Whamit!

Speaker: Itai Bassi and Justin Colley (MIT).
Title: “P-word integrity”: a new condition on ellipsis
Time: Tuesday, October 8th, 1pm - 2pm
Location: 32-D461

Abstract: We observe a curious restriction on the distribution of pronouns in ellipsis contexts, which can be illustrated with the following examples:

1) a.  I called Sheryl on Monday, and called her on Tuesday again.

    b.*I called Sheryl on Monday, and her on Tuesday again.

    c.  I called Sheryl₁ on Monday, and HER₂/HIM/Mary on Tuesday

The generalization that emerges is that if a head is gapped, its complement (if overt) must be contrastively focused. But we will show that the story is more complicated than that. We will propose that the facts follow from a phonological constraint on ellipsis: don’t elide sub-parts of a string that forms a phonological word. We will discuss the consequences of such a principle for the theory of ellipsis.

This is work in progress and we would appreciate any feedback!

Speaker: Nabila Louriz (Hassan II, Casablanca) and Michael Kenstowicz (MIT)
Title: Verb formation in French loanwords in Moroccan Arabic
Time: Wednesday, October 9th, 5pm - 6:30pm
Location: 32-D831

Abstract: This presentation deals with verb formation of borrowings from French into Moroccan Arabic. The purpose is to examine the way the morphology of a Semitic language operates on loanwords from a Romance language. Two categories of borrowed verbs are distinguished in the corpus under investigation. First, there are verbs that are directly borrowed from French verbs (e.g. Fr. caler> MA [kala] “to stall”). The second category includes denominal verbs (e.g. Fr. Joint> MA [ʒwan] “joint”> [ʒwana] “to join”, Fr. cravatte> MA [gRafaTa] “tie”> [gəRfəT] “to wear a tie”). The first and second examples seem to operate on the base form of the French verbs and the borrowed noun, respectively. On the other hand, the third example appears to operate on the consonantal root extracted from the borrowed noun. This paper will explain this asymmetry and attempts a unified account highlighting the interface of morphology and phonology in the formation of loan verbs.

Speaker: Ethan Wilcox (Harvard)
Title: Neural Network Models and the Argument from the Poverty of the Stimulus: The Case of Filler—Gap Dependencies
Time: Thursday, October 10th, 12:30pm - 1:50pm
Location: 32-D461

Abstract: Recurrent Neural Networks (RNNs) are one type of neural model that has been able to achieve state-of-the-art scores on a variety of natural language tasks, including translation and language modeling (which is used in, for example, text prediction). In this talk I will assess how these models might way in to linguistic debates about the types of biases required to learn syntactic structures. By treating these models like subjects in a psycholinguistics experiment, I will demonstrate that they are able to learn the filler—gap dependency, and are even sensitive to the hierarchical constraints implicated in the dependency. Next, I turn to “island effects”, or structural configurations that block the filler—gap dependency, which have historically played a role in the Argument from the Poverty of the Stimulus. I demonstrate that RNNs are able to learn some of the “island” constraints and even recover some of their pre-island gap expectation. These experiments demonstrate that linear statistical models are able to learn some fine-grained syntactic rules, however their behavior remains un-humanlike in many cases.

Speaker: Daniel Asherov (joint work with Ezer Rasin & Roni Katzir)
Location: 32-D461
Time: Friday, October 11, 2:00-3:00pm
Title: Learning alternations from allophonic distribution

When learning a phonology of a language, the learner faces the task of discovering the language’s phonological alternations: the ways in which the shape of morphemes changes in a predictable way, usually to accommodate phonotactic requirements of the language. 

For example, in the grammar of Yamato Japanese s alternates with ʃ: the (bound) morpheme /hanas/ surfaces with ʃ before i[hanaʃ-imasu ‘talk non-past-polite’] and s otherwise [hanas-u ‘talk non-past’]. This alternation serves to repair the sequence si, which is prohibited in Japanese.

Computational learners of phonology make conflicting predictions with respect to the type of evidence required for positing an alternation such as the s~ʃ alternation in Japanese. 

One group of learners does not take economy into consideration (e.g. RIP/CD, Tesar & Smolensky 2000). Given a set of surface forms with the Yamato Japanese pattern, they are expected to correctly detect the prohibition on si. However, they can only learn the correct repair (s → ʃ /_i) if they are presented with direct evidence for the alternation in form of non-identical occurrences of the same morpheme (e.g. hanas-u + hanaʃ-imasu), with the information that these forms belong to the same paradigm.

Another group of learners is one where learners balance restrictiveness and economy (e.g. MDL, Rasin & Katzir 2016). Such a learner will only posit the prohibition *si when it would make the grammar and the description of the data more compact. In the case of a language like Yamato Japanese, the grammar can be made more compact by eliminating the sound ʃ from underlying representations and deriving it in the phonology. Such a learner predicts that it should be possible in principle to infer both the phonotactic prohibition (*si) and the correct repair (s → ʃ /_i) based on only Japanese-like surface forms, with no direct evidence for the alternation.

I will present an artificial language study with adult human learners which aims to test the applicability of these computational learners to human learning in a lab setting. I will start by presenting the design and results of a pilot study testing the conflicting predictions of the computational learners. We will then discuss a new design for a follow-up experiment.

References

Rasin, E., & Katzir, R. (2016). On Evaluation Metrics in Optimality Theory. Linguistic Inquiry, 47(2), 235–282.

Tesar, B., & Smolensky, P. (2000). Learnability in optimality theory. MIT Press.

A group of MIT linguists spent the last two weekends in the Passamaquoddy communities of Motahkomikuk and Sipayik (Indian Township and Pleasant Point), meeting with elders and learning about the Passamaquoddy language.  We’re very grateful to Newell Lewey (MITILI alum), Roger Paul (current MITILI student) and their families for introducing us to amazing people and places, and to the elders for giving us so much of their valuable time.  We’re looking forward to heading back north!  

The participants in the trips were Tanya Bondarenko (who took the attached photos), Colin Davis, Yadav Gowda, Peter Grishin, Tracy Kelley, Anton Kukhto, Cora Lesure, Elise Newman, Roger Paul, Norvin Richards, Ruoan Wang, and Stanislao Zompi.