It's one of the most common assumptions that people make
about baby sign language -- that babies can sign before they can speak.
But is it true?
The answer is: It might be true, but we don't know for sure,
and skepticism is warranted.
Studies confirm that babies start practicing spoken language at an early age. And currently there is no compelling evidence that true signing emerges before genuine speech.
So let's take a closer look, starting with what we know
about the development of language.
1. By the age of 6 months, many babies have begun repeating our speech sounds. They may also understand a variety of everyday words.
Babies are born with a package of powerful mechanisms for
learning to speak. It's a basic human trait, the hallmark of our species, and
experiments show that babies begin learning about language while they are still
in the womb.
Infants can hear their mothers' voices during the latter
stages of pregnancy, and, after birth, they show the ability to distinguish
between a voice speaking their mother's native language, and a voice speaking a
different, unfamiliar language.
From such beginnings, babies begin to analyze what we say to
them. They begin to perceive that our streams of speech are divided up into
distinct words. Eventually, by observing and interacting with us, they work out
what specific words mean.
So when do babies transition from speaking mere syllables, to speaking genuine words?
That's tricky to document, because a baby's early attempts
are going to be inexact.
Babies aren't going to produce words with the magnificent elocution
of James Earl Jones. They are still mastering the muscles required to speak.
Their first attempts are going to be approximations. At what point do we count
an imperfect utterance as a real word?
The problem doesn't arise when the word itself is very easy
for babies to say. And here, enter "mama": the word for
"mother" in languages around the world.
The word reappears in languages as different as Swahili and Chinese, most likely because it's incredible easy for the human infant to master. "Ma" is often the first syllable that a baby repeats, so families in a variety of places and cultures have used it to signify "mother."
Similarly, the baby-friendly syllables "da" and "pa" have often been used to signify "father," though there are cultures where "papa" means mother and "mama" means father (Hendery and McConvell 2013).
But most vocabulary isn't as easy for babies to pronounce, and that's where we can run into grey areas.
For example, is "baba" a good enough approximation of the word "bottle"? Is that a true word?
If you're a parent who notices that your baby says "baba" when she wants her bottle, then you'll want to respond accordingly. It would be confusing if you pretended that you didn't understand her. Babies need confirmation when they get things right. It's how they learn.
But sometimes "baba" is just a random bit of babbling, and researchers who study the emergence of speech must find ways to sift those instances out. They need to establish a set of objective criteria for recognizing an utterance as a spoken word.
2. Can babies sign before they can speak? To answer this question, we need to apply similar standards about what counts as a spoken word or sign.
When babies begin to use signs, are they as competent as
adults? Far from it. Signing, like speaking, depends on important motor skills.
A baby's first signs are going to be approximations, just as a baby's first
spoken words are approximations. Children are still mastering the muscles
required to gesture.
So if we want to compare the timing of signing and speaking,
we need to make sure we're applying similar standards. Whatever allowances we
make for a child approximating a spoken word, we should also make for a child
approximating a sign. And vice versa.
If you see a baby approximate the sign for "drink,"
you may be impressed. But then you should also be impressed when that same baby approximates the spoken word for
Here's a video clip where a baby is doing both.
It's a wonderful example of a baby's early attempts at
The mother says the baby is signing "bottle,"
though of course the gesture is only approximate. To me, it looks like a rough
match for the ASL sign for "drink." Either way, the baby appears to
be producing a rudimentary sign.
But notice that the baby is also saying, "ba-ba,"
which seems to be an approximation of the spoken word for "bottle."
So this baby is showing off two language skills at once -- communicating
simultaneously with speech and sign.
Yet what's the title of this video clip? The title that it
was uploaded with on youtube?
language…asking for a bottle."
The signing is highlighted in the title, while the speech
isn't mentioned at all. Why?
Maybe the person who uploaded the video was simply
more interested in the development of signing.
But clearly, we're kidding ourselves if we watch a video
clip like this and regard it as evidence that babies can sign before they can speak. We can't accept approximate attempts to sign as language, and yet
dismiss similarly approximate attempts to speak as pre-linguistic babbling.
If we did that, it would be easy to support the claim that
signing comes first. But the support would
be phony -- an artifact of the way we measure things.
3. Based on studies conducted to date, we can't tell if signing
emerges any earlier than speech.
To see what I mean, consider the best available research on the developing of signing: Studies that track babies who learn a sign language because they or they parents are deaf.
When do these babies produce their first, signed words?
In one study, researchers followed the development of 11 babies raised by deaf, signing parents. On average, these children produced their first recognizable signs at around 8.5 months postpartum (Bonvillian et la 1983).
Compare that with a survey about spoken language acquisition among babies who weren't deaf or raised by deaf parents. Researchers asked more than 1500 parents to recollect the timing of their babies' first words. The results? Most parents said their babies had spoken their first words by the age of 10 months.
On the face of it, these studies might appear to confirm that babies learn to sign before they learn to speak.
But notice how small the signing study was. We can't draw conclusions about the relative timing of signing, not on the basis of a single study involving just 11 babies.
Maybe, by chance, the 11 babies selected for this study were faster-than-average learners. We need a larger sample to be sure the results don't reflect chance factors and statistical error.
Moreover, the studies used different methods, making direct comparisons difficult. One study collected data on a monthly basis. The other asked parents to recollect what had happened months before. And the researchers used different criteria for counting a baby's communicative efforts as a word.
We see similar problems when trying to evaluate research about the number of words that babies learn over time.
For instance, Diane Anderson and Judy Reilly recruited families where both parents and infants were deaf, and the babies were exposed to signing from birth.
The parents were asked to fill out monthly questionnaires for five months total. What signs have you seen your baby produce this month?
Then the researchers calculated the average number of signs that babies produced during a given age range.
