ling 610

Clarification about squib requirements

I got confused about scheduling --- I thought there were more weeks left in the semester than there are.  So the original December schedule was an impossible schedule. I've fixed it.

We'll have no more regular assignments. Your remaining duties are two:

1. A draft of the squib to me sometime between now and December 15.
2. The final version of the Squib on December 15.

A comment on point 1: the closer your draft is to December 15, the less help I can be with it, due to time constraints. I think the ideal time to turn in the draft is around December 5.

A comment on point 2: I am going to be somewhat unforgiving about late squibs, and I will grant incompletes only under very special conditions. I am being strict because I think it is in your best interests to get this done on time and be done with it. It's a small thing. Get it behind you so that you can concentrate on bigger things.

Multiverses and possible-worlds realism

This is spooky stuff ...

Here is a link to the Scientific American article on multiverses and possible-worlds realism:

Max Tegmark. May 2003. Parallel universes. Scientific American.

(Please let me know if that link breaks. It's not a home copy.)

Wikipedia has a cool article on multiverses.

Intensionality: Some topics for discussion

Intensionality: Some topics for discussion

This is the handout that I plan to use to guide us through the basics of attitude predicates (and intensionality more generally). I'll distribute it today in class.

Assignment 7 (belatedly)

Amy Rose kindly reminded me that I forgot to post a copy of Assignment 7 (which is due today, but which I gave out in hardcopy last week). Apologies. Here it is (yay!):

Assignment 7

The neural basis for Generalized Quantifier comprehension

Corey T. MacMillan, Robin Clark, Peachie Moore, Chrsitian Devita and Murray Grossman from Upenn have very interesting data with respect to the neural basis for the comprehension of Generalized Quantifiers. Using fMRI, they find that a particular area in the brain is more activated when first order-order quantifiers (at least three) are processed than when higher-order quantifiers (more than half) are. More specific, the data show in line with their hypothesis that processing higher-order quantifiers involves a more complicated strategy (i.e. using a "numerosity component to identify the number property, a mechanism that manipulates these numbers, and a working memory component that retains the properties while they are being manipulated"). In other words, they find (if I understand them correctly) that the area that is involved with working memory is more activated in the case of higher-order predicates since you can't only rely on the quantity mechanism. For more details, Download clark.pdf

Erik-Jan

En route to UCLA on Thursday

Until recently, I totally flaked on the fact that I can't be in class on Thursday because I will be en route to Los Angeles to give a colloquium at UCLA. I've known about this trip since the semester's start. So much for being in touch with the entailments of my beliefs...

What to do?  Here is a suggestion: On Thursday, you guys meet at the usual time and place for a self-led discussion. I'll supply you with some questions, topics, and the like beforehand, and you can just see where the material takes you.

Will this work? Or would it be more productive to give you the day off?  If we take the day off, should we try to schedule a make-up?

Let me know what you think --- in the comments section of this message or by email.

Readings on propositional attitudes

After Thanksgiving, we'll study propositional attitude predicates. I've compiled a packet of essential readings on this topic. I'll distribute it in class tomorrow, but you might like to have an electronic copy as well (password protected):

Readings on propositional attitudes and intensionality

We'll have five classes for this material. (There are six after Thanksgiving, but one will be devoted to exploring the entity domain.)

Thus far this semester, we have moved fast, trying out useful tools and ideas without pausing long enough to decide whether they are keepers. For this final unit, we'll slow down --- digging as deep as we can into this topic given the time we have.

There is a lot of material in the above packet. I'll have some specific advice tomorrow about how to study it. My lectures and our discussions will be designed to make this study more rewarding. But everything in the packet is worth reading again and again, year in and year out. This will be just a start.

December overview (with November update)

I've posted an overview of December, and I've updated the overview of November.

Summary of assignments:

• We'll have a short assignment over Thanksgiving (distributed on Tuesday).
• After Thanksgiving, the week's assignment is a squib draft.
• One more regular assignment after that (#8), then you have a nine days for squib writing. (The final draft is due December 15).

And we'll fill our final meetings with an in-depth look at attitude predicates.

Exam 2 commentary

My answers to exam 2.

We won't have time to go through this in class, but please contact me with any questions you have about it.

