The world is full of languages and dialects—more than 7,000. Across these languages, many possible sounds can be combined into words. While there may be similarities in words between closely related languages, for years linguists have believed that the relationship between the sound of a word and its meaning is completely arbitrary. Recently Morten H. Christiansen, Psychology, collaborated with fellow researchers to investigate whether that belief might need to be reappraised.
“We looked at nearly two thirds of the world’s languages from across every continent,” Christiansen says. “And we compared from 40 to 100 different words for concepts or things—like body parts, kinship terms, celestial bodies, and pronouns such as I and you. These are all words that anthropologists have identified as especially important in all languages. Contrary to what people thought, we found that there are reliable sound-meaning patterns across the languages of the world. People will tend to use the same sounds when they come up with a word for a particular concept.”
Sound-Meaning in Languages
Christiansen and his fellow researchers found that if a language has a word for the color red, it is likely to have an r sound in it. Likewise, the word for nose will probably include a nasal or n sound. This relationship between sound and meaning is called a signal. Along with these positive signals, the researchers also found tendencies to avoid certain sounds—or negative signals. The words for the pronouns I and you, for example, tend to avoid p, t, and s sounds.
These probabilistic biases are not universal, Christiansen points out. There are always some languages that do not conform to the tendency. “But we have very reliable evidence based on big data. No one has ever looked at this many words across so many languages before. So we know these signals are there.”
In their analysis, the researchers controlled for the most likely explanations for the existence of signals: contact between languages, resulting in the borrowing of words or sounds and linguistic heritage which results in words or sounds passed down through related languages. They showed that these explanations could not account for the existence of signals across so many languages.
“Given that the signals keep showing up independent of cultural relationship—or any other type of relationship—between the languages suggests it’s common to all humans,” says Christiansen. “It may have to do with our biology. Perhaps because of the connections in our brain, the way we perceive the world and the way we process sound, some labels just may seem more natural such as the inclusion of an r sound in the word for the color red.”
How Language Works
Understanding the reasons for sound-meaning signals is just one of many research questions into which Christiansen is currently delving. As the director of the Cognitive Neuroscience Lab at Cornell, his interests range across all the language sciences and into other disciplines as well. His research integrates theoretical insights from the areas of language evolution, language acquisition, and language processing. “I think we need to look at language in the combined context of how we use it, how it’s acquired, and how it has evolved,” he says. “I try to get convergent evidence using different methods. To answer a particular question, for instance, we might combine data from neuroimaging, cultural transmission experiments, and studies of language acquisition in children, and so on.”
Christiansen’s latest book, Creating Language: Integrating Evolution, Acquisition and Processing (MIT Press, 2016)—co-authored with Nick Chater from the University of Warwick, UK—is an overview of his work for the last 20 years. It grew out of his concern that the language sciences approach questions relating to the evolution, acquisition, and processing of language as completely independent of one another. “In this book I’m trying to reintegrate the language sciences to provide a complete account of how language works,” Christiansen says.
“We need to look at language in the combined context of how we use it, how it’s acquired, and how it has evolved.…To answer a particular question…we might combine data from neuroimaging, cultural transmission experiments, and studies of language acquisition in children.”
One chapter Christiansen is especially excited about focuses on some of his newest research on how we understand language. “Language happens in the here and now,” he says. “We have all experienced losing the thread of a conversation simply because our attention was diverted for even a moment.”
Many models for investigating language ignore that aspect of language processing. This is problematic because our ability to process what we hear is limited by a number of factors, Christiansen explains. To begin with, auditory sound traces disappear in less than a tenth of a second. Also, in normal speech we produce about 10 to 15 phonemes (distinct units of sound) per second, which is extremely fast. To make things even more difficult, our ability to remember sequences of items is very poor—no more than four or five random sets of letters or digits, or even words, at a time. “People have known these things for a long time,” Christiansen says, “but they haven’t taken them into account when thinking about theories of language. We suggest that the solution to this now-or-never bottleneck is that the system essentially chunks or compresses things as quickly as possible and passes it up to a higher level of linguistic representation.”
This means when we process language, we chunk together sound-based units in order to remember more of them. Very quickly we must move those units up to another level and form them into words. Then before there are too many words to remember, we must move the words up another level to form them into sentences. If we don’t manage to do this chunking, we quickly lose the meaning of what is being said.
Danish, a Language Dominated by Vowels
In some languages, chunking is easier to do than in others, Christiansen says. This may help explain the difficulty inherent in learning Christiansen’s own native language, Danish. “It’s well known that Danish is hard to learn for adults,” he says. “But it turns out it’s hard for Danish children to learn as well. That’s really interesting because it counters a common assumption that all languages are equally easy for children to learn as a first language.”
Christiansen is planning a research project with colleagues in Denmark to delve into the mysteries of Danish. He already has some ideas about why children have problems learning it. One may be because Danish has many vowel sounds while consonant sounds are deleted. “With those long stretches of vowel sounds in Danish, children have a hard time figuring out where one word stops and the next begins,” he says. “By the time they’ve processed what was said, it’s just too late, so they miss out on some of the words in the utterance.”
The study of language has important implications on many levels. Along with applications to the treatment of language problems in children and adults, it also is a defining characteristic in our species. “Language is integral to what it means to be human,” Christiansen says. “Imagine how devastating it would be if you woke up tomorrow and didn’t have the ability to use language. It allows us to have culture and knowledge that can be transmitted across generations. It’s something very special.”