🗣️ 19. The lexicon: where the words come from, multiword expressions and register

The vocabulary of Ukrainian is East Slavic at its core, with a learned Church Slavonic layer, a deep Polish layer from the 14th–18th centuries, older Turkic and German loans, a large Russian layer from the 18th–20th centuries, and an English layer that has grown fast since the 1990s. Ukrainian lexicography and editing have a long purist tradition that judges loans, especially Russian ones; usage moves on its own and often faster. This chapter describes the layers (§§1–2), how loans are judged and how usage actually moves, measured in corpora (§§3–5), the vocabulary that the war has brought since 2014 (§6), multiword expressions and idioms (§8), and the registers (§7). §9 says what Mova does with the Ukrainian lexicon.

The counts are of 2026-11-06 over three samples of the Leipzig Corpora Collection (CC BY; 1 million sentences each): ukr_news_2024 (14.2 million words), ukr_newscrawl_2018 (14.8 million words, crawled in 2018, texts partly older) and ukr_wikipedia_2021 (14.6 million words). Word forms were lowercased with apostrophes unified and counted as whole words; a phrase was counted as a whole-word sequence. The UD treebanks used are UD Ukrainian IU 2.18 (counts only) and ParlaMint 2.18 (quotable). Transliteration as in 8. Regional varieties, diaspora and surzhyk.

1. Size and shape of the vocabulary

Dictionaries. The academic Slovnyk ukrajins'koji movy in 11 volumes (SUM-11, 1970–1980) was the standard explanatory dictionary of the Soviet period; its 20-volume successor (SUM-20) has been appearing since 2010. SUM-11 is the dictionary that Kysliuk (2024), Braha (2024) and Taranenko (2019) compare present usage with. The largest machine-readable lexicon is VESUM (Великий електронний словник української мови), with over 430,000 lemmas, from which more than 6.5 million word forms are generated (Starko & Rysin 2025); the release Mova holds has 442,458 lemma lines. VESUM is licensed CC BY-NC-SA, so Mova uses it for tests and counts only.

Frequency. As in every language, a few function words make up a large part of running text. The forty most frequent word forms of the 2024 news sample, in order, are: на, у, в, що, і, з, та, не, до, за, це, для, про, як, а, він, від, також, України, які, але, його, є, року, вони, ми, із, щоб, час, під, може, буде, по, тому, після, ви, я, якщо, ще, зараз. Three features are typically Ukrainian: the euphonic pairs у/в, і/й, з/із/зі split one preposition or conjunction over several forms (4. Connected speech); the conjunctions та and і compete for 'and'; and the genitive року 'of the year' is in the list because dates are said in the genitive (24 лютого 2022 року).

Native core. The core vocabulary (body parts, kinship, nature, basic actions and qualities) is East Slavic and largely shared with Belarusian and Russian, though often with a different word chosen as the main one: говорити 'speak' and розмовляти 'converse' beside Russian говорить, разговаривать; дивитися 'look' (Russian смотреть), гарний 'beautiful', цікавий 'interesting', тиждень 'week', шукати 'look for', дякувати 'thank'. Where Ukrainian and Russian have the same root, they often differ in sound by the regular changes of 9. Older Ukrainian and the old forms alive today §2 (хліб/хлеб, кінь/конь). Many of the words that make Ukrainian look different from Russian are shared with Polish, Belarusian or Czech: дякувати (Polish dziękować), цікавий (ciekawy), мусити (musieć), тиждень (tydzień).

2. The layers of loans

2.1 Church Slavonic

Church Slavonic gave the learned and religious vocabulary from the 10th century on: благо- compounds (благодать 'grace', благословення 'blessing'), глава 'chapter' beside native голова 'head', храм 'temple', владика 'bishop', святиня 'sacred place', пре- and пред- (прекрасний, предмет, предок). It is the most deeply integrated layer and is not felt as foreign (09 §2).

