Here's the quick and dirty explanation....
Sounds are created by vibrations...the audible spectrum for humans ranges from approx 20Hz to 20000Hz. A Hz is one cycle per second, meaning at 20 Hz your speaker must ocillate 20 times a second.
Now, the perfect speaker would be able to reproduce all audible frequencies by itself. This type of speaker doesn't exist, so the audible spectrum must be cut into bits and pieces and then applied to the proper speakers....the sum of which should be able to reproduce the full audible spectrum.
A High pass filter removes the low frequency information and passes only the higher frequencies, while a low pass filter only allows the lower frequencies to pass through.
Filters are not absolute...meaning, for example, an 80 Hz high pass filter doesn't just pass 80Hz and up. All filters have a slope. The frequency response through the filter continues past the crossover frequency, but "rolls off" meaning the filter attenuates the frequency past the crossover point.
The filter's slope is listed in dB/octave. Most slopes start at -6 and continue down to 64 or more dB/octave....the most used values are often -6, -12, -18, -24 and so on dB/octave. As the slope increases, the filter is deemed to be more steep. Thus a -6 dB/octave crossover is not as steep as a -24 dB/occtave crossover.
I hope that clears things up...hope I was not too elementary for you. Anyone feel free to correct any inaccuracies as it's late here.
For more info:
http://www.eatel.net/~amptech/elecdisc/