Human Languages

Human languages connect people across generations, countries and continents. Some are spoken by hundreds of millions of people, while others sustain much smaller communities. Size alone tells only part of the story: languages also differ in their historical relationships, international reach, place in education and visibility online. Ten charts explore these contrasts, from the largest Asian languages to Indigenous languages of the Americas and multilingual societies. A reference table brings together 37 major languages, their families and their first- and second-language speaker estimates.

Speaker comparisons use a consistent 2025 snapshot. Country language counts use the dated 2019 series; older distance and Indigenous-language estimates are identified separately. Website figures are dated 22 October 2026. Language families describe historical relationships between languages, not biological ancestry between people.

Statistics

  1. Major Asian languages by family, 2025
  2. Selected language distances from English
  3. Large Indigenous language communities of the Americas
  4. Ten largest native-speaker populations, 2025
  5. Living languages in ten countries, 2019
  6. First- and second-language speakers, 2025
  7. Language families among 37 widely spoken languages
  8. Native-speaker size and second-language reach, 2025
  9. Access to education in a familiar language
  10. Leading website content languages, October 2026
  11. Language speaker reference table

1. Major Asian languages by family, 2025

Indo-EuropeanSino-TibetanAfro-AsiaticAustronesianJaponic

Mandarin has by far the largest speaker population among these six Asian languages, with almost 1.2 billion people. Hindi follows with roughly 609 million, while Standard Arabic, Bengali and Indonesian each reach several hundred million. Japanese has a smaller total, but still serves a community of around 126 million people. Asia therefore contains several very different kinds of large language community.

Hindi and Bengali belong to the same broad Indo-European family, despite their distinct regional centres and literary traditions. Mandarin belongs to Sino-Tibetan, Indonesian to Austronesian, Standard Arabic to Afro-Asiatic and Japanese to Japonic. Their prominence shows that Asia’s linguistic landscape cannot be reduced to one dominant family. Nor does population alone explain a language’s reach: Indonesian brings together speakers from a highly multilingual archipelago, whereas Japanese is much more concentrated among first-language speakers. The relationship between a country’s population and its leading language is especially complex where many regional languages remain important in daily life.

Source: 2025 speaker estimates, published compilation of Ethnologue data; Ethnologue. Millions of speakers. Figures are rounded; independently rounded totals can differ from the sum of the components. Multilingual people may appear under several languages. Selected languages associated with Asia; totals include speakers worldwide. Standard Arabic is the formal variety, not all Arabic varieties combined.

Shaheed Minar and the Bengali language

Dhaka’s Shaheed Minar commemorates those killed during the Bengali Language Movement demonstrations of 1952. Bengali is one of the large Asian languages shown above, with a substantial speaker population centred on Bangladesh and eastern India. The monument connects that demographic scale with the importance of language in public life and collective identity. A language is more than a means of communication: it can carry literature, memories and demands for recognition. The memorial also gives a human setting to the statistics, reminding visitors that the right to use a familiar language has sometimes been defended at considerable personal cost.

Jubair Bin Iqbal · CC BY-SA 4.0

2. Selected language distances from English

Scots sits closest to English among the selected languages, followed by German. All three belong to the Germanic branch of Indo-European, reflecting a shared linguistic history. French and Spanish are more distant relatives within the Romance branch, while Russian, Hindi and Irish belong to other branches of the same broad family. Shared ancestry can remain visible even after languages have developed very different sounds and grammar.

Irish provides a useful contrast between geographic proximity and linguistic relationship. Ireland is close to England, yet Irish is Celtic rather than Germanic. Mandarin, Arabic and Japanese belong to different families and sit farther from English in this comparison. Everyday familiarity can tell a different story: English contains many words borrowed from French, even though its historical family is Germanic. Contact, borrowing and common ancestry therefore contribute different layers to the way languages resemble one another. A familiar-looking word does not necessarily mean two languages share a particularly recent common ancestor.