2.2 Polish (and through Polish: Latin, German, Czech)

From the 14th to the 18th century the Ukrainian lands were in the Grand Duchy of Lithuania and the Polish–Lithuanian Commonwealth, and the written prosta mova took a large layer from Polish (09 §3.1). Much of it was lost in the east under Russian rule and kept in Galicia, which was Austrian–Polish until 1918, so that today the layer is both general and, in part, western. Examples of the general layer, all standard: мусити 'must', цікавий 'interesting', кохати 'love', шлюб 'marriage', ґудзик 'button' (Polish guzik), ґанок 'porch' (ganek), брама 'gate', цегла 'brick' (cegła), шухляда 'drawer' (szuflada), фарба 'paint' (farba, from German Farbe), рахунок 'account, bill' (rachunek, from German Rechnung), шлях 'road' (szlak, from German Schlag), власний 'own'. Middle Ukrainian also had барзо 'very', же 'that', тежъ 'also', зась 'and, however', which did not survive into the standard. Many internationalisms came through Polish and kept a Polish shape in the western tradition (клюб, кляса, фльот in the 1928 spelling; 08 §3.2).

2.3 Turkic, Greek, German

• Turkic loans came through centuries of contact with the steppe peoples and the Ottoman and Crimean states: козак 'Cossack', отаман 'chieftain', майдан 'square' (Persian through Turkic), базар 'market', кавун 'watermelon', гарбуз 'pumpkin', тютюн 'tobacco', чабан 'shepherd', баштан 'melon field', торба 'bag'.
• Greek, mostly through Church Slavonic, gave religious words and names: ангел, ікона, церква (ultimately Greek), Олександр, Дмитро, Ярина (the 1928 orthography records Ярина, Дмитро as the forms of names "long borrowed": Pravopys 1928 §64).
• German, directly (through crafts, towns and the Magdeburg law of the cities) and through Polish: дах 'roof' (Dach), ґрунт 'ground, soil' (Grund), майстер 'master craftsman' (Meister), рахунок, фарба, шлях (above), and in western speech many more (фест 'very', from fest).

2.4 Russian

Russian was the state language of the Russian Empire and the dominant language of the USSR, and for two and a half centuries it was the main source of new words: administrative, technical and scientific terms, Soviet political vocabulary, and internationalisms in Russian shape (проект with е, which the 2019 orthography changed to проєкт). A large part of this layer is established and unobjectionable: Starko & Rysin (2025) define the "Russianisms" of their dictionary work as the units "predominantly regarded by Ukrainian linguists, lexicographers, and editors as such that are best replaced with existing Ukrainian equivalents", and exclude the "numerous words borrowed from or through Russian that have become established". The objectionable part is described in §3.

The influence goes both ways. Taranenko (2019) observes that since 1991 the pressure of Russian on Ukrainian has weakened, Ukrainian influence on the Russian spoken in Ukraine has become more visible, and Ukrainian speakers distance themselves more deliberately from Russian models; the asymmetry is also in evaluation, since Ukrainianisms in metropolitan Russian are mostly perceived as low, ironic or hostile, while Russian hypocoristics (Альоша, Діма, Свєта) are widely used in Ukrainian speech.

2.5 English

English has been the main source of new loans since the 1990s, as in most European languages: business and technology (менеджер, менеджмент, офіс, маркетинг, бренд, стартап, онлайн, офлайн, контент), the internet (мем, фейк, хейт, тролінг, булінг), everyday life (шопінг, вікенд, лайфхак), and a stream of slang. The layer and its adaptation are measured in §5.

3. How loans are judged: purism and the norm

Ukrainian normative work has a strong purist tradition. Its main target is Russian influence, which under Soviet rule was also a matter of policy: Starko & Rysin (2025) recall that the Soviet period brought an influx of Russianisms and interference with the structure of the language, and that the dictionaries of the Ukrainianisation of the 1920s were banned; Taranenko (2019) notes that SUM-11 explained some words after Russian models (for містечко, село, козак, стрілець) and gave 'Russian' as the first meaning of руський.

