How do children
learn their first language? Most people don’t remember learning how to speak,
so our answers to these questions are largely from observation. My situation is
slightly different.
While I don’t
remember learning to understand English, I do remember learning to speak it.
You see, I was born with severe speech apraxia, which is basically a fancy way
of saying that my brain didn’t know how to properly move the body parts needed
for speech. When my frustration in my inability to communicate led me to
develop my own simplistic sign language, my parents took my to see a speech
pathologist. That was right around my second birthday.
For the next
three years I learned to talk sound by sound. To me, the process seemed like a
bunch of games. I distinctly remember being given multicolor licorice, with the
colors designating not flavor, but how difficult that given piece was to chew
(an exercise in muscle strengthening). We started with the simple sounds, which
I now can name as vowels, as well as both labial and dental nasals and stops.
Next came the velar and palatal stops as well as voiced labial and dental
fricatives. And finally, I learned the hardest sounds, the voiceless fricatives
and approximants, with the true test being in the [ɹ] sound.
It’s no coincidence
that when you hear a toddler speak that “fish” might become “pish”, or that “three”
is often pronounced “twee.” We learn sounds in order of difficulty. I’ve always
found it rather ironic that my speech teacher was named Ruth; I can’t think of
a harder name to say.
This whole
anecdote has led up to this idea that we might learn to understand and speak with
different trajectories. I believe speech is learned from the bottom up. Babies
typically make sounds even before they can talk. At first they just cry, and
soon after begin to make accidental sounds as they move about. Then comes
babbling: “Babababa.” Until the babbling turns into a collection of different
syllables: “Mababada.” And this is all just practice for putting actual sounds
together to make words, and then those words together to make cohesive thoughts
or sentences. There is an order that we learn sounds, both naturally and when
taught, and young children will simplify words with place assimilation until
they learn the proper sound.
However,
before any cohesive sentences we must first have a level of understanding of a
language. The theory of language acquisition proposed by Noam Chomsky is that
of the Universal Grammar. It is
contains the idea that we all have an idea of innate and biological core grammatical
categories that assist children with language development and adults with
language processing. Under this theory a child would know instinctively how to
combines types of word to share thoughts due to a few rules true for all human
languages, and the only task they have is to simply learn the words of their
language.
However, this is
not the only theory of language acquisition out there and I am not entirely
convinced of its veracity, especially considering different languages have different sentence
structures, phonology, semantics, and more. Another theory, called morphology acquisition, which I’d tend
to agree with, is that we learn to understand a language, we do so from the top
down, by which I mean we learn to identify morphemes within sentences. Babies
hear everyone talking around them in complete sentences, to them just a long, cohesive
string of sounds. With enough repetition, we can break those sentences down
into words, and from there we can further derive meaning from particular
morphemes. These morphemes, I believe, lend comprehension to a language and
form the basis for our speech. Thus, our language learning would be from
pulling a sentence apart into morphemes and sounds, and our speech would arise
out of combining simple sounds to build sentences back up.
To consider: Do
you agree with the Universal Grammar theory, or the morphology acquisition
theory of language acquisition? Is it reasonable to say that we learn to
understand language from the top down but to speak from the bottom up as
described? Do you agree we learn sounds in order by their simplicity, or do you
believe we do so in a different order?
Sources:
It seems to me that perhaps these theories aren't mutually exclusive. Based on the morphology acquisition theory as you've described it, it seems that perhaps as babies hear these long strings of sounds (sentences), and begin to notice patterns in the way certain sounds are combined and the context in which these particular sound combinations are spokeb, they are essentially building an innate understanding of grammatical properties. Noam Chomsky says that the child must be exposed to certain "linguistic data" in order to form an understanding of the common structural basis of all languages, or the Universal Grammar. So perhaps these theories fit together in a sense: the way that this linguistic data (hearing sounds in environmental contexts) is processed in the brain forms the basis of the Universal Grammar for language development.
ReplyDeleteIn response to your question: "do you agree we learn sounds in order by their simplicity, or do you believe we do so in a different order", I definitely agree. It is important to note the difference here between cognitively learning sounds for their meaning and physically learning to produce those sounds. Your experience in speech therapy combined with the fact that babies seem to learn to speak simpler sounds first (bababa, mababa, etc.) seems to suggest that speech production certainly has an order and trajectory, as you've described. But as you mentioned, your language comprehension and understanding of the meaning behind those sounds seemed to develop in a different way than your ability to produce them. This makes me think that perhaps the ability to comprehend language follows a top-down model as you've described, yet speech production is learned from the bottom-up. This is a really cool new way of thinking about this concept, and I think you're definitely on to something!