Handouts on quantifiers and type-shifting

Two handouts for tomorrow, both of which I will distribute in class:

Quantifier scopes via lexical shifts
[This handout shows how to capture clausemate scope ambiguities by changing the verb's meaning but leaving the surface syntax alone. Page 2 compares this account with a QR account.]

Partee's nominal type-shifting functions
[This presents the essentials of Partee's classic 1987 paper. It will have to stand as our only look at type-shifting. I hope it conveys a sense for what these tools are like.]

I've added Partee 1987 to the archive of downloadable papers.

Handout: Quantifier basics

A handout on the basics of quantifiers:

• their types
• some meanings for quantifier determiners
• approaches to quantifiers in object position
• scope ambiguities
• monotonicity (with a dash of polarity item licensing)
• conservativity

Exam 2

Exam 2.

I'll hand this out today in class. It is due November 17 (this Thursday).

Horn on Seuss on Grice

Larry Horn just sent me a copy of Dr. Seuss's pean to Grice (click the image for a PDF of the illustrated page):

And Larry has a commentary on this poem (click here, or the image, for the full text [password protected]):

Deliberately vague pronominals

Tom Roeper and I are coauthoring a paper right now.  Here is the latest version of the opening. The passage is of interest for us because the denotations of the pronouns (and in turn the phrases that contain them) are deliberately vague between referring to Tom and referring to me (Chris).

If I forget to buy a crucial item on my grocery list, I might, frustrated with myself, mutter one of the sentences in (1) to express mild self-disapprobation.

(1)	a.	Oh, you fool!
	b.	You idiot!
	c.	You nincompoop/dumbass/screwball!

If my coauthor overheard my muttering, he would recognize that I was in a heightened emotional state, but he would do me a disservice if he reported to others that I thought I was a fool (idiot, nincompoop).  I did not, after all, use any of the fully sentential forms in (2).

(2)	a.	Oh, you are a fool!
	b.	You are an idiot!
	c.	You are a nincompoop/dumbass/screwball!

With the sentences in (1), I express a momentary attitude linked to a situation. With the sentences in (2), if I intend myself to be the addressee, I characterize myself in more general terms.

Downloadable squibs

I added some squibs to the downloadable-paper collection: The two by Karttunen that I handed out today, Fiengo and Lasnik's 'Nonrecoverable deletion', and 'Metalinguistic autoreference' (the one-liner).

Drop me a note if you need the password again.

Information we didn't know was there

A notice posted in the toilets at my friend's workplace reads as follows:

"Please throw hygienic products into the basket! Or anything else that might clog it..."

We see a pronoun that refers to an entity not present in the preceding linguistic context, but prominent in the actual situation (the extra-linguistic context) of the utterance. This raises some questions:

• How should we enrich our model of the context to account for pronoun resolutions in cases like these?
• Can we always randomly refer to entities in our immediate context in a similar fashion?

Defining the context

With respect to the Kaplan-based theory of indexicals we’ve used in class, it has been worrying me that the members of the context tuple are defined in such a way that the notion of context itself is included in the definitions of its components. So, to determine the denotation of the first person indexical 'I', we refer to the “the speaker of the context” (c_S), or, in the case of location-referring indexicals like 'here', to “the location of the context” (c_L). This seems to work, up to a point, but what seems to me to be crucially lacking from definitions formulated in this manner is any explicit notion of how to define the “context” in the first place, an issue which came up for example, in considering the question “Multiple Indexicals” on Assignment 5. In reply to another question posted at this weblog regarding the need for more complex context-determined expressions within the model (“The place of the context” 3 Nov), Chris replied that we will in fact need to appeal to the common ground, as well as to the QUD to understand “(the linguistically important parts of) the context”. Intuitively, the notion of context seems a purely pragmatic one, but these latter two notions seem to me to be aspects of the context that are of a fundamentally different nature than those implicated in interpreting indexicals. Which brings me to my next, slightly broader question, of what exactly the status of the “context,” as defined in a Kaplanian theory, is. Is it pragmatics, or is it part of the semantics? If the former, isn’t it problematic for our semantics that we can’t interpret any indexical-containing expressions without direct recourse to the context? How is it possible to maintain the sort of strict division between semantic (eg entailments, set relations) and pragmatic (eg implicature and, arguably, presupposition) meanings upon which a robust theory of pragmatics relies if we must refer directly, in at least this one case, to pragmatic information in order to compute semantic meaning?  So-called “dynamic semantic” theories such as DRT and File Card Semantics seem to incorporate into their frameworks elements of apparently pragmatic meaning, which has often led me to wonder why they are conceived of as semantic theories at all. Am I missing some essential aspect of the working of such theories, or is it the case that their existence simply point to the fact that a strict semantic/pragmatic division such as the one described above is in actuality not a viable one?

