top of page

Selected Projects
(updated: April 2022)

A revolutionary reading experience

Wordtune Read is an AI reader that summarizes long documents so you can understand more, faster.

Whether you are reading a lengthy blog post, analyzing a report or going through a paper, Wordtune helps you digest what's important. Cut through the word count to find the words that count.

Click here and start reading! 

 

By the hard working team members of AI21Labs!

Untitled-3.png

Express yourself! 

Wordtune is an AI-powered writing companion that helps you translate your thoughts into words.
Its technology understands exactly what you are trying to say and provides tailored suggestions to make your writing more clear, compelling, and authentic.

Download it here for free and tell me what you think! 

 

By the hard working team members of AI21Labs!

Untitled-1 (1).png

When people interfere

When producing object relatives clauses, Hebrew speakers can either use a resumptive pronoun (that’s the book that the boy moved it) or not use one (that’s the book that the boy moved _). This study investigates whether and how their choice is affected by semantic similarity between planned arguments, e.g. the examples above vs. that's the girl that the boy moved (her).

Fadlon, Morgan, Meltzer-Asscher, Ferreira (Submitted, in revision)

DSC_0024.JPG

Looseness has consequences

Restrictive relative clauses are more semantically integrated with the nominal head they modify than non-restrictive relatives are (“loose clauses”, Jespersen 1954). Does this difference modulate the way English and Hebrew speakers form these dependencies upon production? (TL;DR: it does. Paper is in prep, but see CUNY 17 poster here).

Fadlon, Morgan, Meltzer-Asscher, Ferreira 

DSC_0079.JPG

This pronoun is being so extra

Speakers of both intrusive resumption languages (e.g. English) and grammaticized resumption languages (e.g. Hebrew) use more resumptives when creating a dependencies across syntactic islands. A cross linguistic study combining production, comprehension, acceptability and structural priming experiments will hopefully help us understand if this apparent similarity is due to the implementation of distinct (i.e. grammatical vs. extra-grammatical) mechanisms. Preliminary results are intriguing.

Morgan, Fadlon, Meltzer-Asscher, Ferreira

DSC_0072.JPG

Is it for free?

There is ample evidence indicating that when comprehenders process long distance dependencies (such as the one in that’s the singer that the enthusiastic fan stalked _ last week) they anticipate the position in which the filler (singer) gets interpreted again. But what happens when a pronoun occupies this position (that’s the singer that the enthusiastic fan stalked him last week)? A self-paced reading study of Hebrew, where the latter option is acceptable, suggests that having to change the manifestation of this position from gap to pronoun incurs processing difficulty. 

Fadlon, Keshev, Meltzer-Asscher

DSC_0013.JPG

I see rojo

Bilinguals have the ability to control and manage their two languages. Is the language control system tied only to habitual associations between cues to language choice or constantly attunes itself so that in theory any cue that happens to predict L1 or L2 can be exploited efficiently? A mixed-paragraph study answers this question by comparing how visual cues for within-sentence switching are utilized across two bilingual populations: Hebrew-English and English-Spanish.

Fadlon, Li, Prior, Gollan (Submitted)

DSC_0018.JPG
bottom of page