The Economics of the Internet Meme

Write a blog about memes? Challenge accepted.

Sadly, I don't have the time to devote to exploring a truly profound topic - the democratization of entertainment by the internet. Memes definitely do that; I like Success Kid way better than most TV shows that corporations spend millions of dollars to produce. But nevermind that.

I want to dive into what determines quality of a meme. In case you aren't an internet junkie, an internet meme is a recurring, humorous joke that takes many different forms as it spreads virally across the internet. Examples include "lolcats" (cute pics of cats, with funny, grammatically hideous captions), "This is Sparta" jokes, "Sh*t ____ Say" videos, and most recently "What People Think I Do / How It Really Is."

But why do some take off, while others start to suck after just a few days? There is already widespread agreement that the “What People Think I Do” (WPTID) meme is tiresome and annoying. It was overdone. But why?

For a case study, take my favorite meme: rage comics. I LOVE this meme with all my heart. Here are some examples (warning, may be inappropriate for young'uns) - they consist of cut-and-pasted cartoon images that illustrate funny, awkward, and bizarre situations in in the context of everyday life. What makes them funny?

To answer that, we have to look at the producers who create the meme, and their internal decision process If you want to make a rage comic, you have to come up with a good idea, a little plot line, and some creative ways of pasting together little MS paint images. It’s not terribly hard I’m sure, but it’s going to require a good fifteen or twenty minutes to think something up.

For any meme you make, there is some effort or cost required. Maybe it’s shooting and editing video footage for “Sh*t ____ Say” or maybe it’s just thinking of a clever way for High Expectations Asian Father to demand more As. Let’s lump that wasted time, mental effort, etc. into a variable called marginal cost. Let’s also define “marginal benefit” as the amount of satisfaction you derive from having made something . In order to follow through with publishing your meme, your marginal cost has to be less than marginal benefit, (or, your reward must exceed your cost). That's called the "MB > MC" test.

It seems obvious (in other words, I assume) that marginal benefit is directly correlated, usually, with the comedic quality of the comic: the funnier it is, the more satisfaction you’ll get from having made it. But, as I’ll note in a second, the actual comedic value isn’t the only determinant of marginal benefit.

But what happens when the marginal cost gets substantially diminished? Suddenly, a much lower marginal benefit threshold must be crossed in order to motivate one to create a meme. That’s exactly what happened with the WPTID meme. Anyone can think of four or five stereotypes associated with their profession – this requires almost no effort or cleverness. The marginal cost (in time, mental effort, etc.) of making one of these memes is pathetically low. Thus, swathes or people with ideas that are only a tiny little bit funny, if at all, end up making and publishing them.

Moreover, the marginal benefit not as strongly correlated with the funniness of the meme! For any other meme, you will be satisfied with it if it is funny. But for WPTID, you’ll probably be really satisfied with it just because it applied to you and your little circle. When marginal benefit has such a pathetic correlation with funniness, then funniness is no longer a prerequisite for passing the “MB > MC” test. That’s one huge problem with WPTID… you make a meme because you relate to it, not because other people will. Sure, people relate their creations to their own experiences and personalities all the time, but not as narcissistically as they do in WPTID.

Now go waste your life browsing through awesome memes. 

Hotness Theory

Update (Feb. 18, 2012): Sadly, it seems that the evidence I have used to back up this hypothesis is equally well explained by plain old "survivability" signaling, so my whole idea here is probably unfounded. 

We humans are oddly attracted to certain traits. By and large, there are easy explanations for these attractions. The way evolution makes us adapt, it makes sense for features than indicate higher propensities for survival or reproduction to become attractive… but that’s not clearly the case all the time. Take, for instance, facial symmetry and facial averageness. The more symmetrical your face is, and the closer to average the ratios and shapes on you face are, the more attractive your face is. to me, that’s just weird. That’s the question I want to probe here.

Of course there are ways to speculate biologically on why this may be. Symmetry and averageness might indicate something good about your DNA, and people who are attracted to people with good DNA are more likely to have kids with good DNA. Still, I think the biological adaptation explanations don’t have too much going for them. I think the adaptive pressure isn’t biological, as much as economic. I guess I could tag this post with “evolutionary economics.”

Before trying to speculate on another explanation, let’s note some interesting facts that beg for explanations. In general (and those are two really important words, because exceptions abound) attractive people are more intelligent, have more friends, and get better jobs. Evolution’s a bit of a jerk; contrary to popular opinion, if you’re ugly, chances are you got the short end of the stick in other areas of life too.

So why? Allow me to speculate. Society can be thought of as a network. Every person is a node, attached to his or friends, who are attached to theirs, etc. This network serves as a mechanism for sharing information and ideas.

The stronger this network is, i.e., the stronger and more numerous the connections, the more efficiently it transmits stuff. It’s also more robust, and spontaneous interactions between two bright people bouncing heads are more likely. All in all, a better network means a more effective species.

