From the conclusion of [1]:
Moreover, a significant number of people still hold to the view that early australopithecine bipedalism was fully human-like. I have often felt there is a bias in favor of viewing early hominid bipedalism as characterized by completely extended lower limbs because it is difficult for modern humans to walk with bent knees and hips. It seems inconceivable that such a manner of progression could last for more than the briefest of geologic times before evolving into our superior way of doing things. Returning to my simplistic analogy to cetacean evolution, I think if we were whales we would have great difficulty understanding how an ancestor could survive a million years while being such a poor swimmer. I have tried to overcome this bias. Along with others, I believe the bipedal adaptation first arose to improve access to food sources close to the ground, movement between such sources, or both. Bipedalism probably persisted in this nascent but effective state for a million years, with no indication that it would be anything other than an evolutionary sidelight. Only later did some unknown event impel one of the creatures with this adaptation to abandon the trees more completely than any of its predecessors had done and become a tool-making hunter or tuber-gatherer.