When two solid objects slide over each other, friction results from the interactions between the asperities of the (invariably rough) surfaces. Lubrication happens when viscous lubricants separate the two surfaces and carry the load such that solid-on-solid contacts are avoided. Yet, even small amounts of low-viscosity lubricants can still significantly lower friction through a process called boundary lubrication. Understanding the origin of the boundary lubricating effect is hampered by challenges in measuring the interfacial properties of lubricants directly between the two surfaces. Here, we use rigidochromic fluorescent probe molecules to measure precisely what happens on a molecular scale during vapor-phase boundary lubrication of a polymer bead-on-glass interface. The probe molecules have a longer fluorescence lifetime in a confined environment, which allows one to measure the area of real contact between rough surfaces and infer the shear stress at the lubricated interfaces. The latter is shown to be proportional to the inverse of the local interfacial free volume determined using the measured fluorescence lifetime. The free volume can then be used in an Eyring-type model as the stress activation volume, allowing to collapse the data of stress as a function of sliding velocity and partial pressure of the vapor phase lubricant. This shows directly that as more boundary lubricant is applied, larger clusters of lubricant molecules become involved in the shear process thereby lowering the friction.