S-Babble

Archived Schedule: 2022-2023

Fall 2022

9/27: JJ Lim (UCSD Linguistics) - Dissociating genitive case assignment from possessor agreement: Evidence from Khalkha Mongolian nominalised clauses

In many Altaic languages, a single functional head (e.g. D) is commonly analysed to be responsible for genitive (GEN) case on subjects and possessor agreement morphology on nominalised clauses (NCs). Using evidence from adjunct clauses in Khalkha Mongolian, I argue that GEN subject licensing should be dissociated from possessor agreement on NCs. I show that GEN subject licensing only takes place when D is present and the DP is suitably licensed, i.e. assigned a theta-role/referential index (Kornfilt 2003, 2006). I further show that accusative (ACC) and GEN subjects behave similarly, which does not support the idea that they are located in different phases.

10/4: Round Robin

All Babblers are invited to give short informal presentations of data, ideas, or questions related to syntax and/or semantics. We encourage you to use this opportunity for group feedback and discussion on any topics or puzzles that you've been musing over lately. Names of those who would like to present will be drawn randomly, and each presenter will have 5-10 minutes (depending on the number of presenters) for presentation and discussion. You are still welcome and encouraged to attend even if you do not have a topic you would like to present.

10/11: Josh Wampler (UCSD Linguistics) - Part of what do does is seem dynamic: Reanalyzing dynamic/stative mismatches in specificational clefts

Traditional wisdom holds that specificational clefts such (1) are ungrammatical.

(1) *What Josh did was know the answer

The claim is that the main verb do in the wh-clause selects for a dynamic predicate (e.g., run, build a bridge, compose music), which leads to an aspectual mismatch when it attempts to compose with a stative predicate in the post-copular phrase (e.g., know the answer, believe a lie, exist) (Ross 1972, Dowty 1979). A closer look at naturally occurring data, however, reveals that people do produce acceptable clefts with stative predicates. In all attested cases, the stative predicate is interpreted as a consequent state (2), or as a relevant factor contributing to a consequent state (3).

(2) You can't have a circumstance whereby you have the former Defence Minister and now Leader of the Opposition standing around and people making jokes about people drowning. Our Pacific neighbours and partners don't think that's funny. What they do is see that Australia wasn't stepping up in a way that's appropriate and treating them with respect that they deserve. 

(3) Twenty tickets will be given away… The only thing you’ll have to do is know the answer to such questions as “Who plays Randy in Crank 2?”

In (2), seeing that Australia wasn't stepping up is a consequence of standing around and making jokes. In (3), knowing the answer is a relevant contributing factor to being given a ticket. To explain these data, I propose a reanalysis of the contribution of do in specificational clefts. Do does not select against stative predicates; its effects are instead pragmatic in nature. Do requires that the post-copular predicate be a part of a salient culminated process, a type of complex eventuality composed of a preparatory phase, a culminating event and a consequent state (Moens & Steedman 1988). By their nature, statives cannot be interpreted as the culminating event, as these events are inherently dynamic. For a stative to be felicitous, it must be interpretable as either the consequent state or as part of the preparatory phase. The mistaken claim that do is ungrammatical with stative predicates in specificational clefts results from a lack of context. Without context, it is difficult for hearers to infer the required culminated process. In (1), there is no clue as to the relevance of knowing the answer, while in (3), the relevance is clear. This work suggests that language is sensitive to more elaborate event representations than current theories (semantic or syntactic) allow for, as well as highlighting the importance of context for linguistic theorizing. Future work will attempt to generalize these findings to other do-containing constructions in English, and look at similar constructions cross-linguistically.

10/18: Emily Clem (UCSD Linguistics) - The expression of time in Amahuaca switch-reference clauses

Since the term switch-reference was coined by Jacobsen (1967), it has been recognized that, in addition to their reference-tracking function, switch-reference markers often encode other meanings, such as the temporal relationship between clauses. In this talk, I use data from original fieldwork to explore the temporal meanings contributed by switch-reference markers in Amahuaca, a language of the Panoan family spoken in the Peruvian Amazon. I also discuss how these markers combine with a series of morphemes that express temporal lapses between events to arrive at temporally ambiguous interpretations of switch-reference clauses and contrast this with the behavior of these temporal lapse markers in matrix environments.

