It's certainly a philosophical question, and needs to be addressed as such.
This is probably the best place to start:
Classical memeticists attempted to (unsuccessfully) bypass the existing debates about mind and meaning, using universally Darwinist conceptions of the mind and its functions. They frequently reinvented the wheel but worse because of that. Here's Keith Henson (writing in 2003), a member of the memetics-digest mailing list (in response to a proposed "internalist-externalist split" in memetics at the time):
Does a computer file of "War and Peace" read differently from a paper copy? Is it somehow shorthand or a facsimile? I don't think so.
A base pair listing of a gene is an exact representation of the physical structure of a gene. You could get 4 different kinds of those kid-toy snap together type beads and code a gene that way too. The only thing that is common to these different ways of representing a gene is the information. If you look in a journal article, nobody ever makes a disclaimer that this is a paper representation of the hemoglobin gene, they just label the paper listing as Hemoglobin. The least complicated thing you can do is to take this practice at face value and admit the only invariant element is the information.
By close analogy, the form, i.e., paper, brains, or steam engines in which the meme "separate condenser" is encoded are utterly divergent. But
something in each of them can get into the brain of an engineer and change the way he designs steam engines in exactly the same way. That something is a meme. The only thing common over these different forms is information. QED, memes are information in the same sense genes are above.
We've seen the same kinds of debates happening about irony, humour, and whether Pepe is a hate symbol. I think the same misguided intuition (that we must locate and isolate meaning or information in order to study it) is at work in a lot of debates about whether irony is intrinsic or extrinsic to a meme. A huge amount of reinventing without consulting prior research seem to always take place in subcultures as they mature enough to become reflective about their own culture and communities. That's the case whether they're political, aesthetic, comedic, etc.
Memecultures discussing memes developing their own theories of memes is uniquely exciting, because when those new and independently developed theories converge with old and preexisting theories, you can interpolate those theories to generate new evidence about how thinking about memes works. Nonetheless, we here in this research community should examine and consult as much of the preexisting research as possible, so that we can facilitate and encourage such synthesis.