Why do we swing our arms when we walk? This seems like a simple enough question, but if you were to ask two locomotor biomechanists you might get two different answers, both of which sound pretty good. For example, one might answer that arm swinging saves metabolic energy while another might emphasize its importance in stability. Both of these ideas have some empirical support, and if you've ever tried preventing your arms from moving during walking you probably sensed that something about arm swinging made things easier. Recently, Steven Collins, Peter Adamczyk and Arthur Kuo of the University of Michigan combined experimental work and mathematical modeling to more clearly assess the role that active arm swinging plays in the energetics and mechanics of walking.

For their experimental work, Collins and his co-authors instructed 10 subjects to walk using four different types of arm swing: (1) normal, (2) bound, in which subjects' arms were physically restrained from moving, (3) held, in which subjects held their own arms still and (4) anti-normal, where subjects actively swung their arms out of phase relative to normal. By analyzing the exhaled gases of the walkers, the scientists measured metabolic rates of each subject while walking on a treadmill at 1.25 m s−1 adopting each mode of arm swing (in random order). Seven subjects also walked over a force plate using the different forms of arm swing so their locomotor mechanics could be compared.

Several variables were clearly affected by arm swinging mode. Not surprisingly, energy expenditures were lowest in the normal condition and increased 7% for bound, 12% for held and 26% for anti-normal. Vertical ground reaction moments were even more substantially affected. Peak values were lowest during normal arm swinging, approximately 60% higher during the bound and held modes and nearly 3-fold greater when arms were swung out of phase relative to normal. Experiments also revealed that joint torques and power required at the shoulder and elbow joints were very small during arm swinging.

In combination with the modeling work, these experimental data support a number of intriguing ideas. For example, the small torques and work rates in the upper arm suggest that arm swinging during walking requires little effort and that the small amount of muscular energy that is required is more than made up for by the energetic savings it provides (relative to not swinging the arms). Further, the energetic benefits likely stem from reduced ground reaction moments during normal walking. Because forces from the ground are transmitted up through the leg, limb muscle actions must be used to counteract them, and the smaller these forces, the smaller the amount of muscle energy required. The authors point out that speed likely impacts the role of arm swinging during walking, but determining the details will require more work.

As the authors conclude, arm swinging is essential to energy savings during human walking. And while our two hypothetical biomechanists might still favor different notions for why we swing our arms during walking, I know I'm now going to have a lot more to say if anyone ever asks me that question.

Dynamic arm swinging in human walking
Proc. R. Soc. Lond., B