10/25: Axel Arturo Barceló Aspeitia (UNAM) - The semantics of complex cardinal expressions

Ionin and Matushansky (2004, 2006, 2018), Zweig (2006); Danon (2012); De Sciullo (2014), Rothstein (2013, 2016) among others, have presented elegant and sophisticated accounts of the semantics of complex cardinals, where these have internal linguistic structure and are derived compositionally from simpler cardinals. Despite their differences, all of them are based on a basic distinction between additive and multiplicative cardinal expressions. In this talk I will argue that making such a distinction makes the syntactic, semantic and cognitive nature of numerical composition much more complex that necessary. In contrast, I propose a simpler account where, instead of two separate operations of addition and multiplication, the composition of complex numerical expressions involves a single operation of integer multiplication-addition/division-remanent, which is more cognitively basic that the integer operations involved in the traditional accounts. This account, I will argue, better captures what we know about  numerical cognition and the morpho-syntactic behavior of cardinal expressions.

11/15: Seoyeon Jang (UCSD Linguistics) - Korean echo question marker tako↑ is a non-compositional unit (NOTE: talk will be held only on Zoom)

In Korean, echo questions (EQs) are marked by the clause-final particle (CFP) tako↑ (-tako + H% boundary tone, represented by ↑). In this presentation, I argue that tako↑ has to be understood as a non-compositional unit that uniquely characterizes EQs and triggers the EQ meaning, although ta, ko, and ↑ are frequently attested outside EQs. I start by assuming that Korean EQs denote a question (a set of propositions) and bear a presupposition that there is a proposition already conveyed in the discourse (by uttering a sentence bearing it) and is an answer to the question. Then, I show that such EQ meaning cannot be driven by compositionally putting the meanings of ta, ko, and ↑ together for the following reasons: (i) ta always occurs in declarative clauses; (ii) ko is a simple conjunction operator; (iii) ↑ does not guarantee question interpretation; and (iv) none of the three pieces bears a presupposition.

Winter 2023

2/7
Emilio Gonzalez (UC San Diego, Linguistics)
The pragmatic order of directional prepositional phrases

Directional prepositional phrases within the verb phrase appear to enact certain interpretations dependent on their ordering. McInnerney (2022) argues for a VP flat-structure analysis of these constructions, but due to the aforementioned dependency, postulates an additional FormSequence operation (FSQ) to produce ordering for CI (or SEM, LF) interpretation. My goal is to build off McInnerney’s structural analysis, and question the necessity of FSQ. In order to do that I’d like to pursue an intuition that the systematic interpretation tied to the ordering of these directional prepositional phrases results from pragmatic/discourse principles that interact with the full semantic interpretation of these constructions after they are received by the CI interface. Evidence for this theoretical assertion will come in the form of English evidence. The work will be to demonstrate the existence of these pragmatic principles at the discourse level, and their well-formed absence in sentences which ought not to require them. 

2/14
Thomas Morton (UC San Diego, Psychology)
The cat’s (and) dog’s bear: Planning congruent and incongruent structures in English possessives

