Language Projects Have a Different Take on Learning a Language

Researchers Gary Dell, Cynthia Fisher, and Jennifer Cole from the Biological Intelligence research theme are showing that language learning never stops.

According to Gary Dell, he and his fellow linguistics and psycholinguistics researchers at the Beckman Institute have a motto.

“We claim that language learning never stops,” he said. “That’s our motto, that we’re constantly making small adjustments in the way that we understand and produce language to reflect our current circumstances.”

That may not seem like a controversial creed but it is one that goes against the grain of traditional thinking about the process of learning a language. Dell and fellow Biological Intelligence (BioIntel) research theme faculty members Jennifer Cole and Cynthia Fisher are among those whose work demonstrates this motto as they explore the theme of humans updating their grammatical knowledge over the lifespan. Cole is one of those in the field of linguistics who is challenging conventional thinking about language learning.

“For a long time linguists have thought that you acquired the basic grammatical knowledge of your language when you were a child and then by the time you hit puberty that knowledge is pretty much complete,” Cole said. “You may add new words but the structures are finished and then you just use them for the rest of your life.

“In recent years there has been increasing evidence that in fact we continue to update our grammatical knowledge. So our competence as speakers of a language, the knowledge that we have that allows us to communicate with language, is continuously changing, and reflecting our experience as we use the language – as we speak and as we listen.”

Cole is a professor of linguistics at the University of Illinois while Dell and Fisher are professors of psychology. The three are currently collaborating on a project involving perception and production in the domain of phonology, which refers to how the sounds of language are put together to form words. The project investigates an aspect of the phonological processing system called phonotactic constraints, which are language specific generalizations about how words are pronounced.

Cole’s role in the project is focused on linguistic analysis and speech perception experiments, while Dell concentrates on speech production and computational modeling, and Fisher on speech production and the studies involving children.

The research line is demonstrating that language learning is an ongoing process that, as the researchers have written, “adapts to recent experience, while continuing to reflect the accumulated experience of a lifetime of speaking and listening.” Their research is showing that phonotactic constraints can be changed, to a large extent in fact, by experience.

"In recent years there has been increasing evidence that in fact we continue to update our grammatical knowledge. So our competence as speakers of a language, the knowledge that we have that allows us to communicate with language, is continuously changing, and reflecting our experience as we use the language – as we speak and as we listen."
– Jennifer Cole

Dell said their projects involving phonotactic patterns incorporate experiments that look at language production, language perception, and language and infants.

“In all of those experiments what we are studying is the learning of what are called phonotactic patterns, meaning how the sounds of language are put together to form words,” Dell said. “For example, in English the “ng” sound only occurs at the ends of syllables, so we say sing, but we would never reverse that. It sounds weird. The fact that it sounds weird to you reflects your knowledge of phonotactics.

“In other languages syllables can begin with an “ng”. It’s just that English has its own special rules. These are learned early in life. What we do in the laboratory is we expose subjects to syllables that follow artificial phonotactics.”

The project employs experiments involving both children and adults as subjects, linguistic analyses with linguistic studies of the properties of languages of the world, and computational modeling. The laboratory experiments use a method called artificial grammar learning, with experiments designed using nonsense words to test whether subjects will follow phonological rules created by the researchers.

Dell said that in perception experiments, they have found that test subjects are very accurate when it comes to following the artificial rule given in the experiment. He said a key finding of the project has been that language learning is not only ongoing, it is also implicit.

“It shows the subjects have implicitly learned, which means they are learning without being aware that they are learning,” he said. “That is what the grant is about – it reveals the implicit learning process that adapts our linguistic system so that we are able to produce and understand language.”

"That is what the grant is about – it reveals the implicit learning process that adapts our linguistic system so that we are able to produce and understand language."
– Gary Dell

The project was first funded in 2003 and was renewed in 2008. Dell said the seminal research in this area was reported in 2000 in the Journal of Experimental Psychology  in a paper titled Speech Errors, Phonotactic Constraints, and Implicit Learning: A Study of the Role of Experience in Language Production.

The most recent published paper came out in 2009 in Cognition, and was titled Speech errors reflect phonotactic constraints in recently spoken syllables but not in recently heard syllables. Dell said the paper looked at the issue of perception to production transfer.

“If you hear syllables that follow one of these artificial phonotactic constraints, then the question is will it influence your own speech production?” he said. “Will it influence your speech errors? The answer is, largely, no. The learning that occurs is internal to the speech production system, or internal to the speech perception system.”

Cole said the research is pushing the limits as far as understanding what kinds of sound patterns listeners can implicitly learn from simple auditory exposure. They are also currently running experiments that will test hypotheses about linguistic elements that may cut across different languages.

Cole said she and her graduate students Erin Rusaw and Karen Lichtman are currently testing whether subjects “can learn typologically rare sound patterns as easily as common ones, and whether in this rapid, implicit learning paradigm, subjects can associate new sound patterns with the voices of particular speakers, simulating what goes on when children grow up in a multi-lingual or multi-dialectal environments.

“Our research shows that adult subjects are unconsciously updating their grammar in a sense to reflect the new pattern that is found within the experimental language,” Cole added. “They lose the update later because it is not reinforced out in their real world, but it gives us an idea about the mechanism by which we continue to adapt language as we go through life.”