Suppose that we are given a bit pattern 1001011100011110, which is 16 bits long. What is this number? Well, it could be an ASCII 2-byte character string! Or it could be an unsigned integer, which would be 38,686. But it might also be a sign-magnitude integer, in which case it would be -5918, or a 1's complement which would be -26849, or a 2's complement, which would be ... Well, how do we find out what it would be?

The answer is quite simple. A given bit pattern may be interpreted in a variety of ways, and we need to be told somehow what the interpretation is before we can correctly use it.