In "Reading Images: Multimodality, Representation and New Media" Kress discusses the fact that words have different meanings in different societies and cultures, but that pictures have direct meaning because they are visual. He uses the example of looking for food in an American airport, where he sees the sign for Bar & Grille. Grill represents fresh food and grilled food, and bars in America represent alcohol. Writing can only represent so much, for example: reading someone say "you did what?" in a tone of incredulity as opposed to "you did what?" she said incredulously. Words can only be used to describe things for which words already exist. To draw something from our imagination we would have no words to describe it if it were totally new. A picture can show the object exactly as we see it in our heads.
I agree with Kress when he says that "Words are (relatively) vague, often nearly empty of meanings; by contrast images are full, ‘plain’ with meaning" because often there are words with more than one meaning. In one part of the world a friendly greeting could mean something completely disrespectful in another part of the world. Things that you dream up in your imagination might not have words to describe them, and thus no way to tell someone else about them. The only way to do that would be to draw a picture. In The Wild Things by Dave Eggers, he draws pictures of the monsters in the margins to show what they look like because words don't do them justice.