A rotary encoder consists of three main parts: light source, optical encoder disc and light sensors. The light source provides the light source to several light sensors. There is an optical encoder disc that selectively blocks light from light source to light sensor. If the light sensor is receiving light, it is first state of electronic circuit. If the light from light source to the sensor is blocked by optical encoder, then sensor is not receiving light and this is second state of sensor output. Therefore, each sensor has two ouputs possible, high or low. Output from one sensor is called a "bit" so for 4 sensors, it is 4 bit output.
The light encoder disc is a plastic disc with some area opaque and other areas transparent. Depending upon its rotating position, it passes light to some sensors and blocks light to other remaining sensors.
The combination of states of multiple light sensors at a given time due to a certain rotary position of rotary encoder disc is the multi-bit output of that rotary encoder. This multi-bit logic output is used to sense the position, direction of rotation and rotating speed of the rotary encoder.
You can read more about rotary encoders at the links given.