There is a famous illustration that shows a duck or, if you tilt your head slightly to the left, a rabbit. That reversible figure has come to symbolize the work that Noa Attali has been doing in the language science department's colloquially named AmbiLab, or Ambiguity Lab.

"There's so much that we're saying all the time that has multiple potential meanings," says Attali. "But we often arrive effortlessly — without even thinking about it — at that one meaning that the speaker meant, maybe even without thinking about the alternative. And so how do we do that?"

These questions have guided Attali since she arrived in Irvine in 2019 from Rutgers, where she had found herself torn between two possible paths as an undergraduate. "I took English classes because I love books. I love reading," she says. But she was also taking classes in subjects including linguistics, cognitive science, behavioral economics, and language science.

"I had to make this big decision between English and the psychology of language," says Attali. She ended up double majoring, doing both an English honors thesis on Shakespeare and an honors thesis on cognitive science with Karin Stromswold in her language lab. When it came time to apply to grad school, she assumed that linguistics departments would be too theory focused for her and set out to find cognitive science departments in which people were studying language. She found Lisa Pearl.

"From the second I came, she welcomed me with open arms," says Attali. In the spring of 2019, she visited the UCI campus for an orientation day, where she met with Pearl - a professor of language science - and Greg Scontras, who is now her second advisor, as well as Richard Futrell, both now associate professors of language science. "They were just so charming," recalls Attali. "Everyone seemed so different from other faculty that I had seen. They surprised me a little bit."

Attali followed Pearl to the newly created language science department, where she's been pursuing a Ph.D. focused on how we understand one another when a statement could have multiple interpretations. She's curious how listeners decide what speakers mean, especially in naturalistic, spontaneous conversations, where things get messier than they look in a lab.

"This is a question at the heart of how we as humans use language to communicate with each other," says Pearl.

The interdisciplinary nature of UCI's language science department enables Attali to tackle this question. "On the one hand, we have these very rigorous methods like computational cognitive modeling. It's mathematical," says Attali. "On the other hand, there's also this flexibility that my advisors brought. We use any method we need to answer the question, and we use multiple methods and combine them together to whatever you need in order to answer the question."

Building on research that Pearl and Scontras had done previously, Attali zeroed in on a specific grammatical format that's been dubbed an "every-negation" sentence. "Every something didn't something," says Attali, giving the example, "'Every horse didn't jump over the fence.' It could mean either 'none of the horses jumped over the fence' or 'not every horse jumped over the fence.'" It's an ostensibly subtle difference, but an important one when it comes to conveying an absolute meaning.

"Doing a Ph.D. means taking a subject so specific that it strains human beings' ability to care about it," laughs Attali, remembering something she heard prior to starting her Ph.D.

In her research, she's found literature advising against the use of every-negation sentences. "There are all these old papers and philosophers of language a long time ago saying that this kind of sentence isn't okay. You shouldn't produce these potential ambiguities because they're confusing, and it's not right — this is a wrong way to use language, or an illogical way to use language," she says. But in daily conversations, it turns out, people are rarely stumped by such phrasing. "Ambiguity is not a problem, because there's so much information. The context is so rich and includes clues about what the speaker meant."

Attali pored over radio and corpus archives of transcripts, looking for examples of every-negation sentences."We took those cases and we saw what was going on in the context, what was going on with the prosody, and how do people interpret them with the context or without the context, with the prosody, without the prosody." Attali found patterns with prosody, the melody and ups and downs that occur when we speak naturally. Speakers tend to emphasize the quantifier to clarify meaning. "If they said, 'Everything's not okay,' that's more likely to mean not everything's okay," she says. Likewise, there are patterns with context: if the preceding context established the expectation that everything is okay, then the speaker is more likely to mean that not everything’s okay.

"Noa has a passion for working with empirical data and using quantitative methods, which is something she’s been able to hone further in our program and is something fairly unique in the area of language science she’s working in," says Pearl.

Attali defended her Ph.D. in August — the first student to do so after completing the full departmental curriculum (the program's two prior Ph.D. recipients arrived in 2019 ABD and defended their theses in 2021 and 2022). "We’re so delighted that Noa’s been part of our language science Ph.D. program," says Pearl, who notes that Attali is already influencing others. "Noa's drive and perseverance throughout her graduate career has been a guiding light for the more junior members of our department, who often look to her for inspiration on how to succeed in the face of specific challenges."

Every grad student doesn't do that.

-Alison Van Houten for UCI Social Sciences

-photo by Luis Fonseca, UCI Social Sciences