Building Blocks of Talk

When you tell someone to stop babbling, your meaning is obvious: Quit talking gibberish and speak clearly.

Give the youngest babblers a break, though. Scientists now suspect that the inveterate babbling of 7- to 12-month-old babies contains remarkably systematic vocal features that, soon enough, render them able to utter such parental favorites as “No,” “More,” and “Why?” A better understanding of babbling may even allow researchers unprecedented glimpses into the controversial prehistory of human languages.

In fact, the gift of gab appears to grow out of roots that go much deeper than babbling. One line of research finds that by age 3 months, babies communicate by combining vocalizations with facial expressions. Infants exploit their hand-to-mouth existence—whatever their hands grasp gets jammed into their mouths—to learn how to communicate with precisely timed words and gestures, according to other investigators.

“Words are put into play by movements of the human body,” says psycholinguist John L. Locke of the University of Cambridge in England. “Physical effects [on learning to speak] may be more pervasive, and languages less arbitrary in structure, than linguists have previously supposed.”

Jaw movements

Babbling research has come of age over the past 20 years. Investigators have established that between ages 7 and 10 months, infants begin to alternately lower and raise the jaw while making consonant and vowel sounds. Babbling of repeated sequences of consonant and vowel combinations then takes off.

For instance, opening and closing the lips with the tongue flattened allows a baby to say “baba.” Open the lips and thrust the tongue tip off the mouth’s roof to get “dada.” Bunch the tongue up against the back of the mouth for “gaga.”

Little by little, yet with astounding speed, these manipulations of speechlike sounds usher children across the threshold of talk.

During the same period that babbling research got off the ground, several maverick teams of linguists searched for speech sounds that are applied comparably in all or most of the world’s languages. They reported such universal language consistencies as the use of single, distinct consonants and syllables that consist of a consonant followed by a vowel.

Enough of these shared linguistic properties exist to allow for the reconstruction of precursor languages all the way back to a prehistoric “mother tongue,” according to some of these renegade researchers.

Mainstream linguists reject such claims. They argue that rampant changes and innovations by speakers of a language erase any of the features it might have shared with other languages more than about 5,000 years ago. The relentless pace with which people tweak their native tongues renders hopeless the search for linguistic universals, according to this view (SN: 6/9/90, p. 360).

Support for the mavericks’ argument now comes, literally, from the mouths of babes. Peter F. MacNeilage and Barbara L. Davis, both psycholinguists at the University of Texas at Austin, say that a handful of basic sound patterns lurks within babies’ babbling, infants’ first words, and the reconstructed words of long-dead languages. Moreover, reconstructed words that have similar meanings across different language families usually contain matching arrangements of these fundamental sounds, MacNeilage and Davis report in the April 21 Science.

“Our findings provide totally independent support for the theory that there was one original language, a mother tongue,” MacNeilage says. “It’s wrong to assume that related languages cover up their tracks every 5,000 years.”

Strings of speech sounds

MacNeilage and Davis statistically analyzed strings of speech sounds in the babbling of six babies and the first words of 10 slightly older infants. Tiny microphones attached to bibs worn by the diminutive participants, ages 6 to 18 months, recorded their vocalizations as they played with an English-speaking parent. The researchers also examined extensive lists of words from 10 modern languages, including English, Japanese, and Swahili.

MacNeilage and Davis identified three sequences of sound patterns common to babbling and words in general. Each consists of a consonant-vowel combination: lip consonants leading into vowels generated in the center of the mouth with a flattened tongue (such as “mama”); tongue-front consonants preceding vowels produced at the front of the mouth (such as “dada,” the vowel pronounced as in “daddy”); and tongue-back consonants followed by back-of-the-mouth vowels (such as “gogo”).

Additional studies directed by MacNeilage and others have identified these three sound sequences in the babbling of infants exposed to native French, Swedish, Japanese, Portuguese, and Quechua, a language spoken in Ecuador.

A fourth, more complex sound sequence appeared in infants’ first words and across all 10 languages, but not in babbling. Sounds in this three-part sequence start with a lip consonant, followed by a vowel and then a tongue-front consonant (as in “mad”). Other researchers have noted that people have an easier time articulating speech sounds from the front to the back of the mouth, such as “pug,” than the reverse, such as “gap.”

The physical arrangement of the mouth and the vocal tract—which by a baby’s third month lengthens and bends dramatically—encourages certain sequences of oral movements, or frames, during speaking, MacNeilage theorizes. The first words of human ancestors may have been built out of these frames and thus may have sounded much like the first words of modern infants, he proposes.

Attempts to string together frames in more complex ways—as in the front-to-back-of-the-mouth articulations of budding talkers—then fueled the evolution of more complex words, in MacNeilage’s scenario. He refers to this two-step process as the frame-content theory of speech evolution. According to the theory, mouth movements essential for simple speech sounds would have laid the groundwork for devising a meaningful vocabulary.

The three consonant-vowel frames found in babies’ babbling also frequently appear in lists of words from 27 ancestral languages from different parts of the world, the Texas researchers say. Maverick linguists reconstructed these lists from groups of related modern languages. In addition, reconstructed ancestral words contain many more instances of the front-to-back sequence of consonant and vowel sounds observed in infants’ first words than of back-to-front sequences.

MacNeilage says the new data challenge claims by critics of ancient language reconstruction that the sound similarities of words with the same meaning in different languages often arise by chance.

Frame content

Among scientists who study babbling and other types of infant vocalizations, the frame-content theory has attracted considerable interest. MacNeilage set out his position in the August 1998 Behavioral and Brain Sciences and received generally favorable comments in 27 published peer reviews from psycholinguists and developmental psychologists.

