Hungarian Crossword Puzzles

These are fully checked grids in the style of the New York Times: every letter belongs to both a vízszintes (across) and a függőleges (down) answer, a number sits in the first square of each answer, and the clues come in two lists. Hungarians know the numbered grid well — Füles has printed one every week since 1957 — but the Hungarian newspaper puzzle usually carries a megfejtés, a quotation or punchline spelled out in highlighted rows that the solver sends in. These grids have no megfejtés: the answers are the whole puzzle.

Hungarian is the fifth language built here from nothing in a day. The shelf is two 15×15s, two 13×13s and six 11×11s — 540 entries, 540 different words — and six of the ten filled from the most frequent 15% of the vocabulary, two from the top 30% and two from the top 45%. No answer appears in two puzzles, which Greek, Bulgarian and Finnish could not manage: all three had to let three-letter answers recur across their shelves. Hungarian did not, because its short words are real words rather than abbreviations — VÉR (blood), TŐR (dagger), ÍNY (gum), EKE (plough) and ÁSZ (ace) are the three-letter answers a Hungarian setter reaches for, and 151 different ones appear across the ten puzzles.

Every answer is a dictionary form or a word a Hungarian setter would clue as it stands. Hungarian piles suffixes onto a stem — 22,857 entries in the word list end in the inessive -BAN or -BEN alone — but the clued vocabulary these grids fill from was built from dictionary headwords, so a noun is asked for in the nominative (HÁZ, not HÁZBAN) and a verb in the third-person singular present, the form a Hungarian dictionary lists (OLVAS, not OLVASTAM).

The grid follows the rule every Hungarian puzzle prints: long and short vowels are not told apart. Á and A share a square, as do É and E, Í and I, Ó and O, Ú and U — and Ő and Ö, Ű and Ü. What does stay apart is O and Ö, U and Ü, which are different letters and different words: ÖR is a guard and OR a nose, TÖR breaks and TOR is a feast. Ö appears in 113,040 entries of the word list and Ü in 61,308, so the grid alphabet is the Latin letters plus those two.

The fold makes Hungarian a punning language on the grid. 988 spellings in the Hungarian dictionary stand for two or more different words once the accents are set aside — VÁR (castle) and VAR (scab), VÉR (blood) and VER (beats), ÁGY (bed) and AGY (brain), TEVE (camel) and TÉVÉ (television) — so the crossing letters say which one you have, and the clue says which one was meant. The digraphs CS, GY, LY, NY, SZ, TY, ZS, DZ and DZS are single letters of the Hungarian alphabet but take one square per character, as in every Hungarian crossword: SZÉK is four squares.

After the fold every letter sits in the one-byte Latin encoding the .puz format uses, so Hungarian puzzles export to .puz and open in any standard crossword app, as well as printing to PDF.

The Hungarian word list starts from Magyar Ispell, the open Hungarian spelling dictionary behind LibreOffice's spell-checker, and adds every form that occurs at least four times in Hungarian film subtitles, ranked by how often each form appears across the subtitles, a 300,000-sentence Hungarian news corpus and a million sentences of Hungarian Wikipedia. Old subtitle files often spell Ő and Ű as Õ and Û, a relic of a code page that had no room for them; the word list reads those as the letters they mean. That gives 808,177 entries, thickest at nine letters (107,537) and still 21,531 at fifteen; auto-fill takes the most frequent words first and works down the list only when nothing common fits.

There is no Hungarian Wiktionary in the open dictionary exports this site's other clue corpora come from, so every Hungarian clue started from the English Wiktionary's Hungarian entries: 22,149 clues written from scratch, in Hungarian, for words Wiktionary defines in English, using the definitions as a guide rather than translating them. The clues in the ten puzzles themselves were written last, one puzzle at a time, with the whole grid in view.