This paper describes a document preparation system with speech commands which uses the EDR Electronic Dictionaries. The system employs a multimodal user interface, which allows keyboard inputs for writing documents and speech inputs for commanding the system to retrieve linguistic information from the EDR electronic dictionaries. The system consists of three components: a text editor, a speech processor, and a linguistic information retriever. It uses the EDR electronic dictionaries as a source of large scale linguistic knowledge that is referred to by the user preparing documents.