This guide adapts rules and examples from Microsoft’s 55-page Estonian Localization Style Guide (originally written for software/UI localization). The underlying linguistic rules apply universally — to legal contracts, medical documents, marketing copy, and any Estonian translation work. Restructured and reformatted as a general Estonian translator reference by ChatsControl.

Estonian Translation Style Guide — Voice, Word Choice & Common Pitfalls (Legal, Medical, Marketing, IT)¶

TL;DR¶

• Estonian translation across all spheres (legal, medical, marketing, IT) requires a clear, friendly, concise register — formal bureaucratic Estonian carries an alien, machine-like feel in modern consumer-facing content.
• Flexibility means rewriting source rather than literal translation — applies to marketing copy, medical instructions, software UI; less to legal contracts where literal accuracy is paramount.
• Avoid overly formal Latinate verbs (modifitseerima → muutma; sünkroniseerima → sünkroonima; navigeerima → liikuma); use everyday Estonian forms.
• Decline product names (Wordi, Exceli, Office’isse) per Estonian grammar — refusing to decline produces unnatural translatorese; use modifiers (nupp, käsk) only when declension fails.
• Address users with lowercase teie; use gender-neutral forms (lapsevanem, kolleegid, võistkond) and people-first disability language (puudega inimene).

Register and tone for modern Estonian translation¶

Register is the level of formality, warmth, and conversational ease the target text projects. Modern Estonian readers across consumer-facing spheres expect language that feels like everyday conversation rather than technical or commercial formality.

Three principles define the modern Estonian register for consumer-facing content:

• Warm and relaxed. Sounds like honest conversation, not a formal notice. Less formal, more grounded — matching how Estonians actually speak.
• Crisp and clear. Written for scanning first, reading second. Simplicity is the default; precise, well-chosen words aid clarity.
• Ready to help. Anticipates what the reader needs and offers it at the right moment, rather than burying it under qualifications.

Why this matters: Bureaucratic register damages outcomes across spheres. In marketing copy it kills conversion — readers bounce when text sounds like a tax form. In patient-facing medical materials it reduces comprehension and compliance. In software UI it creates friction at every interaction. In consumer-facing legal documents (terms of service, privacy notices) regulators increasingly demand plain language. Only sworn legal translation and pure technical specifications retain the older formal register.

Audience targeting: technical vs. consumer vocabulary¶

The same source text requires different vocabulary depending on who reads the translation. Use technical terms for technical audiences; for consumers use common words. A clinical drug monograph for prescribers uses precise pharmacological terminology; the patient leaflet for the same drug uses everyday Estonian. A software API reference uses developer jargon; the end-user help article uses plain Estonian.

This applies in every sphere. Legal translation for corporate counsel uses Latinisms and procedural shorthand; consumer-facing versions need plain-Estonian framing. Medical translation for clinicians keeps Greek/Latin nomenclature; for patients it switches to common terms. IT translation uses developer jargon in engineer-facing docs, natural Estonian in end-user help.

Flexibility: when to translate literally vs. when to rewrite¶

Flexibility is the translator’s discretion to depart from literal source structure when the literal rendering produces unnatural Estonian. The rule: understand the whole intention of the sentence, paragraph, or page, then rewrite as if composing it yourself for an Estonian reader. This frequently means restructuring sentences, splitting them, omitting redundant content, or shifting from English-style subordinated structures to Estonian-style direct phrasing.

The Estonian versions reshape syntax rather than substituting words into the English skeleton. The first inverts the actor-object relationship — the icon shows the user which emails are safe, rather than the user looking for the icon. The third drops “experience” entirely, replacing it with kasutage (“use”) plus the qualified noun teenust — Estonian readers don’t need the abstract noun “experience” to understand the offer.

Why this matters: Source-faithful translation produces translatorese — text that reads as translated. Required in sworn legal translation and certified document translation (birth certificates, court rulings, disputed contracts) where literal accuracy is mandated. Harmful in marketing translation (lost conversion), patient-facing healthcare materials (lost clarity), and software UX (lost engagement). Knowing where the boundary sits is core translator judgment — the billable skill that distinguishes professional work from raw machine output.

