In a brief history of the discovery of antiferroelectricity in liquid crystals, the important role played by tristable switching, i.e. an electric field induced phase transition from antiferroelectric SCA* to ferroelectric SC*, has been emphasised and the antiferroelectric herringbone structure of SCA* has been presented. Then we have explained how to identify the subphases in the SC* region, e.g. SCγ*, SCα*; the clarification of the subphase structures is essential for understanding antiferroelectricity in liquid crystals. After summarizing the evidence for the SCA* structure presented, we have suggested the pair formation of transverse dipole moments in adjacent smectic layers as the cause of its antiferroelectricity, showing that the smectic layer is much closer to the usual picture of molecules lying on equidistant planes; the packing entropy due to the sinusoidal density wave character stabilizes ferroelectric SC*. The competition between the interactions stabilizing SCA* and SC* is responsible for the occurrence of several varieties of ferrielectric and antiferroelectric subphases, which constitutes the Devil's staircase. We have further suggested that the essentials of the SCα* phase are its considerably reduced ability to form SCA* and SC*. At least when the spontaneous polarization is zero, SCα* is a smectic C-like phase with molecular tilting that is non-correlated on the visible wavelength scale When the spontaneous polarization is not zero, as suggested by Prost and Bruinsma recently, a novel type of Coulomb interaction between smectic layers due to the collective polarization fluctuations causes the antiferroelectricity in the high-temperature region of SCα*; the competition between this antiferroelectricity and the SC* ferroelectricity may form another staircase, causing the complexity in SCα*. Applications and some future problems have been described in the final section.