By using a systematically empirical angle on metaphors, I shall demonstrate how metaphors really work: they are predictable, well-defined, and they do not tell us anything new at all.

Metaphors are often seen as unpredictable and as man's creative, innovative and intuitive power. My aim is here to show the opposite: in a survey of 6,000 computer metaphors it is shown that metaphors in specialised language are very predictable and have functions that do not tell people anything new about technology. Metaphors do not tell us anything new - we are not interested - metaphors always tell us what we already know.

Based on my empirical survey I shall demonstrate that metaphors are always based on our common sense interpretation of the world: it is based on our experiences from our everyday life. But metaphors are not indigenous to everyday language, as often supposed. Metaphors are indigenous to specialised language, in other places they are special occurrences - a statement, which is in opposition to well known and often discussed cognitive linguistics. Metaphors are always extensions of everyday life, and they do not differ very much in the way they appear in different environments.

I shall demonstrate how metaphors work, how they help us see unfamiliar artifacts and occurrences as familiar and as everyday events. I shall show how they do not tell us anything new. How metaphors work is predictable, a statement I shall exemplify. Metaphors are not so creative as supposed, they are not manifestations of man's innovative thoughts and of man's intuition.