This week I’m talking with Dr. Ailís Cournane of The University of Toronto. We are discussing first language acquisition, or FLA. We spoke about the way babies and infants acquire their mother tongue(s), the first language(s) they learn, and how they go about doing that.
Have you ever wondered if your infant understands your baby-talk? Whether you could raise a perfectly trilingual child? Or if it matters that your kid can’t remember that we don’t say “gooses” or “meeses”?”. Understanding how our children acquire language through FLA is the first step in answering some of these questions. Even if you don’t have a child, you were one once, so let’s find out what the kid in us thinks about how we all learned our first language
Like so many aspects of child rearing, having one’s own child tends to make people feel like experts on how babies develop. When it comes to first language acquisition, we can turn to researchers like Doctor Cournane to help understand out the science behind how our babies develop, and what kind of universal truths actually exist. Even if you don’t have a child, you were one once, so let’s find out what the kid in us thinks about how we all learned our first language.Michael Iannozzi : What first got you interested in studying first-language acquisition?
Ailís Cournane : I originally worked on language change, and I continually saw reference to the role of the child learner in driving language change [Language change is the way language evolves over time, which usually means over generations]. Children are thought to reanalyse the language when they learn and build their own [internal and subconscious] mental grammar. However, despite seeing that theory everywhere, and it being widely accepted, no one had explored it sufficiently in how languages change. So I got into child language because I’m interested in change. Development and change have much in common.
MI : How is first-language acquisition defined?
AC : FLA [First-Language Acquisition] describes the process and properties of the language of infants/toddlers/young children as they acquire their mother tongue(s). The child begins with no language (but with the capacity for language!), and with social input from speakers around her, she gradually builds her language(s). I say “language(s)” because many infants are exposed to more than one language, so they acquire these simultaneously.
MI : How is the learning of a first language different from the learning of second language?
AC : There are a few important differences. Firstly, with FLA there is no other language in place. If you are a child acquiring English you are building your first language using just your language capacity and exposure to older speakers. For Second Language Acquisition [SLA] you already have a language in place! So when you learn English as a teenager, for example, you learn it in relation to your mother tongue (say, Mandarin). The English you learn as an L2 competes in some ways with the Mandarin you learned first [An L2 is someone’s 2nd language. The one they acquired second]. Secondly, SLA appears to require more motivation and explicit learning (classes, drills, forcing yourself to practice with native speakers, etc.) than FLA. SLA also seems to show less defined milestones (or stages) than FLA.
MI : Who are the main sources of input for an infant’s first language [L1] acquisition?
AC : The primary caregivers are the main source of input during infancy and early childhood. Very early on, when most infants are relying heavily on their mother, the maternal input is usually strongest. Older siblings also play a role early on in development. Once the child attends daycare or preschool the peer-group begins to play a larger role.
MI : From whom does a child learn their accent?
AC : Well, at first a child – especially an oldest or only child – models their language on their primary caregivers’ language, from who they are hearing most of their language input. However, children very quickly adapt to their peer-group as soon as they enter daycare or elementary school. This is why people with immigrant parents do not share their parents’ accent, but instead sound like their peers. For example, my parents are from Ireland, but my brothers and I grew up in Montreal. We have Canadian English accents with the features expected of Montreal Anglophones. We occasionally have some Irish influence in our language, but we sound much, much, more like our peers than our parents.
MI : Are the parts of a child’s language (accent, grammar, pronunciation, etc.) learned separately/from different sources, or is it all learned concurrently?
AC : Concurrently, however, the focus or main concentration of developmental changes may be in different areas at different times. For example, since words are made up of several sounds the infant needs to begin to crack the sound system of a language before they can really grasp words (let alone complex words or sentences). That said, sounds are contained in words so the child is also necessarily learning about words when they are focusing on sound development. There are very complex interactions at work.
MI : As children learn their first language, they all make some mistakes, what do they tell us about how the children are learning to speak?
AC : I like to call mistakes or errors “divergent” or “creative” analyses, because these analyses are productive and systematic and emerge from aspects of how the child is learning [Meaning the mistakes a child makes like calling geese, “gooses”, or moose “meese”, make sense even though they aren’t correct. In other words, the mistakes are patterned, and can be explained].