The modern tools of this tradition are usage guides, handbooks and, increasingly, software. VESUM lists over 9,300 single-word Russianisms with their replacements, marked :bad so that the spelling and grammar checker built on it (Pravopysnyk LanguageTool) can flag them (Starko & Rysin 2025). The largest groups are:

(Starko & Rysin 2025; the двох- → дво- replacement was codified in the 2019 orthography.) The GRAC column is the point. These forms are used in edited writing hundreds of thousands of times; a description of Ukrainian has to say both things, that normative work rejects them and that they are frequent.

The same tag :bad is used for misspellings and for some ordinary colloquial forms (шо 'what', нічо 'nothing'), next to слідуючий, співпадати, тормозити, получка (VESUM, counted 2026-11-06). In the release Mova holds, 13,876 of 442,458 lemma lines carry :bad, 4,070 :slang, 358 :vulg, 762 :obsc, 9,526 :arch and 1,465 :alt. A tag of this kind is a normative judgement, useful for an editor, and not a fact about how much the word is used.

In this manual a loan or calque is described by measured usage: how often it occurs, in which texts, and whether that is changing. The next two sections do that for Russian and English.

4. Russianisms: measured

4.1 Calques in edited writing

The table gives the frequency in news and Wikipedia of pairs in which normative guides prefer the second form (Leipzig samples, 2026-11-06; whole-word counts of the forms shown).

Three things stand out. The classic calques (приймати участь, слідуючий, являтися, на протязі) are already rare in edited news (about 1% of their pairs for приймати участь, much less for the others) and were rare in 2018 too. The phrases that guides call bureaucratic Russianisms but that have no single-word rival (в якості, мати місце, на сьогоднішній день) are still in use, falling slowly. And в зв'язку з against у зв'язку з is not a Russianism at all but the euphonic rule of у/в (04); its fall from 8% to 1.6% of the pair shows how strictly the news now applies that rule. (Вірно 'faithfully; correctly' and правильно 'correctly' are partly different words; the figures are given because guides list the pair.)

The UA-GEC corpus of essays corrected by professional editors (Syvokon et al. 2023) gives the editors' side: of 46,248 corrections, 2,397 (5.2%; 4.3 per 1,000 tokens) are labelled calque, "word-for-word translation from other languages"; the authors attribute much of this to the roughly 30% of writers who were not native speakers of Ukrainian.

4.2 Russian in speech

Spoken Ukrainian, and especially the mixed speech described in 08 §5, has far more Russian than edited text. In the news sample, Russian discourse words occur mostly in quoted speech: вообще 4, тіпа 2, спасибі 53 against дякую 876, вибачаюсь 8 against перепрошую 27 (2024). In UD Ukrainian ParlaMint, which transcribes plenary speech of 2003–2022 (with code-switched passages removed by its editors), the affirmative particle да occurs 14 times against так 209 (all uses of так), and спасибі twice against дякую 29 times (counts of 2026-11-06). In surzhyk да is the norm (Hentschel 2023).

4.3 Derussification as a process

Taranenko (2019) describes the period since 1991 in two sub-periods, before and after 2014, and lists signs of counteraction to Russian models: Russian-mediated forms of names give way to direct ones (Kyiv not Kiev, Mykolaiv not Nikolayev, Dmytro not Dmitriy in Latin script); typical surnames in textbook examples change from Іванова, Петрова, Сидорова to Петренко, Сидоренко; and dictionary definitions written after Russian models are revised. Taranenko (2023) records the years after 2022 as a further derussification of the literary norm. The measurement in §4.1 shows the written side of this process: the calques that normative work has targeted longest are close to disappearing from edited text.

5. Anglicisms: measured

5.1 Growth

The older loans (менеджер, офіс, бренд) are stable; the internet and volunteer vocabulary (донат 'donation', донатити 'to donate', онлайн, хейт) grew sharply between 2018 and 2024. Донатити, with the Ukrainian verb suffix -ити and the perfective задонатити, is fully integrated morphologically: an English noun has become a Ukrainian aspect pair in a few years, carried by the fundraising for the army since 2022.

