sound like Q. It sounds like K.
Daniel's prompt uses transliteration the way most people do — as a catch-all for writing Hebrew in English letters. But the technical term for what the Academy actually does is romanization.
Hilbert: Most people say transliteration. The Academy doesn't. I used to set type for a cartographic publisher in the late seventies. We did maps for the Survey of Israel. The style guide said romanization. We had a whole drawer of Qs that we stopped using after the first reform discussions. Just sat there.
A drawer of Qs.
Hilbert: Lead type. You don't throw it out. You never know when a client wants the old style. We kept them sorted by point size.
The Q as physical inventory. That's the material history of this debate.
Hilbert: The Ws too. And the doubled letters. Mevaseret Ziyyon took up half a line. The compositor hated it. Every time a sign order came in with Ziyyon on it, he'd sigh and reach for the Z drawer. Then the reform came and he threw the extra Zs back in the case and said good riddance.
The two thousand six reform wasn't just a linguistic decision. It changed what was in the type drawers.
Hilbert: It changed how much lead you needed for a sign. That's the part the Academy never had to think about.
Hilbert's right about the terminology, and I think it sharpens Daniel's question. The reason different practices emerged is that different institutions were doing different things — some transliterating, some romanizing — and calling it all the same word. The Academy moved from one to the other in two thousand six. The signs didn't all move with it.
The signs are the lagging indicator. The standard changed. The physical objects didn't. And now the state is paying to reconcile them.
The forward-looking question is whether this contract actually produces the enforcement mechanism that's been missing since nineteen fifty-seven. The Academy has always had the authority. What it's never had is a way to make the databases and the signs and the railway and the municipalities all point to the same list. If this project works, it won't be because the linguistics got better. It'll be because someone finally built the pipeline.
The open question is whether a year and two hundred fifty thousand shekels is enough to audit every government system. That's a lot of databases. The notice says work starts October fourth. Objections close Thursday. So the state is serious enough to set a deadline for objections but not serious enough to have started the work yet.
The objection window is the most Israeli detail in the whole story. You can object to the government's plan to standardize spelling, but only until eleven thirty at night on a Thursday.
I'd love to see what an objection looks like. Dear Prime Minister's Office, I object to the removal of the Q. It has served this nation faithfully.
Somebody out there is writing that objection right now. They've been waiting for this moment since two thousand six.
We should thank Daniel for this one. It's the kind of topic that looks tiny from the outside and opens up into a hundred years of argument about what a name is for.
Thanks to our producer Hilbert Flumingtop for keeping the show running — and for the drawer of Qs, which I will be thinking about for the rest of the day.
This has been My Weird Prompts, the human-AI collaboration podcast.
If you've got a strong opinion about the letter Q, or any other letter, email us at show at my weird prompts dot com.
We'll be back soon.
