Whamit!

The Weekly Newsletter of MIT Linguistics

Archive for the ‘Talks’ Category

Phonology Circle 5/14 - Koichi Tateishi (Kobe College/MIT)

Speaker: Koichi Tateishi (Kobe College/MIT)
Title: Trimoraicity and Monomoraicity: Cases in Japanese
Date/Time: Monday, May 14, 2018, 5:00-6:30pm
Location: 32-D831
Abstract:

This presentation is about Japanese syllables, which appear to be one of the best-studied areas in phonology, starting from McCawley (1968) and traditional Japanese linguistics papers preceding it. I will point out that /N/, a moraic nasal, which is never a nuclear component of a syllable and hence can never be accented according to McCawley and later works, actually stands out as an independent syllabic nucleus at some morphological peripheries. This syllabic /N/ can be accent-bearing and can undergo Initial Lowering, another signatory tonal phenomenon that is typically observed only with a syllabic nucleus. The presentation also points out that the syllabic /N/ is only for the borrowings and mimetics, while we find an independent phenomenon in the Yamato (native Japanese) stratum that derives a string that appears to derive a syllabic /Q/, moraic obstruent, and that this constitutes a counterargument to Ito and Mester’s (1995) strata-dependent reranking hypothesis.

Share

Phonology Circle 5/7 - Thomas Schatz and Naomi Feldman (UMD/MIT)

Speakers: Thomas Schatz and Naomi Feldman (UMD/MIT)
Title: A simple framework to study how phonological structure can emerge from the interaction of social, physical and cognitive evolutionary pressures
Date/Time: Monday, May 7, 2018, 5:00-6:30pm
Location: 32-D831
Abstract:

Nowak and Krakauer (1999) proposed a framework to study how combinatoriality in language can emerge from evolutionary pressures to communicate in the presence of noise in the communication channel. I will present this framework and discuss possible extensions that might lead to functional accounts for certain phonological phenomena. I will focus in particular on an extension of the framework that adds a pressure to limit the production costs of words in the language, for which I will present a few preliminary results.
This is very preliminary work in collaboration with Matthias Hofer and Naomi Feldman. The main object of the presentation will be to get feedback on the potential of the framework and to advertise the project to students with a formal background in phonology - which both Matthias and me lack - who might be interested in collaborating with us.

Share

Syntax Square 5/8 - Carolyn Spadine (MIT)

Speaker: Carolyn Spadine (MIT)
Title: Evaluating Syntactic Approaches to Interrogative Flip: Test cases from English and Malayalam
Date and time: Tuesday May 8, 1-2pm
Location: 32-D461
Abstract:

“Interrogative flip” describes a phenomena in which elements that appear to orient to the speaker in a declarative utterance shift perspective and orient to the addressee in an interrogative context — evidentials, perspective-sensitive anaphora, modals, adverbs, predicates of personal taste, and others have been reported to show this behavior. In proposing a mechanism for encoding discourse-pragmatic information in syntax, interrogative flip is one of the core phenomena that Tenny and Speas 2003 intend to address, and the same problem has been subsequently taken up in Woods 2014, Zu 2018, and many others.

This talk presents preliminary work on two constructions that display interrogative flip, and examines the ways in which existing syntactic approaches to modeling interrogative flip account for or fail to account for this data. The first is discourse participant-oriented modifiers in English, as in (1):

1. a. [As a film critic], this movie deserves an Oscar.
b. [As a film critic], does this movie deserve an Oscar?

In (1a), the preferred and perhaps only interpretation of the bracketed constituent is that the speaker is a film critic, whereas in (1b), English speakers report both speaker- and addressee-oriented interpretations for the same constituent. A similar but more constrained pattern emerges for embedded instances of these modifiers, posing a challenge for some proposals. The second comes from a reportative evidential marker ennu (glossed as REP) in Malayalam (2a), which can either scope under or over the question particle, yielding two different interpretations — either a question about a report heard by the addressee (2b), or a declarative report of a question overhead by the speaker (2c).

2. a. prime minister varunnu ennu
prime minister come.PROG REP
“I heard that the Prime Minister is coming”
b. prime minister varunnu enn-oo?
prime minister come.PROG REP-Q
“Did you hear if the Prime Minister is coming?”
c. prime minister varunn-oo ennu
prime minister come.PROG-Q REP
“I heard someone ask if the Prime Minister is coming”

In both cases, I suggest the data supports the general pattern that existing proposals intend to account for, but also raise concerns about the specific structures proposed to implement them.

Share

LingPhil Reading Group 4/23 - on Schlenker (2012)

Title: on Schlenker (2012) 
Date and time: Monday April 23th, 1-2pm
Location: 32-D831

Vincent will be presenting Philippe Schlenker’s paper ‘Maximize Presupposition and Gricean Reasoning’.


As per usual, reading the paper is not mandatory, although feel free to read it if you’re feeling brave.

Share

Phonology Circle 2/4 - Jaehyun Son (Duksung Women’s University)

Speaker: Jaehyun Son (Duksung Women’s University)
Title: Pitch Accent Systems in Korean
Date and time: Monday, 23 April 2018, 5:00-6:30pm
Location: 32-D831
Abstract:

Research on the Korean accent has been carried out within the Korean linguistics community, but in that context, the Korean accent system has traditionally been compared to the tone system of Chinese, in which pitch contours are syllabic. In contrast, Japanese researchers have proposed that the Korean accent system should be analyzed from the point of view of word-level and phrase-level accentual systems seen in Japanese dialects. One possible reason for this difference of opinion is that recently in Japan, despite the growing influence of the accentual systems of Tokyo Japanese and the dialects of other major cities, a great variety of smaller dialects have been observed and documented, and as a result of this work researchers have discovered accent types that have played a crucial role in uncovering the history and evolution of the Japanese accentual system. In Korea, on the other hand, accent has been lost in the regions surrounding and including Seoul (the national capital) but there are still dialects, mainly in the south-eastern regions of the Korean peninsula, that retain an accentual system and can shed light on the history of accent in Korea. For the present study, I took the Japanese-oriented view rather than the traditional Chinese-oriented view and analyzed the accentual systems of Korean dialects using data from a purely synchronic field survey of several locations across the Korean-speaking region. The field survey includes dialects that have already been documented by Korean and Japanese researchers, but by including the whole Korean-speaking region in its scope and using a new theoretical framework, the current study was able to highlight the shortcomings of previous work. The current study presents the Korean accent types and their geographical distribution. Moreover, by comparing the various accent types, it was possible to look back and investigate how the Korean accent system has evolved up to the present day.
Share

CompLang 4/23 - Hendrik Strobelt (IBM)

Speaker: Hendrik Strobelt (IBM)
Title:  Visualization for Sequence Models for Debugging and Fun
Date and time: Monday, April 23, 5-6pm
Location: 46-5165
Abstract:

Visual analysis is a great tool to explore deep learning models when there is no strong mathematical hypothesis yet
available. I will present two visual tools where we used design study methodology to allow exploration of
patterns in hidden state changes in RNNs/LSTMs (LSTMVis) and exploration of Sequence2Sequence models (Seq2Seq-Vis).
Both model types have shown superior performance for NLP like language modeling or language translation.
Examples about both tasks will be shown on a variety of models.

As beautiful distraction, we also utilize data science methods to investigate large data in a more artistic way.
Formafluens is such a data experiment where we analyze a large collections of doodles made by humans in the Google
Quickdraw tool.

Share

LF Reading Group 4/17 - Hanzhi Zhu (MIT)

Speaker: Hanzhi Zhu (MIT)
Title: Conditionals in although constructions
Date and time: Wednesday, April 18, 1-2pm
Location: 32-D461
Abstract:

In this talk, I’ll be looking at biclausal constructions with although/even though which convey the truth of two propositions as well as the oddness of their juxtaposition:

1. John went out for a walk, even though it’s raining.
2. Although Bailey is rich, she doesn’t give to charity.

The link between although constructions and conditionals has been explored in previous accounts, in which “although p, q” is analyzed as presupposing “normally, if p then ¬q”. However, these accounts ignore the compositional contribution of even, which appears in these constructions in English as well as cross-linguistically. Lund (2017), borrowing from Guerzoni and Lim’s (2007) account of even if, proposes an account in which “although p, q” asserts a conjunction and has a scalar likelihood presupposition: ¬p and q is less likely/expected than p and q. I’ll present a counterexample to this account which favors having a presupposition even closer to Guerzoni and Lim’s proposal for even if: “although p, q” presupposes that if ¬p, q is less likely/expected than if p, q. I’ll also discuss further consequences of this proposal regarding the role of the additive presupposition of even.

Share

LingPhil Reading Group 4/9 - on Stalnaker 2004

Title: on Stalnaker (2004)
Date and time: Monday April 9th, 1-2pm
Location: 32-D831

This week’s paper is Stalnaker’s Assertion Revisited: On the Interpretation of Two-Dimensional Modal Semantics, available here

Christopher will be presenting the paper.

Share

Phonology Circle 4/9 - Yunjing Li (Tianjin\MIT)

Speaker: Yunjing Li (Tianjin Foreign Studies University & MIT)
Title:  Rule Interaction in Mandarin Tonal Phonology
Date and time: Monday, April 9th, 5:00-6:30pm
Location: 32-D831
Abstract:

In Mandarin Chinese, there are some rules governing tone sandhi changes. In some cases, more than one rule may be applicable, hence rule interaction occurs.

This talk will introduce some basic facts about Mandarin tonesand their representation, followed by a description of the interaction between the Third Tone Sandhi Rule and the Neutral Tone Rule in disyllabic words. The ordering of these two rules causes phonological opacity. An analysis in the framework of Harmonic Serialism is proposed.
Share

Syntax Square 4/10 - Sze-Wing Tang (The Chinese University of Hong Kong/MIT)

Speaker: Sze-Wing Tang (The Chinese University of Hong Kong/MIT)
Title: On the Syntax of Sentence-final Elements
Date and time: Tuesday, April 10, 1-2pm
Location: 32-D461
Abstract:

Insights of Ross (1970) of the analysis of the clausal periphery have been revived under the cartographic approach (Rizzi 1997, 2004, Cinque 1999, see also Speas 2004, Tenny 2006, Hill 2007, Miyagawa 2012, 2017, and Wiltschko and Heim 2016). The goal of this talk is twofold. First, it is argued that there should be two distinct syntactic layers in the clausal periphery that are dedicated to “grounding” and “responding” (Wiltschko and Heim 2016), respectively, by examining the grammatical properties of the Mandarin sentence-final particle (“SFP”) ma and Cantonese SFP ge and the “h-family”. Second, it is argued that some sentence-final expressions, such as tags in tag questions in English should be in the highest syntactic position and form a coordination structure with a silent head, in the sense of Kayne (2016). A hierarchical structure/ordering “Proposition > SFP > Tag” is proposed, which may serve as a working hypothesis to study the syntax of speech act cross-linguistically.

Share

LF Reading Group 4/11 - Frank Staniszewski (MIT)

Speaker: Frank Staniszewski (MIT)
Title: Wanting, Acquiescing, and Neg-raising
Date and time: Wednesday, April 11, 1-2pm
Location: 32-D461
Abstract:

I argue that neg-raised (NR) readings for negated sentences containing want are the result of want expressing an underlying weak (existential) quantificational force, which gives rise to the globally strong meanings under negation. To derive the universal interpretation that is attested for non-negated want, then, I adopt Bassi & Bar-Lev (2016)’s treatment of bare conditionals, and hypothesize that want undergoes strengthening in a manner analogous to Free Choice disjunction, as analyzed in Fox (2007).

As evidence for this view, I examine a puzzling paradigm discussed in Homer (2015), in which want appears to show scopal ambiguity w.r.t. the presuppositional adverbial no longer. I show how assuming an underlying existential semantics for want, motivated by a new observation about the data, provides a solution to the puzzle.

Homer’s puzzle: Assuming that the negative adverbial no longer presupposes that the proposition denoted by the clause that is in its scope used to be true, sentence (1a) is ambiguous between narrow and wide scope of want w.r.t. no longer (Homer 2015).

(1) a. Consumers no longer want to be kept in the dark about food.
b. I no longer want to be called an idiot.

Homer suggests that on its most salient reading, want takes wide scope over no longer, as it is not assumed that consumers ever had a desire to be kept in the dark about food, or that the speaker of (1b) used to want to be called an idiot. The absence of want from the presupposition of no longer on the most natural reading is taken to be evidence that want can QR over no longer, which is consistent with additional evidence that want may be a ‘mobile positive polarity item’ (PPI).

I provide evidence against a QR approach, and suggest that want is indeed within the scope of no longer in sentences like (1a-b). While I agree that they don’t presuppose that consumers used to have a desire to be kept in the dark (or be called an idiot), the meaning of want is not entirely absent from the presupposition. Instead, (1a-b) appear to require the weaker assumption that consumers in some way used to ‘be willing to’ or ‘be OK with’ being kept in the dark (or being called an idiot).

In the spirit of von Fintel and Iatridou (2017)’s discussion of weak variants of imperatives, I refer to these asacquiescence readings, which in addition to being detected in the presupposition of no longer, can also be detected in sentences like (2a-b).

(2) a. If you want to wait here for a minute, I’ll be right back.
b. Do you wanna give me a hand with this box?

The acquiescence readings in (2a-b), as well as the attested NR readings for want in other DE environments (sentential negation, scope of no NP, restrictor of comparatives/superlatives) suggest that an analogy with free choice is on the right track. There are, however, some DE environments (restrictor of no NP, additional questions/conditionals) that don’t show the predicted pattern. I address these, and other problems, and suggest possible solutions. I also explore how this analysis could extend to other priority modals, like should and for-infinitival relative clauses.

Share

LingLunch 4/12 -  Matthew Tyler (Yale) and Michelle Yuan (MIT)

Speaker: Matthew Tyler (Yale) and Michelle Yuan (MIT)
Title: Nominal-clitic case mismatches (WCCFL Practice Talk) 
Date and time: Thursday, April 12, 12:30-1:50pm
Location: 32-D461
Abstract: 

When arguments are clitic-doubled, the clitic and the nominal it doubles typically bear the same case feature. However, recent theoretical work on clitic-doubling, including but not limited to so-called ‘Big DP’ analyses (e.g. Uriagereka 1995, Nevins 2011, Kramer 2014) treats clitics and nominals as separable entities in the syntactic derivation. Given this background, we propose that under the right syntactic conditions, nominals and their clitics should be able to mismatch in case features. In particular, we identify and investigate two classes of mismatch, which form the mirror image of each other. In Choctaw (Muskogean), nominals acquire case features that their associated clitics lack, while in Yimas (Lower-Sepik; data from Foley 1991), clitics acquire case features that their associated nominals lack. We argue that these mismatches are the consequence of case-assignment operations that target nominals or clitics individually.

The availability of such targeted case-assignment operations is contingent on the language creating the right syntactic configurations. In Choctaw, nominals may be individually targeted for a round of NOM/ACC case-assignment because clitics are doubled at a low position on the clausal spine—by the time that NOM/ACC case is assigned at TP, clitics have already separated from their nominal associates. And in Yimas, while nominals are morphologically unmarked (ABS), clitics may be individually targeted for a round of ERG/ABS case-assignment because they adjoin to the same functional head: following Yuan (2017), multiple clitics adjoined to the same head may employ case-assignment as a dissimilation strategy.
 
 

 

 
Share

LingPhil Reading Group 4/2 - on Stalnaker 1978

Title: on Stalnaker (1978)
Date and time: Monday April 2nd, 1-2pm
Location: 32-D831

This week’s paper is Robert Stalnaker’s Assertion. Pre-read is not required.

Mallory will be presenting the paper.