5.2 Slang

Youth slang from English (крінж 'cringe', вайб 'vibe', рофл, краш 'crush', чіл) is common in speech and social media and almost absent from news: крінж 7, вайб 15, краш 3 in the 2024 sample. A news corpus is the wrong place to measure slang; the point here is only that it has not entered edited writing.

5.3 Adaptation: г, х or ґ for English h and g

English h and g are the hardest sounds to adapt, because Ukrainian has [ɦ] (г), [x] (х) and [g] (ґ). The 2019 orthography gives rules; usage varies. Hordiienko (2024) describes the inconsistency in the media: хакер/гакер 'hacker', хакатон/гакатон, лайфхак/лайфгак. The Leipzig counts:

The established loans keep х (хакер, хобі, хокей); a new word can go either way (лайфгак has 38% in 2024). Brand names are mostly left in Latin script: Google 624 against 8 Cyrillic forms, YouTube 468 against Ютуб 14 in 2024.

6. War vocabulary since 2014 and 2022

The war has changed the active vocabulary more than anything since independence. Taranenko (2023) and Kysliuk (2024) describe it from media texts; the Leipzig samples show its scale.

Kinds of change (Taranenko 2023; Kysliuk 2024):

• New meanings of old words (neosemantisms). Kysliuk (2024) compares dictionary meanings with present use and proposes new dictionary senses: приліт (SUM: 'the action of flying in') now 'a hit of missiles, shells or drones on a civilian or military object'; пташка 'little bird' now colloquially 'any military aircraft or drone'; сушка 'drying; a small dry bagel' now 'a Su fighter jet'; тушка 'small carcass' now 'a Tu-95 bomber'; вата 'cotton wool' now 'people with a (post-)Soviet mentality' (from 2014). She classes бульбашка 'bubble', град 'hail' (the Grad rocket launcher), мопед 'moped' (the Shahed drone), приліт as normative and neutral; бавовна, нуль 'the zero line, the front', пташка, сушка, тушка as colloquial; and вата as pragmatically marked.
• Military terms and slang in general use: БПЛА/дрон, арта 'artillery', бат 'battalion', двохсотий 'killed' and трьохсотий 'wounded' (from the Soviet–Afghan war), the ironic verbs мінусувати 'to minus = to kill' and утилізувати 'to dispose of' (Taranenko 2023).
• Names for the enemy and for Russia: рашисти, орки, the -стан formations (Рашистан, Оркостан), русня; and the construction на Росії 'in Russia' modelled on the Russian на Украине, used by media since 2014 (Taranenko 2023). Taranenko also records a graphic change: the media began writing some proper names connected with Russia in lower case.
• Official names changed with the phases of the war: АТО (anti-terrorist operation, April 2014–April 2018), then ООС (joint forces operation); держава-агресор, держава-терорист (Taranenko 2023).
• Renaming of streets and places (decommunisation since 2015, derussification since 2022).

7. Registers as deltas

Each register below is described as a delta over the literary standard (the convention of 8. Regional varieties, diaspora and surzhyk §1.2): which rules of orthography, pronunciation, morphology and syntax change and which words are added or replaced. They are bands of one continuum, and a speaker moves between them; chapter 37 describes register in discourse. The surzhyk delta is in 08 §5.3, the song delta in 08 §6, the archaic and church delta in 09 §6.

7.1 Permitted colloquial

Colloquial forms of Ukrainian origin are normal in speech at every level of education and appear in fiction and dialogue; the standard is the written and formal register, not the only Ukrainian. Normative tools sometimes tag such forms with the same label as Russianisms (VESUM :bad on шо, §3); this manual keeps them apart. Surzhyk is a separate delta, not the bottom of the colloquial one: it is defined by its source (Russian words and forms in a Ukrainian frame) and its lexical regularities, not by informality.

