Forgive me for questionining the fundamentals and this question therefore being not very smart, but I am currently beating my head against the concept of sound generation itself.
I understand so far that every sound is generated as a wave and therefore it can only be done by sinus waves and not just passing single constant values to the soundcard. Instead you have to generate the wave at your samplerate and bitrate with frequency and amplitude and then pass that collection to your card to actually get that sound.
So then, why does the square wave make a sound? It's not a wave, it is exactly that which you should not do, namely just alternating fixed values at different frequencies to get different sounds. That's not a wave at all. Is it because square waves are just a mathematical construct and are not possible in nature, so you just get the sharpest waves possible with your hardware?