Vowel

• 1. Introduction

  ›

  Vowels are speech sounds produced without any constriction in the vocal tract, and they sit at the heart of nearly every spoken language on earth. Without them, syllables lose their center, words collapse into ambiguity, and the melody of human speech falls silent. But what exactly makes a vowel a vowel? That question turns out to be far less settled than most people assume. Linguists argue over two competing definitions, tongue position turns out not to be quite what early phoneticians thought it was, and languages around the world push vowels into territory that defies easy categorization. The word vowel itself comes from the Latin vocalis, meaning "vocal" or relating to the voice. That etymology points toward something real: vowels carry the voice, carry the pitch, carry the weight of a syllable. The story of how linguists have tried to describe and measure them is a story of careful observation repeatedly revised by better instruments. The cardinal vowel system developed by Daniel Jones and the International Phonetic Alphabet's forty-plus symbols for vowel qualities each represent attempts to pin down something that turns out to be surprisingly fluid. And at the edges of this story sit languages like Nuxalk, where words can be constructed from consonants alone, and Polynesian languages, where vowels flow freely one after another across long strings of syllables.

• 2. Two Definitions, One Sound

  ›

  Kenneth Pike, writing in 1943, proposed separating two ideas that had been tangled together for centuries. He coined the term "vocoid" for a phonetically defined vowel, meaning any sound produced with an open vocal tract and no significant air pressure buildup above the glottis. The word "vowel" he reserved for a phonologically defined sound: one that forms the peak of a syllable. That distinction matters because the two categories do not perfectly overlap. The sounds j and w, as in the English words "yet" and "wet", are produced without much constriction in the vocal tract, which makes them phonetically vowel-like by Pike's framework. But they appear at the onset of syllables, which places them phonologically among the consonants. Maddieson and Emmory, writing in 1985, measured the constriction of semivowels across a range of languages and found that j and w are produced with a narrower constriction than true vowels, offering articulatory grounds for treating them as consonants after all. The situation gets stranger with rhotic dialects of English. Words like bird, learn, and girl are analyzed by some phoneticians as containing no vowel at all, only a syllabic consonant /ɹ̩/. Others analyze the same words as containing a rhotic vowel /ɝː/. The English language, in other words, cannot fully settle this question even within a single accent.

• 3. Height, Backness, and the Tongue

  ›

  Peter Ladefoged offered a pointed correction that has been known since 1928 but is still not widely understood outside linguistics. Early phoneticians who described vowels in terms of tongue height believed they were documenting the highest point of the tongue. They were not. They were actually describing formant frequencies, the acoustic resonances of the vocal tract. That gap between perceived articulation and acoustic reality runs through the traditional three-parameter model that Daniel Jones built the cardinal vowel system upon: tongue height, tongue backness, and lip roundedness. Height, in the traditional model, describes whether the tongue sits close to the palate, as in the high vowels i and u, or low in the mouth as in open vowels like a. The IPA recognizes six degrees of vowel height for full vowels, plus the reduced mid vowel ə. The Kensiu language, spoken in Malaysia and Thailand, contrasts true mid vowels with both close-mid and open-mid vowels, a six-way height distinction that holds even for nasal vowels. The Bavarian dialect of Amstetten goes further, with thirteen long vowels analyzed across four distinct height categories. No other language is known to contrast more than four degrees of vowel height. Height is also the one feature that all studied spoken languages use contrastively. No other vowel parameter, not backness, not rounding, appears in every language. Backness describes the tongue's horizontal position: front vowels like i carry the tongue forward, while back vowels like u pull it toward the rear of the mouth. The IPA marks five degrees of backness, but no known language contrasts more than three independent of height.

• 4. Lips, Nasality, and Other Dimensions

  ›

  Lip rounding, the third of the traditional parameters, is visible in a way that tongue position is not, which may explain why it became central to vowel description so early. In most languages, rounding reinforces backness: the higher a back vowel, the more intense the rounding tends to be. French and German depart from this pattern by having front rounded vowels, and Vietnamese has back unrounded vowels. Japanese offers a particularly striking case: its /u/ uses exolabial rounding, in which the lips compress with the corners drawn inward, rather than the endolabial or pursed rounding typical of back vowels in English. Swedish and Norwegian are the only two known languages in which the distinction between exolabial and endolabial rounding is contrastive. Nasalization adds another layer. Vowels in English are often nasalized under the influence of neighboring nasal consonants, as in the word hand hæ̃nd, but these are not the same as nasal vowels, which contrast with oral counterparts as separate phonemes. French distinguishes /ɑ/ from /ɑ̃/, and Polish and Portuguese maintain similar contrasts. Phonation type opens yet another axis. Most languages have only voiced vowels, but several Native American languages, including Cheyenne and Totonac, have both voiced and devoiced vowels in complementary distribution. In Japanese and Quebec French, vowels between voiceless consonants are often devoiced. The Mon language uses creaky voice together with high tone, creating what is called a register complex, where the combination of phonation and pitch together signals a phonemic distinction. At the extreme edge of pharyngealization, the strident vowels of the Khoisan languages involve raising the larynx and constricting the pharynx so severely that either the epiglottis or the arytenoid cartilages vibrate instead of the vocal cords.