I agree with you that we learn sounds in order of simplicity, but I don't think we understand them in that order. Babies begin making noise by crying and gurgling, etc. Then, they slowly progress to making what we would call speech sounds. I'm not sure if gurgling is any easier than the [m] sound, but for some reason we make speech sounds later than natural, emotional sounds. This probably arises from when humans could only communicate by grunting or crying or laughing. We intuitively understand what each of these sounds mean, but once we dive into the realm of speech sounds, there meanings are not immediately obvious. Obviously we understand harder-to-produce combinations of sounds before we can produce them. But we might even understand these combinations of sounds - these morphemes - before we understand easier morphemes. I don't think that understanding and producing sounds are linked too closely. As you pointed out, you could understand English before you could speak it. A baby can recognize their mother crying or laughing, but not necessarily when she says [ma]. Arguably, the emotional sounds are more complex (think of the different types of crying), yet we naturally know what they mean.
ReplyDeleteIn response to the above comment I think there is a certain level of mutual exclusivity between the morphology acquisition and universal grammar theory. By this I mean that if one theory is completely correct then they other cannot be. However, I think there is room for elements of each model. It is possible that we know certain things innately (and this innate knowledge is what allows babies to say truncated by meaning phrases when they are first learning to speak), and that we also learn some elements of language from the top down. As for your question on sound acquisition, I agree with your blog post that we learn simpler sounds first. However, I would like to add the caveat that we learn simpler sounds within our own languages first (so if I am an Arabic speaker I may learn sounds that are more complex than english sounds but that do not occur in Arabic).
ReplyDeleteWhile I have no linguistic evidence to back this up my intuition is that, as many people have been saying, we seem to learn in both a bottom up and top down fashion. I again argue that we do this simultaneously. The idea that we can learn both way seems to be the one that makes the most sense, seeing as babies learn how to form increasingly complex sounds, but also very quickly also have an intentionality behind their words, which supports the idea that they are also learning about the ideas that are behind words. This idea obviously supports a top down understanding with a simultaneous bottom up speaking prowess.
ReplyDeleteAnd I think it is important to consider the fact that possibly the while we do seem to learn certain sounds first because they are easier, it might also be a matter of prevalence. Again, without linguistic data it seems to me that vowel sounds are some of the easier sounds to make, but they also occur more frequently than specific consonant sounds (seeing as every word needs a vowel and there are not fewer vowel sounds). Which maybe forces the question do we learn prevalent sounds first, and those become ones we think of as easier because we have had more time and experience with them?
I think the morphology acquisition theory has particular merit. This theory is aligned with information processing theories and constructivists who believe that children behaviors are a result of behavioral input. Studies, such as the infamous "gavagai" experiment have shown that young babies are capable of detecting word-structures from a string of morpheme-sounds. That being said, I think there is also strong evidence for a modularized view of language acquisition, which would support Chomsky's theory. It may be the case that a domain-specific learning mechanism is necessary for language acquisition, which, as others have pointed out, is not mutually exclusive with the morphological acquisition theory.
ReplyDeleteI am really interested in the speech therapy you went through! When I read your post, I imagined the scene where Eliza Doolittle is learning the voiced H by using a candle to check if the H was voiced. (See image: http://imgur.com/YOK8Ei1) Personally, I had to go through speech therapy after an injury, and it was very different from the speech therapy from the movie and what you described—I am sure my injury was very different. My speech therapy involved practicing pronunciation of words and learning how to properly use the left side muscles of my mouth. I definitely think that speech is acquired from the bottom up, as our brains make neural connections that form the building blocks of words, phrases, and sentences. However, I think that when it comes to understanding language, we also learn from the bottom up in physiologically terms, mostly because neurologically, it is hard to see how we can learn how to understand language without repeated creation and use of neural connections. In a linguistics sense, I definitely think we learn top down, taking sentences and breaking them down into parts and creating schema, which we then place in those neural connections.
ReplyDeleteFascinating, Griffin, absolutely fascinating! I'm curious to know if, from time to time, you experience vestiges of speech apraxia say when you become very nervous or in exceptionally cold temperatures. What an amazing thing to overcome. Your blog reminds me of a conversation I had with a friend just yesterday. He mentioned how, during a recent trip to Brazil, he had little difficulty understand people's Portuguese though he himself speaks very elementary Spanish. We've all heard this before; adults claiming with certainty that they understand a foreign language spoken to them even if they themselves cannot speak it. In my opinion, this is the one traits of our child brains that adults never relinquish and it strikes at the heart of morphology acquisition (top down) theory. As adults we tend to over analyze that which we know. Once we learn a language (from the bottom up) we lose our bottom up capability to over analysis. But in our brains' never ending quest to understand that which we desperately wish to comprehend, we seem never to let go of our top down skills. I never quite put that together until just now. Thanks :)
ReplyDeleteI'm not on the same page with you when you describe the morphology acquisition theory as a "top down" process. If anything, shouldn't it be a bottom-up process? Toddlers acquire and understand each morpheme before being able to form coherent sentences. For example, by no means the sole way to acquire morpheme, reading words off flashcards or tracing word by word in a book are common ways that kids are taught vocabulary and sentence construction. Going from morpheme to sentences can be summarized more as a bottom up process.