Source: eLinguistics comparison tool. Retained undated recorded scores on a 0–100 scale; lower means closer in this tool. The calculator was unavailable during this update, so these values were not revalidated. They are not a universal measure, a percentage, a learning-difficulty score or a measure of people’s ancestry.

3. Large Indigenous language communities of the Americas

Quechua-speaking communities across the Andes and Mayan-speaking communities in Mexico and Guatemala stand out for their scale. Guarani also has millions of speakers, centred on Paraguay and nearby areas. Aymara and Nahuatl form another substantial group, showing how several Indigenous languages remain part of everyday life across large populations rather than surviving only in historical records.

The distribution is strongly concentrated in Latin America. Mexico appears repeatedly through Nahuatl, Mixtec, Zapotec, Otomi and Totonac, while Quechua and Aymara connect communities across the Andean region. This diversity exists within countries as well as across borders: sharing a national identity does not necessarily mean sharing a home language. The smaller communities in this selection still represent hundreds of thousands of speakers, but their circumstances can differ considerably between regions and generations. Maintaining a language depends on its continued use in families, schools, public life and work, alongside the choices and opportunities available to younger speakers.

Source: Native Languages of the Americas. Approximate undated estimates, not a 2026 census. Quechua and Mayan are grouped language communities; several other entries encompass varieties. These group totals must not be treated as a ranking of individual standardised languages.

4. Ten largest native-speaker populations, 2025

Mandarin’s native-speaker population is larger than those of Spanish and English combined. Its roughly 990 million first-language speakers give it a clear lead, followed by Spanish at about 484 million and English at 390 million. Hindi, Portuguese and Bengali also have exceptionally large home-language communities, each with hundreds of millions of speakers.

The ranking highlights different histories of geographic expansion. Spanish and Portuguese connect populations on opposite sides of the Atlantic, while Mandarin and Japanese are much more concentrated in East Asia. English has a substantial native-speaking population, but its position here is lower than its standing when additional-language speakers are included. South Asia is represented by Hindi, Bengali and Western Punjabi, illustrating the region’s demographic weight without treating all its inhabitants as speakers of one language. These large communities coexist with many smaller languages. The prominence of a national or international language does not mean that regional languages have disappeared from homes, neighbourhoods or cultural life.

Source: 2025 Ethnologue native-speaker compilation retained for this dated ranking; native-speaker listing. Millions. Western Punjabi’s 90 million is retained from the native-speaker series; the separate total-speaker table does not report its L1/L2 breakdown.

5. Living languages in ten countries, 2019

EstablishedImmigrant

Papua New Guinea leads this historical comparison by a wide margin, followed by Indonesia and Nigeria. These countries show that linguistic diversity is not simply a consequence of having the largest national population. India and China contain very large populations, yet each has fewer recorded languages than Papua New Guinea in this series. A country can contain many relatively small language communities alongside languages used more widely for communication.

The United States and Australia have a different balance from most of the other countries shown, with larger immigrant-language contributions. Their linguistic landscapes combine Indigenous languages with languages carried by migration. Mexico, Cameroon and Brazil also have substantial established-language diversity, even where a single language dominates national administration or much of the media. These patterns reveal why national borders provide only a partial picture of language. Within one country, local traditions, settlement histories and movement between regions can produce very different combinations of languages in everyday use.

Source: Ethnologue 2019 country-series compilation; Papua New Guinea: Guinness, 2019 Ethnologue edition. Historical series; sorted by established + immigrant languages. Categories follow this historical source and are not the same as Indigenous versus non-Indigenous. Not presented as current counts.

Cultural diversity at the Goroka Show

The Goroka Show brings a visible expression of Papua New Guinea’s cultural diversity to the language comparison above. The country contains hundreds of language communities, many rooted in particular places and local histories. Public performances can bring people together while retaining distinct traditions of dress, music and movement. Cultural groups and languages do not always correspond one to one, so a photograph cannot reveal exactly how many languages are represented. It nevertheless provides a setting for understanding why the country’s remarkable language count is more than an abstract total: it reflects many communities maintaining their own ways of communicating and remembering.

