|
|
|
|
|
by hackerlight
836 days ago
|
|
> Imagine if the positions of the words in the above sentence were randomly jumbled up. But "could care less" isn't random. It is an idiom that has the same meaning as "couldn't care less". If you fed it into a LLM it would know what you mean because meaning is created from global context. Meaning is not some kind of programming language where you input the rules of grammar and the definition of each constituent word, and then out pops the meaning of the sentence. It is impossible to derive meaning that way because meaning is constructed by shared consensus about what collections of words mean in different contexts according to common usage. |
|
That is what I meant by 'English is lax enough about its grammar that "the point still gets through"'. 'Could care less' being wrong but semantically understood is exactly along the lines of 'could of' being wrong but semantically understood as 'could've', or the frequent confusion between 'their' and 'they're', or even any other confusion between homophones in written text.
Certainly, most Anglophones know enough English to read past these sorts of mistakes and still understand the underlying meaning (i.e. semantics) from context, but they are all incorrect, full stop.