Children seem to acquire new languages as easily as they do airborne illnesses, summer tans, and BFFs. Adults, with some exceptions, do not.
A number of hypotheses have been proposed to explain this depressing developmental trajectory: a biological window that slams shut at puberty, a more gradual lifetime decline in the brain’s plasticity, even an increase in the size of the linguistic “chunks” we pay attention to as our memory spans increase throughout childhood.
My own favorite, though, suggests that adults face an architectural problem: because our brainscapes are already cluttered with connections, we have trouble making new neural connections, the ones necessary for language proficiency. It’s the difference between building a highway through virgin meadowlands and constructing it right down the center of Manhattan.
Or, think of it as writing on blank paper instead of newspaper, as researcher Joe Z. Tsien put it: “The difference is not how dark the pen is, but that the newspaper already has writing on it.”
But Tsien’s explanation wasn’t a metaphor for language learning in adults; I encountered it in a New York Times article about preternaturally aged mice who’d lost the ability to form long-lasting memories.
Neurons in the brain fire, or generate a brief electrical pulse, to communicate with each other. When they fire, chemicals released by the neuron travel across a small gap and link to receptors on another neuron. Different chemicals are able to link to different receptors; thus, the nature of these receptors matters. After puberty, it turns out, humans produce fewer of one kind of receptor, NR2B, and more of another, NR2A.
Neuroscientists at Georgia Health Sciences University modeled adult human brains by bestowing upon an unlucky group of mice, via genetic modification, an overabundance of NR2A. They then compared the performances of mutant mice and ordinary mice on tasks that included recognizing familiar objects (as measured by how long the mice spent exploring novel versus familiar objects) and forming new associations (think Pavlov’s dogs, but with tones instead of bells, mild electric shocks instead of food, and freezing in place instead of salivating). Importantly, the mutant mice performed just as well as the ordinary ones—at first. But after a 24-hour delay, the mutants failed to recognize familiar objects or to associate tones with electric shocks.
So just what about the preponderance of NR2A was hurting the mutant mice’s ability to form long-lasting memories? To find out, researchers sent an electric current through slices of tissue (still living, though their hosts had been sacrificed) from the hippocampus, a region of the brain important for memory formation. The researchers hoped to determine the plasticity, or malleability, of the mutant mice’s hippocampal neurons: How responsive would they be to (simulated) communication from another neuron?
They expected to find differences between ordinary and mutant mice in the degree to which communicative channels could be strengthened. What they actually found, however, was a difference in the degree to which connections could be weakened: the mutant mice did not have a building problem; they had a pruning problem. Long-term learning, for these mice, was like drawing lines on newsprint—or building a highway through Manhattan.
Genetically engineered mice are not men, and even the most learned among them understands no Mandarin. But this study offers some of the best evidence that overcoming existing connections may be the real hurdle adults face when confronted with a new language. Indeed, it suggests that two factors might be working against us: a larger number of existing connections to reroute and a decreased ability to reroute them.
Should these findings be replicated, the next question will be one posed by Joanna Workman, a psychobiologist at the University of British Columbia, who was kind enough to give me her take on the study (“compelling”). It would be really exciting, she told me, to make an older brain young again using a similar manipulation.
Permission required for reprinting, reproducing, or other uses.