They only seem like mistakes when we compare them to adult grammar norms, but they aren’t actually mistakes, they’re (for example) the child uncovering rules of the language and applying those rules (sometimes to exceptions). For example, children sometimesover-apply the regular past tense (-ed) to irregular past tense (e.g., goed for went, or eated for ate). This shows that the child understands how to productively form a past tense in English – that’s a big achievement, and shows pattern recognitions and the ability to generalize a rule.
MI : Are the mistakes a child makes in speaking all the same “kind” of mistake? Is a mispronunciation the same “kind of mistake” as eated or goed?
AC : Not necessarily. Mispronunciation, for example, can have either or both physiological (muscle-control, shape of the developing vocal tract, coordination, etc.) and cognitive (understanding of the sound system of the language, planning, etc.) causes.
Omissions of grammatical words (e.g., saying wan go, leaving out the pronoun “I” and the “to” infinitive marker, I wanna go), and what these mistakes mean, are notoriously debated. Does the child omit them because they’re not salient in the sound signal of the language [Is the child not hearing the other parts?]? Or because they are more grammatically complex and abstract? Or because they can be omitted pragmatically (i.e. they’re not necessary to be understood when speech happens in context, as most daily speech does, especially for a child)? Or some combination of these? [It is pretty clear what “want eat” means coming from an infant at supper, even if grammatically it isn’t “correct”]
MI : Is there an “order” to the way children learn a language? Do they learn certain parts first and last?
AC : Yes. Some of this order is logically determined – sentences are made up of words and words are made up of sounds, so you can’t jump right to learning sentences if you haven’t figured out something about the sound system of the language you are learning. Thus, simplifying somewhat, the infant’s first task is to break into the sound system of the speech around them (or the gestural system of the sign language around them). Part of learning the sound patterns in a language is learning where word boundaries are in the stream of speech. Our speech is a continuous acoustic stream with no boundaries, but our mental grammar knows where to put boundaries [This is often why, when we hear a language we aren’t familiar with, we think they are speaking quickly. Because we can’t hear where words end].
We learned to do this as infants by solving what is called the “segmentation problem”. This problem refers to how the child learns where in the continuous stream one word ends and the next begins. Current research mostly argues that children rely heavily, or solely, on monitoring transitional probabilities between sounds. Sounds that frequently pattern together in the speech stream are thought to pattern together as words. Only by having some understanding of the phonology of the language, e.g. which sounds pattern together and how, can the child progress to associating meanings to words and learning how words can be combined together into complex words and sentences.
MI : Is that order the same across languages? What are the differences in children who learn different first languages?
AC : Yes, as far as we know the order is remarkably similar across diverse languages. The infant, unlike someone trying to learn a second language, has no knowledge of any previous language, so much of what constrains development is determined by how the child learns, and if the child has any learning or language impairments. The child must figure out the sound system, word forms and word-patterns, word meanings, grammatical rules (syntax), etc. The task is broadly the same one, despite the language that is being acquired varying. American Sign Language, despite being in a different mode (gestural-visual, rather than oral-aural), is known to be very similar in development to spoken languages when we consider milestones – babbling, first words, first word combinations (early sentences), overgeneralization of rules, etc. That said, most languages have not been sufficiently studied in development, the focus has been primarily on Western European languages and other prominent/widely-spoke languages (e.g., Japanese, Mandarin).
MI : How do adults/parents/caregivers change their speech when talking with children/infants?
AC : Caregivers often use what is called “Child Directed Speech”, or CDS. This is also called Infant-Directed Speech or Motherese. CDS has distinct phonetic features – it is higher pitch than the speaker’s regular voice, stress patterns are exaggerated, and vowels are held for a longer duration. These phonetic features are perceived as having a “happy affect”. Infants and young children respond preferentially to happy affect.
There is also some evidence that adults simplify their word choices to represent basic-level categories, for example, a mother might call a tiger a “kitty” when talking to her young child. Further, there is also some evidence that adults might simplify the sentences containing words the child is poised to learn. For example, a father might say “You want water?” to his 1-year old rather than, “Do you want a glass of water?”. Adults are thought to subconsciously complexify their child-directed language as the child grows linguistically.