Anselmo Lastra · CC BY 2.0

6. First- and second-language speakers, 2025

First languageSecond language

English’s global reach depends heavily on people who learned it after their first language. Its second-language population is almost three times its native-speaking population. Mandarin presents the reverse pattern: most of its speakers learned it as a first language. These two paths produce the world’s largest overall language communities through very different combinations of home use and wider communication.

French also gains much of its reach from additional-language speakers, whereas Spanish and Portuguese draw more heavily on large native-speaking communities. Hindi has substantial populations in both categories. Standard Arabic occupies a distinctive position because people generally learn a local Arabic variety at home and acquire the formal standard through education. The balance matters for how languages are encountered: some are central to family and neighbourhood life, while others connect people across regional or national boundaries. A person can participate in both kinds of community, using different languages for different parts of daily life.

Source: 2025 speaker estimates, published compilation of Ethnologue data; Ethnologue. Millions of speakers. Figures are rounded; independently rounded totals can differ from the sum of the components. Multilingual people may appear under several languages. Bars add the published L1 and L2 components; Bengali totals 285 million here versus the independently rounded published total of 284 million. Standard Arabic is treated as an acquired formal variety.

7. Language families among 37 widely spoken languages

Indo-European — 15Other families — 7Afro-Asiatic — 6Sino-Tibetan — 3Austronesian — 3Dravidian — 3

Indo-European contributes the largest number of entries among these widely spoken languages. Its presence spans several branches, including Germanic, Romance, Indo-Aryan and Iranian. That breadth connects major languages of Europe with those of South Asia, while the worldwide reach of English, Spanish, French and Portuguese extends the family’s presence across other continents.

Afro-Asiatic includes Standard Arabic and several regional Arabic varieties, alongside Hausa and Amharic. Sino-Tibetan is represented by Mandarin, Yue and Wu, while Indonesian, Tagalog and Javanese belong to Austronesian. Tamil, Telugu and Kannada represent the Dravidian family. The remaining entries include Japanese, Korean, Thai, Turkish, Swahili and Nigerian Pidgin, illustrating how very large speaker communities occur outside the most numerous families in this selection. Counting languages and counting speakers reveal different patterns: a family can contribute only a few entries yet contain an exceptionally large population. Equally, families with many smaller languages can be largely absent from a list dominated by the biggest communities.

Source: 2025 speaker estimates, published compilation of Ethnologue data; Ethnologue. Millions of speakers. Figures are rounded; independently rounded totals can differ from the sum of the components. Multilingual people may appear under several languages. Each of the 37 listed languages with at least 50 million total speakers counts once. Shares describe this selection, not all world languages or people. Creole classification follows the source.

8. Native-speaker size and second-language reach, 2025

Every bubble is labelled; colours indicate language families. Larger bubbles mean more total speakers.
Indo-EuropeanSino-TibetanAfro-AsiaticAustronesianJaponic

A large native-speaking population does not necessarily mean a language is used mainly by people who learned it later. Mandarin combines the largest native base with a relatively small second-language share, while English has a smaller native base but a much higher share. Spanish also depends predominantly on native speakers, despite its broad international distribution.

Indonesian occupies a different position: it has fewer native speakers than Japanese, yet a much greater proportion of people use it as an additional language. This reflects its role in communication across a multilingual country. Hindi lies between the strongly native-dominated and strongly second-language-dominated examples. Standard Arabic is the clearest case of an acquired formal variety, rather than a language normally learned first in the home. Together, these examples show several routes to widespread use. Population growth within a home-language community is one route; adoption for education, administration and communication between communities is another, and the two can operate together.

