Muscle contraction is commonly associated with the cross-bridge and sliding filament theories, which have received strong support from experiments conducted over the years in different laboratories. However, there are studies that cannot be readily explained by the theories, showing (i) a plateau of the force-length relation extended beyond optimal filament overlap, and forces produced at long sarcomere lengths that are higher than those predicted by the sliding filament theory, (ii) passive forces at long sarcomere lengths that can be modulated by activation and Ca²⁺, which changes the force-length relation, and (iii) an unexplained high force produced during and after stretch of activated muscle fibers. Some of these studies even propose "new theories of contraction" to explain discrepant results. While some of these observations deserve evaluation, many of these studies present data that lack a rigorous control, and experiments that cannot be repeated in other laboratories. This article reviews these issues, looking into studies that have used intact and permeabilized fibers, myofibrils, isolated sarcomeres and half-sarcomeres. A common mechanism associated with sarcomere and half-sarcomere length non-uniformities, and a Ca²⁺-induced increase in the stiffness of titin is proposed to explain observations that derive from repeatable studies.