Systems that can be described with (relatively) fewer symbols are easier to replicate, but at the same time, their transmission can be blocked equally easily. Let's put this on a scale, and call it the meme-score, or m-score for short.
There are two important consequences that arise from this:
1. Systems that are corollaries to other systems and vice-versa (those that support other systems) will have a better m-score. If a system (S1) is a corollary to another (S2), it generally implies that S1 can be described using fewer symbols than S2, assuming that the symbols are from the same set (ex: numbers, alphabets)
2. Larger systems become deceptively hard to put a check to. They hold within themselves sub-systems that contradict each other. Most large systems continue replicating with inertia. They survive not because they've cracked replication, but because they manage to survive within a collection of systems. But at the same time, the larger systems actively benefit from this parasitic (Excuse me!) behaviour.
Having established these two consequences, we can now proceed to define what would make for a stable and viral system:
"A set of symbols that can be used to describe any other set of symbols"
Take for example a human in a society, "A set of symbols that can be used to describe any other set of symbols", is just the persons philosophy on life. If a person subscribes wholly to a line of philosophy and actively adjusts all their interactions and behaviours to conform to it, their m-score will be high and will hardly fluctuate once stable.