How many ingredients does it take to call it a salad?

DoctorWhookah@sh.itjust.works to Asklemmy@lemmy.ml – 240 points –

My significant other ate cucumbers and onion with some ranch. I called it a cucumber onion salad. She says there aren't enough ingredients to call it a salad, because "it takes multiple ingredients". I pointed out she had three and asked what the minimum is. She refuses to answer so I ask Lemmy.

183

You are viewing a single comment

So teeeeechnically, a salad is a dish composed of mixed ingredients. You could make the argument that you mix any two set of chopped ingredients and bingo bongo, it's a salad.

However, I like to think that dishes' ingredients aren't a taxonomic thing, they're a probabilistic thing. In other words, there's no such thing as "not salad" or "salad", only shades of saladness.

  • Serve it cold? Ok it's saladier

  • It's made up of chopped ingredients? Saladier still

  • Those ingredients are mostly vegetables? Getting pretty saladish

  • They're mixed together? Even more salad like

  • They've got some sort of dressing mixed in? Now it's very likely a salad!

... and so on. To me, your SO'a dish has a pretty high Salad Probability^tm

Is there anyway to convert a Salad Probability^tm into a Salad Factor score?

Give me weights for the coefficients and I'll construct a matrix

Here's some example weights for a salad factor

Lettuce - 10

Spinach - 9

Arugula - 7

Cabbage - 7

Tomato- 6

Carrots- 6

Cucumber - 5

Onion - 4

Olives (black) 4

Anchovies - 1

There are a few missing points in there IMO, like which of your ingredient is cooked, or how are they sliced? Graped carrots rises the score, but cook them and it's less likely to be a salad. Diced radish? Not in my salad, especially not cooked, but thinly sliced raw radish definitely belongs. And don't even get me started on tomatoes.

Damn, I'm not sure the two are compatible then. The salad factor score is meant to be super easy so people don't get overwhelmed by all the possibilities and variations.

What we need is a salad categorizing multilayer neural network

It wouldn't be hard to train on obvious salads, but what about the non obvious ones? Who do we trust to properly label these as salad or not a salad?

We don't! That's the joy of it, just like people do, our algorithm will constantly waffle back and forth and argue with itself over whether these things are salads

I think this is one of those words / concepts where one needs to invoke Wittgenstein's "family resemblance" idea. You're not going to find some exact set of criteria that define what people do and don't consider a salad. They instead have a "family resemblance".

Your probably idea is not a bad way of describing how that works.

I know, I was being humorous but it is in fact the way most categorization works. Very seldom is it a taxonomy; the way we recognize faces, voices, shapes, etc ... it's all probabilistic.