Whaleship Abstraction?

Semantics at The Lady Killigrew Cafe?

[Thanks Rajesh!]

a puzzle

Traditionally, pragmatics has been crucial to giving meanings to sentences like the following:

 1. I am the person who is spilling the sugar.[1]

 The meaning of the word “I”, on the traditional model, is the speaker of the utterance within a particular context assignment.

 Now consider the following scenario. Suppose that John is grocery shopping. The bag of sugar he has placed into his shopping cart is torn, and a slow trickle of sugar is spilling from the bag onto the floor of the grocery store. Suppose further that John, while backtracking through an aisle of the store, notices a small trail of sugar on the ground. He does not know that it is his bag that is ripped. Since John is somewhat of a good Samaritan, he decides to follow the trail of sugar to find the person spilling it, and let that person know of the problem. As he follows the trail, he utters:

 2. I should find the person who is spilling the sugar.

 He follows the trail around the store, and notices that in a short amount of time, the trail is getting thicker. In light of this observation, John utters,

 3. The person who is spilling the sugar is just ahead of me.

He speeds through the store, thinking the culprit is just around the next corner. After his third pass through the same aisle, John stops. He looks behind him, and notices that the trail is slightly thicker than the one he sees when he looks forward. All of a sudden, John has a realization, and utters sentence 1.

Notice, however, that when we give analyses to sentences 1-3 something strange happens. The semantics will assign the definite description “the person who is spilling the sugar” to John. As I said above, the context will assign “I” to the speaker, namely John. Thus we can use simple definitional substitution to derive the following:

1’. John is John.

2’ John should find John.

3’ John is just ahead of John.

 If we factor into account John’s belief context we run into serious trouble. Now assume the disquotation principle (When S utters p, S sincerely believes p). But when John utters 2, he does believe 2, but does not believe 2’. John does not believe he should find himself. When John utters 3, he believes 3, but does not believe the impossible claim purported by 3’. Finally when John utters 1, something happens in his conscious mind, something not equivalent to the trivial fact that John is identical with himself. Rather, John has realized something.

 Given this little puzzle, I put forth the following cluster of questions:
What went wrong? Did either the pragmatic or semantic elements of the analyses get something wrong? Should it be the job of semantics, pragmatics, or both to give a solution to this puzzle? Does there need to be special mode of analysis for belief contexts? Sould uncovering such a mode of analysis be a linguitic concern?   

---

[1] The example is taken, and perhaps slightly modified, from John Perry’s The Problem of the Essential Indexical and other Essays, Oxford University Press, 1993.

Irene Heim's colloquium

Irene Heim's colloquium on Friday, 'Features on anaphoric pronouns', was really well timed for us. This post is for questions and comments relating to her talk.

Bare plurals in context

We know that the pragmatics we are developing allows us to disambiguate several indexicals and this kind of things. For know, we are putting everything in the context. But I wonder if there is more stuff we can specify in the context.

There are several things that belong to the environment of the type of discourse and that out of it, in another discourse, would be ungrammatical. For example, working on bare plurals I found that next phenomena:

1. (*)Padres  de alumnos discutían sobre  el  asunto a  la  entrada  del  colegio.

parents of students  talked   about  the topic  at the entrance of the school

‘Some parents of the students were arguing about the topic at the entrance of the school’

2. (*)Representantes sindicales se      reunieron con  el  comité de empresa para discutir el

representatives syndical  se meet.pl.3 with the committee of enterprise to discuss the

problema.

issue.

‘The/some syndical representatives met with the enterprise committee to discuss the issue’

  (1) and (2) are perfectly fine in the news. Any reporter would utter those when giving the highlights.. But those are pretty odd in any other situation. Actually, the journalist would never utter them after the highlights (different rules in the same context?, two contexts in a context?, or two different utterances in an utterance?). How can we address that. Can we? Furthermore, do we want to?

here here

I'm wondering how to classify a kind of 'here' which seems midway between indexical and deictic. The instances I'm thinking of are when we have a group of people from one place, say, UMass, who then travel to another place, say, Boston, and one of them says: "Leah moved here from Boston" (the meaning being that L moved from Boston to the greater UMass area).