This study sought out to investigate the low-level effect of Plan Reuse (Rosenbaum & Jorgensen, 1992) on syntactic structure building. Plan Reuse is a general planning principle where speakers prefer to keep with a selected plan’s parameters to avoid some penalty in changing the plan. This study seeks to compare the planning of two syntactic structures: conjoined possessives, like the cat’s and dog’s bear, and embedded possessives, like the cat’s dog’s bear. Conjoined possessives can be broken down in semantic terms of the cat and dog having the bear, while embedded possessives in terms of the cat having the dog, the dog having the bear. Conjoined possessives are constructed with two incongruent operations, conjunction and embedding, while embedded possessives are constructed sequentially with one congruent operation, embedding. We propose that the congruent nature of embedded possessives will facilitate production according to Plan Reuse. A way to neutralize Plan Reuse would be to introduce an intervening word in the phrase to disrupt the strategy; for instance, placing an adjective on the second noun (e.g., the cat’s blue dog’s bear vs. the cat’s dog’s bear). In a 2x2 within-subjects design eliciting prenominal possessives, we compare two conditions: the type of structure elicited, either conjoined or embedded, and the status of an adjective on the second noun, either present or absent. In an online study, participants were instructed to describe one of two scenes presented. The two scenes contrast in a way to encourage elicitation of the target stimuli. Production onset latency (from the time the stimuli is presented) and error rates were measured. We predict a higher rate of error for the embedded possessives than the conjoined possessives due to their increased semantic complexity. If plan reuse is a major factor in the production of embedded possessives, we predict a significant interaction of embedding*adjective-present due to a disruption of Plan Reuse. We found (n = 67) a significant positive effect on embedded structures (p < 0.001) and on the adjective status (p <  .05). However, there is no significant effect of onset latency, but there is a visible difference in means within the embedded structure (adj abs: 1900 ms, adj pres: 1823 ms) and no visible difference within the conjoined structure (both: 1869 ms). While more participants are needed to see the true effect, a tentative framing of this difference in means is that it reflects a strategy to avoid difficult planning, that is, speakers are shortening their planning scope to be word-by-word in anticipation of difficult production. This would be reflective of a motor planning pattern called precrastination where gestures are planned to reduce initial-load at the expense of later effort (Fournier et al., 2019). Bock (1982) proposes that language production is split between two processes: an automatic, habitual, process and a cognitively-controlled planning process. This view aligns with modern approaches to motor control (Pacherie & Mylopoulos, 2021). While this study sought to explore low-level planning advantages, perhaps it will lead to further exploration of high-level sequence planning and the dynamic interaction between the two. 
Bock, J. K. (1982). Toward a cognitive psychology of syntax: Information processing contributions to sentence formulation. In Psychological Review (Vol. 89, Issue 1, pp. 1–47). https://doi.org/10.1037/0033-295x.89.1.1
Fournier, L. R., Stubblefield, A. M., Dyre, B. P., & Rosenbaum, D. A. (2019). Starting or finishing sooner? Sequencing preferences in object transfer tasks. Psychological Research, 83(8), 1674–1684.
Pacherie, E., & Mylopoulos, M. (2021). Beyond automaticity: The psychological complexity of skill. Topoi. An International Review of Philosophy, 40(3), 649–662.
Rosenbaum, D. A., & Jorgensen, M. J. (1992). Planning macroscopic aspects of manual control. Human Movement Science, 11(1), 61–69. 

2/21
Lorenzo Pinton (MIT)  [on Zoom only]

"Numerous" relative clauses: permutation invariance, anti-restrictiveness, triviality 

It’s been observed that gather-like and numerous-like predicates give rise to different felicity patterns when combined with plural quantifiers (Kroch, 1974; Dowty et al., 1987; Champollion, 2010; Amiraz, 2021):

1)  a.    All the students gathered.

     b.  #All the students are numerous.

In this work, I aim to provide an analysis for similar data brought about by restrictive relative clauses:

2)  a.    Jack only talked to the students that gathered.

     b.  #Jack only talked to the students that were numerous.

While gather can be felicitously applied in a restrictive construction (2a), numerous cannot (2b). First, I will argue, through Italian data, that the problem is really tied to restriction, rather than relative clauses in general. Second, I will claim that predicates like numerous have a specific property, permutation invariance (i.e. the fact that such predicates only care about the cardinality of a group, and not about the specific members that compose that group). This property is problematic when numerous is combined with pluralized predicate, where pluralization is defined as the star operator (Link, 1984). In particular, I will show that when a pluralized predicate modified by numerous combines with the definite article the, it generates triviality, which leads to infelicity (Gajewski, 2002). A positive outcome of this solution is that it  predicts the puzzling data in (3), namely the fact that (2b) becomes good if students is modified by another predicate:

3)  Jack only talked to the gathered students that were numerous.

In fact, we can assume that, when gathered students is not pluralized, it will feed numerous plural individuals (since it is a collective predicate), without leading to triviality. In the presentation I will discuss the conclusion that this solution seems to suggest: namely that pluralization is a rather free operation, which is syntactic in nature and its application is governed by certain logical properties, like avoiding triviality. I would like to conclude showing some problems that might be lurking in the proposed solution, and possible extensions of this solution to the plural quantifiers puzzle in (1) that the literature has focused on so far.

3/7
Open House Round Robin

Spring 2023

4/4
Andrew Hedding (University of Washington) [on Zoom only]
Focusing Non-Interrogative Wh-words in Peras Tu'un Savi

In many languages, bare wh-words can also function as non-interrogative indefinites in specific environments (Haspelmath, 1997). Recent work has proposed that the interpretation of these ambiguous wh-words also interacts with focus: specifically, focus on a wh-word forces an interrogative interpretation, while the lack of focus forces a non-interrogative interpretation (Hengeveld et al., 2021). In this talk, I present data that support two main empirical claims: (i) in Peras Tu’un Savi, some bare wh-words can be interpreted non-interrogatively in non-veridical contexts; (ii) in the right context, the same bare wh-words can be interpreted non-interrogatively, even when syntactically focused (pace, Hengeveld et al., 2021). I argue that focusing a bare wh-word triggers a verum focus interpretation (i.e., focus to emphasize the truth of a proposition) and conclude by showing that a similar pattern holds when existential indefinites (e.g., someone) are focused in English, suggesting the possibility of a unified analysis. 