Share

Phonology Circle 2/4 - Gašper Beguš (Harvard)

Speaker: Gašper Beguš (Harvard)
Title:  Learning the Blurring Process
Date and time: Monday, April 2nd, 5:00-6:30pm
Location: 32-D831
Abstract:

Many artificial grammar learning experiments have provided strong evidence for the assumption that learning biases influence phonological typology: typologically rare processes have been shown to be more difficult to learn. Most of the experiments, however, fail to control for diachronic influences: in many cases, the observed typology can be explained by diachronic factors equally well. In this talk, I present a method for controlling for diachronic influences when testing learning biases. Unnatural processes provide a crucial solution to this problem. I show that a statistical model of diachronic development (that I call Bootstrapping Sound Changes) identifies a crucial mismatch in predictions between the learning and diachronic bias approaches. This mismatch allows me to design experiments such that diachronic factors are controlled for. I present results from two experiments that test learnability of complex vs. unnatural processes and suggest that learnability differences have to influence observed typology in cases that cannot be explained by historical factors. I also discuss implications of this approach for phonological theory in general.
Share

CompLang 4/2 - Idan Blank (MIT)

Speaker: Idan Blank (MIT)
Title:  When we “know the meaning” of a word, what kind of knowledge do we have?
Date and time: Monday, April 2nd, 5:00-6:00pm
Location: 46-5156
Abstract: 

Understanding words seems to require both linguistic knowledge (stored form-meaning pairings and ways to combine them) and world knowledge (object properties, plausibility of events, etc.). In this talk, I will pose some challenges for common distinctions between these knowledge sources. First, I will ask whether rich information about concrete objects could be, in principle, learned from just the co-occurrence statistics of different words even in the absence of non-linguistic (e.g., perceptual) information. To this end, I will introduce a domain-general approach for leveraging such statistics (as captured by distributional semantic models, DSMs) to recover context-specific human judgments such that, e.g., “dolphin” and “alligator” appear relatively similar when considering size or habitat, but different when considering aggressiveness. Second, I will probe DSMs for “syntactic”, abstract compositional knowledge of verb-argument structure (e.g., “eat”, but not “devour”, can appear without an object). I will demonstrate that these syntactic properties of verbs can often be predicted from distributional information (i.e., without explicit access to “syntax”), indicating that DSMs capture those aspects of verb meaning that correlate with verb syntax. Nevertheless, only a small fraction of distributional information is needed for predicting verb argument structure - the rest appears to capture semantic properties that are relatively divorced from syntax. In fact, the overall similarity structure across verbs in a DSM is independent from the similarity structure across verbs as determined by their syntax, and both kinds of similarity are needed for explaining human judgments. Together, these two studies attempt to push against the upper bound on the potential complexity of distributional word meanings.
Share

LingLunch 4/5 - Stanislao Zompí (MIT)

Speaker: Stanislao Zompí (MIT)
Title: Ergative is not inherent: Evidence from *ABA in suppletion and syncretism
Date and time: Thursday, April 5, 12:30-1:50pm
Location: 32-D461
Abstract: 

This is a practice talk for GLOW 41. The abstract can be found here.

Share

Syntax Square 3/20 - Abdul-Razak Sulemana (MIT)

Speaker: Abdul-Razak Sulemana
Title: Obligatory Controlled Subjects in Buli
Date and time: Tuesday March 20, 1-2pm
Location: 32-D461
Abstract:

It has long been noted that PRO makes no contribution to PF, as it is phonetically null. Several approaches to control have been developed based on this conclusion. These theories could be put into two broad categories. Under one (Bresnan 1978, 1982, Chierchia 1984, Dowty 1985, Jackendoff and Culicover 2003, a.o) this has been taken as evidence that there is no syntactic representation of this element. Under the other (Chomsky 1981, Manzini 1983, Landau 2000, 2001, 2013, 2015 a.o) PRO is syntactically present but its nullness is due to the licensing properties of the controlled structure. In this paper, I present data from Buli a Mabia (Gur) language spoken in Sandema (Ghana) that argue against theories that deny the syntactic presence of PRO. I argue that Buli is a language where PRO is overtly expressed and conclude that phonetic nullness is not an inherent property of PRO.

Share

LingPhil Reading Group 3/12 - on Goldstein

Title: on Goldstein- Free Choice and Homogeneity
Date and time: Monday, March 12th, 1-2pm
Location: 32-D831

 

David Boylan will be presenting. No preread required.

Share

Syntax Square 3/6 - Colin Davis (MIT)

Speaker: Colin Davis
Title: Parasitic Gaps and the Structures of Multiple Movement
Date and time: Tuesday March 6, 1-2pm
Location: 32-D461
Abstract:

In this talk, I present some work in progress about the structure of derivations where multiple A’-movement chains overlap. These derivations show interesting complexities that do not (and could not) arise in derivations with only one A’-movement (Pesetsky 1982, Richards 1997). Towards deepening our understanding of this issue, I use Nissenbaum’s (2000) findings about parasitic gap licensing as a diagnostic for the multiple specifier structures created by successive-cyclic movement through vP in these derivations.

This test reveals a puzzle: While Richard’s (1997) theory of specifier formation predicts tucking-in structures at vP in these scenarios, I show via parasitic gap licensing that (at least sometimes) tucking-in fails to occur. Observations about parasitic gaps in superiority violating D-linking from Nissenbaum provide another instance of the same puzzle. It seems to be the case that the structure at vP, tucked-in or not, reflects the final order of the moved phrases. This is exactly what we predict under the hypothesis of Order Preservation (Fox & Pesetsky 2005). However, it remains mysterious how derivations can ‘know’ what vP configurations to form based on what the final result of the derivation will be. I do not have a good solution, but I hope discussing these puzzles will help.

Share

LF Reading Group 3/7 - Itai Bassi (MIT)

Speaker: Itai Bassi
Title: Fake Indexicals without feature transmission
Date and time: Wednesday March 7, 1-2pm
Location: 32-D461
Abstract:

In a footnote, Partee (1989) mentioned that 1st person pronouns can be semantically bound (“fake indexicals”), pointing to sentence (1). That footnote generated a line of research (Kratzer 1998, Kratzer 2009, Wurmbrand 2017; Heim 2008) according to which bound variables (can) enter the syntactic derivation lacking interpreted phi-features, and inherit features from their binder at the PF branch, as a result of some “feature transmission” mechanism(s).

(1) I am the only one around here who will admit that I could be wrong
—> the speaker is the only individual in {x: x is willing to admit that x could be wrong}

In this talk I offer a formal syntax-semantics for this construction which derives a bound reading for (1) while maintaining that the bound “I” has its person feature interpreted, rendering feature transmission unnecessary. My proposal is to reduce (1) to focus constructions like (2), for which there are alternatives to the feature-transmission story (Bassi and Longenbaugh 2017, a.o.). I will thus propose, building on a suggestion made in Bhatt (2002), that the construction in (1) involves silent association with focus. In addition, I show how my proposal can account for the contrast between (1) and the minimally different (3), which does not have a bound reading for “I” and constitutes a problem for existing feature-transmission analyses (Wurmbrand, Kratzer).

(2) Only I will admit that I could be wrong

(3) I met the only one around here who will admit that I could be wrong (no bound reading)

Share

Invited talk 3/8 - Athulya Aravind (MIT)

Speaker: Athulya Aravind (MIT)
Title: Principles of presupposition in development
Time: Thursday March 8th, 12:30-2:00pm 
Place: 32-D461
Abstract:

Natural language affords us the means to communicate not only new information, but also information that we are already taking for granted, our presuppositions. The proper characterization of presuppositions–the way they enter into the compositional semantics and the way they fit into the exchange of information in communicative situations–has been at the center of long-standing debate. One class of theories treat presuppositions as categorically imposing restrictions on the conversational common ground: presuppositions must signal information that is already mutually known by all participants. While principled and elegant, these theories are often thought to be empirically inadequate, as the common ground requirement is not always met in everyday conversation. A second class of theories, therefore, adopt weaker and less categorical approaches to the phenomenon that are nonetheless a better fit to the empirical facts. 

 
This talk compares these two classes of approaches to presupposition in terms of their implications for language acquisition. I argue that children initially adopt a view of presuppositions as uniformly placing restrictions on the conversational common ground, even in situations where these requirements may be bent. More tellingly, I show that children initially lack the ability to use presuppositions in ways that violate the common ground requirement. The observed two-step developmental trajectory supports a common ground theory of presuppositions, according to which the “rule of thumb” is that presuppositions are already common knowledge, and informative uses involve strategic violations of this rule. In turn, the acquisition data vindicate some of the theoretical idealizations whose empirical validity is masked in part due to the pragmatic sophistication of adult language users.
Share

MIT Colloquium 3/9: Sandhya Sundaresan

Speaker: Sandhya Sundaresan (Leipzig)
Title: An Alternative Treatment of Indexical Shift: Modelling Shift Together Exceptions, Dual Contexts, and Selectional Variation
Date and time: Friday March 9, 3:30-5:00pm
Location: 32-155
Abstract:

I present the following three types of evidence that challenge both context-overwriting and quantifier-binding approaches to indexical shift (the phenomenon where the denotation of an indexical is interpreted, not against the utterance context, but against the index associated with an intensional verb). (I) Systematic exceptions to Shift Together (the constraint that all shiftable indexicals in a local intensional domain must shift together) in Tamil, varieties of Zazaki and Turkish, and potentially also Late Egyptian; (II) novel evidence from imperatives in Korean and supporting secondary data from imperatives in Slovenian, showing that the utterance context continues to be instantiated even in putatively shifted environments; and (III) results from personal fieldwork in Tamil dialects and secondary data from 26 languages (from 19 distinct language families) showing that there is structured selectional variation in the intensional environments in which indexical shift obtains and, furthermore, that such variation is one-way implicational. The following desiderata emerge: 1. Shift Together holds whenever possible, but systematic exceptions may nevertheless obtain; 2. the utterance-context is never overwritten; 3. indexical shift is an embedded root phenomenon that privileges speech predicates. To capture these, I develop an alternative model of indexical shift with the following properties. The context-shifter is not a context-overwriting operator, but a contextual quantifier. At the same time, unlike with standard quantificational approaches to shifting, this contextual quantifier (or “monster”) is a distinct grammatical entity severed from the attitude verb. Specifically, I present evidence from nominalization patterns and complementizer deletion to show that the monster is encoded on the complementizer selected by the attitude verb. I then propose that selectional variation for indexical shift ensues as the result of the monster being encoded on structurally distinct types of complementizer head, each selected by a different class of attitude verb (as has also been recently proposed in the literature).

Share

CompLang 2/26 - Thomas Schatz (UMD/MIT)

Speaker: Thomas Schatz (UMD/MIT)
Title: Leveraging automatic speech recognition technology to model cross-linguistic speech perception in humans
Date and time: Monday, February 26 5:00-6:00pm
Location: 46-3310
Abstract: 

Existing theories of cross-linguistic phonetic category perception agree that listeners perceive foreign sounds by mapping them onto their native phonetic categories. Yet, none of the available theories specify a way to compute this mapping. As a result, they cannot provide systematic quantitative predictions and remain mainly descriptive. In this talk, I will present a new approach that leverages Automatic Speech Recognition (ASR) technology to obtain fully specified mapping between foreign and native sounds. Using the machine ABX evaluation method, we derive quantitative predictions from ASR systems and compare them to empirical observations in human cross-linguistic phonetic category perception. I will present results both where the proposed model successfully predicts empirical effects (for example on the American English /r/-/l/ distinction) and where it fails (for example on the Japanese vowel length contrasts) and discuss possible interpretations.
Share

Invited talk 2/27 - Shota Momma (UC San Diego)

Speaker: Shota Momma (UC San Diego)
Title: Aligning parsing and generation
Time: Tuesday, February 27th, 1:00-2:30pm 
Place: 32-D461
Abstract:

We use our grammatical knowledge in at least two ways. On one hand, we use our grammatical knowledge to say what we want to convey to others. On the other hand, we use our grammatical knowledge to understand what others say. In either case, we need to assemble sentence structures in a systematic fashion, in accordance with the grammar of our language. Despite the fact that the structures that comprehenders and speakers assemble are systematic in an identical fashion, the cognitive systems that assemble the mental representation of sentence structures in comprehension and production might or might not be the same. The potential existence of two independent systems of structure building doubles the problem of linking the theory of linguistic knowledge and the theory of linguistic performance, making the integration of linguistics and psycholinguistic harder. In this talk, I will discuss whether it is possible to design a single system that builds mental representations of sentence structures in comprehension, i.e., parsing and in production, i.e., generation. I will discuss existing and new experimental data pertaining to how sentence structures are assembled in real-time comprehension and production, and attempt to show that the unification between parsing and generation is possible.

Share

LF Reading Group 2/28 - Milena Sisovics (MIT)

Speaker: Milena Sisovics
Title:  Embedded imperatives and voluntatives in Mongolian  
Date and time: Wednesday February 28, 1-2pm
Location: 32-D461
Abstract:

Imperatives have long been considered immune to embedding, though recent research has found numerous counterexamples across languages. I introduce novel data showing that Mongolian likewise allows for genuine embedding of (hearer-directed) imperative as well as a speaker-directed non-assertive speech acts (“voluntative”). I propose a uniform analysis of voluntatives and imperatives (collapsed under the term “jussives”) as necessity modals. Moreover, I address the interesting question as to the analysis of jussive subjects: Subjects of embedded jussives are interpreted relative to the reported context in that they denote the reported speaker (in voluntatives) or hearer (in imperatives), rather than the actual discourse participants. I demonstrate how an analysis of imperative and voluntative clauses as PRO clauses can derive this fact, and provide arguments why such an analysis is preferable over an alternative, indexical shift analysis of embedded jussive subjects.

Share

Invited talk 3/2 - Lyn Tieu (Western Sydney)

Speaker: Lyn Tieu (Western Sydney)
Title: Semantic theory and meaning acquisition
Date and time: Friday, March 2nd, 3:30-5:00pm
Location: 32-D461
Abstract: 

The overarching goal of my research program is to understand the nature of language by drawing on different empirical sources of data in a way that is informed by linguistic theory. In this talk, I will present three examples of recent work that highlight the interplay between theoretical issues in semantics and empirical facts about meaning acquisition. I will begin with the acquisition of the polarity-sensitive item ‘any’ in English (Tieu 2013). In this case, corpus and experimental data from child language suggest a potential learnability problem. A possible solution can be found in current linguistic theory; specifically, the cross-linguistic typology of polarity-sensitive items (Chierchia 2013) may provide a restricted hypothesis space for the learner to consider. In the second example, I discuss a series of recent developmental studies of various kinds of implicatures (Tieu et al. 2016, Tieu et al. 2017, and others). In this case, the child data suggest the presence of a developmental stage where children have mastered some implicatures but not others. Crucially, here too the relevant distinction turns out to quite naturally align with one made in current linguistic theories, specifically about scalar alternatives (Katzir 2007; Fox & Katzir 2011). In the final example, I discuss ongoing work that experimentally investigates the semantic contribution of co-speech gestures (gestures that accompany speech) and pro-speech gestures (gestures that replace spoken words) (Tieu et al. 2017, 2018, ongoing). In this case, experimental work with adults suggests people can very rapidly learn new meanings and project presuppositions from a single exposure to novel iconic gestures that they have never seen before, suggesting productive rules for triggering presuppositions.

Share

Invited talk 2/23 - Shayne Sloggett (Northwestern)

Speaker: Shayne Sloggett (Northwestern)
Title: Understanding reflexives: Combining psycholinguistic and theoretical perspectives
Time: Friday, February 23th, 3:30-5pm 
Place: 32-D461
Abstract:

Should evidence from real-time sentence comprehension routines inform our grammatical models? In this talk I present evidence that they should, drawing on recent findings from investigations of on-line reflexive reference resolution. Reflexive pronouns in English are grammatically constrained to refer to local referents which agree with the reflexive’s morphosyntactic features (Chomsky, 1986; Pollard & Sag, 1992; Reinhart & Reuland, 1993; i.a.), and many studies in sentence processing have found that comprehenders initially consider only these referents when determining a reflexive’s antecedent (Nicol & Swinney, 1989; Sturt, 2003; Dillon, Xiang, Dillon, & Phillips, 2009; i.m.a.). However, some recent work has found that comprehenders entertain long-distance reference when local referents present a particularly poor morphosyntactic match (Parker & Phillips, 2017). This raises an interesting question: are such findings the result of a processing mistake, or rooted in grammatical principles? I present evidence from two sets of studies which suggest the latter, demonstrating that native English-speaking comprehenders are less willing to entertain long-distance interpretations in the presence of an indexical (first/second person) pronoun. Notably, this pattern of behavior bears a striking resemblance to the “person blocking” phenomenon associated with the Mandarin long-distance reflexive “ziji” (Huang & Liu, 2001). In light of these findings, I propose a unified treatment of Mandarin and English reflexives which claims that comprehenders consider long-distance referents which could act as logophoric antecedents. Thus, by approaching the study of reflexive reference from theoretical and psycholinguistic perspectives simultaneously, insights from each inform the other. Moreover, this proposal has strong implications for both our understanding of real-time reflexive comprehension, and our grammatical models of reflexive binding. First, it suggests that the adoption of a long-distance interpretation in English does not reflect a processing “error”, but rather the function of a grammatically available, but otherwise dispreferred, alternative (c.f. Parker & Phillips, 2017). Second, it assigns a processing-based explanation to the fact that English speakers generally do not take such interpretations, rather than a grammatical one. Finally, the proposed model is more consistent with a specification of Binding Theory in terms of locality and c-command, rather than predicates, supporting recent theoretical work on French (Charnavel & Sportiche, 2016) which has reached similar conclusions.

Share

LingPhil Reading Group 12/11 - Matt Mandelkern (Oxford)

Speaker: Matt Mandelkern (Oxford)
Title: Constructing the conditional
Date and time: Monday, December 11, 1-2pm
Location: 32-D769
Abstract:

I argue, contra the dominant line in the linguistics literature, that conditionals validate Conditional Excluded Middle in full generality, and invalidate Duality:

Conditional Excluded Middle: ‘If p, q, or if p, not q’ is always true

Duality: ‘If p, q iff Not: if p, might not q’

I argue that we can, however, give a semantics for conditionals which makes these predictions, without abandoning the insights of Kratzer’s restrictor analysis.

