Processing Text as Spoken Word