• 5. Acoustics and the Spectrogram

  ›

  Spectrograms made the relationship between tongue position and acoustic output visible for the first time in ways that early phoneticians could not have imagined. The vocal tract acts as a resonant cavity, and the position of the jaw, lips, and tongue shifts its parameters, producing different formant values that show up as dark bands on a spectrogram. The first formant, F1, tracks vowel height: open vowels like ɑ have high F1 frequencies, while close vowels like i and u have low ones. The second formant, F2, tracks vowel backness: front vowels like i show a much higher F2 than back vowels. In open vowels, the high F1 value forces a corresponding rise in F2 frequency, which is why some analysts prefer to plot F2 minus F1 against F1 to represent frontness. In the third edition of his textbook, Peter Ladefoged recommended that F1 vs. F2 minus F1 approach. By the fourth edition, he had switched to a simple plot of F1 against F2, and this remained his recommendation through the fifth edition. Katrina Hayward, comparing both methods, concluded that plotting F1 against F2 minus F1 is unsatisfactory because of what it does to the placement of central vowels. The simple F1-F2 plot has since been used to map vowel qualities across a wide range of languages, including RP, American English, Singapore English, Brunei English, North Frisian, Turkish Kabardian, and various indigenous Australian languages. Rounded vowels are identified acoustically by a lowering in F2, which reinforces their perceived backness. R-colored vowels, the rhotic vowels of American English, are marked by a distinctly lowered F3, the third formant.

• 6. Languages at the Extremes

  ›

  Nearly all languages have at least three phonemic vowels: classical Arabic, some Malayic languages of Borneo including Banjarese, and Inuktitut all use the minimal system /i/, /a/, /u/. A handful of Arrernte, Circassian, and Ndu languages have been argued to have only two vowels, /ə/ and /a/. At the opposite end, Standard Danish has eleven to thirteen short vowels, and classical French carries eleven vowel qualities plus four nasal vowels. Spanish, a closely related Romance language, manages with just five pure vowel qualities: /a, e, i, o, u/. The Jinhui dialect of Wu Chinese has been reported to have eleven basic vowels plus restricted /ɨ/, not counting its seven nasal vowels. Identifying which language has the most vowels is not straightforward, because long vowels, nasal vowels, and different phonation types may or may not be counted separately. Some languages use vowels in ways that challenge the assumption that they distinguish words rather than grammatical forms. In the Semitic languages, internal vowel changes mark inflection: English "man" becomes "men" in the plural, but that is an isolated case. In most languages, vowels serve mainly to distinguish separate lexemes. Words without any vowels exist across several language families. In the Miyako language of Japan, words such as ss for "dust" and ff for "a comb" contain no voiced sounds at all. Nuxalk, a language of the American Pacific Northwest, allows entire lexical words made of consonants: scs means "seal fat". Whole sentences can be built on the same principle in Czech: Strč prst skrz krk, meaning "stick a finger through your neck", consists entirely of syllabic consonants. On the other side, Hawaiian and Polynesian languages generally allow long strings of consecutive vowels; the Hawaiian word aeāea, the name of a small green fish, runs three syllables of pure vowels.

• 7. Writing Vowels Down

  ›

  The Hebrew alphabet and the Arabic alphabet do not ordinarily mark vowels in ordinary writing, because vowels are frequently unnecessary for identifying a word in those languages. These systems are technically called abjads rather than alphabets. The Masoretes devised a vowel notation system for Hebrew Jewish scripture that remains in use today, forming part of oral tradition and the basis for many Jewish and Christian bible translations. Latin script handles vowels differently, but imperfectly. The original Latin alphabet made no written distinction between the letters that became u and v; one letter represented both the approximant w and the vowels u and ʊ. Modern Welsh still uses this same letter to represent those same sounds. English spelling preserves the trace of a dramatic historical event known as the Great Vowel Shift. After printing was introduced to England and spelling was largely stabilized, a series of changes in vowel pronunciation continued into recent centuries without any corresponding update to the written forms. The result is the well-known inconsistency between English spelling and pronunciation, including the tendency of English speakers to mispronounce foreign words and names. Languages that lack enough vowel letters in the Latin alphabet cope through various strategies: Swedish uses modified letters like o-umlaut, some Scandinavian languages have constructed additional symbols by altering standard Latin vowel letters, and many languages use combinations of letters to represent sounds that no single letter can capture. The IPA, by contrast, provides twenty-eight symbols for essential vowel qualities, along with a further set of diacritics for denoting variations, making explicit what writing systems built for everyday use have always left partially implicit.