MI : Is this change helpful for the learning children?
AC : It seems to be helpful, but not necessary. There are cross-cultural differences in the ways adults, and particularly caregivers, interact with children. We know that CDS is something that infants and young children respond to, and it may help exaggerate word boundaries and other features of the speech stream and thus help with learning words, but helping is different from being necessary. Most of our research has been conducted on children learning a language in western societies in recent times, so it is fair to say that it this point we know more about this learning situation than all other extant learning situations.
MI : Do different children learn their first language differently?
AC : Yes. There is variation across children, but this needs to be viewed as secondary to very robust trends and similarities. Thus, there is more that is the same about how children learn their first language than there is that differs. With that in mind, let’s talk about what varies.
First, children may have a disorder that affects language development (e.g., Down Syndrome, Autism, Williams Syndrome), or they may have a hearing impairment that impairs their access to spoken language (but not to sign language if they are exposed to a sign language). Then, even among children who are considered typically-developing with respect to language, the time course of development varies. Some children reach milestones much earlier than others, all within the normal range. For example, some children will use their first words by 10 months of age, while others may not use their first words until 24 months old. This is comparable to how some children start walking at 9 months while others start walking at 18 months. Language delay is only diagnosed in children over age 4 because children can vary widely in early development without any cause for alarm.
The child’s personality also affects development. Some children will be quiet and cautious, while others will be quite loquacious. The nature of linguistic interaction with caregivers matters as well. Some studies have shown that socioeconomic status (SES) is a significant factor in vocabulary and syntactic development. Families in the US were followed and their speech analysed; caregivers of lower socioeconomic status were less likely to ask their young children open ended questions (e.g. “What are you drawing?”) and more likely to address young children with yes/no questions (e.g., “Are you drawing?”), and more likely to use prohibitive language (e.g., “Don’t…”) compared to higher SES caregivers. Children from higher SES families on average had larger vocabularies, and reached syntactic development milestones somewhat earlier.
MI : How does a child change the language of their parents?
AC : My research directly addresses this, but it is an area that has previously received a lot of theoretical attention, but very little data-driven research. In other words, many researchers think children play a role in how languages change, but it’s unclear from real-world research whether that is true or not. We know that teenagers are “early adopters”, and most likely to embrace and spread language changes (think about, for example, use of quotative “be like” as in “He was like, ‘thank you’”), however do these new variants in the language emerge from child-language innovations?
I look at this question for modal expressions (words that express possibility: must, can, might, maybe, probably, etc). Children do appear to make analyses that are compatible with how languages change over time. However, compatible is not necessarily causal, so stay tuned!
MI : Does your research affect the way you now speak with children?
AC : I don’t think so, except perhaps that I’m more and more comfortable with lots of different children the more I work with kids. That said, working on child language definitely affects how I listen to children! I love talking to children not just for the content of what they say, but for the linguistic form of what they say.
MI : What is your favourite part of studying the way children learn languages?
AC : The linguistic creativity! They are using their developing language to live their wee lives, and they make many creative sentences when they try to express themselves. The problem itself is fascinating and complex – how does a language-less infant go from that pre-linguistic stage to becoming a fully-linguistic adult? Language is complex and systematic, and it is so easy to take it for granted. When you have to explicitly think about what the little language learner is doing you are regularly struck by how marvellous it is that humans can learn language at all.A great thank you to Ailís Cournane for taking us back in time, and teaching us how we once learned language.
One of the great advantages of studying linguistics, or language more broadly, is that you are constantly amazed at the complexities of language, and how subconsciously and effortlessly you are able to handle that complexity.
The fact that the little child version of me was able to learn “Tractor” as one of my first words, is quite remarkable. Infant Michael was able to find out where the word began and ended (with a “t” sound and an “r” sound), pick out the sounds in between, and attach that collection of sounds to a massive machine my grampa drove through the field with me, and then produce those sounds without being asked when I saw it the next week. Not bad work considering I still struggle after 10 years to use dont correctly in French.
Take care eh,