Suppose I order a blegg from a mail-order catalog. As it turns out, the object I received is blue and is furry, but it is cube-shaped, does not glow in the dark, and contains neither vanadium nor palladium. I am disappointed and attempt to return the object, claiming that it is not, in fact, a blegg. The seller refuses to give me a refund or exchange the object for another. Annoyed, I decide to take the seller to court.
Would I win the lawsuit?
(This is why arguments over definitions have real-world consequences.)
Sorry, just… no. I realize it’s been four years, but I had to create an account just to register my disapproval. The question remains, what did you want from the blegg? Vanadium or Palladium? Its glow-in-the-dark property? A gestalt effect arising from the combination of certain salient features? What does any of this have to do with consensus-based definitions?
I find it quite interesting that despite the two above posts having very strongly contradictory points they both a large number of upvotes are are both at 100% positive (15 and 8 at the time of writing). I wonder whether the community’s opinion has shifted over the years, or whether lw voters just think both points are well put and are very reluctant to downvote things based on disagreeing with a point.
I think it’s very much a case of well-put arguments.
I can certainly see how pragmatically, definitions can clearly matter. Heck, we have laws that are very picky about them, because we need a very specific set of rules so that crinimals/those falsely accused are clearly in one category or another, and to make the law above debate.
At the same time, asking whether something is against the law, whether it fits into the category of “murder” for example, is simly arguing whether it is case considered worthy, by those who wrote the law, of punishment.
Both arguments are very well explained by the two comments.
I hope you’ll forgive me for creating an account two years later after you created one four years later.
If the catalog offered a “blegg” for sale, with no further information, then the definition of “blegg” itself would very much be an important issue in the ensuing lawsuit. If “blegg” generally means something that contains vanadium and is egg-shaped, but the one sold and delivered is neither, then the actual definition would be important in determining whether the buyer was scammed. The question is not “what did you want from the blegg”, it’s “what is a blegg, and is that what was sent by the company”.
If I order a “bicycle” from a catalog and the product comes with one triangular wheel and no seat, then it very much matters that the commonly used definition of “bicycle” is “a conveyance with two round wheel and a place to sit”, and that would be a valid basis for suing the fraudulent company. It doesn’t matter “what I want from” the bicycle. I might want a convenient mode of human-powered transportation, or I might want a frame of welded metal. But the fact is the company did not deliver what was advertised.
Since we have been debating this topic regularly for 11 years, I’ll chime in to keep it going.
I do not disagree that definitions have real-world consequences, and I don’t think EY was ever trying to imply in his writings that they do not. Of course, if you compress meaning down into an word that stands for multiple characteristics, what characteristics are included in a particular use-case become important when two humans must achieve business together using those definitions.
However, no one is cheated by an ebay seller and is still intellectually confused afterward about the fact—I mean the fact itself—that the seller has left out pertinent information. When the one says that the seller sold a blegg that was not a blegg, they are actually asserting that the seller’s item did not include important characteristics of bleggness and thus they were cheated. They grasp the problem in the same instant.
If I buy a microwave at an estate sale, I take it home and make sure the electronics aren’t fried. If the estate sale organizers tried the microwave and found that it did not work, but sold it anyway and then refused to refund, they are committing fraud by implying a characteristic that was not present, of which they were fully aware. If an Amazon seller sends me a book and it is blindingly obvious that it was stolen (middle school library markings that have not even been crossed out or perhaps marked as a textbook exclusively for an overseas market), then I’m not confused about why I am mad. The item included a detrimental characteristic that was not specified in the listing. (The same essential problem—added or subtracted characteristics—in both cases, but this second one is not actually a definition problem—a stolen book is still fully a “book.”)
It is very easy and intuitive for me to think these things about items I have spent my money on, and if a judge ruled that the car I bought online at full market value for a running car was delivered without a motor, but I’m still on the hook for paying full price, “Because the ad never claimed it had a motor or could run,” I would not only spend years griping, but all my friends would probably agree I had been cheated by both the seller and the judge.
The point about rationality literature is not particularly to point out where the world and our mental models meet and agree. It’s to point out where the world and our mental models clash and break down, and our mental models win (incorrectly), like when we ask, “But is it a blegg!?” not for the sake of a real world dispute, but because the blue, fuzzy, etc. thing can’t be gotten out of our mind until we have labeled it and put it to rest.
In other words, you are right but not particularly useful.
