So, perhaps:
Gorrilla A saw woman W who fed man C and man D
Gorilla B saw woman W who fed man E and man F
— is what is not possible, right? The men have to be either C and D or E and F, so, they are different from each other (C is not D) but they must be the same set from the point of view of both A and B. W is the same either way.
Thanks for the question.
Here’s what Carlson says about the example himself:
This sentence “has no reading describing a situation in which Koko the gorilla saw a woman who fed John and Kong the gorilla saw a woman who fed Paul.”
So the intended reading could involve only two men. And this isn’t made explicit in the quote, but the woman who fed John and the woman who fed Paul don’t have to be the same.
Thank you for your interesting presentation, and for the challenging data on slide 14 – “not obvious” is a very appropriate characterization.
I would not rely too much on Beck’s distinction based on German “ander-” and “verschieden-“: The data are much less clear than it seems at first, especially in spoken language and in colloquial German. There does seem to be some difference between the two versions of German “different”, but it does not completely align with the semantic distinction that is being made. It might be important to note that Beck’s paper was written before corpora became available.
Have you looked into David Lahm’s HPSG account of “different”? How does the coverage of his analysis compare to yours?
Okay, I’ll look into the difference between ander and verschieden in more detail.
Yes, I’ve read David Lahm’s paper in the SALT 26 proceedings. What is presented in that paper is a plausible account of singular different, and it isn’t meant to apply to plural different, so the empirical coverage of his theory and that of my theory are disjoint.
I’m not sure. I’m not entirely convinced that “different” in this sentence is anteceded by “the two students”. I wonder if the antecedent of “different” here couldn’t be the unexpressed subject of “solving”.
I’m still not convinced. Are you sure this “different” is not what I call a simple, normal “different”, i.e. an adjective that means something like “various”? In other words, can this sentence be used in a situation where there are only two kinds of problem, say Kind A and Kind B, and Student 1 has a talent that helps us with problems of Kind A, and Student 2 has a talent that helps us with problems of Kind B?
(Incidentally, the “different” in this example is arguably what I call the singular different, if the floating quantifier “each” in the example cannot be elided. But the singular different is supposed to be sensitive to scope islands just like the plural different.)
Yes, I intended such a situation. But I agree that it is degraded without “each”.
In that case, it does seem that relative clauses are not always scope islands. Thank you.
Telugu, a South Indian language uses reduplication as a strategy for getting ‘the internal reading’. The word ‘veeree’ (different) has ‘the external reading’ while reduplicated ‘veeree veree’ has ‘the internal reading’.
Thanks a lot for the information. Am I correct in guessing that “veeree veree” can have an internal reading only when its antecedent is a plural expression like “Anna and Bill”, and not when a singular expression corresponding to something like “every student”?
Yes. With only plural. A sentence like “Every student read veeree veeree book” is ungrammatical. Even “Each student read veeree veeree book” is not possible.
But the question is : How do we represent the simple and the reduplicated ones in the Lexicon? Do we need a separate lexical entry or have a morphological rule (reduplication) which also brings in the semantic change?
Thank you. That’s an interesting question. In his 2011 Linguistics and Philosophy paper, Brasoveanu claims that the singular “different” (veeree in the case of Telugu) crosslinguistically tends to be the same lexical item as the “different” that receives external readings. That leads him to propose separate lexical entries for the singular different and the plural different, and I’ve been assuming that that’s the right way to go. But in languages where there is an obvious morphological similarity between the singular different and the plural different, it may make sense to try to derive one lexical entry from the other.
Brasoveanu’s paper has an appendix that shows how singular different and plural different are expressed in various languages, and for what it’s worth, Japanese is another language in which the plural different (“betsu-betsu no”) looks like a reduplicated form of external-reading different (“betsu no”). (I’m not sure if the latter can be given an internal reading even when the antecedent is singular.)
The question I wanted to ask was: could we use just one analysis of “different”, so that we don’t need the “normal adjective”? Would something go wrong if we use the lexical entry proposed in the talk?
Thanks a lot for the question, and I’m sorry I couldn’t answer it. Amazingly, I accidentally hit the button on my laptop that disables the microphone while fumbling to turn up the volume!
