Below a speed limit, hydromagnetic jump shocks propagate through the interstellar medium without dissociating the molecules. The resulting excitation of the molecules is modelled here, where several factors, including the shock shape and magnetic field strength (i.e. the Alfven speed), are explored. The speed limit for molecular hydrogen is in the range approximately 24-39 km s-1 for low Alfven speeds but reaches 50 km s-1 for an Alfven speed of 10 km s-1. High Alfven speeds may well occur within active regions, such as already found for HH 34, where the results of multiple outflow events are being observed. Origins for the low but significant level of pre-shock ionization, a prerequisite for a jump shock, are considered. The excitation state of H-2 in local thermodynamic equilibrium is calculated. For planar shocks, extremely low (but non-zero) H2O fractions are indeed necessary to match the observed OMC-1 column densities. However, gas-grain cooling then dominates and the planar model is still inadequate. Bow-shock models overcome these problems. A fast bow shock possesses an oblique tail region where the normal flow speed is below the molecular dissociation limit and warm H-2 survives. A paraboloidal bow in gas of density 10(7) cm-3 then allows an H2O abundance of 3 x 10(-6) for OMC-1. Bows with longer tails reproduce the H-2 data with an arguably acceptable H2O abundance of 3 x 10(-5). The planar and bow models result in overproduction of CO rotational emission unless CO abundances less than 10(-5) are admitted. Even then, a secondary source must still be invoked to explain the CO line ratios. The degree of excitation of the H2O molecule provides a means of distinguishing shock structures and shock physics. The H2O excitation differences are estimated here, and are measurable by the Infrared Space Observatory. Differences in linewidths between types of jump and continuous shocks are also predicted.