
(Photo by Ana Tablas on Unsplash)
RICHARDSON, Texas — Have you ever wondered how babies learn to talk? A new study reveals that the adorable sounds infants make in their first year of life are far from random – they’re calculated explorations into the realm of vocal communication. The research, conducted by a team of researchers from across the United States, provides valuable insights into the early stages of speech development and how babies practice making various sounds.
When we think about babies learning to talk, we often focus on their first words. However, long before babies say “mama” or “dada,” they go through an important stage of vocal exploration. During this time, they produce a variety of sounds that aren’t quite words but are crucial stepping stones towards speech. These early vocalizations are called “protophones,” and they include three main types: squeals (high-pitched sounds), growls (low-pitched or harsh sounds), and vocants (vowel-like sounds).
The researchers were particularly interested in how babies produce these different types of sounds over time. They noticed that babies often make certain types of sounds in clusters – for example, a baby might make several squeals in a row, then switch to making vocants for a while. This clustering behavior suggests that babies might be practicing these different sound categories, which is an important step in developing the ability to produce the varied sounds needed for speech.
To investigate this idea, the research team conducted a large-scale study involving 130 typically developing infants. They recorded these babies in their homes once a month throughout their first year of life, using special all-day recording devices. This approach allowed the researchers to capture a comprehensive picture of how babies vocalize in their natural environment.
The finding of the study, published in PLOS ONE, were eye-opening. The researchers discovered that infants showed clear patterns of clustering their vocalizations from a very young age. In fact, across the entire first year of life, babies demonstrated non-random patterns in producing squeals and growls. This means that rather than making these sounds randomly, infants seemed to be actively engaging in practicing specific sound categories.
“We observed in these studies that infant vocalizations are not produced randomly; they form a pattern, producing three categories of sounds in clusters,” says Dr. Pumpki Lei Su, an assistant professor of speech, language, and hearing from the University of Texas at Dallas, in a media release. “The home recordings we analyzed included times when adults were interacting with their child and when children were on their own, showing that children explore their vocal capabilities with or without language input from an adult.”

Interestingly, the study found that this clustering behavior was present even in the youngest babies, challenging the idea that vocal play only begins around three to four months of age. The researchers observed significant clustering of both squeals and growls in infants as young as zero to two months old.
Another surprising finding was that the amount of clustering didn’t necessarily increase steadily with age. Instead, it varied across different age groups, with some periods showing more clustering than others. For example, squeal clustering seemed to peak around seven to eight months of age.
These findings support the idea that babies are active participants in their own vocal development. Rather than simply making random sounds, infants appear to be exploring and practicing different vocal categories from a very early age. This exploration is thought to be crucial for developing the ability to produce the wide range of sounds needed for speech.
“We tend to think babies are passive recipients of input. And certainly, parents are their best teachers. But at the same time, they’re doing a lot of things on their own,” Su says.
The study’s results have important implications for our understanding of language development. They suggest that the foundation for speech begins much earlier than we might have thought, with babies actively working on their vocal skills from the first months of life. This research also highlights the importance of allowing babies plenty of opportunities for vocal play and exploration.
For parents and caregivers, this study offers a new perspective on those adorable baby sounds. Those squeals, growls, and coos aren’t just cute noises – they’re signs of your baby hard at work, laying the groundwork for future speech. So, the next time you hear your baby making a series of high-pitched squeals or low growls, remember that they’re not just playing around – they’re practicing important skills that will help them become talkers.
Paper Summary
Methodology
The researchers used a clever approach to study baby vocalizations in a natural setting. They gave parents special recording devices called LENA recorders, which are small enough to fit in a baby’s clothing pocket. These devices recorded the baby’s sounds for up to 16 hours a day, once a month, throughout the baby’s first year.
From each day-long recording, the researchers randomly selected 21 five-minute segments to analyze. They trained a team of graduate students to listen to these segments and identify different types of baby sounds, focusing on squeals, growls, and vocants.
To determine if babies were clustering their vocalizations, the researchers used a statistical test called Fisher’s exact test. This test helped them figure out if the babies were making certain types of sounds more often in some time periods than in others, which would indicate clustering.
Key Results
The study found strong evidence of vocal clustering. About 40% of the recordings showed significant clustering of squeals, and about 39% showed clustering of growls. When looking at either squeal or growl clustering, 61% of recordings showed significant clustering.
Looking at individual babies, 87% of infants showed significant clustering of both squeals and growls at some point during their first year. Every single baby in the study showed significant clustering of at least one type of vocalization.
The researchers also looked at how clustering changed with age. They found that clustering occurred at all ages, even in the youngest babies (0-2 months old). Interestingly, the 3-4 month age range, which is often thought of as a key period for vocal play, actually showed the lowest amount of clustering.
Study Limitations
The researchers noted several limitations to their study. First, their method of coding baby sounds was simplified to make it possible to analyze such a large amount of data. This means they might have missed some of the complexity in baby vocalizations.
Second, it can be challenging to consistently identify different types of baby sounds, especially in very young infants. For example, it’s sometimes hard to tell the difference between a growl and a grunt (a sound babies make when they’re moving or straining).
Lastly, while the study included a large number of babies, most of them were young, white females. This limits how well the results might apply to all babies.
Discussion & Takeaways
This study provides strong evidence that babies actively practice making different types of sounds from a very young age. This practice seems to be an important step in developing the ability to produce the varied sounds needed for speech.
The findings challenge some existing ideas about vocal development. For example, they suggest that vocal play starts earlier than previously thought and continues throughout the first year, rather than being concentrated in a specific period.
The research also highlights the importance of babies’ own exploration in learning to talk. While interaction with caregivers is crucial, this study shows that babies do a lot of vocal practice on their own.
For parents and caregivers, these findings underscore the value of giving babies plenty of opportunities for vocal play. Those seemingly random baby noises are actually important practice for future speech.
Funding & Disclosures
This research was funded by grants from the National Institutes of Health, specifically the National Institute on Deafness and Other Communication Disorders and the National Institute of Mental Health. The funders didn’t play a role in designing the study, collecting or analyzing data, deciding to publish, or preparing the manuscript.
One of the authors, D. K. Oller, is an unpaid member of the LENA Scientific advisory board. However, this doesn’t change the authors’ adherence to PLOS ONE’s policies on sharing data and materials.








What if they speak clearly formed words? My 7 day old daughter said “hi”. She also said five more clear words before the age of 3 months. Then more words after 6 months. She was speaking 3 word sentences by 1 year.