Share

LingPhil Reading Group 12/4 - on Pietroski 2015

Title: Discussion of Pietroski 2015: Framing Event Variables
Date and time: Monday December 4, 1-2pm
Location: 7th Floor Seminar room
Abstract:

I find myself writing two papers that are due around the same time. This one develops an objection, based on Davidson’s (1967a) analysis of action reports, to truth-theoretic accounts of linguistic meaning. The other one is about the relevance of Liar Paradoxes for such accounts; see Pietroski (forthcoming). In both papers, the first numbered sentence is (1).

(1) The first numbered sentence in “Framing Event Variables ” is false.

But here, (1) is simply a reminder of a familiar difficulty for the idea that declarative sentences of a Human Language have truth conditions; where for these purposes, a Human Language is a spoken or signed language that any biologically normal child can acquire given an ordinary course of linguistic experience. In my view, (2) and (3) present further difficulties for this idea.

(2) Alvin chased Theodore gleefully and athletically but not skillfully.

(3) Theodore chased Alvin gleelessly and unathletically but skillfully.

Following Davidson and others, I think action reports have “eventish ” logical forms like (2a-3a).

(2a) ∃e[Chased(e, Alvin, Theodore) & Gleeful(e) & Athletic(e) & ~Skillful(e)]

(3a) ∃e[Chased(e, Theodore, Alvin) & Gleeless(e) & Unathletic(e) & Skillful(e)]

We can stipulate that these existential generalizations—sentences of an invented language—have recursively specifiable truth conditions, and that ‘Chased(e, x, y)’ is true of <α, β, γ> if and only if α was an event of β chasing γ. But as we’ll see, there are good reasons for denying that (2) and (3) exhibit the specified truth conditions. There are potential replies. But I argue that they are implausible, especially given the many independent illustrations (e.g., via Kahneman and Tversky) of how human judgments are affected by linguistic framing.

The discussion will be led by Christopher Baron.
Share

LingLunch 12/7 - Suzana Fong (MIT)

Speaker: Suzana Fong (MIT)
Title: A featural and edge-based analysis of hyper-raising
Date and time: Thursday, December 7, 12:30-1:50pm
Location: 32-D461
Abstract: 

Hyper-raising (HR) consists in raising a subject of an embedded finite clause into the subject or object position in the embedding clause (Ura 1994, Tanaka 2002, Yoon 2007, Nunes 2008, Halpert & Zeller 2015, Halpert 2016, Bondarenko 2017, Deal 2017, Zyman 2017, a.o.). Because the clause a DP hyper-raises from is a finite CP, this introduces a challenge to common assumptions about the phasehood of this type of domain. Moreover, both the position a DP hyper-raises from and the landing site in the matrix clause are case-marked. This introduces a challenge to common assumptions about case assignment.

To circumvent the phase problem, I follow van Urk (2015)’s featural definition of syntactic positions. Specifically, I propose that the complementizer of HR sentences has A-features (i.e. not A-bar). The A-features in the HR complementizer trigger the movement of the subject to the edge of the embedded clause, [Spec, CP] (cf. Tanaka 2002, Takeuchi 2010, Zyman 2017, a.o.). As a consequence of being at the edge of a phase, the embedded subject is accessible to a probe in the dominating phase. This allows the embedded finite subject to be accessed by a matrix probe (T or v).

The postulation of features in C will be argued for by HR to object in Mongolian. As first discussed by Hiraiwa (2005) and Bondarenko (2017) regarding Japanese and Buryat, respectively, an embedded finite subject can receive accusative case while remaining in the embedded clause. I will call ‘medial-raising’ this variety of HR, where the accusative subject does not exit the embedded clause. In a canonical, non-HR sentence, if the embedded nominative subject contains a locally-bound reflexive, the reflexive cannot be bound by the matrix subject. However, in the medial-raised counterpart, binding is possible. [Spec, CP] is a position that can account for the dual properties of medial-raising: it is still inside the embedded clause, but it extends the binding domain of the medial-raised subject, and allows it to receive accusative case from the matrix v. [Spec, CP] will also be relevant to provide an explanation to the interaction between HR and (seemingly) long distance scrambling in Mongolian.

Also following van Urk (2015), I assume that probes can also come in a composite, A/A-bar variety. If HR is triggered by A-features in C and if there can be composite probes, we may expect for there to be an instance of HR that is triggered by composite probes. I tentatively analyze data from Kipsigis and Imbabura Quechua as this type of HR. In these languages, what seems to undergo HR is an embedded argument that is lower than the subject. This creates an additional minimality problem. If this lower argument bears the features that the composite probe in C is looking for, while the subject does not, the minimality question is avoided.

To circumvent the case problem, I first try to show that HR consists of a multiple case checking (Béjar & Massam 1999, a.o.). I propose that the same DP can be assigned more than one case, as long as it is able to move from one domain of case assignment into another (cf. Levin 2016). Under this view, we can characterize HR as an instance of movement across domains of case assignment.

Share

Ling-Lunch 11/30 - Michelle Yuan (MIT)

Speaker: Michelle Yuan (MIT)
Title: Anaphors and antipassives in Inuktitut
Date and time: Thursday, November 30th, 12:30-1:50pm
Location: 32-D461
Abstract: 

The Anaphor Agreement Effect (AAE) refers to the cross-linguistic (possibly universal) inability for anaphors to be cross-referenced with co-varying phi-agreement (Rizzi 1990, et seq.). In this talk, I investigate the AAE in Inuktitut (Inuit; Eskimo-Aleut) as a window into its antipassive construction and clause structure more generally. It has been claimed that the Inuit languages avoid phi-agreement with anaphoric objects by resorting to the antipassive construction, which demotes the object to an oblique (Woolford 1999; Sundaresan 2015). However, a closer look reveals that the interaction of anaphors and antipassives in Inuktitut is both more nuanced and more complex than previously assumed.

Despite surfacing with identical case morphology, I argue that anaphors enter the derivation with inherent case, while true (non-anaphoric) antipassive objects receive structural Case (building on Spreng 2006, 2012). Thus, Inuktitut obviates the AAE by enclosing its anaphors in a PP layer, not by antipassivization. Evidence for this distinction comes from the possibility of case stacking on anaphors (not otherwise possible in Inuktitut), as well as from the parallel behaviour of phi-agreement and verbal antipassive morphology in anaphoric contexts. I also explore the wider implications of this analysis for Inuktitut clause structure and argument licensing.

Share

MIT Colloquium 12/1 - Nina Sumbatova (RSUH)

Speaker: Nina Sumbatova (Russian State University for the Humanities)
Title: Referential, Radical Alliterative, and other uncommon instances of gender agreement
Time: Friday, December 1st, 3:30-5pm
Place: 32-155
Abstract:

According to the widely cited definition by Charles Hockett, “genders are classes of nouns reflected in the behavior of associated words” (Hockett 1958: 231). In other words, genders are defined via agreement. At the same time, in most languages with gender systems, the speakers cannot determine the gender of an arbitrary noun neither by its meaning nor by any formal properties – it is a lexical feature of the noun and should be kept in the lexicon.

However, there is a whole number of languages whose gender systems are described as totally semantic as well as some languages where (almost) all nouns are assigned to particular genders by applying some simple and highly regular formal rules. In this talk, I shall discuss some languages of these two types.

First, I shall present the data of Dargwa (East Caucasian), a language with a “semantic” gender system. I hope to show that gender assignment in Dargwa is rather based on the properties of the NP referents than on the lexical features of the nouns.

Second, I shall discuss the possibility of systems where gender is assigned solely by a set of rules and probably does need to be stored in the lexicon. The language of Landuma (Mel < Niger-Congo) is a unique example of a language where the choice of agreement markers seems to be conditioned by a purely phonological rule. The “gender” system in Landuma needs a serious discussion, since it looks as a violation of the well-known principle of phonology-free syntax (Zwicky 1969; Zwicky, Pullum 1986, etc.).

The data of Dargwa and Landuma were obtained in course of fieldwork in Daghestan (Russian Federation) and in the Republic of Guinea, respectively.

Share

LingPhil Reading Group 11/20 - on Koev (under review)

Title: Discussion of Koev (under review): Parentheticality and Assertion Strength
Date and time: Monday November 20, 1-2pm
Location: 7th Floor Seminar room
Abstract:

The traditional picture in linguistics and philosophy is that descriptive content and illocutionary force are cleanly separated. This paper argues that this orthodoxy is incorrect: descriptive content can modify the strength of the main assertion of the sentence, thus blurring the boundary between semantics and pragmatics. One prominent such case are sentences with SLIFTING PARENTHETICALS (e.g. The dean, Jill said, greeted the secretary; Ross 1973), which grammaticalize an intriguing interaction between compositional meaning and speech act function. In such sentences, the main clause (the non-parenthetical part of the sentence) pragmatically depends on the parenthetical, which provides evidential information. The discourse effect of this setup is that slifting parentheticals modulate the strength with which the main clause is asserted (cf. Urmson 1952; Asher 2000; Rooryck 2001; Jayez & Rossari 2004; Davis et al. 2007; Simons 2007; Murray 2014; Maier & Bary 2015; AnderBois 2016; Hunter 2016). Building on Lewis (1976) and Davis et al. (2007), this paper develops a probabilistic dynamic model that captures the role of parentheticality as a language tool for qualifying commitments. The model also derives three properties that set apart slifting sentences from regular embedding constructions (e.g. Jill said that the dean greeted the secretary), i.e. (i) the fact that slifting parentheticals invariably express upward-entailing operators, (ii) the fact that they modify root clauses and do not occur in subordinate clauses, and (iii) the fact that slifted claims are more difficult to reject or doubt by the speaker than claims expressed in complement clauses.

The discussion will be lead by Maša

Share

Phonology Circle 11/20 - Koichi Tateishi (Kobe College & MIT)

Speaker: Koichi Tateishi (Kobe College & MIT), joint work with Shinobu Mizuguchi (Kobe University)
Title: Focus Prosody in Japanese Reconsidered
Date/Time: Monday, 20 November 2017, 5:00-6:30pm
Location: 32-D831
Abstract:
This study, a part of ongoing joint project on production/perception of Japanese intonation and focus, discusses topics that have hardly been touched upon in relation to them. Most previous studies on Japanese accent and pitch have focused on mechanisms of accentuation, how lexical accents manipulate pitch/F0, how intonation patterns are parsed into prosodic phrases, and how focus affects intonational patterns and prosodic phrasing. However, except for sporadic mention by, for example, Pierrehumbert and Beckman (1988), Selkirk and Tateishi (1988, 1991), Sugahara (2003), and Ishihara (2011a), the discussions are solely on accented words and Downstep (Catathesis), and discussions on unaccented words are hardly found. In this study, we did preliminary experiments on perception/production of focus under various accentual conditions, to examine whether the so-called effects of focus in Japanese (and other languages perhaps) apply to other accentual environments. Due to the preliminary nature of our study, we do not intend to say that the points below are valid, but our interim conclusion is a very moderate one, namely that our results accord with what has been being said about the effects of focus in recent years (Ishihara (2003, 2011ab, 2016, 2017), Sugahara (2003), among others), with several findings that may be novel.
  1. An unaccented prosodic phrase also undergoes Focal Boost of F0, which means that focal prosody is not an accent.
  2. Related to this, the boost of F0 due to focus is not so strong that it nullifies previous F0-manipulating effects, such as Downstep. (Ishihara (2003), Shinya (1999))
  3. Because of the Boost and other factors (such as Lowering of F0 due to GIVENness), Downstep-like effects CAN be observed after an unaccented word with focus. (somehow related to Sugahara’s (2003) results, but a little different)
  4. When there is a string of enough length with GIVEN information contents, there appears to be PRE-focal lowering. This lowering appears to lower the pitch of the focus following it. (There have been studies on GIVEN strings AFTER focus, but not BEFORE in our understanding.)
  5. Given 1.-4., the manipulation of F0 by focus is affected by so many factors, and perhaps it is not appropriate to say that focus initiates a new Major Phrase (or Phonological Phrase). (Ishihara’s series of work)
Share

LFRG 11/22 - Moshe Bar-Lev - rescheduled to 11/29

LF Reading Group meeting with a talk from Moshe Bar-Lev on Homogeneity will take place on November 29th 1-2pm.

Share

LingPhil Reading Group 11/13 - on Romero & Han 2004

Title: Discussion of Romero & Han 2004: On Negative Yes/No Questions
Date and time: Monday November 13, 1-2pm
Location: 7th Floor Seminar room
Abstract:

Preposed negation yes/no (yn)-questions like Doesn’t John drink? necessarily carry the implicature that the speaker thinks John drinks, whereas non-preposed negation yn-questions like Does John not drink? do not necessarily trigger this implicature. Furthermore, preposed negation yn-questions have a reading “double-checking” p and a reading “double-checking” ≠ p, as in Isn’t Jane coming too? and in Isn’t Jane coming either? respectively. We present other yn-questions that raise parallel implicatures and argue that, in all the cases, the presence of an epistemic conversational operator VERUM derives the existence and content of the implicature as well as the p/≠ p-ambiguity.

The discussion will be led by Milica.

Share

Phonology Circle 11/13 - Erin Olson (MIT)

Speaker: Erin Olson (MIT)
Title: Influence of schwa duration on stress/pitch patterns in Passamaquoddy
Date/Time: Monday, 13 November, 5:00-6:30pm
Location: 32-D831
Abstract:

Passamaquoddy-Maliseet (Eastern Algonquian; Maine and New Brunswick) is well-known for the fact that reduced vowels are in many cases invisible to the stress/pitch accent system (LeSourd 1988, 1993; Hagstrom 1995). While previous analyses have assumed that this invisibility is due to some sort of structural deficiency of these vowels, I assume that it is at least in part due to the phonetics of stress/pitch accent. I will present a preliminary phonetic study of pitch and vowel duration in the language, based on data from the Passamaquoddy-Maliseet Dictionary Project (Language Keepers & Passamaquoddy-Maliseet Dictionary Project 2016). I will show that the syllables marked as stressed in the literature crucially involve a rise in pitch, and I hypothesize that the odd behaviour of schwa can be derived from the fact that it is too short to host either this rise in pitch or a fall from one pitch to the next. A brief sketch of this hypothesis in OT is provided.

Share

Syntax Square 11/14 - Michelle Yuan (MIT)

Speaker: Michelle Yuan (MIT)
Title: Last Resort licensing in Inuktitut
Date and time: Tuesday November 14, 1-2pm
Location: 32-D461
Abstract: 

The traditional Case Filter (Chomsky 1981, et seq.) regulates the syntactic distribution of nominal arguments; nominals cannot appear in positions where they are inaccessible for licensing (e.g. Case assignment). It has also been argued that nominals in such positions may nonetheless be later licensed by a Last Resort mechanism that rescues the derivation. In much of this literature, Last Resort Case assignment is modeled as the countercyclic insertion of a Case-bearing head directly onto the nominal in question (e.g. Harley & Noyer 1998, Rezac 2011, Levin 2015, van Urk 2015).

I argue for the existence of this kind of Last Resort Case assignment in Inuktitut (Eskimo-Aleut), thus providing evidence in favour of this idea more generally. Along the way, I also show how this analysis sheds light on various other aspects of Inuktitut morphosyntax. In Inuktitut, regular nominal licensing strategies may be blocked or disrupted in certain environments. When this happens, these nominals uniformly receive a certain kind of oblique case. Crucially, this oblique case morphology must be countercyclically inserted once the clause structure has been fully built. In Inuktitut, this derivational logic is particularly transparent: phi-agreement, the language’s regular nominal licensing strategy, is in C0, meaning that failure to license a lower nominal via phi-agreement can only be seen once the CP layer is Merged. Finally, I argue that, despite surface similarities, this Last Resort Case-insertion process is distinct from default case (e.g. Schutze 2001).

Share

LF Reading Group 11/15 - Tanya Bondarenko (MIT)

Speaker: Tanya Bondarenko (MIT)
Title: Results, Repetitives and Datives: towards an account of the crosslinguistic variation
Date and time: Wednesday November 15, 4-5pm
Location: 32-D461
Abstract:

Repetitive adverbs like ‘again’ are ambiguous in sentences with dative (= indirect) arguments in some languages (ex., in English, (1)), but not in others (ex., in Russian, (2)):

(1) Thilo gave Satoshi the map again.
REP(etitive): ‘Thilo gave Satoshi the map again, and that had happened before.’
RES(titutive): ‘Thilo gave Satoshi the map, and Satoshi had had the map before.’
(Beck & Johnson 2004: 113)

(2) Maša opjat’ otdala Vase kartu.
Masha.NOM again gave Vasja.DAT map.ACC
REP(etitive): ‘Masha gave Vasja the map, and that had happened before.’
RES(titutive): *’Masha gave Vasja the map, and Vasja had had the map before.’

Following the structural ambiguity approach to repetitives ((von Stechow 1996), (Beck 2005), among others), I will argue that the availability of the restitutive reading in constructions with dative arguments depends on the syntactic configuration. The restitutive reading in ditransitives has been previously attributed either to presence of a small clause in the structure (Beck & Johnson 2004), or on the presence of an applicative phrase (Bruening 2010). I will explore another hypothesis: that the restitutive reading arises due to existence of a prepositional phrase in the structure.

