Author Archives: Michael Becker

Lau colloquium Friday March 12 at 3:30

Ellen Lau, University of Maryland, will present “New ways forward in neurolinguistics: more thought, less words” in the Linguistics colloquium series at 3:30 Friday March 12. An abstract follows. All are welcome!

Register here: https://umass-amherst.zoom.us/meeting/register/tJAsde6vrTMjHtCx7Az5AhCgXukmgiFYM5ti

Abstract:
Figuring out the neural underpinnings of language processing is hard; we have to hold onto so many pieces from different disciplines that it’s easy for a few basics to fall off our ‘stack’. I’ll discuss a few that I’ve recently remembered myself, and that give me hope that I could actually have a few coherent things to say when I’m supposed to teach what is known about the neuroanatomy of language processing. First, we neuroscientists of language too often conflate language with non-linguistic thought and conceptual knowledge systems. We forget that when we observe ‘semantic’ neural responses, they may often reflect the activity of these non-linguistic conceptual systems, and we miss too many relevant insights about these systems from research in other domains of cognitive science. Changing my ways, here I will draw from theories about parietal cortex’s role in binding object representations in visual scene perception, to hypothesize that its responses during language comprehension reflect something like the binding of conceptual properties to discourse files. Second, following others in the field (Fedorenko et al. 2020, Matchin & Hickok 2020), I’ll note that neuroscience of language has been missing explicit representational theories about stored language knowledge—the ‘lexicon’—and that this has been blocking our progress on neuroanatomical models of syntax. The folk idea that language is a bunch of simple sound-to-meaning pairs (‘words’) is rooted deep in our collective psyche, and even when we publicly disavow it, ‘under the hood’ it continues to shape how we develop our neuroscience of language theories and how we reason about our experiments. Looking at some problematic cases in my own past work, I’ll argue that as a field we need to stop using vague/incoherent terminology like ‘word’ and ‘lexico-semantic’, and instead commit to sketching out explicit assumptions about how our rich language knowledge is organized every time we embark on an investigation of the neural basis of language production or comprehension.

Georgi colloquium March 5

Doreen Georgi, University of Potsdam, presented “How to account for resumptives in movement chains: insights from Igbo” in the Linguistics colloquium series March 5. An abstract follows.

Abstract:
In this talk I will address the general problem of how we can model the pronunciation of lower chain links when the lower copy is realized in a reduced form such as a resumptive pronoun. There are two main approaches in the literature: BigDP/stranding and spell-out approaches. Both approaches have quite general (conceptual as well as empirical) short-comings, and hence none of them can be considered the standard / widely accepted approach. Van Urk (2018) provides new arguments that favor a spell-out approach that makes use of partial copy deletion. His arguments are based on cross-linguistic patterns of phi-mismatches between the pronounced chain links. In the talk, I will present novel data from my recent (co-authored) work on resumption in Igbo (Benue-Kwa, Nigeria). The phi-mismatch pattern in Igbo is more complex than previously described patterns; in fact, it raises new challenges for a spell-out approach (and also for a BigDP approach) to chain link realization.

Culbertson colloquium Friday February 26 at 2:30

Jennifer Culbertson, University of Edinburgh, will present “Experimental evidence for learning biases in word and morpheme order” in the Linguistics colloquium series at 2:30 Friday February 26. Notice the different time – one hour earlier than usual. An abstract follows. All are welcome!

Register here: https://umass-amherst.zoom.us/meeting/register/tJAsde6vrTMjHtCx7Az5AhCgXukmgiFYM5ti

Abstract:
Recent research has suggested that the cross-linguistic and language-internal frequencies of particular word and morpheme orders might be shaped by constraints on processing combined with learned distributional information (e.g., Hupp et al. 2009, Futrell et al. 2015, Hahn et al. 2020). In this talk I discuss a set of three experiments investigating this claim using artificial language experiments. In the first two sets of experiments, I show that at least some constraints on nominal word and morpheme order in fact reflect universal learning biases, present across populations, independent of their native language. I argue that these biases are driven by simplicity and aspects of meaning, not frequency or other distributional information. In the third set of experiments, I address a well-known claim about the so-called suffixing preference, namely that it results from processing or perception of sequential information. By comparing behavioral results across language populations, I show that is likely not the case. Rather, speakers’ perception adapts to the affix order of their language.

UMass linguists at Biased Questions

UMass visitors, students, and faculty participated in Biased Questions: Experimental Results & Theoretical Modelling, ZOOM Conference Leibniz-Zentrum Allgemeine Sprachwissenschaft (ZAS), February 4–5, 2021, Berlin.

  • Discourse biases
    María Biezma (UMass Amherst)
  • Biased questions in English: an acquisition path
    Rebecca Woods (Newcastle University), Tom Roeper (UMass Amherst)
  • Bias in English and German negative polar questions
    Anja Arnhold (University of Alberta Canada & Universität Konstanz), Bettina Braun (Universität Konstanz), Maribel Romero (Universität Konstanz)
  • Focused NPIs in statements and questions
    Sunwoo Jeong (Seoul National University), Floris Roelofsen (University of Amsterdam)

Pater on local COVID-19 data

Joe Pater published an article “Western Massachusetts counties now have their highest numbers of new Covid-19 cases since the spring (And why are you learning about that here?)” on his blog Nov. 13th. The article discusses the steep rise in local new COVID-19 cases in the fall, as well as the deficiencies in local data reporting. This work was covered in a Nov. 21 Hampshire Gazette article, and Pater was interviewed about in on Bill Newman’s morning show on WHMP on Nov. 25th. A summary can also be found in this CHFA news piece.