Source: 2025 speaker estimates, published compilation of Ethnologue data; Ethnologue. Millions of speakers. Figures are rounded; independently rounded totals can differ from the sum of the components. Multilingual people may appear under several languages. Seven selected languages. X: first-language speakers; Y: L2 divided by L1 + L2. Bubble size: published total speakers, on a visibility-adjusted scale rather than exact area proportionality. Colours identify families. “Mandarin” and “Standard Arabic” are shortened labels.

9. Access to education in a familiar language

Lack access — 40Remaining population — 60

Language can create a barrier to learning before a lesson’s subject has even been understood. UNESCO estimates that forty per cent of people lack access to education in a language they speak and understand fluently. For affected learners, following instructions, asking questions and demonstrating knowledge can all become harder when the language of teaching is unfamiliar.

The challenge is especially important in multilingual societies, where a single national teaching language may differ from many children’s home languages. Migration and displacement can bring additional language needs into the same classroom. Learning an international or national language can open valuable opportunities, but learners also need a way to understand what they are being taught while acquiring it. Multilingual education can connect these goals by building on existing language knowledge and introducing additional languages with support. The central issue is whether learners can participate fully in education, rather than having their knowledge obscured by a language barrier.

Source: UNESCO, Languages matter, 2025; article updated January 2026. Global estimate concerning access to education in a language spoken and understood fluently. The 60% segment is the arithmetic remainder; it is not a measure of educational quality, attendance or literacy.

The Rosetta Stone and access to meaning

The Rosetta Stone offers a historical example of making a message available through more than one written form. Its decree appears in Egyptian hieroglyphs, Egyptian demotic script and Greek. The repeated content later helped scholars understand Egyptian writing. Beside the education comparison above, it illustrates the enduring importance of access to meaning across language barriers. Using more than one language can connect different audiences without requiring everyone to begin with the same knowledge. The stone also distinguishes languages from scripts: its three writing systems represent two languages, reminding us that the appearance of writing and the language behind it are different things.

BabelStone · CC BY-SA 3.0

10. Leading website content languages, October 2026

English accounts for by far the largest presence among the website content languages measured here, appearing on almost half of sites with an identified language. Spanish and German follow at around six per cent each, with Japanese and French next. The gap between English and the following languages is much wider than the difference between their speaker populations alone would suggest.

German, Italian, Dutch and Polish have a substantial web presence despite being smaller than many Asian languages by total speaker numbers. Chinese accounts for a much smaller share of the measured websites than its enormous speaker population might lead readers to expect. Website publishing therefore reflects more than demographic size. Businesses, institutions and publishers make language choices around their audiences, and international communication can favour languages used beyond their native communities. Access to useful online information may consequently feel very different depending on which languages a person reads, even when that person belongs to a large language community.

Source: W3Techs content-language survey, 22 October 2026. Percentage of surveyed websites whose content language is known; not internet users, traffic, pages or all online content. Languages can overlap on multilingual sites, so these percentages are not a pie-chart partition.

11. Language speaker reference table

37 languages with at least 50 million total speakers in the 2025 compilation. Scroll horizontally and vertically. Web figures: 22 October 2026. “Not recorded” is not zero. Second-language shares use the summed rounded L1/L2 components; missing and <1 million components are not treated as exact values.

