A century of research has described the development of walking based on periodic gait over a straight, uniform path. The current study provides the first corpus of natural infant locomotion based on spontaneous activity during free play. Locomotor experience was immense: 12- to 19-month-olds averaged 2368 steps and fell 17 times/hour. Novice walkers traveled farther faster than expert crawlers, but fall rates were comparable, suggesting that increased efficiency without increased cost motivates expert crawlers to transition to walking. After walking onset, natural locomotion dramatically improved: Infants took more steps, traveled farther distances, and fell less. Walking was distributed in short bouts with variable paths—frequently too short or irregular to qualify as periodic gait. Nonetheless, measures of periodic gait and natural locomotion were correlated, indicating that better walkers spontaneously walk more and fall less. Immense amounts of time-distributed, variable practice constitute the natural practice regimen for learning to walk.

We examined several factors that affect people’s ability to perceive possibilities for action. In Experiment 1, 24 participants crossed expanses of various sizes in three conditions: leaping, a familiar, launching action system; arm-swinging on monkey bars, an unpracticed skill that uses the arms rather than the legs; and crawling on hands and knees, a disused skill that involves all four limbs. Before and after performing each action, participants gave verbal judgments about the largest gap they could cross. Participants scaled initial judgments to their actual abilities in all three conditions. But they considerably underestimated their abilities for leaping, a launching action, and for arm-swinging when it was performed as a launching action; judgments about crawling, a non-launching action, and arm-swinging when it was performed as a non-launching action were more accurate. Thus, launching actions appear to produce a deficit in perceiving affordances that is not ameliorated by familiarity with the action. However, after performing the actions, participants partially corrected for the deficiency and more accurately judged their abilities for launching actions—suggesting that even brief action experience facilitates the perception of affordances. In Experiment 2, we confirmed that the deficit was due to the launching nature of the leaping and arm-swinging actions in Experiment 1. We asked an additional 12 participants to cross expanses using two non-launching actions using the legs (stepping across an expanse) and the arms (reaching across an expanse). Participants were highly accurate when judging affordances for these actions, supporting launching as the cause of the underestimation reported in Experiment 1.

“Cruising” infants can only walk using external support to augment their balance. We examined cruisers’ understanding of support for upright locomotion under four conditions: cruising over a wooden handrail at chest height, a large gap in the handrail, a wobbly unstable handrail, and an ill positioned low handrail. Infants distinguished among the support properties of the handrails with differential attempts to cruise and handrail-specific forms of haptic exploration and gait modifications. They consistently attempted the wood handrail, rarely attempted the gap, and occasionally attempted the low and wobbly handrails. On the wood and gap handrails, attempt rates matched the probability of cruising successfully; but on the low and wobbly handrails, attempt rates under- and over-estimated the probability of success, respectively. Haptic exploration was most frequent and varied on the wobbly handrail, and gait modifications—including previously undocumented “knee cruising”—were most frequent and effective on the low handrail. Results are discussed in terms of developmental changes in the meaning of support.