I think we need to give “different” a semantically simple lexical entry too anyway. Consider a sentence like “They are different songs”, the intended meaning being “These songs are different from each other”. The lexical entry that I proposed for the plural different cannot be used in a sentence like this, on the assumption that the expression immediately following a copula is not a quantifier.
By the way, is the work on quantifier coordination that you mentioned already available as a written work?
It wouldn’t be a virtual conference without a few technical mishaps!
I see, maybe it depends on other analyses. I guess I was imagining something like: some(X, songs(X), every(y, y in X, every(z, other(z, X, y), different(z, y)))).
I haven’t written that work up yet.
I use a scope DAG in the following forthcoming work, but the re-entrancy in the DAG isn’t the main point of the paper… in fact, the example with a DAG is relegated to the appendix! (Figures 2 and 4 should help to explain the notation in the appendix.)
Let me clarify what I meant to say using the example “They are different songs”. If the phrase “different songs” in this example is interpreted as a generalized quantifier, then the verb “are” will have to be interpreted as claiming identity between the denotation of the subject and the denotation of the complement. (And if we are to use the lexical entry for the plural different that I proposed, the sentence as a whole will have to be interpreted as saying “Each member of the contextually given set is identical to some song that is different from all the songs that are identical to some other member of the set”.) This strikes me as counterintuitive; the sentence doesn’t sound like it’s asserting some identity.
Thanks a lot for the information about the paper. I’ll study it.
Thank you for the clarification. I don’t think there is any such work on Telugu quantification. I was only reacting to your examples. But it’s very interesting. I can think of doing some work on it.
Thank you for the clarification. It’s nice to interact with you. I don’t think there is any such work on Telugu quantification. I was only reacting to your examples. But it’s very interesting. I can think of doing some work on it.
???
Thank you, Shuichi!
On slide 29, did you mean to say: “The two **women** cannot be different from each other”? Or did I misunderstand.
Yes I think I misunderstood…
So, perhaps:
Gorrilla A saw woman W who fed man C and man D
Gorilla B saw woman W who fed man E and man F
— is what is not possible, right? The men have to be either C and D or E and F, so, they are different from each other (C is not D) but they must be the same set from the point of view of both A and B. W is the same either way.
Thanks for the question.
Here’s what Carlson says about the example himself:
This sentence “has no reading describing a situation in which Koko the gorilla saw a woman who fed John and Kong the gorilla saw a woman who fed Paul.”
So the intended reading could involve only two men. And this isn’t made explicit in the quote, but the woman who fed John and the woman who fed Paul don’t have to be the same.
Hi Shûichi,
Thank you for your interesting presentation, and for the challenging data on slide 14 – “not obvious” is a very appropriate characterization.
I would not rely too much on Beck’s distinction based on German “ander-” and “verschieden-“: The data are much less clear than it seems at first, especially in spoken language and in colloquial German. There does seem to be some difference between the two versions of German “different”, but it does not completely align with the semantic distinction that is being made. It might be important to note that Beck’s paper was written before corpora became available.
Have you looked into David Lahm’s HPSG account of “different”? How does the coverage of his analysis compare to yours?
Okay, I’ll look into the difference between ander and verschieden in more detail.
Yes, I’ve read David Lahm’s paper in the SALT 26 proceedings. What is presented in that paper is a plausible account of singular different, and it isn’t meant to apply to plural different, so the empirical coverage of his theory and that of my theory are disjoint.
Would the following be a counterexample to the island constraint?
The two students each have a talent which comes out when solving different kinds of problem.
I’m not sure. I’m not entirely convinced that “different” in this sentence is anteceded by “the two students”. I wonder if the antecedent of “different” here couldn’t be the unexpressed subject of “solving”.
I see. How about:
The two students each have a talent that proves useful for different kinds of problem.
… except “useful” arguably also has a dropped argument (“useful for the student”). How about:
I try to solve problems collectively with my two students. They each have a talent that helps us with different kinds of problem.
I’m still not convinced. Are you sure this “different” is not what I call a simple, normal “different”, i.e. an adjective that means something like “various”? In other words, can this sentence be used in a situation where there are only two kinds of problem, say Kind A and Kind B, and Student 1 has a talent that helps us with problems of Kind A, and Student 2 has a talent that helps us with problems of Kind B?