“I think the frame-content hypothesis has real merit,” remarks psycholinguist D. Kimbrough Oller of the University of Maine in Orono, who didn’t write a published review of MacNeilage’s paper. “This provides an interesting addition to the debate over the reconstruction of ancient languages.”

Beginning around 20 years ago, Oller helped pioneer the study of speechlike sounds infants make while babbling. He summarizes research to date on infant vocalization in The Emergence of the Speech Capacity (2000, Erlbaum).

Before 6 months of age, infants emit a variety of sounds that appear to prepare them for babbling, Oller notes. These include squeals, growls, incomplete speech sounds, and the lip-rattling blasts known as raspberries. Such vocalizations, as well as smiling, gain in frequency when infants younger than 6 months gaze at their mothers.

Moreover, youngsters begin to coordinate vocalizations with their facial expressions months before they start to babble, presumably to communicate with their caregivers, Oller and his coworkers reported in the March 1999 Developmental Psychology.

The scientists videotaped 12 infants, each at age 3 months and again at age 6 months, interacting with one of their parents. Each parent was instructed to divide time between playing with the child and staring at a picture on the wall while maintaining a neutral facial expression.

In these situations, both 3-month-old and 6-month-old infants paired specific vocalizations with their facial expressions, Oller holds, although the older kids did so more frequently. For instance, a 3-month-old child confronted with a wall-gazing parent blurted out growling noises just after adopting an angry expression. The ominous sounds faded out shortly before the youngster stopped looking upset.

Vocalizing that precedes babbling may even contain rudimentary types of syllables, according to psychologist Kathleen Bloom of the University of Waterloo in Ontario. She argues that from around 3 to 6 months of age, infants make “syllabic” sounds that have more in common with babbling than with growls and most of the other prebabbling sounds cited by Oller. Syllabics consist of sustained vowellike and consonantlike sounds with acoustic frequencies comparable to those measured for babbling, Bloom asserts.

In a series of studies from 1988 to 1996, Bloom reported that parents and other adults talk more to infants who make syllabic sounds, view those infants more favorably, and tend to regard infants’ syllabic sounds as attempts at communication.

Syllabic sounds increase in length and acoustic complexity from around 2 months to 4 months of age but then become shorter and less distinct for about a month before regaining momentum toward babbling, according to a long-term study of 13 infants that was directed by psychologist Hui-Chin Hsu of the University of Georgia in Athens.

Vocalization, from prebabbling to speech, requires the coordination of more than 70 muscles and many different body parts, ranging from the diaphragm to the lips, Hsu’s team says. The temporary setback in the ability to vocalize probably results from major structural changes in the vocal tract that occur at around 4 months of age, upsetting the balance of the vocalization system that infants had been working to master.

Vocalization may thus develop in fits and starts that vary greatly from one child to another, the researchers propose in the March Infant and Child Development. Other researchers have applied this perspective, known as dynamic-systems theory, to motor development (SN: 3/20/99, p. 184).

A contrasting view of speech development posits the existence of progressively unfolding, genetically controlled stages of vocal ability that diverge little from one child to the next. This approach draws inspiration from the influential theory that infants have a genetically programmed ability to understand grammatical rules (SN: 5/3/97, p. 276: https://www.sciencenews.org/sn_arc97/5_3_97/bob2.htm).

Oller sees much value in dynamic-systems theory. However, he and many other psycholinguists remain skeptical about reports of syllabic vocalizing by 3- to 6-month-olds. Syllabic sounds are poorly defined acoustically, Oller contends.

Seamless combination

From a dynamic-systems perspective, infants and young children learn to communicate by seamlessly combining two movement systems, one for speech and the other for gestures, according to Jana M. Iverson of the University of Missouri in Columbia and Esther Thelen of Indiana University at Bloomington. It’s time to reconsider the popular notion of gestures as a decorative sideline to the real business of saying what you mean, the two psychologists argue. Instead, gabbing and gesticulating go hand in hand.

“In adults, language and movement are very closely related in the brain,” Iverson says. “The question then becomes, How did they get that way?”

Mental activity arises through bodily interactions with the world, a process that forges a deep connection between talk and gestures, Iverson and Thelen argued in the November/December 1999 Journal of Consciousness Studies. For example, as Iverson and a colleague reported in 1998, even blind speakers gesture while talking to blind listeners. Arm movements apparently play an integral role in a speaker’s ability to express ideas, whether or not anyone else sees the motions.

In learning to communicate, infants probably employ some type of general ability to produce rhythmic sequences of both hand movements and speech sounds, Iverson suggests. Rhythmic arm and hand movements emerge at about 6 months of age, at roughly the same time as rhythmic babbling.

The tight linkage of mouth and hand movements also explains the emergence of so-called “manual babbling” in both hearing and deaf infants, Iverson says. First reported for deaf babies in 1991, manual babbles consist of cycles of repeated gestures that have no meaning. For instance, a child might rhythmically jab the index finger of the left hand into the palm of the right hand.

Researchers initially explained this as evidence for an innate language-learning facility that works through gestures instead of speech in deaf babies. However, a 1995 study found that hearing infants with no exposure to sign language babble with their hands as well as their mouths.

Manual babbling represents a rhythmic behavior that appears during a child’s transition to achieving fine-motor control over the hands, just as rhythmic vocal babbling heralds a shift to speech-worthy control over the vocal tract, Iverson argues.

By adulthood, people unthinkingly time each of their gestures to coincide precisely with the word or phrase it accompanies, she notes. Investigators have found that when a speaker stutters over a word, an accompanying gesture tends to freeze until the moment that speech resumes.

Still, comprehensive explanations of speech development—and the capacity for understanding grammar in particular—elude scientists. If the babies babbling into microphones and repetitively poking their palms have an opinion on the matter, they’re not saying.