Consonant
Consonants are the sounds that frame everything you say. Right now, the words entering your ears are mostly made of them. Every language on Earth uses them, though no two languages share exactly the same set. And hidden inside that universal fact is a world of astonishing variety.
The recently extinct Ubykh language packed 84 consonants into its sound system alongside only 2 or 3 vowels. The Taa language, spoken in southern Africa, reaches 87 under one analysis and 164 under another. Meanwhile, Rotokas, a language of Papua New Guinea, gets by on just six. How can the same category of human sound encompass such extremes? And what makes a consonant a consonant in the first place?
To answer that, linguists reach back to a Greek grammarian named Dionysius Thrax, who named these sounds sýmphōna, meaning "sounded with." His reasoning pointed at vowels: consonants, he believed, could only be pronounced alongside them. Centuries of language study have complicated that story considerably. What that journey reveals about the sounds that shape all human speech is what this documentary explores.
At its most basic, a consonant is produced with complete or partial closure of the vocal tract. When you say p or b, your lips seal completely. When you say f or s, air is forced through a narrow channel, generating friction. When you say m or n, the breath reroutes through the nose entirely.
One sound sits awkwardly in this picture: h. It is classified as a consonant, but unlike every other consonant, it is pronounced without any stricture in the vocal tract at all. It is the exception that sharpens the definition of everything around it.
Beyond the mechanics of closure, each consonant carries several distinguishing properties. The manner of articulation describes how air escapes: stops block airflow entirely, fricatives squeeze it through a narrow gap, and nasals redirect it. The place of articulation marks where the blockage happens. Bilabial consonants like p use both lips; alveolar consonants like t press the tongue against the gum ridge; velar consonants like k engage the back of the tongue against the soft palate.
Phonation adds another layer. When the vocal cords vibrate fully during a consonant, that sound is voiced; when they stay still, the consonant is voiceless. That single difference separates b from p, d from t, and g from k. The timing of when the cords start vibrating, called voice onset time, is itself a phonetically measurable feature, and aspiration is one of its properties.
Length is a feature that English speakers rarely notice as meaningful, but in Italian, Japanese, and Finnish, the difference between a short consonant and a long one, called geminate, changes the meaning of a word entirely. Estonian and some Sami languages go further still, distinguishing three phonemic lengths: short, geminate, and a long geminate.
In a syllable, the most sonorous part is the nucleus, which is typically a vowel. Consonants occupy the margins: the onset before the vowel and the coda after it. This CV pattern, consonant followed by vowel, has been argued to be the only structure found in most of the world's languages, and perhaps the primary structure in all of them.
But the border between consonant and vowel blurs in ways that have kept phonologists debating for generations. Segments called semivowels, semiconsonants, or glides fall in a genuinely ambiguous zone. The j in the English word "yes" behaves like a consonant in forming the onset of a syllable, yet its articulation closely resembles a vowel. Some phonologists treat the j in "yes" and the j in "yield" as underlying instances of the vowel /i/, with the word "bit" phonemically as /bit/ and "yield" as /i̯ii̯ld/.
Syllabic consonants create a different puzzle. In Czech and in languages of the Democratic Republic of the Congo, fricative and trilled segments occupy the nucleus of a syllable, the slot that normally belongs to a vowel. Peter Ladefoged and Ian Maddieson called these segments "fricative vowels," describing them as phonetically consonants that phonemically behave as vowels. In Mandarin, these sounds are historically allophones of /i/ and are spelled that way in Pinyin.
In Nuxalk, a Salishan language, it is genuinely difficult to identify the nucleus of a syllable or even to confirm that all syllables have one. A sequence like /sx̩s/, meaning "seal fat," illustrates why. Miyako, spoken in Japan, presents similar challenges, with forms like /f̩ks̩/, meaning "to build," and /ps̩ks̩/, meaning "to pull." Many Slavic languages allow the trill r̩ and the lateral l̩ as syllabic nuclei, as documented in the phenomenon sometimes called "words without vowels."
Dionysius Thrax shaped the vocabulary of consonants for more than two thousand years. Writing in Classical Greek, he divided consonants into two groups: hēmíphōna, meaning "half-sounded," which covered continuants, and áphōna, meaning "unsounded," which covered plosives.
His term sýmphōna passed into Latin as cōnsonāns, meaning "sounding together," the direct ancestor of the English word "consonant." It was a calque, a translation built piece by piece from Greek into Latin.