Word choice: approved terminology and conversational vocabulary¶

Approved terminology is the project-specific bank of fixed translations for key terms, product names, technical concepts, and recurring phrases. Every serious translation project has one, explicit (glossary, termbase) or implicit (translator’s accumulated decisions). Consistency within the bank matters more than the individual choice — switching between teave and informatsioon, or muutma and modifitseerima, in one document signals carelessness.

Short, everyday words are preferred over long formal ones wherever both exist and the audience is non-specialist. Estonian shares English’s preference for short, familiar words in conversational content — but the avoidance pattern is different (Estonian avoids Latinate verbs with -eeri- suffixes rather than English’s Latinate roots).

For the word “app”, choose context-appropriate Estonian: rakendus (full application) or minirakendus (mini-app/widget). Don’t default-translate.

Why this matters: Terminology consistency is non-negotiable in legal translation (a defined term in a contract must render identically across all 200 pages — variant renderings create ambiguity opposing counsel will exploit), medical translation (drug names, dosage units, anatomical terms must be invariant — a synonym swap can produce a dispensing error), and IT/software translation (UI labels, menu items, error codes must match help documentation word-for-word or users can’t find what they need). Choosing short vs. formal forms is a separate decision from consistency — but once chosen, apply uniformly.

Word-for-word translation: why direct mapping fails¶

Word-for-word translation substitutes each source word with its dictionary equivalent while preserving source word order and syntax. In Estonian translation from English it produces stiff, unnatural — sometimes ridiculous — text, because Estonian and English distribute information differently. Estonian relies on rich morphology (14 cases, productive compounding, verb aspect); English leans on word order and prepositions. Mapping one onto the other without restructuring produces the recognizable translated-from-English feel.

The remedy: read for paragraph-level meaning, then compose Estonian sentences that carry the same meaning naturally. Split sentences or omit descriptors to make the text snappier. Pay particular attention to the word “experience” — in most cases it should not be translated literally as kogemus into Estonian.

The last example is instructive: the “correct” version splits one English sentence into two Estonian sentences, replaces the abstract “experience” with the concrete seminar (“seminar”), and converts the post-modifier (“led by…”) into a separate clause. None of those moves are dictionary lookups — they require thinking about how an Estonian writer would naturally frame the content.

Why this matters: Word-for-word translation is the dominant failure mode of inexperienced translators and unedited machine output. In legal contracts it produces clauses that translate every term but obscure who owes what, creating dispute risk. In medical instructions it separates action from actor in ways that confuse patients — a known source of compliance errors. In marketing copy it produces headlines that read as foreign — technically Estonian but emotionally flat. In software UI it produces labels users hesitate over because the phrasing doesn’t match how they’d describe the action.

Words and phrases to avoid in modern Estonian¶

Why this matters: These overly formal forms appear in legal templates and government forms out of institutional habit but feel alien in modern consumer products, patient-facing medical materials, brand-led marketing, and user-friendly software. A privacy policy reading “Süsteem sünkroniseerib teie andmed” sounds like a Soviet-era technical manual; “Süsteem sünkroonib teie andmed” reads as the product talking to its user. A patient leaflet saying “navigeerima rakendusse” lands differently than “liikuma rakenduses” — the second matches how Estonians actually speak. These substitutions are among the highest-leverage edits a translator can make.

Inclusive language¶

Gender-neutral vocabulary¶

Estonian doesn’t have grammatical gender, but many compound nouns include gendered roots (mees “man”, naine “woman”). Use gender-neutral alternatives wherever they exist.

When making generalizations, use gender-neutral plural noun forms (inimesed, isikud, õppurid).

Handling English singular they¶

English commonly uses they/their/them for generic single-person reference. Decide each instance based on meaning: if it refers to one person, translate as singular tema/ta; if it refers to several, plural nemad/nad. Don’t default to plural.

Accessibility and people-first language¶

Focus on people, not disabilities. Don’t use words that imply pity (kannatama, vaevama). Don’t mention a disability unless it’s relevant.

Keep paragraphs short and sentence structure simple — aim for one verb per sentence. Read text aloud and imagine it spoken by a screen reader. Spell out words like ja, pluss, and umbes — screen readers can misread special characters (&, +, ~).