Suppose I order a blegg from a mail-order catalog. As it turns out, the object I received is blue and is furry, but it is cube-shaped, does not glow in the dark, and contains neither vanadium nor palladium. I am disappointed and attempt to return the object, claiming that it is not, in fact, a blegg. The seller refuses to give me a refund or exchange the object for another. Annoyed, I decide to take the seller to court.
Would I win the lawsuit?
(This is why arguments over definitions have real-world consequences.)
Sorry, just… no. I realize it’s been four years, but I had to create an account just to register my disapproval. The question remains, what did you want from the blegg? Vanadium or Palladium? Its glow-in-the-dark property? A gestalt effect arising from the combination of certain salient features? What does any of this have to do with consensus-based definitions?
I find it quite interesting that despite the two above posts having very strongly contradictory points they both a large number of upvotes are are both at 100% positive (15 and 8 at the time of writing). I wonder whether the community’s opinion has shifted over the years, or whether lw voters just think both points are well put and are very reluctant to downvote things based on disagreeing with a point.
I think it’s very much a case of well-put arguments.
I can certainly see how pragmatically, definitions can clearly matter. Heck, we have laws that are very picky about them, because we need a very specific set of rules so that crinimals/those falsely accused are clearly in one category or another, and to make the law above debate.
At the same time, asking whether something is against the law, whether it fits into the category of “murder” for example, is simly arguing whether it is case considered worthy, by those who wrote the law, of punishment.
Both arguments are very well explained by the two comments.
I hope you’ll forgive me for creating an account two years later after you created one four years later.
If the catalog offered a “blegg” for sale, with no further information, then the definition of “blegg” itself would very much be an important issue in the ensuing lawsuit. If “blegg” generally means something that contains vanadium and is egg-shaped, but the one sold and delivered is neither, then the actual definition would be important in determining whether the buyer was scammed. The question is not “what did you want from the blegg”, it’s “what is a blegg, and is that what was sent by the company”.
If I order a “bicycle” from a catalog and the product comes with one triangular wheel and no seat, then it very much matters that the commonly used definition of “bicycle” is “a conveyance with two round wheel and a place to sit”, and that would be a valid basis for suing the fraudulent company. It doesn’t matter “what I want from” the bicycle. I might want a convenient mode of human-powered transportation, or I might want a frame of welded metal. But the fact is the company did not deliver what was advertised.
Since we have been debating this topic regularly for 11 years, I’ll chime in to keep it going.
I do not disagree that definitions have real-world consequences, and I don’t think EY was ever trying to imply in his writings that they do not. Of course, if you compress meaning down into an word that stands for multiple characteristics, what characteristics are included in a particular use-case become important when two humans must achieve business together using those definitions.
However, no one is cheated by an ebay seller and is still intellectually confused afterward about the fact—I mean the fact itself—that the seller has left out pertinent information. When the one says that the seller sold a blegg that was not a blegg, they are actually asserting that the seller’s item did not include important characteristics of bleggness and thus they were cheated. They grasp the problem in the same instant.
If I buy a microwave at an estate sale, I take it home and make sure the electronics aren’t fried. If the estate sale organizers tried the microwave and found that it did not work, but sold it anyway and then refused to refund, they are committing fraud by implying a characteristic that was not present, of which they were fully aware. If an Amazon seller sends me a book and it is blindingly obvious that it was stolen (middle school library markings that have not even been crossed out or perhaps marked as a textbook exclusively for an overseas market), then I’m not confused about why I am mad. The item included a detrimental characteristic that was not specified in the listing. (The same essential problem—added or subtracted characteristics—in both cases, but this second one is not actually a definition problem—a stolen book is still fully a “book.”)
It is very easy and intuitive for me to think these things about items I have spent my money on, and if a judge ruled that the car I bought online at full market value for a running car was delivered without a motor, but I’m still on the hook for paying full price, “Because the ad never claimed it had a motor or could run,” I would not only spend years griping, but all my friends would probably agree I had been cheated by both the seller and the judge.
The point about rationality literature is not particularly to point out where the world and our mental models meet and agree. It’s to point out where the world and our mental models clash and break down, and our mental models win (incorrectly), like when we ask, “But is it a blegg!?” not for the sake of a real world dispute, but because the blue, fuzzy, etc. thing can’t be gotten out of our mind until we have labeled it and put it to rest.
In other words, you are right but not particularly useful.