ReplyDeleteI would definitely agree that I side more with the morphology acquisition theory. However, I would say that there is some universal "grammar" we all share to some extent. For example, the idea of nouns to describe concrete object or more broadly, "things", or verbs to describe actions is fundamental to human language. Kids tend to point at things and utter nouns to describe them or use verbs to describe movements.
Thanks for sharing your experience and your thought, Griffin!
ReplyDeleteI also think the process of acquiring language can’t be simply sorted out—it’s neither a bottom to top nor a top to bottom process. Intuitively, I think children learn both the words and the structure at the same time. Of course, there are ‘priorities’: children might learn some simple words or structure first, but the order of learning, I think, is not fixed. It might also differ from children to children and from language to language. For example, children might understand ‘nouns to describe concrete object’ (from Joseph) first, because moms sometimes teach infants by pointing at objects and repeating words. Children might also learn some simple grammar on an early stage, for example, they need to understand the questions “What is this?”, “What’s this?”… Lastly, as you pointed out, understanding and speaking can probably be separate. (And may also occur at the same time).
Between the theory of universal grammar and morphology acquisition, I definitely agree with morphology acquisition. I don’t believe that we all have an innate ability to understand the basic grammatical rules of a language. It’s definitely learned through experiences such has being around others and watching television. I have a brother and sister who are six and five years old respectively, so I remember pretty clearly how they learned English. As for speaking, they definitely started with simple sounds and moved up from there. My brother first called me “maga” before he finally learned to say “Marco.” As for understanding, I believe they first started with understanding certain nouns and verbs, like “tree” or “run.” Adjectives were also learned fairly quickly. The rest of the language such as articles and prepositional phrases and other parts were learned through experience with others. Of course, at their age, their language skills still aren’t perfect. My sister still usually says the past tense of buy as “buyed.” As they get older, their vocabulary will continue to expand and they will learn how to properly use every aspect of the English language.
ReplyDeleteIf my understanding is correct, I think I agree more with the idea of morphology acquisition. We've all heard young children attempt to make sentences that make very little sense grammatically, but, based on the words they're uttering, we're able to determine what this child actually means.
ReplyDeleteThe child knows the important words to say, but they don't necessarily know the order in which to say them. This is something that they learn later on, and I wouldn't say it comes naturally at all. You can look at the way we order adjectives - big, red house vs. red, big house. It comes intuitively to me now, but a child just learning how to use adjectives might string them together in a way that doesn't seem natural. I would still understand their meaning, but it wouldn't seem to fit together in a grammatically satisfying way.
The issue is very interesting, and I believe it's definitely something that any linguist with a serious interest in understanding how language works. But to be fair, Chomsky's Universal Grammar theory does not in any way say that "the only task they have is to simply learn the words of their language.” Children also need to learn grammatical rules by exposing to outside input. I agree with the Universal Grammar theory exactly because “ different languages have different sentence structures, phonology, semantics,” as the author pointed out, but more importantly, in these different systems of phonological and syntactic systems we find striking similarities that would be otherwise inexplicable.
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteThis topic of language acquisition has been used by many computer scientists to model the next artificial intelligence. In trying to create a machine that would imitate the adult human mind, Alan Turing investigated the process that brought it to the state that it is in. He noticed three central components: the initial state of the mind at birth, the education it is subjected to, and non-educational experiences. Turing believed that we could use the human mind’s model for learning language on a programme. In trying to simulate the adult mind, one should create a programme that would simulate a child’s and then apply the appropriate course of education on it. Central to this argument was the common belief that an infant’s mind was something like a notebook as one buys it from the stationers, with very little underlying mechanism and many blank sheets.
ReplyDeleteI do not believe Turing was right. In fact, a lot of new data shows that infants use computational strategies to detect patterns in language input, which leads to the discovery of morphemes. Does this make Turing’s hypothesis, modeling a computer after a child’s mind, unusable? The metaphor remains interesting, but I wonder how useful it is now.
I agree with your assertion that language comprehension occurs by breaking down sentences into smaller morphemes. However, while I don’t think it is necessarily wrong that we learn to speak first with simplicity and then upward, I am not sure if there can be such a huge divide in these two processes. Hypothetically, the smaller morphemes we break down sentences into would be our first words or sounds that we build up in complexity. Yet, if that is true, are we really comprehending in sentences as much as the morphology acquisition claim states. I feel that one might need to further analyze both the neural and speech muscles used in both language understanding and speaking in order to make a more definitive speculation. Great blog with a lot of food for thought!
ReplyDelete