Share

Ling-Lunch 11/16 - Tom Roeper (UMass-Amherst)

Speaker: Tom Roeper (UMass-Amherst)
Title: Strict Syntax/Semantic Interfaces and Ellipsis: an explanatory role for acquisition theory
Date and time: Thursday November 16, 12:30-1:50pm
Location: 32-D461
Abstract: 

It has often been said that linguistic theory should account for acquisition. Modern work extends this claim to the more important goal: account for steps on the acquisition path itself, where micro-operations are often visible. The abstract claim has particular relevance when it can resolve long-standing questions in theory. We show that the acquisition path for ellipsis must entail a shift from an Open Interface, dependent upon inference, to a Strict Syntax/semantic Interface which involves a perfect correspondence of syntax and semantics and minimizes pragmatics:

Strict Interface Hypothesis: The acquisition device assumes that there should be an isomorphic connection between LF-semantics, syntax, and phonology.

This is based upon an acquisition principle:

Minimization Goal: Minimize pragmatic inference and maximize the information determined by explicit grammar.

One tradition, epitomized in the work of Hardt (1999), claims that a broad empty category like pro and general cognitive biases toward parallelism (Jackendoff and Culicover (2005)) are needed to account for ellipsis. Another tradition is captured in Kennedy’s (2008) Syntax-Semantics Identity Hypothesis:

A recovered syntactic representation generates an LF form with variables that allow sloppy readings.

Merchant (2015, to appear) and Craenenburg (2014) argue for a “hybrid” account which is theoretically unsatisfying. They claim the system is strict, but then the system assumes pronominal forms (or just do that) when necessary. The literature has also shown an ongoing debate about less than full reconstruction in syntactic representations, given various deviations known as “mismatches”. The acquisition path perspective makes this outcome natural.

The empirical basis of this argument works with evidence about VP ellipsis, NP-ellipsis, and Partitive interpretation. Children overgenerate possible reconstructions beyond sloppy identity and then must somehow retreat. eg. They allow John patted his dog and so did his grandfather to mean grandfather pats an uncle’s dog. This creates a learnability problem since interpretations both include and exceed the adult grammar. Evidence on early VP-ellipsis from Santos, and Wijnen and Roeper (2003) on NP-ellipsis and Hulk and Schleeman (2014) on partitives, and Perez et al (to appear) on empty pro objects support this account.

The acquisition path claim is: Children will step by step, following Minimize Pragmatics, replace a pronominal representation with an explicit syntactic representation and a fixed link to semantic representations.

This Strict Interface perspective is deliberately much more stringent than a claim that interfaces involve “Third Factor” effects (Chomsky (2005) which weakens learnability. A big challenge is to develop a notation that makes the connection transparent and simple for the child and natural within UG.

Share

CompLang 11/16 - David Alvarez Melis (MIT CSAIL)

Speaker: David Alvarez Melis (MIT CSAIL)
Title: Interpretability for black-box sequence-to-sequence models
Date/Time: Thursday, November 16th, 5:00-6:30pm
Location: 46-5165
Abstract:

Most current state-of-the-art models for sequence-to-sequence NLP tasks have complex architectures and millions —if not billions—of parameters, making them practically black-box systems. Such lack of transparency can limit their applicability to certain domains and can hamper our ability to diagnose and correct their flaws. Popular black-box interpretability approaches are inapplicable to this context since they assume scalar (or categorial) outputs. In this work, we propose a model to interpret the predictions of any black-box structured input-structured output model around a specific input-output pair. Our method returns an “explanation” consisting of groups of input-output tokens that are causally related. These dependencies are inferred by querying the black-box model with perturbed inputs, generating a graph over tokens from the responses, and solving a partitioning problem to select the most relevant components. We focus the general approach on sequence-to-sequence problems, adopting a variational autoencoder to yield meaningful input perturbations. We test our method across several NLP sequence generation tasks.
Share

MIT Colloquium 11/17 - Michelle Sheehan (Anglia Ruskin)

Speaker: Michelle Sheehan (Anglia Ruskin)
Title: Different routes to partial control: German/English = French + Icelandic
Time: Friday, November 17th, 3:30-5:00 pm
Venue: 32-155
Abstract:
 
One of the biggest challenges to the movement theory of control (Hornstein 1999, et seq.) is the existence of partial control (PC), whereby (descriptively speaking) big PRO denotes the controller plus some other contextually given individual(s) (Landau 2000, 2003, 2004):
 
  1. Johni wants PROi+ to part company now. 
Early accounts of this phenomenon attribute it to an imperfect control relation between the controller and PRO, mediated by C, which allows PRO to be syntactically singular but semantically plural (Landau 2000, 2004). More recent accounts, however, recognise the problems such an account and attribute PC to either the semantics of attitude predicates (Pearson 2013, 105, Landau 2015) or the presence of an associative marker on the embedded predicate (Landau 2016). On the latter kinds of approaches, note, PC seems less problematic for the movement theory of control. Moreover, Boeckx, Hornstein and Nunes (2010) propose an account of PC which is straightforwardly compatible with the movement theory of control: PC involves exhaustive control with a covert comitative.
 
In this paper, I give experimental data from a number of languages to show that PC remains problematic for the movement theory of control in some (but not all) languages. That is to say that PC is not a uniform phenomenon. While there are languages in which PC does reduce to exhaustive control plus a covert comitative (French), there are also languages in which it only sometimes does (German, English, European Portuguese) and others in which it never does (Icelandic, Russian). There are thus two types of PC, which I label fake PC and true PC. Moreover, in instances of true PC in languages with rich case/agreement morphology, we can see that PRO is actually plural in instances of true PC, and also that it has its own case specification. This raises serious problems for the claim that control is always derived via movement.
 
As a solution to this puzzle, I propose that there are two syntactic routes to control: movement and failed movement. Where a complement CP lacks case/phi-features, an argument in spec TP is indeed free to move to the thematic domain of the matrix clause, according to PIC2 (Chomsky 2001). This gives rise to an exhaustive control reading. Following van Urk and Richards (2015), however, I claim that where the complement CP has case/phi-features, as it can in Icelandic, Russian, European Portuguese and German, this movement is blocked as the clause is an intervener. In such cases, a matrix thematic head can only agree with an embedded subject, but the latter cannot move. A distinct DP is therefore externally merged with that thematic head. At PF, the two arguments form a partial chain, leading to deletion of the lower copy and at LF the two arguments are required to be non-distinct, leading to partial control. 
 
An extended abstract is also available.
Share

LingPhil Reading Group 11/6 - on Cariani 2013

Title: Discussion of Cariani 2013: `Ought’ and Resolution Semantics
Date and time: Monday November 6, 1-2pm
Location: 7th Floor Seminar room
Abstract:

I motivate and characterize an intensional semantics for ‘ought’ on which it does not behave as a universal quantifier over possibilities. My motivational argument centers on taking at face value some standard challenges to the quantificational semantics, especially to the idea that ‘ought’-sentences satisfy the principle of Inheritance. I argue that standard pragmatic approaches to these puzzles are either not sufficiently detailed or unconvincing.

The discussion will be led by David.

Share

Syntax Square 11/7 - Andrew Peters (University of Toronto)

Speaker: Andrew Peters (University of Toronto)
Title: Keeping Finiteness Syntactic: Propositions, bare events and the lack of a finite distinction in Mandarin Chinese
Date and time: Tuesday November 7, 1-2pm
Location: 32-D461
Abstract: 

Lacking morphological diagnostics, arguments for or against the presence of a finiteness contrast in Mandarin Chinese (MC) have focused primarily around the distribution of modal verbs and aspectual markers, and the analysis of control and raising. Lin (2011, 2015), for example, posits (1) that MC raising verbs take finite TP complements, control taking non-finite TP complements, (2) that epistemic modals are raising verbs and root modals are control verbs, and (3) that root and epistemic modals’ differing distribution with respect to aspect requires a finiteness distinction. This talk departs from Lin’s argument by positing that the fundamental clause-type distinction in Mandarin is between propositions and bare events. This distinction relies on a PROP(osition) feature (Cowper, 2005; 2016) which I argue is hosted on AspP in Mandarin, and is responsible for taking bare events and returning propositions. Epistemic modals require propositions in their prejacent (Kratzer, 1991; Butler 2003 inter alia), while root modals may take bare events in Mandarin (Grano, 2015). Thus, finiteness may remain a purely syntactic distinction involving the licensing of structural case (Cowper, 2016), and need not be defined semantically. Lacking further evidence for its existence, I posit that there is no finiteness distinction in Mandarin. 

Share

LF Reading Group 11/08 - Verena Hehl (MIT)

Speaker: Verena Hehl (MIT)
Title: From times to worlds
Date and time: Wednesday November 8, 1-2pm
Location: 32-D461
Abstract:

In this work in progress I will examine the meaning of sentences with German eher ’sooner’/’rather’ which exhibits temporal and modal meanings, in light of Gergel’s 2010, a.o. analysis of English rather, taking into account the historical development of both particles. I will develop an analysis for German eher that has at its core the notion of a scale, basically an ordered set of degrees. The degree approach simplifies the formal account of how this particle came to be used in temporal as well as modal domains. Temporal and modal readings of eher have in common that they consist of a gradable property which combines with a comparative morpheme. They are, however, distinguished in terms of the scale they operate on, leading to different LF attachment sites for the temporal vs. the modal reading.

Share

LingPhil Reading Group 10/30 - on Jacobson 1995

Title: Discussion of Jacobson 1995: On the Quantificational Force of English Free Relatives
Date and time: Monday October 30, 1-2pm
Location: 7th Floor Seminar room
Abstract:

The late 70’s and early 80’s witnessed considerable debate as to the correct syntactic analysis of free relatives in English and other languages with a similar construction: is the internal structure of an NP free relative basically like that of an ordinary NP, or is its internal structure instead like that of other wh constituents such as wh questions? The underlying concern surrounding this debate was whether the gap in a free relative could be analyzed as the result of wh movement; this question in turn, of course, bore on the status of Subjacency and on the feasibility of reducing a large class of phenomena to wh movement. But the correct syntactic analysis of free relatives also has significant implications for the syntax/semantics map and for the theory of NP meanings, and it is to this question that this paper is addressed. In particular, I wikk present some ecidence suggesting that English free relatives do indeed have the internal strcture of other wh constituents — they contain no overt lexical head and therefore also contain no overt quantificational element.1 Just how and why, then, du these have NP-type meanings and — given the claim that there is no overt lexical quantifier — what is it that supplies them with their particular quantificational force?

The discussion will be led by Kelly.

Share

Phonology Circle 10/30 - Michael Kenstowicz (MIT)

Speaker: Michael Kenstowicz (MIT)
Title: The Accent of Surnames in Kyengsang Korean: A Study in Analogy
Date/Time: Monday, October 30th, 5:00-6:30pm 
Location: 32-D831
Abstract:
(Joint work with Hyang-Sook Sohn)

The canonical Korean name consists of a monosyllabic surname followed by a disyllabic given name: Kim, Yu-na. The surname is normally not used alone and must be combined with a given name or title. As a result, underlying contrasts among high, low, and rising accents are partially neutralized in this phrasal context. This presentation reports and analyzes the tonal contours that arise when Kyengsang speakers are tasked with inflecting the surname by itself (Kim, Yu-na -> Kim, Kim-i, Kim-ɨl, etc.)—a type of wug-test. The results suggest that speakers take recourse to the most reliable rules underlying the lexical distribution of the accents relative to specific strata of the lexicon as well as to default rules of word-level inflection.

Share

Syntax Square 10/31 - Rafael Abramovitz (MIT)

Speaker: Rafael Abramovitz (MIT)
Title: Verb-stranding vP Ellipsis in Russian: Evidence from unpronounced subjects
Date and time: Tuesday October 31, 1-2pm
Location: 32-D461
Abstract: 

In this presentation of work in progress, I contribute to the debate on the status of verb-stranding verb phrase ellipsis (VVPE) in Russian with novel evidence from the interpretation of unpronounced subjects. I argue that a heretofore unnoticed construction involving a silent subject in clauses embedded under an attitude verb is the result of VVPE, supporting Gribanova (2013, 2017)’s conclusion that Russian has this type of ellipsis, contra Erteschik-Shir et. al. (2013) and Bailyn (2014). Based on a surprising ban on quantificational elements in the subject of the ellipsis sites, I propose a modification to Takahashi and Fox (2005)’s definition of parallelism domains to allow them to contain certain instances of unbound traces. The ban on quantificational subjects, as well as vehicle change from R-expressions to pronouns (Fiengo and May 1994), follows naturally from this modification.

Share

LF Reading Group 11/01 - Frank Staniszewski (MIT)

Speaker: Frank Staniszewski (MIT)
Title: Modal ambiguity and neg-raising
Date and time: Wednesday November 1, 1-2pm
Location: 32-D461
Abstract:

In a constrained range of environments, some modal operators appear to show an ambiguity in modal force between weak (existential) readings and strong (universal) readings. In English, this weak/strong ambiguity has been discussed with respect to For-Infinitival Relative clauses (Hackl and Nissenbaum (2010)). In this presentation based on work-in-progress, I will examine additional environments in which modal ambiguity can be detected. These include modal mismatches in sluicing (Klein (1985), Merchant (2001), Rudin (2017)), as well as instances in which weaker than expected meanings for want and should can satisfy the presupposition of anymore (based on observations by Keny Chatain p.c.). I will explore the hypothesis that these ambiguities are the result of an underlying weak semantics for the modal operators in question, and that this is the also source of their strengthened ‘neg-raised’ meanings under negation.

Share

Ling-Lunch 11/2 - Asia Pietraszko (UConn)

Speaker: Asia Pietraszko (UConn)
Title: Obligatory clause nominalization in Ndebele
Date and time: Thursday November 2, 12:30-1:50pm
Location: 32-D461
Abstract: 

Clausal DP shells (CPs embedded directly in a DP) have been argued to have a different distribution than bare CPs (e.g. in-situ vs derived sentential subjects), which in turn led to treatments of clausal DP shell as a last resort phenomenon (e.g. Hartman 2012). In this talk, I present evidence for the existence of obligatory clausal DP shell. In Ndebele, a Bantu language of Zimbabwe, DP-contained CPs appear in non-derived positions, rendering the last resort status of a clausal DP shell unlikely in this language. Furthermore, I show that a DP shell must appear even in contexts where it is problematic, namely in N-complement clauses and relative clauses (Ndebele does not allow adnominals of category D). Instead of “dropping” the DP-shell in those contexts, extra functional structure is necessary to combine the clause with the head noun.

Share

LingPhil Reading Group 10/23 - on Fodor & Sag 1982

Title: Discussion of Fodor & Sag 1982: Referential and Quantificational Indefinites
Date and time: Monday October 23, 1-2pm
Location: 7th Floor Seminar room
Abstract:

The formal semantics that we have proposed for definite and indefinite descriptions analyzes them both as variable-binding operators and as referring terms. It is the referential analysis which makes it possible to account for the facts outlined in Section 2, e.g. for the purely instrumental role of the descriptive content; for the appearance of unusually wide scope readings relative to other quantifiers, higher predicates, and island boundaries; for the fact that the island-escaping readings are always equivalent to maximally wide scope quantifiers; and for the appearance of violations of the identity conditions on variables in deleted constituents. We would emphasize that this is not a random collection of observations. They cohere naturally with each other, and with facts about other phrases that are unambigously referential. We conceded at the outset of this paper that the referential use of an indefinite noun phrase does not, by itself, motivate the postulation of a referential interpretation. Our argument has been that the behavior of indefinites in complex sentences cannot be economically described, and certainly cannot be explained, unless a referential interpretation is assumed. It could be accounted for in pragmatic terms only if the whole theory of scope relations and of conditions on deletion could be eliminated from the semantics and incorporated into a purely pragmatic theory. But this seems unlikely.

The discussion will be led by Keny.

Share

Phonology Circle 10/23 - Edward Flemming (MIT)

Speaker: Edward Flemming (MIT)
Title: Stochastic Harmonic Grammars
Date/Time: Monday, 23 October, 5:00-6:30pm 
Location: 32-D831
Abstract:

MaxEnt grammars have become the tool of choice for analyzing phonological phenomena involving variation or gradient acceptability. MaxEnt grammars are a probabilistic form of Harmonic Grammar in which harmony scores (sums of weighted constraint violations) of candidates are mapped onto probabilities (Goldwater & Johnson 2003). But MaxEnt grammar is not the only proposal for deriving probabilities from Harmonic Grammars – one alternative is Noisy Harmonic Grammar (Boersma & Pater 2016), in which variation is derived by adding random ‘noise’ to constraint weights, and further variants of this scheme are discussed by Hayes (2017). I will report on an investigation into the properties of these stochastic grammar formalisms, with the ultimate goal of identifying predictions that could be used to evaluate them empirically.

In spite of the superficial differences between MaxEnt and Noisy Harmonic Grammar, they can both be formulated as NHGs where the noise is added to the harmony scores of each candidate. This common format facilitates analysis and comparison of the two formalisms. The differences identified are difficult to assess empirically, but in the absence of evidence in favor of NHG, MaxEnt is to be preferred for its simplicity and tractability.

