One thing to keep in mind is that in the early days of computers (as we know them anyway) the registers in a CPU were very small. In fact, the CPU that I first learned assembly programming on was an Intel 8088 and the registers in that chip were only 8 bits wide. Not to mention there were only 4 registers that you could really use for processing at one time, so the reality was that a byte couldn't be any more than 8 bits wide and you couldn't process but one bit in a register at a time, or have more than 4 bytes in the main CPU at once. (In reality you could, but you could get yourself into trouble really easily, so that's a topic best left for another day)

As CPU's have developed they have increased in size from the 8-bit models that I first worked with to the current 64-bit to 128-bit models that are buzzing around today. With the expansion of the width of registers in a CPU came an increase in the the number of registers. So the increase in the width of the registers has allowed the number of bits used to make up a piece of data to increase. Take Java's unicode representations for instance. All characters are represented as 16 bits in Java. The current ASCII representations are accounted for in the lower 7 bits, but by increasing the number of bits available Java is able to use all 16 to represent virtually every character in every language.