4/18
Group brainstorming: S-Babble this quarter and next year
[on Zoom only]

4/25
JJ Lim (UC San Diego, Linguistics)
Nominalisations without DP: Dissociating genitive case assignment and possessor agreement

5/2
Ivano Caponigro (UC San Diego, Linguistics)
Referring and quantifying without nominals: Headless Relative Clauses across languages

5/9
Josh Wampler (UC San Diego, Linguistics)
Why being acceptable with states is hard to do: A new perspective on English main verb "do" 

English main verb do has traditionally been seen as selecting for a non-stative complement:

 1a. Event: Josh answered the question, and Feifei did so, too.

 1b. State: #Josh knew the answer, and Feifei did so, too.

A broader look at the data, however, reveals that do can acceptably appear with stative predicates:

2a. What you have to do is know when to raise your hand.

2b. Why loving yourself is important, and how to do it.

2c. We retrieve data as needed; occasionally it's the police who do the needing.

2d. Q: What should I do to get into heaven?

        A: Be a good person.

If do requires non-stativity, these examples should encounter a mismatch, leading to ungrammaticality. Why, then, are they so acceptable? In this talk, I propose a novel solution to this puzzle. I first explore traditional ways of resolving this issue (coercion, ambiguity), as well as several recent proposals that are less traditional (Houser 2010, Miller 2013, Bruening 2019), concluding that none can satisfactorily explain the full breadth of the data. The problem is not with do, as has traditionally been thought, but in the interaction between stative predicates (and, interestingly, certain eventive ones as well) and the environments do often appears in. As an illustration, many states are odd when modified by an adjunct:

3a. #Josh is tall [when he is outside].

3b. #Josh will believe in faeries [for your birthday].

One common environment for do is in referring expressions such as do so/do it/do that, which are known to prefer to occur with additional information (compared to their antecedent) (Huddleston & Pullum 2002; Miller 2011, 2013; Bruening 2019, Flambard 2021).

4a. Josh will buy Feifei some flowers. #He'll do so.

4b. Josh will buy Feifei some flowers. He'll do so [because it will make her smile].

This preference for additional information has nothing to do with do, but is a consequence of discourse pragmatic factors (e.g., (4a) is redundant). Given that states often resist modification, and that these referring expressions prefer it, it is no surprise that we often find the combination of the two unacceptable. Yet, with an appropriately plausible adjunct, stative predicates can be acceptable.

5. Josh believes in faeries, and he's done so [for as long as I can remember].

There are many other reasons that states may appear unacceptable in context (e.g., conflicts with tense/aspect, conflicts with inferences of inception/termination/causation/control/intention, conflicts with relevance). Crucially, all of these reasons are independent of do. Do is completely grammatical with states, and all other eventualities. All appearance of unacceptability is epiphenomenal, and can ultimately be traced back to issues of plausibility rooted in our perception and experience with eventualties in the world.

5/23
Seoyeon Jang (UC San Diego, Linguistics)

Semantic evidence for the non-compositional nature of the echo question marker tako in Korean

Korean echo questions (EQs) are uniquely characterized by the clause-final particle (CFP) "tako," which overtly distinguishes EQs from other types of clauses, such as ordinary interrogative and declarative clauses. Since "ta," "ko," and "tako" with falling final intonation can occur on their own as CFP in sentences other than EQs and since rising intonation is attested in sentences other than EQs, "tako" in EQs has been assumed as compositionally derived by combining the CFPs "ta" and "ko" (Noh 1995; H. Lee 2010; E. Lee, Madigan & Park 2016; Sohn 2020). I examine the formal semantic contributions of "ta," "ko," and "tako" (falling) outside EQs and show that it is impossible to combine them according to independently motivated formal principles and derive any of the formal semantic analyses for EQs that have been suggested. Therefore, I conclude that "tako" (rising) should be analyzed as a morphosyntactic unit in contemporary Korean from a synchronic perspective.

6/6
Andy Kehler (UC San Diego, Linguistics)
Predicting structure that never comes: Toward resolving the great ellipsis debate