7.2 Slang and rap

Slang is a news-poor register: a news corpus records only what reaches edited writing. Rap, like song (08 §6), has an alternative grammar of its own; this manual describes it by the delta categories but holds no corpus to measure it.

7.3 Officialese (канцелярит)

7.4 Other register markers

Church language adds Church Slavonic words and old forms (єси, отче, во віки віків; 09 §6.7–6.8).

Address and politeness are a register marker of their own: the vocative (пане Голово, шановні колеги), the polite Ви, and the choice between nominative and vocative for name and patronymic (09 §6.1). In the 2024 news sample ви (28,088) is about seven times as frequent as ти (3,658), much of it in direct address to readers.

Feminine personal nouns are a register question that usage has settled fast. The 2019 orthography describes the suffixes that form them (-к-: авторка, дизайнерка, директорка, редакторка; -иц-(я): порадниця; -ин-(я): майстриня, філологиня) as ordinary word formation (Pravopys 2019 §32.4). In the news samples:

Forms that were rare or stylistically marked in 2018 have become the neutral way to name a woman in a profession in edited news by 2024. The codified norm and usage moved together here, the norm a step ahead. Word formation is in 17. Derivation: suffixes.

Off-dictionary vocabulary of the classics. Tsymbaliuk-Skopnenko (2026) points out the opposite gap: words and phrases from canonical works of Nechui-Levytskyi, Lesia Ukrainka, Kobylianska, Kotsiubynskyi and Vynnychenko that the academic dictionaries do not record or do not label, so that dictionary-based tools treat the core of the literary canon as off-norm. Her argument is that codification depends above all on the frequency of elements in the canonical texts, and that the picture of the literary language is distorted if the idiolects of these authors are left out.

8. Multiword expressions

8.1 What counts

Ukrainian phraseology traditionally takes a narrow view: only idioms proper (phraseologisms) are its object, while proverbs and sayings, clichés, etiquette formulas and other fixed expressions are collected separately or not at all (Sytar et al. 2026; Tsymbaliuk-Skopnenko 2022). Ukrainian grammar describes other kinds under its own names: light-verb constructions as periphrastic predicates, adpositional idioms as "secondary compound prepositions". The PARSEME framework, adopted for the first annotated Ukrainian corpus of multiword expressions, brings them together.

8.2 The Ukrainian MWE corpus

Sytar, Shvedova & Kanishcheva (2026) annotated 12,078 sentences (198,555 tokens) of interviews from Ukrajins'kyj tyžden' (2013–2016, from GRAC) with 5,993 multiword expressions:

Inter-annotator agreement on a 20-file subset was an MWE-based F-score of 54, highest for adpositional (0.845) and adverbial (0.754) idioms, lowest for light-verb and inherently adpositional verb constructions. The corpus also contains prepositional units not described in grammars or dictionaries before, such as на момент свого відкриття 'at the time of its opening' and коливатися в коридорі 60–70% 'fluctuate between 60 and 70%'.

8.3 Idioms

Idioms are many and vivid, and much of the stock is shared with Belarusian, Polish and Russian; some are specifically Ukrainian (скакати в гречку; дати гарбуза 'refuse a suitor', lit. 'give a pumpkin'; пекти раків 'blush', lit. 'bake crayfish'; теревені правити 'chatter'). Their behaviour in text is what makes them hard for a machine: they inflect (дала гарбуза, дадуть гарбуза), allow material inside (скакав, кажуть, у гречку) and change word order.

Petruniv, Makogon & Kyslyi (2026) built SimIdioms, a corpus of aligned Ukrainian–English idiom pairs from phraseological dictionaries (among them Bilonozhenko's), with figurative meanings, example sentences from the UberText fiction corpus and semantic-transparency scores, and evaluated six large language models on 65,723 translations. All models were substantially worse translating into Ukrainian than out of it; naming the source idiom in the prompt helped from Ukrainian into English but hardly in the other direction, which the authors read as a bottleneck of morphological generation rather than of idiom recognition; transparency was only weakly correlated with quality.

