Approaching Memes

a lecture given 10/21

One thing I did in October was give a lecture for a class taught by my friend Nina at the University of Washington. It was a lot of fun, and I found it useful in two ways: first, as an opportunity to condense and centralize my thinking about memes, and second, as a demonstration of how interesting other people find this stuff. Sometimes I forget that memes are fun, and the questions of the students right after the talk — which were interesting and incisive — reminded me of how much fun it is to think about memes. Apart from the practical significance of studying internet culture (geopolitical, technological, etc.) it’s a joyful topic, and I’m fortunate to share that joy with people.

Here’s the lecture, presented in Substack form and somewhat edited with my thoughts since then.

my introductory slide from the lecture

Memes as cultural artifacts (I’m wondering whether to use “artifact” instead of “text” lately) are relevant in three ways. First, you can read them and get a meaning like, “oh, this is political satire,” or “oh, this meme is about the experience of x or y.” That meaning is interesting to analyze, and just as rich as what you can gather from books, movies, or art. Memes riff on decades of artistic tradition and pull together influences from far and wide.

Second, you can read a meme the way a scientist reads the data from one of those GPS trackers they put on whales. The viral spread of a meme hopping from community to community and platform to platform, bearing the marks of each milieu it moves through, tells you about the overall structure of a media ecosystem — how information moves, whose voice is listened to, how messages morph. In the absence of a coherent quantitative record for how these platforms work (either because of their scale, or because of purposeful obfuscation on the part of the companies) the focused, in-depth study of memes as representative artifacts can be really useful and is an avenue of study that should be explored (I don’t know how to make graphs so don’t think I can do it). Further, memes are often the anchors on which audience conceptions of a platform’s vibe are moored: a meme’s sudden emergence on the feed and its viral spread is a highly visible demonstration of the social internet’s power and reach. Whether accurate or no, users’ ideas about how TikTok or Facebook work are deeply informed by the memes they have observed pop off and then die off. Viral spread is never some willy-nilly irrational force, it is always structured.

I showed pie charts from this report I did for Know Your Meme looking at the trends in site of origin over the years. The 2014 graph is here below, but there’s a whole lot more on the site and by tracing memes tend to originate you really do see how the social internet has changed.

Third, I think memes can be studied as formal scaffoldings of discourse. This last point is one I used to lump in with the second, but I think the focus here is different enough to warrant its own subheading now. Studying a meme like a GPS’ed humpback whale means studying its transit across and participation in external structures. Studying a meme as a formal scaffolding, however, means turning to its internal structure and thinking about how a meme arranges thoughts and feelings, words and images. I’m thinking of Stanley Fish and his reader response theory: that the experience of reading a text, and being aware of other selves clustered around it, carries a form of meaning. I love his reading of Milton’s Paradise Lost, where he calls it a “script” for a reading experience rather than a “poem,” if I recall correctly. Fish says that Milton’s use of a more Latin-like sentence structure in the poem (where the sentences are hard to understand because you have to read like three lines until you find the verb, which is more how Romance languages do grammar than English) helps make the poem difficult to interpret — which is the point. Fish says Paradise Lost is supposed to frustrate you and leave itself open to interpretation so that you feel the stress and doubt that comes with living the kind of Christian life Milton writes about. It’s kind of the ultimate “show, don’t tell” — the poem makes you feel the way the people in it feel.

The Latinate sentences in Milton are an example of an internal formal choice that structures a reader response. If I were less Fishpilled I would just say “style,” but I think that term carries the connotation of individual authorship, which is not true of most memes. The operations which a meme format makes us perform — like scanning an image’s changing labels and captions to find a meaning and then implicitly understanding that meaning as something attached after the fact instead of inherent in the image — are analogous to the operations of Milton’s Latinate sentences. These are not natural ways for the human brain to think, but rather terms set by a creator’s choices.

If thoughts and feelings must be fit into these internal structures in order to be communicable, then the shapes and vibes of meme formats (or literary styles) really do matter. Water hits different out of a gallon jug than it does from a champagne glass. Or maybe that metaphor isn’t strong enough, and I should say water hits different when it’s in glacier form than when it’s in cloud form. These kinds of internal structuring, as evidenced by the length I have gone on here, are really what I’m most interested in.

If you’re a long-time or medium-time reader of this blog, you’ll likely recognize my whole thing about “layers” in memes, which I see as the most relevant internal formal structure. The way a meme works (particularly the object-labelling memes which have been the focus of my research) is by depicting time as stacked and nested layers. What’s on the bottom (like the Distracted Boyfriend stock photo) happened before what’s in the middle (like the labels applied to it). What’s on top (the platform’s framing of the meme in a scroll) happened last. The layers further up (and the furthest up is your own screen, where you get to like, repost, or just ignore the meme) have power over the ones further down. We find meaning and humor in a meme by comparing these layers, juxtaposing and judging them.

In addition to a moment in time, each layer represents a social context with a particular speaker, implied audience, and set of understandings. Seeing it even more straightforwardly, each layer is the result of a different screen/machine that displayed the image… going from top to bottom: your screen, my screen, Twitter’s screen, the meme-maker’s screen, the screen on the camera of the photographer back in 2015 in northern Spain. The meme stacks these translucent screens one on top of the other.

My contention is that this kind of internal structuring, which results from choices made by people as well as conditions imposed by technology, alters the way we think about stuff. In the same way reading too many novels might give you “main character syndrome,” a communicative form that encourages you to employ the comparison of one take, time, or social context against another as the primary way of finding meaning does make you think differently. In a meme, meaning is structured as relative rather than absolute, multiple rather than singular. This differs significantly from the kind of meaning you get from non-meme images or texts which live on other interfaces (page, picture-frame).

I concluded my talk with a list of the actual elements in or around a post that are readable, the ways a meme speaks to you. I divided the list into “within the post” and “around the post,” which I think maps onto my “internal structure” versus “external structure” division.

Within the Post (what the meme is)

• Username and poster’s voice and reputation

• Likes, shares, comments, quantifying reach and approval, signalling bandwagon effect

• References/allusions to shared experiences, other pieces of media

• The meme/posting format (and its associations)

• The traditions of the platform or community (e.g. greentext, photo dump)

• Calls to action (e.g., “RT if you agree”)

• Technical traces indicating a tool, generator, or era (e.g., “looks like AI,” or “Impact font”)

• The behaviors the meme demands of a user for proper interpretation

Around the Post (what the meme is evidence of)

• The post is the result of algorithmic placement (it knows you)

• The post is the result of your own user curation of algorithm (you know it)

• The post is the result of network effects (your friends/niche must have liked it)

• Placement within search results or platform-generated lists (SEO)

• Placement within user-created archives/compilations

• Content moderation processes, or lack of them

I doubt this list is complete. Maybe you can think of more things to put in it. Thank you for reading.