The occurrence of spurious frequency components due to the presence of an input signal to the systemís ADC (analogue to digital converter) whose frequency was greater than 1/2 the sampling rate (SR).  

If S is the sampling rate and F is a frequency higher than 1/2 the sampling rate, then a new frequency will be produced FA where FA = S - F.  

If SR = 44.1 Khz and F = 28 KHz we will get a new signal (FA) of 16.1 KHz included in the sound file.

Aliasing is actually hard to achieve with 
new sound cards, etc. Although easy enough to hear with "digital" telephone answering machines, computer systems try to protect the user from the experience. This example was created using a 5000 Hz sampling rate and then sweeping the sine tone from 2400 Hz (below Nyquist frequency) to 9600 Hz. It is also an 8 bit sample so the timbre is a bit "gritty".  The original signal was a 2400 Hz sine tone.  All the other tones derive from the foldover or aliasing as the sampling rate effectively  becomes less than twice the input frequency.