The 4K notch is a common audiological phenomenon in which there is a dip in a person's hearing sensitivity centered on or near 4000 Hz. This dip is caused by damage to the hair cells in the cochlea, and thus presents as sensorineural hearing loss.
This often occurs due to noise exposure, and in the past was said to be specifically indicative of noise exposure. However, it appears that this effect can be caused by almost anything, including viral infections, head trauma, and perilymph fistulas.
While one should be careful about blithely diagnosing noise-induced hearing loss simply because of a 4K notch, given the frequency with which hearing loss is caused by noise and the fact that a noise-induced hearing loss should have an identifiable cause, it is often fairly safe to make this diagnosis. One feature of the 4K notch is that it is most often caused by broadband noise, such as large machinery or firing a gun (interestingly, audiologists can often identify hunters from their audiograms because they are 1. young men with 2. a 4K notch 3. that is more severe in one ear than the other).
This notch has been known of since at least 1860, when Joseph Toynbee noted what he called a diminution in hearing of the 5th fork -- refering to the fifth tuning fork in a set, corresponding to the fifth octave above middle C. This octave, in today's measurement, falls at 4096 Hz. This was also sometimes referred to as the C5 dip.
Despite the name, the exact location of the notch is somewhat uncertain, and is frequently observed anywhere from 3000 Hz to 8000 Hz. Somewhere in the range of 4000 is common, but the frequency of reporting a 4 kHz notch is in large part because the usual audiometric test will only target 1, 2, 4, and 8 kHz (and some lower frequencies), meaning that a dip actually centered on 6000 Hz will show up as a dip centered on 4000 Hz.
Despite this being the focus of no small amount of research, we have no good idea why this area of the cochlea is particularly prone to damage. It may be that the structures of the ear either 'artificially' amplify sounds in this range, or that the structures in the cochlea where the 4K waves are the strongest is for some reason weaker than the surrounding areas. One popular theory is simply that the hair cells responsible for sensing 4000 Hz happen to lie just past the first curve in the cochlea, which means that the sound waves are hitting them first and before the dampening that protects later curves. Regardless, the damage seems to be most severe to the outer hair cells. A similar effect has been found among most, if not all, mammals.