Why this matters: Inclusive vocabulary is increasingly contractually required in enterprise translation (Microsoft, EU institutions, multinational HR materials), public-sector medical translation, and government-facing legal translation. Estonian’s compound-formation flexibility makes gender-neutral alternatives readily available — there’s no excuse for defaulting to gendered forms in modern content.

Grammar and orthography¶

Abbreviations¶

Avoid using too many abbreviations and don’t use unfamiliar abbreviations — they make text harder to comprehend. If a shorter string is needed, prefer shorter synonyms or rephrase. Don’t abbreviate words that are already short (mai, juuni, juuli, märts).

Common Estonian abbreviations: a (aasta), AS (aktsiaselts), dets (detsember), dr (doktor), e (ehk), eKr (enne Kristuse sündi), e.m.a (enne meie ajaarvamist), hr (härra), j.a (juures asuv), jne (ja nii edasi), jt (ja teised), k.a (käesoleval aastal; kaasa arvatud), kd (köide), lk (lehekülg), lp (lugupeetud), mag (magister), m.a.j (meie ajaarvamise järgi), min (minut), mln (miljon), mld (miljard), mnt (maantee), nn (niinimetatud), nr (number), nt (näiteks), OÜ (osaühing), p (punkt), pKr (pärast Kristuse sündi), prof (professor), ptk (peatükk), s.a (sel aastal), s.o (see on), st (see tähendab), tel (telefon), tk (tükk), Tln (Tallinn), tn (tänav), Trt (Tartu), u (umbes), v.a (välja arvatud), vt (vaata).

Common technical abbreviations:

For more detail, see Lühendiraamat by Martin Ollisaar (Valgus, 2006).

Acronyms¶

Localized acronyms get an Estonian equivalent when widely accepted (GDP → SKP, U.A.E → AÜE, U.S.A → USA). International acronyms widely used in their English form stay (USB, LAN-ühendus).

If a US English acronym isn’t widely used and isn’t explained in the original, localize the full term and leave the original acronym in brackets:

For unfamiliar English acronyms, prefer the full localized term over the acronym throughout.

Capitalization¶

Estonian uses capitals far less than English does. Don’t follow English capitalization patterns in Estonian translation.

1. Sentences. First letter capitalized. After a colon, the next sentence generally continues lowercase.
2. Proper nouns. Names of programs, modules, wizards, and similar elements are capitalized. For phrasal names, capitalize only the first word: Salvesta nimega (not Salvesta Nimega). Trademarks like Microsoft Dynamics keep every word capitalized.
3. Text in brackets. If the bracketed text isn’t a complete sentence, start with a lowercase letter.

Capitalize: toponyms (countries, regions, states), companies and organizations, names of people, media editions and documents, literary and art works, historical events, events, goods, plant species, rewards.

Do NOT capitalize: months, weekdays, holiday names, job titles, titles of honor, ranks, scientific degrees, designations and appellations, derivatives of names (shakespearelik väljend), the sentence after a colon or dash, list items (with exceptions), language names.

Compounds¶

Estonian forms compounds productively by combining stems. Keep them understandable — overly long compounds diminish usability. Note that compound vs. non-compound changes meaning: lapse põlv (“child’s knee”) vs. lapsepõlv (“childhood”).

Contractions¶

Estonian contractions apply only to the verb olema in negative sentences:

Genitive and product name declension¶

English forbids attaching genitive -s to product names (Microsoft’s would imply modification). Estonian rules go the other way: adding case endings to product names is consistent with grammar and sounds natural. Estonians don’t interpret declension as modification — all names decline.

When a name can’t be declined naturally, use a modifier and decline the modifier instead. But excessive modifier use sounds unnatural — decline names whenever possible.

Estonian’s 14 cases all apply to names: nominative (Microsoft), genitive (Microsofti), partitive (Microsofti), allative (Microsoftile), adessive (Microsoftil), ablative (Microsoftilt), illative (Microsofti), inessive (Microsoftis), elative (Microsoftist), translative (Microsoftiks), terminative (Microsoftini), essive (Microsoftina), abessive (Microsoftita), comitative (Microsoftiga).