Proverbs and sayings (прислів'я, приказки) form a large separate stock, collected since the 19th century (Nomys's collection of 1864 is the classic one). They are quoted in speech in shortened form (Не кажи гоп… for Не кажи «гоп», поки не перескочиш 'don't say "hop" until you have jumped over') and are a register of their own in folk and political speech.

8.4 Collocations

Collocations, the habitual combinations that are not idioms (брати участь 'take part', ухвалити рішення 'adopt a decision', здобути освіту 'get an education'), are where calques show most (§4.1: приймати участь), and where Ukrainian dictionaries of combinability are the editor's tool (Dydyk-Meush 2019; Trub 2021, cited here from their cards). UA-GEC counts unnatural collocations separately from calques: 459 of its 46,248 corrections (1.0%; Syvokon et al. 2023).

9. How Mova handles the Ukrainian lexicon

Mova has no Ukrainian lexicon of its own: no list of loans by origin, no register labels, no multiword-expression lexicon and no idiom list for Ukrainian. The English machinery of the English manual's chapter 19 (global::expressions_in_sentence, the register lexicons global/data/register-*.tsv, the pupil dictionary) is English only.

• The delta principle, and the missing layer. A register (colloquial, slang and rap, officialese, song, church) is the standard model plus a lexicon and grammar delta (§7). Mova has no delta layer yet. The nearest hook is VESUM's style tags (:slang, :bad, :arch, :alt, :vulg, :obsc) and its replacement lists, which already pair many delta words with standard ones; kg does not read them (next point).
• VESUM through kg. The kg crate reads VESUM (kg/src/morph.rs) for analysis and synthesis of word forms. Lemma keys are lemma|part_of_speech (word:uk heap). The style tags of §3 (:bad, :slang, :arch, :alt, :vulg, :obsc) and VESUM's suggested replacements are not read: Gram::from_vesum skips unknown tags (kg/src/gram.rs). A Russianism or slang word that VESUM lists is parsed like any other; one it does not list stops the parse ("word not in VESUM"). VESUM's licence (CC BY-NC-SA) keeps this a research and test path.
• Government. kg does read VESUM's government tags (rv_*: the cases a verb or preposition governs), so collocations of the type verb + case (впливати на + accusative) are partly encoded; light-verb constructions and idioms are not.
• Wiktionary. The raw Ukrainian Wiktionary dump (2026-09-01) is held, and tools/wikt-count has a uk mode that counts its pages, definitions and etymology sections (Data, libraries and licences). It is CC BY-SA: for learning and tests only. No Ukrainian lexicon, etymology table or idiom list is built from it.
• The English parser on UD Ukrainian (LAS 73.43 / 78.17 own / gold tags on IU dev, 2026-11-06) learns the UD relations fixed, flat and compound where the treebank has them; it has no lexicon and does not know that скакати в гречку is one unit.
• What would be needed. A Ukrainian lexicon with an open licence (from Wiktionary, CC BY-SA, or Wikidata lexemes, CC0) carrying origin, register and variant labels; the replacement pairs of §3 kept as labels with measured frequencies rather than as corrections; an MWE lexicon trained on the Ukrainian PARSEME corpus (Sytar et al. 2026); and the measurements of §§4–6 rerun on each new corpus sample so that the description stays dated. None of this is built.

