The bulk update request #5228 is pending approval.

remove implication fruit (39192) -> plant (314334)

Reason: Fruits are seed-bearing structures which are often depicted without the presence of plants. A fruit alone is not a plant.

For instance, the following image depicts a watermelon but does not depict a plant. However, watermelon implies melon, melon implies fruit, and fruit implies plant -- so the image is erroneously tagged plant.
post #4140574

Analogy (updated):
It is no more proper to tag plant from the presence of pineapple on pizza than it would be to tag caprine due to the presence of a distinctly goat- or sheep-like skull in the desert.
post #3129661

Response to potential objection:
Some users might think,

Why is this deimplication necessary? If you want to find pictures of dragons with leafy plants, but you don't want pictures of dragons eating pineapple on pizza, just search for dragon plant -fruit!

However, the fruit -> plant implication causes problems which could not be easily solved in this manner. If a dragon is tending to his plants in the forest, and there happens to be a blueberry bush in the background, then searching dragon plant -fruit will exclude this result even though it's probably the exact sort of thing the user wanted to find with a search like dragon plant.

Put another way, the fruit -> plant implication makes the search terms dragon plant fruit and dragon fruit identical, so that one cannot easily distinguish the pictures which show dragons with fruits from the pictures which show dragons with fruits and the plants that produce those fruits.

Do we really need to argue semantics? A fruit comes from a plant.

I wont out right oppose the bur, but I dont see how that would help anyone.

Isn't a fruit a part of a plant, even when you remove it from the rest of the plant?
I don't have any strong feelings on whether or not it should be an implication though, but I do think it is slightly more reasonable that a living part of a plant would imply plant than a dead part of an animal implying that animal species.

wolfmanfur said:
Do we really need to argue semantics? A fruit comes from a plant.

I wont out right oppose the bur, but I dont see how that would help anyone.

The benefits are clarity and specificity in search results. People who search for plant are not trying to find pictures of fruits totally removed from the plants from which they came.

Let's say I remember a cute picture of a dragon tending to his plants, I might try to search for this with the terms dragon plant.

Because of the fruit -> plant implication, my initial search results will be cluttered with results like:
post #3273934 (pineapple pieces on pizza)
post #4124106 (apple in mouth)
post #3668516 (berries next to tail)

...and so on, and so forth. It's important to tag the above images with pineapple, apple, and berry, but the fruit -> plant implication means all of these pictures are unhelpfully tagged with plant as well.

The plant tag could be quite a helpful search term, but it requires an excess of additional operators to be used in any meaningful searches -- which dissuades regular users from getting much use out of it.

themasterpotato said:
Isn't a fruit a part of a plant, even when you remove it from the rest of the plant?

A fruit is a plant-part, but I think the implication is misleading because a depiction of a fruit is usually not a meaningful depiction of a plant.

A drawing of a wolf eating dried banana chips or peach slices should be tagged banana or peach, but those images are not particularly helpful when a user is searching for a wolf with a plant.

themasterpotato said:
I don't have any strong feelings on whether or not it should be an implication though, but I do think it is slightly more reasonable that a living part of a plant would imply plant than a dead part of an animal implying that animal species.

My analogy probably could be improved. It is perhaps more reasonable that a wolf with a bag of dried banana chips be tagged plant, but it still isn't very reasonable on the whole.
Plus, setting aside the ontology of plants, the implication has functional problems. It makes the plant tag less useful by cluttering search results.

Any opinions on the current implication fruit_humanoid -> plant_humanoid?

Are we also going to unimply flower from plant? That's likewise just a structure that forms part of a flowering plant, and can be depicted separately.

faucet said:
Are we also going to unimply flower from plant? That's likewise just a structure that forms part of a flowering plant, and can be depicted separately.

The issue is that we don't consider a single petal to be a flower, but we do consider a tiny part of a fruit to be a fruit (for instance, the pineapple on pizza example above). A complete flower is still a significant part of a plant.

There's also the fact that we use the word flower for both the reproductive part of a plant, and an entire group of plants.

faucet said:
Are we also going to unimply flower from plant? That's likewise just a structure that forms part of a flowering plant, and can be depicted separately.

scth said:
Any opinions on the current implication fruit_humanoid -> plant_humanoid?

I would want to look at these tags more thoroughly, to form an opinion on them. If anyone has used the flower or fruit_humanoid tags in any substantive way and has an opinion on their implications, I'd love to hear -- your experience is probably far more informed than mine.

In any case, my initial position is that it's reasonable to hold that fruit -> plant is an unhelpful implication, without staking out a claim on whether flower -> plant or fruit_humanoid -> plant_humanoid is useful. Having a parallel set of implication rules would be nice, but there is no reason why it must be that way. If it turns out that flower -> plant causes far fewer spurious search results than fruit -> plant, then it's totally reasonable to leave the flower -> plant implication intact whilst unimplying plant from fruit.

What troubles me most about the fruit -> plant implication is cases like the examples I posted above (post #4140574, post #3273934, post #4124106, and post #3668516). The question I pose is whether users want to see pictures of fruit -- and no other plant parts -- when they search for the plant tag. My inclination is to say no.

I have updated the original post with a response to a possible objection, which is to simply add an operator to remove fruit results from searches for plants.

This is an unsatisfactory solution, for reasons I have now outlined in the original post.

No. You're going to have the same issue with grass. You also don't need to use -fruit.

Just search https://e621.net/posts?tags=dragon+%7Egrass+%7Etree+%7Ecactus+%7Eshrub

Remove or add additional plants you desire.