Action potentials are pulses that propagate along the membrane of excitable cells and play a crucial functional role in the behavior of many living organisms. These pulses are typically described as a purely electrical change in transmembrane potential difference, and the mathematical description of action potentials is based on the view that the excitable membrane can be fully represented by an equivalent circuit. However, this approach has come under criticism since many experimental facts are neither readily explained nor predicted by the electrical theory. One of the central points of criticism of the electrical framework is that it does not contain nonelectrical manifestations of the action potential. These pulse components, however, exist and include, for instance, a propagating cell-surface deformation with a nanoscale amplitude. In the first part of the talk, I will elaborate on the mechanical aspect, and demonstrate that in excitable plant cells (Chara braunii), a mechanical deformation in the micrometer range is observed. Thus, this preparation is useful to study cell mechanical changes during excitation in great details. I will further show that a minimalistic curvature model captures the transient cellular shapes, suggesting potential macroscopic parameters, namely, surface tension, bending rigidity, and pressure difference across the surface, that are modified during an action potential. In the second part of the talk I will describe striking similarities between action potentials and nonlinear sound waves that propagate within lipid monolayers near phase transition. These observations have motivated us to develop of a comprehensive theory of sound in lipid membranes, suggesting that sound may play an important role in the mechanism of action potentials. The model demonstrates the following similar properties: (1) correspondence of time, velocity, and voltage scales, (2) sigmoidal response to stimulation amplitude (an ‘all-or-none’ behavior), (3) electric and nonelectric manifestations, (4) conserved pulse shape upon using different types of stimulations, and (5) annihilation upon collision. Falsifiable predictions are made, suggesting that crucial computational information may be overlooked by focusing on electrical measurements only.