Modifiers (descriptors before UI element names)¶

When a variable tag or English-form product name can’t take a clean case ending, use a modifier (nupp “button”, käsk “command”, suvand “option”) and decline that. The modifier goes BEFORE the name, not after. Only the modifier is declined — the UI element name appears exactly as it shows in the interface.

Modifier use is required with the verb klõpsake (“click”) and optional with valige (“select”).

Common descriptors by software type:

For long lists of UI elements after click, prefer valima (nominative case) over klõpsama (partitive case) — eliminates the need for modifiers. This applies only in imperative mood (valige); other forms (valides, valisite) require genitive.

Note: suvand has the specific meaning of “optional parameter or setting” and can’t be used as a modifier for buttons. Use nupp.

Nouns and inflection¶

Estonian nouns inflect for 14 cases and two numbers. Plural partitive endings (-id, -sid, -e, -i, -u) depend on stem shape — consult Eesti keele käsiraamat for the full rule set. Short plural stems exist for several cases (plural illative, inessive, elative, allative, adessive, ablative, translative) for many nouns.

Numbers¶

The main rule: numbers 1–10 written as words, numbers larger than 10 as numerals. If a sentence mixes numbers above and below 10, use numerals for all.

Prepositions¶

Estonian uses few prepositions (enne, keset, piki, tänu, ilma); most relations are expressed via case endings on nouns. Don’t drop prepositions out of English-influence habit, but don’t add them where Estonian uses a case.

Pronouns¶

Address the user with second-person plural teie (Teams and Skype use sina). When you do use the pronoun, make it lowercase. Often the pronoun isn’t needed at all — Estonian’s verb morphology already encodes person.

Avoid using me (“we”) when referring to the computer or system, even if the English uses “we”. Use passive voice or impersonal phrasing instead.

Why this matters: Verb-only constructions (no pronoun) are denser, more idiomatic Estonian — apply them across legal contracts (clearer attribution: sõlmite lepingu vs. te sõlmite lepingu), medical instructions (cleaner imperatives: võtke ravimit vs. te võtke ravimit), and software UI (less screen clutter). Adding pronouns out of English habit is the most common translatorese signal.

Punctuation¶

Key rules:

• No punctuation at the end of menu and command names.
• Check boxes and radio buttons don’t end with a period unless they’re complete sentences (even if the English source has a period).
• Headings of list boxes and tables never end with a period.
• Column and row headings start capitalized, no end punctuation.
• Error messages always end with a period or question mark (even if the English source doesn’t). Exclamation points and other punctuation only in exceptional cases.

Decimal separator¶

The decimal value is separated from the integer with a comma: 5,6 km (not 5.6).

Colon¶

Use the colon to introduce a list of single words or phrases. List items start lowercase, separated by comma or semicolon; the last item ends with a period.

RTF-vormingus faili loomiseks tuleb teha järgmist:
• avada tekstiredaktoris (näiteks Microsoft Wordis) soovitud dokument (või tuleb see ise koostada),
• valida Fail > Salvesta nimega.

When a bulleted list contains full sentences, don’t use a colon — end the introducing phrase with a period:

RTF-vormingus faili loomiseks tehke järgmist.
• Avage tekstiredaktoris (näiteks Microsoft Wordis) soovitud dokument (või koostage see ise).
• Valige Fail > Salvesta nimega.

Dashes¶

Estonian uses only the hyphen and en dash — never the em dash. The em dash is purely an English convention; in Estonian, replace it with an en dash with spaces.

Hyphen. Word break or one-word compound formation:

En dash. Minus sign, ranges, and parenthetical/emphasizing phrases:

Ellipsis¶

Although Estonian grammar allows ellipsis with or without preceding space, in UI translations use it without preceding space — character limits and avoidance of ellipsis appearing alone on a line.

Be consistent throughout the product — pick spaced or unspaced ellipsis and stick with it.

Period¶

Sentences end with a period; next sentence starts capitalized. A caption isn’t followed by a period if it’s one word or a phrase (Sissejuhatus); if it’s a complete sentence or directly precedes another caption, use the period.

Error messages (except questions) always end with a period. Don’t use a period after abbreviations unless the abbreviation could be misread as another word.

Apostrophe¶

The preferred type is the curly apostrophe (‘). For UI localization, the type depends on the CAT tool — use whichever the apostrophe key produces by default rather than requiring a key combination.

Quotation marks¶

Quotation marks are used: when referring to a phrase as it appears in the UI; for product/service names that should be separated from the rest of the sentence for clarity; for titles of help topics, articles, book chapters; in place of bold when the UI doesn’t allow bold; when describing a word rather than referring to its meaning.

Estonian allows several quotation mark types, including curly „ ” (Alt+0132, U+201E and Alt+0147, U+201C) and keyboard ” ” (Shift+2). Use one type consistently throughout a product.

If UI element references aren’t highlighted, rephrase or use quotation marks:

Parentheses¶

No space between the parentheses and the text inside them:

Sentence fragments¶

Estonian’s morphology limits sentence fragment use compared to English, but use them where possible — they’re a useful way to get straight to the point.

Symbols and nonbreaking spaces¶

Replace the ampersand (&) with the word ja. Use & only when part of a (trademarked) name.

Use nonbreaking space to prevent line breaks in numeric strings:

Error messages¶

Error messages inform the user of an error that needs attention. Apply the same voice principles — clear, polite, addressing the user directly, never robotic.

Standard error phrases¶

Placeholders in error messages¶

When localizing error messages with placeholders, identify what will replace each placeholder to ensure grammatical correctness. Placeholder letters convey meaning:

• %d, %ld, %u, %lu →
• %c →
• %s →

Examples: “Checking Web %1!d! of %2!d!” → “Checking Web of .” | “INI file “%1!-.200s!” section” → “INI file “” section.”

Why this matters: Standardized error-message phrasing matters across medical software (clinicians need predictable error patterns to react correctly to safety-critical conditions), legal document processing tools (consistent error vocabulary supports audit trails), and enterprise IT (helpdesk troubleshooting depends on users describing the same error consistently).

Keys and shortcuts¶

Localized key names¶

Many keys have established Estonian names:

• Alt — muuteklahv
• Backspace — tagasilükkeklahv
• Caps lock — suurtähelukk
• Ctrl — juhtklahv
• Delete — kustutusklahv
• Down arrow — allanool
• Enter — sisestusklahv
• Esc — paoklahv
• Insert — lisamisklahv
• Left arrow — vasaknool
• Num lock — numbrilukk
• Right arrow — paremnool
• Shift — tõstuklahv
• Spacebar — tühikuklahv
• Tab — tabeldusklahv
• Up arrow — ülesnool
• Windows key — Windowsi klahv
• Menu key — menüüklahv

Key names generally left in English¶

Used with the descriptor klahv: Print Screen → klahv Print Screen, Scroll Lock → klahv Scroll Lock, Pause → klahv Pause, Page Down → klahv Page Down, Page Up → klahv Page Up, Home → klahv Home, Break → klahv Break.

Standard shortcut keys (selection)¶

Pronunciation of English terms¶

Unlocalized English terms generally use English pronunciation, but established Estonian pronunciations should be used (server). Adapt to Estonian phonetic system if the original sounds awkward.

Acronym pronunciation¶

Acronyms resembling words are pronounced as words: RADIUS → raadius, RAS → rass, ISA → isa, LAN → lann, WAN → vann, WAP → vapp, MAPI → mapi, POP → popp, OWA → ova.

Letter-by-letter pronunciation for non-word acronyms: URL → uu-err-ell, IP → ii-pee, TCP/IP → tee-tsee-pee/ii-pee, XML → iks-emm-ell, HTML → haa-tee-emm-ell, SQL → ess-kuu-ell.

URL pronunciation¶

Drop the “http://” prefix; read the rest. Pronounce www as vee-vee-vee, dot as punkt. Example: http://www.microsoft.com/et-ee → haa-tee-tee-pee-koolon-kaldkriips-kaldkriips-vee-vee-vee-maikrossoft-punkt-komm-kaldkriips-ee-tee-sidekriips-ee-ee.

Copilot prompts: best practices for localization¶

Copilot prompts are functional — translation quality directly affects AI response quality. Apply these principles when localizing predefined prompts:

• Be clear and specific. Source prompts are usually questions or requests starting with action verbs. Make target prompts read as natural Estonian questions or requests. Avoid vague language.
• Keep it conversational. Use simple natural Estonian. Use an informal tone of voice and form of address.
• Be polite and professional. Kind respectful language; no slang or jargon.
• Use quotation marks to mark content for Copilot to write, modify, or replace.
• Pay attention to punctuation, grammar, and capitalization. Clear communication enables collaboration.
• Handle entity tokens correctly. Tokens like <entity type='file'>file</entity> aren’t localizable, but their position must make sense in target syntax. Some prompts use display text as examples — translate those if DevComment indicates.
• Be consistent. Similar English prompts → similar Estonian prompts.

Examples:

Reference materials: authoritative Estonian dictionaries¶

Use these references for orthography, grammar, and terminology when this guide doesn’t specify. Standard authorities consulted by professional Estonian translators across spheres.

Normative sources (must be followed):

1. Eesti keele käsiraamat (Eesti Keele Sihtasutus, 2007) — Estonian language handbook. Available online at eki.ee/books/ekk09.
2. Eesti õigekeelsussõnaraamat ÕS 2018 (Eesti Keele Sihtasutus, 2018) — Standard Estonian spelling dictionary. Available online at eki.ee/dict/qs.
3. Eesti ortograafia (Eesti Keele Sihtasutus, 2015) — Standard orthographic reference.
4. Eesti Keele Instituut — The Estonian Language Institute. portaal.eki.ee
5. Sõnaveeb — Word resource portal. sonaveeb.ee

Supplementary sources (informative):

1. Väikesed tarbetekstid. Käsiraamat (Maire Raadik, 2014) — Handbook on small functional texts.
2. Keeleseadus (Estonian Language Act) — Legal framework for Estonian language use. riigiteataja.ee/akt/114032014046
3. E-keelenõu — Online language advice. eki.ee
4. Lühendiraamat (Martin Ollisaar, Valgus, 2006) — Abbreviation reference.

FAQ¶

What’s the modern register for Estonian translation across professional contexts?¶

Clear, friendly, concise, with the look and feel of original Estonian writing. This applies to medical patient materials, marketing copy, software UI, and consumer-facing legal documents. Pure technical/legal contracts retain more formality but clarity still beats bureaucratic register.

When should I deviate from literal translation in Estonian?¶

When literal translation produces unnatural Estonian — which happens in most marketing, healthcare patient communication, and software UI translation. Less applicable in sworn legal translation and certified document translation where source-faithful rendering is mandated. Translator judgment determines the boundary.

Which Estonian vocabulary should I avoid in modern translation?¶

Overly formal Latinate verbs and bureaucratic phrasing (modifitseerima, sünkroniseerima, navigeerima, protsess, pukseerima, klaviatuuri otsetee). Replace with everyday Estonian forms (muutma, sünkroonima, liikuma, toiming, lohistama, kiirklahv). See full table in “Words and phrases to avoid” above.

How should I address users in Estonian translation?¶

Use the second-person plural teie (lowercase). Teams and Skype historically use second-person singular sina. Avoid we when referring to the computer or system (use passive voice instead). Avoid pronouns when sentence structure already defines the actor.

How should product names be handled in Estonian?¶

Decline them per Estonian grammar (Wordi, Exceli, Office’isse, Microsofti). Estonian readers expect declension and don’t interpret case endings as name modification. Use modifier nouns (nupp, käsk, suvand) only when a name cannot be declined naturally — excessive modifier use sounds unnatural.

What authoritative Estonian language references should I use?¶

Eesti keele käsiraamat (online at eki.ee/books/ekk09/), ÕS 2018 spelling dictionary, Eesti ortograafia (2015), the Estonian Language Institute portal at portaal.eki.ee, and Sõnaveeb. For supplementary reference: Maire Raadik’s Väikesed tarbetekstid (2014) and the Language Act.

Sources¶

• Microsoft Globalization Style Guides hub
• Eesti keele käsiraamat (online)
• ÕS 2018 — Eesti õigekeelsussõnaraamat
• Eesti Keele Instituut
• Sõnaveeb
• Estonian Language Act (Keeleseadus)