Historically
Dr. Victor Zue, at MIT, has said that context cannot be programmed, at least now. At the earliest, he feels we can expect a 50-year wait. Also at the Massachusetts Institute of Technology, Dr. Noam Chomsky looked for a monogenesis, a deep structure "architecture," a grammatical core to all language. The notion of a language function at the heart of all language, a human similarity that we could capture as a "magic" key to language and to computer emulation of language, captures the imagination. With a core technology for "sorting" words and thoughts, for translating accurately, for reducing language to "chunks," or strings, computer technology would begin to look more like human functions. The history of this computational search begins with Mikail Bongard in the 1970s, Wittengenstein in the 1940s, Chomsky, Zue, and, more modernly, almost everybody.
But, the search differs from one expert thinker to another. Zue refers to the context of words. Chomsky refers to grammars. Grammars too, however, usually isolate words: the right word or its spelling in particular cases. For instance, do you properly say "I" or "me" and when? Which grammatical choices create the right, correct (prescription) choice and how, therefore, does grammar create meaning?
For Zue, the Microsoft sentence might illustrate.
"The spaceship photographed Seattle flying to Mars."
Context programming could, as would "common sense," respond with something like: "Sentence violation; Seattle cannot fly." The context of how a word's meaning fits into a "logical" sentence makes determining multiple meanings very difficult. The idea that computer language can emulate "common sense" has framed a great deal of the development work.
Currently
Today's literature is filled with discussion about natural-language programming. The National Institute for Standards and Technology (NIST) seeks high-risk solutions for natural-language functions. But the single-word paradigm still prevails. Search-engines, resume text-search, work-products, consistent medical text processing and billing, web search, voice functions, machine translation, or any technology which interfaces with human language run straight into "walls" when language functions move past single-word tagging.
theExact Word's® Patent
The fixed word-order solutions provide new directions for existing word-based, language-based problems. Solving these problems without semantic requirements allows a new paradigm for integrating solutions developed to date with word-order, context-based language strings.
Go to the Heart of the Matter--Free
Text Search
Go to the Brainy
Read Free--Text Search