A 5-layer model and techniques are proposed to hide information in text documents which are less redundant than images, video, and audio. The model consists of presentation layer, encrypt layer, carrier-content layer, carrier-code layer and carrier-format layer. At the presentation layer, the message to hide is encoded to unreadable symbol sequence, then encrypted at the encrypt layer. At the carrier-content layer, the output of encrypt layer is embedded in the carrier syntactical, semantic and pragmatic planes respectively. Then the output of encrypt layer can be embedded in multiple carrier coding at carrier-code layer and in diverse carrier formats at carrier-format layer. The layers are independent, each of them employs own control message and data as parameters, and executes different operation. All layers are optional, one or more can be skipped. The hiding techniques and approach at each layer are introduced, and a prototype system is implemented for the Microsoft Word document.