LanguageLanguage familyBranchEstimate yearFirst-language speakers (millions)Second-language speakers (millions)Published total (millions)Second-language share (%)Web content share (%)Notes
EnglishIndo-EuropeanGermanic20253901138152874.549.5Figures rounded.
Mandarin ChineseSino-TibetanSinitic2025990194118416.4Not recordedWeb survey groups Chinese varieties together (1.2%); no variety-level allocation.
HindiIndo-EuropeanIndo-Aryan202534526460943.3<0.1Figures rounded.
SpanishIndo-EuropeanRomance20254847455813.36.1Figures rounded.
Modern Standard ArabicAfro-AsiaticSemitic20250335335100.0Not recordedFormal variety acquired after a home Arabic variety; web survey uses broader Arabic category (0.6%).
FrenchIndo-EuropeanRomance20257423831276.34.6Figures rounded.
BengaliIndo-EuropeanIndo-Aryan20252424328415.10.1Published rounded total differs from summed rounded components.
PortugueseIndo-EuropeanRomance2025250172676.44.1Figures rounded.
IndonesianAustronesianMalayo-Polynesian20258018026069.21.2Figures rounded.
RussianIndo-EuropeanBalto-Slavic202514510825342.73.4Figures rounded.
UrduIndo-EuropeanIndo-Aryan20258017025068.0<0.1Figures rounded.
Standard GermanIndo-EuropeanGermanic2025765813443.35.9Figures rounded.
JapaneseJaponic—202512421261.64.9Figures rounded.
Nigerian PidginEnglish CreoleKrio2025511612195.9Not recordedFigures rounded.
Egyptian ArabicAfro-AsiaticSemitic2025843511929.4Not recordedFigures rounded.
MarathiIndo-EuropeanIndo-Aryan202583169916.2<0.1Figures rounded.
VietnameseAustroasiaticVietic202586119711.30.9Figures rounded.
TeluguDravidianSouth-Central202583139613.5<0.1Figures rounded.
HausaAfro-AsiaticChadic202558369438.3Not recordedFigures rounded.
TurkishTurkicOghuz2025856916.6Not recordedFigures rounded.
Western PunjabiIndo-EuropeanIndo-Aryan2025Not reportedNot reported90Not recordedNot recordedFirst/second-language split not reported in this source.
SwahiliNiger–CongoBantu20254838795.4<0.1Figures rounded.
TagalogAustronesianMalayo-Polynesian202533548762.1<0.1Figures rounded.
TamilDravidianSouth2025798869.2<0.1Published rounded total differs from summed rounded components.
Yue ChineseSino-TibetanSinitic2025851861.2Not recordedWeb survey groups Chinese varieties together (1.2%); no variety-level allocation.
Wu ChineseSino-TibetanSinitic202583<183Not recordedNot recordedWeb survey groups Chinese varieties together (1.2%); no variety-level allocation. L2 below one million; share not calculated from an unknown exact value.
Iranian PersianIndo-EuropeanIranian202565178320.7Not recordedWeb survey uses broader Persian category (1.0%). Published rounded total differs from summed rounded components.
KoreanKoreanic—202581<182Not recorded0.9L2 below one million; share not calculated from an unknown exact value.
ThaiKra–DaiZhuang–Tai202527447162.00.3Figures rounded.
JavaneseAustronesianMalayo-Polynesian2025Not reportedNot reported69Not recordedNot recordedFirst/second-language split not reported in this source.
ItalianIndo-EuropeanRomance2025633664.52.8Figures rounded.
GujaratiIndo-EuropeanIndo-Aryan2025585627.9<0.1Published rounded total differs from summed rounded components.
Levantine ArabicAfro-AsiaticSemitic2025583604.9Not recordedPublished rounded total differs from summed rounded components.
AmharicAfro-AsiaticSemitic202535256041.7<0.1Figures rounded.
KannadaDravidianSouth202544155925.4<0.1Figures rounded.
BhojpuriIndo-EuropeanIndo-Aryan202553<153Not recordedNot recordedL2 below one million; share not calculated from an unknown exact value.
Sudanese ArabicAfro-AsiaticSemitic202541115221.2Not recordedFigures rounded.

Source: 2025 speaker estimates, published compilation of Ethnologue data; Ethnologue. Millions of speakers. Figures are rounded; independently rounded totals can differ from the sum of the components. Multilingual people may appear under several languages. Other datasets: Indigenous languages, 2019 country counts, language distance, education access and web content. Image context: Shaheed Minar; Rosetta Stone. Images are credited individually to their Wikimedia Commons file and licence pages.