 

Share

Syntax Square 10/24 - Sherry Chen (MIT)

Speaker: Sherry Chen (MIT)
Title: Double topicalization, I find really interesting
Date and time: Tuesday October 24, 1-2pm
Location: 32-D461
Abstract: 

Whether Chinese topics are syntactically constrained remains a debatable issue. In this talk, I present on-going work suggesting that in double topicalization constructions, the base- generated topic must precede the moved topic, and if both topics are derived via movement, the two intersecting movement paths must be in a “nested” relation (i.e. the Path Contain ment Constraint, Pesetsky (1982)). This challenges the view that Chinese topics are only constrained by a semantic “aboutness” relation with the comment clause (cf. Xu and Lan- gendoen, 1985), and provides new cross-linguistic evidence for PCC effects (see Appendix).

For the second part of this talk, I discuss a new puzzle regarding topicalization in embedded environments. Specifically, I observe that for certain predicates in non-asserted contexts (Hooper & Thompson, 1973; Miyagawa, to appear), embedded topicalization is possible only if the matrix subject binds a pronoun or a reflexive ziji in the embedded clause. 

Share

Ling-Lunch 10/26 - Deniz Özyıldız (UMass-Amherst)

Speaker: Deniz Özyıldız (UMass-Amherst)
Title: The interaction between factivity and prosodic structure in Turkish attitude reports
Date/Time: Thursday, October 26, 12:30-1:50pm
Location: 32-D461
Abstract:

This talk articulates two puzzles raised by the interaction between the availability of the factive inference in Turkish attitude reports, and their prosodic structure.

Puzzle 1: Where is a trigger?
For a class of attitude reports, we observe the following contrast. 1a) and 1b) are string identical. The position of the sentences Nuclear Pitch Accent (on the matrix verb in 1a), on the embedded object in 1b), seems to correlate with the availability of the factive inference (available in 1a), unavailable in 1b)).

1a) Aybike [Dilara’nin Ankara’da oldugunu] BILIYOR.
  Aybike Dilara.GEN Ankara.LOC be.NMZ know.PRES
  Aybike knows that Dilara is in Ankara (factive).

1b) Aybike [Dilara’nin ANKARA’da oldugunu] biliyor.
  Aybike Dilara.GEN Ankara.LOC be.NMZ know.PRES
  Aybike believes that Dilara is in Ankara (non-factive).

Faced with this contrast, we must ask: Is it the prosodic structure of an attitude report that is driving the availability of the factive inference? Or is it the availability of the factive inference that has an effect on prosodic structure? The former option is appealing, but I argue for the latter. 1a) and 1b) are associated with two distinct semantic representations. Prosody follows. But how?

Puzzle 2: Getting from presupposed to given?
In out of the blue contexts, non-factive attitude reports must be realized with embedded NPA.

2a) What’s up?
  Aybike [Dilara’nin ANKARA’da oldugunu] saniyor.
  Aybike Dilara.GEN Ankara.LOC be.NMZ believe.PRES
  Aybike believes that Dilara is in Ankara.

2b) What’s up?
  #Aybike [Dilara’nin Ankara’da oldugunu] SANIYOR.
  (Out because the embedded clause isn’t given.)

On the other hand, in out of the blue contexts, factive attitude reports are realized with matrix verb NPA.

3a) What’s up?
  Aybike [Dilara’nin Ankara’da oldugunu] unuttu.
  Aybike Dilara.GEN Ankara.LOC be.NMZ forget.PST
  Aybike forgot that Dilara is in Ankara.

3b) What’s up?
  #Aybike [Dilara’nin ANKARA’da oldugunu] unuttu.
  (Out because a contrastive interpretation is triggered, but not licensed in context.)

There is mixed evidence as to whether presupposing a proposition p makes the clause that denotes p given (which would explain why the NPA migrates to the matrix verb in 3). Kallulli (2006, 2010) argues that it does. Others argue that presupposition and givenness are two independent dimensions of meaning (Wagner 2012, Rochemont 2016, Buring 2016). Do the Turkish facts offer a way out? We will see.

Share

CompLang 10/26 - Jon Gauthier (MIT BCS)

Speaker: Jon Gauthier (MIT BCS)
Title: What does NLP tell us about language?
Date/Time: Thursday, October 26, 5:00-6:30pm
Location: 46-5165
Abstract:

Many state-of-the-art models in natural language processing achieve top performance in challenging shared tasks while doing little to explicitly model the syntax or semantics of their input. In light of these results, I will attempt to pitch two of my own past projects in natural language processing with a philosophical spin. I first present a model which reduces syntactic understanding to a by-product of more general pressures of semantic accuracy. I next present evolutionary simulations in which word meanings spontaneously arise due to nonlinguistic cooperative pressures. We close with potentially heretical and hopefully interesting discussion on the ideal relationship between AI engineering and cognitive science.
Share

MIT Colloquium 10/27 - Mark Baker (Rutgers)

Speaker: Mark Baker (Rutgers)
Title: Allocutive Agreement and Indexical Shift in Magahi: A Wedge into the Ghostly Operators at the Clausal Edge
Time: Friday, October 27th, 3:30-5pm
Place: 32-155
Abstract:
In this talk I focus on the phenomenon of allocutive agreement in Magahi, an Indo-Aryan language of North Eastern India, and how it interacts with indexical shift.  Allocutive agreement is known to occur in a fairly small set of languages (Basque, Japanese, Tamil, …) in which the finite verb varies in form depending on who the sentence is addressed to.  Like previous researchers, I argue that there is a syntactic representation of the addressee (“Hr”) in the periphery of the clause which can be the goal of an Agree operation.  More specifically, I claim that allocutive agreement in Magahi happens then finite V+T moves (optionally) from T to Fin, where it gains access to Hr.  Unlike other languages, allocutive agreement in Magahi can happen in a wide range of embedded clauses. Interesting in itself, this also makes it possible to study the interaction between allocutive agreement and the phenomenon of indexical shift: using pronouns like “I” and “you” in embedded clauses to refer to the subject and object of the matrix clause.  I show that indexical pronouns in Magahi shift if and only if allocutive agreement shifts. I use this to develop an analysis in which “Hr” (and similarly “Sp”, a representation of the speaker) is the vehicle of indexical shift—not a sui generis nonnominal context shifting operator, as in most previous analyses of indexical shift.  More specifically, indexical shift happens in the special case where the “Hr” of an embedded clause is controlled by the goal argument of a higher clause rather than by the “Hr” of the higher clause, and then binds a second person pronoun in the embedded clause.  This brings the analysis of indexical shift more in line with how logophoric pronouns have been analyzed in West African languages in the tradition of Koopman and Sportiche 1989 (contrary to the distinction drawn between the two by Deal (2017)).  I give further support to the syntactic control-and-binding style analysis of indexical shift by showing that the antecedent of a shifted first person pronoun needs to be not just the semantic author of the content of the embedded clause, but the syntactic subject of the matrix clause in both Magahi and Sakha—parallel to logophoric phenomena and complementizer agreement in Niger-Congo languages.
Share

Ling-Lunch 10/12 - Itai Bassi and Nick Longenbaugh (MIT)

Speaker: Itai Bassi and Nick Longenbaugh (MIT)
Title: Features on bound pronouns: an argument for a semantic approach [NELS practice]
Date/Time: Thursday, October 12, 12:30-1:50pm
Location: 32-D461
Abstract:

Phi-features (person, gender, number) seem not to be semantically interpreted when they appear on bound pronouns. For example, in (1) the bound “my” appears to contribute an unrestricted variable, rather than a variable restricted to the speaker.

(1) Only I did my homework
(bound reading: no x other than the speaker did x’s homework: for any x)

On one influential approach (Kratzer 1998, Heim 2008, a.o.), bound pronouns enter the derivation without phi-features, which are then transferred to them at PF from the binder. The spelled out phi-features are therefore absent at LF. Recent works (Sauerland 2013, Jacobson 2012, Spathas 2010 a.o.) have taken a more semantic approach which denies this ‘LF-PF mismatch’ view. Those works have argued that phi-features on bound pronouns in cases like (1) do get semantically interpreted, except that they don’t contribute to focus alternatives. This talk provides a novel argument for the semantic approach. The argument comes from the observation that (non-trivial) phi-features appear on donkey pronouns - pronouns that show covariance without c-command:

(2) Only the woman who is dating ME introduced me to her parents
(sloppy reading available)

We show why cases like (2) are a problem for the morpho-syntactic approach and how they are derived on the semantic approach. We further demonstrate how our implementation of the semantic approach is advantageous over the morpho-syntactic approach in accounting for the phenomenon of split binding (Rullmann 2004).

Share

MIT Colloquium 10/13 - Benjamin Spector (CNRS)

Speaker: Benjamin Spector (CNRS)
Title: Plural predication, vagueness and principles of language use
Time: Friday, October 13th, 3:30-5:00 pm
Venue: 32-155
Abstract:

(Based on joint work w/ Manuel Kriz)

The interpretation of plural definites (among others) displays two somewhat unexpected properties: non-maximality and homogeneity.

- Non-maximality refers to the fact that, sometimes, plural definites have less-than-universal quantificational force:

(1) Non-Maximality

[Context: A job interview]
The committee members smiled.
>> Can be appropriately used if, say, 8 out of 10 committee members smiled.

- Homogeneity refers to the fact that plural definites tend to have (near)-universal force in affirmative sentences, but only existential force in the scope of negation:

(2) Homogeneity
(a) John read the books on the reading list.
>> He must have read roughly all of them.
(b) John didn’t read the books on the reading list.
>> He must have read roughly none of them.

These two properties are not restricted to plural definites, but are pervasive over other types of constructions (embedded interrogatives, conditionals, singular predication over complex objects, etc.). They are also both removed when the quantifier ‘all’ is added, as in The committee members all smiled.

We show how, by adopting (a) an underspecified semantics for plural predication making it intrinsically vague, and (b) certain principles of language uses, we can account for both properties in a way that makes very specific and apparently correct predictions. Time-permitting, I will discuss the interactions of homogeneity/non-maximality with binding and quantification.

Share

Ling-Phil Reading Group 10/2 - on Mourelatos 1978

Title: Discussion on Mourelatos 1978: Events, processes, and states
Date and time: Monday October 2, 1-2pm
Location: 7th Floor Seminar room
Abstract:

The familiar Vendler-Kenny scheme of verb-types, viz., performances (further differentiated by Vendler into accomplishments and achievements), activities, and states, is too narrow in two important respects. First, it is narrow linguistically. It fails to take into account the phenomenon of verb aspect. The trichotomy is not one of verbs as lexical types but of predications. Second, the trichotomy is narrow ontologically. It is a specification in the context of human agency of the more fundamental, topic-neutral trichotomy, event-process-state.The central component in this ontological trichotomy, event, can be sharply differentiated from its two flanking components by adapting a suggestion by Geoffrey N. Leech and others that the contrast between perfective and imperfective aspect in verbs corresponds to the count/mass distinction in the domain of nouns. With the help of two distinctions, of cardinal count adverbials versus frequency adverbials, and of occurrence versus associated occasion, two interrelated criteria for event predication are developed. Accordingly, Mary capsized the boat is an event predication because (a) it is equivalent to There was at least one capsizing of the boat by Mary, or (b) because it admits cardinal count adverbials, e.g., at least once, twice, three times. Ontologically speaking, events are defined as those occurrences that are inherently countable.

The discussion will be lead by Christopher.

Share

Phonology Circle 10/2 - Chiyuki Ito (Tokyo University of Foreign Studies, ILCAA & MIT)

Speaker: Chiyuki Ito (Tokyo University of Foreign Studies, ILCAA & MIT)
Title: A Sociophonetic Study of the Ternary Laryngeal Contrast in Yanbian Korean
Date/Time: Monday, 2 October, 5:00-6:30pm
Location: 32-D831
Abstract:

This paper investigates the three-way tense-lax-aspirated laryngeal contrast for stop consonants in Yanbian Korean. Data from 61 speakers (DoB 1935-1992) finds three phonetic correlates dependent on laryngeal type, place of articulation, tone, gender, age, and sub-dialect. VOT has significantly shortened over apparent time. The difference between lax and tense is relatively small but still distinguished reliably.

We compare our Yanbian result with the sound changes reported for Seoul Korean, and analyze them based on the dispersion theory of contrast (Flemming 2004). We show that both the Yanbian and Seoul Korean developments were motivated by the same mechanism: minimizing articulatory effort (*Long VOT) while maximizing the number of contrasts (Maximize Contrasts) and the distinctiveness of contrasts (Mindist=VOT: Xms). An economy constraint (*Over-specify) is postulated to ban the ternary laryngeal contrast in the VOT dimension, which is over-differentiated, suggesting a change from ternary → binary distinction.

We point out that another alternative acoustic cue for the relevant laryngeal contrast existed before the VOT merger took place in both dialects. Thus, we conclude that Korean never had a laryngeal contrast based solely on the VOT dimension, and that the sound change that occurred in Seoul Korean is not a (typical case of) tonogenesis.

Share

Ling-Lunch 10/5 - Sarah Zobel (Tübingen & MIT)

Speaker: Sarah Zobel (University of Tübingen and MIT)
Title: On the semantic variability of weak adjunct “as”-phrases
Date/Time: Thursday October 5, 12:30-1:50pm
Location: 32-D461
Abstract:

Stump (1985) discusses the behavior of free adjuncts, like the sentence-initial adjuncts in (1). He observes that certain free adjuncts (which he calls “weak”) are interpreted as restrictors of co-occurring quantifiers (i.e., modals, adverbs of quantification, Gen/Hab), see (1a), while others (which he calls “strong”) only allow for a non-restrictive, causal interpretation, see (1b).

(1) a. Playing with his toys, Peter is often content. (weak)
   (Possible: When Peter is playing with his toys, he is often content.)

 b. Having three toy cars, Peter is often content. (strong)
   (Not possible: When Peter has three toy cars, he is often content.)

My work focuses on English “as”-phrases (and German “als”-phrases), which are weak adjuncts according to Stump’s distinction. Compare (1a) and (2).

(2) As a passenger of Lufthansa, Peter is often content. (weak)
  (Possible: When Peter is a passenger of Lufthansa, he is often content.)

Starting out with a straightforward analysis of the interactions between “as”-phrases and co-occurring operators (inspired by Stump 1985), I take a closer look at the restrictive possibilities of “as”-phrases (and weak adjuncts in general), and I show that these are in fact more restricted than previously assumed.

Share

CompLang 10/5 - Kasia Hitczenko (UMD & MIT)

Speaker: Kasia Hitczenko (UMD & MIT)
Title: Exploring the efficacy of normalization in the acquisition and processing of the Japanese vowel length contrast
Date/Time: Thursday, October 5, 5:00-6:30pm
Location: 46-5165
Abstract:

Infants must learn the sound categories of their language and adults need to map particular acoustic productions they hear to one of those learned categories. These tasks can be difficult because there is often a lot of overlap between the acoustic realizations of different categories that can mask which sounds should be grouped together. Previous work has proposed that this overlap is caused, at least in part, by systematic and predictable sources of variability, and that listeners could learn about the structure of this variability and normalize it out to help learn from and process the incoming sounds. In this work, we further explore this idea of normalization, by applying it to the problem of Japanese vowel length contrast – a contrast that current computational models fail to learn due to high overlap between short and long vowels. We find that, at least in the way it is implemented here, normalizing out systematic variability does not substantially improve categorization performance over leaving acoustics unnormalized. We then present an alternative path forward by showing that a strategy that uses both acoustic cues and non-acoustic top-down information in categorization is better able to separate the short and long vowels.
Share

LPRG 9/25 - on Stone 1994

Title: Discussion on (Stone 1994) The reference argument of epistemic must
Date and time: Monday September 25, 1-2pm
Location: 7th Floor Seminar room
Abstract:

Epistemic must is used to present a conclusion. In this paper, I explore the hypothesis that this should be modeled computationally using the notion of argument presented by Simari and Loui. An utterance of must p in conversational context K is interpreted as asserting that the argument ‹A,p› is justified in K. The parameter A provides a set of reasoning rules which, along with factual premises from which they derive p, must be salient in K for the utterance to be felicitous.

Simari and Loui’s formulation describes a relationship of defeat between arguments. Thus, in this account as in previous ones, the conclusions presented by epistemic must may be defeasible. This proposal improves on previous accounts in three key respects. First, the criterion that the argument be justified ensures that the speaker believes p when uttering must p. Second, the requirement that the speaker intend the hearer to recover the argument helps to explain the distribution and of must in discourse and the accommodation sometimes involved in understanding uses of must. Third, the link between the claim made by must and a specific argument correctly predicts the variation in apparent force of the modal in different contexts: it varies according to the strength of the argument and the speaker’s intentions in providing the argument.

Because this interpretation for must incorporates restrictions based on salience into a frame- work designed to be relatively tractable, it may be uniquely suited for implementation.

The discussion will be lead by Maša Močnik.