'Here' in this context doesn't require any pointing on behalf of the speaker, and it doesn't refer to the place of utterance. It seems more that speaker and addressee are pretending they're in a more typical place of utterance (the typical location being Amherst, not Boston), and the 'here' is referring in virtue of this pretense. What do we do about such cases?

The larger implication concerns the interface of semantics and pragmatics w.r.t. the theory of proforms. In our semantic interpretation, we treat some as indexicals, some as deictic elements, some as free variables. But it seems that the pragmatic background can interfere with this neat trichotomy and give us neither-here-nor-there cases like the 'here' above.

Pragmatics and the language faculty

My question about pragmatics is really quite general: what is the status of “pragmatic competence” vis-à-vis the notion of a language faculty? We are usually encouraged to think of the language faculty as essentially modular, in Jerry Fodor’s sense (domain specific, informationally encapsulated, etc.). This seems quite plausible with regard to phonology and syntax. It is hard to see how, e.g., the speaker’s application of Principle C could possibly be taken to reflect some more general cognitive capacity, or how information from other parts of the cognitive system could possibly affect the application of the binding principles in comprehension or production. The same could plausibly be said of semantics; for example, the composition rules FA, PA, and PM seem to be domain specific, if anything is. 

 But what about pragmatics? It seems plausible that any being that has both a well-developed “Theory of Mind,” in the sense in which this term is used by developmental psychologists and animal behaviorists, and an assumption of cooperativity, will apply the Gricean maxims. For example, it would not be too surprising to find that animals to which we want to attribute communication, but not language, apply Gricean maxims in interpreting the communicative attempts of their fellow species members. If this is right, then why do we consider pragmatics to be part of linguistics?

The place of the context...

Grice suggests that generalized conversational implicatures can be cancelled either explicitly (though the addition of a clause) or as a consequence of the context.  If the context plays a role in determining the felicity of various implicatures, do we need to somehow include a more explicit understanding of the context within our model?  Is it ever necessary to include features or variables (beyond indexicals) within our composition that require reference to a more explicit context?

Implicatures in the wild

Masashi writes in with a relevance/quantity implicature from real life:

I heard an episode about Brahms which is like the Car Talk example in Section 9.2.1. A young composer handed in his compositions to Brahms for comments. After looking at them in silence for a while, Brahms said, "Where did you buy this music paper?"

Or and Calculating the conversational implicature

I failed to follow the reasoning d. in Section 10.12. If Speaker B “is not in a position to know for certain that p obtains, that q obtains, or that (p ∧ q) obtains”, then it seems that he can only narrow down to three of the four possible worlds without violating quality.

Related to it, another question arises. Suppose that Speaker B's answer is (p ∧ ¬q), “Ali is intelligent and Ali is not hard-working”. Comparing with a shorter answer, p (“Ali is intelligent”), the former is stronger, but sounds odd. The second half of Grice's maxim of quantity, "(2) Do not make your contribution more informative than is required" will come to account for it, but how can we make explicit calculation of the appropriateness of informativeness?

Assignment 5 commentary

My answers to Assignment 5.

November overview

This page provides a broad overview of November.

After we get through quantification and type-shifting (two essentials), we have some time to play with.

For the Tuesday before Thanksgiving, I scheduled some time for exploring the entity domain: kinds, plural sums, events, situations, etc. It will have to be fast and somewhat self-contained, but I hope it will at least make you aware of what people are doing with D_e.

After that, we have six classes. I want to use them to cover one or two topics in depth. Your input would be greatly appreciated. My hope is that we can address issues that are relevant to people's final squib-like projects. This is another motivation for question 3 of Assignment 6.

Quality and quantity ratings

A handout on quality and quantity ratings.

This handout is basically an excerpt from one of the talks I gave at ZAS in September. Here is a link to the fuller version. And here is a link to the associated paper.

Assignment 6

Assignment 6. This is technically due on November 8, but see the sheet itself for some suggestions about how to proceed.

This one's unlike the rest --- it's a demand for a question/hunch, a reading assignment, and a lowkey beginning to the final squib-like project.