Do you prefer the sweetness of ice cream or the bitterness of broccoli rabe? And if you didn’t know, how would you find out?
For me, ice cream is cloyingly sweet. It leaves me bloated and gives me a sugar craving the next day. But broccoli rabe, with garlic and olive oil, in the context of a meal, is perfect.
Assuming I had never tried either food, what if were to conduct a side-by-side comparison to discover my favorite? I’d taste some ice cream, then some broccoli rabe, then some ice cream, going back and forth until my preference became clear.
Ice cream would win. The sweet food will always prevail in a pairwise comparison because of the way it interacts with the bitter food when they’re crowded together in short-term memory.
Though I don’t have a sweet tooth, my palate still responds to sweetness. The momentary pleasure leaves me wanting more, and once I’ve had something as sugary as ice cream I can’t enjoy the bitterness of broccoli rabe for a while afterwards. But the inevitable victor of this duel is the opposite of my true preference. If I had to pick only one to keep in my life, it would be broccoli rabe.
Most of us know it’s not a great idea to pit a savory food against a dessert in a taste test. One food can change the way another food tastes, giving the dessert an unfair advantage. But in other areas of life, we assume that a side-by-side comparison is the best way, maybe the only way to identify our favorite item in a set of options. The strategy is broken, but we think it’s dependable.
I was comparing two versions of a painting, one simpler and the other more complex. Flipping back and forth between them, I always preferred the simpler version. It took me three days to realize that the complex version was the better composition.
What happened? When looking at the images in rapid succession – specifically then – the simpler one always stood out as a relief for my eyes, quicker to scan, easier to comprehend. But if I looked at the complex image by itself, for an extended time, I could see that its complexity was a virtue, not a flaw. The piece wasn’t too busy. It just didn’t perform as well in a contest of first reactions.
When editing a photograph, we might boost the color saturation and then do a before/after comparison. Often we find that the enhanced image catches our attention; returning to the original feels like a letdown. But if we had looked at the original by itself, never seeing the enhanced version, we would not have said “this image needs more saturation.” It’s saturated enough. The need for more saturation only arises when we offer it to ourselves as an option and then take it away.
When editing an audio recording, we might make it louder by a small amount – a quarter decibel. In an A/B comparison, a listener might not realize that one version is louder than the other, but they might feel that one is richer, fuller, somehow more alive – that’s the magic of loudness. Taking the listener’s feedback, we’d make it louder, and louder, and louder again.
Other edits we might make, like changing the EQ or adding reverb, can have a side effect: they sometimes make the recording a bit louder. And that makes us think it’s gotten better. Unless we “level match” the input and output, it’s hard to know if the character of the sound has improved, or if we just like the output better because it’s slightly louder.
Side-by-side comparisons can help us make decisions. But even though we’re sampling the options one after the other, not simultaneously, we’re still not sampling them independently. The memory of one option influences how we experience the other option.
We want to pick the option that would give us the best experience if we came to it fresh, but the comparison only tells us which option seems better in the artificial context of flipping back and forth.
A taste test between ice cream and broccoli rabe? You might laugh if I told you I’d done this. But how many times have you chosen the figurative “ice cream” in a pair of other things – versions of a sentence, versions of an idea – that seemed like perfectly good candidates for a side-by-side comparison? ■