Existing heliopause models are critically rediscussed under the new aspect of possible plasma mixing between the solar wind and the ambient ionized component of the local interstellar medium (LISM). Based on current kinetic plasma theories, effective diffusion rates across the heliopause are evaluated for several models with turbulence caused by electrostatic or electromagnetic interactions that could be envisaged in this context. Some specific cases that may lead to high diffusion rates are investigated, especially in regard to their LISM magnetic field dependence. For weak fields (less than 10−7 G), macroscopic hydrodynamic instabilities, such as of Rayleigh-Taylor or Kelvin-Helmholtz-types, can be excited. The resulting plasma mixing rates at the heliopause may amount to 20–30% of the impinging mass flow. Recently, an unconventional new approach to the problem for the case of tangential magnetic fields at the heliopause was published in which a continuous change of the plasma properties within an extended boundary layer is described by a complete set of two-fluid plasma equations including a hybrid MHD-formulation of wave-particle interaction effects. If a neutral sheet is assumed to exist within the boundary layer, the magnetic field direction is proven to be constant for a plane-parallel geometry. Considering the electric fields and currents in the layer, an interesting relationship between the field-reconnection probability and the electric conductivity can be derived, permitting a quantitative determination of either of these quantities. An actual value for the electrical conductivity is derived here on the basis of electron distribution functions given by a superposition of Maxwellians with different temperatures. Using two-stream instability theory and retaining only the most unstable modes, an exact solution for the density, velocity, and magnetic and electric fields can be obtained. The electrical conductivity is then shown to be six orders of magnitude lower than calculated by conventional formulas. Interestingly, this leads to an acceptable value of 0.1 for the reconnection coefficient. By analogy with the case of planetary magnetopauses, it is shown here for LISM magnetic fields of the order of 10−6 G or larger that field reconnection processes may also play an important role for the plasma mixing at the heliopause. The resulting plasma mixing rate is estimated to amount to an average value of 10% of the incident mass flow. It is suggested here that the dependence of the cosmic-ray penetration into the heliosphere on the distribution of reconnecting areas at the heliopause may provide a means of deriving the strength and orientation of the LISM field. A series of observational implications for the expected plasma mixing at the heliopause is discussed in the last part of the paper. In particular, consequences are discussed for the generation of radio noise at the heliopause, for the penetration of LISM neutrals into the heliosphere, for the propagation of cosmic rays towards the inner part of the solar system and for convective electric field mergings into the heliosphere during the course of the solar cycle, depending on the solar cycle variations. With concern to a recent detection of electrostatic plasma waves by plasma receivers on Voyagers 1 and 2, we come to an interesting alternate explanation: the heliopause, rather than the heliospheric shock front, could be responsible for the generation of these waves.