Attractive people are well-connected hubs – they have stronger and more numerous connections. The more hubs like this there are, the better the network. The features that make them attractive could be arbitrary, or they could complement other adaptive purposes, but the important part is that they are “agreed” upon signals. Notice that I mean "agreed" in the sense that everyone’s DNA agrees to be attracted to those features. It’s not a conscious decision. Like language: no one sits down and invents common words. Words, like attractive traits, just sort of happen.

This only works if lots of people are attracted to the same thing; so the adaptive pressure is placed on the group for everyone to become attracted to the same things. In our case, those things include facial symmetry and facial averageness. So we have a social network that is strong, because it contains some nodes that are very well connected. I might have to write a follow-up post arguing why a network with a few well-connected hubs is better than one with the same number connections, but more equally spread. Eh, another time. I’ve got a report to finish tonight. Just assume I'm right, please.

Now, this explains the intelligence-attractiveness link pretty well – network hubs have a higher relative benefit from intelligence than average person. Society gets more bang for its buck when network hubs have more intelligence, because they can put it to better use is transmitting information and ideas across the network.

So that’s my hypothesis. What do you think? Any criticisms? Flawed reasoning or assumptions? What would you add? How could this be tested? Remember: you being the gorgeous human being that you are, it’s your duty to participate in the sharing of ideas, so leave a comment below!

Visual Thinking, and Awesome Shapes

In which I try to picture hypercubes.

We are bound to a three-spatial-dimension world. As you know, that means objects reach out in three directions of motion: forward, side-to-side, and up-and-down. You could also say, "length, width, depth." Our brains are nicely built to think about 3d (or 2d or 1d) spaces and objects, though. That's quite nice for artists, and others who like graphic representations for ideas, and people who just like to think about shapes.

But what about that pesky 4th spatial dimension? What about that speculative, hypothetical, additional dimension that isn't length, width, or depth? I don't know whether or not our brains are fundamentally, physically capable of imagining a shape in four spatial dimensions. But I like to try.

Here's something I thought about recently: half of a square's perimeter is the derivative of its area, with regard to the length of its side, right?

Perimeter = 4 * x.

Area = x^2.

d/dx(x^2) = 2 * x, which is half of the a square's perimeter (4 * x)/2.

Yes. Right.

And half of a cube's surface area is the derivative of its volume, with regard to the length of its side, right?

Area = 6 * x^2.

Volume = x^3.

d/dx(x^3) = 3 * x^2, which is half of its area.

Now, picture those two equations in your mind. Why is half of perimeter the derivative of area? Imagine a square is sitting comfortably in a corner. If you increase the length of its side, x, its area will grow outward, along both of the sides that are not adjacent to the corner.

The SAME mental picture applies to the cube. It is sitting in the corner of a room, and when you increase the length of x, it grows in the regions of its three faces that aren't adjacent to the walls.

Terminology attack: as a working definition, let's take the sides (or faces) of a square (or cube) that actually expand when the length of the square or cube increases, and call those the "growth regions." A square has two growth regions that are straight lines. A cube has three growth regions that are perfect squares.

Now, here's my idea. Just try to extrapolate and imagine the derivative of a hypercube's, uh, "inside stuff." (Some call it hyper-dimensional volume. Mathematically, it's the continuation of the square to cube pattern: x^4) Well, the derivative of x^4 = 4 * x^3. That means that the "growth region" is composed of four separate 3d cubes. Also note that in some conceptual way, these cubes must be perpendicular to each other in 4-space just like a square's growth lines are 90 degrees apart and a cube's growth faces are, well, 90 degrees apart...sort of.

It makes sense... a square "grows" through two sides, a cube grows through three faces, and a hypercube grows through four regular cubes.

Now, onto some really cool, cool stuff. Let's start from square 1. Imagine a circle growing inside of a square, and then growing larger than the square. The circle's edge first intersects the square's edge at the very center of each of the square's edges.

But now suddenly imagine you're in three dimensional space. You have a cube, in which a sphere is growing, and the sphere outgrows the cube. The edge of the sphere intersects the cube first at the very center of each of the cube's faces. Now imagine the region where the sphere and cube intersect, as the sphere gets larger. Well, that intersection is a growing circle on each face of the cube. It starts as a tiny dot when the sphere barely touches the edge, and grows until the diameter of the circle exceeds the side-length of the face of the cube.

And now... bang. Imagine what would happen if the equivalent of a four-dimensional sphere, a hypersphere, was birthed in the center of a hypercube, and grew inside of it, and continued to expend its way out of the cube! Remember those four 3d cubes that compose the "growth region" of the hypercube? The intersection of hypersphere against the edge of the hypercube would compose 3d sphere inside of the 3d "edge" cubes, which would start as tiny dots in the center and grow their way out, until they became larger than the cube itself.

There's your intellectual and artistic chewing gum for the day (night?). Tell me if you can see hypercubes in your head now. I'll be really jealous.