(Incidentally, the “different” in this example is arguably what I call the singular different, if the floating quantifier “each” in the example cannot be elided. But the singular different is supposed to be sensitive to scope islands just like the plural different.)
Yes, I intended such a situation. But I agree that it is degraded without “each”.
In that case, it does seem that relative clauses are not always scope islands. Thank you.
Telugu, a South Indian language uses reduplication as a strategy for getting ‘the internal reading’. The word ‘veeree’ (different) has ‘the external reading’ while reduplicated ‘veeree veree’ has ‘the internal reading’.
Thanks a lot for the information. Am I correct in guessing that “veeree veree” can have an internal reading only when its antecedent is a plural expression like “Anna and Bill”, and not when a singular expression corresponding to something like “every student”?
Yes. With only plural. A sentence like “Every student read veeree veeree book” is ungrammatical. Even “Each student read veeree veeree book” is not possible.
But the question is : How do we represent the simple and the reduplicated ones in the Lexicon? Do we need a separate lexical entry or have a morphological rule (reduplication) which also brings in the semantic change?
Thank you. That’s an interesting question. In his 2011 Linguistics and Philosophy paper, Brasoveanu claims that the singular “different” (veeree in the case of Telugu) crosslinguistically tends to be the same lexical item as the “different” that receives external readings. That leads him to propose separate lexical entries for the singular different and the plural different, and I’ve been assuming that that’s the right way to go. But in languages where there is an obvious morphological similarity between the singular different and the plural different, it may make sense to try to derive one lexical entry from the other.
Brasoveanu’s paper has an appendix that shows how singular different and plural different are expressed in various languages, and for what it’s worth, Japanese is another language in which the plural different (“betsu-betsu no”) looks like a reduplicated form of external-reading different (“betsu no”). (I’m not sure if the latter can be given an internal reading even when the antecedent is singular.)
The question I wanted to ask was: could we use just one analysis of “different”, so that we don’t need the “normal adjective”? Would something go wrong if we use the lexical entry proposed in the talk?
Thanks a lot for the question, and I’m sorry I couldn’t answer it. Amazingly, I accidentally hit the button on my laptop that disables the microphone while fumbling to turn up the volume!
I think we need to give “different” a semantically simple lexical entry too anyway. Consider a sentence like “They are different songs”, the intended meaning being “These songs are different from each other”. The lexical entry that I proposed for the plural different cannot be used in a sentence like this, on the assumption that the expression immediately following a copula is not a quantifier.
By the way, is the work on quantifier coordination that you mentioned already available as a written work?
It wouldn’t be a virtual conference without a few technical mishaps!
I see, maybe it depends on other analyses. I guess I was imagining something like: some(X, songs(X), every(y, y in X, every(z, other(z, X, y), different(z, y)))).
I haven’t written that work up yet.
I use a scope DAG in the following forthcoming work, but the re-entrancy in the DAG isn’t the main point of the paper… in fact, the example with a DAG is relegated to the appendix! (Figures 2 and 4 should help to explain the notation in the appendix.)
https://arxiv.org/abs/2006.03002
Let me clarify what I meant to say using the example “They are different songs”. If the phrase “different songs” in this example is interpreted as a generalized quantifier, then the verb “are” will have to be interpreted as claiming identity between the denotation of the subject and the denotation of the complement. (And if we are to use the lexical entry for the plural different that I proposed, the sentence as a whole will have to be interpreted as saying “Each member of the contextually given set is identical to some song that is different from all the songs that are identical to some other member of the set”.) This strikes me as counterintuitive; the sentence doesn’t sound like it’s asserting some identity.
Thanks a lot for the information about the paper. I’ll study it.
Thank you for the clarification. I don’t think there is any such work on Telugu quantification. I was only reacting to your examples. But it’s very interesting. I can think of doing some work on it.
Thank you for the clarification. It’s nice to interact with you. I don’t think there is any such work on Telugu quantification. I was only reacting to your examples. But it’s very interesting. I can think of doing some work on it.