On Wednesday, the U.S. Patent and Trademark Office issued a new speech recognition patent to Apple Computer, Inc. This patent (U.S. Patent No. 7,127,394, entitled "Assigning meanings to utterances in a speech recognition system") includes 29 claims for computer-implemented methods, computer-readable storage media, and an apparatus pertaining to speech recognition. The first claim is a method comprising "determining a set of speech rules which match a spoken sequence of words by searching a current language model," the language model generated "from a plurality of speech rules according to a current operating context" and the words are received through an audio input. The claim also describes the plurality of speech rules as comprising "a language model and an expression; and evaluating said expressions in said current language model to assign a meaning to said spoken sequence of words." According to the patent, "prior art speech recognition systems suffer from many deficiencies that prohibit incorporating such technology into non-dedicated devices such as a personal computer," including the overall complexity of algorithms used for speech recognition, problems with accuracy, the requirement of a large vocabulary of words, dynamically adding words to the vocabulary, and transforming the spoken commands into data that can be used by the system. To potentially overcome these deficiencies, the disclosed technology within the patent is to "provide a means for associating meanings with spoken utterances in a speech recognition system" and "provide an improved method for associating expressions (e.g. actions and variable values) to speech rules in a speech recognition system." According to the Mac News Network Article (link below), "it's apparent that this patent could easily apply to a future iPhone."

U.S. Patent No. 7,127,394: LINK
Mac News Network Article: LINK