“Speech” is a high-reliability data-transfer system that employs data redundancy, compression, disambiguation, and collision detection techniques to transfer information between participants in the protocol.

At its lowest level, it is composed of “phonemes” which are combinations of fundamental tones in low (200-800Hz) and high (2000-8000Hz) frequency ranges and specific sets of overtones. These may be transmitted over any suitable medium, typically a gas-filled chamber or “Room.” Transmission over other media are not covered in this application.

“Phonemes” are connected according to certain “articulation rules” to form “syllables,” “morphemes,” “words,” and “sentences.” Morphemes are subsets of words. They can be standardized to represent certain kinds of meanings such as the function a given word plays in a sentence. Grammatical rules can govern how morphemes interact with words, and how words interact with each other, ensuring data-redundancy and reliability of transmission.

Phonemic and morphemic similarity between words can be exploited in series of sentences to achieve pleasing tonal and emotional effects. Moreover, the timing of syllables, combined with specific fundamental frequencies, can exploit alternate channels to enhance the communicated data bandwidth. Ultimately, several transmitters of “Speech” could theoretically emit data simultaneously for additional data transmission channels.

Multiple sets of rules for encoding data into “speech” could be defined. These need not be mutually intelligible.

The environment for speech is a chamber filled with a suitable transmission medium. All entities that wish to participate must be present. When one entity wishes to transmit, it should first monitor the medium to ensure that no other entity is transmitting. When the transmission medium is clear of other communications, the entity may transmit. If two entities transmit at the same time, both must stop, wait a randomly determined amount of time, and then begin the protocol again.

Modifications to this protocol may be adapted in situations where certain entities have a higher priority for transmission than others. Such entities may choose to increase the intensity of their transmissions, deliberately wait a shorter time to begin retransmission in case of message collision, or ignore the protocol altogether and go on transmitting.
Error correction is incorporated into the protocol. When a receiving entity senses an error in a received transmission, it may transmit an agreed-upon morpheme. The transmitting entity should retransmit the message, preferably in an alternative format, thus ensuring that the same error in transmission or reception is not repeated.

Phonemes can be encoded and represented as graphical symbols or “glyphs” which can then be arrayed linearly and in other arrangements to represent their order and organization. (This application will be covered in the separate patent application for “Writing.”) }