Sources for this chapter

• Starko, Vasyl & Andriy Rysin. 2025. An electronic Ukrainian dictionary as a derussification tool. eLex 2025, pp. 824–838. CC BY-SA 4.0 (starko-2025-dictionary-derussification); paraphrased.
• Taranenko, Oleksandr. 2019. Novi tendenciji v poširenni rosijs'kyx elementiv. Movoznavstvo 2019(2–3). CC BY-SA 4.0 (taranenko-2019-rosiiski-elementy); paraphrased.
• Taranenko, Oleksandr. 2023. Rosijs'ko-ukrajins'ka vijna i ukrajins'ka mova. Movoznavstvo 2023(3–4). CC BY-SA 4.0 (taranenko-2023-viina-i-mova); paraphrased.
• Kysliuk, Larysa. 2024. Nova leksyka vojennoho času ta jiji leksykohrafuvannja. Leksykohrafičnyj bjuleten' 33: 33–48. CC BY 4.0 (kysliuk-2024-voienna-leksyka).
• Tsymbaliuk-Skopnenko, Tetiana. 2026. Problema opysu pozaslovnykovoji leksyky j frazemiky ukrajins'koji xudožn'oji klasyčnoji literatury. Movoznavstvo 2026(1). CC BY-SA 4.0 (tsymbaliuk-2026-pozaslovnykova-leksyka); paraphrased.
• Tsymbaliuk-Skopnenko, Tetiana. 2022. Frazeohrafuvannja v akademičnomu tlumačnomu slovnykarstvi. Leksykohrafičnyj bjuleten' 31: 93–101. CC BY 4.0 (tsymbaliuk-2022-frazeohrafuvannia).
• Sytar, Hanna, Maria Shvedova & Olha Kanishcheva. 2026. Ukrainian multiword expressions corpus: creation, annotation, and linguistic analysis. MWE 2026. CC BY 4.0 (sytar-2026-mwe-corpus).
• Petruniv, Yaryna, Iuliia Makogon & Roman Kyslyi. 2026. SimIdioms: a corpus and benchmark for Ukrainian idiom translation. UNLP 2026. CC BY 4.0 (petruniv-2026-simidioms-corpus-benchmark-ukrainian-idiom).
• Syvokon, Oleksiy, Olena Nahorna, Pavlo Kuchmiichuk & Nastasiia Osidach. 2023. UA-GEC: grammatical error correction and fluency corpus for the Ukrainian language. UNLP 2023. CC BY 4.0 (syvokon-2023-ua-gec).
• Hentschel, Gerd. 2023. Ukrainian and Russian in the lexicon of Ukrainian Suržyk. Russian Linguistics 48(1). CC BY 4.0 (hentschel-2023-surzhyk-lexicon).
• Braha, I. I. 2024. Sociolinhval'ni javyšča v ukrajins'kyx slovnykax. CC BY 4.0 (braha-2024-sotsiolinhvalni-slovnyky).
• Hordiienko, O. Je. 2024. Fonetyko-hrafične osvojennja ta pravopys novitnix zapozyčen'. Ukrajins'ka mova 2024(4): 97–108 (hordiienko-2024-novitni-zapozychennia). Library card CC BY 4.0; the article prints CC BY-NC-ND 4.0, so paraphrased only.
• Cards only, no claim rests on their text here: Dydyk-Meush 2019 (dydyk-2019-meush-kombinatornyi-slovnyk), Trub 2021 (trub-2021-slovnyk-spoluchuvanosti).
• Ukrajins'kyj pravopys 2019 §32.4 and 1928 §64. Official documents (uk-gram-pravopys-2019, uk-gram-pravopys-1928).
• Goldhahn, Dirk, Thomas Eckart & Uwe Quasthoff. 2012. Building large monolingual dictionaries at the Leipzig Corpora Collection. LREC 2012. Corpora ukr_news_2024, ukr_newscrawl_2018, ukr_wikipedia_2021 (CC BY), uk.
• VESUM (dict_corp_vis.txt.bz2, CC BY-NC-SA 4.0), counts only, uk-vesum-dict-uk.
• UD Ukrainian IU 2.18 (CC BY-NC-SA 4.0, counts only), UD Ukrainian ParlaMint 2.18 (CC BY-SA 4.0), ud-treebanks-v2.18.
• Mova: kg/src/morph.rs, kg/src/gram.rs, tools/wikt-count/src/main.rs; Data, libraries and licences.

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/languages/uk/19-lexicon-mwe-register/