Immunological memory crucially depends on CD4 T cells. In contrast with B cells, we find no decisive evidence that CD4 T cells are permanently altered by antigen stimulation. We propose that the memory response is derived from an increase in frequency of resting naïve-like CD4 T cells with a half-life of years (or months in rodents), rather than the currently proposed specialized T-cell types that have a known lifespan of days. In addition, residual antigen will significantly influence the longevity of a memory response. Our model offers a new insight into immunological memory that could assist the development of CD4 T cell-based vaccines.