Models as Metaphors
In classical linguistics a metaphor is a figure of speech that explains one thing in terms of another. The cognitive linguist George Lakoff revolutionized contemporary linguistics when he proposed that metaphor is "an ontological mapping across conceptual domains. Metaphor is not just a matter of language, but of thought and reason. The language is secondary; the mapping is primary."
Data models are like metaphors because they map one domain onto another. In fact, the logical data model is mapped in two directions.
- We use the logical model going downward to design the physical database so that it can record significant facts about the operations of the enterprise. In this direction the data model is a metaphor for the physical model and database. This is the traditional use of data models in systems development projects, and the goal is to map the business data requirements onto physical data stores.
- We use the logical model going upward to map the complex ideas implicit in the logical model onto a simpler conceptual model. The conceptual model can then be used for system integration by helping to understand how one application or business area fits into the larger enterprise.
A model is a visual metaphor. Instead of using figurative language to explain one thing in terms of another it uses a combination of visual shapes, colors, captions, and relationship symbols to map data requirements onto the physical or conceptual domains.
It may sound strange to talk about visual metaphors but that's only because we are still thinking of the classical definition of metaphor as a part of the figurative language of poetry and literature. In order to see just how common visual metaphors have become one need only look at examples from commercial advertising. Visual metaphors have become so common that the conceptual mapping is almost subconscious.
Consider the following Heinz Ketchup Advertisement. The image of a ketchup bottle sliced up like a tomato clearly expresses the message that Heinz ketchup is as fresh as a tomato.
The purpose of a data model is not just to show a graphical image of the relational data tables. The ultimate purpose is to map the business requirements onto the physical database so the application systems can generate meaningful information. The data model must bear a likeness to the cognitive models in the minds of the business people who analyze the various states of the business environment. Data models are visual metaphors for the purpose of sharing a body of factual statements about some situation or state of affairs. The visual image of the boxes and lines is not as important as the mapping between the business requirements and the functioning database in the information system.
Data models show their meaning without explicitly telling it. The data model diagram lets us take in the meaning of the model in one gulp. In that way the model is mystical; its meaning is directly apprehended. Of course the data model can generate the explicit mappings needed to create the physical database but this is done behind the scenes. Comprehension of the data model appeals to our mystical side.
 George Lakoff and Mark Johnson, "The Metaphorical Structure of the Human Conceptual System," Cognitive Science, Vol 4, No. 2 (1980), pp. 195-208.
 Andrew Ortony, Metaphor and Thought, New York: Cambridge University Press (1993), p. 208.
# # #