Despite significant advancements in memory research—from the level of the entire brain to specific regions and even to the connections between neurons—the exact physical location and mechanism by which memories are stored remain unresolved. Moreover, most existing studies have focused primarily on output data such as behavioral responses, leaving research on the memory system itself rather limited. In contrast, the recently proposed "meshcode" theory posits binary encoding within cells via the talin protein, suggesting that physical symbols of memory may exist at the molecular level. If certain proteins function as physical symbols that store memory, the earliest memory system might have directly linked symbols and meanings in an ideogrammatic manner, and the first concept to be encoded would have been "self," based on feedback mechanisms. Furthermore, the varied feedback systems found in living organisms can be viewed as another name for the memory system. By examining how these feedback systems, which originate from self/non-self recognition, expand, we can infer the structure of the memory system. This approach not only reconstructs the structural algorithms of memory but also implies that memory may be distributed throughout the organism rather than limited to the brain. In addition, the expansion of self-recognition feedback systems may serve as a universal foundation for the formation of self-awareness and consciousness. Such insights offer a framework forreinterpreting the origins of memory and consciousness beyond the limits of neuroscience, while also revealing that today’s artificial intelligence is constrained by the absence of self-recognition algorithms. In the future, the design of AI or humanoids based on self-recognition could open the door to the emergence of consciousness.