But his foundational claim that consonants can only be pronounced with a vowel does not hold across all languages. In the Salishan language family, plosives occur without any accompanying vowel. The modern concept of "consonant" has since moved away from requiring co-occurrence with a vowel altogether, reflecting the full range of sounds that human languages actually produce.
The International Phonetic Alphabet emerged from a later problem: the number of speech sounds across the world's languages far exceeds the number of letters in any single writing system. English illustrates the mismatch neatly. The alphabet has fewer consonant letters than the language has consonant sounds. Digraphs like those for the sounds in "this" and "thin" patch the gap, but even those overlap in unpredictable ways. In the IPA, the two sounds in "this" and "thin" are unambiguously ð and θ, respectively, each with its own symbol and no confusion possible.
Five consonants come closer to universality than any others: the three voiceless stops /p/, /t/, and /k/, and the two nasals /m/ and /n/. Nearly every language on Earth includes them. But "nearly" hides a catalogue of exceptions that reveals how flexible the human vocal tract can be.
Arabic and several other languages near the Sahara Desert lack /p/. Mohawk and some other North American languages lack both /p/ and /m/. The Wichita language of Oklahoma lacks /n/ at the phonemic level, though it does appear phonetically as an allophone of /ɾ/. On Bougainville Island and around Puget Sound, languages such as Makah lack both nasals m and n entirely, except in special speech registers like baby-talk.
Ubykh, with its 84 consonants, lacks the plain velar /k/ in native words, a gap it shares with the related Adyghe and Kabardian languages. The Pacific Northwest coast shows a regional pattern where historical /k/ became palatalized: Saanich, for instance, has /tʃ/ and /kʷ/ but no plain /k/. The same shift happened in the Northwest Caucasian family, where historical /k/ became /kʲ/ in Ubykh and /tʃ/ in most Circassian dialects.
At the other extreme, Xavante and Tahitian have no dorsal consonants whatsoever. The Nǁng click language lacks /t/. Colloquial Samoan lacks both alveolars, /t/ and /n/. Nearly all Australian languages lack fricatives entirely. Despite these gaps, most languages of the world include at least one fricative, with /s/ being the most common, and the approximant /w/ is widespread. Among consonant types, the most commonly used in rhotic American English are /n/, /ɹ/, and /t/, in that order.
Most consonants in most languages run on the same engine: air pushed out of the lungs by the diaphragm. These are pulmonic egressive consonants, and they account for the vast majority of sounds across the world's languages.
Very few natural languages operate outside this system. Some use ejectives, in which air trapped above the closed glottis is compressed and released. Some use implosives, where the larynx drops to create a different pressure differential. And some use clicks, sounds produced by manipulating air in the mouth without involving the lungs at all.
Clicks are most famously associated with languages of southern Africa, including Taa, whose click-rich consonant inventory reaches 87 or more. The Central dialect of Rotokas, at the opposite end of the spectrum, manages its entire phonological system with just six consonants, none of them clicks, ejectives, or implosives. The range between Rotokas and Taa represents the full documented breadth of what human languages have chosen to make distinctive in their sound systems, and linguists have developed the full apparatus of the IPA specifically to map that range without ambiguity.
Common questions
What is a consonant in linguistics?
A consonant is a speech sound produced with complete or partial closure of the vocal tract. Most consonants are pulmonic, using air from the lungs and diaphragm. They contrast with vowels, which form the nucleus of a syllable.
Which language has the most consonants in the world?
The Taa language has the most, reaching 87 consonants under one analysis and 164 under another. The recently extinct Ubykh language had 84 consonants alongside only 2 or 3 vowels.
Which language has the fewest consonants?
The Central dialect of Rotokas has the smallest number of consonants in the world, with just six. It is spoken in Papua New Guinea.
Where does the word consonant come from?
The word consonant comes from the Latin cōnsonāns, meaning "sounding together," which was itself a calque of the Greek sýmphōnon. The Greek grammarian Dionysius Thrax coined the Greek term because he believed consonants could only be pronounced alongside vowels.
What are the most universal consonants across the world's languages?
The three voiceless stops /p/, /t/, and /k/, and the two nasals /m/ and /n/ appear in nearly all of the world's languages. However, even these five are not completely universal: Arabic lacks /p/, and languages such as Makah lack both /m/ and /n/.
What is the difference between voiced and voiceless consonants?
Voiced consonants are produced when the vocal cords vibrate fully during articulation; voiceless consonants are produced when the cords do not vibrate at all. This distinction separates pairs such as b from p, d from t, and g from k.
All sources
4 references cited across the entry