I’m not sure I buy the glacier example (nor any similar ones).
The sentence “A glacier is a river of ice” may also be read as shorthand for “A glacier is a river, except of ice”. This is itself still a sort of shorthand; the full and precise formulation would be: “A glacier is like a river in all important respects, except that instead of being made of water, as actual rivers are, it is made out of ice.”
Under this interpretation, if I say “A glacier is a river of ice”, I am not actually saying that a glacier is a river; in other words, the following implication does not hold (or at least, I am not claiming, nor implying, that it holds):
“A glacier is a river of ice” → “A glacier is a river”
Or, to put it in yet another way, “river” is one sort of thing, and “river of ice” is another sort of thing, and the latter is not simply a particular kind of the former.
So I am not extending the concept of “river”, you see. Under this interpretation, what I am doing instead is a sort of one-time generalization, where I take some named concept (in this case, “river”), and I generalize it—without, however, naming the more general concept, because I do not necessarily intend to re-use it to generate other analogues. (By analogy with certain programming concepts, we might say that I am creating a “locally-scoped anonymous template”.)
Other examples:
An open-faced sandwich is a sandwich, except without one of the bread layers. (It is not the case that an open-faced sandwich is a sandwich.)
A cobbler is a pie, except without the bottom crust. (It is not the case that a cobbler is a pie.)
A tricycle is a bicycle, except with three wheels. (It is not the case that a tricycle is a bicycle.)
Hearing aids are eyeglasses, except for hearing. (It is not the case that hearing aids are eyeglasses.)
A dishwasher is a washing machine, except for dishes. (It is not the case that a dishwasher is a washing machine.)
Note that under this interpretation, no “general” or “extended” version of the concept is ever created (the template is anonymous, and is discarded as soon as it “goes out of scope”—which is to say, as soon as it has been used to create the new concept). There is thus no need to ask the questions of what this new, “general”/“extended” concept means, to what else it may or may not apply, how to differentiate between uses of it and any specific version, etc.
EDIT: To add to what I say above: I think it is likely that this account more realistically describes how people form new concepts. Actual generalization—which is to say, the creation of a persistent, named generalized version of some concrete concept—is, I suspect, more rare than this sort of “concept cloning”.
The concept cloning you talk about is definitely a thing. It’s a good fit for the river/glacier example.
I think there is a general form of the “river of X” concept, such that people know what you mean if you say “river of (new substance)” (and can reuse their intuitions about other rivers(2)), but this form isn’t named.
There are cases where “river” is used metaphorically without modifying it to “river of X”. Consider: “The glacier is a river flowing slowly down the mountain”. It’s poetic/metaphorical language, not considered to be literally true, but considered to be true in some metaphorical sense.
The water/ice example is a case where ice is considered to be, literally, a kind of water. In the case of adoptive/biological parents, the “parent” concept is metaphorically extended to include adoptive parents, with the original being renamed to “biological parent”. These are examples of the kind of extensions I’m talking about, where the extended one becomes a canonical concept.
I think both of this and the above post are treating concepts as more platonic than they are in practice. Take for instance “one legged duck”, it has the same structure as “open faced sandwich” but most people would say it is in fact still a duck, even though ducks have two legs. I think language concepts are often pointing at fuzzy categories, and so it’s natural to not be sure how far they extend, to extend them in ways that stretch further than other people’s extensions, etc.
Some people’s boundary for river includes water, but some people’s doesn’t, etc. The dictionary definition is a good starting place, but rarely sufficient for how humans use language.
Take for instance “one legged duck”, it has the same structure as “open faced sandwich” but most people would say it is in fact still a duck, even though ducks have two legs.
Yes, of course. I didn’t say anything that would contradict that, even by implication. I certainly wouldn’t claim that all concepts which might be described with the same very general syntactic structure behave in the way I described; that would be a rather bizarre claim, wouldn’t it?
EDIT:
The dictionary definition is a good starting place, but rarely sufficient for how humans use language.
I made no reference to the dictionary definition of anything, though.
The point I was trying to make was that for instance in your open faced sandwich example, there are many people who would say it’s still obviously a sandwich, just like the duck is still a duck.
