I investigated many other languages before settling on Smalltalk. I even prototyped a few new ones.
They all had their pros and cons, but no other language was as concise, as flexible and well-documented. No other practical system came anywhere near close to Smalltalk in terms of stability and proven utility.
Of course, there are things I don't like about Smalltalk, but they're only cosmetic details. The system itself can't really be improved upon in any meaningful way - it's so flexible that any annoying detail can easily be overcome within the system without having to break compatibility.
Frankly, a better question would be "why did anybody else bother trying to fix a system that was already working perfectly 37 years ago?" - and the answer to that would be "to sell you a shiny new lemon!".