Deniz Özy?ld?z’s defense December 3 10am

Deniz Özy?ld?z will defend his dissertation at 10am EST, Thursday, December 3. The title of Deniz’s dissertation is “Attitude Events”. The advisors are Vincent Homer and Rajesh Bhatt, and the committee includes Maria Biezma, Seth Cable and Kristine Yu. A short abstract follows.

Register here: https://umass-amherst.zoom.us/meeting/register/tJ0vc–spj8tHdSqeLytldtLoK11Oyw7z9PM

Abstract

In this talk, we place the verb “think” and its complement clauses under the microscope and see that “think” with a declarative may describe a state, in (1), and that “think” with a question must describe an event, seen by comparing (2) and (3). The state is a belief, and the event, here, a deliberation.

(1) Anna thinks that she should invite Brian.
(2) #Anna thinks whether she should invite Brian.
(3) Anna is thinking whether she should invite Brian.

Aspectual properties of attitude reports, then, interact with properties that attitude verbs have in virtue of their ability to embed clauses, creating non-trivial differences in meaning as well as apparent restrictions in the distribution of embedded questions. To account for such interactions, we must structure attitude eventualities with structures provided by embedded clause denotations, and so, we work towards a system in which it is possible to do so.

Gribanova colloquium Friday November 13 at 3:30

Vera Gribanova, Stanford University, will present “Negative concord, genitive of negation, and clausal ellipsis in Russian” in the Linguistics colloquium series at 3:30 Friday November 13. An abstract follows. All are welcome!

Register here: https://umass-amherst.zoom.us/j/91820615234 

Abstract:
In this talk, I present an in-progress investigation of interactions between the syntax of polarity in Russian and polarity-sensitive items — negative concord elements (NEG-words) and DPs marked with the genitive of negation (GoN-DPs) — in the environment of clausal ellipsis. Though both NEG-words and GoN-DPs must generally co-occur with clausemate negation in Russian, it has been known for some time that the syntactic licensing conditions for these two phenomena are in fact distinct (Franks and Brown 1995; Brown 1999). In the first part of the talk, I provide a syntax for these licensing conditions and demonstrate that this syntax, in conjunction with the application of clausal (TP) ellipsis, gives rise to the differences we observe between NEG-words and GoN-DPs as fragment answers: NEG-words can be licensed as fragment answers in the absence of an overt expression of negation in the antecedent, but GoN-DPs cannot. These differences in behavior follow from three interrelated commitments: first, that in Russian there is a low position for polarity, association with the expression of sentential negation, and a high one, which is null but semantically interpretable (Brown and Franks, 1995; Brown, 1999; Gribanova, 2017); second, that there can be fronting of the NEG-word to the left periphery in conjunction with TP ellipsis (Giannakidou, 1998; Merchant, 2004); and third, that Russian NEG-words are licensed by the higher instance of polarity (Laka, 1994; Zeijlstra, 2008) while GoN-DPs are licensed by an AGREE relation with the low expression of negation (Franks and Brown 1995, Brown 1999, Harves, 2002, Abels, 2005).
In the second part of the talk, this unified picture meets with a set of challenges that arise from the interaction between Gon-DPs and contrastive polarity ellipsis (Kazenin 2006; Gribanova 2017), in which clausal ellipsis is combined with the fronting of a contrastive DP to the left periphery, preceding a polar particle (‘yes’ or ‘no’). For some native Russian speakers, such configurations give rise to violations of the case connectivity effect usually associated with the phrasal remnant: genitive patients under negation in the antecedent can, in a narrow set of circumstances, correspond to an accusative patient remnant outside the ellipsis site. Although these effects seem to contradict prominent ideas about the identity relation necessary to license ellipsis (Chung 2013, Merchant 2013), I point out that they might be better understood in light of recent work that takes the domain for identity in clausal ellipsis (e.g. in sluicing) to be smaller than has traditionally been assumed (Rudin 2019, Anand, Hardt, McCloskey in progress).

Breen colloquium Friday October 30 at 3:30

Mara Breen, Mount Holyoke, will present “Hierarchical linguistic metric structure in speaking, listening, and reading” in the Linguistics colloquium series at 3:30 Friday October 30. An abstract follows. All are welcome!

Register here: https://umass-amherst.zoom.us/meeting/register/tJUldemurz4oGdAo6hV69nh4k3y82zRiLVZB

Abstract
In this talk, I will describe results from experiments exploring how hierarchical timing regularities in language are realized by speakers, listeners, and readers. First, using a corpus of productions of Dr. Seuss’s The Cat in the Hat—a highly metrically and phonologically regular children’s book, we show that speakers’ word durations and intensities are accurately predicted by models of linguistic and musical meter, respectively, demonstrating that listeners to these texts receive consistent acoustic cues to hierarchical metric structure. In a second experiment, we recorded event-related potentials (ERPs) as participants listened to an aprosodic production of The Cat in the Hat. ERP results reveal separable electrophysiological indices of metric and phrasal processing, demonstrating top-down realization of metric structure even in the absence of explicit prosodic cues. In a third experiment, we recorded ERPs while participants silently read metrically regular rhyming couplets where the final word sometimes mismatched the metric or prosodic context. These mismatches elicited ERP patterns similar to responses observed in listening experiments. In sum, these results demonstrate similarities in perceived and simulated hierarchical timing processes in listening and reading and help explain the processes by which listeners use predictable metric structure to facilitate speech segmentation and comprehension.

UMass redesign of Praat logo

Praat (Boersma & Weenink 1991), an open source program for phonetic analysis and phonological grammar modeling, recently received its first logo redesign in nearly thirty years. The new version of the beloved mouth-above-ear logo is the work of our linguistics major Larry (Sichen) Lyu, class of 2022. Bravo, Larry!

Pictured: the old (left) and new (right) Praat logo.