San José State University
Thayer Watkins
Silicon Valley
& Tornado Alley

The Basic Structure of Language

Phonetic alphabets are wonderful linguistic inventions. They made literacy far, far easier than the symbolic character systems of writing such as the Chinese characters. For example, the Hungûl system of Korean is known as the morning alphabet because a speaker of Korean can in one morning learn the letters and, because Korean words are written perfectly phonetically, becomes literate. Imperfectly phonetic spelling, as for English, require a much longer time to get acquainted with the idiosyncratic forms.

The success of the phonetic alphabets have led people, including linguists, to presume that the basic building blocks of a language are the individual phonemes. Recent evidence coming out of attempts to create computer-generated speech indicates that this is not the case.

A team of linguists and engineers were engaged in a project to create a computer system for scanning documents and converting the files of characters into speech. The character recognition phase of the project was achieved and the team turned its attention to converting the character groups into spoken words. The strategy that seemed reasonable was to record humans reading text and cut out individual character sounds for assembly into words. For example, the recording of a word, say toad would be cut into [t], [o] and [d]. In theory then these could be reassembled as [d][o][t] for the pronunciation of the word dote. But the reassembled word was unintelligible. The reason was soon found. When a speaker articulates the word toad the speech mechanism anticipates the pronunciation of the [o] while articulating the [t] and anticipates the following [d] while articulating the [o]. Thus the pronunciation of phonemes is context-dependent. Thus the reassembly of the phonemes of toad in an attempt to pronounce dote fails.

Linguists went on to show that a listener can distinguish between the supposedly same initial consonants of different syllables in a small fraction of the time required to articulate that consonant.

What this means is the that the fundamental building blocks of a language are the syllables. It is to be noted that auxiliary modifications of the pronunciation of multisyllable words are of the syllable rather than the phoneme. In the case of the tonal language such as Chinese the tone is tied to the syllable rather than the phoneme. Thus the basic chunks of a language are the syllables. Some languages are written using a syllabary rather than an alphabet. Typically a syllabary will have one hundred to two hundred distinct characters rather than the thirty of a typical alphabet. This imposed a greater burden in learning to read and write a language. The reasonable compromise would be to have an alphabet but teach the pronunciation of syllables to children rather than starting with the individual letters. The emphasis on individual letters is almost always misguided because parents teach their children first the name of the letters rather than their supposed sound value; i.e., that h is aitch rather than huh.

HOME PAGE OF applet-magic
HOME PAGE OF Thayer Watkins