However, I realized I don’t know enough about the philosophy of language to meaningfully contribute to the discussion, so I’m going to bow out.
I’m not sure I buy the glacier example (nor any similar ones).
The sentence “A glacier is a river of ice” may also be read as shorthand for “A glacier is a river, except of ice”. This is itself still a sort of shorthand; the full and precise formulation would be: “A glacier is like a river in all important respects, except that instead of being made of water, as actual rivers are, it is made out of ice.”
Under this interpretation, if I say “A glacier is a river of ice”, I am not actually saying that a glacier is a river; in other words, the following implication does not hold (or at least, I am not claiming, nor implying, that it holds):
“A glacier is a river of ice” → “A glacier is a river”
Or, to put it in yet another way, “river” is one sort of thing, and “river of ice” is another sort of thing, and the latter is not simply a particular kind of the former.
So I am not extending the concept of “river”, you see. Under this interpretation, what I am doing instead is a sort of one-time generalization, where I take some named concept (in this case, “river”), and I generalize it—without, however, naming the more general concept, because I do not necessarily intend to re-use it to generate other analogues. (By analogy with certain programming concepts, we might say that I am creating a “locally-scoped anonymous template”.)
Other examples:
An open-faced sandwich is a sandwich, except without one of the bread layers. (It is not the case that an open-faced sandwich is a sandwich.)
A cobbler is a pie, except without the bottom crust. (It is not the case that a cobbler is a pie.)
A tricycle is a bicycle, except with three wheels. (It is not the case that a tricycle is a bicycle.)
Hearing aids are eyeglasses, except for hearing. (It is not the case that hearing aids are eyeglasses.)
A dishwasher is a washing machine, except for dishes. (It is not the case that a dishwasher is a washing machine.)
Note that under this interpretation, no “general” or “extended” version of the concept is ever created (the template is anonymous, and is discarded as soon as it “goes out of scope”—which is to say, as soon as it has been used to create the new concept). There is thus no need to ask the questions of what this new, “general”/“extended” concept means, to what else it may or may not apply, how to differentiate between uses of it and any specific version, etc.
EDIT: To add to what I say above: I think it is likely that this account more realistically describes how people form new concepts. Actual generalization—which is to say, the creation of a persistent, named generalized version of some concrete concept—is, I suspect, more rare than this sort of “concept cloning”.
The concept cloning you talk about is definitely a thing. It’s a good fit for the river/glacier example.
I think there is a general form of the “river of X” concept, such that people know what you mean if you say “river of (new substance)” (and can reuse their intuitions about other rivers(2)), but this form isn’t named.
There are cases where “river” is used metaphorically without modifying it to “river of X”. Consider: “The glacier is a river flowing slowly down the mountain”. It’s poetic/metaphorical language, not considered to be literally true, but considered to be true in some metaphorical sense.
The water/ice example is a case where ice is considered to be, literally, a kind of water. In the case of adoptive/biological parents, the “parent” concept is metaphorically extended to include adoptive parents, with the original being renamed to “biological parent”. These are examples of the kind of extensions I’m talking about, where the extended one becomes a canonical concept.
I think both of this and the above post are treating concepts as more platonic than they are in practice. Take for instance “one legged duck”, it has the same structure as “open faced sandwich” but most people would say it is in fact still a duck, even though ducks have two legs. I think language concepts are often pointing at fuzzy categories, and so it’s natural to not be sure how far they extend, to extend them in ways that stretch further than other people’s extensions, etc.
Some people’s boundary for river includes water, but some people’s doesn’t, etc. The dictionary definition is a good starting place, but rarely sufficient for how humans use language.
Yes, of course. I didn’t say anything that would contradict that, even by implication. I certainly wouldn’t claim that all concepts which might be described with the same very general syntactic structure behave in the way I described; that would be a rather bizarre claim, wouldn’t it?
EDIT:
I made no reference to the dictionary definition of anything, though.
The point I was trying to make was that for instance in your open faced sandwich example, there are many people who would say it’s still obviously a sandwich, just like the duck is still a duck.
However, I realized I don’t know enough about the philosophy of language to meaningfully contribute to the discussion, so I’m going to bow out.