Among a group of 5 babies between the ages of 15 and 16 months, the average (median) number of signs produced was 64, with individuals ranging between 30 and 102 signs (Anderson and Reilly 2002, p.89).
By contrast, in a study of speech development, researchers assessed a group of 16-month-old babies -- 64 infants total -- and found that the average (median) number of words spoken was about 40. Individual babies ranged between fewer than 10 words and more than 175 (Fenson et al 1994, pp. 37-39).
Once again, a mere glance at the average number of words produced in each study -- 64 words versus 40 -- makes it appear that signing has the advantage.
But once again, we can't directly compare these studies.
The average number of signs derives from a very small sample, and the range of individual variation is large in both studies. So it's impossible to know if there is a real, underlying difference between the groups. It might just be a statistical blip.
As language researcher Richard Meier (2016) notes, it's just too soon to say whether or not babies can learn to sign before they can learn to speak. We need more research to decide the question (Meier 2016).
Why are we so eager to believe that babies can sign before they can speak?
This is a fascinating question in its own right, and I don't
know the answer. But I have my suspicions.
There is the hype factor, if course. It's exciting to hear
bold claims about baby signing, and they might make us forget the obvious
-- that lots of babies perform equally
impressive feats with spoken language.
But it seems to me there is more behind it. I get the
impression that people think signing is intrinsically easier than speaking,
that producing symbolic gestures is less demanding than producing symbolic
Why would this be true?
People seem to assume that speech is harder. That it's
harder to figure out, or harder to physically reproduce. Or both. But we really have no basis for assuming this.
We might if we were comparing the task of learning speech
with the task of learning a small set of pantomime gestures -- gestures that
"act out" the thing being referenced.
It's relatively easy to figure out what pantomime gestures
mean. That's why people who lack a common language often resort to pantomime.
But baby sign language programs aren't typically based on
pantomime. Instead, they involve teaching babies signs borrowed from real sign
languages, like ASL. And these signs are like the words we use in English and
other spoken languages:
A few signs offer clues about their meaning, but most
do not. As I explain elsewhere, the link between most signs and their referents is arbitrary one.
Then there's the question of physical dexterity. Is it
physically easier to make the signs of ASL? Easier than it is to form the
spoken words of English?
The answer might vary depending on which signs and spoken
words we compare. But overall, we have no compelling scientific evidence to
prove that signing is any less challenging for babies. The dexterity required
to form signs is substantial, so much so that babies don't typically manage to
reproduce signs in their correct form.
Finally, I wonder if some of people are influenced by the example of ape language studies.
I know I thought of it when I first heard of baby sign language.
Admittedly, it was a professional reflex on my part, because of my background in behavioral ecology, primates, and the evolution of intelligence. But I suspect many of my readers have heard about the signing apes.
Years ago, researchers concluded that nonhuman primates lacked the necessary anatomy to mimic human speech. But they wanted to see if apes had the ability to learn language. So they tried teaching them to communicate with hand gestures and visual symbols, and they were successful.
A number of individuals -- including the chimpanzees Washoe and Sarah, and the bonobo, Kanzi -- have shown they can communicate ideas and answer questions by using non-vocal symbols.
There is a certain parallel with babies. The human infant's vocal tract is also shaped differently, limiting the range of speech sounds that a baby can make. Yet we know they understand a great deal. So perhaps babies, like nonhuman apes, are anatomically incapable of speech, yet capable of learning symbolic gestures.
It's an appealing analogy for those of us interested in zoology and human evolution. But with greater scrutiny, it falls apart.
First, the premise about vocal tracts is wrong: You don't need an adult human's vocal tract to speak.
For example, research suggests that it would be possible to speak intelligible English (and other languages) with the vocal anatomy of a macaque monkey (Fitch et al 2016). You can listen to a reconstruction here.
So the reason these creatures don't speak isn't that their vocal tracts fail them. Instead, it's likely a failure of the brain: Nonhuman primates lack the precise neural control needed to produce speech (Fitch et al 2016).
By contrast, human babies show a lot of control quite early in life. Between 12 to 20 weeks postpartum, human babies are already capable of matching the vowels they hear adults make (Kuhl and Meltzoff 1996). And by the time canonical babbling emerges, babies have mastered the articulation of many common syllables.
Second, it's wrong to assume that it's easy for nonhuman apes (or human babies) to mimic adult human hand gestures. It isn't.
The human hand is a marvelous, species-specific adaptation. As noted, many linguistic hand gestures require substantial dexterity and control. Such skills aren't part of a nonhuman ape's repertoire, so researchers shifted long ago from trying to teach apes symbolic hand gestures. Instead, they focus on systems that use visual symbols on a board.
So if we really wanted to apply an ape language analogy, we'd be teaching our babies to communicate by pointing to pictograms.
But pointing is useful for much more than this. When babies learn to point, they experience a major breakthrough in communication. They can direct our attention to things that capture their interest, and, if we're responsive, it's a golden learning opportunity.
When your baby points to a horse, you can immediately provide the label, saying "that's a horse." Experiments suggest that babies quickly learn new labels this way.
And that brings us to the really important lesson from research on language acquisition: Children learn language best when we talk with them, one-on-one, and use nonverbal cues -- like pantomime and pointing -- to get our meaning across.
So whether your baby learns to speak first or sign first, you can give your baby's language development a boost by enriching your talk with lively, easy-to-parse gestures.
More information about language and learning
To read more about the remarkable importance of gesture for learning, see this evidence-based review. There I talk about the role of gesture not just in learning language, but also in learning mathematics and physics.
And for more about baby signing, be sure to read this article, where I tackle several questionable claims made about baby sign language programs, and share the fascinating research on "transparent" parental communication.
Finally, for other articles about the emergence of language, see these: