History · 38 languages · 18 scripts

The world's 38 biggest languages,
and the design of how they're written

38 languages have 50 million or more speakers in Ethnologue's 2026 ranking, led by English (1.49 billion), Mandarin Chinese (1.18 billion), Hindi (611 million), Spanish and Modern Standard Arabic. They are written in 18 scripts of seven basic kinds — from the 26-letter Latin alphabet to thousands of Chinese characters and Korea's purpose-designed Hangul.

By , FounderUpdated Every figure sourced

Quick answer: which languages have 50 million or more speakers?

  1. English — 1.49 billion (372 million native)
  2. Mandarin Chinese — 1.18 billion (988 million native)
  3. Hindi — 611 million (347 million native)
  4. Spanish — 561 million (487 million native)
  5. Modern Standard Arabic — 335 million
  6. French — 334 million (75 million native)
  7. Bengali — 274 million (234 million native)
  8. Portuguese — 269 million (252 million native)
  9. Indonesian — 255 million (78 million native)
  10. Urdu — 246 million (78 million native)

…and 28 more, from Russian to Bhojpuri. Source: Ethnologue 2026 [1,2].

Explore all 38

Ethnologue ranks individual languages, not "macrolanguages", so Arabic and Chinese appear as several separate varieties — Modern Standard, Egyptian, Levantine and Sudanese Arabic; Mandarin, Cantonese and Wu Chinese[1,2]. Counts include second-language speakers, which is why English, spoken natively by 372 million, leads with 1.49 billion[2].

Languages
3850M+ speakers each
Scripts
187 basic kinds
Language families
12Indo-European is the largest
Speaker counts, added
8.1Bpeople counted once per language

Showing 38 of 38 languages

First language vs second language

For 9 of the 38, more people speak the language as a second language than as a first: English, Modern Standard Arabic, French, Indonesian, Urdu, Nigerian Pidgin, Swahili, Tagalog (Filipino), Thai. These are the bridge languages of trade, school and government[2].

English is the biggest of these bridge languages, and most of its speakers learned it as a second language. If you’re one of them, our free English level test places you on the CEFR scale from A1 to C2 and gives you a certificate for your CV and LinkedIn.

English#1 English · 1.49B total · 372M first · 1.12B second · 75% learned it as a second language
  1. 1English1.49B
  2. 2Mandarin Chinese1.18B
  3. 3Hindi611M
  4. 4Spanish561M
  5. 5Modern Standard Arabic335M
  6. 6French334M
  7. 7Bengali274M
  8. 8Portuguese269M
  9. 9Indonesian255M
  10. 10Urdu246M
  11. 11Russian210M
  12. 12German133M
  13. 13Japanese126M
  14. 14Nigerian Pidgin121M
  15. 15Egyptian Arabic118M
  16. 16Marathi99M
  17. 17Vietnamese97M
  18. 18Telugu96M
  19. 19Swahili95M
  20. 20Hausa95M
  21. 21Turkish94M
  22. 22Western Punjabi90M
  23. 23Tagalog (Filipino)87M
  24. 24Tamil86M
  25. 25Cantonese (Yue Chinese)86M
  26. 26Wu Chinese83M
  27. 27Persian (Iranian)82M
  28. 28Korean82M
  29. 29Amharic78M
  30. 30Thai71M
  31. 31Javanese69M
  32. 32Italian66M
  33. 33Gujarati62M
  34. 34Kannada59M
  35. 35Levantine Arabic58M
  36. 36Sudanese Arabic54M
  37. 37Yoruba53M
  38. 38Bhojpuri53M
Millions of speakers, Ethnologue 2026. Bars share one scale; “Share learned” stretches each bar to 100% to show the split.

Added together, their speaker counts reach 8.1 billion, because many people speak two or three of these languages and are counted once for each[2].

“How are you?” in 38 languages

The same question, written in each language's own script — and spoken. Press play to hear all of them one after another, or pick one to stop the stage; each language page compares it side by side with any other and cites its phrasebook source.

How are you?

How are you?

English →
1 / 38

Synthetic voice (ElevenLabs Eleven v3, voice “Sia”). Pronunciation is approximate and may not match a native speaker; Cantonese and Wu Chinese are not voiced because the model reads Chinese characters with Mandarin sounds.

Where writing came from

Most of the world's alphabets descend from a single invention. In the second millennium BCE, Semitic speakers in Egypt and Sinai adapted Egyptian hieroglyphs into Proto-Sinaitic, considered the earliest alphabet; its Phoenician descendant spread around the Mediterranean[3,4,5].

Hover, tap or tab to any script · arrow keys walk the tree
Egyptian hieroglyphsOracle-bone scriptProto-SinaiticChinese charactersPhoenicianSouth ArabianGreekAramaicGe'ezKana (Japanese)Simplified & traditionalLatinCyrillicNabataeanBrahmiArabicGuptaSouthern BrahmiPallavaHangul (invented 1443)Persian · NastaliqDevanagariBengaliGujaratiTamilTeluguKannadaThaiJavanese
नमस्ते
DevanagariAbugida (syllabic)

Took shape in use from the 7th century CE and in its present form by about the 11th century.

→ → → → → →

Phoenician gave rise to Greek, and through it to Latin and Cyrillic, and to Aramaic, whose Nabataean branch became Arabic[5,6,7]. Ancient South Arabian, another offshoot of Proto-Sinaitic, became Ge'ez[8,9].

India's Brahmi, the ancestor of every Indian script and, through Pallava, of Thai and Javanese, may have been inspired by Aramaic, but its origin is still debated, so that link is drawn dashed[10,5,11]. Chinese writing arose on its own from the oracle bones of the Shang dynasty[12], and Hangul was invented outright in 1443[13].

Seven kinds of script

Writing systems differ in what one symbol stands for. The 18 scripts on this page use seven basic designs; see chunking for why units like Hangul's syllable blocks are easier to read.

  • HxpgLatin alphabet · Hxpg

    Alphabet

    One letter for each consonant and vowel.

    Latin alphabet · Cyrillic alphabet

  • سلامArabic script · salām

    Abjad (consonants)

    Letters write consonants; short vowels are optional marks.

    Arabic script · Persian alphabet (Perso-Arabic) · Nastaliq (Urdu and Shahmukhi)

  • नमस्तेDevanagari · namaste

    Abugida (syllabic)

    Each consonant carries a vowel, changed by marks attached to it.

    Devanagari · Bengali–Assamese script · Gujarati script · Telugu script · Kannada script · Tamil script · Thai script · Ge'ez (Ethiopic) script · Javanese script (Aksara Jawa)

  • 你好Chinese characters · nǐ hǎo

    Logographic

    Each character writes a word or meaningful part of one.

    Chinese characters (simplified) · Chinese characters (traditional)

  • 日本語Japanese · nihongo

    Mixed system

    Several writing systems are used together in one sentence.

    Japanese (kanji, hiragana, katakana)

  • 한글Hangul · hangeul

    Featural alphabet

    Letter shapes encode how the sound is made.

    Hangul

Ten ways to write 0–9

The digits 0–9 used worldwide began in India, where they developed out of Brahmi numerals; they reached the Islamic world through writers such as al-Khwarizmi and Europe through Fibonacci's Liber abaci of 1202[14,15]. Many scripts still keep their own forms of the same ten digits.

  • Western (Hindu-Arabic)14 Latin-script languages, plus Korean and most of the world
    1947
    1. 00
    2. 11
    3. 22
    4. 33
    5. 44
    6. 55
    7. 66
    8. 77
    9. 88
    10. 99
  • Arabic scriptModern Standard Arabic, Egyptian Arabic, Levantine Arabic, Sudanese Arabic
    ١٩٤٧
    1. ٠0
    2. ١1
    3. ٢2
    4. ٣3
    5. ٤4
    6. ٥5
    7. ٦6
    8. ٧7
    9. ٨8
    10. ٩9
  • Persian and UrduUrdu, Western Punjabi, Persian (Iranian)
    ۱۹۴۷
    1. ۰0
    2. ۱1
    3. ۲2
    4. ۳3
    5. ۴4
    6. ۵5
    7. ۶6
    8. ۷7
    9. ۸8
    10. ۹9
  • DevanagariHindi, Marathi, Bhojpuri
    १९४७
    1. ०0
    2. १1
    3. २2
    4. ३3
    5. ४4
    6. ५5
    7. ६6
    8. ७7
    9. ८8
    10. ९9
  • Bengali–Assamese scriptBengali
    ১৯৪৭
    1. ০0
    2. ১1
    3. ২2
    4. ৩3
    5. ৪4
    6. ৫5
    7. ৬6
    8. ৭7
    9. ৮8
    10. ৯9
  • Gujarati scriptGujarati
    ૧૯૪૭
    1. ૦0
    2. ૧1
    3. ૨2
    4. ૩3
    5. ૪4
    6. ૫5
    7. ૬6
    8. ૭7
    9. ૮8
    10. ૯9
  • Telugu scriptTelugu
    ౧౯౪౭
    1. ౦0
    2. ౧1
    3. ౨2
    4. ౩3
    5. ౪4
    6. ౫5
    7. ౬6
    8. ౭7
    9. ౮8
    10. ౯9
  • Kannada scriptKannada
    ೧೯೪೭
    1. ೦0
    2. ೧1
    3. ೨2
    4. ೩3
    5. ೪4
    6. ೫5
    7. ೬6
    8. ೭7
    9. ೮8
    10. ೯9
  • Tamil scriptTamil
    ௧௯௪௭
    1. ௦0
    2. ௧1
    3. ௨2
    4. ௩3
    5. ௪4
    6. ௫5
    7. ௬6
    8. ௭7
    9. ௮8
    10. ௯9
  • Chinese and JapaneseMandarin Chinese, Japanese, Cantonese (Yue Chinese), Wu Chinese
    一九四七
    1. 〇0
    2. 一1
    3. 二2
    4. 三3
    5. 四4
    6. 五5
    7. 六6
    8. 七7
    9. 八8
    10. 九9
  • JapaneseMandarin Chinese, Japanese, Cantonese (Yue Chinese), Wu Chinese
    一九四七
    1. 〇0
    2. 一1
    3. 二2
    4. 三3
    5. 四4
    6. 五5
    7. 六6
    8. 七7
    9. 八8
    10. 九9
  • Thai scriptThai
    ๑๙๔๗
    1. ๐0
    2. ๑1
    3. ๒2
    4. ๓3
    5. ๔4
    6. ๕5
    7. ๖6
    8. ๗7
    9. ๘8
    10. ๙9
  • Ge'ez (Ethiopic) scriptAmharic
    not positional
    1. —0
    2. ፩1
    3. ፪2
    4. ፫3
    5. ፬4
    6. ፭5
    7. ፮6
    8. ፯7
    9. ፰8
    10. ፱9
  • Javanese scriptJavanese
    ꧑꧙꧔꧗
    1. ꧐0
    2. ꧑1
    3. ꧒2
    4. ꧓3
    5. ꧔4
    6. ꧕5
    7. ꧖6
    8. ꧗7
    9. ꧘8
    10. ꧙9

What scripts teach product owners

Writing systems are the oldest interfaces humans have designed, refined over centuries by millions of readers. Six lessons carry straight into apps — and into the conventions users bring with them.

  1. Store meaning, draw appearance

    In Devanagari the vowel sign ि is drawn before its consonant but spoken after it. Unicode stores text in spoken order and leaves drawing to the font — a clean split between data and presentation that every design system should copy[16].

    See it in Hindi →
  2. Direction is not one-way

    Arabic, Persian and Urdu run right to left, but numbers inside them run left to right, so one line can change direction twice. Interfaces built for these languages must handle both[17].

    See it in Arabic →
  3. Case is a local rule

    Uppercasing 'i' gives 'İ' in Turkish, not 'I'. Most of the world's scripts have no case at all. Never assume a transformation works the same way in every language[18].

    See it in Turkish →
  4. Spaces are optional

    Thai and traditional Javanese do not separate words with spaces, so line-breaking needs a dictionary. Truncating text at the nearest space breaks in these languages[19,20].

    See it in Thai →
  5. Tools shape forms

    The reed pen's square dot became the unit of Arabic calligraphy, and round South Indian letters are often linked to writing on palm leaves. Every interface is shaped by the medium it is made for, too[21,22].

    See it in Tamil →
  6. A system beats a set of one-offs

    Hangul builds every syllable from 24 letters by rule, and Ge'ez makes seven vowel forms of each consonant by regular marks. Readers learn the pattern once — the same reason consistent components make apps easier to learn[13,23].

    See it in Korean →

The full table

8.06Bspeaker counts, addedclick a slice to filter the table
Shares of the 38 languages' combined speaker counts (Ethnologue 2026). Someone who speaks two of these languages is counted in both, so shares describe speaker counts, not people.
Languages with 50 million or more speakers, Ethnologue 2026 (millions). Click a heading to sort. Each row opens that language's full page.
Own name
1EnglishEnglish1.5B372M1.1BLatin alphabetIndo-European
2Mandarin Chinese普通话1.2B988M194MChinese charactersSino-Tibetan
3Hindiहिन्दी611M347M264MDevanagariIndo-European
4Spanishespañol561M487M75MLatin alphabetIndo-European
5Modern Standard Arabicالعربية الفصحى335M0335MArabic scriptAfro-Asiatic
6Frenchfrançais334M75M258MLatin alphabetIndo-European
7Bengaliবাংলা274M234M43MBengali–Assamese scriptIndo-European
8Portugueseportuguês269M252M18MLatin alphabetIndo-European
9IndonesianBahasa Indonesia255M78M177MLatin alphabetAustronesian
10Urduاُردُو246M78M168MNastaliqIndo-European
11Russianрусский210M133M77MCyrillic alphabetIndo-European
12GermanDeutsch133M76M57MLatin alphabetIndo-European
13Japanese日本語126M124M2MJapaneseJaponic
14Nigerian PidginNaijá121M5M116MLatin alphabetEnglish-based creole
15Egyptian Arabicمصري118M83M35MArabic scriptAfro-Asiatic
16Marathiमराठी99M83M16MDevanagariIndo-European
17VietnameseTiếng Việt97M86M11MLatin alphabetAustroasiatic
18Teluguతెలుగు96M83M13MTelugu scriptDravidian
19SwahiliKiswahili95M4M91MLatin alphabetNiger–Congo
20HausaHausa95M58M36MLatin alphabetAfro-Asiatic
21TurkishTürkçe94M86M7MLatin alphabetTurkic
22Western Punjabiپنجابی90M——NastaliqIndo-European
23Tagalog (Filipino)Tagalog87M33M54MLatin alphabetAustronesian
24Tamilதமிழ்86M79M8MTamil scriptDravidian
25Cantonese (Yue Chinese)粵語86M85M1MChinese charactersSino-Tibetan
26Wu Chinese吴语83M83M—Chinese charactersSino-Tibetan
27Persian (Iranian)فارسی82M65M17MPersian alphabetIndo-European
28Korean한국어82M82M1MHangulKoreanic
29Amharicአማርኛ78M39M39MGe'ez (Ethiopic) scriptAfro-Asiatic
30Thaiภาษาไทย71M27M44MThai scriptKra–Dai
31Javaneseꦧꦱꦗꦮ69M——Javanese scriptAustronesian
32Italianitaliano66M60M6MLatin alphabetIndo-European
33Gujaratiગુજરાતી62M58M5MGujarati scriptIndo-European
34Kannadaಕನ್ನಡ59M44M15MKannada scriptDravidian
35Levantine Arabicشامي58M55M3MArabic scriptAfro-Asiatic
36Sudanese Arabicعربي سوداني54M43M11MArabic scriptAfro-Asiatic
37YorubaYorùbá53M48M5MLatin alphabetNiger–Congo
38Bhojpuriभोजपुरी53M53M—DevanagariIndo-European
The world's 38 most spoken languages, each with 50 million or more speakers (Ethnologue 2026), and the scripts they are written in
The 38 languages at a glance. Free to share with a link to this page.

How we built this

  • Which languages: every individual language with 50 million or more total speakers in Ethnologue's 2026 ranking [1,2]. Ethnologue's own page publishes the top 20 to one decimal; ranks 21–38 use its 2026 figures rounded to the million. A dash means Ethnologue gives no first/second-language split.
  • Phrases: each “How are you?”, “Hello” and “Thank you” comes from a published phrasebook — Omniglot for most languages, the British Council for Nigerian Pidgin, and learner guides for Sudanese Arabic — and is cited on its page. Spoken varieties without a fixed spelling (Levantine and Sudanese Arabic, Nigerian Pidgin) vary from writer to writer.
  • History and design: dates and claims are cited to Encyclopaedia Britannica, the Unicode Standard 17.0, UNESCO, government records, museums and primary sources. Where scholars disagree — the age of the Ram Khamhaeng inscription, the date of Halmidi, Brahmi's origin — the page says so.
  • Maps: Natural Earth's public-domain 1:110m country outlines; shading marks whole countries where a language is official or widely spoken, and pins mark where it began. Small places appear as dots.
  • Type: every script is set in a Noto typeface, subset to the characters each page shows, so the letters you see are drawn by a font made for that script.

Frequently asked questions

What are the most spoken languages in the world?

By total speakers (first and second language), Ethnologue's 2026 ranking puts English first with about 1.49 billion, followed by Mandarin Chinese (1.18 billion), Hindi (611 million), Spanish (561 million) and Modern Standard Arabic (335 million).

How many languages have more than 50 million speakers?

38 languages have 50 million or more total speakers in Ethnologue's 2026 figures. Ethnologue counts individual languages, so varieties of Arabic and Chinese such as Egyptian Arabic, Cantonese and Wu are listed separately.

Which language has the most native speakers?

Mandarin Chinese, with about 988 million first-language speakers. English has more speakers in total, but most of them (1.12 billion) learned it as a second language.

Which writing system do the most languages use?

The Latin alphabet: 14 of the 38 languages are written mainly in it, including English, Spanish, French, Portuguese, Indonesian, Vietnamese, Turkish and Swahili. Brahmic scripts descended from ancient Brahmi come next, used across South and Southeast Asia.

Why are Hindi and Urdu listed as separate languages?

Spoken Hindi and Urdu are largely mutually intelligible — linguists call the shared base Hindustani — but they have separate standards: Hindi is written in Devanagari and draws learned words from Sanskrit; Urdu is written in Perso-Arabic Nastaliq and draws them from Persian and Arabic. Ethnologue counts them separately.

What was the first alphabet?

Proto-Sinaitic, created in the second millennium BCE by Semitic speakers in Egypt and Sinai who adapted Egyptian hieroglyphs, is considered the earliest alphabet. Through its descendant Phoenician it is the ancestor of Greek, Latin, Cyrillic, Arabic and, possibly, India's Brahmi.

Which major languages are written right to left?

Among the 38 languages with 50 million or more speakers, Modern Standard Arabic, Urdu, Egyptian Arabic, Western Punjabi, Persian (Iranian), Levantine Arabic, Sudanese Arabic are written right to left, in the Arabic script or its Persian and Nastaliq styles. Numbers inside right-to-left text still run left to right.

Where do the numbers 0–9 come from?

They began in India, developing from Brahmi numerals, and reached Europe through the Arabic world — Fibonacci's Liber abaci of 1202 popularised them — which is why they are called Hindu-Arabic numerals. Scripts such as Devanagari, Bengali, Thai and Arabic still use their own forms of the same ten digits.

Sources

  1. Ethnologue — What are the top 200 most spoken languages? (2026)
  2. Wikipedia — List of languages by total number of speakers (Ethnologue 2026 figures, retrieved October 2026)
  3. Omniglot — Proto-Sinaitic script
  4. Omniglot — Phoenician alphabet
  5. Encyclopaedia Britannica — Alphabet (writing)
  6. Omniglot — Aramaic alphabet
  7. Omniglot — Nabataean alphabet
  8. Omniglot — Ancient South Arabian script
  9. Omniglot — Ge'ez (Ethiopic) script
  10. Encyclopaedia Britannica — Brahmi
  11. Omniglot — Pallava script
  12. Encyclopaedia Britannica — Chinese languages
  13. Encyclopaedia Britannica — Hangul
  14. Encyclopaedia Britannica — Hindu-Arabic numerals
  15. Encyclopaedia Britannica — Liber abaci (1202)
  16. The Unicode Standard 17.0, chapter 12: South and Central Asia-I (Devanagari, Bengali, Gujarati, Tamil, Telugu, Kannada)
  17. The Unicode Standard 17.0, chapter 9: Middle East-I (Arabic)
  18. The Unicode Standard 17.0, chapter 7: Europe-I (Latin, Greek, Cyrillic)
  19. The Unicode Standard 17.0, chapter 16: Southeast Asia-I (Thai)
  20. The Unicode Standard 17.0, chapter 17: Southeast Asia-II (Javanese, Philippine scripts, Kawi)
  21. Asian Art Museum, San Francisco — The history of Islamic calligraphy
  22. Wikipedia — Ola leaf (palm-leaf writing and rounded scripts)
  23. The Unicode Standard 17.0, chapter 19: Africa (Ethiopic)

Checked October 25, 2026. Each language page lists the sources for its own history, phrases and script. Found an error? Tell us.

Pass it on

Know someone who'd enjoy this?

Send it to them — one tap, no sign-up. Good ideas travel best between friends.

Keep exploring

Where to go from here

Join free

Get the ideas, not the noise.

Sign up with Google and we’ll send you short UX psychology tips by email — why people tap, stall and leave, and what to do about it. Unsubscribe anytime. Prefer something to keep? Get the free Laws of UX guide as a PDF.

Get the free Laws of UX PDFFree · no spam · unsubscribe anytime
Free English level test

Speak English as a second language? Find your level.

A free, CEFR-aligned test of grammar, vocabulary, reading and listening from A1 to C2, with a certificate that states your level for your CV, LinkedIn and job applications. Not proctored, and not for visas or university admission.

  • Credential ID
  • QR verification
  • LinkedIn-ready
Free PDF · 74 pages · 2 editions

The Laws of UX, in one field guide

60 research-backed principles in 6 categories — what each one means, why it works on the human brain and how to use it on your own screens. Written by Sushant Singh, founder of Science of Apps, who has scaled consumer apps to lacs of users and spent 10+ years leading product across startups and digital businesses in the US, EU & India.

Pick your edition
Free. No card, no paywall — just your email.
Free PDF · 74 pages

Get Laws of UX by email

Choose your edition

You’ll also get occasional UX psychology tips. Unsubscribe anytime. Privacy policy

Free PDF · 90 pages · 2 editions

100+ Subscription Paywalls, decoded

110 real paywalls from top apps, taken apart, rated out of 5 and decoded with the Laws of UX — plus the frameworks, copy and checklists to ship your own. By Sushant Singh, founder of Science of Apps.

  • 110 paywalls rated
  • 34 Laws of UX
  • 12 frameworks

Which paywall scored 4.8 out of 5, and which scored 2.6? It’s all inside.

Pick your edition
Free. No card, no paywall — just your email.
Free PDF · 90 pages

Get 100+ Paywalls, Decoded by email

Choose your edition

You’ll also get occasional UX psychology tips. Unsubscribe anytime. Privacy policy