How cells ‘remember’ regulatory decisions across divisions remains a central question in gene regulation and developmental biology. The concept of ‘epigenesis’, introduced by Conrad Waddington, has evolved to describe mitotically heritable changes in gene expression that occur without alterations in DNA sequence. A stricter definition further requires that the initiating trigger — for example, long non-coding RNA expression — becomes dispensable once the expression state is established. What qualifies as ‘epigenetic’ has long been the subject of lively debates, including discussions I encountered at conferences early in my career as a graduate student. They highlight a fundamental question in the field: is epigenetic information continuously enforced, or can it be memorized and propagated independently of the original signal and, if so, how?

Mammalian X chromosome inactivation (XCI) is a paradigm for epigenetic regulation. In female embryos, two parental X chromosomes coexist within the same nucleus, with one of the two female X chromosomes maintained in a compact, silent state. In the 1990s, the discovery of the long non-coding RNA Xist triggered a wave of research into the mechanisms underlying the establishment and maintenance of XCI. Two landmark studies probed whether XCI can persist after the initiating signal, that is, Xist, is removed.