Share

Phonology Circle 9/25 - Kasia Hitczenko (UMD & MIT)

Speaker: Kasia Hitczenko (UMD & MIT)
Title: Exploring the efficacy of normalization in the acquisition and processing of the Japanese vowel length contrast
Date/Time: Monday, September 25, 5:00-6:30pm
Location: 32-D831
Abstract:

Infants must learn the sound categories of their language and adults need to map particular acoustic productions they hear to one of those learned categories. These tasks can be difficult because there is often a lot of overlap between the acoustic realizations of different categories that can mask which sounds should be grouped together. Previous work has proposed that this overlap is caused, at least in part, by systematic and predictable sources of variability, and that listeners could learn about the structure of this variability and normalize it out to help learn from and process the incoming sounds. In this work, we further explore this idea of normalization, by applying it to the problem of Japanese vowel length contrast – a contrast that current computational models fail to learn due to high overlap between short and long vowels. We find that, at least in the way it is implemented here, normalizing out systematic variability does not substantially improve categorization performance over leaving acoustics unnormalized. We then present an alternative path forward by showing that a strategy that uses both non-acoustic and acoustic cues to categorize the sounds is better able to separate the short and long vowels.
Share

Ling-Lunch 9/28 - Naomi Francis (MIT)

Speaker: Naomi Francis (MIT)
Title: There’s something odd about presupposition-denying even
Date/Time: Thursday September 28, 12:30-1:50pm
Location: 32-D461
Abstract:

This talk will explore a puzzle about even. Even can be used in denials of presuppositions, but only when it is below negation in the surface string, as shown in (1).

1. A: Did Kenji bring his wife to the picnic?
  B: He isn’t even married!
  B’: #He’s even unmarried/a bachelor!

I show that this asymmetry is not straightforwardly reducible to independent properties of even or of presupposition denials, but is instead a property unique to sentences that are both presupposition denials and contain even. I present a solution to this puzzle that makes crucial use of the additive presupposition of even. This presupposition is controversial; I argue that the evidence used to challenge its presence does not show what it is usually claimed to show, and that when examined more closely these data turn out to be an argument in favour of the additive presupposition rather than against it. I show that this asymmetry is not unique to English and sketch crosslinguistic implications of the proposed analysis.

Share

CompLang 9/28 - Ray Jackendoff (Tufts)

Speaker: Ray Jackendoff (Tufts)
Title: Morphology and Memory
Date/Time: Thursday, September 28, 5:00-7:00pm
Location: 46-5165
Abstract:

(in collaboration with Jenny Audring)

We take Chomsky’s term “knowledge of language” very literally. “Knowledge” implies “stored in memory,” so the basic question of linguistics is reframed as

What do you store in memory such that you can use language, and in what form do you store it?

Traditionally – and in standard generative linguistics – what you store is divided into grammar and lexicon, where grammar contains all the rules, and the lexicon is an unstructured list of exceptions. We develop an alternative view in which rules of grammar are lexical items that contain variables, and in which rules have two functions. In their generative function, they are used to build novel structures, just as in traditional generative linguistics. In their relational function, they capture generalizations over stored items in the lexicon, a role not seriously explored in traditional linguistic theory. The result is a lexicon that is highly structured, with rich patterns among stored items.

We further explore the possibility that this sort of structuring is not peculiar to language, but appears in other cognitive domains as well. The differences among cognitive domains are not in this overall texture, but in the materials over which stored relations are defined – patterns of phonology and syntax in language, of pitches and rhythms in music, of geographical knowledge in navigation, and so on. The challenge is to develop theories of representation in these other domains comparable to that for language.

Share

Linguistics-Philosophy Reading Group 09/18 - on Ninan; Kelly Gaus and Thomas Byrne

Speaker: Kelly Gaus and Thomas Byrne
Title: Discussion on (Ninan 2015) Two Puzzles about Deontic Necessity
Date and time: Monday September 18, 1-2pm
Location: 7th Floor Seminar room
Abstract:

The deontic modal must has two surprising properties: an assertion of must p does not permit a denial of p, and must does not take past tense complements. I first consider an explanation of these phenomena that stays within Angelika Kratzer’s semantic framework for modals, and then offer some reasons for rejecting that explanation. I then propose an alternative account, according to which simple must sentences have the force of an imperative.
Share

Phonology Circle 9/18 - Suyeon Yun (UToronto)

Speaker: Suyeon Yun (University of Toronto)
Title: Allophonic variation of the word-initial liquid in Korean dialects
Date and time: Monday, September 18, 5:00-6:30pm
Location: 32-D831
Abstract:

This study presents large-scale production data of the word-initial liquid allophones in North and South Korean dialects. In South Korean dialects, liquids cannot occur in word-initial position, and Sino-Korean word-initial liquids undergo deletion (e.g., /Lipjəl/ → [ipjəl] ‘parting’) or nasalization (e.g., /Lotoŋ/ → [notoŋ] ‘labor; Iverson & Kim 1987, a.o.). It is reported, however, that the initial liquid /l/ or /r/ in recent English loanwords may be realized as [ɾ] (e.g., Lee 1999) or as [l] (Seo 2004). On the other hand, in North Korean dialects, word-initial liquids are retained (e.g., Bae 2011) albeit based on a small number of Sino-Korean words. This study examines the allophonic variation of the word-initial liquid in North Korean as well as in South Korean, involving a considerable number of words, both Sino-Korean and loanwords, and participants that consist of different age groups. 35 speakers of North Korean defectors who speak Northern Hamkyeong dialect and 20 speakers of Seoul Korean read 41 /L/-initial words in isolation. Results show that the most common variant of the word-initial liquid is the tap [ɾ], and there are also several conditioned and free allophonic realizations, including the trill [r], approximant [ɹ], lateral [l], nasal [n], and stop [t]. The current data serve as an interesting case of variation, in which one phonemic liquid /L/ shows a lot of variability in its phonetic realizations as a result of the interactions between universal phonological constraints (rhotics occur less frequently and stops occur more frequently before /i/ than before other vowels (cf. Hall & Hamann 2010)), different lexical stratifications ([n] occurs more frequently in Sino-Korean than other loans), and speakers’ exposure to L2 ([r] occurs more frequently for Russian L2 speakers and [l] occurs more frequently for English L2 speakers).
Share

Syntax Square 9/19 - Elise Newman (MIT)

Speaker: Elise Newman (MIT)
Title: E2P2: The Extended Extended Projection Principle
Date and time: Tuesday, September 19, 1:00-2:00pm
Location: 32-D461
Abstract:

Current definitions of EPP properties stipulate a preference for checking features via specifier creation rather than merging as a complement. I argue that this is an unnecessary stipulation, and that we can generalize EPP properties such that they may be satisfied by either specifier creation or complementation. I will use facts about English V to T movement and do-support to motivate this extension of the EPP. I propose a syntactic account of V to T movement, which builds off of Matushansky (2006), in conjunction with this new extension of EPP (henceforth E2P2) to explain not only facts about the English auxiliary system, but also facts about the interaction between negation and auxiliary verbs, and negation and non-finite T. I will discuss how this view can also extend to T to C movement, as well as notions about Anti-locality.
Share

LFRG 09/20 - Keny Chatain (MIT)

Speaker: Keny Chatain (MIT)
Title: Defining local contexts for anaphora
Date and time: Wednesday September 20, 1-2pm
Location: 32-D461
Abstract:

Schlenker (2008, 2010) has shown how one can define an incremental notion of local context for presupposition. This removes the need for encoding projection behaviour in particular lexical entries, as in Dynamic Semantics. In subsequent unpublished work, Schlenker attempts to apply the same ideas to define an incremental notion of local context relevant to anaphora. The proposal, while broad in its empirical coverage, falls short of predicting existential/universal readings of donkey anaphora. It also predicts that universal quantifiers introduce singular discourse referents, just as indefinites.

In work-in-progress, I pick up on this project and show how one could deal with the exceptionality of indefinites, taking stock on their alternative semantics, and sketch ideas on how to produce the weak/strong distinction.

Share

Ling-Lunch 9/21 - Jessica Rett (UCLA)

Speaker: Jessica Rett
Title: Explaining ‘EARLIEST’
Date and time: Thursday, September 21, 12:30-1:50pm
Location: 32-D461
Abstract:

The semantics of degree constructions has motivated the implementation of a MAX operator, a function from a set of degrees to its maximal member (von Stechow 1985, Rullmann 1995, a.o.). This operator is unsatisfying: it’s arbitrary (cf. MIN), and therefore not explanatory. There have thus been several calls to reduce MAX to a more pragmatic principle of maximal informativity (Dayal 1996, Beck & Rullmann 1999, Fox & Hackl 2007, von Fintel et al. 2014).

Intriguing differences between before and after have caused some to posit an EARLIEST operator in the temporal domain (Beaver & Condoravdi 2003, Condoravdi 2010). This operator is unsatisfying for similar reasons (cf. LATEST), and some have suggested it, too, can be redefined in terms of informativity (Rett 2015). However, recent cross-linguistic evidence (reported here) complicates the reduction of EARLIEST to `maximize informativity’: while counterparts of `before’ and `after’ across languages share many foundational semantic properties, they appear to differ in a principled way in how certain before constructions are interpreted. I discuss this and other related observations with respect to the future of a domain-general `maximize informativity’ program.

Share

ESSL/LAcqLab 9/22 - Laurel Perkins (UMD)

Speaker: Laurel Perkins (University of Maryland)
Title: Perceiving Transitivity: Consequences for Verb Learning
Date and time: Friday, September 22nd, 2pm-4pm
Location: 32-D461
Abstract:

There is a paradox in language acquisition concerning the perception of the input. If learners can veridically parse the input, then there is nothing to learn from it; but if they cannot parse the input, then it is unclear how they avoid faulty inferences about structure, or even learn from it at all (Valian 1990, Fodor 1996). In this talk, I examine how children deal with their input, given only partial knowledge of the target grammar. Specifically, I focus on the intersection of transitivity, wh-movement, and verb learning.
Infants can use a verb’s distribution in transitive and intransitive clauses to draw inferences about its meaning (e.g. Fisher et al., 2010) and its argument-taking properties (Lidz, White, & Baier, 2017). In this talk I’ll address two questions about the nature of these inferences. First, are infants’ inferences about verb meaning best characterized as one-to-one matching between arguments in a clause and participants in an event described by that clause (Naigles, 1990; Fisher et al., 2010)? To differentiate this participant-to-argument matching hypothesis from other possibilities, we investigate whether children think an intransitive clause could be a good fit for a two-participant event. Second, at early stages in development, infants may not recognize transitivity in certain “non-basic” clauses, like What did Amy fix? (Gagliardi, Mease, & Lidz, 2016). If a learner does not yet recognize that what stands for the object of fix, might she erroneously infer that fix does not require an object? We probe when infants are able to recognize the transitivity of non-basic clauses like wh-object questions, and how infants who do not yet have that ability might learn to “filter” non-basic clauses from the data they use for verb learning. Thus, learners may be able to overcome the limits of partial knowledge by unconsciously filtering data that may lead to faulty inferences about their grammar.
Share

Ling-Lunch 9/7 - Colin Davis (MIT)

Speaker: Colin Davis (MIT)
Title: Intermediate Stranding, Constraints on Movement, and Cyclic Linearization
Date and time: Thursday, September 7, 12:30-1:50pm
Location: 32-D461
Abstract:

I argue that certain rarely considered facts about stranding point us towards a particular understanding of some general issues in syntactic theory, namely: The theory of phasal domains/spellout, and the nature of movement operations. In particular, using data from English, West Ulster English, Afrikaans, Polish, and Russian, I examine scenarios where material is pied-piped with one step of movement and stranded with a subsequent step, stranding that material at an intermediate position in the clause. I show that this phenomenon of intermediate stranding is subject to the following descriptive generalization:

(1). Intermediate stranding is only possible when the stranded material is, or can be, linearly to the right of the material that continues to move leftward.

I argue that the Cyclic Linearization theory of phase spellout (Fox & Pesetsky 2005, Ko 2014) and a theory of movement as driven by Probe-Goal Agree (Chomsky 1995, Ko 2014, van Urk 2015), and thus constrained by c-command, precisely predicts exactly this generalization, while the commonly held theory of phases, and a view of movement as free and untriggered, cannot. These results also point us to a new perspective on what constrains movement out of moved elements.

Share

MIT Colloquium 9/8 - Alan Yu (UChicago)

Speaker: Alan Yu (UChicago)
Title: Are individual differences in cue weight strategies contrast-specific?
Time: Friday, September 8th, 3:30-5:00 pm
Place: 32-155
Abstract:

Phonological categories are generally defined by multiple acoustic dimensions. Recent studies have found that listeners of the same gender and dialect group give differential weighting to dimensions in phonetic categorization. What factors govern the differences between individuals remain a largely unanswered question. Variability may stem from differences in individual perceptual experience or the influence of some contrast-independent cognitive mechanism that modulates speech processing strategies. This study examines the relationship between perceptual behaviors across three categorization tasks concerning three sets of phonological contrasts in English (i.e. the effects of VOT and f0 on voicing identification, the effects of spectral and duration information on i/ɪ identification, and the effects of vowel on sibilant identification) and show that, while individuals vary in their cue weightings for each contrast, the weight settings across contrasts are not random. Implications of these findings for models of speech perception and language variation and change will be discussed.
Share

ESSL/LAcqLab 08/31 - Marie-Christine Meyer

Speaker: Marie-Christine Meyer (ZAS Berlin)
Title: Priming Three Categories of Implicature (joint work with Roman Feiman, UCSD)
Date and time: Thursday August 31, 11:00 - 12:00
Location: 32-D831
Abstract:

Priming is a powerful and widely used method; its applications range from establishing the existence of implicit stereotyping (e.g., Banaji & Hardin 1996) to arguing for the reality of syntactic representations (e.g., Raffray & Pickering 2010). To the young field of experimental pragmatics, however, the priming paradigm has only been introduced very recently (Bott & Chemla 2016). In this talk, we report on a series of experiments in which we try to prime the computation of three categories of inferences: scalar implicature, exactly-readings of numerals, and (disjunctive) Free Choice inferences. As we will see, it is theoretically attractive to derive these inferences through the same mechanism. The existence of a priming effect between these three categories of inferences would provide strong evidence in favor of the cognitive reality of this theoretical assumption. Before we can produce and believe in such evidence, however, a few more questions need to be answered. Can the results of Chemla & Bott be replicated? What are the kinds of stimuli that work best for the paradigm and the linguistic material? Are we really priming the mechanism in question, or a confound? And finally: What are the mechanisms involved in these inferences for which we find evidence in the form of cross- and within-category priming effects? We will present our results and discuss their relevance to these questions, as well as to the theory of (recursive) scalar implicature.
Share

Ling-Lunch 5/22 - Jay Keyser

The day has come and Prof. Jay Keyser will give a special Ling-Lunch talk today!

Speaker: Jay Keyser
Title: Music, Poetry, Painting and Easter Eggs
Date/Time: Monday, May 22/12:30-1:50pm [notice the exceptional time!]
Location: 32-D461
Abstract:

This talk takes the view that modernism in the so-called sister arts of music, poetry and painting resulted from the abandonment of sets of rules that characterized each genre and that were shared by the artist and his/her audience. Rules governing meter and tonal music are reasonably well understood. I propose a way to think about “rules” for the third genre, painting. These rules define a natural aesthetic, ’natural’ in that the rules are shared by the artist and his or her audience in the way that the rules of one’s natural language are shared by speaker and listener.

I suggest that the esoteric direction that the sister arts took in the period cultural historians call “Modernism” is a direct result of abandoning the natural, i.e. shared aesthetic for private formats whose origins can be found in the 14th century.

Finally, I will speculate on the similarity between what happened to the arts at the turn of the 20th century and what happened in science after the publication of Principia Mathematica two centuries earlier.

Share

MIT @ WAFL 13

Kenyon Branan (4th-year grad student) is off to the International Christian University in Tokyo (Japan) to present The i-boundedness of A-scrambling at WAFL 13. The 13th Workshop on Altaic Formal Linguistics will take place later this week, on May 25—28. Shigeru Miyagawa (faculty) is among the organizers.

Share

[cancelled] Ling-Lunch 5/18 - Colin Davis

Speaker: Colin Davis (MIT)
Title: Cyclic Linearization and Intermediate Stranding: English Possessor Extraction and Beyond
Date/Time: Thursday, May 18th/12:30-1:50pm
Room: 32-D461
Abstract:

In this work, I argue that the Cyclic Linearization view of phases (CL, Fox & Pesetsky 2005, Ko 2005, 2014) accurately constrains pied-piping/stranding, incorporating unexamined facts from the English possessor extraction (PE) construction, first noted in Gavruseva & Thornton (2001). McCloskey (2000) showed that West Ulster English all can be stranded by wh-movement, not only in the base position, but in an intermediate one also, which he takes as evidence for successive-cyclic movement:

(1). What (all) did he say [CP t (all) that we should buy t (all)]?

Intriguingly in contrast, Postal (1974) noted that English prepositions cannot be stranded in intermediate positions:

(2). (To) who do they believe [CP t (*to) that the students spoke t (to)]?

