On Friday 10/14 at 3:30 in ICL S331, Kate Davidson (Harvard) will give a talk entitled, “What is the value of symbolic abstraction?” The presentation will be both in-person and available through Zoom. The abstract can be found below. All are welcome!

Abstract:
Although language is often taken to be a paradigmatic case of the use of arbitrary symbols to communicate ideas, it is also clear that linguistic interactions in both signed and spoken languages frequently incorporate elements of iconic depiction. How exactly these two aspects of language interact, and what, if anything, sets apart and motivates the development of symbolic vs. iconic language, is an area of active research on spoken and signed languages and gesture studies. I will briefly overview various approaches to formally modeling the contributions of iconic and symbolic meaning, and make the case that while both are pervasive in natural languages, only symbolic abstraction supports reasoning over alternatives, necessary for many domains in compositional semantics including negation and questions. In contrast to symbolic descriptions, iconic depictions must be reanalyzed as symbolic in order to participate in alternative building structures, providing a clearer motivation for both depictive and descriptive aspects of language, based on their separate compositional capacities.

On Wednesday, 10/12 @noon in Tobin 521B, we’ll hear from Annika Andersson from Linnaeus University who will talk about “ERP Studies of Cross-Linguistic Influence on Second Language Processing of Syntax and Semantics.”  The abstract for Annika’s talk is below.

Studies comparing second language (L2) processing to first language (L1) processing as measured with event-related potentials have investigated syntactic, morphosyntactic, and semantic processing. Typically, these studies find L2-processing of syntax and morphosyntax to be related to mainly age of acquisition (AoA) but also to proficiency (see Steinhauer et al., 2009 for an overview). The earlier the language is acquired the more similar the processing, such that when the L2 is acquired prior to 3 years of age a native like biphasic ERP response (LAN/P600) can be found (Weber-Fox & Neville, 1996). With high L2 proficiency, even with later AoA some studies show L2 processing of syntax to be native like (Bowden et al., 2007). In contrast to syntactic processing, for semantic processing only proficiency seems to be a relevant predictor and only for quantitative differences in the ERP effect, in this case the N400 (Moreno & Kutas, 2005). In addition to AoA and proficiency, cross-linguistic influence (CLI) has been investigated on morphosyntactic processing (Dowens et al., 2010; Sabourin & Stowe, 2008). In these studies, shared structures (gender agreement) in L1 and L2 support more native like processing of the L2 construct.

In this talk, I will present two studies where we investigated the effects of CLI on L2-processing. First, we compared processing of Swedish word order, verb second, in two groups of learners matched on proficiency and AoA, but whose L1 have similar word order (German) or not (English) (Andersson et al. 2019). In this study of word order processing, we replicated the previous findings of CLI effects on morphosyntactic processing. Second, we compared processing of fine-grained L2 verb semantics (placement verbs) that were either shared or not shared with the L1 (German and English). In contrast to previous ERP studies of semantic processing we found qualitative differences between semantic processing in a first and second language. In addition, these differences could be explained by CLI (Andersson & Gullberg, 2022). These studies show the importance of including CLI as a predictor of L2 processing but also that previous studies not finding qualitative differences in L1 and L2 semantic processing could be due to restricting the investigation into shared
semantic concepts.

On 10/6 from 12:00 pm to 1:00 pm Eastern Time in Computer Science Building Room 150/151, Karthik Narasimhan (Prinston) will give a talk. More information can be found here.

Bio:

Karthik Narasimhan’s research spans the areas of natural language processing and reinforcement learning, with a view towards building intelligent agents that learn to operate in the world through both experience and existing human knowledge (ex. text). Karthik is especially interested in developing autonomous systems that can 1) acquire an understanding of language through interaction with their environment and 2) utilize textual knowledge to drive their decision making.

This Friday (Oct 7), we will have a special meeting of the Language Acquisition Research Center (LARC) to welcome people that might be interested in acquisition research with bilingual populations but have no previous experience in the field.
The talk is open to the public and it is meant for people with no previous background in studying the acquisition of additional languages.
We will meet at 11:30 in ILC RM 451.
Join us!

Language Acquisition Research with Bilingual Speakers for Beginners
Luiz Amaral

This is a presentation for undergraduate and graduate students that have no previous experience with language acquisition research focusing on bilingual populations.
In this talk we will discuss some basic elements necessary to propose a research project that focuses on how people learn additional languages. We will talk about some core theoretical questions in the field of bilingual/second language acquisition and discuss topics from establishing a participant group, to choosing interesting linguistic structures, deciding on research questions, and understanding some commonly used methods for data elicitation, coding and analysis. This introduction will provide a general understanding of research projects currently going on with adult second language learners, heritage speakers, bilingual children in language shift environments, and much more. No previous background is necessary, and we welcome students from other areas, such as education, communication disorders, and language pedagogy to join us.

On Wed 10/5 at noon in Tobin 521B, Andrew Cohen and Jeff Starns will give a talk that could fit the title, “Thrilling Theories of How to Interpret Data are at Least as Important as the Data.”  You are welcome to join us on zoom if you can’t be there in person.

https://umass-amherst.zoom.us/j/97623669473?pwd=bU9aTVo1c1U3ZVBuVDJ3QmFwUmVHUT09
Meeting ID: 976 2366 9473
Passcode: Cog2223

We will be recording the talk such that Andrew and Jeff can share it with others who aren’t able to make that time.  Please let me know if you have questions or if there are other things that could make tomorrow’s talk or this year’s series more inviting or accessible.

