It has been proposed that the Si(557)-Au surface exhibits spin-charge separ
ation in a one-dimensional electron liquid. Two narrowly spaced bands are f
ound which exhibit a well-defined splitting at the Fermi level. That is inc
ompatible with the assignment to a spinon-holon pair in a Luttinger liquid.
instead, we propose that the two bands are associated with two nearly dege
nerate atomic chains, or a chain of step atoms with two broken bonds. Such
an assignment explains why the surface is metallic despite an even number o
f electrons per unit cell.