A Puzzle: West Ulster English all and English prepositions are both strandable elements in principle. Why then is only the first capable of IS? This, as McCloskey noted, is mysterious.

Another IS context is the English PE construction, a colloquial option for many speakers. Long-distance PE out of non-subject DP, such as whose money in (3), requires IS of that DP in the embedded spec-CP. Why does English tolerate IS in PE derivations, but not with prepositions? This fact compounds the puzzle.p>

(3). Who did they say[CP [ts money John stole t]? (PE with object pied-piping)

Solution: Chomsky (2001, inter alia) and CL both offer theories of how phases and their spellout determine the properties of successive-cyclic movement. Whereas Chomsky’s conception of phases does nothing to rule out preposition IS, leaving Postal’s puzzle and related facts mysterious, I argue that CL gets the facts right, predicting (1-3). This solution also predicts a generalization about IS, stated in (4).

(4). Intermediate Stranding Generalization (ISG, Predicted by CL)
IS is possible when what pied-pipes, and then is stranded, was adjoined to the right of the mover.

This generalization fits the fact that IS is possible for the strandable all of West Ulster English and [‘s NP] in possessor-extracting English, as these items follow the moving wh-word. (4) rules out IS of prepositions, which precede a mover they are adjoined to. I argue that (4) is cross-linguistically robust, fitting all cases of IS I’ve so far found.

Share

Save the date: Ling-Lunch 5/22 - Jay Keyser

Mark you calendar! Jay Keyser will give a special Ling-Lunch talk on Monday, May 22.

Speaker: Jay Keyser
Title: Music, Poetry, Painting and Easter Eggs
Date/Time: Monday, May 22/12:30-1:50pm [notice the exceptional time!]
Location: 32-D461
Abstract:

This talk takes the view that modernism in the so-called sister arts of music, poetry and painting resulted from the abandonment of sets of rules that characterized each genre and that were shared by the artist and his/her audience. Rules governing meter and tonal music are reasonably well understood. I propose a way to think about “rules” for the third genre, painting. These rules define a natural aesthetic, ’natural’ in that the rules are shared by the artist and his or her audience in the way that the rules of one’s natural language are shared by speaker and listener.

I suggest that the esoteric direction that the sister arts took in the period cultural historians call “Modernism” is a direct result of abandoning the natural, i.e. shared aesthetic for private formats whose origins can be found in the 14th century.

Finally, I will speculate on the similarity between what happened to the arts at the turn of the 20th century and what happened in science after the publication of Principia Mathematica two centuries earlier.

Share

Phonology Circle 5/8 - Ting Huang (MIT)

Speaker: Ting Huang (MIT)
Title: On the Unreleased Stops in Taiwanese Southern Min
Date/Time: Monday, May 8, 5:00–6:30pm
Location: 32-124
Abstract:

Unreleased stops, lacking a burst, have been claimed to have low perceptibility and are more likely to neutralize place contrasts (Stevens 1994; Ohala 2001), which has been supported by examining no-burst VC fragments spliced from released stops. The common consensus is that stop perception relies on two sources of acoustic cues: (i) formant transition and (ii) spectral frequency of burst noise. It is clear that the latter cue is absent in unreleased stops, and therefore the hypothesis is that the cues of formant transition will be enhanced. This study investigates the acoustic correlates of VC (where C=unreleased stops p̚, t̚, k̚) in Taiwanese Southern Min. We argue that the cues of VC are not totally diminished or undistinguishable. Moreover, different morphoprosodic structures (VC-V vs. VC#V vs. VC#C) further complicate the dispersion of stop contrasts in this study, which will also be discussed in this talk. A further comparison of stop cues exhibits that the coronal and dorsal are contrastive in terms of vowel duration and quality. This may be related to the asymmetrical behaviors of consonant-to-consonant assimilation in several languages (e.g. English and Korean) and diachronic changes of stop contrasts in Chinese Phonology.
Share

LFRG 5/10 - Daniel Margulis

Speaker: Daniel Margulis (MIT)
Title: Quantifier float with overt restriction
Date and time: Wednesday May 10, 1-2pm
Location: 32-D461
Abstract:

Example (1) demonstrates quantifier float. The quantifier each intuitively quantifies over individual parts of the subject they, but the two are not linearly adjacent. The Hebrew quantifier kol has to be overtly restricted, even when it floats: (2) is ungrammatical without exad ‘one’ or a full NP like student.

1. They have each read a different book.

2. hem kar’u kol *(exad) sefer axer
they read each one book other
“They each read a different book.”

I will discuss several syntactic and semantic puzzles posed by the construction in (2) and their implications for the analysis of quantifier float.

Share

MIT Colloquium 5/12 - Rajesh Bhatt (UMass Amherst)

Speaker: Rajesh Bhatt (UMass Amherst)
Title: Polar Questions, Selection and Disjunction: clues from Hindi-Urdu ‘kyaa’
Time: Friday, May 12th, 3:30-5:00 pm
Place: 32-155
Abstract:

[joint work with Veneeta Dayal, Rutgers] Hindi-Urdu has an optional marker ‘kyaa’ that appears in polar and alternative questions. We delineate its properties distinguishing from the homophonous thematic ‘kyaa’ (what); in particular we locate it in ForceP. We demonstrate that its distribution in embedded environments is similar to that of embedded inversion in English. Then we use `kyaa’ to argue that projection of alternatives (as in Alternative and Inquisitive Semantics) is constrained by the syntax. In particular, A-bar movements lead to `closure’ of alternatives, making them inaccessible. Consequently we expect a bleeding relationship between such movements and operations that depend upon alternatives such as alternative questions. Finally we also explore interactions between the intonational marking of Y/N questions and syntax.
Share

Phonology Circle 5/1 - Rafael Abramovitz

Speaker: Rafael Abramovitz (MIT)
Title: A Case for Morpheme Structure Constraints from Koryak Labials
Date/Time: Monday, May 1, 5:00–6:30pm
Location: 32-D461
Abstract:

One of the central departures of Optimality Theory and its descendants from earlier models of generative phonology is the principle of the Richness of the Base (ROTB), which holds that the set of inputs to the grammar lacks language-specific properties (Prince and Smolensky 2004). Since the set of ranked constraints is the only locus of crosslinguistic variation in these models, morpheme structure constraints (MSC) (Stanley 1967, Chomsky and Halle 1968 et. seq.) are inadmissible. In this talk, I present an argument against this view based on the distribution of labials in Koryak (Chukotko-Kamchatkan). In this language, v and w contrast prevocalically (1-2), but are neutralized to w elsewhere (3).
  1. wutku ‘here’ vs. vutq-ə-vut ‘darkness’
  2. e-wejulʔ-et-ke ‘not scared’ vs. ɣənt-ev-e ‘you ran’
  3. waɲav-at-ə-k ‘to speak’, but waɲaw ‘word’, a-waɲaw-ka ‘without words’
For morphemes like the root in (3), we can set up the root-final segment in the UR as v, but morphemes with an underlying final w, giving rise to a putative pattern *waɲaw ~ waɲaw-at-ə-k, do not exist. While these facts are straightforwardly captured by an MSC banning w morpheme-finally, as well as by equivalent machinery like morpheme-level filtering, analyses assuming ROTB without intermediate filtering are unable account for them.
Share

Syntax Square 5/2 - Rafael Abramovitz

Speaker: Rafael Abramovitz (MIT)
Title: Outward-Sensitive Phonologically Conditioned Allomorphy in Koryak
Date and time: Tuesday May 2, 1—2pm
Location: 32-D461
Abstract:

In realizational theories of morphology like Distributed Morphology, syntactic operations are taken to apply to structures that lack phonological information, which then needs to be inserted at some later point in the derivation. A question we can then ask is whether there are any principled restrictions on how this insertion proceeds. One influential answer comes from Bobaljik (2000), who argues that vocabulary insertion is cyclic and phase-based: vocabulary items are inserted bottom-up within phases, and bottom-up from phase to phase. This makes predictions about restrictions on allomorphy determined at vocabulary insertion because it entails that, when a morpheme is undergoing insertion, only phonological information is present about the nodes below it, and only morphosyntactic information is available about those above it. This view predicts, then, that outward-looking allomorphy can be only conditioned by morphosyntactic features, and inward looking allomorphy by only conditioned by phonological/morphological features. In line with this prediction, cases of outward-looking phonologically conditioned allomorphy are very scarce, the only clear example that I know of coming from Deal and Wolf (2014). In this presentation of work in progress, I will provide partial support for cyclic spellout. In particular, I will present 3 cases of allomorphy from the Koryak (Chukotko-Kamchatkan, Kamchatka) verb-word, and will argue that they are outward-looking and phonologically conditioned, contrary to one of the claims of cyclic spellout. Based on this, I will argue that the phase-internal part of cyclic spellout is either false as a universal or unfalsifiable: these patterns of allomorphy pattern cannot be captured by it, but a theory of phonology sufficiently powerful to account for them negates cyclic spellout’s predictive power. However, I will argue that the predictions of phase-by-phase cyclicity are, in fact, borne out: in none of these cases can morphemes trigger allomorphy across a phase boundary, even if they are linearly adjacent to each other.
Share

LFRG 5/3 - Athulya Aravind

Speaker: Athulya Aravind (MIT)
Title: Against a unified treatment of obligatory presupposition effects
Date and time: Wednesday May 3, 1-2pm
Location: 32-D461
Abstract:

Since its original conception, the principle Maximize Presupposition (Heim 1991, Sauerland 2003) has been recruited to explain why the use of certain presupposition triggers is obligatory in contexts that satisfy their presuppositions (1).

1. a. The/#A sun is shining.
b. I washed both/#all of my hands.
c. Does your dog have a bushy tail/#bushy tails?

In this talk, I re-examine one type of “obligatory presupposition” environment, involving additive particles (2), and argue that they do not fall within the purview of Maximize Presupposition (contra e.g. e Amsili and Beyssade, 2006, Chemla 2008, Singh 2008).

2. a. Sue went to the party. John went to the party #(too).
b. Jenn went to the movies yesterday. She went #(again) today.

Building on previous work (Krifka 1999, Saebo 2004, Bade 2016), I will first propose an account for the effects in (2), on which insertion of additives is one strategy (among many) to circumvent inconsistencies arising from uncancellable exhaustivity inferences. I will then present experimental results from both adults and children that offer support for a non-unified treatment of obligatory presupposition effects.

Share

Ling-Lunch 5/4 - Amanda Rysling

Speaker: Amanda Rysling (UMass Amherst)
Title:  Preferential early attribution in segmental perception
Date/Time: Thursday, May 4th, 12:30-1:50pm
Location: 32-D461
Abstract:

Recognizing the speech we hear as the sounds of the languages we speak requires solving a parsing problem: mapping from the acoustic input we receive to the sounds and words we recognize as our language. The way that listeners do this impacts the phonologies of the world’s languages.

Most work on segmental perception has focused on how listeners successfully disentangle the effects of segmental coarticulation. An assumption of this literature is that listeners almost always attribute the acoustic products of articulation to the sounds whose articulation created those products. As a result, listeners usually judge two successive phones to be maximally distinct from each other in clear listening conditions. Few studies (Fujimura, Macchi, & Streeter, 1978; Kingston & Shinya, 2003; Repp, 1983) have examined cases in which listeners seem to systematically “mis-parse” (Ohala, 1981; et seq.), hearing two sounds in a row as similar to each other, and apparently failing to disentangle the blend of their production. I advance the hypothesis that listeners default to attributing incoming acoustic material to the first of two phones in a sequence, even when that material includes the products of the second phone’s articulation. I report studies which show that listeners persist in attributing the acoustic products of a second sound’s articulation to a first sound even when the signal conveys early explicit evidence about the identity of that second sound. Thus, in cases in which listeners could have leveraged articulatory information to begin disentangling the first sound from the second, they did not do so. I argue that this behavior arises from a domain-general perceptual bias to construe temporally distributed input as evidence of one event, rather than two.

These results support a new conceptualization of the segmental parsing problem. Since listeners necessarily perceive events in the world at a delay from when those events occurred, it may be adaptive to attribute the incoming signal to an earlier speech sound when no other determining information is available. There are cases in which listeners do not disentangle the coarticulated acoustics of two sequential sounds, because they are not compelled to do so. Finally, I argue that this has affected the phonologies of the world’s languages, resulting in, for example, predominantly regressive assimilation of major place features.

Share

Syntax Square 4/25 - Nick Longenbaugh

Speaker: Nick Longenbaugh (MIT)
Title: Towards a unified treatment of the φ-Agree/Move correlation
Date/Time: Tuesday April 25, 1-2pm
Location: 32-D461
Abstract:

The past two decades have seen an explosion of research into the cross-linguistic manifestation of φ-Agreement and the basic principles at stake in its operation. Much of the original impetus for this investigation came from a desire to understand the precise correlation between φ-Agree and movement. If anything, however, ensuing discoveries have muddied the waters in this domain. While it is almost universally acknowledged that φ-Agree and Move are related (see van Urk 2015 for discussion and a formalization of this link), there has been a steady retreat from the strong position of the early 1990s that φ-Agree is parasitic on movement.

(1) Specifier-head agreement (Kayne 1989): If AgrX is an agreement head and DP a phrase bearing φ-features, morphological agreement obtains only if the following structural configuration obtains: [AgrXP DP [AgrXP AgrX […DP…]]]

A principle like (1) is especially successful for capturing agreement phenomena in the vP domain, e.g., past-participle agreement (PPA) in Romance and Scandinavian (Kayne 1985, 1989a; Christensen and Taraldsen 1989), but a wealth of cross-linguistic data (in at least Tsez (Polinsky and Potsdam 2001; English (Chomsky 2000, 2001), Icelandic (Sigurðsson 1996, 2008; Boeckx 2010), Hindi-Urdu (Boeckx 2004; Bhatt 2005), Basque (Etxepare 2006; Preminger 2009)) supports a “long-distance” φ-Agree operation, as in Chomsky 2000, 2001.

(2) Agree (Chomsky 2000, 2001): An Agree relation obtains between a head H and a phrase XP, provided:
(i) Matching: XP bears valued features that are a superset of the unvalued features on H
(ii) Locality: There is no YP asymmetrically c-commanding XP that satisfies matching

Given that (2) formally dissociates Agree and Move, the link so commonly observed between them must be added back in, a task that usually falls to ad-hoc “EPP” features, either stipulated to be present on heads or probes themselves. This state of affairs leaves unanswered a number of fundamental questions, both theoretical and technical: How do we handle cases like PPA and other apparent instances of Spec-Head agreement? Can we predict which probes trigger movement, or must this be stipulated in an ad-hoc, language specific way? Most crucially, why should Agree and Move ever be correlated in the first place?

It is to these questions that this talk will be addressed. Beginning with a case study of PPA, I argue for the conclusion that every φ-probe has the postulated “EPP-property,” so that φ-Agree must trigger movement unless independent factors intervene to block it. This allows us to remove “EPP” as a feature of heads or probes, and to predict straightforwardly whether φ-Agree triggers movement. I then explore two consequences of this proposal. The first is that those “null subject” languages where T has φ-probe have both the traditional EPP (T must project a specifier) and null expletives (following Chomsky’s 1981 proposal), a result I argue for on independent grounds following Bresnan & Kanerva (1989) and Sheehan (2010). The second is a new theory of expletive there that treats it as a semantically vacuous oblique pronoun. By analogy to the behavior of oblique pronouns in Icelandic and cross-linguistically, I show this treatment better captures the distribution of there in English and cross-linguistically with fewer stipulations than competing treatments.

Share

LFRG 4/26 - Keny Chatain

Speaker: Keny Chatain (MIT)
Title: Relative clauses; interactions with modals and definite article choice in Fering and Akan
Date and time: Wednesday April 26, 1-2pm
Location: 32-D461
Abstract:

In this talk, I propose a different semantics for relative clauses that brings them closer to the semantics of conjoined sentences. This accounts for the intuition that in many cases (albeit not all), a sentence with a relative clause is paraphrasable as a conjunction of two clauses:

1. I saw a dog that was limping.

2. I saw a dog and it was limping.

I explore two consequences of this idea in two unrelated areas.

First, I show that this semantics can improve over the standard Grosu & Krifka (2007)’s account of intensional relative clauses (as in (3)), in that it avoids postulating type-shifters that are specific to this construction and it does not posit higher-order abstraction.

3. The gifted mathematician you claim to be should be able to solve this problem in no time.

Second, I show that this semantics can explain interactions between relative clauses and the choice of the definite article in languages with the weak/strong definite article distinction. While strong articles are standardly taken to be anaphoric to previously mentioned entities (Schwarz 2012), they can appear in combination with a restrictive relative clause even when no previous referent is available. However, using the strong form in contexts where they are not licensed is not possible in all languages that make the weak/strong distinction: while Fering can, Akan and Haitian creole cannot. I account for this split in terms of the syntax/semantics interface given here.

Finally, I will discuss other patterns that may fall out from the proposal, and patterns that probably won’t.

