We explore how fracture permeability in confined tight carbonates evolves due to flow of reactive fluids. Core plugs of the Capitan Massive Limestone are saw-cut to form a smooth axial fracture that is subsequently roughened to control the fracture surface topography. Either distilled water or distilled water-ammonium chloride solutions are circulated through these plugs, where fracture roughness, inlet fluid pH, and confining stresses are controlled. Throughout the experiment we measure the fluid flow rate and chemical composition of the effluent fluid. Mass balance, conducted on the effluent fluid mass and on dissolved mineral components, independently constrains the mineral mass removal. We use an idealized lumped parameter model of asperity supported fractures undergoing simultaneous stress corrosion cracking-induced diffusion and free-face dissolution to infer theoretical rates of aperture loss or gain. This model incorporates the roles of confining stress, fracture contact area, and composition and reactivity of the permeating fluid while identifying zones of diffusion-dominated mass transfer within the fracture. These theoretical rates of aperture strain are compared to those inferred from the experimentally determined permeability evolution and permeating fluid mineral mass balance. By measuring in regimes of both increasing and decreasing permeability we quantitatively constrain the transition between fracture-gaping and fracture-closing modes of behavior. We parameterize this transition in permeability evolution by the ratio of mechanically to chemically controlled dissolved mass fluxes. The transition from regimes of closing to regimes of gaping occurs at unity () when stress and chemically driven mass fluxes are theoretically equal.