International Phonetic Alphabet

In Paris in 1886 a small group of teachers and linguists formalized a meeting of like-minded educators into what was soon called in English the "International Phonetic Association" or IPA. This group created a standardized format for expressing the phonetic sounds used in the various spoken human languages. This format also became known informally as the "IPA," or "International Phonetic Alphabet." It is an international method for expressing in written form the sounds of spoken words.

Increasingly the IPA is also becoming the single, universally-accepted method of making phonetic spellings, even though some dictionaries still cling to older or idiosyncratic systems. The coming of the computer age has helped to advance standardization across languages, as the symbols of the IPA have Unicode representations, thus giving users of Latin and non-Latin alphabet computers alike a simple recourse to the phonetic symbols.

Organic Units of Sound

The "phoneme" is the linguistic equivalent of an atom - the smallest division of human sound that still results in distinguishable speech. The "phone" is the sound produced by a speaker. The "phoneme" is the sound the speaker's brain hears and requests from the speaker. They are often the same thing, but not always. A person who speaks with an accent, but does not perceive it as such, may be "hearing" phonemes in his head and producing "phones" through the mouth that are somewhat different from one another. This subjective aspect of linguistics is presumed not to be present when considering the recordation of speech sounds using the IPA, and phones and phonemes are taken to be the same thing.

Thus, consonant and vowel sounds are phonemes. Some phonemes combine with others to create "diphthongs," which may be regarded in a given language as a single unit of sound expression made up of more than one sound. Sometimes they are official diphthongs, as in the name Oedipus. Sometimes they are happenstance, as when a person with a Southern accent says "I am" in a form that is closer to "Ah Ay-um." The vowel sound in "am" has been pronounced as a diphthong of a long "a" and an "uh".

The IPA comprehends phonemes as well as intonation and the separation of sounds into separate syllables (or words). Diphthongs are usually spelled as a combination of IPA characters, but they may be represented by single IPA characters, modified by diacritical markings of some sort.

Human language contain sub-phoneme occurrences called "allophones." In the metaphor, they might be sub-atomic particles, like electrons. They are separate occurrences of distinct pronunciations of the same phoneme. For example, the word "dart" is pronounced in standard American English with a discernable "r" sound. In Boston, the "r" disappears altogether, and the word sounds like "dot" (or maybe "daht") to the non-native ear. (The native speakers can tell the difference, however. It's another instance of the disconnect between phonemes and phones.) In the upper Midwest, on the other hand, the "r" in "dart" seems to dominate the word. It is prolonged and rounded, perhaps even savored until the explosive "t" finally brings it to an end.

In Japanese, "r" and "l" fall into the same phoneme to the native speaker's brain, explaining why Japanese speakers of English confuse "r" with "l" from time to time, or perhaps more correctly, pronounce them in the same way. The native English speaker may find it incredible that the Japanese ear does not hear a difference between the two letters, and the native speaker of Japanese finds it amazing that every English speaker can hear a distinct sound. A similar example comes from Swedish, in which (for example) the word "seven" ("sju") contains a consonant - represented by "sj"-- that Swedes assert will always unmask a foreigner, no matter how good his Swedish might be. If you are not a native speaker of Swedish, they say, you will not pronounce "sju" correctly. These "sub-atomic" sounds, these allophones that are beneath the surface of the phoneme, can often be precisely described in the IPA, but usually do not need to be. Such precision is usually only necessary for therapeutic or scientific applications.

Extensions to the IPA comprehend yet more sounds found in human communication, like clicks, pops, and the gnashing of teeth. They describe non-vocal sounds as well as several pathological speech patterns (such as lisping and the nasalization of a cleft palate). The extensions to the IPA are rarely employed in translation or language study, and only when the language in question contains a non-vocal sound.

How the IPA is Organized

The IPA has changed several times since its early formulation at the end of the 19th century. The last major revision to the core IPA was in 1986, with minor changes since then. The extensions to the IPA were added in 1990.

The core IPA contains 107 letters, plus 56 prosody marks and diacritics. A diacritic is a mark, like an accent mark, cedilla, tilde, hacek, umlaut or stroke through a letter, to cause it to be pronounced differently from the way the same letter would be pronounced without it. "Diacritic" derives from the Greek word for "distinguishing." Prosody is the quality of length, rhythm, stress or intonation of a letter or syllable. "Prosody" derives from the ancient Greek word for "words sung to music."

A common convention employed in using the IPA is to separate sounds with slashes or brackets of different kinds, rather than quotation marks or some other punctuation. This is to avoid confusion, although the usage is not always consistent. In general, phonemes are enclosed in slashes, phones in square brackets. For example, "keep" is pronounced as [khip], where the superscript indicates aspiration of the { k } sound. This is a reproduction of the phone. Native speakers of American English automatically aspirate the { k } in "keep." If you look it up in the dictionary, however, the phonetic spelling (i.e., the phoneme) will be /kip/ -- without the aspiration. That's because the aspiration is descriptive of how the sound comes out, but not definitional for the word. The conceptual sound of that word is /kip/ -- without recording any other nuances. Be a bit cautious, however, as not all usage of slashes and brackets will follow the convention consistently (though it should!).

Another way to look at this is that words in square brackets are precisely transcribed according to what is actually said - the phone - as in [lAV] for "love"). Words are found in slashes if they are more generally described or described according to how they are pronounced in concept (as in the dictionary phonetic spelling of "love" as /lAV/. When the phoneme and the phone are the same, the slashes and brackets will have the same content.

Angle brackets (for example, { A } for the broad "ah" sound) are used to set off specific phonetic letters and symbols, referring to the sound they represent, and not their name in English or some other language. So { ɛ } is the sound of an epsilon, not "ɛ", which would be name of the character epsilon.

These conventions are then organized into various "charts" for individual languages and dialects. Each chart specifies the sounds required by that language to form its component words, and each such sound is then specified in terms of the IPA phonetic system. The orthography (i.e., the alphabet employed in writing the language) is then "translated" into IPA characters to approximate how they are sounded.

The language charts also provide one or more diagrams, based on a schematic concept of the mouth in profile, indicating where each sound is generated, particularly the vowel sounds. Often more than one vowel sound can be defined for each letter representation in that language's alphabet. (For example, linguists identify approximately 24 "lexical sets" of vowels in English, indicating (absent dialects and special accents) 24 discrete vowel sounds which the five vowel letters and their combination have to convey. Even so, by no means is every possible IPA consonant and vowel found in English or in any other language. English requires only about 22 of the IPA symbols to express all the consonants it requires for its language chart.

The ultimate objective of the IPA language charts is to have one unambiguous phonetic symbol for each distinct sound in the "standard" articulation of a language, regardless of context or the number of letters (or diacritical marks) used in the language to represent them.