According to the four-ears model of communication, when we talk we transmit four messages and, when we listen, we can hear with four different ears (thus, of course, why the model is called four-ears). I believe the number of ears and messages are more than four, but this model nevertheless can be very helpful in improving understanding.
(Click the image above to enlarge the four-ears diagram.)
- Content (What are the facts and data I'm stating)
- Self-disclosure (What i'm revealing about myself)
- Appeal (What I'm seeking from the other person)
- Relationship (How I relate to, think about, and regard the other person)
Because there are four kinds of ears with which we can listen, and four beaks (that's what Schulz von Thun called the output part of the communication) with which we can transmit, the potential for miscommunication can be very high. In fact, if you were only to consider the single channels, any communication has 16 different combinations! My statement of a fact can be heard by any of your four ears, as can my communication of self-disclosure, appeal, and relationship.
Here's an example from Wikipedia which I have edited a bit for clarity:
The classic example of Schulz von Thun is the front-seat passenger who tells the driver: "You, the traffic lights are green". The driver will understand something different according to the ear with which he hears. With the content ear, he will understand the "fact" that the traffic lights are green. He could also understand it as "Come on, drive!" command, the appeal ear. Or in the relationship ear he could hear a helping message like "I want to help you." Or the listener may hear "I am in a hurry" [which] reveals something about the speaker--self-disclosure ear. The communication on the four layers can be meant differently and also be understood differently. So [for example,] the sender can stress the appeal of the statement and the receiver can mainly receive the relationship part of the message.
There are many communication models of which this is only one. I like using models such as this for a