Why is Supersymmetry Necessary in String Theory?

["pi"mp]

I was recently watching a talk by Witten and he mentioned that one of the magical things about string theory is that it forces us to accept certain symmetries of nature, as opposed to choosing them as we do in QFT. Can anyone give an enlightening explanation of this? I do have very basic knowledge of string theory but not enough over-arching knowledge to appreciate that claim.

 

[atyy]

Perhaps he is referring to supersymmetry. I don't know if there is a proof that it is absolutely necessary, but bosonic string theory has an unstable vacuum, whereas superstring theory does not.

http://www.superstringtheory.com/experm/exper4a1.html
"The biggest problem with bosonic string theory (aside from the lack of fermions) is that the lowest energy state was a tachyon, or a particle mode with negative mass squared. This means the vacuum state of the theory is unstable.

In the mid-seventies Gliozzi, Scherk and Olive realized that they could implement a rule to consistently discard certain states from the RNS model, and after this truncation, known as the GSO projection, was made on the string spectrum in ten spacetime dimensions, the ground state was massless, and the theory was tachyon free."