On 9/29 at noon in Tobin 521B we will hear from Gaja Jarosz, (recently promoted to) Professor of Linguistics, who will present work on phonology and exceptionality that has been done in collaboration with UMass graduate students Brandon, Andrew, Maggie, Seoyoung, and Max.

We will also be set up such that you can join by zoom.  The information for all of our meetings this academic year will be:
https://umass-amherst.zoom.us/j/97623669473?pwd=bU9aTVo1c1U3ZVBuVDJ3QmFwUmVHUT09
Meeting ID: 976 2366 9473
Passcode: Cog2223

Everyone is welcome!

Lisa Davidson (NYU) will present, “How phonemic and non-phonemic glottals co-exist: evidence from Hawaiian” on Friday September 30th, 2022 at 3:30pm as part of the Linguistics colloquium series. The presentation will be both in-person, at N400 in ILC, and available through Zoom. Abstract can be found below. All are welcome!

How phonemic and non-phonemic glottals co-exist: evidence from Hawaiian

While studies have examined the linguistic conditions that affect the implementation of phonemic glottal stops (e.g. Triqui, DiCanio 2012, German and Polish, Malisz et al. 2013), there has been less focus on the phonetic implementation of phonemic and non-phonemic uses of glottalization within the same language. This study examines both phonemic and non-phonemic glottal elements in Hawaiian conversational speech to determine whether prosodic factors influence how these two types of glottal elements are employed within the same language. The data comes from a 1970-80s Hawaiian language radio show. Phonemic glottal stops were only produced as a full glottal closure 7% of the time, and segmental factors and word position indicate that creaky realizations are more extensive when there are identical flanking vowels (e.g. /ho?okahi/ ‘one’), and that they occur earlier in the [V?V] sequence when the /?/ is in word-initial position (e.g., /ka#?ulu/ ‘the breadfruit’). A prosodic analysis of the words containing phonemic glottal stop that were parsed using the computational prosodic grammar in Parker Jones (2010) shows that full closures were more likely in prosodic word-initial position (e.g. {(ki:)}{(?a.ha)} ‘cup’). For non-phonemic glottalization at word boundaries, the main factor conditioning the presence of a glottal element was being followed by single-vowel grammatical markers (e.g. [nui#o] ‘big POSS’). 

For the phonemic glottal stop, a full closure may help indicate prosodic word boundaries, which could resolve cases where stress assignment does not disambiguate possible parses, e.g., {(? ho:)}{?o.(‘a.ka)} or {(? ho:.?o)}{(‘a.ka)}, ‘to open’. The preponderance of non-phonemic cases in the content word+single vowel grammatical marker environment may be to ensure that a critical single-vowel grammatical marker is not perceptually subsumed by the preceding vowel. Moreover, non-phonemic glottalization most often occurs where a content word that might begin with a phonemic glottal stop would not be expected, which may allow for both types of glottal elements to co-exist in the language without perceptual confusion or segmentation difficulties. 

(joint work with ‘?iwi Parker Jones)

On 10/14 at 10:00am at ILC N400, Tyler Knowlton (UPenn) will give a talk entitled “Non-conservative quantifiers are unlearnable”. Tyler will be joining us over zoom, and everyone is welcome! The abstract of the talk and the zoom link can be found below.

Join Zoom Meeting
https://umass-amherst.zoom.us/j/91355871706

Title: Non-conservative quantifiers are unlearnable
Abstract: Perhaps the most well-known typological universal in semantics is the “conservativity” of quantificational determiners: roughly, the observation that only the noun phrase a quantifier combines with matters for the truth of the sentence. For example, only fish (and their properties) matter to the truth of every fish swims. Accordingly, no language has a “non-conservative” quantifier like equi, where equi fish swims means “the fish and the swimmers are numerically equivalent.” The hypothetical equi fails to be conservative because the number of swimmers also matters. This robust cross-linguistic generalization has been argued to reflect a fundamental fact about quantifier semantics, as opposed to general cognitive constraints, communicative pressures, or historical accident. In particular, conservativity has been used to question the standard, relational conception of quantifier meanings. But if conservativity has a specifically linguistic source, then non-conservative quantifiers are predicted to be unlearnable. I’ll present a series of experiments that bear out this prediction. Adult participants fail to learn two novel non-conservative quantifier meanings, even when explicitly taught, but succeed at learning their conservative counterparts. Moreover, this learnability asymmetry disappears when the intended meanings are taught as novel verbs instead of novel quantifiers. These results suggest that the conservativity universal is tied to learnability, which supports semantic theories on which conservativity reflects a deep fact about the human language faculty.

This is an invitation to the meetings of LARC (Language Acquisition Resource Center) where we are now organizing to resume sessions after Covid complications last year.

The Center focuses on language acquisition of first, second, bilingual, heritage, multilingual, dialect and indigenous languages, besides language revitalization and language disorders.  No specific theoretical background in linguistic theory or acquisition theory is required or expected to participate at the meetings.  Our meetings have always been full of lively discussions and discussion of broader topics is encouraged.  It is an excellent way to get a sense of what is going on in acquisition in any of these areas.

This Thursday (9/22), Nicholas Sharp will give a talk from 12:00 pm to 1:00 pm Eastern Time in Computer Science Building Room 150/151. More information can be found here.

Bio:

Nicholas Sharp is a researcher in geometry processing, computer graphics/vision, and 3D machine learning. Nicholas’s work seeks new algorithms and new representations to make computing with geometric data easy, efficient, and reliable. Currently, Nicholas is a Senior Research Scientist at NVIDIA, based out of Seattle, WA.