Share

MIT Colloquium 4/28 - Jonathan Bobaljik (UConn)

Speaker: Jonathan Bobaljik (UConn)
Title: On Some Universals(?) of Case and Agreement
Time: Friday, April 28th, 3:30-5:00 pm
Place: 32-155
Abstract:

The distribution of the major case and agreement alignments has been held to reflect a tetrachoric (implicational) universal: languages may show the same alignment in both case and agreement, but if they diverge, then it is always the case that case alignment is ergative-absolutive, while agreement alignment is apparently nominative-accusative. The reverse combination is unattested. After reviewing the explanation of this universal in Bobaljik 2008 (cf. Baker 2008, Legate 2008), I examine alleged counter-examples, arguing that the universal survives scrutiny, once the distinction between accusative case and differential object marking is made clear. The proposed explanation makes use of the grouping of cases known as the Dependent Case Hierarchy: {nom/abs} < {erg/acc} < {dat/obl}. Dependent Case Theory may play a central role in the explanation of another asymmetry between case and agreement, specifically, in explaining the the typological observation that “Split-S” and other “active” alignments are surprisingly much rarer as case alignments than as alignments of bound person marking. The account, developed in joint work in progress with Mark Baker, relies on the observation that where active agreement systems can be readily described, an active case pattern cannot arise as a core alignment under DCT. Such patterns can only arise as the interaction of one of the core alignment patterns with independent aspects of the grammars of individual languages. In developing that account, we predict a further, and as far as we are aware previously unobserved, asymmetry between what Bittner & Hale called “accusative active” and “ergative active” languages.
Share

Ling-Lunch 4/20 - Ian Roberts

Speaker: Ian Roberts (University of Cambridge/UConn)
Title: Verb Movement and Cartography in English and Romance
Date/Time: Thursday, April 20th, 12:30-1:50pm
Location: 32-D461
Abstract:

I begin by presenting the recent important proposals in Schifano (2014, 2015a,b) showing that across Romance there are at least four distinct landing sites in the TMA zone of the clause for finite lexical verbs, all of them higher than the position of the English lexical verb and lower than the V2 landing site. I then show, on the basis of the interaction of verb- and object-placement with very low adverbs in the Cinque (1999) hierarchy, that English has low vP-fronting, to SpecVoiceP. Both Romance and English verb-movement license a very low event variable (arguably a condition on “anchoring” the clause to the utterance situation, in Wiltschko’s 2014 sense). I then briefly consider the English auxiliary system, drawing largely on Harwood (2013). Finally, I briefly consider three further kinds of system: the fully analytical TMA system of Haitian Creole, and two kinds of V-initial system, comparing Welsh/Irish with Niuean. A range of quite simple parameters governing V-movement and licensing the TMA field emerges.
Share

Explanatory Adequacy in Formal Semantics Reading Group 4/21 - Itai Bassi (MIT)

Speaker: Itai Bassi (MIT)
Title: Katzir, R., & Singh, R. (2013). Constraints on the lexicalization of logical operators. Linguistics and Philosophy 36:1–29.
Date/Time: Friday, April 21, 2–3pm
Location: 32-D831

Abstract:
We revisit a typological puzzle due to Horn (Doctoral Dissertation, UCLA, 1972) regarding the lexicalization of logical operators: in instantiations of the traditional square of opposition across categories and languages, the O corner, corresponding to ‘nand’ (= not and), ‘nevery’ (= not every), etc., is never lexicalized. We discuss Horn’s proposal, which involves the interaction of two economy conditions, one that relies on scalar implicatures and one that relies on markedness. We observe that in order to express markedness and to account for a bigger typological puzzle, namely the absence of lexicalizations of ‘XOR’ (= exclusive or), ‘all-or-none’, and many other imaginable logical operators, one must restrict the basic lexicalizable elements to a small set of primitives. We suggest that an ordering based perspective, following Keenan and Faltz (Boolean semantics for natural language, 1985), makes the stipulated primitives that we arrive at more natural. We also propose a modification to Horn’s proposal, based on recent work on implicatures, in which only the implicature condition is operative and in which markedness is part of the definition of the alternatives for scalar implicatures rather than an independent condition.
Share

MIT Colloquium 4/21 - Gaja Jarosz (UMass Amherst)

Speaker: Gaja Jarosz (UMass Amherst)
Title: Sonority Sequencing in Polish: Interaction of Prior Bias and Experience
Time: Friday, April 21st, 3:30-5:00 pm
Place: 32-155
Abstract:

Recent work on phonological learning has questioned the traditional view that innate principles guide and constrain language development in children and explain universal properties cross-linguistically. In this talk I focus on a particular universal, the Sonority Sequencing Principle (SSP), which governs preferences among sequences of consonants syllable-initially. Experimental evidence indicates that English, Mandarin, and Korean speakers exhibit sensitivity to the SSP even for consonant sequences that never occur syllable-initially in those languages (such as [nb] vs. [bn] in English). There is disagreement regarding the implications of this finding. Berent et al. (2007) argue that these results can only be explained with reference to an innate principle; however, Daland et. al (2011) show that computational models capable of inferring statistical generalizations over sound classes can detect evidence for these preferences based on related patterns in the language input (and therefore no reference to innate principles is required). Building on these studies, I argue that English is the wrong test case: it does not differentiate predictions of these two hypotheses. I examine learning of syllable structure phonotactics in Polish, a language with very different sonority sequencing patterns from English. Polish provides a crucial test case because the lexical statistics contradict the SSP, at least in part. I review developmental evidence indicating that children acquiring Polish are nonetheless sensitive to the SSP, producing larger sonority rises more accurately in spontaneous production (Jarosz to appear). I then present results from two experiments investigating adult Polish native speakers’ phonotactic knowledge. The findings indicate that Polish native speakers’ phonotactic preferences are sensitive to the SSP and that this SSP sensitivity is not predicted by the computational models that succeeded for languages like English, Mandarin, and Korean. This suggests a crucial role of an inherent bias or a constraint on generalization from the input. At the same time, native speakers’ sonority-sequencing preferences are not entirely expected on the basis of SSP alone, suggesting an important role for experience as well. I discuss implications of these prior bias – experience interactions for modeling of phonological learning.
Share

Phonology Circle 4/10 - Benjamin Storme

Speaker: Benjamin Storme (MIT)
Title: Cyclicity in Standard French: the role of stem-base perceptual similarity
Date/Time: Monday, April 10, 5:00–6:30pm
Location: 32-D461
Abstract:

In Standard French, stems in derivatives behave regularly or cyclically depending on the phonological shape of the suffix: stem-final mid vowels behave regularly if the suffix starts with a non-schwa vowel or a glide, and cyclically otherwise. I compare two approaches to explain this pattern: a syllable-based analysis (cf van Oostendorp 2004 on a similar pattern in Dutch), and a perceptually-based analysis. The syllable-based analysis predicts that cyclic application entails identical syllabification of the base-final consonant and its correspondent in the stem. The perceptually-based analysis predicts that cyclic application entails greater perceptual similarity of the base-final consonant and its correspondent in the stem. A comparison of the results of two experiments (a small experiment based on a syllabification task and an experiment based on a discrimination task) suggests that the perceptually-based analysis is superior: it can better explain the difference between liquid-initial suffixes (before which stems behave cyclically) and glide-initial suffixes (before which stems behave regularly). The results of this study are relevant for two debates in phonology: whether phonotactics are better explained in syllable-based or perceptually-based terms (Steriade 1999), and whether phonetic detail plays a role in cyclicity (Steriade 2000).
Share

LFRG 4/12 - Neil Banerjee

Speaker: Neil Banerjee (MIT)
Title: A problem with future-shifting
Date and time: Wednesday April 12, 1-2pm
Location: 32-D461
Abstract:

The English verbs hope and want are future-shifters in that they allow their non-future complements to be interpreted as occurring in the future.

(1) a. Paul hopes to win the championship.
b. Sam wants to live in Boston.

Assuming that non-finite clauses behave like bound present tense, Abusch (2004) builds the future shift into the lexical semantics of future-shifting verbs. Work by Lekakou and Nilsen (2008), as well as Klecha (2016) suggests that the difference, while lexical, can be made to fall out from Condoravdi’s (2001) diversity condition and the modal base of the attitude report. This gives us a lexical semantics where the future is introduced because of the modal base of the attitude report. Verbs compatible with non-doxastic modal bases are predicted to be future-shifters. Independent evidence from want suggests that it can indeed have a non-doxastic modal base, and is also a future-shifter. But the prediction then is that hope and want should then have the same truth conditions in the following case.

(2) a. I have what I want
b. *I have what I hope

Evidence from other future shifters in English (other attitude verbs, antecedents of conditionals, probability reports) suggests that locating the source of futurity in the modal base may not be on the right track. I leave finding the right track as a puzzle for the future.
Share

Ling-Lunch 4/13 - Ezer Rasin

Speaker: Ezer Rasin (MIT)
Title: Severing stress from phonology
Date/Time: Thursday, April 13th, 12:30-1:50pm
Location: 32-D461
Abstract:

According to the consensus view in generative linguistics, the cognitive module known as ‘phonology’ is responsible for various phonological computations, including the computation of word stress, tone, and segmental processes. I will present two differences between stress and segmental phonology to motivate a modular decomposition of phonology, where the computation of stress is carried out in a separate module with a limited interaction with the rest of phonology:

1) Information Encapsulation: drawing on observations by de Lacy (2006) and Blumenfeld (2006), I propose a universal asymmetry between stress and segmental processes. Segmental processes are often sensitive to the position of stress (In American English, for example, [t] is flapped between a preceding stressed vowel and a following unstressed vowel, as in políDical vs. politícian) but the computation of stress is never directly sensitive to segmental information: stress patterns like ‘stress the rightmost vowel followed by a velar’ are unattested, and can be excluded in the modular architecture if the input to the stress module excludes representations of segmental features.

2) Weak Generative Capacity: Heinz (2014) observes that the computational complexity of attested stress patterns goes beyond that of segmental patterns. In particular, stress patterns can require exactly one primary stress per word, but segmental patterns that require exactly one e.g. sibilant per word are unattested. This difference places stress and segmental phonology in two different domains of the Chomsky hierarchy of formal languages, a hallmark of modularity.

Share

Explanatory adequacy in formal semantics 4/14 - Keny Chatain

The reading group on explanatory adequacy in formal semantics continues this week with a discussion of the paper “The logical primitives of thought” by Piantadosi, S. T., Tenenbaum, J. B., and Goodman, N. D. led by Keny Chatain.

Speaker: Keny Chatain (MIT)
Title: Piantadosi, S. T., Tenenbaum, J. B., and Goodman, N. D. (2016). The logical primitives of thought: Empirical foundations for compositional cognitive models. Psychological review, 123(4):392–424 (link)
Date/Time: Friday, April 14, 2:00-3:00pm
Location: 32-D831
Abstract:

The notion of a compositional language of thought (LOT) has been central in computational accounts of cognition from earliest attempts (Boole, 1854; Fodor, 1975) to the present day (Feldman, 2000; Penn, Holyoak, & Povinelli, 2008; Fodor, 2008; Kemp, 2012; Goodman, Tenenbaum, & Gerstenberg, 2015). Recent modeling work shows how statistical inferences over compositionally structured hypothesis spaces might explain learning and development across a variety of domains. However, the primitive components of such representations are typically assumed a priori by modelers and theoreticians rather than determined empirically. We show how different sets of LOT primitives, embedded in a psychologically realistic approximate Bayesian inference framework, systematically predict distinct learning curves in rule-based concept learning experiments. We use this feature of LOT models to design a set of large-scale concept learning experiments that can determine the most likely primitives for psychological concepts involving Boolean connectives and quantification. Subjects’ inferences are most consistent with a rich (nonminimal) set of Boolean operations, including first-order, but not second-order, quantification. Our results more generally show how specific LOT theories can be distinguished empirically.
Share

Phonology Circle 4/3 - Kevin Ryan (Harvard)

Speaker: Kevin Ryan (Harvard)
Title: Onset vs. rime effects in phrasal weight
Date/Time: Monday, April 3, 5:00–6:30pm
Location: 32-D461
Abstract:

Prosodic end-weight (PEW) refers to the specifically phonological aspect of end-weight, whereby prosodically heavier constituents tend to be preferred domain-finally, all else being equal (i.e. controlling for semantics, frequency, morphosyntactic complexity, etc.). This tendency can be seen in coordination (“X and Y” or “Y and X”?) among numerous other constructions, is widespread (though not universal) cross-linguistically, and is amply supported by experiments, including wug-tests. Several explanations have been put forth for PEW, including final lengthening, complexity deferral (for reasons related to processing), metrical optimization, phonotactic optimization, and (esp. in my own work) stress-weight alignment in sentential prosody. I maintain that the stress-weight interface best explains the core properties of PEW, while the other factors are either irrelevant or at least largely orthogonal to it. One area in which the stress-weight analysis illuminates PEW concerns its differing treatment of onset vs. rime segments. For instance, in the nucleus and coda, greater sonority correlates with greater weight, while in the onset, the generalization is reversed: Greater obstruency patterns as heavier. This reversal is also evident from other types of weight systems (with phonetic rationales in Gordon 2005, Ryan 2014). Thus, I propose that PEW instantiates the same stress-weight interface that is well-documented for stress, meter, etc., a generalization of Weight-to-Stress (Prince 1983 et seq.). The proposed generalization is formalized as a stringent weight hierarchy (e.g. moraic sonorant X > moraic X > X), partly to avoid monsters, but stringency can only be maintained if one recognizes a natural class that is the union of onset obstruents (which cannot be analyzed as moraic in English) and rime segments (which are moraic), among other issues.
Share

Syntax Square 4/4 - Colin Davis

Speaker: Colin Davis (MIT)
Title: English Possessor Extraction, Pied-piping, and Cyclic Linearization
Date and time: Tuesday April 4, 1-2pm
Location: 32-D461
Abstract:

In my previous syntax square, I introduced possessor extraction in English. This essentially undocumented possibility in colloquial speech stands in contrast to canonical pied-piping wh-movement of possessors in English.
  1. Who do they say[[_’s cat] is cute]? (Possessor extraction)
  2. [Whose cat] do they say [_ is cute]? (Pied-piping)
English possessor extraction cannot happen all the time, however. In this talk, I go on to analyze the phenomenon’s restrictions. For example, non-subject DPs must be pied-piped to the edge of their clause for PE out of them to be licit, producing a unique instance of partial pied-piping. In a very general sense, pied-piping to an intermediate position provides a nice piece of overt evidence for successive-cyclic movement through intermediate specifiers of CP.
  1. *Who do they think [John likes [_’s cake]]? (No PE from object in-situ)
  2. Who do they think[[_’s cake] John likes _]? (PE from pied-piped object)
I argue that this pied-piping and a number of other details result from an adjacency condition between possessor and the saxon genitive (cf. Gavruseva & Thornton 2001) which interacts with phase-by-phase linearization of syntactic structure (Fox & Pesetsky 2005, Ko 2005, 2014). Along the way, this analysis finds a explanation for the fact that successive-cyclic movement through spec-vP cannot strand anything in English, a curious gap in the paradigm of McCloskeys’s all-stranding and true of P-stranding in English generally. This finding leads to a number of broader predictions about stranding and its interaction with movement and the nature of specifiers (cf. Ko). The interaction of English possessor extraction with existential constructions also leads to a novel argument from linearization that expletive there originates in vP (Biberauer & Richards 2005, Deal 2009).
Share

LFRG 4/5 - Colin Davis

Speaker: Colin Davis (MIT)
Title: English possessor extraction and LF pied-piping
Date and time: Wednesday April 5, 1-2pm
Location: 32-D461
Abstract:

The colloquial speech of many English speakers permits what looks like possessor extraction, which A’-moves a possessor (1) without pied-piping the rest of the DP (2).

1. Who do they think [[_’s fat cat] is cute]? (Possessor extraction)

2. [Whose fat cat] do they think [_ is cute]? (Standard pied-piping)

This movement is interesting in light of the fact that English is a language that otherwise obeys the Left Branch Condition (Ross 1967), which describes a lack of extraction of the leftmost constituent of a nominal phrase. I argue that despite appearances, the possessum is in fact covertly pied-piped in (1), meaning that there really is no Left Branch Condition exception here. Some evidence for this comes from parasitic gaps, where a possessum stranded in an embedded clause can bind a parasitic gap in the matrix clause, as in (3).

3. [Who did you say[_’s haircut is awful] despite wanting help from PG]?

If who moved alone and didn’t carry haircut into the matrix clause, we expect the PG to be bound by who and so refer to a person. If there was full pied-piping we predict whose haircut to bind the PG, giving a silly reading where you want help from a haircut. By the judgments of most speakers, it turns out that the silly reading is the most salient for sentences like this, with the non-silly reading being absent or difficult. Importantly, we only expect the silly reading to be available if the possessum was covertly pied-piped, binding the PG. von Stechow (1996) argues against Nishigauchi (1990), saying that covert pied-piping does not exist, or at least is not interpreted. In (3), covert pied-piping is interpreted. I also apply the logic of covert pied-piping to sluicing in answers to possessor-extracting questions, and some puzzles regarding free relatives, which don’t pattern as expected.
Share